search

Zhefan-Xu / onboard_detector

[IEEE RA-L'24] Dynamic Obstacle Detection and Tracking (DODT) algorithm for Autonomous Robots (C++/ROS)
MIT License
120 stars 15 forks source link

no detect result #2

Open wtq20230227 opened 7 months ago

wtq20230227 commented 7 months ago

dynamitic_detect

2024-04-12 16-15-14屏幕截图 2024-04-12 16-13-58屏幕截图 I use the camera when D455. After I run the dynamic obstacle detection program, there are all topics that should be subscribed to and published displays in rqt_raph, but the running results are not displayed, and there is a problem with the display of the terminal running dynamic obstacle detection, how can I solve this problem?

Zhefan-Xu commented 7 months ago

Are you able to run with the provided rosbag?

wtq20230227 commented 7 months ago

Are you able to run with the provided rosbag?

yes，I can run the provided rosbag, How can I solve this problem?

Zhefan-Xu commented 7 months ago

I would suggest to check your topic data by comparing with the rosbag. It seems that you don't have a local position?

wtq20230227 commented 7 months ago

I would suggest to check your topic data by comparing with the rosbag. It seems that you don't have a local position?

I solved the problem that the odometer had no information. He can run. But always after running for a while, the picture in rviz gets stuck, and the terminal shows that detector_node-1 process has died DD_failed

wtq20230227 commented 7 months ago

I would suggest to check your topic data by comparing with the rosbag. It seems that you don't have a local position?

I have solved the problem that "the process has died". The reason is that the size of depth image produced by D455 is different from the size produced by D435i. But now ,I meet a new problem. The image ,depth image, U-depth image and bird-eye view are shown in rviz, but there is no box plot containing the point cloud. I don’t know the reason. Can you tell me what information I should provide and where to investigate next? DD_no_pointcloud- 2024-04-14 09_55_00

Zhefan-Xu commented 7 months ago

The pointcloud shows the dynamic objects. Those results shown on the screen are called candidates which means they are not necessarily dynamic.

wtq20230227 commented 5 months ago

The pointcloud shows the dynamic objects. Those results shown on the screen are called candidates which means they are not necessarily dynamic.

In my test scenario, there is a moving person who walks at a speed greater than the threshold of 0.15 in the parameter list, but there are still no boxes and point clouds in the rviz. My testing environment is outdoors, will this have an impact? If there is no impact, where might the problem be? Can you help me analsy the reason?