A New Algorithm for Backlight Image Enhancement
Chengxu Ma, Shangyou Zeng, Daihui Li · 2020
Backlight images will have influences on many aspects of our lives, such as our shooting effect, the judgment of the images in the public security systems, etc. So it is very important to solve the backlight images for building a smart city. When processing the backlight images, we need to ensure that the backlight area of the images is effectively enhanced, and the image information of the non-backlight area cannot be lost. In this paper, we design a new function for backlight images. We first enhance the R, G and B channels of the backlight images locally to make the images more colorful. At the same time, we transform RGB brightness adjustment model into HSV brightness adjustment model and use the new function to adaptively adjust the V channel in the local enhanced images. The method of self-adaptive brightness adjustment can ensure that it can be applied to most backlight images. Finally, the traditional image enhancement methods are compared from the perspective of vision and information entropy to verify the effectiveness of this function.