作者：付振振, 李蓓智, 杨建国, 周亚勤

作者单位：东华大学机械工程学院, 上海 201600

关键词：图像处理;细颈测量;边界跟踪;圆弧链码;识别规则

摘要：

对于挠性接头细颈不宜接触测量和测量效率低下的问题,提出一种基于圆弧边界识别的视觉测量方法。对原始图像滤波、二值化及边界跟踪后,通过所提出圆弧链码识别规则,提取形成细颈的非整圆轮廓,准确计算细颈最小厚度的方位。为减少细颈表面纹理、光照不均的影响,在分析测量原理误差的基础上,对细颈最小厚度处两端延伸区域进行连续一维测量,剔除粗差值后取均值作为实际测量结果。测试结果表明:提出的细颈测量及图像处理方法可使细颈重复测量准确度0.7 m,满足细颈加工的测量要求。

Research on measurement and method of image processing for the thin neck of flexible joint

FU Zhenzhen, LI Beizhi, YANG Jianguo, ZHOU Yaqin

College of Mechanical Engineering, Donghua University, Shanghai 201600, China

Abstract: A method of visual measurement based on arc boundary recognition was proposed to implement non-contact and inefficient measurement on the thin neck of flexible joint. After the original image was processed by filtering, binarization and boundary tracking, the criterion for arc chain code recognition was used to identify the incomplete round arc boundaries that constituted the thin neck so as to position accurately the thinnest part of the thin neck. To eliminate the measurement error caused by rugged surfaces and uneven illumination, the sequential one-dimension measurement was applied to the areas extending from the two ends with the minimum thickness of the thin neck. The mean value removed of gross errors was regarded as the final measurement results. According to the measurement results, the repeated measuring accuracy of the measuring and image processing methods is less than 0.7 μm, which can satisfy the demand for measuring processed thin necks.

Keywords: image processing;measurement on thin neck;boundary tracking;arc chain code;recognition criterion

2015, 41(10): 22-26  收稿日期: 2015-01-07;收到修改稿日期: 2015-03-12

基金项目: 国家863计划项目(2012AA041309)

作者简介: 付振振(1989-),男,河南商丘市人,硕士研究生,专业方向为机器视觉与检测。

参考文献

[1] 姜黎,吴伟仁,张之敬,等. 微小型结构件显微图像边缘的自动识别[J]. 光学精密工程,2013,21(1):224-232.
[2] 赵渊博,赵慧洁. 挠性接头薄筋厚度在线测量图像分割方法研究[J]. 光学精密工程,2005,1(13):153-157.
[3] 杜跃军,周安石,闫志强. 挠性接头细颈尺寸在线测量技术的研究[J]. 中国惯性技术学报,1994,2(1):39-44.
[4] 樊星. 挠性接头细颈厚度测量装置的研制[D]. 哈尔滨:哈尔滨工业大学,2006.
[5] 裘祖荣,陈浩玉,胡文川,等. 嵌入式线结构光角度视觉检测及误差补偿[J]. 光学精密工程,2013,21(10):2480-2487.
[6] Hamedon Z, Mori K, Abe Y. In-situ measurement of three-dimensional deformation behaviour of sheet and tools during stamping using borescope[J]. Journal of Materials Processing Technology,2014,214(4):945-950.
[7] 华才健,蹇红梅,张燕. 基于机器视觉的火花塞间隙检测技术[J]. 中国测试,2014(5):108-110.
[8] 张显全,王继军,蒋联源. 基于Freeman链码的圆识别方法[J]. 计算机工程,2007,33(15):196-198,
[9] 王成. 一种基于链码的圆形目标识别方法[J]. 信息与电脑:理论版,2013(1):113-114.
[10] 王福生,齐国清. 二值图像中目标物体轮廓的边界跟踪算法[J]. 大连海事大学学报,2006,32(1):62-64.
[11] Canny J. A computational approach to edge detection[J]. Pattern Analysis and Machine Intelligence, IEEE Transactions,1986(6):679-698.
[12] 尚雅层,陈静,田军委. 高斯拟合亚像素边缘检测算法[J].计算机应用,2011,31(1):179-181.