喷丝板单孔图像的检测算法  

Detecting Algorithm of Spinneret Hole Image

作　　者：张佳全[1] 吴庆妍[2] 冯玮琛[3] 

机构地区：[1]吉林大学物理学院,长春130012 [2]吉林大学公共计算机教学与研究中心,长春130022 [3]吉林大学数学学院,长春130012

出　　处：《吉林大学学报（理学版）》2010年第5期835-839,共5页Journal of Jilin University:Science Edition

基　　金：吉林省科技厅重点项目(批准号:20080317)

摘　　要：运用计算机图像处理方法检测化纤喷丝板单孔孔径和堵塞.使用二维转台和CCD摄像头采集喷丝板单孔图像,并提出一种新的图像检测算法:先对喷丝孔图像进行傅里叶变换,再采用Sobel算子边缘检测和Sobel算子边缘细化法,依据圆孔几何特征判定喷丝孔是否堵塞,依据其图像所占像素量确定堵塞面积比,并求得未堵塞喷丝孔的直径.实际应用表明,所提算法合理、有效,达到了一定的测量精度。The occluded spinneret hole and the diameter of spinneret were detected by means of computer image processing.Meanwhile,an algorithm was introduced to detect the spinneret hole image which was collected by a CCD camera,including image collection,square Gray-scale transformation,edge detection based on Sobel algorithm,binarization based on distance-preserving method,an edge thinning algorithm with Sobel operator.On the basis of geometrical characteristics of the circle,we judged if spinneret holes had been occluded,and determined clogging area proportion via the pixel of the spinneret hole image and solved the diameter of the hole that had not been occluded.The application indicates that this algorithm is reasonable,efficient and accurate.

关 键 词：图像处理 喷丝孔 边缘检测 算法 圆孔识别 面积比 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]