Vision Based Tracking of Moving Target for Autonomous Ground Vehicle

Abstract

The tracking of moving objects using a camera mounted on an autonomous ground vehicle time to keep the target within the image on an autonomous ground vehicle is addressed. The Continuously Adaptive Mean Shift (CAMSHIFT) algorithm is used to identify the moving target based on the colour distribution of the target histogram. The CAMSHIFT algorithm adaptively adjusts the track window`s size and the color distribution pattern of targets during tracking and uses a background-weighted histogram to distinguish the target from the background. A control scheme is designed for yaw and position control to keep the tracked object in the field of view. The system is practically implemented and tested using the Arduino platform and a cheap low cost web based camera.

The tracking of the blob is done by moving the camera to desired orientation to follow the moving blob to keep it in the FOV. From the experiment we were able to get PAN angle and TILT angle of the camera orientation, which has been validated against the Inertial Measurement Unit.