基于亚像素边缘检测的LED芯片定位算法研究
|
摘要
LED芯片的定位精度直接决定了LED制造装备的生产质量和效率,为了提高LED芯片的定位精度,提出了一种基于亚像素边缘检测的芯片定位算法。该算法首先采用Gamma变换方法增强图像的对比度,并利用Blob算法获取芯片有效区域;接着采用Canny算法进行芯片亚像素边缘轮廓的提取;最后,通过拟合芯片边缘轮廓,获取芯片位置中心,完成芯片位置的精确识别。该算法不需要人工训练模板进行匹配,提高了边缘提取的定位精度,实验表明,该算法能在平均4ms内完成一颗芯片的识别,且重复精度达到0.1pixel,满足LED芯片高速高精度定位需求。
The position of LED chips directly determines the accuracy of LED detection equipment.To improve the accuracy of the chip positioning,and to minimize the effect of template matching,the sub-pixel positioning algorithm based on the edge extraction was put forward in this paper.Firstly,the algorithm uses the gamma transform to enhance the contrast between the chip and the background area.Then,it uses Blob algorithm to get the chip area of interest.Finally,the Canny edge algorithm is used to extract the sub-pixel edge profile and the chip edge profile was fitted to locate the chip position.The proposed algorithm does not need artificial training to match template,so the position accuracy of the edge extraction can be improved.Through actual measurement and test,the results show that the image repeatability positioning precision is up to 0.1pixel,and the high-speed and high-precision positioning demands for LED chip has been realized.
The position of LED chips directly determines the accuracy of LED detection equipment.To improve the accuracy of the chip positioning,and to minimize the effect of template matching,the sub-pixel positioning algorithm based on the edge extraction was put forward in this paper.Firstly,the algorithm uses the gamma transform to enhance the contrast between the chip and the background area.Then,it uses Blob algorithm to get the chip area of interest.Finally,the Canny edge algorithm is used to extract the sub-pixel edge profile and the chip edge profile was fitted to locate the chip position.The proposed algorithm does not need artificial training to match template,so the position accuracy of the edge extraction can be improved.Through actual measurement and test,the results show that the image repeatability positioning precision is up to 0.1pixel,and the high-speed and high-precision positioning demands for LED chip has been realized.
引文
[1]李钊宝.LED晶片检测中的电极定位图像处理及算法研究[D].杭州:中国计量学院,2013.Li Zhaobao.Image processing and algorithm research on LED chip electrode positioning[D].Hangzhou:China Jiliang University,2013.
[2]吴晓.一种LED芯片椭圆定位方法[J].红外与激光工程,2013,42(z1):98-102.Wu Xiao.Positioning system of LED chips based on ellipse[J].Infrared and Laser Engin.,2013,42(z1):98-102.
[3]吴晓.LED芯片粗精定位系统[J].红外与激光工程,2013,42(9):2478-2484.Wu Xiao.LED chips coarse and fine locating system infrared and laser engineering[J].Infrared and Laser Engin.,2013,42(9):2478-2484.
[4]李蒙,方家乐,朱颖合.LED芯片图像的势函数标记分水岭分割[J].机电工程,2010,27(7):81-84.Li Meng,Fang Jiale,Zhu Yinghe.Marker-controlled watershed segmentation of potential function for LED die image[J].J.of Mechanical&Electrical Engin.,2010,27(7):81-84.
[5]许新征.丁世飞,史忠植,等.图像分割的新理论和新方法[J].电子学报,2010,38(2):76-82.Xu X Z,Ding S F,Shi Z Z,et al.New theories and methods of image segmentation[J].Acta Electronica Sinica,2010,38(2):76-82.
[6]王庆海,赵凤霞,李纪峰,等.基于Blob分析的玻璃纤维织物缺陷检测方法研究[J].郑州大学学报(工学版),2015,36(6):90-93.Wang Qinghai,Zhao Fengxia,Li Jifeng,et al.Research on glass fiber fabric defect detection method based on Blob analysis[J].J.of Zhengzhou University(Engineering Science),2015,36(6):90-93.
[7]李钊宝,汪新新,贾茜媛.基于机器视觉的工件尺寸高精度测量方法[J].计算机测量与控制,2015,23(6):1866-1870.Li Zhaobao,Wang Xinxin,Jia Qianyuan.A method of highprecision machine vision measurement of size of workpiece[J].Computer Measurement&Control,2015,23(6):1866-1870.
[8]李蒙.LED芯片检测系统视觉图像分析技术研究[D].杭州:杭州电子科技大学,2010.Li M.A study on the vision image analysis technology of LED chip detection system[D].Hangzhou:Hangzhou Electronic Science and Technology University,2010.
[2]吴晓.一种LED芯片椭圆定位方法[J].红外与激光工程,2013,42(z1):98-102.Wu Xiao.Positioning system of LED chips based on ellipse[J].Infrared and Laser Engin.,2013,42(z1):98-102.
[3]吴晓.LED芯片粗精定位系统[J].红外与激光工程,2013,42(9):2478-2484.Wu Xiao.LED chips coarse and fine locating system infrared and laser engineering[J].Infrared and Laser Engin.,2013,42(9):2478-2484.
[4]李蒙,方家乐,朱颖合.LED芯片图像的势函数标记分水岭分割[J].机电工程,2010,27(7):81-84.Li Meng,Fang Jiale,Zhu Yinghe.Marker-controlled watershed segmentation of potential function for LED die image[J].J.of Mechanical&Electrical Engin.,2010,27(7):81-84.
[5]许新征.丁世飞,史忠植,等.图像分割的新理论和新方法[J].电子学报,2010,38(2):76-82.Xu X Z,Ding S F,Shi Z Z,et al.New theories and methods of image segmentation[J].Acta Electronica Sinica,2010,38(2):76-82.
[6]王庆海,赵凤霞,李纪峰,等.基于Blob分析的玻璃纤维织物缺陷检测方法研究[J].郑州大学学报(工学版),2015,36(6):90-93.Wang Qinghai,Zhao Fengxia,Li Jifeng,et al.Research on glass fiber fabric defect detection method based on Blob analysis[J].J.of Zhengzhou University(Engineering Science),2015,36(6):90-93.
[7]李钊宝,汪新新,贾茜媛.基于机器视觉的工件尺寸高精度测量方法[J].计算机测量与控制,2015,23(6):1866-1870.Li Zhaobao,Wang Xinxin,Jia Qianyuan.A method of highprecision machine vision measurement of size of workpiece[J].Computer Measurement&Control,2015,23(6):1866-1870.
[8]李蒙.LED芯片检测系统视觉图像分析技术研究[D].杭州:杭州电子科技大学,2010.Li M.A study on the vision image analysis technology of LED chip detection system[D].Hangzhou:Hangzhou Electronic Science and Technology University,2010.