作者：刘磊, 陈爱军, 程楼, 丁佳为

作者单位：中国计量大学计量测试工程学院, 浙江 杭州 310018

关键词：图像处理;波纹阻火盘;尺寸测量;图像拼接;亚像素边缘检测

摘要：

针对目前波纹阻火盘采用人工取点的方式对波高进行测量引起的效率低、误差大等问题，提出一种基于机器视觉的波纹阻火盘表面波高测量方法。该方法先利用线阵CCD相机配合二维运动平台扫描的方式获取图像，再利用改进的FLANN自适应特征点匹配筛选算法优化拼接过程，提高线阵图像配准准确度，然后采用基于Zernike矩的亚像素边缘检测算法准确获取边缘定位，并结合该文提出的波高尺寸测量原理以及最小二乘法，确定最大波高对应的两个像素点的坐标，最后根据线阵相机标定结果计算得到实际波高尺寸。利用波纹阻火盘线阵图像进行实验，并与影像仪法测量结果进行对比，实验结果表明该测量方法测量误差在±0.02 mm以内，满足实际工业生产需要。

A machine vision-based method for measuring surface wave heights of corrugated fire retardant disk
LIU Lei, CHEN Aijun, CHENG Lou, DING Jiawei
College of Metrology and Measurement Engineering, China Jiliang University, Hangzhou 310018, China
Abstract: Aiming at the problems of low efficiency and large error caused by the measurement of wave height by the manual method of picking up the current of the corrugated fire retardant disk, a method for measuring the surface wave height of the corrugated fire retardant disk based on machine vision is proposed. The method firstly uses the linear array CCD camera to acquire images with the two-dimensional motion platform, and then uses the improved FLANN adaptive feature matching point screening algorithm to optimize the stitching process, improve the line array image matching accuracy, and then adopt the Zernike moment based sub-pixel. The edge detection algorithm accurately obtains the edge pixels, and combines the wave height measurement principle and the least squares method proposed in this paper to determine the coordinates of the two pixel points corresponding to the maximum wave height. Finally, the actual wave height size is calculated according to the calibration result of the line array camera. Experiments were carried out using the image of the corrugated fire retardant line array, and compared with the results of the image instrument method. The experimental results show that the average detection accuracy of the measurement method is within ±0.02 mm, which meets the needs of actual industrial production.
Keywords: image processing;corrugated fire retardant disk;size measurement;image stitching;subpixel edge detection
2020, 46(6):44-50  收稿日期: 2019-07-31;收到修改稿日期: 2019-09-11
基金项目:
作者简介: 刘磊(1995-),男,浙江金华市人,硕士研究生,专业方向为图像处理、深度学习
参考文献
[1] 唐君萍. 飞机复杂零件上大量小尺寸导孔的快速视觉检测技术[D]. 南京: 南京航空航天大学, 2015.
[2] 周应东, 廖晓波, 蔡勇. 一种机器视觉的多管工件内外径测量方法[J]. 机械设计与制造, 2018(S2): 119-122
[3] 由隔爆外壳“d”保护的设备最大试验安全间隙测定方法：GB 3836.11—2008 [S]. 北京: 中国标准出版社, 2008.
[4] MUJA M, LOWE D G. Fast approximate nearest neighbors with automatic algorithm configuration [C]//Proceedings of IEEE Conference on Computer Vision Theory and Applications. Lisbon: IEEE Computer Society, 2009(1): 331-340.
[5] DAVID L G. Distinctive image features from scale-invariant keypoints[J]. International Journal of Computer Vision, 2004, 60(2): 91-110
[6] 孙文昌, 宋建社, 杨檬, 等. 基于熵和独特性的角点提取算法[J]. 计算机应用, 2009, 29(S2): 225-227
[7] 丁尤蓉, 王敬东, 邱玉娇, 等. 基于自适应阈值的FAST特征点提取算法[J]. 指挥控制与仿真, 2013, 35(2): 47-53
[8] 余博, 郭雷, 赵天云, 等. Freeman链码描述的曲线匹配方法[J]. 计算机工程与应用, 2012, 48(4): 5-8
[9] 崔继文, 谭久彬. 基于Zernike矩的亚像素边缘定位算法[J]. 光学技术, 2005, 31(5): 779-782
[10] 李慧鹏, 刘缤艳, 魏晓马, 等. 基于Zernike矩的圆形插针特征点提取算法研究[J]. 半导体光电, 2019, 40(1): 129-133,139
[11] 杨春晖. 基于标准规的几何尺寸参照测量技术研究[D]. 太原: 中北大学, 2010.
[12] 贺俊吉, 张广军, 杨宪铭. 基于交比不变性的镜头畸变参数标定方法[J]. 仪器仪表学报, 2004, 25(5): 597-599