Design For Infrared Small Target Tracking System Based On FPGA

Real-time infrared small target tracking technology has been widely used in military and civilian applications. However, the small target information such as size, texture and shape is lacking and the traditional detection and tracking algorithms are highly complex. It is difficult to get a stable real-time tracking of small targets in infrared image. In this paper, a real-time small target tracking system is designed and implemented. FPGA (Field Programmable Gate Array) is used as the core processor of the system because of its advantages of parallel processing and rapidly computing. A small target tracking algorithm which is effective and easy for hardware implementation is proposed and implemented in FPGA. The main contents are as follows:(1) BT.656 standard and involved video scanning methods are discussed. Then the color space theory and different YCbCr sampling methods are introduced. Finally, the working principle and the design process of FPGA are specifically analyzed. The necessary development tools used in this paper are introduced and the chips are determined.(2) An improved Top-hat transform is proposed based on the analysis of the imaging characteristics of small target for the limitations of classic Top-hat transform in small target detection. The improved transform is applied in image preprocessing to restrain the background while highlighting the small target. On this basis, with the motion estimation of the targets in the video sequence and the spatial filter based on the characteristics, the false targets which are not within the scope of the motion region can be removed so that the position of the small target could be calculated effectively and stably. Experimental results show that the algorithm not only has better tracking performance but also meets the real-time requirements.(3)In this paper, the system architecture is designed with D/A and A/D converter codec chip, SRAM for storing video data, and Cyclone II series EP2C70F672I8 FPGA chip. Modules of the tracking system, such as image acquisition, storage, and core algorithm, are implemented based on the hardware architecture. Considering the great resource consumption and large amount of computation of traditional sliding window operation, a new sliding window calculation method is designed. The new template window operation is completed by accessing the multiple lines of pixel grey-values with dual-port RAM, and it is applied in Top-hat preprocessing and detection-tracking module, saving the system hardware resources. Experimental results show that the proposed system achieves a good realization of the real-time small infrared target tracking.