3D Dense Visual SLAM

Implemented Dense SLAM with point based fusion for 3D localization and mapping. Implemented Projective Iterative Closest Point algorithm for Visual Odometry and glued it with point based fusion for Simultaneous Localization and Mapping.

ScreenCapture_2022-04-15-11-19-03_edited.jpg

Screenshot new.png

ScreenCapture_2022-04-15-11-20-08_edited.jpg

ScreenCapture_2022-04-15-11-19-03_edited.jpg

In this algorithm projective data association is implemented where 3D point cloud is projected on to a 2D plane and projective nearest neighbors are considered to be related. Based on this data association, Rotation and Translation matrices are calculated to get the camera motion transform. The whole procedure is iterated until convergence.

For Mapping we project the points based on the estimated camera pose and after filtering we calculate the weighted averages of properties like positions, normals and colors.

3D Dense Visual SLAM

​

For Mapping we project the points based on the estimated camera pose and after filtering we calculate the weighted averages of properties like positions, normals and colors.

​