Research on Night Scene Image Enhancement Algorithm Based on Half-pixel Constraint

Abstract

Due to the relatively concentrated distribution of gray value in night scene image, it is often necessary to improve the visual effect of the image by widening and lifting the gray value, so as to further clarify the details and facilitate the subsequent processing operation. To solve the problem that the traditional contrast enhancement algorithm often enlarges the noise when processing the image, which affects the visual quality of the enhanced image, a half pixel constrained image enhancement algorithm is proposed. The algorithm divides the night scene image into bright and dark regions according to the gray value corresponding to the half pixel of the image, and uses bi-directional histogram equalization to realize the half pixel constraint in the bright region, so as to suppress the signal-to-noise ratio, while the traditional double histogram equalization is still used in the dark region. The experimental results show that the algorithm can better preserve the details of the original night scene image, and has the ability to deal with the noise interference.