The Applications Of Vehicles Target Tracking Algorithm In Real-Time Based On SOPC

In recent years, the intelligent vehicle research, which has been got the extensive attention at home and abroad, is currently in exploratory research stage. As the core of intelligent vehicle research, the key technology of visual target tracking of vehicles is the environment anti-jamming ability and real-time of algorithm. With car as the carrier, by the research of target tracking algorithm, this paper presents a target tracking system which is based on SOPC design. The experimental results show that the method used in this system can achieve the visual tracking of vehicles, and can get a better real-time, so it has certain practical value in the relevant research.First, combining the research content and extensive reference, made a search for target tracking algorithm in depth and summarized the predecessor's research achievements, and have a good mastery of research situation.Second, combining the environmental anti-jamming and the real-time of the algorithm's hardware implementation got the improved Mean Shift as the target tracking algorithm and made a pipelined design on the hardware.Third, made a parallel design of the algorithm by using DSP Builder, and achieved the target tracking of model system of vehicle in real-time. Combining the FPGA hardware development environment, converted the model built by DSP Builder to VHDL code automatically and completed the FPGA hardware environment adaptation. In this way, instead of Verilog or VHDL FPGA hardware code in hand written, the system shortened the development cycle and improved electronics design efficiency of FPGA realizing DSP system.Fourth, based on the SOPC system research, combining FPGA hardcore and comprehensive application of Nios II soft nuclear, realized design of the algorithm on chip system, and set the system-level design, embedded software development, programmable logic design at an organic whole.Finally, compared the method used in the system with the method using Matlab software programming for the accuracy and real-time test, the experimental results show that the income of tracking trajectory of the two methods is basic consistent, but the method in this system is about four times on Matlab software programming method in real-time, which improve the real-time greatly.