def add_difficulty_to_annos_v2(info):
v_path = info['velodyne_path']
num_features = info['pointcloud_num_features']
points = np.fromfile(os.path.join(data_root, v_path),
dtype=np.float32,
count=-1).reshape([-1, num_features])
annos = info['annos']
dims = annos['dimensions'] # lwh format
loc = annos['location'] # xyz in camera
rots = annos['rotation_y'] # rad in camera
rect = info['calib/R0_rect']
Trv2c = info['calib/Tr_velo_to_cam']
gt_boxes_lidar = np.concatenate([loc, dims, rots[..., np.newaxis]], axis=1)
gt_boxes = copy.deepcopy(gt_boxes_lidar)
gt_point_table = box_np_ops.points_in_rbbox(points, gt_boxes)
gt_point_count = gt_point_table.sum(axis=0)
annos["num_points"] = gt_point_count
gt_boxes = copy.deepcopy(gt_boxes_lidar)
gt_boxes[:, 3:6] = gt_boxes[:, 3:6] + np.array([1.2, 0.5, 8])
gt_point_table = box_np_ops.points_in_rbbox(points, gt_boxes)
gt_point_count = gt_point_table.sum(axis=0)
annos["difficulty"] = gt_point_count
def save_pointcloud(self, idx):
point_color = self.w_config.get("PointColor")[:3]
point_color = (*point_color, self.w_config.get("PointAlpha"))
point_color = np.tile(np.array(point_color), [self.points.shape[0], 1])
point_size = np.full([self.points.shape[0]], self.w_config.get("PointSize"), dtype=np.float32)
if 'annos' in self.info:
gt_point_table = box_np_ops.points_in_rbbox(self.points, self.gt_boxes)
self.gt_filled_mask = gt_point_table.sum(axis=0)
gt_point_mask = gt_point_table.any(1)
point_size[gt_point_mask] = self.w_config.get("GTPointSize")
gt_point_color = self.w_config.get("GTPointColor")
gt_point_color = (*gt_point_color[:3], self.w_config.get("GTPointAlpha"))
point_color[gt_point_mask] = gt_point_color
range = np.array([self.w_config.get("CoorsRange")])
# range = np.zeros(range.shape, dtype=range.dtype)
# range[0] = [-30, -38.4, -10.5, 51.92, 38.4, 9.5]
bbox = box_np_ops.minmax_to_corner_3d(range)
self.w_pc_viewer.boxes3d("bound", bbox, GLColor.Green)
self.w_pc_viewer.remove("dt_boxes/labels")
self.w_pc_viewer.remove("dt_boxes")
if self.detection_annos is not None and self.w_config.get("DrawDTBoxes"):
detection_anno = self.detection_annos[idx]
self.draw_detection(detection_anno)
self.w_pc_viewer.scatter(
"pointcloud", self.points[:, :3], point_color, size=point_size)
def noise_per_object(gt_boxes,
points=None,
valid_mask=None,
rotation_perturb=(5.0 / 180) * np.pi,
center_noise_std=0.15,
global_random_rot_range=(2.0 / 180) * np.pi,
num_try=100):
num_boxes = gt_boxes.shape[0]
rotation_perturb = [-rotation_perturb, rotation_perturb]
center_noise_std = [center_noise_std, center_noise_std, center_noise_std]
if valid_mask is None:
valid_mask = np.ones((num_boxes, ), dtype=np.bool_)
center_noise_std = np.array(center_noise_std, dtype=gt_boxes.dtype)
loc_noises = np.random.normal(scale=center_noise_std,
size=[num_boxes, num_try, 3])
rot_noises = np.random.uniform(rotation_perturb[0],
rotation_perturb[1],
size=[num_boxes, num_try])
global_rot_noises = np.random.uniform(-global_random_rot_range,
global_random_rot_range,
size=[num_boxes, num_try])
point_masks = box_np_ops.points_in_rbbox(points, gt_boxes)
enable_grot = False
if global_random_rot_range > (0.01 / 180) * np.pi:
enable_grot = True
if not enable_grot:
selected_noise = noise_per_box(gt_boxes[:, [0, 1, 3, 4, 6]],
valid_mask, loc_noises, rot_noises)
else:
selected_noise = noise_per_box_v2_(gt_boxes[:, [0, 1, 3, 4, 6]],
valid_mask, loc_noises, rot_noises,
global_rot_noises)
loc_transforms = _select_transform(loc_noises, selected_noise)
rot_transforms = _select_transform(rot_noises, selected_noise)
points_transform_(points, gt_boxes[:, :3], point_masks, loc_transforms,
rot_transforms, valid_mask)
box3d_transform_(gt_boxes, loc_transforms, rot_transforms, valid_mask)
def plot_pointcloud(self):
point_color = self.w_config.get("PointColor")[:3]
point_color = (*point_color, self.w_config.get("PointAlpha"))
point_color = np.tile(np.array(point_color), [self.points.shape[0], 1])
point_size = np.full(
[self.points.shape[0]],
self.w_config.get("PointSize"),
dtype=np.float32)
self.w_pc_viewer.draw_bounding_box(self.w_config.get("CoorsRange"))
if self.gt_boxes is not None and len(self.gt_boxes) > 0:
gt_boxes = copy.deepcopy(self.gt_boxes)
# gt_boxes[:, 3:6] = gt_boxes[:, 3:6] + np.array([1.2, 0.8, 8])
# gt_boxes[:, 2] = gt_boxes[:, 2] + np.array([1])
gt_point_table = box_np_ops.points_in_rbbox(self.points, gt_boxes)
self.gt_filled_mask = gt_point_table.sum(axis=0)
gt_point_mask = gt_point_table.any(1)
point_size[gt_point_mask] = self.w_config.get("GTPointSize")
gt_point_color = self.w_config.get("GTPointColor")
gt_point_color = (*gt_point_color[:3], self.w_config.get("GTPointAlpha"))
point_color[gt_point_mask] = gt_point_color
self.w_pc_viewer.remove("gt_boxes/labels")
self.w_pc_viewer.remove("gt_boxes")
if self.gt_names is not None and len(self.gt_names) > 0 and self.w_config.get("DrawGTBoxes"):
self.plot_gt_boxes_in_pointcloud()
self.w_pc_viewer.remove("dt_boxes/labels")
self.w_pc_viewer.remove("dt_boxes")
if self.detection_annos is not None and len(self.detection_annos) > 0 and self.w_config.get("DrawDTBoxes"):
detection_anno = self.detection_annos[self.current_idx]
self.draw_detection(detection_anno)
self.w_pc_viewer.remove("anchors")
if self.plot_anchors:
self.plot_anchors_in_pointcloud()
self.w_pc_viewer.scatter(
"pointcloud", self.points[:, :3], point_color, size=point_size)
def remove_points_outside_boxes(points, boxes):
masks = box_np_ops.points_in_rbbox(points, boxes)
points = points[masks.any(-1)]
return points
def remove_points_in_boxes(points, boxes):
masks = box_np_ops.points_in_rbbox(points, boxes)
points = points[np.logical_not(masks.any(-1))]
return points