基于机器视觉的碳毡装配检测算法系统设计
Design of carbon felt assembly inspection algorithm system based on machine vision
冯雨 1方成刚 1张文东 2程丽娟3
作者信息
- 1. 南京工业大学 机械与动力工程学院,江苏 南京 211800
- 2. 北京西融储能科技有限公司,北京 102600
- 3. 镇江艾牧智能科技有限公司,江苏 镇江 212000
- 折叠
摘要
针对碳毡装配过程中碳毡占位、边缘毛刺复杂,人工检测效率低、成本高、接触产生不良影响等问题,设计了一种基于机器视觉的碳毡装配检测算法系统.使用Python语言调用相机获取图像并进行预处理.随后,对图像应用改进的Sobel算子进行初步的边缘检测,获取边缘的粗略坐标信息.接着,采用改进的插值法亚像素边缘检测算法对这些粗略坐标进行细化,得到更精确的坐标信息.然后,利用最小二乘法对这些精确坐标进行直线拟合,得到图像中的四个角点.最终,通过对这些角点的位置关系进行判断和分析.实验结果表明,本系统可以稳定进行碳毡装配检测,具有一定的鲁棒性.
Abstract
A machine vision-based carbon felt assembly detection algorithm system was designed to address issues in the assembly process,such as carbon felt misplacement and complex edge burrs.Manu-al inspection was found to be inefficient,costly,and prone to undesirable contact effects.The system uti-lizes Python to interface with a camera for image acquisition and preprocessing.Subsequently,an en-hanced Sobel operator is applied to perform initial edge detection,obtaining approximate edge coordi-nates.Then,a refined interpolation-based sub-pixel edge detection algorithm is employed to enhance the accuracy of the coordinates.Following this,the least squares method is utilized to fit precise coordi-nates,thereby identifying the four corner points in the image.Finally,the positions of these corner points are analyzed to make determinations.Experimental results demonstrate that this system can consistently conduct carbon felt assembly detection with a certain level of robustness.
关键词
机器视觉/边缘检测/亚像素边缘/图像处理Key words
machine vision/edge detection/subpixel edge/image processing引用本文复制引用
出版年
2024
NETL
NSTL
万方数据