Design and Implementation of Infrared Image Preprocessing System based on FPGA

Infrared image processing is a conventional topic, but the hardware design and implementation for infrared image processing algorithms are full of challanges, especially for special purpose infrared detectors which have customized interfaces and protocols. Moreover, some customized infrared detectors only generate the infrared image with very poor quality. Therefore, in this paper, we design and implement an infrared image preprocessing system based on FPGA for customized infrared detectors. We propose a hardware architecture that includes the whole process from the input of detector, image processing, to final infrared video stream output and display. Our hardware architecture is able to adapt to different customized infrared detectors, which has the characteristics of real-time, low-power, dynamic parameter configuration, high processing efficiency and stable imaging effect. We have implemented it on Xilinx FPGA. The infrared imaging effect is significantly improved with limited FPGA resources (about only 1.6M LUT, 420 BRAM for all algorithm models on Xilinx zc706 evaluation board), the image processing speed can reach to 225Hz with the resolution of 320*256. Our preprocessing system can generate real-time video for display with 1920*1080@60Hz.