Software Design Of Dual-Spectrum Image Detection System Based On DSP

Corona and spark discharge may appear on high voltage power network,substation and distribution etc. UV and visible double spectrum image system isdesigned to detect corona discharge by using solar blind ultraviolet detectiontechnology. Highly digital signal processing gives an impetus to real-time imageprocessing. With the usage of DSP, multi-channel image fusion can be achieved inhigh speed, as it has the virtue of programmable, high speed, pipeline architecture,parallelism and stable.This thesis focuses on system software design of the UV and visible doublespectrum image system. Firstly, the paper introduces the concept, significance andintemational and internal developmental status of the UV signal detection and highlydigital signal process. Secondly, based on the analysis of TMS320DM642's structurecharacteristics, design thought of system software and working process of eachsoftware module are deeply discussed, and Mini-Driver of video date acquisition anddisplay is compiled by using DSP/BIOS of CCS. With system image's signal andnoise characteristic, related image processing algorithms are deeply analyzed and arefound the best proposal to this system after doing some experiments on DSP circuitboard. For the system real-time needs, program is optimized to elevate running speed.Thirdly, to extend the remote control function of system, experiment of real-timedigital image processing based on Ethernet is carried out on EVM.The system debugging result indicates that system software design and imageprocessing scheme are correct and valid.