Object Recognition of RGBD data with SVM and PCL

Built perception pipeline for object recognition of PCL point cloud against SVM model using ROS system in Gazebo simulation.

Responsible for SVM model training and PCL point cloud filtration, segmentation, clustering, feature extraction, and classification. Developed C++ and Python multi-node architecture, custom messages, services, launch files, and shell script with parameter server initialization of environment variables. Achieved 100% correct classification in real time across 3 simulated test environments.

Experience with ROS, PCL, C++, Python, SVM, Sklearn, Gazebo, Point Cloud Filtration, Segmentation, and Clustering.

Full write-up and code available here