Performance analysis of real time object tracking system based on compressive sensing

Abstract

The main focus of Video Surveillance missions is to amass and verify data regarding appearance of object and position of target object. Such missions typically involve a high degree of covertness. Hence, for such sensitive applications there is desideratum for designing an unmanned surveillance system by using wireless visual sensor network. Because of bandwidth and energy utilisation bound of sensor nodes such systems necessitate less bandwidth and energy aware designs to ensure longevity of system. Therefore this paper proposes a compressive sensing based real time object tracking surveillance system that reduces bandwidth utilization by minimizing the amount of data required for processing. The system aims at minimizing the time required for image reconstruction, enhancing the quality of reconstructed image and reliable tracking of the moving object by reducing computational complexities in real time scenario. The paper attempts to reduce computational complexities and improve the quality of reconstructed image using Smooth Projected Landweber reconstruction technique. It also focuses on reducing the noise accumulated due to randomness of channel by operating under various modulations. Further, Kalman filter is used to track the object's path. To test the reliability of the proposed method, the performance of the system is evaluated under noisy channel using different modulation schemes.