Obstacle Detection using Cascade Classifiers

Course Project under course CS365A - Artificial Intelligence Programming

• Trained Cascade Classifiers on features (HOG, Haar and LBP) to detect cars in images
• Worked with the KITTI Vision Benchmark Suite - Object Detection Evaluation 2012 dataset
• Partitioned Point Cloud data to identify obstacles using Velodyne Height Map algorithm (“http://wiki.ros.org/velodyne_height_map")
• Processed LIDAR scan data using graph based technique to identify ROIs (Regions of Interest)

Check out the detailed report here and poster here.