摘要
板形缺陷是板材矫直过程中确定矫直机弯辊量的重要因素,本文利用图像识别技术采集缺陷板材图像,然后对其进行滤波去噪和边缘检测处理,最后通过径向基网络对特征值进行训练,从训练结果识别板材缺陷,识别率达到了97.5%。有助于中厚板矫直机弯辊量的精确设定,提高了矫直板材的质量。
Shape defects are an important factor to determine roller bending value in the leveling process of plate. The image of shape defects plate is acquired by image recognition technique, then filter de-noising and edge detection technique are used for dealing the images, finally, according to the training result of the radial basis function rbf network to the eigenvalue, plate defects is identified, and accuracy rate reaches 97.5% , which is helpful to set roller bending value accurately and improve plate quality.
作者
柴文婷
胡鹰
李雯琼
王效岗
朱丽丽
CHAI Wen- ting;HU Ying;LI Wen- qiong;WANG Xiao- gang;ZHU Li- li(Taiyuan University of Science and Technology, Taiyuan 030024 , China;Taiyuan Institute of Technology, Taiyuan 030008 , China)
出处
《重型机械》
2017年第6期46-49,共4页
Heavy Machinery
基金
重型机械教育部工程研究中心"重型机械"研究生创新资助计划项目(20172010)
关键词
矫直机
图像识别
径向基网络
leveler
image recognition technique
RBF network
