Error happened while running 00 sequence in the KITTI dataset

[xiangfuli]

Hello Ruan, I have to say it's a nice work with beautiful codes.

I've successfully compiled and run the code using the provided launch file. However, as the times goes, when the program deals with the LiDAR data indexed in 979, error always happened.

The message in system log shows that segfault at 8 ... error 4 in slamesh. Do you have any clue why this error happened?

Thanks.

[RuanJY]

Would you please tell me the memory size your PC has, which sequence you are running, and whether you can replicate the problem if you use slamesh_kitti_meshing.launch?

[xiangfuli]

Yes. I was running the mesh script. Every time the process stopped when dealt with the 979th frame in 00 sequence.

The memory size is 32GB and the memory each process can use was set to be unlimited.

[xiangfuli]

I can also reproduce the error in another computer.

[RuanJY]

Please give me a screenshot of the Terminal when you launch the node and when the error appears. Besides, can you run slamesh_kitti_odometry.launch in the same dataset? If it works well but slamesh_kitti_meshing.launch fails, the problem may due to the mesh msg.

[xiangfuli]

The odometry.launch also stops at the frame index 979. Below is the console output and system log output.

[RuanJY]

Because the node reads the point cloud file succesfully but filters all points, the reason may be your file was broken and all points are NaN point.

[JACKLiuDay]

Because the node reads the point cloud file succesfully but filters all points, the reason may be your file was broken and all points are NaN point. RUAN, Jianyuan @. … ---- Replied Message ---- From Xiangfu @.> Date 8/29/2023 02:17 To @.> Cc @.> , @.> Subject Re: [RuanJY/SLAMesh] Error happened while running 00 sequence in the KITTI dataset (Issue #7) The odometry.launch also stops at the frame index 979. Below is the console output and system log output. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.> Disclaimer: This message (including any attachments) contains confidential information intended for a specific individual and purpose. If you are not the intended recipient, you should delete this message and notify the sender and The Hong Kong Polytechnic University (the University) immediately. Any disclosure, copying, or distribution of this message, or the taking of any action based on it, is strictly prohibited and may be unlawful. The University specifically denies any responsibility for the accuracy or quality of information obtained through University E-mail Facilities. Any views and opinions expressed are only those of the author(s) and do not necessarily represent those of the University and the University accepts no liability whatsoever for any losses or damages incurred or caused to any party as a result of the use of such information.

Hi, bro. I want to make the mesh grid to 0.05cm, which param should I change. Could you give me some some advice.

[RuanJY]

For example, If grid=1.6m, and num_test=6, and full_cover=true, then mesh grid is 1.6/(6-1)=0.32m You can adjust one of them or both of them, for example, only adjust num_test=11, then mesh gird is 0.16m, the time-cost should increase. In indoor scenario, you can also adjust the grid to a smaller one, like 1m, 0.6m

[JACKLiuDay]

For example, If grid=1.6m, and num_test=6, and full_cover=true, then mesh grid is 1.6/(6-1)=0.32m You can adjust one of them or both of them, for example, only adjust num_test=11, then mesh gird is 0.16m, the time-cost should increase. In indoor scenario, you can also adjust the grid to a smaller one, like 1m, 0.6m

Hi, thank you for your kind reply. I will try it with SLAMesh.

[RuanJY]

Added a warning when there is no valid points.