基于机器视觉的轮廓度检测算法研究

Research on Contour Detection Algorithm Based on Machine Vision

提出一种基于形状模板的匹配方法和奇异值分解的迭代匹配方法的轮廓度误差视觉测量算法,该算法可以检测任意形状的玻璃、塑料或金属薄片产品的轮廓度在加工时所产生的误差。首先,在图像中使用基于Sigmoid函数的亚像素边缘检测方法来拟合产品的亚像素边缘,接着使用基于形状的模板匹配方法来进行粗匹配,在粗匹配的基础上使用SVD算法得到图像的旋转平移矩阵,进行精匹配。最后计算对应点之间的距离作为轮廓度误差。

A visual measurement algorithm of contour error based on shape template matching method and singular value decomposition iteration matching method was proposed.The algorithm can detect the contour error of glass,plastic or metal sheet with arbitrary shape during processing.Firstly,the sub-pixel edge detection method based on Sigmoid function is used to fit the sub-pixel edge of the product.Then the shape-based template matching method is used for rough matching.On the basis of rough matching,the rotation translation matrix of the image is obtained by SVD algorithm,and the fine matching is carried out.Finally,the distance between the corresponding points is calculated as the contour error.