Study Of Backlight Compensation And Color Correction Algorithms For HD Camera

It's endless pursuit of camera quality. In the broadcast and consumer electronics, high definition videos and high-resolution images have become a common mass consumption. And, with the updating of the consumer experience, we also make similar requests on industrial cameras, traffic monitoring cameras and aerospace cameras. As the key to improve the quality of the camera, the image pre-processing algorithms in the camera play a decisive role.Based on the application demands of high-definition aerospace cameras with independent research, image pre-processing algorithms are optimized in this paper. And focus on the issues of unreal brightness and color of the image when shooting in backlight or harsh external environment, backlight compensation and color correction algorithms have been optimized to enhance the camera quality.The main research works in this paper are as follows. Firstly, based on the demands of high-definition aerospace cameras and the defects of image pre-processing framework in existing cameras, this paper presents an improved image pre-processing framework, adding color correction and backlight compensation function modules specially, and improves the ability of cameras to adjust brightness and color adaptively. Secondly, based on the comprehensive analysis of existing backlight compensation algorithms, combined with the specific needs of aerospace cameras, a dichotomy backlight compensation method based on the ratio of image backlight is presented. First of all, according to the BBHE algorithm, this paper proposes a method for image segmentation based on the ratio of image backlight, to ensure the brightness of an image which is compensated with segmentation keep continuously and natural. Then, after the segmentation of the backlight and background area, compensation is achieved using BUBO HE for the two subsets. Taking advantage of the thought that the HE degree is adjustable for BUBO HE algorithm, this paper makes different levels of brightness compensation for different regions based on backlight ratios. That is, the lighter background area maintains its basic brightness, and the backlight area is substantially compensated. The algorithm in this paper is improved according to the conventional HE, with better compensation effect and broader applicability. Thirdly, according to the needs of aerospace cameras and disadvantages of the existing algorithms of color correction, a method based on the weighted standard deviation of color is proposed to optimize color correction. In the first place, a method for determining color distortion is proposed, to achieve adaptive distortion decision. And then, this paper improves the gray world method, and proposes a method based on the weighted standard deviation of the color to optimize gain coefficients. Meanwhile, in order to reduce the impact of the color temperature of the light source, the traditional linear correction method in which the three color channel correction coefficients are processed equally is improved, and a new method which deals the three color channel correction coefficients with different weights according to different color casts(corresponding to different color temperature of light source) is presented. The method in this paper not only improves the accuracy of the color correction but also has a wider adaptability.