一种小波变换的图像边缘检测方法

An image edge detection method based on wavelet transform

边缘检测可以大幅度地减少原图的数据量,消除许多没有意义的信息,保留图像重要的结构属性。本文使用Haar小波滤波器,利用小波变换将原灰度图像进行一层分解,产生4个子图像,对水平高频子图像和垂直高频子图像分别进行Sobel算子检测处理,将低频子图像和对角细节图像分别置零处理,然后对处理过的4个子图进行重构,得到的图像与Canny算子对原图提取的特征图像进行了对比。实验结果表明,该方法得到的结果较好。

Edge detection can drastically reduce the amount of original data,to eliminate a lot of meaningless information,image important structural properties. This article use the Haar wavelet filter,using wavelet transform to the original gray image layer decomposition,the image produced by four,levels of high frequency images and vertical Sobel operator to detect high frequency sub images respectively,and the low frequency image and zero diagonal detail image processing,then to reconstruct the processed by four images and Canny operator compares the features of image were extracted from the original image. The experimental results show that the method of the result is good.