AtsushiSakai/PythonRobotics

Add VSLAM or 3D lidar SLAM algorithms

Opened this issue · 6 comments

devyani-code commented

Visual SLAM and 3D Lidar SLAM algorithms are widely used in the industry. Concepts like visual odometry, object tracking using 3D lidars, etc., can be helpful to many people.

AtsushiSakai commented

I think it’s a good idea. However, it might be difficult to implement sample code for 3D LIDAR or Visual SLAM without using any external libraries... But PR is welcome.

kartikmadhira1 commented

I think adding core concepts in 3D computer vision first would be a good starting point - Stereo, E and F matrices, PnP, Triangulation and Bundle Adjustment without libraries. These can then be re-arranged for a bigger VSLAM module.

  • 👍1
devyani-code commented

Okay! Will start with Camera geometery basics and stereo vision.

2black0 commented

image detector, descriptor, and matching is the starting point of visual odometry and visual slam, even now some deep learning method are detector-free.

  • 👍1
aviralgarg05 commented

I would like to solve the issue for hacktoberfest

bishwa9 commented

Are we talking about a specific VSLAM algorithm to be implemented here. There are multiple parts to it, and since there's interest, it could be split up as well.

Also, unsure how testing is usually done in his repo for algorithms, but testing a VSLAM would require data. Would we be pulling external datasets?