This release consists on updated UML diagrams and revised implementation of previously designed stubs.
The project detects human obstacles and tracks them while outputting the location information (x, y, z) in inches, preferably in autonomous vehicle systems. The system work such that the distance from any camera reference frame is converted to the robot reference frame and made know to the robot for perception tasks.
The API consists of a detection and tracking sub-parts,
In detection, a video stream is initialized and human obstacles are detected using face recognition in OpenCV by employing a ResNet detection.
In tracking, the detected obstacles are assigned unique IDs for tracking and their location (x, y, z) in robot reference frame is calculated.