|
AbstractIn recent years,with the rapid developments of the deep sub-micron integrated circuit technology and SOC (System on Chip) technology, the performance of the IC chips have been greatly improved, algorithm modules which have a higher complexity can be integrated into the single chip. At present, ASIC which is based on the implementation of machine vision has become one of the most popular research topics. Video Moving Target Tracking, as an important field in the machine vision, links up an amounts of subjects such as computer science, pattern recognition, artificial intelligence and image processing. Today, Video motion tracking technology has been used not only in the military fields, for instance, radar, navigation and so on. But also apply in the civilian areas including intelligent video surveillance areas, vehicle tracking, automatic driving and human-computer interaction. Therefore, making research to the video motion tracking devices has a significance sense and it is no double that it will has a prosperity prospect in the future.In this paper, an ASIC chip which can make the tracking to the dynamic video object in the static background is proposed. First, the overall structure of the entire chip digital system is constructed according to the specifications, Second, the video image acquisition module, SDRAM controller module, image format conversion module, the value of the filter module, image edge detection module, SAD matching tracking module and the VGA controller module are designed, Regarding to the median filter module, the fast median filtering algorithm is used to improve the chip’s processing speed. For the image edge detection module, applications of the improved Sobel edge detection algorithm can be simplified to reduce the chip area. In the SAD matching tracking module, according to combination of HSV color space,to overcome the impact of the impact of the algorithm by light.Then, using Leda, VCS, Design Compiler and other EDA tools to making code verification, timing simulation, synthesis to the circuit design of each module and the results of the simulation are illustrated. Finally, making layout design and FPGA verification for the chip by using the HeJian 0.18 m CMOS process technology.Simulation and test results show that the moving object size is 16x16 pixels and the maximum speed is 4m/s in the static background, the paper design of the chip processing speed of up to 50 frames/sec or more, the system runs the maximum operating frequency of 100 MHz And accurate tracking of moving targets higher than 90%. Therefore, the development and design of the chip to achieve the desired technical requirements.
|