-
Notifications
You must be signed in to change notification settings - Fork 20
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
how to visualize the experiments results #21
Comments
We have not released the visualization code because the toolkit is quite complicated to setup. The box center have two formats, i.e., nuscenes format and KITTI format. They are converted here. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hello, I have successfully run the code and I have two questions. First, I'd like to know how to visualize the experiments results and how can I render the nuscenes dataset scene like the pictures shown in your paper, and I would be appreciate it if you could provide the corresponding code or a tutorial! Second, I'm trying to project the track 3d-box on the image but there's something wrong.
As you can see, the coordinates of the boxes on the y-axis are offset downwards as a whole.
boxes_3d = predictor_utils.to_numpy(results[0]['boxes_3d'].tensor) # lidar coordinate bboxes_3d = img_metas[0]['box_type_3d'][0](boxes_3d, 9) # convert to <class LiDARInstance3DBoxes> corners = bboxes_3d.corners # n * 8 * 3 for cam_idx in range(6): lidar_to_image = lidar2img[0][cam_idx] image_idx = images[cam_idx] for i in range(len(bboxes_3d)): corners_lidar = corners[i].squeeze(0).cpu().numpy() # (8, 3) corners_lidar = np.concatenate([corners_lidar, np.ones((corners_lidar.shape[0], 1))], axis = 1) # 8 x 4 corners_image = corners_lidar @ lidar_to_image.T # torch.matmul(lidar_to_image, corners_lidar) # corners_image[:, :2] /= corners_image[:, [2]] # x, y / z if np.all((corners_image[:, 2] > 0) == True) == True: track_id = track_ids[i] track_cls = class_names[track_labels[i]] # ['car', 'truck', ...] color = colors[track_labels[i]] # [(0, 255, 255), ...] # 画3d_track_instance image_idx = draw_cube_on_image(image_idx, corners_image[:, :2], color) corners_image = corners_image[:, :2].astype(int)
As the results 'boxes_3d' are in lidar coordinate, I jusr convert them to the and return their corners. Then I project the corners using array 'lidar2img' on the raw images. Could you help me find the error? Thanks!
The text was updated successfully, but these errors were encountered: