def predcit(self, pointclouds): t0 = time.time() #滤处雷达附近的噪点 mask = np.sqrt(np.square(pointclouds[:, :2]).sum(axis=1)) > 5 pointclouds = pointclouds[mask] ret = self.voxel_generator.generate(pointclouds, max_voxels=12000) voxels = ret['voxels'] coords = ret['coordinates'] num_points = ret['num_points_per_voxel'] # print('befor filter voxels shape is ',voxels.shape) if self.bg_filter is not None: voxels_mask = self.bg_filter.filter_bg(voxels, num_points, coords) voxels = voxels[voxels_mask] coords = coords[voxels_mask] num_points = num_points[voxels_mask] # print("after filter voxels shape is ", voxels.shape) voxels, num_points, coords, voxel_mask = voxel_padding( voxels, num_points, coords, max_voxel_num=self.max_voxel_num) # print('after padding voxels shape is ', voxels.shape) # mask_points = np.logical_not(np.all(voxels.reshape(-1, 4) == 0, axis=1)) # points = voxels.reshape(-1, 4)[mask_points] example = { "anchors": self.anchors, "voxels": voxels[np.newaxis, ...], "num_points": num_points[np.newaxis, ...], "coordinates": coords[np.newaxis, ...], 'voxel_mask': voxel_mask[np.newaxis, ...], "metadata": [{ "image_idx": '000000' }] } if self.anchors_area >= 0: # 计算每个grid map坐标位置是否有pillars(非空) dense_voxel_map = box_np_ops.sparse_sum_for_anchors_mask( coords, tuple(self.grid_size[::-1][1:])) dense_voxel_map = dense_voxel_map.cumsum(0) dense_voxel_map = dense_voxel_map.cumsum(1) # 计算每个anchor_bev占有的非空的pillars anchors_area = box_np_ops.fused_get_anchors_area( dense_voxel_map, self.anchors_bv, self.voxel_size, self.pc_range, self.grid_size) anchors_mask = anchors_area >= self.anchors_area if anchors_mask.sum() < 1: anchors_mask = np.zeros(anchors_area.shape[0], dtype=np.bool) print("anchors_mask is zero") example['anchors_mask'] = anchors_mask example_list = example_to_tensorlist(example, device=self.device, float_type=self.float_dtype) with torch.no_grad(): boxes, scores = self.net(*example_list)[0][:2] boxes = boxes.detach().cpu().numpy() scores = scores.detach().cpu().numpy() print("current frame process time is {:.3f}ms".format( (time.time() - t0) * 1000)) return boxes, scores
def generate_example(net, model_cfg, points, device): target_assigner = net.target_assigner voxel_generator = net.voxel_generator grid_size = voxel_generator.grid_size feature_map_size = grid_size[:2] // config_tool.get_downsample_factor( model_cfg) feature_map_size = [*feature_map_size, 1][::-1] anchors_np = target_assigner.generate_anchors(feature_map_size)["anchors"] anchors = torch.tensor(anchors_np, dtype=torch.float32, device=device) anchors = anchors.view(1, -1, 7) anchors_bv = box_np_ops.rbbox2d_to_near_bbox(anchors_np[:, [0, 1, 3, 4, 6]]) res = voxel_generator.generate(points, max_voxels=12000) voxels = res["voxels"] coordinates = res["coordinates"] num_points = res["num_points_per_voxel"] # add batch idx to coords coords = np.pad(coordinates, ((0, 0), (1, 0)), mode='constant', constant_values=0) voxels = torch.tensor(voxels, dtype=torch.float32, device=device) coords = torch.tensor(coords, dtype=torch.float32, device=device) num_points = torch.tensor(num_points, dtype=torch.float32, device=device) # generate anchor mask # slow with high resolution. recommend disable this forever. voxel_size = voxel_generator.voxel_size pc_range = voxel_generator.point_cloud_range coors = coordinates dense_voxel_map = box_np_ops.sparse_sum_for_anchors_mask( coors, tuple(grid_size[::-1][1:])) dense_voxel_map = dense_voxel_map.cumsum(0) dense_voxel_map = dense_voxel_map.cumsum(1) anchors_area = box_np_ops.fused_get_anchors_area(dense_voxel_map, anchors_bv, voxel_size, pc_range, grid_size) anchors_mask = anchors_area > 1 anchors_mask = torch.tensor(anchors_mask, dtype=torch.uint8, device=device) # example['anchors_mask'] = anchors_mask.astype(np.uint8) example = { "anchors": anchors, "voxels": voxels, "num_points": num_points, "coordinates": coords, 'anchors_mask': anchors_mask, } return example
def generate_anchor_mask(self, anchors, anchor_area_threshold=-1): anchors_bv = box_np_ops.rbbox2d_to_near_bbox(anchors[:, [0, 1, 3, 4, 6]]) anchors_mask = None if anchor_area_threshold >= 0: # slow with high resolution. recommend disable this forever. coors = coordinates dense_voxel_map = box_np_ops.sparse_sum_for_anchors_mask( coors, tuple(grid_size[::-1][1:])) dense_voxel_map = dense_voxel_map.cumsum(0) dense_voxel_map = dense_voxel_map.cumsum(1) anchors_area = box_np_ops.fused_get_anchors_area( dense_voxel_map, anchors_bv, voxel_size, pc_range, grid_size) anchors_mask = anchors_area > anchor_area_threshold # example['anchors_mask'] = anchors_mask.astype(np.uint8) return anchors_mask
def __getitem__(self, idx): pc_file = self.pc_files[idx] points = self.read_bin_points(pc_file) metrics = {} t1 = time.time() # [0, -40, -3, 70.4, 40, 1] voxel_size = self.voxel_generator.voxel_size pc_range = self.voxel_generator.point_cloud_range grid_size = self.voxel_generator.grid_size # [352, 400] res = self.voxel_generator.generate(points, self.max_voxels) voxels = res["voxels"] coordinates = res["coordinates"] num_points = res["num_points_per_voxel"] num_voxels = np.array([voxels.shape[0]], dtype=np.int64) metrics["voxel_gene_time"] = time.time() - t1 anchors = self.anchor_cache["anchors"] anchors_bv = self.anchor_cache["anchors_bv"] # anchors_dict = self.anchor_cache["anchors_dict"] # matched_thresholds = self.anchor_cache["matched_thresholds"] # unmatched_thresholds = self.anchor_cache["unmatched_thresholds"] example = { 'pc_file': pc_file, 'voxels': voxels, 'num_points': num_points, 'coordinates': coordinates, "num_voxels": num_voxels, "metrics": metrics, "anchors": anchors } if self.anchor_area_threshold >= 0: # slow with high resolution. recommend disable this forever. coors = coordinates dense_voxel_map = box_np_ops.sparse_sum_for_anchors_mask( coors, tuple(grid_size[::-1][1:])) dense_voxel_map = dense_voxel_map.cumsum(0) dense_voxel_map = dense_voxel_map.cumsum(1) anchors_area = box_np_ops.fused_get_anchors_area( dense_voxel_map, anchors_bv, voxel_size, pc_range, grid_size) anchors_mask = anchors_area > self.anchor_area_threshold # example['anchors_mask'] = anchors_mask.astype(np.uint8) example['anchors_mask'] = anchors_mask return example
def prep_pointcloud(input_dict, root_path, voxel_generator, target_assigner, db_sampler=None, max_voxels=20000, remove_outside_points=False, training=True, create_targets=True, shuffle_points=False, remove_unknown=False, gt_rotation_noise=(-np.pi / 3, np.pi / 3), gt_loc_noise_std=(1.0, 1.0, 1.0), global_rotation_noise=(-np.pi / 4, np.pi / 4), global_scaling_noise=(0.95, 1.05), global_random_rot_range=(0.78, 2.35), global_translate_noise_std=(0, 0, 0), num_point_features=4, anchor_area_threshold=1, gt_points_drop=0.0, gt_drop_max_keep=10, remove_points_after_sample=True, anchor_cache=None, remove_environment=False, random_crop=False, reference_detections=None, out_size_factor=2, use_group_id=False, multi_gpu=False, min_points_in_gt=-1, random_flip_x=True, random_flip_y=True, sample_importance=1.0, out_dtype=np.float32): """convert point cloud to voxels, create targets if ground truths exists. input_dict format: dataset.get_sensor_data format """ t = time.time() class_names = target_assigner.classes points = input_dict["lidar"]["points"] if training: anno_dict = input_dict["lidar"]["annotations"] gt_dict = { "gt_boxes": anno_dict["boxes"], "gt_names": anno_dict["names"], "gt_importance": np.ones([anno_dict["boxes"].shape[0]], dtype=anno_dict["boxes"].dtype), } if "difficulty" not in anno_dict: difficulty = np.zeros([anno_dict["boxes"].shape[0]], dtype=np.int32) gt_dict["difficulty"] = difficulty else: gt_dict["difficulty"] = anno_dict["difficulty"] if use_group_id and "group_ids" in anno_dict: group_ids = anno_dict["group_ids"] gt_dict["group_ids"] = group_ids calib = None if "calib" in input_dict: calib = input_dict["calib"] if reference_detections is not None: assert calib is not None and "image" in input_dict C, R, T = box_np_ops.projection_matrix_to_CRT_kitti(P2) frustums = box_np_ops.get_frustum_v2(reference_detections, C) frustums -= T frustums = np.einsum('ij, akj->aki', np.linalg.inv(R), frustums) frustums = box_np_ops.camera_to_lidar(frustums, rect, Trv2c) surfaces = box_np_ops.corner_to_surfaces_3d_jit(frustums) masks = points_in_convex_polygon_3d_jit(points, surfaces) points = points[masks.any(-1)] if remove_outside_points: assert calib is not None image_shape = input_dict["image"]["image_shape"] points = box_np_ops.remove_outside_points( points, calib["rect"], calib["Trv2c"], calib["P2"], image_shape) if remove_environment is True and training: selected = kitti.keep_arrays_by_name(gt_names, target_assigner.classes) _dict_select(gt_dict, selected) masks = box_np_ops.points_in_rbbox(points, gt_dict["gt_boxes"]) points = points[masks.any(-1)] metrics = {} if training: """ boxes_lidar = gt_dict["gt_boxes"] bev_map = simplevis.nuscene_vis(points, boxes_lidar) cv2.imshow('pre-noise', bev_map) """ selected = kitti.drop_arrays_by_name(gt_dict["gt_names"], ["DontCare"]) _dict_select(gt_dict, selected) if remove_unknown: remove_mask = gt_dict["difficulty"] == -1 """ gt_boxes_remove = gt_boxes[remove_mask] gt_boxes_remove[:, 3:6] += 0.25 points = prep.remove_points_in_boxes(points, gt_boxes_remove) """ keep_mask = np.logical_not(remove_mask) _dict_select(gt_dict, keep_mask) gt_dict.pop("difficulty") if min_points_in_gt > 0: # points_count_rbbox takes 10ms with 10 sweeps nuscenes data point_counts = box_np_ops.points_count_rbbox(points, gt_dict["gt_boxes"]) mask = point_counts >= min_points_in_gt _dict_select(gt_dict, mask) gt_boxes_mask = np.array( [n in class_names for n in gt_dict["gt_names"]], dtype=np.bool_) if db_sampler is not None: group_ids = None if "group_ids" in gt_dict: group_ids = gt_dict["group_ids"] sampled_dict = db_sampler.sample_all( root_path, gt_dict["gt_boxes"], gt_dict["gt_names"], num_point_features, random_crop, gt_group_ids=group_ids, calib=calib) if sampled_dict is not None: sampled_gt_names = sampled_dict["gt_names"] sampled_gt_boxes = sampled_dict["gt_boxes"] sampled_points = sampled_dict["points"] sampled_gt_masks = sampled_dict["gt_masks"] gt_dict["gt_names"] = np.concatenate( [gt_dict["gt_names"], sampled_gt_names], axis=0) gt_dict["gt_boxes"] = np.concatenate( [gt_dict["gt_boxes"], sampled_gt_boxes]) gt_boxes_mask = np.concatenate( [gt_boxes_mask, sampled_gt_masks], axis=0) sampled_gt_importance = np.full([sampled_gt_boxes.shape[0]], sample_importance, dtype=sampled_gt_boxes.dtype) gt_dict["gt_importance"] = np.concatenate( [gt_dict["gt_importance"], sampled_gt_importance]) if