Data collection is done using the Intel realsense camera for 2D data and Velodyne ultra-puck LiDAR camera for 3D data. After collecting data, we analyze it, clean, and label it for use in our web application.

Machine learning models for both 2D and 3D data are trained with labeled data in order for the models to track and label vehicles in real-time. Data will stream through AWS clients to use in the web application.

Website displays labeled real-time stream with graphs based on the data from 2D/3D Machine learning models. Registered users have access to historical data + camera configurations.