摘要
运用机器视觉的相关理论,综合图像处理及模式识别技术,利用数字电子计算机实现理论算法的程序化,最终判别活塞装配是否正确。利用模板匹配、阈值分割、字符分割、特征提取和模式识别等算法,通过适当的处理和分析,对在工作现场采用CCD摄像机采集的活塞图像中的目标文字进行自动检测与识别,最终实现了对活塞装配的机器自动识别。试验证明,该方法具有一定的实用性和准确性。
The application of computer vision theory, digital image processing and pattern recognition technology to inspect the assembling of piston was described. The piston images were collected by CCD industrial cameras in the assembly-line. Then the object images would be inspected and recognized automatically by using proper disposing and analyzing algorithms such as template matching, threshold segmentation , character segmentation, feature extraction and pattern recognition and so forth. The result of the experiments showed that the system was practical and veracious to some extent.
出处
《无损检测》
2004年第6期280-283,共4页
Nondestructive Testing
关键词
活塞装配
字符识别
特征提取
信息处理
Piston assembling
Character recognition
Feature extraction
Information processing
