基于离散小波技术的图像边缘特征提取  

Image Edge Feature Extraction Based on Discrete Wavelet

作　　者：刘春科 徐工[1] 卢鑫 LIU Chunke;XU Gong;LU Xin(College of Architectural Engineering,Shandong University of Technology,Zibo 255000,China;Shandong Gold Mining(Linglong)Co.,Ltd.,Zhaoyuan 265400,China)

机构地区：[1]山东理工大学建筑工程学院,山东淄博255000 [2]山东黄金矿业(玲珑)有限公司,山东招远265400

出　　处：《测绘与空间地理信息》2022年第10期41-43,共3页Geomatics & Spatial Information Technology

基　　金：山东省自然科学基金(ZR2018LD003)资助。

摘　　要：针对提取图像边缘信息完整度问题。传统边缘检测算子检测噪声信息不敏感,对图像进行边缘提取时易把噪声因子定义成边缘点,无法完整提取边缘信息。基于小波变换具有探测信号奇异性的能力和小波的多分辨率特性,采用二维小波变换进行图像边缘特征提取,通过二维小波分解提取图像的小波系数,迭代寻找其局部极值点,对其进行阈值处理并进行小波逆变换能够获取图像的边缘信息。实验结果表明,利用二维小波变换提取图像边缘特征信息能够获得更高的完整度。Aiming at the integrity problem of extracting image edge information,the traditional edge detection operator is not sensitive to noise information.When extracting the edge of the image,it is easy to define the noise factor as edge points,so it is unable to extract the edge information completely.Based on the ability of wavelet transform to detect signal singularity and the multi-resolution characteristics of wavelet,the two-dimensional wavelet transform is used for image edge feature extraction.The wavelet coefficients of the image are extracted through two-dimensional wavelet decomposition,the local extreme points are found iteratively,the threshold is processed and the inverse wavelet transform is carried out to obtain the edge information of the image.Experimental results show that using two-dimensional wavelet transform to extract image edge feature information can obtain higher integrity.

关 键 词：边缘提取 经典边缘检测算子 小波变换 阈值函数 局部极大值 

分 类 号：P209[天文地球—测绘科学与技术]