摘要
本文在图像几何畸变校正应用中,阐述了图像边缘检测器中Sobel算子和LoG(拉普拉斯一高斯)算子的边缘检测特性。为了抑制虚假边缘、提高边缘定位精度和克服椒盐噪声带来的干扰,针对LoG算子,提出了选择性平滑方式清除图像中的椒盐噪声,探讨了LoG算子的改进算法,引入自适应滤波器(LMS算法)自动获取最佳空间系数σ值。试验表明,改进算法弥补了LoG算子的不足,提高了检测效果。
This paper describe two different edge detection operators based on Sobel and Laplacian-Gauss in correcting the geometric distortion system. Then analyse the charac- teristic of these operators using the digital simulation technique at different noise environ- ments. Removed shot noises by selective smoothness and using adaptive filter auto-adjust Gaussian space coefficient to avoid the defect of LoG.
关键词
畸变校正
边缘检测
LOG算子
自适应滤波
Image geometric rectification
Edge detection
LoG operator
Adaptive filter
