改进双边滤波与平均γ矫正的图像增强

Image Enhancement Based on Improved Bilateral Filtering and Averageγ Correction

为克服现有的图像增强算法存在图像失真和低对比度的不足,提出了一种改进双边滤波与平均γ矫正的图像增强算法。鉴于双边滤波具有良好的边缘保持性能,对其进行改进,并将其作为Retinex的中心环绕函数,以获取原始图像的光照图像和反射图像。分别对多尺度Retinex分解的光照图像进行γ矫正,以增强各尺度光照图像的光照效果。然后对γ矫正后的各尺度光照图像取均值图像以及对各尺度的反射图像取均值图像。最后将均值光照图像和反射图像进行重构输出增强图像。实验结果表明,相对于部分最新增强算法,所提出方法的图像增强性能更好,增强图像的对比度和清晰度更高。

In order to overcome the defects of image distortion and low contrast in existing image enhancement methods,an image enhancement method based on improved bilateral filtering and averageγcorrection is proposed.In view of the good edge preserving performance of bilateral filtering,it is improved and taken as the central surround function of Retinex,so as to obtain the illumination image and reflection image of the original image.The illumination images decomposed by multi-scale Retinex areγcorrected respectively,so as to enhance the illumination effect of each scale illumination image.Then,the average illumination image is obtained from allγcorrected illumination images of various scales,and the average reflection image is obtained from all reflection images of various scales.The final enhanced image is obtained from the average illumination image and reflection image through re-construction.The experimental results show that compared with some recent enhancement algorithms,the proposed method shows better image enhancement performance,and the contrast and clarity of the enhanced image got by using this method are higher than those of images got by using other enhancement methods.