摘要
提出一种基于机器视觉与人工神经网络的铸件表面缺陷识别方法,采用计算机图像技术采集和处理生产线上的铸件表面图像信号,采用改进的BP神经网络算法对图像信号进行缺陷识别分析。该方法在为某厂研制的铸件表面缺陷检测系统中使用后,作业时耗平均降低4min/工件,表面缺陷检出准确率平均提高15%,实践表明本文方法是可行、有效的。
This paper proposes an artificial neural network recognition method for cast product surface defects based on machine vision.Computer image technologies are used for the image process of cast product surface defect,and an improved BP algorithm is used for the defects images recognition.This method has been used in a cast product face defect-detecting system,and the job time reduce by 4min per component,while the defects correct inspecting rate improve 15%.The results show that this method is effective.
出处
《重型机械》
2004年第2期51-54,共4页
Heavy Machinery
