摘要
为实现丝饼毛羽的自动检测,提出基于机器视觉的丝饼毛羽检测方法,通过寻找毛羽轮廓点中凸包位置实现丝饼毛羽检测。首先构建具有特定结构特征的卷积核,利用该卷积核对原始图像进行卷积运算,从而获取毛羽特征,并采用阈值处理对其进行二值化;然后对二值化后的毛羽进行轮廓检测,继而筛选轮廓点以减少运算时间;最后利用单方向凸包算法对满足条件的轮廓点进行检测,实现对丝饼毛羽的定位及计数。运用3类典型丝饼毛羽对检测方法进行验证,结果表明,该方法可有效地实现对丝饼毛羽准确定位并计数,且对不同背景的毛羽图片有较强的适应性。
In order to realize the automatic detection of the hairiness of yarn packages,a method for detecting the hairiness of yarn packages based on machine vision was proposed.Firstly,the convolution kernel was constructed,and the hairiness characteristics were obtained by convolving the original image with the convolution kernel,and binarization was performed using threshold processing.Secondly,the contours of the binarized hairiness were detected,and then the contour points were screened to reduce the computing time.Finally,the unidirectional convex hull detection was applied to the outline points which meet the screening conditions,and then the hairiness of the yarn packages was located and counted.Three kinds of typical yarn packages were used to verify the method.The experimental results show that the method can position and count the hairiness number of the yarn packages accurately,and has strong adaptability to the hairiness images of different backgrounds.
作者
景军锋
郭根
JING Junfeng;GUO Gen(School of Electronics and Information,Xi′an Polytechnic University,Xi′an,Shaanxi 710048,China)
出处
《纺织学报》
EI
CAS
CSCD
北大核心
2019年第1期147-152,共6页
Journal of Textile Research
基金
国家自然科学基金项目(61301276)
陕西省重点研发计划项目(2017GY-003)
关键词
丝饼毛羽
凸包检测
卷积运算
轮廓检测
机器视觉
yarn package hairiness
convex hull detection
convolution operation
contour detection
machine vision
