which sequences did you use in KITTY dataset?

Question

which sequences did you use in KITTY dataset?

aiyaqing opened this issue 2 years ago · comments

hello,i try to run you code but i found the TE is 79.43, i use you fpfh data.
I want to use initial kitty dataset, what can i do?

ZhiChen · Answer 1 · Wed May 25 2022 16:16:37 GMT+0800 (China Standard Time)

I retest the result, and it is normal. How do you test the results? Could show me the results you test?

ZhiChen · Answer 2 · Fri May 27 2022 21:06:15 GMT+0800 (China Standard Time)

When you test on the KITTI dataset using fpfh descriptor, you should change the "data_path" in the json file to the path of fpfh dataset, and change the "descriptor" as fpfh. I don't know if you have a wrong configuration here.

ixiaoyaqing · Answer 3 · Mon May 30 2022 19:35:43 GMT+0800 (China Standard Time)

I am sorry that I did not reply because I was busy. Yes, I adjusted the parameters, and the TE of the code I ran was between 60 and 80. Now I'm writing a visual display of the point cloud after registration.

ZhiChen · Answer 4 · Tue May 31 2022 20:45:57 GMT+0800 (China Standard Time)

The printed information is the error of the mis-estimated matching pairs, not the whole error. It means only 64-th and 377-th point cloud pairs are wrongly estimated. The error of the whole dataset is printed at the end of the code. Since you use the cpu-version, which has no attribute 'max_memory_allocate', the following program is not running. You can comment the 'max_memory_allocate' and run it again.