Design And Implementation Of Real-time Infrared Target Detection System Based On FPGA

Target detection technology is an important research direction in image processing.In recent years,people's requirements for the real-time,portability and detection effect of target detection systems have become higher and higher,and software-based target detection systems have been unable to meet the above requirements.Therefore,target detection systems based on hardware platforms have become the future development trend.For this reason,this paper designs a real-time infrared target detection system based on FPGA.The main tasks are as follows:Firstly,we introduced the imaging principles and characteristics of infrared imaging and visible light imaging,the infrared and visible light cameras used in the system,the basic theory of FPGA,and several commonly used image quality evaluation methods.Secondly,the image preprocessing algorithm and target detection and visible light positioning algorithm are introduced.Including,optimized median filter algorithm,the combined algorithm of linear transformation and Laplacian sharpening,the combined algorithm of histogram equalization,gamma transformation and Laplacian sharpening,the combined algorithm of background difference method and threshold segmentation,edge detection algorithm and visible light positioning algorithm.The simulation research is carried out for the above algorithm,which provides a theoretical basis for the subsequent hardware implementation.Thirdly,a real-time infrared target detection system based on FPGA was built,the design and simulation of infrared video acquisition,decoding,buffering,SD card reading image and image display module in the system are completed.The ping-pong operation was used in SDRAM to realize video streaming.Finally,the image processing algorithms proposed in the previous stage were implemented based on FPGA,and the circuit scheme and hardware structure of the algorithm are introduced,and the simulation verification of each image processing algorithm module.At the same time,a comparison chart of the algorithm processing results is given.The research of the thesis has important reference significance for the design of real-time infrared target detection system.Through further improvement,it is expected that a more convenient and fast target detection system could be obtained,which can meet the needs of target detection systems in many occasions.