Basic Info

• Library: OpenCV 4.1
• Language: C++
• Plateform: Mac

Project 1 | Lidar Obstacle Detection

• Created a pipiline for segmentation and clustering of the real time PCD data.
• Real time PCD data is streamed to the pipeline.
• Data is downsampled using Voxel-Grid.
• Implemented Ransac Algorithm for segmentation of road from obstacles.
• Implemented Euclidean Clustering Algorithm using KD-Tree for clusteration of the obstacles.

This is how the results looks like:

Project 2 | Feature Tracking 2D

• Implemented Keypoint Detection, Description Extraction and Keypoints Matching algorithms.
• First keypoints are located in the images.
• Using these keypoints their descriptors are extracted.
• Same keypoints are matched using their descriptors in different images.

Following is a detailed analysis of algorithms used for keypoints detection, descriptor extraction and keypoints matching.