基于多尺度Gabor滤波器的彩色图像边缘检测

Color image edge detection based on multi-scale Gabor filter

为了更好实现对彩色图像进行边缘提取,并保证算法具有更好的稳定性,文章使用多尺度Gabor滤波器的方法,来提取彩色图像边缘和轮廓。构建了3个尺度、16个方向的Gabor滤波器。首先将彩色图像灰度化,利用多尺度Gabor虚部滤波器提取图像灰度变化信息。通过非极大值抑制,并通过高低阈值获取边缘像素点及其候选边缘,最后利用局部边缘连接获取图像边缘轮廓。并将本算法与常用边缘检测算法进行实验性能比较,实验结果表明:提出的算法既能获得较高的定位准确度,又具有很好的噪声鲁棒性,该算法与常用的Roberts等一系列算法相比,检测效果更好,稳定性更强。

In order to realize the edge extraction of color images,and to ensure that the algorithm has better stability.The article used the multi-scale Gabor filter method to extract the edge of the color image and contour,the Gabor filter has three scales and sixteen directions.The first,changes the color image into the gray image,the imaginary part of multi-scale Gabor filter is utilized to obtain the change information of gray image.The second,Through the non-maxima suppression,and then through the high and low threshold to extract the edge pixels and its candidate edge pixels.The last,through the local edge connection to obtain edge contour line.And the performance of this algorithm is compared with the commonly used edge detection algorithm.This experimental conclusion states：this algorithm can obtain high positioning accuracy,and also has good noise robustness.Compared with a series of algorithms,such as Roberts,this method has better detection effect and better stability.