group_ids is not None: sampled_group_ids = sampled_dict["group_ids"] gt_dict["group_ids"] = np.concatenate( [gt_dict["group_ids"], sampled_group_ids]) if remove_points_after_sample: masks = box_np_ops.points_in_rbbox(points, sampled_gt_boxes) points = points[np.logical_not(masks.any(-1))] points = np.concatenate([sampled_points, points], axis=0) pc_range = voxel_generator.point_cloud_range group_ids = None if "group_ids" in gt_dict: group_ids = gt_dict["group_ids"] prep.noise_per_object_v3_( gt_dict["gt_boxes"], points, gt_boxes_mask, rotation_perturb=gt_rotation_noise, center_noise_std=gt_loc_noise_std, global_random_rot_range=global_random_rot_range, group_ids=group_ids, num_try=100) # should remove unrelated objects after noise per object # for k, v in gt_dict.items(): # print(k, v.shape) _dict_select(gt_dict, gt_boxes_mask) gt_classes = np.array( [class_names.index(n) + 1 for n in gt_dict["gt_names"]], dtype=np.int32) gt_dict["gt_classes"] = gt_classes gt_dict["gt_boxes"], points = prep.random_flip(gt_dict["gt_boxes"], points, 0.5, random_flip_x, random_flip_y) gt_dict["gt_boxes"], points = prep.global_rotation_v2( gt_dict["gt_boxes"], points, *global_rotation_noise) gt_dict["gt_boxes"], points = prep.global_scaling_v2( gt_dict["gt_boxes"], points, *global_scaling_noise) prep.global_translate_(gt_dict["gt_boxes"], points, global_translate_noise_std) bv_range = voxel_generator.point_cloud_range[[0, 1, 3, 4]] mask = prep.filter_gt_box_outside_range_by_center(gt_dict["gt_boxes"], bv_range) _dict_select(gt_dict, mask) # limit rad to [-pi, pi] gt_dict["gt_boxes"][:, 6] = box_np_ops.limit_period( gt_dict["gt_boxes"][:, 6], offset=0.5, period=2 * np.pi) # boxes_lidar = gt_dict["gt_boxes"] # bev_map = simplevis.nuscene_vis(points, boxes_lidar) # cv2.imshow('post-noise', bev_map) # cv2.waitKey(0) if shuffle_points: # shuffle is a little slow. np.random.shuffle(points) # [0, -40, -3, 70.4, 40, 1] voxel_size = voxel_generator.voxel_size pc_range = voxel_generator.point_cloud_range grid_size = voxel_generator.grid_size # [352, 400] t1 = time.time() if not multi_gpu: res = voxel_generator.generate( points, max_voxels) voxels = res["voxels"] coordinates = res["coordinates"] num_points = res["num_points_per_voxel"] num_voxels = np.array([voxels.shape[0]], dtype=np.int64) else: res = voxel_generator.generate_multi_gpu( points, max_voxels) voxels = res["voxels"] coordinates = res["coordinates"] num_points = res["num_points_per_voxel"] num_voxels = np.array([res["voxel_num"]], dtype=np.int64) metrics["voxel_gene_time"] = time.time() - t1 example = { 'voxels': voxels, 'num_points': num_points, 'coordinates': coordinates, "num_voxels": num_voxels, "metrics": metrics, } if calib is not None: example["calib"] = calib feature_map_size = grid_size[:2] // out_size_factor feature_map_size = [*feature_map_size, 1][::-1] if anchor_cache is not None: anchors = anchor_cache["anchors"] anchors_bv = anchor_cache["anchors_bv"] anchors_dict = anchor_cache["anchors_dict"] matched_thresholds = anchor_cache["matched_thresholds"] unmatched_thresholds = anchor_cache["unmatched_thresholds"] else: ret = target_assigner.generate_anchors(feature_map_size) anchors = ret["anchors"] anchors = anchors.reshape([-1, target_assigner.box_ndim]) anchors_dict = target_assigner.generate_anchors_dict(feature_map_size) anchors_bv = box_np_ops.rbbox2d_to_near_bbox( anchors[:, [0, 1, 3, 4, 6]]) matched_thresholds = ret["matched_thresholds"] unmatched_thresholds = ret["unmatched_thresholds"] example["anchors"] = anchors anchors_mask = None if anchor_area_threshold >= 0: # slow with high resolution. recommend disable this forever. coors = coordinates dense_voxel_map = box_np_ops.sparse_sum_for_anchors_mask( coors, tuple(grid_size[::-1][1:])) dense_voxel_map = dense_voxel_map.cumsum(0) dense_voxel_map = dense_voxel_map.cumsum(1) anchors_area = box_np_ops.fused_get_anchors_area( dense_voxel_map, anchors_bv, voxel_size, pc_range, grid_size) anchors_mask = anchors_area > anchor_area_threshold # example['anchors_mask'] = anchors_mask.astype(np.uint8) example['anchors_mask'] = anchors_mask # print("prep time", time.time() - t) metrics["prep_time"] = time.time() - t if not training: return example example["gt_names"] = gt_dict["gt_names"] # voxel_labels = box_np_ops.assign_label_to_voxel(gt_boxes, coordinates, # voxel_size, coors_range) if create_targets: t1 = time.time() targets_dict = target_assigner.assign( anchors, anchors_dict, gt_dict["gt_boxes"], anchors_mask, gt_classes=gt_dict["gt_classes"], gt_names=gt_dict["gt_names"], matched_thresholds=matched_thresholds, unmatched_thresholds=unmatched_thresholds, importance=gt_dict["gt_importance"]) """ boxes_lidar = gt_dict["gt_boxes"] bev_map = simplevis.nuscene_vis(points, boxes_lidar, gt_dict["gt_names"]) assigned_anchors = anchors[targets_dict['labels'] > 0] ignored_anchors = anchors[targets_dict['labels'] == -1] bev_map = simplevis.draw_box_in_bev(bev_map, [-50, -50, 3, 50, 50, 1], ignored_anchors, [128, 128, 128], 2) bev_map = simplevis.draw_box_in_bev(bev_map, [-50, -50, 3, 50, 50, 1], assigned_anchors, [255, 0, 0]) cv2.imshow('anchors', bev_map) cv2.waitKey(0) boxes_lidar = gt_dict["gt_boxes"] pp_map = np.zeros(grid_size[:2], dtype=np.float32) voxels_max = np.max(voxels[:, :, 2], axis=1, keepdims=False) voxels_min = np.min(voxels[:, :, 2], axis=1, keepdims=False) voxels_height = voxels_max - voxels_min voxels_height = np.minimum(voxels_height, 4) # sns.distplot(voxels_height) # plt.show() pp_map[coordinates[:, 1], coordinates[:, 2]] = voxels_height / 4 pp_map = (pp_map * 255).astype(np.uint8) pp_map = cv2.cvtColor(pp_map, cv2.COLOR_GRAY2RGB) pp_map = simplevis.draw_box_in_bev(pp_map, [-50, -50, 3, 50, 50, 1], boxes_lidar, [128, 0, 128], 1) cv2.imshow('heights', pp_map) cv2.waitKey(0) """ example.update({ 'labels': targets_dict['labels'], 'reg_targets': targets_dict['bbox_targets'], # 'reg_weights': targets_dict['bbox_outside_weights'], 'importance': targets_dict['importance'], }) return example
def prep_pointcloud(input_dict, root_path, voxel_generator, target_assigner, db_sampler=None, max_voxels=20000, class_names=['Car'], remove_outside_points=False, training=True, create_targets=True, shuffle_points=False, reduce_valid_area=False, remove_unknown=False, gt_rotation_noise=[-np.pi / 3, np.pi / 3], gt_loc_noise_std=[1.0, 1.0, 1.0], global_rotation_noise=[-np.pi / 4, np.pi / 4], global_scaling_noise=[0.95, 1.05], global_random_rot_range=[0.78, 2.35], generate_bev=False, without_reflectivity=False, num_point_features=4, anchor_area_threshold=1, gt_points_drop=0.0, gt_drop_max_keep=10, remove_points_after_sample=True, anchor_cache=None, remove_environment=False, random_crop=False, reference_detections=None, add_rgb_to_points=False, lidar_input=False, unlabeled_db_sampler=None, out_size_factor=2, min_gt_point_dict=None, bev_only=False, use_group_id=False, out_dtype=np.float32): """convert point cloud to voxels, create targets if ground truths exists. """ points = input_dict["points"] if training: gt_boxes = input_dict["gt_boxes"] gt_names = input_dict["gt_names"] difficulty = input_dict["difficulty"] group_ids = None if use_group_id and "group_ids" in input_dict: group_ids = input_dict["group_ids"] rect = input_dict["rect"] Trv2c = input_dict["Trv2c"] P2 = input_dict["P2"] unlabeled_training = unlabeled_db_sampler is not None #image_idx = input_dict["image_idx"] if reference_detections is not None: C, R, T = box_np_ops.projection_matrix_to_CRT_kitti(P2) frustums = box_np_ops.get_frustum_v2(reference_detections, C) frustums -= T # frustums = np.linalg.inv(R) @ frustums.T frustums = np.einsum('ij, akj->aki', np.linalg.inv(R), frustums) frustums = box_np_ops.camera_to_lidar(frustums, rect, Trv2c) surfaces = box_np_ops.corner_to_surfaces_3d_jit(frustums) masks = points_in_convex_polygon_3d_jit(points, surfaces) points = points[masks.any(-1)] if remove_outside_points and not lidar_input: image_shape = input_dict["image_shape"] points = box_np_ops.remove_outside_points(points, rect, Trv2c, P2, image_shape) if remove_environment is True and training: selected = kitti.keep_arrays_by_name(gt_names, class_names) gt_boxes = gt_boxes[selected] gt_names = gt_names[selected] difficulty = difficulty[selected] if group_ids is not None: group_ids = group_ids[selected] points = prep.remove_points_outside_boxes(points, gt_boxes) if training: # print(gt_names) selected = kitti.drop_arrays_by_name(gt_names, ["DontCare"]) gt_boxes = gt_boxes[selected] gt_names = gt_names[selected] difficulty = difficulty[selected] if group_ids is not None: group_ids = group_ids[selected] gt_boxes = box_np_ops.box_camera_to_lidar(gt_boxes, rect, Trv2c) if remove_unknown: remove_mask = difficulty == -1 """ gt_boxes_remove = gt_boxes[remove_mask] gt_boxes_remove[:, 3:6] += 0.25 points = prep.remove_points_in_boxes(points, gt_boxes_remove) """ keep_mask = np.logical_not(remove_mask) gt_boxes = gt_boxes[keep_mask] gt_names = gt_names[keep_mask] difficulty = difficulty[keep_mask] if group_ids is not None: group_ids = group_ids[keep_mask] gt_boxes_mask = np.array([n in class_names for n in gt_names], dtype=np.bool_) if db_sampler is not None: sampled_dict = db_sampler.sample_all(root_path, gt_boxes, gt_names, num_point_features, random_crop, gt_group_ids=group_ids, rect=rect, Trv2c=Trv2c, P2=P2) if sampled_dict is not None: sampled_gt_names = sampled_dict["gt_names"] sampled_gt_boxes = sampled_dict["gt_boxes"] sampled_points = sampled_dict["points"] sampled_gt_masks = sampled_dict["gt_masks"] # gt_names = gt_names[gt_boxes_mask].tolist() gt_names = np.concatenate([gt_names, sampled_gt_names], axis=0) # gt_names += [s["name"] for s in sampled] gt_boxes = np.concatenate([gt_boxes, sampled_gt_boxes]) gt_boxes_mask = np.concatenate( [gt_boxes_mask, sampled_gt_masks], axis=0) if group_ids is not None: sampled_group_ids = sampled_dict["group_ids"] group_ids = np.concatenate([group_ids, sampled_group_ids]) if remove_points_after_sample: points = prep.remove_points_in_boxes( points, sampled_gt_boxes) points = np.concatenate([sampled_points, points], axis=0) # unlabeled_mask = np.zeros((gt_boxes.shape[0], ), dtype=np.bool_) if without_reflectivity: used_point_axes = list(range(num_point_features)) used_point_axes.pop(3) points = points[:, used_point_axes] pc_range = voxel_generator.point_cloud_range if bev_only: # set z and h to limits gt_boxes[:, 2] = pc_range[2] gt_boxes[:, 