摘要
针对经典Sobel模板梯度方向信息不充分和抗噪能力较差的缺点,对其进行改进,并结合灰色关联分析方法,提出一种基于改进Sobel模板的灰色关联分析边缘检测新算法。该方法将经典的Sobel模板从2个方向的3×3阶的规则矩阵扩展成4个方向的不规则矩阵,利用改进的Sobel模板作为参考序列,通过灰色系统理论中的灰色关联分析进行边缘检测。实验结果表明该方法可以获得很好的边缘特性,并具有较强的抗噪能力。
Against the defect of the traditional Sobel operator with the insufficience of gradient information and lack of noise-immuity, a new edge detection algorithm is presented based on the grey relational analysis with improvement. The proposed algorithm puts forward irregular matrixes of detection templates with four directions instead of original 3 × 3 matrixes of detection templates with two directions. The improved Sobel operator is acted as referenced sequence, then the grey degree is calculated to detect edge. Experimental results show that more edge information is observed and the anti-noise capability is upgraded by the method.
出处
《电视技术》
北大核心
2012年第9期22-23,43,共3页
Video Engineering
基金
江苏省重点实验室项目(10KFJJ005)
