Development of Monocular Vision based Targeting System for an Autonomous Defence Vehicle

The present work deals with development of a Target tracking system based on Monocular Vision for an Autonomous Defence Vehicle without a global satellite view. The vehicle is also equipped with ultrasonic sensors to gauge distance from the target at any point of time. In this work, theoretical expressions have been arrived at for control of motors for accurate capture of target without image capture of the targeting pointer itself. This can greatly reduce the costs as the high resolution cameras required for capture of targeting pointer (generally a LASER pointer) are very costly. Theoretical expressions have also been determined for the approach distance in dependence only on camera, motor specifications and the target size. A prototype system was developed along with the necessary real time embedded controller and image processing algorithms for identifying the target and to control targeting system LASER pointer. Experiments were conducted to find out efficacy of developed system. The system is able to detect the target on different approach distances and camera cone angles. It should be noted that the target size will also be calculated on the fly by the robot itself. This makes it possible for deployment of this targeting module without much change to the existing system. Keywords— Autonomous Defence Vehicle, Monocular Vision, Approach Distance, Size Factor of the Target