改进重构的自适应权重Retinex图像增强算法

摘要

In order to eliminate the phenomenon of color distortion and "halo artifacts" in the processing of low illumination color image,a low illumination multiscale Retinex color image enhancement algorithm of improved reconstruction adaptive weighting is proposed. In HSI color space, saturation is managed by linear stretching and reflection component is acquired by using guided filter multiscale Retinex of improved reconstruction adaptive weighting,then reflection component is corrected by Gamma to gain the final brightness component. Finally,the proposed algorithm employes color restoring function to recover image to achieve the final enhanced low illumination color image. The experimental results show that the algorithm has the advantages of markable enhancement effect, high real-time performance,more rich image texture details,avoiding the color distortion.%为了消除低照度彩色图像处理中存在颜色失真、"光晕伪影"等现象,提出改进重构的自适应权重Retinex低照度彩色图像增强算法.在HSI颜色空间中,对饱和度采用线性拉伸处理,利用改进重构的自适应权重引导滤波多尺度Retinex算法获取反射分量,再对反射分量进行Gamma校正来得到最终的亮度分量,最后对图像进行颜色恢复实现最终的低照度彩色图像增强.实验结果表明,该算法具有增强效果好,实时性高,丰富的图像细节和避免颜色失真的优点.