基于改进多尺度形态学的带钢缺陷图像边缘检测

Strip steel surface defects imaging edge inspection based on improved multi-scale morphology

针对目前板带钢表面缺陷在线检测过程中无法准确地检测出所有缺陷边缘问题,根据带钢缺陷的特点,分析了结构元素的选取,提出了一种将多尺度形态学和多结构元素有机结合的边缘检测方法。该方法首先进行多尺度形态学滤波降噪,分别求取0°结构元素、45°结构元素、90°结构元素和135°结构元素带钢缺陷图像边缘;其次通过一定的运算组合,提取多结构边缘;最后对得到的带钢缺陷图像的边缘作二值化处理,再细化边缘得到缺陷图像边缘的最终结果。实验结果表明,该方法较好地解决了边缘检测精度与抗噪性能之间的协调问题,实现了在多个尺度上提取板带钢表面缺陷的边缘。同时能够较好地保留图像中缺陷的边缘细节信息,为带钢表面缺陷在线检测系统中自动分割、缺陷识别等后续处理奠定了基础。

Aiming at detecting all the edge defects in the surface of strip steel, one way to select the structural elements is analyzed. A new edge detection method which combines multi-scale morphology with multi-structure element is proposed. Firstly,multi-scale morphological filtering method is used to get the edge of 4 structural elements;secondly, the image edge is extracted by some operations. Finally, the binary edge image is processed and the final result is obtained by edge thinning. The results show that the method solves the coordination problem between edge detection precision and denoising ability, the edges detail can be preserved well by extracting surface defect of strip steel edge at multiple scales,which establishes a basis for on-line inspection and automatic segmentation.