自动分检轮胎的字符预处理研究

Study on Character Pretreatment in Tire Classification Automation

自动化是现代企业生产必不可少的,在轮胎生产中自动化分检轮胎可以有效提高生产率;自动化分检轮胎过程中轮胎的型号是一个重要的分检信息;为便于在轮胎自动化分检中正确的识别轮胎型号,对现场采集的图像信息必须进行一系列的预处理;重点研究了型号字符在二值化时采用改进的Otsu阈值结合水平、垂直、交叉边缘提取法;提出一种类似水平投影法的圆弧投影法分割轮胎字符;再采用Hough变换对单个字符旋转矫正;最后得到规则的字符图像;结果表明,通过以上处理字符信息能有效的提取出来。

Automation is necessary in modern company. The productivity can be increased by automatic classification. Tire type is important detected information in tire type classification. In order to recognize the tire type correctly in automatic classification, the image must be pretreatmented. A improved Otsu arithmetic and edge detection method are presented in binary image processing; A new cycloid projection method that like the horizontal projection is put forward; And then, revolving correction is used to every words with Hough arithmetic. At last, word image of trim is given. The result shows that information of type can be received by above mentioned methods. The information can be supplied to the subsequent classification.