基于全方位和多尺度结构元数学形态学的木材缺陷图像边缘检测

Edge Detection of Wood Defect Image Based on Mathematical Morphology of Omni Directional Multi-scale Element Algorithm

提出一种基于全方位、多尺度结构元的数学形态学图像边缘检测算法。针对图像中噪声和边缘形态不同,定义了全方位、多尺度的形态学结构元素,并通过形态学运算的加权组合,构造了全方位、多尺度的边缘检测算法。在针对木材缺陷图像的仿真实验中,该方法与经典的边缘检测算子相比不仅具有很好的边缘提取能力,而且有很强的抗噪性。

Based on the mathematical morphology of omni directional multi-scale element, an algorithm of image edge detection was proposed in this paper. Mathematical morphology of omni directional multi-scale dement was defined in order to suppress noise and adapt to different edges in the image. An approach of image edge detection based on the morphology of omni directional multi-scale element was constructed by weight adding combination of morphological operation. The results of simulation demonstrate of wood defect images showed that the method performed better not only in edge detection but also in noise-suppression than classical edge detection operators.