Study On The Enhancement Algorithm For Low Illumination Images

Currently digital video technology has been widely applied to various fields, such as the important place of safety monitoring, traffic management, auxiliary driving, offshore video monitoring, etc. Influenced by illumination environment, monitoring technology image quality can meet the application requirements under the condition of good light conditions in the daytime. But frequent crime at night time, image quality is seriously deteriorated under low illumination; video monitoring has brought serious challenges to night. Firstly, low illumination image rendering a large number of dark space, dark area details blurred, the lost many information. Secondly, the image appeared some specular highlights under artificial light which make the overall image brightness serious inequality. Those problems also reduce the image quality; it is difficult to the problems arising from the monitor with the naked eye observation. Above problem limit monitoring technology function under the condition of low illumination, monitoring application is limited by severe at night. Therefore, the low illumination image enhancement algorithm research has become one of the hot spot of current research. In order to solve the problem of the above several aspects, in this paper, we have investigated the low illumination image enhancement methods and have accomplished the following research work.(1) Research of low illumination enhancement algorithm based on nonlinear transform. In order to overcome the image resolution and uneven illumination, this paper, by using Gaussian smoothing of image smoothing processing, according to the original image of the overall information and local information of image pixel values of the migration processing. Then, using the migration image information and the nonlinear function of image transformation enhances the image details. Through space transform, the image from RGB space to HSV space, realize to the processing of color images, the algorithm run of good real-time performance, can effectively improve the quality of low illumination image, according to the original image information, the adaptive adjustment of the effect of image enhancement, effectively solve the problem of low image resolution and uneven illumination, meet the requirements of the naked eye.(2) Research of low illumination enhancement algorithm based on wavelet transform. The shortcoming of traditional image enhancement algorithm is real-time and adaptive problems. In order to overcome those problems, this paper proposes a low illumination image algorithm based on wavelet transform. First of all, the RGB image to HSV space, and the brightness V image using wavelet transform to image of high and low frequency component separation. Then in the low frequency sub-band of wavelet transform using the illuminate of light component in an image of bilateral filtering estimate and remove the fast in the high frequency sub-band fuzzy transformation to realize the enhancement of edge and texture information and deal with the noise. After the above processing V image, based on the histogram of the objective function is put forward, using Powell combined with simulated annealing optimization algorithm, implements the contrast is rapid, adaptive enhancement processing. Finally, the V component has been enhanced image for color images. The experimental results show that the algorithm can quickly and effectively implement management-level processing a low illumination image.This study can significantly improve the quality of images of low illumination image, make a night monitoring image more clear; According to the image of the light environment adaptive enhancement processing, meet at night video monitoring requirements, has the important theoretical research value and application prospect.