Implementation of Real Time Local Search Particle Filter Based Tracking Algorithms on BeagleBoard-xM

Abstract

This work hybridizes Particle Filter and Local Search algorithm for the realization of intelligent computer vision based target tracking on BeagleBoard-xM, an ARM based embedded platform, which tracks moving targets in a continuous scene operating in real-time. The integration of Local Search algorithms with Particle filter significantly increases the accuracy of tracking. The use of embedded board helps to reduce the space requirement and the cost of design. The implementation of target tracking on embedded platform has many end uses especially in the field of robotics, surveillance, human-computer interaction applications, etc.