5] = pc_range[5] - pc_range[2] prep.noise_per_object_v3_( gt_boxes, points, gt_boxes_mask, rotation_perturb=gt_rotation_noise, center_noise_std=gt_loc_noise_std, global_random_rot_range=global_random_rot_range, group_ids=group_ids, num_try=100) # should remove unrelated objects after noise per object gt_boxes = gt_boxes[gt_boxes_mask] gt_names = gt_names[gt_boxes_mask] if group_ids is not None: group_ids = group_ids[gt_boxes_mask] gt_classes = np.array([class_names.index(n) + 1 for n in gt_names], dtype=np.int32) gt_boxes, points = prep.random_flip(gt_boxes, points) gt_boxes, points = prep.global_rotation(gt_boxes, points, rotation=global_rotation_noise) gt_boxes, points = prep.global_scaling_v2(gt_boxes, points, *global_scaling_noise) bv_range = voxel_generator.point_cloud_range[[0, 1, 3, 4]] mask = prep.filter_gt_box_outside_range(gt_boxes, bv_range) gt_boxes = gt_boxes[mask] gt_classes = gt_classes[mask] gt_names = gt_names[mask] if group_ids is not None: group_ids = group_ids[mask] # limit rad to [-pi, pi] gt_boxes[:, 6] = box_np_ops.limit_period(gt_boxes[:, 6], offset=0.5, period=2 * np.pi) if shuffle_points: # shuffle is a little slow. np.random.shuffle(points) # [0, -40, -3, 70.4, 40, 1] voxel_size = voxel_generator.voxel_size pc_range = voxel_generator.point_cloud_range grid_size = voxel_generator.grid_size # [352, 400] voxels, coordinates, num_points = voxel_generator.generate( points, max_voxels) example = { 'voxels': voxels, 'num_points': num_points, 'coordinates': coordinates, "num_voxels": np.array([voxels.shape[0]], dtype=np.int64) } example.update({ 'rect': rect, 'Trv2c': Trv2c, 'P2': P2, }) # if not lidar_input: feature_map_size = grid_size[:2] // out_size_factor feature_map_size = [*feature_map_size, 1][::-1] if anchor_cache is not None: anchors = anchor_cache["anchors"] anchors_bv = anchor_cache["anchors_bv"] matched_thresholds = anchor_cache["matched_thresholds"] unmatched_thresholds = anchor_cache["unmatched_thresholds"] anchors_dict = anchor_cache["anchors_dict"] else: ret = target_assigner.generate_anchors(feature_map_size) anchors = ret["anchors"] anchors = anchors.reshape([-1, 7]) matched_thresholds = ret["matched_thresholds"] unmatched_thresholds = ret["unmatched_thresholds"] anchors_dict = target_assigner.generate_anchors_dict(feature_map_size) anchors_bv = box_np_ops.rbbox2d_to_near_bbox(anchors[:, [0, 1, 3, 4, 6]]) example["anchors"] = anchors # print("debug", anchors.shape, matched_thresholds.shape) # anchors_bv = anchors_bv.reshape([-1, 4]) anchors_mask = None if anchor_area_threshold >= 0: coors = coordinates dense_voxel_map = box_np_ops.sparse_sum_for_anchors_mask( coors, tuple(grid_size[::-1][1:])) dense_voxel_map = dense_voxel_map.cumsum(0) dense_voxel_map = dense_voxel_map.cumsum(1) anchors_area = box_np_ops.fused_get_anchors_area( dense_voxel_map, anchors_bv, voxel_size, pc_range, grid_size) anchors_mask = anchors_area > anchor_area_threshold # example['anchors_mask'] = anchors_mask.astype(np.uint8) example['anchors_mask'] = anchors_mask if not training: return example if create_targets: targets_dict = target_assigner.assign_v2(anchors_dict, gt_boxes, anchors_mask, gt_classes=gt_classes, gt_names=gt_names) example.update({ 'labels': targets_dict['labels'], 'reg_targets': targets_dict['bbox_targets'], 'reg_weights': targets_dict['bbox_outside_weights'], }) return example
def prep_pointcloud_ikg(input_dict, root_path, voxel_generator, target_assigner, db_sampler=None, max_voxels=20000, class_names=['Pedestrian'], remove_outside_points=False, training=True, create_targets=True, shuffle_points=False, reduce_valid_area=False, remove_unknown=False, gt_rotation_noise=[-np.pi / 3, np.pi / 3], gt_loc_noise_std=[1.0, 1.0, 1.0], global_rotation_noise=[-np.pi / 4, np.pi / 4], global_scaling_noise=[0.95, 1.05], global_loc_noise_std=(0.2, 0.2, 0.2), global_random_rot_range=[0.78, 2.35], generate_bev=False, without_reflectivity=False, num_point_features=4, anchor_area_threshold=1, gt_points_drop=0.0, gt_drop_max_keep=10, remove_points_after_sample=True, anchor_cache=None, remove_environment=False, random_crop=False, reference_detections=None, add_rgb_to_points=False, lidar_input=False, unlabeled_db_sampler=None, out_size_factor=2, min_gt_point_dict=None, bev_only=False, use_group_id=False, out_dtype=np.float32): """convert point cloud to voxels, create targets if ground truths exists. """ points = input_dict["points"] if training: gt_boxes = input_dict["gt_boxes"] gt_names = input_dict["gt_names"] image_idx = input_dict["image_idx"] if training: # print(gt_names) selected = kitti.drop_arrays_by_name(gt_names, ["DontCare"]) gt_boxes = gt_boxes[selected] gt_names = gt_names[selected] # unlabeled_mask = np.zeros((gt_boxes.shape[0], ), dtype=np.bool_) gt_boxes_mask = np.array([n in class_names for n in gt_names], dtype=np.bool_) pc_range = voxel_generator.point_cloud_range if bev_only: # set z and h to limits gt_boxes[:, 2] = pc_range[2] gt_boxes[:, 5] = pc_range[5] - pc_range[2] # should remove unrelated objects after noise per object gt_boxes = gt_boxes[gt_boxes_mask] gt_names = gt_names[gt_boxes_mask] gt_classes = np.array([class_names.index(n) + 1 for n in gt_names], dtype=np.int32) # gt_boxes, points = prep.random_flip(gt_boxes, points) # gt_boxes, points = prep.global_rotation( # gt_boxes, points, rotation=global_rotation_noise) # gt_boxes, points = prep.global_scaling_v2(gt_boxes, points, # *global_scaling_noise) # Global translation #gt_boxes, points = prep.global_translate(gt_boxes, points, global_loc_noise_std) bv_range = voxel_generator.point_cloud_range[[0, 1, 3, 4]] mask = prep.filter_gt_box_outside_range(gt_boxes, bv_range) gt_boxes = gt_boxes[mask] gt_classes = gt_classes[mask] # limit rad to [-pi, pi] gt_boxes[:, 6] = box_np_ops.limit_period(gt_boxes[:, 6], offset=0.5, period=2 * np.pi) #打乱点的顺序 # if shuffle_points: # # shuffle is a little slow. # np.random.shuffle(points) # [0, -40, -3, 70.4, 40, 1] voxel_size = voxel_generator.voxel_size pc_range = voxel_generator.point_cloud_range grid_size = voxel_generator.grid_size # [352, 400] voxels, coordinates, num_points = voxel_generator.generate( points, max_voxels) example = { 'voxels': voxels, 'num_points': num_points, 'coordinates': coordinates, "num_voxels": np.array([voxels.shape[0]], dtype=np.int64) } # if not lidar_input: feature_map_size = grid_size[:2] // out_size_factor feature_map_size = [*feature_map_size, 1][::-1] if anchor_cache is not None: anchors = anchor_cache["anchors"] anchors_bv = anchor_cache["anchors_bv"] matched_thresholds = anchor_cache["matched_thresholds"] unmatched_thresholds = anchor_cache["unmatched_thresholds"] else: ret = target_assigner.generate_anchors(feature_map_size) anchors = ret["anchors"] anchors = anchors.reshape([-1, 7]) matched_thresholds = ret["matched_thresholds"] unmatched_thresholds = ret["unmatched_thresholds"] anchors_bv = box_np_ops.rbbox2d_to_near_bbox(anchors[:, [0, 1, 3, 4, 6]]) example["anchors"] = anchors # print("debug", anchors.shape, matched_thresholds.shape) # anchors_bv = anchors_bv.reshape([-1, 4]) anchors_mask = None if anchor_area_threshold >= 0: coors = coordinates dense_voxel_map = box_np_ops.sparse_sum_for_anchors_mask( coors, tuple(grid_size[::-1][1:])) dense_voxel_map = dense_voxel_map.cumsum(0) dense_voxel_map = dense_voxel_map.cumsum(1) anchors_area = box_np_ops.fused_get_anchors_area( dense_voxel_map, anchors_bv, voxel_size, pc_range, grid_size) anchors_mask = anchors_area > anchor_area_threshold # example['anchors_mask'] = anchors_mask.astype(np.uint8) example['anchors_mask'] = anchors_mask if generate_bev: bev_vxsize = voxel_size.copy() bev_vxsize[:2] /= 2 bev_vxsize[2] *= 2 bev_map = points_to_bev(points, bev_vxsize, pc_range, without_reflectivity) example["bev_map"] = bev_map #if testing(without labels),return example = { #voxels,num_points,coordinates,num_voxels} if not training: return example if create_targets: targets_dict = target_assigner.assign( anchors, gt_boxes, anchors_mask, gt_classes=gt_classes, matched_thresholds=matched_thresholds, unmatched_thresholds=unmatched_thresholds) example.update({ 'labels': targets_dict['labels'], 'reg_targets': targets_dict['bbox_targets'], 'reg_weights': targets_dict['bbox_outside_weights'], }) return example
def prep_main(points, calib, voxel_generator, target_assigner, max_voxels=20000, shuffle_points=False, anchor_area_threshold=1, anchor_cache=None, out_size_factor=2, multi_gpu=False): """ Main preprocessing: coverting points to voxels. Main args: points: np.ndarray(dtype=float32, shape=(N, 3+)) TODO: confirm this. calib: calibration Returns: example: dict{'voxels', 'num_points', 'coordinates', 'num_voxels', 'metrics', 'calib', 'anchors', 'anchors_mask', 'anchors_dict', 'matched_thresholds', 'unmatched_thresholds' } """ metrics = {} if shuffle_points: # shuffle is a little slow. np.random.shuffle(points) # [0, -40, -3, 70.4, 40, 1] voxel_size = voxel_generator.voxel_size pc_range = voxel_generator.point_cloud_range grid_size = voxel_generator.grid_size # [352, 400] t1 = time.time() if not multi_gpu: res = voxel_generator.generate( points, max_voxels) voxels = res["voxels"] coordinates = res["coordinates"] num_points = res["num_points_per_voxel"] num_voxels = np.array([voxels.shape[0]], dtype=np.int64) else: res = voxel_generator.generate_multi_gpu( points, max_voxels) voxels = res["voxels"] coordinates = res["coordinates"] num_points = res["num_points_per_voxel"] num_voxels = np.array([res["voxel_num"]], dtype=np.int64) metrics["voxel_gene_time"] = time.time() - t1 example = { 'voxels': voxels, 'num_points': num_points, 'coordinates': coordinates, "num_voxels": num_voxels, "metrics": metrics, } if calib is not None: example["calib"] = calib feature_map_size = grid_size[:2] // out_size_factor feature_map_size = [*feature_map_size, 1][::-1] if anchor_cache is not None: anchors = anchor_cache["anchors"] anchors_bv = anchor_cache["anchors_bv"] anchors_dict = anchor_cache["anchors_dict"] matched_thresholds = anchor_cache["matched_thresholds"] unmatched_thresholds = anchor_cache["unmatched_thresholds"] else: ret = target_assigner.generate_anchors(feature_map_size) anchors = ret["anchors"] anchors = anchors.reshape([-1, target_assigner.box_ndim]) anchors_dict = target_assigner.generate_anchors_dict(feature_map_size) anchors_bv = box_np_ops.rbbox2d_to_near_bbox( anchors[:, [0, 1, 3, 4, 6]]) matched_thresholds = ret["matched_thresholds"] unmatched_thresholds = ret["unmatched_thresholds"] example["anchors"] = anchors example["anchors_dict"] = anchors_dict example["matched_thresholds"] = matched_thresholds example["unmatched_thresholds"] = unmatched_thresholds anchors_mask = None if anchor_area_threshold >= 0: # slow with high resolution. recommend disable this forever. coors = coordinates dense_voxel_map = box_np_ops.sparse_sum_for_anchors_mask( coors, tuple(grid_size[::-1][1:])) dense_voxel_map = dense_voxel_map.cumsum(0) dense_voxel_map = dense_voxel_map.cumsum(1) anchors_area = box_np_ops.fused_get_anchors_area( dense_voxel_map, anchors_bv, voxel_size, pc_range, grid_size) anchors_mask = anchors_area > anchor_area_threshold # example['anchors_mask'] = anchors_mask.astype(np.uint8) example['anchors_mask'] = anchors_mask return example
