基于投影变换的锂电池喷码字符检测定位

Projection transformation based lithium battery printed characters detection and location

针对锂电池喷码字符检测定位问题,提出一种基于机器视觉的定位方法。该方法首先根据锂电池图像的特点采用梯度特征对喷码字符区域进行初定位,然后通过倾斜校正得到精细定位后的字符区域。在此基础上,将阈值分割后的字符图像与形态学膨胀后的Canny边缘图像进行逻辑与操作,得到待投影字符区域,最后将待投影字符区域进行垂直和水平两次投影得到单个字符的位置信息,完成字符的检测定位。实验结果表明所研究的方法可以较好的实现对单个喷码字符的精确定位,具有一定的应用价值。

According to the task of detecting and locating the lithium battery printed characters,a machine vision based algorithm was proposed.In this algorithm,the printed characters region was firstly coarsely located via the gradient feature and the characteristics of lithium battery images.Then,through the rotation correction we got the fine location of printed characters region.On this basis,a logic AND operation was carried out between the segmented image and Canny edge image after morphological dilation to generate the image region for projection.In the last,the single character was located by two projections with different direction.Experimental results indicated that the proposed method can precisely locate the characters,and it had certain application value.