Comparison study between different automatic threshold algorithms for motion detection

This paper proposes a comparative study between different commonly used global thresholding methods applied on various differential motion detection algorithms. The purpose of comparison is to define the appropriate automatic threshold method for surveillance applications, both in indoor and outdoor areas. In order to achieve that, five threshold methods have been tested on different differential motion detection algorithms, using four scenes with different complex backgrounds. A pixel-based evaluation has been done to determine the best combination.