Design and Implement of Median Filter toward Remote Sensing Images Based on FPGA

In this paper, an improved median filtering algorithm for remote sensing images is proposed and implemented on FPGA. According to the characteristics of FPGA parallel processing, the 5x5 double-sorted median filtering algorithm is optimized, and the 3x3 double-sorted median filtering, 5x5 double-sorted median filtering and bit-level median filtering are implemented by pipelining, which improves the frequency and speed. The final simulation results show that the maximum clock frequency of this design on XC7K325T can reach 419MHz. For remote sensing images of 4096x4096 size, the processing of a single image can spend 40ms, and the throughput rate is 25 images/second, which can meet the requirements of on-orbit real-time processing and can be applied in practice.