Research And Implementation Of Real Time Detection System For Moving Objects

Moving object detection is an important branch in the field of image processing,and it has important applications in intelligent monitoring, industrial design, military navigation and early warning. However, with the increasing of the practical application, the real-time requirement of moving object detection system is more and more high. In many cases, only using the method of motion detection software has been unable to meet the actual needs, need through the hardware platform to achieve real-time detection of moving targets.This thesis focuses on the real-time moving target detection system. Firstly, the thesis introduced several image processing algorithms used in moving target detection system,including median filtering, Gaussian filtering,morphological image processing, and three classical moving target detection. Described in detail the advantages and disadvantages of several algorithms, finally selected the way of combing three frame difference method and Canny edge detection method to detect the moving objects. The algorithm was simulated by OpenCV. This method which has high real-time and strong adaptability to change in the environment can obtain better detection results. It designed the real-time moving target detection system by using FPGA as the main chip, discuss in detail the implementation process of the main part of the hardware and software of the system. The hardware part mainly introduced some major modules including image sensor module, SDRAM module,VGA module and FPGA module.The software part analysed in detail the algorithm implementation process and the process of each module controlled by FPGA.Finally, the system performance and the detection results were tested and the various modules of the system were simulated by Signaltap. The experimental results show that the system has high real-time and good detection effect.