摘要
提出一种新型的非线性边缘检测技术。首先利用标准的Sobel梯度算子对图像的边缘点进行初选,然后用双同心圆窗口算子在候选边缘点集中寻找真正的边缘。通过与其它成熟的边缘检测技术进行比较。结果显示:该算法获得的边缘图像质量优于传统的Sobel算法;检测出的边缘与Canny和LoG算法可比,计算量却比二者有明显降低。
A new nonlinear technology for edge detection was presented.The algorithm performs a preliminary selection of edges using a standard Sobel bradient operator and uses a two-concentric-circular-window operator to reveal real edges within a set of edge candidates.By comparisons with other well-established technologies for edge detection,the results show that the nonlinear algorithm is better than stadard Sobel operator in terms of visual appearance of edges,and that detected edges are surely comparable with Canny and LoG algorithms,with lower computational load.
出处
《光电子．激光》
EI
CAS
CSCD
北大核心
2003年第12期1332-1335,共4页
Journal of Optoelectronics·Laser
