Adaptive Multi-Threshold Image Segmentation Based on Deep Learning and Potential Function Clustering
In order to improve the contrast enhancement effect of remote sensing blurred images and increase clarity,a method based on mean filtering for remote sensing blurred image contrast enhancement is proposed.Firstly,a fast median adaptive mean filtering algorithm is used to denoise the entire remote sensing blurred image.Secondly,combining the fractal self-similarity feature of remote sensing image edge and the change of gray scale gradient,the edge points of the image are extracted.On this basis,the whole area of the image is divided into bright areas and dim areas.Finally,the detail preserving mapping algorithm and perceptual contrast mapping method are used to enhance the contrast of the two regions,respectively,and the overall contrast of the remote sensing blurred image achieving color restoration of the image.The experimental results show that the proposed method can effectively denoise images,with an absolute mean difference of less than 0.85,and exhibits good performance in enhancing image contrast and clarity.
mean filteringremote sensing imagecontrast enhancementdenoisingedge extractioncolor restoration
国家科技期刊平台
NSTL
万方数据
