Study Of Man-In-The-Loop Infrared Imaging Guidance Simulation System And Technology Based On CUDA

People involve in the process of guidance in man-in-the-loop infrared imaging guidance system, which improves the accuracy of target recognition and enhances its antijamming capability significantly, and it is an important branch of precision-guided munitions developing. The development of man-in-the-loop infrared imaging guidance simulation system is of great significance in terms of cost reduction, lowering the risk of experiments, operator training and improving the hit rate. CUDA(Computer Unite Device Architecture) provides a solution of general computing on GPU, which has been applied to accelerate application in many areas. Designing and analyzing the man-in-the-loop infrared imaging guidance simulation system and applying CUDA to its computationally intensive parts is an important respect in ensuring the real time simulation with man in the loop.In this paper, we design man-in-the-loop infrared imaging guidance simulation system and divide mathematical simulation system into three subsystems:missile dynamics solver subsystem, infrared scene generation subsystem and seeker control subsystem according to the real process of man-in-the-loop infrared imaging guidance. To accelerate the simulation process and meet real-time requirements, we analyze and study infrared scene generation subsystem and seeker control subsystem, both of which are computationally intensive. We analyze the modules of infrared scene generation subsystem and seeker control subsystem in the perspective of applying CUDA and make some conclusions. After understanding the way of CUDA program and execution, we focus on accelerating automatic target recognition process in seeker control subsystem with CUDA.We parallel gray-level image matching algorithm and covariance matrix image matching algorithm respectively to reduce the matching time and achieve good results.