def prep_pointcloud(input_dict, root_path, voxel_generator, target_assigner, db_sampler=None, max_voxels=20000, remove_outside_points=False, training=True, create_targets=True, shuffle_points=False, remove_unknown=False, gt_rotation_noise=(-np.pi / 3, np.pi / 3), gt_loc_noise_std=(1.0, 1.0, 1.0), global_rotation_noise=(-np.pi / 4, np.pi / 4), global_scaling_noise=(0.95, 1.05), global_random_rot_range=(0.78, 2.35), global_translate_noise_std=(0, 0, 0), num_point_features=4, anchor_area_threshold=1, gt_points_drop=0.0, gt_drop_max_keep=10, remove_points_after_sample=True, anchor_cache=None, remove_environment=False, random_crop=False, reference_detections=None, out_size_factor=2, use_group_id=False, out_dtype=np.float32): """convert point cloud to voxels, create targets if ground truths exists. input_dict format: dataset.get_sensor_data format """ # t = time.time() class_names = target_assigner.classes points = input_dict["lidar"]["points"] if training: anno_dict = input_dict["lidar"]["annotations"] gt_dict = { "gt_boxes": anno_dict["boxes"], "gt_names": anno_dict["names"], } if "difficulty" not in anno_dict: difficulty = np.zeros([anno_dict["boxes"].shape[0]], dtype=np.int32) gt_dict["difficulty"] = difficulty else: gt_dict["difficulty"] = anno_dict["difficulty"] if use_group_id and "group_ids" in anno_dict: group_ids = anno_dict["group_ids"] gt_dict["group_ids"] = group_ids calib = None if "calib" in input_dict: calib = input_dict["calib"] if reference_detections is not None: assert calib is not None and "image" in input_dict C, R, T = box_np_ops.projection_matrix_to_CRT_kitti(P2) frustums = box_np_ops.get_frustum_v2(reference_detections, C) frustums -= T frustums = np.einsum('ij, akj->aki', np.linalg.inv(R), frustums) frustums = box_np_ops.camera_to_lidar(frustums, rect, Trv2c) surfaces = box_np_ops.corner_to_surfaces_3d_jit(frustums) masks = points_in_convex_polygon_3d_jit(points, surfaces) points = points[masks.any(-1)] if remove_outside_points: assert calib is not None image_shape = input_dict["image"]["image_shape"] points = box_np_ops.remove_outside_points(points, calib["rect"], calib["Trv2c"], calib["P2"], image_shape) if remove_environment is True and training: selected = kitti.keep_arrays_by_name(gt_names, target_assigner.classes) _dict_select(gt_dict, selected) masks = box_np_ops.points_in_rbbox(points, gt_dict["gt_boxes"]) points = points[masks.any(-1)] if training: # boxes_lidar = gt_dict["gt_boxes"] # bev_map = simplevis.nuscene_vis(points, boxes_lidar) # cv2.imshow('pre-noise', bev_map) selected = kitti.drop_arrays_by_name(gt_dict["gt_names"], ["DontCare"]) _dict_select(gt_dict, selected) if remove_unknown: remove_mask = gt_dict["difficulty"] == -1 """ gt_boxes_remove = gt_boxes[remove_mask] gt_boxes_remove[:, 3:6] += 0.25 points = prep.remove_points_in_boxes(points, gt_boxes_remove) """ keep_mask = np.logical_not(remove_mask) _dict_select(gt_dict, keep_mask) gt_dict.pop("difficulty") gt_boxes_mask = np.array( [n in class_names for n in gt_dict["gt_names"]], dtype=np.bool_) if db_sampler is not None: group_ids = None if "group_ids" in gt_dict: group_ids = gt_dict["group_ids"] sampled_dict = db_sampler.sample_all(root_path, gt_dict["gt_boxes"], gt_dict["gt_names"], num_point_features, random_crop, gt_group_ids=group_ids, calib=calib) if sampled_dict is not None: sampled_gt_names = sampled_dict["gt_names"] sampled_gt_boxes = sampled_dict["gt_boxes"] sampled_points = sampled_dict["points"] sampled_gt_masks = sampled_dict["gt_masks"] gt_dict["gt_names"] = np.concatenate( [gt_dict["gt_names"], sampled_gt_names], axis=0) gt_dict["gt_boxes"] = np.concatenate( [gt_dict["gt_boxes"], sampled_gt_boxes]) gt_boxes_mask = np.concatenate( [gt_boxes_mask, sampled_gt_masks], axis=0) if group_ids is not None: sampled_group_ids = sampled_dict["group_ids"] gt_dict["group_ids"] = np.concatenate( [gt_dict["group_ids"], sampled_group_ids]) if remove_points_after_sample: masks = box_np_ops.points_in_rbbox(points, sampled_gt_boxes) points = points[np.logical_not(masks.any(-1))] points = np.concatenate([sampled_points, points], axis=0) pc_range = voxel_generator.point_cloud_range group_ids = None if "group_ids" in gt_dict: group_ids = gt_dict["group_ids"] prep.noise_per_object_v3_( gt_dict["gt_boxes"], points, gt_boxes_mask, rotation_perturb=gt_rotation_noise, center_noise_std=gt_loc_noise_std, global_random_rot_range=global_random_rot_range, group_ids=group_ids, num_try=100) # should remove unrelated objects after noise per object # for k, v in gt_dict.items(): # print(k, v.shape) _dict_select(gt_dict, gt_boxes_mask) gt_classes = np.array( [class_names.index(n) + 1 for n in gt_dict["gt_names"]], dtype=np.int32) gt_dict["gt_classes"] = gt_classes gt_dict["gt_boxes"], points = prep.random_flip(gt_dict["gt_boxes"], points) gt_dict["gt_boxes"], points = prep.global_rotation( gt_dict["gt_boxes"], points, rotation=global_rotation_noise) gt_dict["gt_boxes"], points = prep.global_scaling_v2( gt_dict["gt_boxes"], points, *global_scaling_noise) prep.global_translate_(gt_dict["gt_boxes"], points, global_translate_noise_std) bv_range = voxel_generator.point_cloud_range[[0, 1, 3, 4]] mask = prep.filter_gt_box_outside_range(gt_dict["gt_boxes"], bv_range) _dict_select(gt_dict, mask) # limit rad to [-pi, pi] gt_dict["gt_boxes"][:, 6] = box_np_ops.limit_period( gt_dict["gt_boxes"][:, 6], offset=0.5, period=2 * np.pi) # boxes_lidar = gt_dict["gt_boxes"] # bev_map = simplevis.nuscene_vis(points, boxes_lidar) # cv2.imshow('post-noise', bev_map) # cv2.waitKey(0) if shuffle_points: # shuffle is a little slow. np.random.shuffle(points) # [0, -40, -3, 70.4, 40, 1] voxel_size = voxel_generator.voxel_size pc_range = voxel_generator.point_cloud_range grid_size = voxel_generator.grid_size # [352, 400] voxels, coordinates, num_points = voxel_generator.generate( points, max_voxels) example = { 'voxels': voxels, 'num_points': num_points, 'coordinates': coordinates, "num_voxels": np.array([voxels.shape[0]], dtype=np.int64) } if calib is not None: example["calib"] = calib feature_map_size = grid_size[:2] // out_size_factor feature_map_size = [*feature_map_size, 1][::-1] if anchor_cache is not None: anchors = anchor_cache["anchors"] anchors_bv = anchor_cache["anchors_bv"] anchors_dict = anchor_cache["anchors_dict"] else: ret = target_assigner.generate_anchors(feature_map_size) anchors = ret["anchors"] anchors = anchors.reshape([-1, 7]) anchors_dict = target_assigner.generate_anchors_dict(feature_map_size) anchors_bv = box_np_ops.rbbox2d_to_near_bbox(anchors[:, [0, 1, 3, 4, 6]]) example["anchors"] = anchors anchors_mask = None if anchor_area_threshold >= 0: # slow with high resolution. recommend disable this forever. coors = coordinates dense_voxel_map = box_np_ops.sparse_sum_for_anchors_mask( coors, tuple(grid_size[::-1][1:])) dense_voxel_map = dense_voxel_map.cumsum(0) dense_voxel_map = dense_voxel_map.cumsum(1) anchors_area = box_np_ops.fused_get_anchors_area( dense_voxel_map, anchors_bv, voxel_size, pc_range, grid_size) anchors_mask = anchors_area > anchor_area_threshold # example['anchors_mask'] = anchors_mask.astype(np.uint8) example['anchors_mask'] = anchors_mask if not training: return example # voxel_labels = box_np_ops.assign_label_to_voxel(gt_boxes, coordinates, # voxel_size, coors_range) """ example.update({ 'gt_boxes': gt_boxes.astype(out_dtype), 'num_gt': np.array([gt_boxes.shape[0]]), # 'voxel_labels': voxel_labels, }) """ if create_targets: targets_dict = target_assigner.assign_v2( anchors_dict, gt_dict["gt_boxes"], anchors_mask, gt_classes=gt_dict["gt_classes"], gt_names=gt_dict["gt_names"]) example.update({ 'labels': targets_dict['labels'], 'reg_targets': targets_dict['bbox_targets'], 'reg_weights': targets_dict['bbox_outside_weights'], }) return example
def prep_pointcloud(input_dict, root_path, voxel_generator, target_assigner, db_sampler=None, max_voxels=20000, class_names=['Car'], remove_outside_points=False, training=True, create_targets=True, shuffle_points=False, reduce_valid_area=False, remove_unknown=False, gt_rotation_noise=[-np.pi / 3, np.pi / 3], gt_loc_noise_std=[1.0, 1.0, 1.0], global_rotation_noise=[-np.pi / 4, np.pi / 4], global_scaling_noise=[0.95, 1.05], global_loc_noise_std=(0.2, 0.2, 0.2), global_random_rot_range=[0.78, 2.35], generate_bev=False, without_reflectivity=False, num_point_features=4, anchor_area_threshold=1, gt_points_drop=0.0, gt_drop_max_keep=10, remove_points_after_sample=True, anchor_cache=None, remove_environment=False, random_crop=False, reference_detections=None, add_rgb_to_points=False, lidar_input=False, unlabeled_db_sampler=None, out_size_factor=2, min_gt_point_dict=None, bev_only=False, use_group_id=False, out_dtype=np.float32): """convert point cloud to voxels, create targets if ground truths exists. """ # 这部分用来读取某一帧数据 points = input_dict["points"] # velodyne_reduced, array(N*4) if training: gt_boxes = input_dict["gt_boxes"] # 真值框,位置,尺寸,绝对转角,N*1,一个真值框一行 gt_names = input_dict["gt_names"] difficulty = input_dict["difficulty"] group_ids = None if use_group_id and "group_ids" in input_dict: # False group_ids = input_dict["group_ids"] rect = input_dict["rect"] Trv2c = input_dict["Trv2c"] P2 = input_dict["P2"] unlabeled_training = unlabeled_db_sampler is not None image_idx = input_dict["image_idx"] if reference_detections is not None: # None C, R, T = box_np_ops.projection_matrix_to_CRT_kitti(P2) frustums = box_np_ops.get_frustum_v2(reference_detections, C) frustums -= T # frustums = np.linalg.inv(R) @ frustums.T frustums = np.einsum('ij, akj->aki', np.linalg.inv(R), frustums) frustums = box_np_ops.camera_to_lidar(frustums, rect, Trv2c) surfaces = box_np_ops.corner_to_surfaces_3d_jit(frustums) masks = points_in_convex_polygon_3d_jit(points, surfaces) points = points[masks.any(-1)] if remove_outside_points and not lidar_input: # False image_shape = input_dict["image_shape"] points = box_np_ops.remove_outside_points(points, rect, Trv2c, P2, image_shape) if