Enhancement of Gaussian background modelling algorithm for moving object detection & its implementation on FPGA

Abstract

Gaussian background modelling algorithm is a classic algorithm used for moving object detection in video. This work proposes a technique to enhance the Gaussian background modelling algorithm in order to improve the moving object detection result. Since the potential application scenario requires real-time video processing, a circuit to implement the algorithm is designed on a Xilinx FPGA. The proposed design is of pipeline structure. Experimentation shows that it can perform real-time moving object detection in a live video of 1280×720p@30fps.