【摘要】 丝网印刷样板种类和样式多,每片样板上待测目标也较多,不宜采用一致的梯度阈值参数定位边缘实施测量;此外,传统视觉检测采用定位矩形框建立测量坐标系的方式会引入测量误差,因此,提出一种基于机器视觉的丝网印刷样板的尺寸测量方法。采用一种由粗到细的测量策略,借助每一类待测目标的模板进行信息统计,针对性地设置阈值参数,改善测量精度。在测量阶段,利用基于图像金字塔和归一化互相关函数相结合的分层匹配算法实现多个待测目标的粗糙定位,再使用由模板信息统计获取的阈值参数进行边缘精细定位,建立局部坐标系,完成测量。实验结果表明:在同等实验条件下,此方法能够有效提高测量精度,算法改进后测量结果的平均相对误差由原来的4.02%降到1.47%,并且无需用户介入调整测量参数,适用于不同种类丝网印刷样板的灵活测量。更多还原

【Abstract】 Many types and styles of screen printing samples are used in screen printing plate and there are also many targets to be tested on each sample plate,which make the gradient threshold parameter to locate the edge to perform measurement unsuitable. In addition,the method of using the positioning rectangular frame to establish the measurement coordinate system in the traditional visual inspection can cause measurement error. A method for measuring the size of a screen printing template based on machine vision was proposed. A coarse-to-fine measurement strategy was adopted to analyze the information of each type of target to be measured,and the threshold parameters were set in a targeted manner to improve the measurement accuracy. During the measurement process,a hierarchical matching algorithm based on image pyramid and normalized cross-correlation function was used to achieve rough positioning of multiple targets to be measured,and then the threshold parameters obtained by template information statistics were used to perform fine edge positioning to establish local coordinates to complete the measurement. The experimental results showed that the proposed method can effectively improve the measurement accuracy under the same experimental conditions. The average relative error of the improved algorithm is reduced from 4. 02% to 1. 47%,and the measurement parameters need not be adjusted by users. It is suitable for flexible measurement of different types of screen printing templates.更多还原