remove_environment is True and training: # False selected = kitti.keep_arrays_by_name(gt_names, class_names) gt_boxes = gt_boxes[selected] gt_names = gt_names[selected] difficulty = difficulty[selected] if group_ids is not None: # None group_ids = group_ids[selected] points = prep.remove_points_outside_boxes(points, gt_boxes) if training: # 先去掉真值内的DontCare selected = kitti.drop_arrays_by_name(gt_names, ["DontCare"]) # 去掉DontCare gt_boxes = gt_boxes[selected] gt_names = gt_names[selected] difficulty = difficulty[selected] if group_ids is not None: # None group_ids = group_ids[selected] gt_boxes = box_np_ops.box_camera_to_lidar(gt_boxes, rect, Trv2c) # 相机坐标下的真值框转换成激光雷达坐标下[xyz_lidar,w,l,h,r],一个对象一行 if remove_unknown: # False remove_mask = difficulty == -1 """ gt_boxes_remove = gt_boxes[remove_mask] gt_boxes_remove[:, 3:6] += 0.25 points = prep.remove_points_in_boxes(points, gt_boxes_remove) """ keep_mask = np.logical_not(remove_mask) gt_boxes = gt_boxes[keep_mask] gt_names = gt_names[keep_mask] difficulty = difficulty[keep_mask] if group_ids is not None: # None group_ids = group_ids[keep_mask] gt_boxes_mask = np.array( # 目标类别的对象标签,布尔类型,同样一行一个对象 [n in class_names for n in gt_names], dtype=np.bool_) # 下面用来对去掉DontCare的真值进行采样补充 if db_sampler is not None: # not None # 数据库预处理类里的方法,返回的是该帧数据内各类别用来补充的采样真值数据,包括真值内出现的非目标类别 sampled_dict = db_sampler.sample_all( root_path, gt_boxes, gt_names, num_point_features, random_crop, gt_group_ids=group_ids, # None rect=rect, Trv2c=Trv2c, P2=P2) if sampled_dict is not None: # 下面将原始数据与补充采样数据合并 sampled_gt_names = sampled_dict["gt_names"] sampled_gt_boxes = sampled_dict["gt_boxes"] sampled_points = sampled_dict["points"] sampled_gt_masks = sampled_dict["gt_masks"] # gt_names = gt_names[gt_boxes_mask].tolist() gt_names = np.concatenate([gt_names, sampled_gt_names], axis=0) # gt_names += [s["name"] for s in sampled] gt_boxes = np.concatenate([gt_boxes, sampled_gt_boxes]) gt_boxes_mask = np.concatenate( [gt_boxes_mask, sampled_gt_masks], axis=0) # 真值框是目标类别的标志 if group_ids is not None: # None sampled_group_ids = sampled_dict["group_ids"] group_ids = np.concatenate([group_ids, sampled_group_ids]) if remove_points_after_sample: # False,将采样框所占位置点云去除 points = prep.remove_points_in_boxes( points, sampled_gt_boxes) points = np.concatenate([sampled_points, points], axis=0) # 合并原始点云与采样点云 # unlabeled_mask = np.zeros((gt_boxes.shape[0], ), dtype=np.bool_) if without_reflectivity: # False used_point_axes = list(range(num_point_features)) used_point_axes.pop(3) points = points[:, used_point_axes] pc_range = voxel_generator.point_cloud_range if bev_only: # set z and h to limits, False gt_boxes[:, 2] = pc_range[2] gt_boxes[:, 5] = pc_range[5] - pc_range[2] prep.noise_per_object_v3_( # 每一个对象添加扰动,实际上就是修改每个对象真值框坐标与内部点云坐标 gt_boxes, # 补充过采样数据的真值框 points, # 补充过采样数据的点云 gt_boxes_mask, # 真值框标志(不区分原始数据与采样数据) rotation_perturb=gt_rotation_noise, # 旋转角度范围,平均分布 center_noise_std=gt_loc_noise_std, # 位置正态分布方差 global_random_rot_range=global_random_rot_range, # 全局旋转[0.0,0.0] group_ids=group_ids, # None num_try=100) # should remove unrelated objects after noise per object gt_boxes = gt_boxes[gt_boxes_mask] # 添加扰动后筛选目标类别真值框 gt_names = gt_names[gt_boxes_mask] if group_ids is not None: # Fasle group_ids = group_ids[gt_boxes_mask] gt_classes = np.array( # (gt_num) [class_names.index(n) + 1 for n in gt_names], dtype=np.int32) gt_boxes, points = prep.random_flip(gt_boxes, points) # 随机翻转 gt_boxes, points = prep.global_rotation( # 全局旋转 gt_boxes, points, rotation=global_rotation_noise) gt_boxes, points = prep.global_scaling_v2(gt_boxes, points, # 全局缩放 *global_scaling_noise) # Global translation # 全局定位扰动 gt_boxes, points = prep.global_translate(gt_boxes, points, global_loc_noise_std) bv_range = voxel_generator.point_cloud_range[[0, 1, 3, 4]] mask = prep.filter_gt_box_outside_range(gt_boxes, bv_range) # 过滤超出鸟瞰图范围的真值框 gt_boxes = gt_boxes[mask] gt_classes = gt_classes[mask] if group_ids is not None: # None group_ids = group_ids[mask] # limit rad to [-pi, pi] gt_boxes[:, 6] = box_np_ops.limit_period( # 调整真值转角至目标范围 gt_boxes[:, 6], offset=0.5, period=2 * np.pi) if shuffle_points: # True,打乱全局点云数据 # shuffle is a little slow. np.random.shuffle(points) voxel_size = voxel_generator.voxel_size # [0.16, 0.16, 4] pc_range = voxel_generator.point_cloud_range # [0, -39.68, -3, 69.12, 39.68, 1] grid_size = voxel_generator.grid_size # [432, 496, 1] x,y,z # 生成体素 """ Returns: voxels:[num_voxels, 100, 4] 体素索引映射全局体素特征 coordinates:[num_voxels, 3] 体素索引映射体素坐标 num_points:(num_voxels,) 体素索引映射体素内的点数 """ voxels, coordinates, num_points = voxel_generator.generate( points, max_voxels) example = { 'voxels': voxels, 'num_points': num_points, 'coordinates': coordinates, "num_voxels": np.array([voxels.shape[0]], dtype=np.int64) } example.update({ 'rect': rect, 'Trv2c': Trv2c, 'P2': P2, }) # if not lidar_input: feature_map_size = grid_size[:2] // out_size_factor # [216 248] feature_map_size = [*feature_map_size, 1][::-1] # [1,248,216] if anchor_cache is not None: anchors = anchor_cache["anchors"] anchors_bv = anchor_cache["anchors_bv"] matched_thresholds = anchor_cache["matched_thresholds"] unmatched_thresholds = anchor_cache["unmatched_thresholds"] else: ret = target_assigner.generate_anchors(feature_map_size) anchors = ret["anchors"] anchors = anchors.reshape([-1, 7]) matched_thresholds = ret["matched_thresholds"] unmatched_thresholds = ret["unmatched_thresholds"] anchors_bv = box_np_ops.rbbox2d_to_near_bbox( anchors[:, [0, 1, 3, 4, 6]]) example["anchors"] = anchors # print("debug", anchors.shape, matched_thresholds.shape) # anchors_bv = anchors_bv.reshape([-1, 4]) anchors_mask = None if anchor_area_threshold >= 0: # True coors = coordinates dense_voxel_map = box_np_ops.sparse_sum_for_anchors_mask( # 某处体素是否被采样,否为0,是为1.array[496,432] y,x coors, tuple(grid_size[::-1][1:])) dense_voxel_map = dense_voxel_map.cumsum(0) # 元素值为按列累加到元素的和 dense_voxel_map = dense_voxel_map.cumsum(1) # 元素值为按行累加到元素的和,相当于统计了元素两个维度上之前元素的和 anchors_area = box_np_ops.fused_get_anchors_area( # 统计每个锚框内的体素数量 dense_voxel_map, anchors_bv, voxel_size, pc_range, grid_size) anchors_mask = anchors_area > anchor_area_threshold # 标记超过一个体素的锚框 # example['anchors_mask'] = anchors_mask.astype(np.uint8) example['anchors_mask'] = anchors_mask if generate_bev: # False bev_vxsize = voxel_size.copy() bev_vxsize[:2] /= 2 bev_vxsize[2] *= 2 bev_map = points_to_bev(points, bev_vxsize, pc_range, without_reflectivity) example["bev_map"] = bev_map if not training: return example # 测试数据集不需要创建训练目标 if create_targets: # True targets_dict = target_assigner.assign( anchors, # (248*216*2,7) gt_boxes, # (gt_num, 7) anchors_mask, # (248*216*2,) gt_classes=gt_classes, # (gt_num,) matched_thresholds=matched_thresholds, # (248*216*2,) unmatched_thresholds=unmatched_thresholds) # (248*216*2,) example.update({ 'labels': targets_dict['labels'], # (total_anchors,),所有锚框对应真值类别(1,2...),无对应真值设为0,dontcare设为-1 'reg_targets': targets_dict['bbox_targets'], # (total_anchors, 7),所有锚框对应真值相对锚框的偏差编码,无对应真值设为[0,0,0,0,0,0,0] 'reg_weights': targets_dict['bbox_outside_weights'], # (total_anchors,),所有锚框的外部权重,有对应真值设为1,无对应真值设为0 }) return example
def prep_pointcloud(input_dict, root_path, voxel_generator, target_assigner, db_sampler=None, max_voxels=20000, class_names=['PEDESTRIAN'], remove_outside_points=False, training=True, create_targets=True, shuffle_points=False, reduce_valid_area=False, remove_unknown=False, gt_rotation_noise=[-np.pi / 3, np.pi / 3], gt_loc_noise_std=[1.0, 1.0, 1.0], global_rotation_noise=[-np.pi / 4, np.pi / 4], global_scaling_noise=[0.95, 1.05], global_loc_noise_std=(0.2, 0.2, 0.2), global_random_rot_range=[0.78, 2.35], generate_bev=False, without_reflectivity=False, num_point_features=4, anchor_area_threshold=1, gt_points_drop=0.0, gt_drop_max_keep=10, remove_points_after_sample=True, anchor_cache=None, remove_environment=False, random_crop=False, reference_detections=None, add_rgb_to_points=False, lidar_input=False, unlabeled_db_sampler=None, out_size_factor=2, min_gt_point_dict=None, bev_only=False, use_group_id=False, out_dtype=np.float32): """convert point cloud to voxels, create targets if ground truths exists. """ points = input_dict["points"] pc_range = voxel_generator.point_cloud_range pts_x, pts_y, pts_z = points[:, 0], points[:, 1], points[:, 2] range_flag = ((pts_x >= pc_range[0]) & (pts_x <= pc_range[3]) & (pts_y >= pc_range[1]) & (pts_y <= pc_range[4]) & (pts_z >= pc_range[2]) & (pts_z <= pc_range[5])) points = points[range_flag] if training: gt_boxes = input_dict["gt_boxes"] gt_names = input_dict["gt_names"] ## group_ids ? np.arange(num_gt,dtype=np.int32) num_gt - number of objects (of all categories) in annotated lidar frame group_ids = None if use_group_id and "group_ids" in input_dict: group_ids = input_dict["group_ids"] #unlabeled_training = unlabeled_db_sampler is not None if training: gt_boxes_mask = np.array([n in class_names for n in gt_names], dtype=np.bool_) #print(gt_boxes_mask.shape,gt_boxes.shape,"before") prep.noise_per_object_v3_( gt_boxes, points, gt_boxes_mask, rotation_perturb=gt_rotation_noise, center_noise_std=gt_loc_noise_std, global_random_rot_range=global_random_rot_range, group_ids=group_ids, num_try=100) #print(gt_boxes_mask.shape,gt_boxes.shape,"after") # should remove unrelated objects after noise per object gt_boxes = gt_boxes[gt_boxes_mask] gt_names = gt_names[gt_boxes_mask] if group_ids is not None: group_ids = group_ids[gt_boxes_mask] gt_classes = np.array([class_names.index(n) + 1 for n in gt_names], dtype=np.int32) #need to check the output gt_boxes, points = prep.random_flip(gt_boxes, points) gt_boxes, points = prep.global_rotation(gt_boxes, points, rotation=global_rotation_noise) gt_boxes, points = prep.global_scaling_v2(gt_boxes, points, *global_scaling_noise) # Global translation gt_boxes, points = prep.global_translate(gt_boxes, points, global_loc_noise_std) bv_range = voxel_generator.point_cloud_range[[0, 1, 3, 4]] mask = prep.filter_gt_box_outside_range(gt_boxes, bv_range) gt_boxes = gt_boxes[mask] gt_classes = gt_classes[mask] if group_ids is not None: group_ids = group_ids[mask] # limit rad to [-pi, pi] gt_boxes[:, 6] = box_np_ops.limit_period(gt_boxes[:, 6], offset=0.5, period=2 * np.pi) #assert -np.pi/2 <= g <= np.pi/2 if shuffle_points: # shuffle is a little slow. np.random.shuffle(points) voxel_size = voxel_generator.voxel_size pc_range = voxel_generator.point_cloud_range grid_size = voxel_generator.grid_size voxels, coordinates, num_points = voxel_generator.generate( points, max_voxels) example = { 'voxels': voxels, 'num_points': num_points, 'coordinates': coordinates, "num_voxels": np.array([voxels.shape[0]], dtype=np.int64) } # if not lidar_input: feature_map_size = grid_size[:2] // out_size_factor feature_map_size = [*feature_map_size, 1][::-1] if anchor_cache is not None: anchors = anchor_cache["anchors"] anchors_bv = anchor_cache["anchors_bv"] matched_thresholds = anchor_cache["matched_thresholds"] unmatched_thresholds = anchor_cache["unmatched_thresholds"] else: ret = target_assigner.generate_anchors(feature_map_size) anchors = ret["anchors"] anchors = anchors.reshape([-1, 7]) matched_thresholds = ret["matched_thresholds"] unmatched_thresholds = ret["unmatched_thresholds"] anchors_bv = box_np_ops.rbbox2d_to_near_bbox(anchors[:, [0, 1, 3, 4, 6]]) example["anchors"] = anchors anchors_mask = None if anchor_area_threshold >= 0: coors = coordinates dense_voxel_map = box_np_ops.sparse_sum_for_anchors_mask( coors, tuple(grid_size[::-1][1:])) dense_voxel_map = dense_voxel_map.cumsum(0) dense_voxel_map = dense_voxel_map.cumsum(1) anchors_area = box_np_ops.fused_get_anchors_area( dense_voxel_map, anchors_bv, voxel_size, pc_range, grid_size) anchors_mask = anchors_area > anchor_area_threshold example['anchors_mask'] = anchors_mask if generate_bev: bev_vxsize = voxel_size.copy() bev_vxsize[:2] /= 2 bev_vxsize[2] *= 2 bev_map = points_to_bev(points, bev_vxsize, pc_range, without_reflectivity) example["bev_map"] = bev_map if not training: return example if create_targets: targets_dict = target_assigner.assign( anchors, gt_boxes, anchors_mask, gt_classes=gt_classes, matched_thresholds=matched_thresholds, unmatched_thresholds=unmatched_thresholds) example.update({ 'labels': targets_dict['labels'], 'reg_targets': targets_dict['bbox_targets'], 'reg_weights': targets_dict['bbox_outside_weights'], }) return example
def draw_anchors(self, gt_boxes_lidar, points=None, image_idx=0, gt_names=None): # print(gt_names) voxel_size = np.array(self._voxel_size, dtype=np.float32) # voxel_size = np.array([0.2, 0.2, 0.4], dtype=np.float32) coors_range = np.array(self._coors_range, dtype=np.float32) # coors_range = np.array([0, -40, -3, 70.4, 40, 1], dtype=np.float32) grid_size = (coors_range[3:] - coors_range[:3]) / voxel_size grid_size = np.round(grid_size).astype(np.int64) # print(grid_size) bv_range = coors_range[[0, 1, 3, 4]] anchor_generator = AnchorGeneratorStride( # sizes=[0.6, 0.8, 1.73, 0.6, 1.76, 1.73], sizes=[0.6, 1.76, 1.73], anchor_strides=[0.4, 0.4, 0.0], anchor_offsets=[0.2, -39.8, -1.465], rotations=[0, 1.5707963267948966], match_threshold=0.5, unmatch_threshold=0.35, ) anchor_generator1 = AnchorGeneratorStride( # sizes=[0.6, 0.8, 1.73, 0.6, 1.76, 1.73], sizes=[0.6, 0.8, 1.73], anchor_strides=[0.4, 0.4, 0.0], anchor_offsets=[0.2, -39.8, -1.465], rotations=[0, 1.5707963267948966], match_threshold=0.5, unmatch_threshold=0.35, ) anchor_generator2 = AnchorGeneratorStride( # sizes=[0.6, 0.8, 1.73, 0.6, 1.76, 1.73], sizes=[1.6, 3.9, 1.56], anchor_strides=[0.4, 0.4, 0.0], anchor_offsets=[0.2, -39.8, -1.55442884], rotations=[0, 1.5707963267948966], # rotations=[0], match_threshold=0.6, unmatch_threshold=0.45, ) anchor_generators = [anchor_generator2] box_coder = GroundBox3dCoder() # similarity_calc = DistanceSimilarity(1.0) similarity_calc = NearestIouSimilarity() target_assigner = TargetAssigner(box_coder, anchor_generators, similarity_calc) # anchors = box_np_ops.create_anchors_v2( # bv_range, grid_size[:2] // 2, sizes=anchor_dims) # matched_thresholds = [0.45, 0.45, 0.6] # unmatched_thresholds = [0.3, 0.3, 0.45] t = time.time() feature_map_size = grid_size[:2] // 2 feature_map_size = [*feature_map_size, 1][::-1] # print(feature_map_size) # """ ret = target_assigner.generate_anchors(feature_map_size) anchors = ret["anchors"] anchors = anchors.reshape([-1, 7]) anchors_bv = box_np_ops.rbbox2d_to_near_bbox(anchors[:, [0, 1, 3, 4, 6]]) matched_thresholds = ret["matched_thresholds"] unmatched_thresholds = ret["unmatched_thresholds"] print(f"num_anchors_ {len(anchors)}") if points is not None: voxels, coors, num_points = points_to_voxel( points, self._voxel_size, # self._coors_range, coors_range, self._max_num_points, reverse_index=True, max_voxels=self._max_voxels) # print(np.min(coors, 0), np.max(coors, 0)) dense_voxel_map = box_np_ops.sparse_sum_for_anchors_mask( coors, tuple(grid_size[::-1][1:])) dense_voxel_map = dense_voxel_map.cumsum(0) dense_voxel_map = dense_voxel_map.cumsum(1) anchors_mask = box_np_ops.fused_get_anchors_area( dense_voxel_map, anchors_bv, voxel_size, coors_range, grid_size) > 1 # print(np.sum(anchors_mask), anchors_mask.shape) class_names = [ 'Car', "Pedestrian", "Cyclist", 'Van', 'Truck', "Tram", 'Misc', 'Person_sitting' ] gt_classes = np.array([class_names.index(n) + 1 for n in gt_names], dtype=np.int32) t = time.time() target_dict = target_assigner.assign( anchors, gt_boxes_lidar, anchors_mask, gt_classes=gt_classes, matched_thresholds=matched_thresholds, unmatched_thresholds=unmatched_thresholds) labels = target_dict["labels"] reg_targets = target_dict["bbox_targets"] reg_weights = target_dict["bbox_outside_weights"] # print(labels[labels > 0]) # decoded_reg_targets = box_np_ops.second_box_decode(reg_targets, anchors) # print(decoded_reg_targets.reshape(-1, 7)[labels > 0]) print("target time", (time.time() - t)) print(f"num_pos={np.sum(labels > 0)}") colors = np.zeros([anchors.shape[0], 4]) ignored_color = bbox_plot.gl_color(GLColor.Gray, 0.5) pos_color = bbox_plot.gl_color(GLColor.Cyan, 0.5) colors[labels == -1] = ignored_color colors[labels > 0] = pos_color cared_anchors_mask = np.logical_and(labels != 0, anchors_mask) colors = colors[cared_anchors_mask] anchors_not_neg = box_np_ops.rbbox3d_to_corners( anchors)[cared_anchors_mask] self.boxes3d("anchors", anchors_not_neg, colors=colors)
def prep_pointcloud2(input_dict, root_path, voxel_generator, target_assigner, db_sampler=None, max_voxels=20000, class_names=['Car'], remove_outside_points=False, training=True, create_targets=True, shuffle_points=False, reduce_valid_area=False, remove_unknown=False, gt_rotation_noise=[-np.pi / 3, np.pi / 3], gt_loc_noise_std=[1.0, 1.0, 1.0], global_rotation_noise=[-np.pi / 4, np.pi / 4], global_scaling_noise=[0.95, 1.05], global_loc_noise_std=(0.2, 0.2, 0.2), global_random_rot_range=[0.78, 2.35], generate_bev=False, without_reflectivity=False, num_point_features=4, anchor_area_threshold=1, gt_points_drop=0.0, gt_drop_max_keep=10, remove_points_after_sample=True, anchor_cache=None, remove_environment=False, random_crop=False, reference_detections=None, add_rgb_to_points=False, lidar_input=False, unlabeled_db_sampler=None, out_size_factor=2, min_gt_point_dict=None, bev_only=False, use_group_id=False, out_dtype=np.float32): """convert point cloud to voxels, create targets if ground truths exists. """ points = input_dict["points"] rect = input_dict["rect"] Trv2c = input_dict["Trv2c"] P2 = input_dict["P2"] unlabeled_training = unlabeled_db_sampler is not None image_idx = input_dict["image_idx"] if reference_detections is not None: C, R, T = box_np_ops.projection_matrix_to_CRT_kitti(P2) frustums = box_np_ops.get_frustum_v2(reference_detections, C) frustums -= T # frustums = np.linalg.inv(R) @ frustums.T frustums = np.einsum('ij, akj->aki', np.linalg.inv(R), frustums) frustums = box_np_ops.camera_to_lidar(frustums, rect, Trv2c) surfaces = box_np_ops.corner_to_surfaces_3d_jit(frustums) masks = points_in_convex_polygon_3d_jit(points, surfaces) points = points[masks.any(-1)] now = rospy.get_rostime() print('before shuffle TIME:' + str(now.to_sec())) if shuffle_points: # shuffle is a little slow. np.random.shuffle(points) # [0, -40, -3, 70.4, 40, 1] voxel_size = voxel_generator.voxel_size pc_range = voxel_generator.point_cloud_range grid_size = voxel_generator.grid_size # [352, 400] now = rospy.get_rostime() print('before voxel generate:' + str(now.to_sec())) voxels, coordinates, num_points = voxel_generator.generate( points, max_voxels) now = rospy.get_rostime() print('before encording TIME:' + str(now.to_sec())) example = { 'voxels': voxels, 'num_points': num_points, 'coordinates': coordinates, "num_voxels": np.array([voxels.shape[0]], dtype=np.int64) } now = rospy.get_rostime() print('before encording2 TIME:' + str(now.to_sec())) example.update({ 'rect': rect, 'Trv2c': Trv2c, 'P2': P2, }) now = rospy.get_rostime() print('before anchor_cache TIME:' + str(now.to_sec())) # if not lidar_input: feature_map_size = grid_size[:2] // out_size_factor now = rospy.get_rostime() print('before anchor_fm TIME:' + str(now.to_sec())) feature_map_size = [*feature_map_size, 1][::-1] if anchor_cache is not None: anchors = anchor_cache["anchors"] anchors_bv = anchor_cache["anchors_bv"] matched_thresholds = anchor_cache["matched_thresholds"] unmatched_thresholds = anchor_cache["unmatched_thresholds"] else: ret = target_assigner.generate_anchors(feature_map_size) anchors = ret["anchors"] anchors = anchors.reshape([-1, 7]) matched_thresholds = ret["matched_thresholds"] unmatched_thresholds = ret["unmatched_thresholds"] anchors_bv = box_np_ops.rbbox2d_to_near_bbox(anchors[:, [0, 1, 3, 4, 6]]) example["anchors"] = anchors # print("debug", anchors.shape, matched_thresholds.shape) # anchors_bv = anchors_bv.reshape([-1, 4]) anchors_mask = None now = rospy.get_rostime() print('before create_anchor_mask TIME:' + str(now.to_sec())) #costs time if anchor_area_threshold >= 0: coors = coordinates dense_voxel_map = box_np_ops.sparse_sum_for_anchors_mask( coors, tuple(grid_size[::-1][1:])) dense_voxel_map = dense_voxel_map.cumsum(0) dense_voxel_map = dense_voxel_map.cumsum(1) anchors_area = box_np_ops.fused_get_anchors_area( dense_voxel_map, anchors_bv, voxel_size, pc_range, grid_size) anchors_mask = anchors_area > anchor_area_threshold # example['anchors_mask'] = anchors_mask.astype(np.uint8) example['anchors_mask'] = anchors_mask now = rospy.get_rostime() print('before create_bev TIME:' + str(now.to_sec())) if generate_bev: bev_vxsize = voxel_size.copy() bev_vxsize[:2] /= 2 bev_vxsize[2] *= 2 bev_map = points_to_bev(points, bev_vxsize, pc_range, without_reflectivity) example["bev_map"] = bev_map now = rospy.get_rostime() print('after create bev TIME:' + str(now.to_sec())) if not training: return example
