Design And Implementation Of Real-Time Moving Object Detection System

The detection of moving objects from video frames plays an important and often very critical role in different kinds of machine vision applications including human detection and tracking, traffic monitoring, human-machine interfaces and military applications. Moving object detection is the basis for further complex processing, and play an important roles to other topics' progress in computer vision. How to detect moving object real-timely, steadily and effectively, becomes an important problem that needs to be paid attention and researched.For in the current most of the moving object detection algorithms are realized by software based on the PC, it is slow to deal with processing speed and has limited of application scope. This paper presents an FPGA-based platform for moving object detection system. As a main large scale programmable specific integrated circuit, FPGA meets the requirements of high speed image processing. The system uses high-speed CCD camera to capture real-time images, by improving the detection algorithm for image processing to detect multiple moving objects.This paper focus research on the current major algorithm of moving object detection, and studies the key technologies of the field based on the current research achievements and LBP-based moving object detection algorithm has been improved to make it more suitable for hardware implementation. The algorithm is not only the advantages of background subtraction, but also to overcome the impact of detection of the monotony of light change. The algorithm is simple, and conducive to real-time detection system. Based on the rational utilization of hardware resources, through the design of FPGA hardware circuit, implemented high-speed image capture and image pre-processing, the full and effective use of FPGA hardware parallelism, through implement a hardware structure of a pixel processing, the FPGA at the same time achieve 3 pixels to parallel processing, and hardening of a partial algorithms significantly improve the detection speed. For the algorithm have a better adaptability to environmental issues, we are combination the improved the texture-based on motion object detection algorithm and the adjacent frame difference, so that the result of detection is improved. The detecting object moving system, can be used for other applications in the visual field, and will have a wide application prospect as well.