The Research On Image Enhancement Technology In Video Surveillance System

With the rapid development of network, communications, and image processingtechnologies, video monitoring with its intuitive, user-friendly and content-rich features attracts people’s increasing attention, and is already applied in safety monitoring, fire monitoring, traffic flow and peccancy, and the surveillance for public area. The video monitoring image quality effects on monitoring system directly. But with some practical reasons, the video monitoring system images usually have defects such as poor contrast even blur, it’s difficult to distill useful information from these images. So image enhancement technology is nesscessary for improving image quality and intelligibility.The main works of thesis include:The basic theory of image enhancement was introduced, the concept, basic type and color space of digital image were briefly descript, and the common image enhancement algorithms and their defect were analyzed, the effect of these algorithms was tested by actual images. In order to evaluate the effect of image enhancement, the evaluation criteria was analyzed from subjective and objective aspect.Aiming at enhancing color image with low contrast, the classic multi-scale Retinex algorithm for image enhancement was analyzed, then a new adaptive scale MSR enhancement algorithm was proposed which enhanced image in HSI color space. The Gaussian filter was determined by the variance of luminance component in HSI space, as a result, the detail feature of image was highlighted and meanwhile the stability of color was preserved.In order to enhancing image in particular interesting region, the classical moving target detection methods were summarized, with its combination to image enhancement algorithms, a new image enhancement method based on ROI(Region Of Interest) was proposed, only the target’s area in image was enhancement, so the amount of computation was reduced meanwhile the target area was enhancement.