def vis_fn_kitti(data_dir,
idx,
detection,
x_range=(-35.2, 35.2),
y_range=(-40, 40),
grid_size=(0.1, 0.1)):
itm = detection
output_dict = {
"calib": True,
"image": False,
"label": True,
"velodyne": True
}
if "nusc" in data_dir:
calib, _, label, pc = KittiData(
data_dir, idx, output_dict=output_dict).read_data(num_feature=5)
else:
calib, _, label, pc = KittiData(data_dir, idx,
output_dict=output_dict).read_data()
bevimg = BEVImage(x_range, y_range, grid_size)
bevimg.from_lidar(pc)
for obj in label.data:
bevimg.draw_box(obj, calib, bool_gt=True)
box3d_lidar = itm["box3d_lidar"]
score = itm["scores"]
for box3d_lidar_, score_ in zip(box3d_lidar, score):
x, y, z, w, l, h, ry = box3d_lidar_
obj = KittiObj()
bcenter_Flidar = np.array([x, y, z]).reshape(1, -1)
bcenter_Fcam = calib.lidar2leftcam(bcenter_Flidar)
obj.x, obj.y, obj.z = bcenter_Fcam.flatten()
obj.w, obj.l, obj.h = w, l, h
obj.ry = ry
bevimg.draw_box(obj, calib, bool_gt=False, width=2)
return bevimg
def create_db_file_kitti_wk_fn(idx):
global g_data_dir, g_dbinfos
root_dir = g_data_dir
tag = idx
output_dict = {
"calib": True,
"image": False,
"label": True,
"velodyne": True
}
calib, _, label, pc_Flidar = KittiData(
root_dir=str(root_dir), idx=tag, output_dict=output_dict).read_data()
for i, obj in enumerate(label.data):
name = obj.type
gt_idx = i
gtpc_path = str(root_dir.parent / "gt_database" /
f"{tag}_{name}_{gt_idx}.bin")
gtpc_idx = obj.get_pts_idx(pc_Flidar[:, :3], calib)
gtpc = pc_Flidar[gtpc_idx, :]
box3d_cam = obj
num_points_in_gt = gtpc.shape[0]
write_pc_to_file(gtpc, gtpc_path)
obj_info = {
"name": name,
"gtpc_path": gtpc_path,
"tag": tag,
"gt_idx": gt_idx,
"box3d_cam": box3d_cam,
"num_points_in_gt": num_points_in_gt,
"calib": calib
}
g_dbinfos.append(obj_info)
def create_info_file_kitti_wk_fn(idx):
global g_data_dir, g_infos
root_dir = g_data_dir
infos = g_infos
output_dict = {
"calib": True,
"image": False,
"label": True,
"velodyne": True
}
if root_dir.stem == "testing":
output_dict[
'label'] = False # testing folder do not have label_2/ data.
info = dict()
tag = idx
pc_path = str(root_dir / "velodyne" / f"{tag}.bin")
reduced_pc_path = str(root_dir / "reduced_velodyne" / f"{tag}.bin")
img_path = str(root_dir / "image_2" / f"{tag}.png")
calib, _, label, pc_Flidar = KittiData(
root_dir=str(root_dir), idx=tag, output_dict=output_dict).read_data()
mask = reduce_pc(pc_Flidar, calib)
pc_reduced_Flidar = pc_Flidar[mask, :]
with open(reduced_pc_path, 'wb') as f:
pc_reduced_Flidar.tofile(f)
info["tag"] = tag
info["pc_path"] = pc_path
info["reduced_pc_path"] = reduced_pc_path
info["img_path"] = img_path
info["calib"] = calib
info["label"] = label
infos.append(info)
def vis_fn(idx):
global dataset, data_dir, output_dir
if dataset == "KITTI":
from det3.dataloader.kittidata import KittiData
calib, img, label, pc = KittiData(data_dir, idx).read_data()
elif dataset == "CARLA":
from det3.dataloader.carladata import CarlaData
import numpy as np
pc_dict, label, calib = CarlaData(data_dir, idx).read_data()
pc = np.vstack([
calib.lidar2imu(v, key="Tr_imu_to_{}".format(k))
for k, v in pc_dict.items()
])
elif dataset == "UDI":
from det3.dataloader.udidata import UdiData, UdiFrame
import numpy as np
udidata = UdiData(data_dir, idx).read_data()
calib = udidata["calib"]
label = udidata["label"]
pc_dict = udidata["lidar"]
for k, v in pc_dict.items():
pc_dict[k] = calib.transform(v[:, :3],
source_frame=UdiFrame(
UdiData.lidar_to_frame(k)),
target_frame=UdiFrame("BASE"))
pc_merge = np.vstack([v for k, v in pc_dict.items()])
bevimg = BEVImage(x_range=(-70, 70),
y_range=(-40, 40),
grid_size=(0.05, 0.05))
