摘要
基于梯度的边缘检测方法在处理含有弱边缘的图像时存在问题:选取高梯度阈值会遗漏弱边缘,而选取低梯度阈值将会出现虚假边缘点。本文提出了一种基于非线性灰度变换的弱边缘检测方法。该方法首先对图像作非线性变换,沿边缘方向作平滑后再计算梯度值并进行非最大化抑制,最后参照梯度直方图选取合适的梯度阈值来标识边缘点。实验结果表明本文提出的方法能有效地检测出弱边缘。
Many gradient-based edge detection algorithms have problems with images that contain weak edges： a tight threshold may fail to detect weak edges whereas a loose threshold may cause lots of false edges.A weak edge detection method based on nonlinear transform of gray level is therefore proposed.It first enhances the raw image using a nonlinear function,followed by a smoothing along the edge direction.It then calculates the gradient magnitude and makes the non-maximal suppression.Finally it labels the edge points using a threshold with Reference to the variance of the gradient histogram.Experimental results show that the proposed algorithm can detect weak edges effectively.
出处
《中国体视学与图像分析》
2011年第3期232-236,共5页
Chinese Journal of Stereology and Image Analysis
关键词
弱边缘
非线性变换
阈值
weak edge
nonlinear transform
threshold
variance from median
