非盲模糊图像边缘结构细节增强算法设计  

Design of Algorithm for Enhancing Edge Structure Details in Non-Blind Blurred Images

作　　者：刘昊 朱勇军 LIU Hao;ZHU Yong-jun(China People's Police University,Beijing 102300,China)

机构地区：[1]中国人民警察大学,北京102300

出　　处：《计算机仿真》2025年第2期210-214,共5页Computer Simulation

摘　　要：由于模糊图像清晰度较低、目标边缘轮廓易与背景混为一体,使得目标与背景间难以界定,为此,提出一种非盲边缘结构细节增强算法。利用小波降噪算法去除非盲模糊图像中的噪声点,仅保留边缘特征点,提取特征;设定一个判断边缘点与非边缘点的值,使图像的类内方差最小、类间方差最大;利用多尺度高斯模糊对图像作相减处理,得到多个尺度下边缘结构细节特征,经权值相加后形成边缘结构细节特征图,与原始图像特征融合后即可得到边缘结构细节增强处理后的图像。实验结果表明,所提方法可以针对非盲模糊图像实现理想的边缘结构细节增强,保留更多的原始图像特征,使得增强后的图像在结构、像素灰度以及对比度方面与原始图像接近。Due to the low clarity of the fuzzy image,the target edge contour is easily mixed with the background.It is difficult to distinguish between target and background.Therefore,an algorithm for enhancing non-blind edge details was proposed.Firstly,the wavelet denoising algorithm was used to remove noise points from the non-blind fuzzy image and retain edge feature points.And then,the features were extracted.Moreover,a threshold was set to distinguish between edge points and non-edge points,thus minimizing the intraclass variance and maximizing the interclass variance of the image.Furthermore,using multi-scale Gaussian blur to perform subtraction processing on images,multiple scales of edge structure detail features are obtained.After weighted addition,a detail feature map was formed,which could be fused with the original image features to obtain an enhanced image.Experimental results prove that the proposed method can achieve the ideal enhancement for edge details of non-blind fuzzy image,and retains more original features,thus making the enhanced image close to the original image in terms of structure,pixel grayscale and contrast.

关 键 词：非盲模糊图像 边缘结构细节增强 多尺度高斯模糊 最佳阈值 类内方差 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]