摘要
由于工件表面照度不均、纹理、毛刺等因素的影响,挠性接头薄筋厚度在线测量图像具有背景反射强、噪声干扰大、对比度差等特征。本文根据在线测量图像在灰度、梯度和空间信息等各方面的特点,提出了基于背景区域生长的外挠分割法、灰度梯度的内挠分割法和基于过渡区的内外挠通用分割法。工程测试结果表明,使用文中提出图像分割方法的重复性精度可达0.5μm。
Because of nonuniform illumination,texture,burrs on surface of metal workpiece,an image with strong reflection background,high noise,poor contrast is captured in on-line measurement for tendon thickness of flexible connector. According to the image's feature in gray,gradient and spatial information,three image segmentation methods are presented,i.e.,background region growth for outer flexible connector,gray gradient for inner flexible connector and transition region for both outer and inner flexible connector.Based on the measurement result,0.5 μm repeatability is gained using the image processing method mentioned in the article.
出处
《光学精密工程》
EI
CAS
CSCD
北大核心
2005年第z1期153-157,共5页
Optics and Precision Engineering
关键词
在线测量
图像分割
边缘检测
on-line measurement
image segmentation
edge detection