bevimg.from_lidar(pc_merge, scale=1)
for obj in label:
bevimg.draw_box(obj, calib, bool_gt=True)
bevimg_img = Image.fromarray(bevimg.data)
bevimg_img.save(os.path.join(output_dir, idx + '.png'))
for k, v in pc_dict.items():
bevimg = BEVImage(x_range=(-70, 70),
y_range=(-40, 40),
grid_size=(0.05, 0.05))
bevimg.from_lidar(v, scale=1)
for obj in label:
bevimg.draw_box(obj, calib, bool_gt=True)
bevimg_img = Image.fromarray(bevimg.data)
bevimg_img.save(os.path.join(output_dir, idx + f'_{k}.png'))
fv_img = FVImage()
vcam_T = np.eye(4)
vcam_T[:3, 3] = np.array([-3, 0, 3])
vcam_T[:3, :3] = rotz(0) @ roty(np.pi * 0.1)
calib.vcam_T = vcam_T
fv_img.from_lidar(calib, pc_merge, scale=2)
for obj in label:
fv_img.draw_3dbox(obj, calib, bool_gt=True)
fv_img.save(os.path.join(output_dir, idx + '_fv.png'))
return
bevimg = BEVImage(x_range=(0, 70),
y_range=(-40, 40),
grid_size=(0.05, 0.05))
bevimg.from_lidar(pc, scale=1)
for obj in label.read_label_file().data:
bevimg.draw_box(obj, calib, bool_gt=True)
bevimg_img = Image.fromarray(bevimg.data)
bevimg_img.save(os.path.join(output_dir, idx + '.png'))
def main(is_val, data_dir, vis_dir, result_path):
dt_data = pickle.load(open(result_path, 'rb'))
save_dir = vis_dir
if is_val:
for dt in tqdm(dt_data):
idx = dt['metadata']['image_idx']
idx = "{:06d}".format(idx)
calib, img, label, pc = KittiData(root_dir=data_dir,
idx=idx).read_data()
label_est = KittiLabel()
label_est.data = []
num_dt = len(dt['name'])
for i in range(num_dt):
obj = KittiObj()
obj.type = dt['name'][i]
obj.truncated = dt['truncated'][i]
obj.occluded = dt['occluded'][i]
obj.alpha = dt['alpha'][i]
obj.bbox_l, obj.bbox_t, obj.bbox_r, obj.bbox_b = dt['bbox'][i]
obj.l, obj.h, obj.w = dt['dimensions'][i]
obj.x, obj.y, obj.z = dt['location'][i]
obj.ry = dt['rotation_y'][i]
obj.score = dt['score'][i]
label_est.data.append(obj)
fvimg = FVImage()
fvimg.from_image(img)
for obj in label.data:
if (obj.x > obj.z) or (-obj.x > obj.z):
continue
if obj.type in ['Car', 'Van']:
fvimg.draw_3dbox(obj, calib, bool_gt=True, width=3)
for obj in label_est.data:
if (obj.x > obj.z) or (-obj.x > obj.z):
continue
fvimg.draw_3dbox(obj,
calib,
bool_gt=False,
width=2,
c=(255, 255, int(255 * obj.score), 255))
fvimg_img = Image.fromarray(fvimg.data)
fvimg_img.save(os.path.join(save_dir, "fv_{}.png".format(idx)))
else:
pc_dir = os.path.join(data_dir, "velodyne_points", "data")
img2_dir = os.path.join(data_dir, "image_02", "data")
calib_dir = os.path.join(data_dir, "calib")
calib = read_calib(calib_dir)
for dt in tqdm(dt_data):
idx = dt['metadata']['image_idx']
idx = "{:010d}".format(idx)
pc = read_pc_from_bin(os.path.join(pc_dir, idx + ".bin"))
img = read_image(os.path.join(img2_dir, idx + ".png"))
label_est = KittiLabel()
label_est.data = []
num_dt = len(dt['name'])
for i in range(num_dt):
obj = KittiObj()
obj.type = dt['name'][i]
obj.truncated = dt['truncated'][i]
obj.occluded = dt['occluded'][i]
obj.alpha = dt['alpha'][i]
obj.bbox_l, obj.bbox_t, obj.bbox_r, obj.bbox_b = dt['bbox'][i]
obj.l, obj.h, obj.w = dt['dimensions'][i]
obj.x, obj.y, obj.z = dt['location'][i]
obj.ry = dt['rotation_y'][i]
obj.score = dt['score'][i]
label_est.data.append(obj)
fvimg = FVImage()
fvimg.from_image(img)
# for obj in label.data:
# if obj.type in ['Car', 'Van']:
# fvimg.draw_3dbox(obj, calib, bool_gt=True, width=3)
for obj in label_est.data:
if (obj.x > obj.z) or (-obj.x > obj.z):
continue
fvimg.draw_3dbox(obj,
calib,
bool_gt=False,
width=2,
c=(255, 255, int(obj.score * 255), 255))
fvimg_img = Image.fromarray(fvimg.data)
fvimg_img.save(os.path.join(save_dir, "fv_{}.png".format(idx)))