search

koide3 / fast_gicp

A collection of GICP-based fast point cloud registration algorithms
BSD 3-Clause "New" or "Revised" License
1.27k stars 318 forks source link

compare with the ndt_omp #38

Open sintax2con opened 3 years ago

sintax2con commented 3 years ago

thanks for your excellent work, i wonder the comparison of fast-gicp and ndt-omp, speed and accuracy. will it replace the ndt method?

koide3 commented 3 years ago

In short, fast_gicp is much accurate but a little bit slower than ndt_omp. Please take a look at our paper for details: https://easychair.org/publications/preprint/ftvV

sintax2con commented 3 years ago

Thanks for reply, fast_gicp is truly more accurate than ndtomp. But it may stuck occasionly, seems in compute jacobian. Do you have any idea? image image

koide3 commented 3 years ago

I've not seen this problem before. Can you provide some data for reproducing the problem?

sintax2con commented 3 years ago

https://github.com/sintax2con/fast_gicp_test Thanks for reply!I Uploaded the test file. Please input the same cloud as source and target. It transforms the source cloud with some certain matrix, and evaluates the accuracy by distance and angle error.

koide3 commented 3 years ago

Do you mean running like this? With the following line, I couldn't reproduce the problem: rosrun fast_gicp gicp_align 1.pcd 1.pcd

sintax2con commented 3 years ago

It seems so strange. I test on my laptop and pc and found the same problem. I test gicp and ndt separately , it works fine. Maybe some bug in my system inviroment. Thanks for your reply!