Dear Alexander,
Recently, I have been working with your code, "Long-Term Urban Vehicle Localization Using Pole Landmarks Extracted from 3-D Lidar Scans." While running python ncltpole.py, I visualized the pole detection results and noticed that there are a considerable number of redundant and inaccurate bounding boxes.
Where do you think the issue might be? Do you have any suggestions for improving the results?
Looking forward to hearing from you. Thank you so much!
Best regards,
Dear Alexander,
Recently, I have been working with your code, "Long-Term Urban Vehicle Localization Using Pole Landmarks Extracted from 3-D Lidar Scans." While running python ncltpole.py, I visualized the pole detection results and noticed that there are a considerable number of redundant and inaccurate bounding boxes.
Where do you think the issue might be? Do you have any suggestions for improving the results?
Looking forward to hearing from you. Thank you so much!
Best regards,