Low-light image enhancement based on FPGA

Abstract

This paper studies FPGA image processing, and proposes two methods to improve FPGA image processing acceleration, data reorganization and multi-processing modules. Both methods use ARM as the data control module. Data reorganization improves the transmission efficiency of the AXI bus by merging data, and the multiprocessing module improves the parallelism of the processing system by adding processing modules. Finally, these two methods are used to realize the low- light image enhancement algorithm combining wavelet transform and Retinex algorithm, and compared with the processing effect on matlab, and good results are obtained.