Developed algorithms to estimate the position of a Point Light Source in a scene using Kinect and Point Cloud Library based on code developed by by Zhongshi (Sam) Jiang and Shayan Rezvankhah Instead of PCL's viewer I used OpenGL for rendering.
Steps:
- Record Point cloud using OpenNI Grabber from Kinect
- Calculate Normals of the point cloud
- Colod-based region growing segmentation
- Select specific segments for further processing through a UI
- Estimate the position of Light-Source
- Render the light source as a sphere in the scene using data retrieved from previous step