def prep_pointcloud(input_dict, root_path, voxel_generator, target_assigner, db_sampler=None, max_voxels=20000, max_sweeps=10, remove_outside_points=False, training=True, create_targets=True, shuffle_points=False, remove_unknown=False, gt_rotation_noise=(-np.pi / 3, np.pi / 3), gt_loc_noise_std=(1.0, 1.0, 1.0), global_rotation_noise=(-np.pi / 4, np.pi / 4), global_scaling_noise=(0.95, 1.05), global_random_rot_range=(0.78, 2.35), global_translate_noise_std=(0, 0, 0), num_point_features=4, anchor_area_threshold=1, gt_points_drop=0.0, gt_drop_max_keep=10, remove_points_after_sample=True, anchor_cache=None, remove_environment=False, random_crop=False, reference_detections=None, out_size_factor=2, use_group_id=False, multi_gpu=False, min_points_in_gt=-1, random_flip_x=True, random_flip_y=True, sample_importance=1.0, out_dtype=np.float32): """convert point cloud to voxels, create targets if ground truths exists. input_dict format: dataset.get_sensor_data format """ t = time.time() class_names = target_assigner.classes points = input_dict["lidar"]["points"] indices = input_dict["lidar"]["indices"] origins = input_dict["lidar"]["origins"] if training: anno_dict = input_dict["lidar"]["annotations"] gt_dict = { "gt_boxes": anno_dict["boxes"], "gt_names": anno_dict["names"], "gt_importance": np.ones([anno_dict["boxes"].shape[0]], dtype=anno_dict["boxes"].dtype), } if "difficulty" not in anno_dict: difficulty = np.zeros([anno_dict["boxes"].shape[0]], dtype=np.int32) gt_dict["difficulty"] = difficulty else: gt_dict["difficulty"] = anno_dict["difficulty"] if use_group_id and "group_ids" in anno_dict: group_ids = anno_dict["group_ids"] gt_dict["group_ids"] = group_ids calib = None if "calib" in input_dict: calib = input_dict["calib"] # # Disable these two since we do not do this for NuScenes # if reference_detections is not None: # assert calib is not None and "image" in input_dict # C, R, T = box_np_ops.projection_matrix_to_CRT_kitti(P2) # frustums = box_np_ops.get_frustum_v2(reference_detections, C) # frustums -= T # frustums = np.einsum('ij, akj->aki', np.linalg.inv(R), frustums) # frustums = box_np_ops.camera_to_lidar(frustums, rect, Trv2c) # surfaces = box_np_ops.corner_to_surfaces_3d_jit(frustums) # masks = points_in_convex_polygon_3d_jit(points, surfaces) # points = points[masks.any(-1)] # if remove_outside_points: # assert calib is not None # image_shape = input_dict["image"]["image_shape"] # points = box_np_ops.remove_outside_points( # points, calib["rect"], calib["Trv2c"], calib["P2"], image_shape) # # Very interesting attempt # # I have tried the same and found it doesn't really work # if remove_environment is True and training: # selected = kitti.keep_arrays_by_name(gt_names, target_assigner.classes) # _dict_select(gt_dict, selected) # masks = box_np_ops.points_in_rbbox(points, gt_dict["gt_boxes"]) # points = points[masks.any(-1)] metrics = {} point_indices_to_remove = None if training: """ boxes_lidar = gt_dict["gt_boxes"] bev_map = simplevis.nuscene_vis(points, boxes_lidar) cv2.imshow('pre-noise', bev_map) """ selected = kitti.drop_arrays_by_name(gt_dict["gt_names"], ["Denture"]) _dict_select(gt_dict, selected) if remove_unknown: remove_mask = gt_dict["difficulty"] == -1 """ gt_boxes_remove = gt_boxes[remove_mask] gt_boxes_remove[:, 3:6] += 0.25 points = prep.remove_points_in_boxes(points, gt_boxes_remove) """ keep_mask = np.logical_not(remove_mask) _dict_select(gt_dict, keep_mask) gt_dict.pop("difficulty") # This part is interesting - we will need to do the same if min_points_in_gt > 0: # points_count_rbbox takes 10ms with 10 sweeps nuscenes data point_counts = box_np_ops.points_count_rbbox( points, gt_dict["gt_boxes"]) mask = point_counts >= min_points_in_gt _dict_select(gt_dict, mask) gt_boxes_mask = np.array( [n in class_names for n in gt_dict["gt_names"]], dtype=np.bool_) if db_sampler is not None: group_ids = None if "group_ids" in gt_dict: group_ids = gt_dict["group_ids"] sampled_dict = db_sampler.sample_all(root_path, gt_dict["gt_boxes"], gt_dict["gt_names"], num_point_features, random_crop, gt_group_ids=group_ids, calib=calib) if sampled_dict is not None: sampled_gt_names = sampled_dict["gt_names"] sampled_gt_boxes = sampled_dict["gt_boxes"] sampled_points = sampled_dict["points"] sampled_gt_masks = sampled_dict["gt_masks"] gt_dict["gt_names"] = np.concatenate( [gt_dict["gt_names"], sampled_gt_names], axis=0) gt_dict["gt_boxes"] = np.concatenate( [gt_dict["gt_boxes"], sampled_gt_boxes]) gt_boxes_mask = np.concatenate( [gt_boxes_mask, sampled_gt_masks], axis=0) sampled_gt_importance = np.full([sampled_gt_boxes.shape[0]], sample_importance, dtype=sampled_gt_boxes.dtype) gt_dict["gt_importance"] = np.concatenate( [gt_dict["gt_importance"], sampled_gt_importance]) if group_ids is not None: sampled_group_ids = sampled_dict["group_ids"] gt_dict["group_ids"] = np.concatenate( [gt_dict["group_ids"], sampled_group_ids]) # # Commented out because we have a new way of removing points # if remove_points_after_sample: # masks = box_np_ops.points_in_rbbox(points, sampled_gt_boxes) # point_indices_to_remove = np.flatnonzero(masks.any(-1)) # # # Delay this process so we can use the full point cloud # # # when we do the ray stopping algorithm # # points = points[np.logical_not(masks.any(-1))] # # Paste objects behind so that we don't have to update indices # points = np.concatenate([sampled_points, points], axis=0) points = np.concatenate([points, sampled_points], axis=0) pc_range = voxel_generator.point_cloud_range group_ids = None if "group_ids" in gt_dict: group_ids = gt_dict["group_ids"] # # Disable this one for now (not used in PointPillars anyways) # prep.noise_per_object_v3_( # gt_dict["gt_boxes"], # points, # gt_boxes_mask, # rotation_perturb=gt_rotation_noise, # center_noise_std=gt_loc_noise_std, # global_random_rot_range=global_random_rot_range, # group_ids=group_ids, # num_try=100) # should remove unrelated objects after noise per object # for k, v in gt_dict.items(): # print(k, v.shape) _dict_select(gt_dict, gt_boxes_mask) gt_classes = np.array( [class_names.index(n) + 1 for n in gt_dict["gt_names"]], dtype=np.int32) gt_dict["gt_classes"] = gt_classes gt_dict["gt_boxes"], points, origins = prep.random_flip( gt_dict["gt_boxes"], points, origins, 0.5, random_flip_x, random_flip_y) gt_dict["gt_boxes"], points, origins = prep.global_rotation_v2( gt_dict["gt_boxes"], points, origins, *global_rotation_noise) gt_dict["gt_boxes"], points, origins = prep.global_scaling_v2( gt_dict["gt_boxes"], points, origins, *global_scaling_noise) prep.global_translate_(gt_dict["gt_boxes"], points, origins, global_translate_noise_std) bv_range = voxel_generator.point_cloud_range[[0, 1, 3, 4]] mask = prep.filter_gt_box_outside_range_by_center( gt_dict["gt_boxes"], bv_range) _dict_select(gt_dict, mask) # limit rad to [-pi, pi] gt_dict["gt_boxes"][:, 6] = box_np_ops.limit_period( gt_dict["gt_boxes"][:, 6], offset=0.5, period=2 * np.pi) # boxes_lidar = gt_dict["gt_boxes"] # bev_map = simplevis.nuscene_vis(points, boxes_lidar) # cv2.imshow('post-noise', bev_map) # cv2.waitKey(0) # # Disable this for now (not used in PointPillars anyways) # if shuffle_points: # # shuffle is a little slow. # np.random.shuffle(points) # [0, -40, -3, 70.4, 40, 1] voxel_size = voxel_generator.voxel_size pc_range = voxel_generator.point_cloud_range grid_size = voxel_generator.grid_size # organize points into lists based on timestamps time_stamps = points[ indices[:-1], -1] # counting on the fact we do not miss points from any intermediate time_stamps time_stamps = (time_stamps[:-1] + time_stamps[1:]) / 2 time_stamps = [-1000.0] + time_stamps.tolist() + [1000.0] # add boundaries time_stamps = np.array(time_stamps) # # LL_OCCUPIED, LL_FREE = 0.85, -0.4 # lo_occupied = np.log(0.7 / (1 - 0.7)) # lo_free = np.log(0.4 / (1 - 0.4)) # is there are additional points (from database sampling) num_original = indices[-1] if len(points) > num_original: # split data into two half (indexed and un-indexed) original_points, sampled_points = points[:num_original], points[ num_original:] # compute occupancy and masks # visibility, original_mask, sampled_mask = mapping.compute_visibility_and_masks( # original_points, sampled_points, origins, time_stamps, pc_range, min(voxel_size) # ) logodds, original_mask, sampled_mask = mapping.compute_logodds_and_masks( original_points, sampled_points, origins, time_stamps, pc_range, min(voxel_size) # , lo_occupied, lo_free ) # apply visible mask points = np.concatenate( (original_points[original_mask], sampled_points[sampled_mask])) else: # visibility = mapping.compute_visibility( # points, origins, time_stamps, pc_range, min(voxel_size) # ) logodds = mapping.compute_logodds( points, origins, time_stamps, pc_range, min(voxel_size) #, lo_occupied, lo_free ) # T = len(time_stamps)-1 # visibility = visibility.reshape(T, -1) # if T < (1 + max_sweeps): # visibility = np.pad(visibility, ((0, (1+max_sweeps)-T), (0,0)), 'edge') # with open(f'./utils/mapping/examples/{time.time()}.pkl', 'wb') as f: # ## # pickle.dump(original_points, f) # pickle.dump(sampled_points, f) # pickle.dump(origins, f) # pickle.dump(time_stamps, f) # pickle.dump(pc_range, f) # pickle.dump(voxel_size, f) # ## # pickle.dump(occupancy, f) # pickle.dump(original_mask, f) # pickle.dump(sampled_mask, f) if training: if min_points_in_gt > 0: # points_count_rbbox takes 10ms with 10 sweeps nuscenes data point_counts = box_np_ops.points_count_rbbox( points, gt_dict["gt_boxes"]) mask = point_counts >= min_points_in_gt _dict_select(gt_dict, mask) # [352, 400] t1 = time.time() if not multi_gpu: res = voxel_generator.generate(points, max_voxels) voxels = res["voxels"] coordinates = res["coordinates"] num_points = res["num_points_per_voxel"] num_voxels = np.array([voxels.shape[0]], dtype=np.int64) else: res = voxel_generator.generate_multi_gpu(points, max_voxels) voxels = res["voxels"] coordinates = res["coordinates"] num_points = res["num_points_per_voxel"] num_voxels = np.array([res["voxel_num"]], dtype=np.int64) metrics["voxel_gene_time"] = time.time() - t1 example = { 'voxels': voxels, # 'visibility': visibility, 'logodds': logodds, 'num_points': num_points, 'coordinates': coordinates, "num_voxels": num_voxels, "metrics": metrics, } if calib is not None: example["calib"] = calib feature_map_size = grid_size[:2] // out_size_factor feature_map_size = [*feature_map_size, 1][::-1] # print(f'feature_map_size in prep_pointcloud(): {feature_map_size}') if anchor_cache is not None: # print('having anchor cache') anchors = anchor_cache["anchors"] anchors_bv = anchor_cache["anchors_bv"] anchors_dict = anchor_cache["anchors_dict"] matched_thresholds = anchor_cache["matched_thresholds"] unmatched_thresholds = anchor_cache["unmatched_thresholds"] else: # print('NOT having anchor cache') ret = target_assigner.generate_anchors(feature_map_size) anchors = ret["anchors"] anchors = anchors.reshape([-1, target_assigner.box_ndim]) anchors_dict = target_assigner.generate_anchors_dict(feature_map_size) anchors_bv = box_np_ops.rbbox2d_to_near_bbox(anchors[:, [0, 1, 3, 4, 6]]) matched_thresholds = ret["matched_thresholds"] unmatched_thresholds = ret["unmatched_thresholds"] # print(f'anchors.shape: {anchors.shape}') example["anchors"] = anchors anchors_mask = None if anchor_area_threshold >= 0: # slow with high resolution. recommend disable this forever. coors = coordinates dense_voxel_map = box_np_ops.sparse_sum_for_anchors_mask( coors, tuple(grid_size[::-1][1:])) dense_voxel_map = dense_voxel_map.cumsum(0) dense_voxel_map = dense_voxel_map.cumsum(1) anchors_area = box_np_ops.fused_get_anchors_area( dense_voxel_map, anchors_bv, voxel_size, pc_range, grid_size) anchors_mask = anchors_area > anchor_area_threshold # example['anchors_mask'] = anchors_mask.astype(np.uint8) example['anchors_mask'] = anchors_mask # print("prep time", time.time() - t) metrics["prep_time"] = time.time() - t if not training: return example example["gt_names"] = gt_dict["gt_names"] # voxel_labels = box_np_ops.assign_label_to_voxel(gt_boxes, coordinates, # voxel_size, coors_range) if create_targets: t1 = time.time() targets_dict = target_assigner.assign( anchors, anchors_dict, gt_dict["gt_boxes"], anchors_mask, gt_classes=gt_dict["gt_classes"], gt_names=gt_dict["gt_names"], matched_thresholds=matched_thresholds, unmatched_thresholds=unmatched_thresholds, importance=gt_dict["gt_importance"]) """ boxes_lidar = gt_dict["gt_boxes"] bev_map = simplevis.nuscene_vis(points, boxes_lidar, gt_dict["gt_names"]) assigned_anchors = anchors[targets_dict['labels'] > 0] ignored_anchors = anchors[targets_dict['labels'] == -1] bev_map = simplevis.draw_box_in_bev(bev_map, [-50, -50, 3, 50, 50, 1], ignored_anchors, [128, 128, 128], 2) bev_map = simplevis.draw_box_in_bev(bev_map, [-50, -50, 3, 50, 50, 1], assigned_anchors, [255, 0, 0]) cv2.imshow('anchors', bev_map) cv2.waitKey(0) boxes_lidar = gt_dict["gt_boxes"] pp_map = np.zeros(grid_size[:2], dtype=np.float32) voxels_max = np.max(voxels[:, :, 2], axis=1, keepdims=False) voxels_min = np.min(voxels[:, :, 2], axis=1, keepdims=False) voxels_height = voxels_max - voxels_min voxels_height = np.minimum(voxels_height, 4) # sns.distplot(voxels_height) # plt.show() pp_map[coordinates[:, 1], coordinates[:, 2]] = voxels_height / 4 pp_map = (pp_map * 255).astype(np.uint8) pp_map = cv2.cvtColor(pp_map, cv2.COLOR_GRAY2RGB) pp_map = simplevis.draw_box_in_bev(pp_map, [-50, -50, 3, 50, 50, 1], boxes_lidar, [128, 0, 128], 1) cv2.imshow('heights', pp_map) cv2.waitKey(0) """ example.update({ 'labels': targets_dict['labels'], 'reg_targets': targets_dict['bbox_targets'], # 'reg_weights': targets_dict['bbox_outside_weights'], 'importance': targets_dict['importance'], }) return example
def prep_pointcloud(input_dict, root_path, voxel_generator, target_assigner, db_sampler=None, max_voxels=70000, class_names=['Car'], remove_outside_points=False, training=True, create_targets=True, shuffle_points=False, reduce_valid_area=False, remove_unknown=False, gt_rotation_noise=[-np.pi / 3, np.pi / 3], gt_loc_noise_std=[1.0, 1.0, 1.0], global_rotation_noise=[-np.pi / 4, np.pi / 4], global_scaling_noise=[0.95, 1.05], global_loc_noise_std=(0.2, 0.2, 0.2), global_random_rot_range=[0.78, 2.35], generate_bev=False, without_reflectivity=False, num_point_features=4, anchor_area_threshold=1, gt_points_drop=0.0, gt_drop_max_keep=10, remove_points_after_sample=True, anchor_cache=None, remove_environment=False, random_crop=False, reference_detections=None, add_rgb_to_points=False, lidar_input=False, unlabeled_db_sampler=None, out_size_factor=2, min_gt_point_dict=None, bev_only=False, use_group_id=False, out_dtype=np.float32): """convert point cloud to voxels, create targets if ground truths exists. """ points = input_dict["lidar"]["points"] if training: gt_boxes = input_dict['lidar']['annotations']["gt_boxes"] gt_names = input_dict['lidar']['annotations']["gt_names"] # difficulty = input_dict["difficulty"] group_ids = None if use_group_id and "group_ids" in input_dict: group_ids = input_dict["group_ids"] # rect = input_dict["rect"] # Trv2c = input_dict["Trv2c"] # P2 = input_dict["P2"] unlabeled_training = unlabeled_db_sampler is not None calib = None # print(gt_dict) # print("+++++++++++++++1111111+++++++++++++++") if "calib" in input_dict: calib = input_dict["calib"] if remove_environment is True and training: selected = kitti.keep_arrays_by_name(gt_names, class_names) gt_boxes = gt_boxes[selected] gt_names = gt_names[selected] # difficulty = difficulty[selected] if group_ids is not None: group_ids = group_ids[selected] points = prep.remove_points_outside_boxes(points, gt_boxes) if training: # print(gt_names) selected = kitti.drop_arrays_by_name(gt_names, ["DontCare"]) gt_boxes = gt_boxes[selected] gt_names = gt_names[selected] # difficulty = difficulty[selected] if group_ids is not None: group_ids = group_ids[selected] # if remove_unknown: # remove_mask = difficulty == -1 # """ # gt_boxes_remove = gt_boxes[remove_mask] # gt_boxes_remove[:, 3:6] += 0.25 # points = prep.remove_points_in_boxes(points, gt_boxes_remove) # """ # keep_mask = np.logical_not(remove_mask) # gt_boxes = gt_boxes[keep_mask] # gt_names = gt_names[keep_mask] # difficulty = difficulty[keep_mask] # if group_ids is not None: # group_ids = group_ids[keep_mask] gt_boxes_mask = np.array( [n in class_names for n in gt_names], dtype=np.bool_) if db_sampler is not None: sampled_dict = db_sampler.sample_all( root_path, gt_boxes, gt_names, num_point_features, random_crop, gt_group_ids=group_ids ) if sampled_dict is not None: sampled_gt_names = sampled_dict["gt_names"] sampled_gt_boxes = sampled_dict["gt_boxes"] sampled_points = sampled_dict["points"] sampled_gt_masks = sampled_dict["gt_masks"] # gt_names = gt_names[gt_boxes_mask].tolist() gt_names = np.concatenate([gt_names, sampled_gt_names], axis=0) # gt_names += [s["name"] for s in sampled] gt_boxes = np.concatenate([gt_boxes, sampled_gt_boxes]) gt_boxes_mask = np.concatenate( [gt_boxes_mask, sampled_gt_masks], axis=0) if group_ids is not None: sampled_group_ids = sampled_dict["group_ids"] group_ids = np.concatenate([group_ids, sampled_group_ids]) if remove_points_after_sample: points = prep.remove_points_in_boxes( points, sampled_gt_boxes) points = np.concatenate([sampled_points, points], axis=0) # unlabeled_mask = np.zeros((gt_boxes.shape[0], ), dtype=np.bool_) if without_reflectivity: used_point_axes = list(range(num_point_features)) used_point_axes.pop(3) points = points[:, used_point_axes] pc_range = voxel_generator.point_cloud_range if bev_only: # set z and h to limits gt_boxes[:, 2] = pc_range[2] gt_boxes[:, 5] = pc_range[5] - pc_range[2] prep.noise_per_object_v3_( gt_boxes, points, gt_boxes_mask, rotation_perturb=gt_rotation_noise, center_noise_std=gt_loc_noise_std, global_random_rot_range=global_random_rot_range, group_ids=group_ids, num_try=100) # should remove unrelated objects after noise per object gt_boxes = gt_boxes[gt_boxes_mask] gt_names = gt_names[gt_boxes_mask] if group_ids is not None: group_ids = group_ids[gt_boxes_mask] gt_classes = np.array( [class_names.index(n) + 1 for n in gt_names], dtype=np.int32) gt_boxes, points = prep.random_flip(gt_boxes, points) gt_boxes, points = prep.global_rotation( gt_boxes, points, rotation=global_rotation_noise) gt_boxes, points = prep.global_scaling_v2(gt_boxes, points, *global_scaling_noise) # Global translation gt_boxes, points = prep.global_translate(gt_boxes, points, global_loc_noise_std) bv_range = voxel_generator.point_cloud_range[[0, 1, 3, 4]] mask = prep.filter_gt_box_outside_range(gt_boxes, bv_range) gt_boxes = gt_boxes[mask] gt_classes = gt_classes[mask] if group_ids is not None: group_ids = group_ids[mask] # limit rad to [-pi, pi] gt_boxes[:, 6] = box_np_ops.limit_period( gt_boxes[:, 6], offset=0.5, period=2 * np.pi) if shuffle_points: # shuffle is a little slow. np.random.shuffle(points) # [0, -40, -3, 70.4, 40, 1] voxel_size = voxel_generator.voxel_size pc_range = voxel_generator.point_cloud_range grid_size = voxel_generator.grid_size # [352, 400] # max_voxels: maximum number of voxels voxels, coordinates, num_points = voxel_generator.generate( points, max_voxels) example = { 'voxels': voxels, 'num_points': num_points, 'coordinates': coordinates, "num_voxels": np.array([voxels.shape[0]], dtype=np.int64), } # if not lidar_input: feature_map_size = grid_size[:2] // out_size_factor feature_map_size = [*feature_map_size, 1][::-1] if anchor_cache is not None: anchors = anchor_cache["anchors"] anchors_bv = anchor_cache["anchors_bv"] matched_thresholds = anchor_cache["matched_thresholds"] unmatched_thresholds = anchor_cache["unmatched_thresholds"] else: ret = target_assigner.generate_anchors(feature_map_size) anchors = ret["anchors"] anchors = anchors.reshape([-1, 7]) matched_thresholds = ret["matched_thresholds"] unmatched_thresholds = ret["unmatched_thresholds"] anchors_bv = box_np_ops.rbbox2d_to_near_bbox( anchors[:, [0, 1, 3, 4, 6]]) example["anchors"] = anchors # print("debug", anchors.shape, matched_thresholds.shape) # anchors_bv = anchors_bv.reshape([-1, 4]) anchors_mask = None if anchor_area_threshold >= 0: coors = coordinates dense_voxel_map = box_np_ops.sparse_sum_for_anchors_mask( coors, tuple(grid_size[::-1][1:])) dense_voxel_map = dense_voxel_map.cumsum(0) dense_voxel_map = dense_voxel_map.cumsum(1) anchors_area = box_np_ops.fused_get_anchors_area( dense_voxel_map, anchors_bv, voxel_size, pc_range, grid_size) anchors_mask = anchors_area > anchor_area_threshold # example['anchors_mask'] = anchors_mask.astype(np.uint8) example['anchors_mask'] = anchors_mask if generate_bev: bev_vxsize = voxel_size.copy() bev_vxsize[:2] /= 2 bev_vxsize[2] *= 2 bev_map = points_to_bev(points, bev_vxsize, pc_range, without_reflectivity) example["bev_map"] = bev_map if not training: return example if create_targets: targets_dict = target_assigner.assign( anchors, gt_boxes, anchors_mask, gt_classes=gt_classes, matched_thresholds=matched_thresholds, unmatched_thresholds=unmatched_thresholds) example.update({ 'labels': targets_dict['labels'], 'reg_targets': targets_dict['bbox_targets'], 'reg_weights': targets_dict['bbox_outside_weights'], }) example["points"]= input_dict["lidar"]["points"] example['gt_boxes'] = input_dict['lidar']['annotations']["gt_boxes"] example['gt_names'] = input_dict['lidar']['annotations']["gt_names"] return example