Transplantation And Implementation Of The Millimeter Wave Imaging Processing Algorithm Based On Multiple Execution Platform

Passive millimeter wave imaging system is an imaging system which can detect the radiation of each object in the millimeter wave band and image based on difference of radiation energy . In addition to the characteristic that it can penetrate the clothing to detect and image objects hidden, it has no external radiation,doing no harm to human health. Thus ,in the future ,it can be widely used in human security and other scenes.In the passive millimeter wave imaging system of our research group, it cost the DSP signal processing unit a long time to implement some algorithms, difficult to meet the real-time requirements of the system. To solve this problem, combined with some research results of the algorithm of our research group, these algorithms have been implemented is PC execution platform and corresponding software has been designed and developed; In order to make the system more flexible, these algorithms in the FPGA execution platform have also been implemented and corresponding hardware has been designed and developed .The research contents of this paper are as follows:(1) Firstly, some research results of millimeter wave imaging algorithm of our research group have been introduced and analysed, which are channel equalization algorithm based on scene calibration, S curve mapping algorithm based on Logistic function, denoising algorithm with PM model and YK model combined, interpolation algorithm for sampling channel scene calibration.Based on the research and analysis of the algorithm, we have laid the theoretical foundation for the software implementation of the PC execution platform as well as the hardware design and implementation of the FPGA execution platform.(2) Then, software of millimeter wave imaging processing algorithm based on PC execution platform has been designed and developed:the software has not only realized 4 algorithms of channel equalization, S curve mapping, denoising and interpolation, it also provides nice user interface and file interface, serial interface, network interface,algorithm selection and parameter configuration, display after image processing, servo control and other functions. In addition, the software uses the multi-thread pipeline image processing technology, which greatly improves the processing speed of the image sequence and improves the real-time performance of the system.(3) Finally, hardware of the millimeter wave imaging processing algorithm based on the FPGA execution platform has been designed and implemented. The design process includes circuit design, PCB mapping, physical production and the realization of the 4 algorithms. In addition, the function of serial communication and inter board cascade has also been realized on the hardware so as to facilitate the later expansion of hardware.The above work has been verified by millimeter wave imaging system. The software and hardware designed and developed based on these algorithms can be used in millimeter wave imaging system, and the algorithm is less time-consuming, which improves the real-time performance of millimeter wave imaging system.