Design Of Moving Tracking System Based On FPGA And Machine Vision

Along with the development of science and technology, robots play a broad role in society. They have replaced human beings in performing repetitive and dangerous tasks. In robots, machine vision is an important part of automatic control. At the present stage, in most machine vision, computers are used for image acquisition, image processing, motion analysis and motion tracking. However, in some applications, e.g. unmanned aerial vehicles (UAV), highly flexible mobility and ultra-low power consumption are required, which resulted that the computer cannot be used in them. Therefore, there is an urgent need for a motion control system with higher performance, lower power consumption and smaller size compared with computer.The embedded vision system with a FPGA (field-programmable gate array) cores has the advantages of high performance, low power consumption and compact structure, which is one of the current research hotspots.In this paper, a FPGA based visual tracking system was designed. The logic derivations and code writings of image preprocessing and motion tracking algorithm were given. Then the algorithm level simulation was carried out by MATLAB software, which determined the feasibility of this design scheme. In addition, the whole hardware circuit is designed. The selection of the hardware components, the top-level schematic of whole system and overall circuit board were layout.In this design, a lot of image preprocessing was done, which effectively decreased the post target-tracking algorithm calculation and the whole calculation. At the same time, the sparse optical flow field was used as the motion-tracking algorithm, which can take advantage of the parallel computing character of FPGA. Therefore, the speed and accuracy of the motion detection and tracking were effectively increased. In addition, the DDR3-1600 memory chip was used on the printed circuit board (PCB), which is the latest and most popular memory product up to now. Comparing with other memories, it can significantly simplify the wiring, reduce the requirements of integrity of the PCB signal, and ensure the high data throughput of the system.