Skip to content

Pointnet based lidar odometry prediction in 2D space

License

Notifications You must be signed in to change notification settings

cdbharath/odom-pointnet

Repository files navigation

Pointnet based lidar odometry prediction in 2D space

Odometry prediction with KITTI Lidar dataset

Image 1 Image 2

This work implements the following steps

  1. Extract lidar scans
  2. Convert to depth image with spherical projection
  3. KD-Tree based depth completion
  4. Find feature matches with ORB features
  5. Reproject back to euclidean space
  6. PointNet based architecture for realtime odometry prediction

Results

Image 1 Image 2
Graph in the left shows the odometry prediction of the point-based model relative to ground truth in KITTI city 0. The graph in the right shows the performance for ICP for the same city.

Reference

  1. LodoNet: A Deep Neural Network with 2D Keypoint Matchingfor 3D LiDAR Odometry Estimation

About

Pointnet based lidar odometry prediction in 2D space

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published