Research Of Moving Objects Detection And Tracking System Based On FPGA

Currently, the moving target detection and tracking technology based on video images is one of the most important researches in the field of machine vision research, which is the foundation and key technology of intelligent monitoring, human-computer interaction and robot navigation and other areas of application. FPGA hardware processor has characteristics of fast processing speed, design flexibility and low power consumption etc, so it can improve the speed of image processing algorithms, especially suitable for video image processing area. Therefore, the research based on the FPGA hardware processing of moving target detection and tracking system has both important theoretical significance and practical value.In this thesis, several moving target detection and tracking algorithms are studied. Firstly, the basic principles and characteristics of three motion target detection algorithm are introduced, namely the frame difference method, the background minus division and optical flow method. According to the tracking characteristics, the tracking methods are classified into feature-based tracking, region-based tracking, active-contour-model-based tracking, and model-based tracking. Finally, analysis showed that, the frame difference method is suitable to realize on the FPGA device, and the tracking algorithm which based on the target characteristics is chosen in this thesis. And matlab software is used to simulate the frame difference and centroid-based feature tracking algorithm.This system is based on FPGA hardware development platform to introduce the detailed design principle of the various hardware modules of this program, using frame difference method to detect moving target and chosing the centroid of the moving target as the tracking characteristics, detailing design principle of the various hardware modules based on the program. First, configuring the CMOS image sensor internal registers via the I2 C bus configuration module, then image data acquisition module receives sensor data and frame control signal. The image conversion module converts the Bayer format data into RGB format. The multi-port SDRAM controller module stores the generated RGB format data to realize the image data of the two frame buffer. The target detection module adopts frame difference method to detect the moving target. The target tracking module uses centroid matching method to track the detected target. The VGA control module displays detection and tracking results on a VGA monitor. At last, the Quartus II software integrates the various modules together to analyze, synthesize, lay out, generate the configuration file and download it to the DE2-115 board. This system achieves the goal of moving object detection and tracking in real time under static background.