Abstract: Aiming at the problem of difficult recognition for stacked texture-less workpieces in industrial field, an improved geometric template matching algorithm featuring holes on the workpiece surface is proposed, and the hinge is taken as an example for workpiece recognition. The weighted average method is used to gray-scale the color image, and then the Canny algorithm is used for edge detection. The rotating calipers algorithm is used to obtain the minimum area bounding rectangle of the contour, and geometric constraints are performed to obtain the hole contour corresponding to the hole, and randomized incremental algorithm calculates the minimum enclosing circle of the hole contour to obtain the coordinates of hole feature center. The proposed improved geometric template matching algorithm is used to identify the workpieces according to the geometric constraints between the hole features. The fake workpieces are removed according to whether there is edge between the hole features. The experimental results show that the proposed algorithm has a good recognition effect on stacked texture-less workpieces with holes. The recall rate of workpieces recognition is 98.3% and the false positive rate is 0.9%. It provides a method for the recognition of texture-less workpieces with holes.

Key words: stacked texture-less workpieces, workpiece recognition, hole feature, improved geometric template matching, geometry constraint

摘要： 针对目前工业现场弱纹理堆叠工件识别困难的问题，提出一种以工件表面孔洞为特征的改进几何模板匹配算法，以合页为例进行工件识别。首先采用加权平均法对彩色图像进行灰度化处理，再采用Canny算法进行边缘检测；其次采用旋转卡壳算法求取轮廓的最小面积外接矩形，进行几何约束后得到孔洞对应的孔轮廓，并采用随机增量法计算孔轮廓的最小外接圆得到孔特征圆心坐标。采用提出的改进几何模板匹配算法，即根据孔特征之间的几何约束进行工件识别，根据孔特征之间是否存在边缘剔除误识别工件。实验结果表明，提出的算法对带孔弱纹理堆叠工件具有良好的识别效果，工件识别查全率为98.3%，误检率为0.9%，为带孔弱纹理工件的识别提供方法。

关键词: 弱纹理堆叠工件, 工件识别, 孔特征, 改进几何模板匹配, 几何约束