Research On Video Image Processing Technology Of UV-visible Dual-spectrum Camera

With the increase of transformer substation and power transmission station and the rapid development of high-speed rail, the maintenance and inspection of high voltage equipment and transmission line have become a research hot spot. In the line maintenance, not only corona discharge phenomenon should be detected, but also the location of malfunction should be fixed. Ultraviolet camera can only detect the discharge signal, but cannot identify the specific discharge location. Ultraviolet-visible double spectrum camera can be used to identify and locate the corona signal. In this paper, video image processing techniques of ultraviolet-visible double spectrum camera are researched.This paper introduced the working principle of ultraviolet-visible double spectral camera, and the image shot by double spectrum camera was analyzed. The time-recursive filtering algorithm and median filtering algorithm were selected to restrain the noise of the video image to prevent noise from interfering discharge signal.In double spectrum image matching, the proper registration object was found, and the image translation and scaling parameter were acquired by characteristic image registration method. The matched images were fused to realize the detection and location of corona discharge. And the algorithms were transplanted into the hardware platform to make a comprehensive test.The results show that the video image processing technology of ultraviolet-visible double spectral camera can eliminate the noise signal effectively and obtain accurate registration parameters. The fusion effect is perfect. The test results in the hardware platform meet the design requirements, and the real-time performance is good.