基于物理模型与边界约束的低照度图像增强算法

Enhancement Algorithm for Low-lighting Images Based on Physical Model and Boundary Constraint

针对低照度下图像降质严重的问题,该文提出一种基于边界约束与图像亮度的低照度图像增强算法。该算法首先通过改进的边界约束对伪雾图进行透射率估计,并对其进行优化;同时从伪雾图"雾"的形成原理出发,利用低照度图像的亮度分量进行伪雾图大气光值的估计;最后将增强后的伪雾图反转,即得到增强后的低照度图像。实验结果表明,针对低照度下的图像,该算法可以有效地提升对比度和亮度,过增强现象得到改善;效果优于对比算法,且复杂度低。

The aim of this paper is to achieve a low-lighting between fog image and inverted low illumination image. The image enhancement method by using the similarity transmittance estimation of the pseudo-fog image is estimated by the improved boundary constraint, and then it is optimized. Based on the formation principle of pseudo fog, the light intensity of pseudo fog map is estimated by using the brightness component of low illumination image. The enhanced pseudo fog image is reversed to obtain the enhanced low illumination image. Extensive experimental results using natural low-lighting images indicate that the proposed method perform better than contemporary algorithms in terms of several metrics, including the intensity, the contrast. The proposed algorithm can effectively suppress the wrong phenomenon caused by enhanced with low complexity.