结合图像增强和改进Canny算子的遥感图像边缘检测  

Remote Sensing Image Edge Detection Combined with Image Enhancement and Improved Canny

作　　者：陈顺 孟青青 李登峰 CHEN Shun;MENG Qingqing;LI Dengfeng(School of Mathematics and Computer,Wuhan Textile University,Wuhan 430200,China)

机构地区：[1]武汉纺织大学数学与计算机学院,武汉430200

出　　处：《河南大学学报（自然科学版）》2020年第5期623-630,共8页Journal of Henan University:Natural Science

基　　金：国家自然科学基金项目(61471410)。

摘　　要：针对遥感图像对比度低和含有大量混合噪声,采用传统Canny算子难以有效地提取遥感图像边缘信息,本文提出一种结合图像增强和改进Canny算子的遥感图像边缘检测算法.首先利用NSCT(非下采样轮廓波变换)将图像分为高低频子图,采用高斯滤波器对高低频分量进行去噪预处理,并重构得到预处理后的图像;然后运用Top-hat变换对预处理后的图像进行图像增强突出细节信息;最后采用改进Canny算子得到边缘图像,并去除部分孤立点得到最终边缘图像.实验结果表明,本文算法相比传统Canny算子具有更好的去噪效果,有效增强了图像对比度,可以保留遥感图像更多真实边缘,改进了传统Otsu方法,提高计算效率.Aiming at the low contrast of remote sensing image and containing a lot of mixed noise,it is difficult to extract the edge information of remote sensing image effectively by traditional Canny operator.This paper proposes a remote sensing image edge detection algorithm combining image enhancement and improved Canny operator.Firstly,the image is divided into high and low frequency subgraphs by NSCT(non-downsampledcontourlet transform).The Gaussian filter is used to denoise the high and low frequency components,and the preprocessed image is reconstructed.Then the Top-hat transform is applied.The pre-processed image is used to enhance the detail information.Finally,the improved Canny operator is used to obtain the edge image,and some isolated points are removed to obtain the final edge image.Experimental results show that the proposed algorithm has better denoising than the traditional Canny operator.The effect effectively enhances the image contrast,preserves more real edges of the remote sensing image,and improves the traditional Otsu method to improve computational efficiency.

关 键 词：CANNY算子 TOP-HAT变换 OTSU方法 边缘检测 

分 类 号：O174.2[理学—数学]