A simple object detection module for use with the Xtion3 camera
Functions from point cloud library http://pointclouds.org/ are used to perform:
- Voxel Grid Filtering of scene
- Plane segmentation using RANSAC model, and removing these inliers from scene's point cloud
- Euclidean Clustering to group points belonging to each object
- Computing centroids of clusters to detect object location
This perception module is part of my master thesis on a reinforcement learning problem in robotics, details can be found here https://github.com/d-misra/Master-Thesis