基于分数阶微分的边缘检测

Edge Detection Based on Fractional Differential

为了提高图像边缘提取的信噪比,更有效和准确检测图像边缘,由信号的微分特性得出分数阶微分算子较传统1阶和2阶微分算子具有更高的信噪比,然后根据经典的G-L分数阶微分定义推导出的分数阶差分方程,构建了近似的分数阶Tiansi模板。实验证明,基于分数阶微分的边缘提取算子,可以有效提取边缘和有比传统的算子更高的信噪比。

In order to improve signal noise ratio （SNR） of edge extraction, and to detect edge more effectively and exactly, according to fractional order differential difference function which was deduced from classical fractional differential G- L definition, an approximate fractional order differential Tiansi module was constructed. Experiments show that fractional differential operator can effectively extract edge information and has higher SNR than traditional operators.