摘要
Abstract
In the process of image enhancement,it is usually necessary to denoise and smooth the image,which often leads to blurring or weakening of the image edges,however.Therefore,a method for multi-scale detail enhancement of images under adaptive guided filtering is proposed.Adaptive guided filtering is adopted to filter the image,which effectively preserves the edge details of the image while denoising and smoothing the image.Subsequently,the filtered image is split into detail layer and base layer.The multi-scale detail enhancement techniques are adopted to improve the sharpness of local details of the images in detail layer,and the Gaussian kernel functions of different scales are applied for convolution operations to generate image sequences with different levels of smoothness,thus performing multi-scale detail enhancement.The base layer with enhanced contrast and the detail layer with enhanced details are fused according to a certain weight,so as to obtain multi-scale detail enhanced images.The experimental results show that the proposed method can enrich the detailed information of the image,ensure the integrity and natural transition of the overall structure of the image,and make the color of the image more full,and that the information entropy and average gradient indicators have shown excellent performance,which strongly proves the effectiveness of the proposed method in multi-scale detail enhancement of images.关键词
引导滤波/自适应/多尺度/细节增强/高斯核函数/加权融合/平滑处理Key words
guided filtering/adaption/multi-scale/detail enhancement/Gaussian kernel function/weighted fusion/smooth processing分类
信息技术与安全科学
