基于多层融合和细节恢复的图像增强方法

Image enhancement method based on multi-layer fusion and detail recovery

针对部分图像在光照不均匀、过亮或过暗下出现的对比度低、细节不可见等问题,提出了一种基于多层融合和细节恢复的图像增强方法。首先在HSV图像空间中将V通道等价复制为Retinex模型增强层、亮度增强层、细节突出层三层。在Retinex增强层中,利用加权引导滤波和形态学结合来消除光晕现象,并通过改进Retinex模型增强图像亮度和细节;在亮度增强层中,通过自适应归一化函数进一步增强亮度;在细节突出层中,人工蜂群算法优化改进局部线性增强模型来突出图像细节。最后根据Gamma校正特性和邻域像素关系,提出细节恢复方案避免融合后造成的部分细节模糊。实验数据表明,该算法能更有效地突出图像细节和提高对比度,并与现有算法在客观量化方面进行对比,综合性能更为优越,尤其在清晰度指标上远高于其他算法。

This paper proposed an image enhancement method based on multi-layer fusion and detail recovery,to solve the image deterioration such as low contrast and blurred details in undesirable illumination environments.Firstly,this paper copied the V channel equivalently into three layers in HSV color space:Retinex enhancement layer,brightness enhancement layer,detail enhancement layer.In Retinex enhancement layer,this paper combined with weighted guided image filtering and morphology to eliminate halo phenomenon.It improved Retinex model to enhance brightness and details of images.In detail enhancement layer,this paper used artificial bee colony algorithm to optimize improved model of local linear to obtain more details.Finally,this paper performed Gamma correction and pixel arrangement to avoid partial fuzzy details caused fusion.The experimental results show that the proposed method can more effectively highlight image details and improve the contrast.The comprehensive performance is superior while comparing with the related methods in terms of objective quantification,especially in Tenengrad index.