摘要
针对珍珠大小分选时,人工劳动强度大、精度低,而机械分选容易对珍珠表面造成损伤的问题,提出基于机器视觉的珍珠大小检测方法.构建背光成像方式减少珍珠表面光泽和纹理的影响,采用直方图增强和高斯滤波等预处理算法,提高珍珠图像对比度、降低噪声。以最大类间方差法分割图像背景,通过基于行扫描的边缘轮廓提取方法提取珍珠边缘轮廓特征,计算珍珠的平均直径。实验结果表明,该方法对珍珠大小检测误差小于0.25 mm,检测率为50颗/min。
In order to solve the problem of pearl size time-sharing,manual sorting labor intensity is high,sorting accuracy is low,and mechanical sorting is easy to damage the pearl surface,a pearl size detection method based on machine vision is proposed.In order to reduce the influence of luster and texture of pearl surface,a backlight imaging method is constructed.Histogram enhancement and Gaussian filtering are used to improve the contrast of Pearl image and reduce the noise.The maximum inter class variance method is used to remove the image background.The edge contour feature of pearl is extracted by the edge contour extraction method based on line scanning,and the average diameter of pearl is calculated.The experimental results show that the average diameter measurement error of the device is less than 0.25 mm,and the detection rate was 50 particles/min.
作者
高凯
焦航
严楠
GAO Kai;JIAO Hang;YAN Nan(College of Mechanical and Electrical Engineering,Xi'an Polytechnic University,Xi'an Shaanxi 710600,China;Ningbo Polytechnic,Ningbo Zhejiang 315800,China)
出处
《信息与电脑》
2021年第10期24-26,共3页
Information & Computer
基金
宁波市产业技术创新重大项目(项目编号:2018B10010)。
关键词
机器视觉
珍珠
轮廓提取
直径
machine vision
pearl
contour extraction
diameter
