Hardware acceleration of infrared small target detection based on FPGA

High-speed small targets detection system based on infrared images plays an important role in aviation, remote sensing, military and other fields. The real-time detection performance is an important indicator to measure the high-speed small target detection system. With the development of small target detection technology, the complexity of the algorithm continues to increase, and the detection system with a single CPU as the processor cannot meet the real-time requirements. To promote real-time performance of high-speed small targets detection system, this paper uses the heterogeneous acceleration platform consisting of FPAG and ARM to implement and optimize the detection algorithm. Based on two hardware design methods named HDL and High-Level Synthesis, some typical small target detection algorithms are implemented with the hardware acceleration technology in this paper. Besides, an infrared small target detection and display system based on the Xilinx Zynq-7000 heterogeneous platform is built and verified. The experimental results show that the method proposed in this paper can not only detect small infrared targets with a high accuracy rate, but also achieve a detection rate of 92.5 frames per second, with a running power consumption of 0.372W. Overall, the computation time reduced 25.9% compared to the CPU platform.