Video Sequence Image Processing And Intelligent Matching System

Video monitoring is widely used in many kinds of fields such as transportation, finance, medicine and military with many advantages. Video sequence intelligent processing not only reduces the physical burden of human monitoring, but also improves the accuracy and automation degree of monitoring processing as well as reduces the storage space. Video sequence processing depends on to establish human visual computer theory, also needs to solve the problems including image enhancement processing and image matching technology.This paper mainly studied the video sequence image processing and matching theory, based on image enhancement processing technology and image matching algorithm, built a video sequence processing system.First, we obtain streaming video information and get image sequences through the network video recorder, then enhance the images and take template matching, to finally achieve the video sequence enhancement processing and intelligent matching software.In image enhancement processing, with the problems of reducing image quality, we mainly studies the image grey level transformation, image smoothing and sharpen processing technologies.Affected by light, noise and many other factors, sequence images gain degradation or quality reduction problem. The purpose of image enhancement is to improve the input image quality, and make it a clear image, thus lay a solid foundation for subsequent image matching part. This paper researched and implemented the image histogram equalization, neighborhood average method and median filtering technique based on the airspace method, and image gradient sharpening and Laplace mask sharpen. In the image matching, the theoretical basis and classification method and common technology of the image matching technology had been deeply analysed. Image matching uses similarity measure and binding conditions to find the match point between two or more images, aiming to eliminate various environmental differences. This paper mainly studies image matching method of SIFT features, detailed introduces the algorithm thought and method, thenmake a performance contrast experiment based on the gray scale image matching algorithm. The experimental results show that the SIFT matching has high stability and robustness for transformation as illumination, rotation, scale and shelter, and use the algorithm as a matching methodin the video sequence image processing and the matching system.