摘要
本文把经典的Prewitt边缘检测算法的不足之处做了改进,在检测之前先通过中值滤波(MF)去掉了噪声和干扰的影响,然后再通过八个方向模板的Prewitt算法对图像进行边缘检测,最后的阈值选取利用了Ostu法,检测得到的图像边缘比改进前的Prewitt算法、Robert、Canny、Laplacian等算法检测的图像边缘完整、清晰,而且也去除了伪边缘。
This article improves the deficiencies of the classic Prewitt algorithm of edge detection. We firstly remove the noise and interference through median filter(MF) before the detection, then take the edge detection for images by using Prewitt algorithm of eight direction template, and finally select the threshold by using the Ostu method. The detected image edge is more complete and clearer than that of the Prewitt algorithm, Robert algorithm, Canny algorithm and Laplacian algorithm, and also can remove false edges.
出处
《科技广场》
2015年第1期57-60,共4页
Science Mosaic