摘要
针对轴承图像的特点,提出一种在图像中检测轴承内外径尺寸的新方法。首先,构造2个可变卡尺,分别将轴承内外边缘覆盖;其次,将卡尺所在圆环映射到矩形图像中,并建立其映射关系;然后,在映射图像中提取每个卡尺对应轴承区域的边界,取一点作为边界点;最后,通过映射关系找到边界点对应原图的坐标,并用这些点进行最小二乘稳健拟合获得圆参数。试验表明:该方法检测精度高、速度快、灵活性强,可满足轴承在线检测的需求。
Aiming at characteristics of bearing images,a new method is proposed for dimension detection of inner and outer diameter of bearings in image. Firstly,two variable calipers are constructed to cover inner and outer edges of bearings respectively. Secondly,the circle where the caliper located is mapped into rectangular image,and the mapping relationship is established. Then,each caliper in mapping image corresponding to boundary of bearing area is extracted; a point is selected as boundary point. Finally,the coordinates of boundary points corresponding to original image is found out through mapping relationship,and the circle parameters are obtained through least squares robust fitting of these points. The experiment shows that the method has high detection precision,fast speed and strong flexibility,the online detection requirements of bearing are met.
出处
《轴承》
北大核心
2018年第2期43-45,50,共4页
Bearing
关键词
滚动轴承
可变卡尺
映射
稳健拟合
rolling bearing
variable caliper
mapping
robust fitting
