Error Detection Method of Image and Graphic Dynamic Registration
-
摘要: 针对零件加工精度检测要求较高、检测环境复杂,提出了一种新型的加工精度检测方法。通过动态调整被检测零件的三维设计图形的空间位姿,使调整后的图形空间位姿与工业相机采集到的被检测零件的高质量图像处于同一拍摄角度,保存零件此时的三维设计图形的空间位姿图像,使用改进后的SIFT-Harris算法将采集到的零件高质量图像与保存的空间位姿图像实现配准, 进而实现零件与其对应的三维设计图形尺寸的对比检测。通过实验测试表明:本文提出的图像与图形动态配准的误差检测方法,具有较高的精度和速度,并且检测精度符合误差分布概率,对加快误差检测技术的发展具有一定的推动和探索意义。Abstract: Aiming at the requirements of the highaccuracy and the complicated detection environment in the processing detection, a new method of processing detection accuracy was proposed. By dynamically adjusting the spatial position of the 3D design graphics of the detected part, the adjusted spatial position of the graphics and the high-quality image of the detected part collected by the industrial camera are at the same shooting angle, and the 3D design graphics of the part at this time are saved for the spatial pose image, the improved SIFT-Harris algorithm is used to register the high-quality image of the part and the saved spatial pose image, thereby realizing the contrast detection of the part and its corresponding 3D design graphic size. The experimental test shows that the error detection method of dynamic registration of images and graphics has high accuracy and efficiency, and the detection accuracy accords with the probability of error distribution, which has certain promotion and exploration significance for accelerating the development of error detection technology.
-
Key words:
- industrial camera /
- images and graphics /
- dynamic registration /
- error detection
-
表 1 图像处理过程
实验一 实验二 原图 二值化 边缘检测 Hough 圆检测 确定圆心 表 2 实验一检测数据
序号 像点的像方坐标 物点的物方空间坐标 欧拉角 ① (627,567) (625,570,50) (−0.0001,0.0034,3.1149) ② (675,910) (679,915,49) ③ (514,776) (514,776,65) 表 3 实验二检测数据
序号 像点的像方坐标 物点的物方空间坐标 欧拉角 ① (629,673) (616,681,59) (−0.0013,1.0216,3.1086) ② (639,883) (653,915,50) ③ (613,659) (639,806,50) 表 4 图形位姿变换后以其XY面保存的图像
实验验号 图象 实验一 实验二 表 5 特征点精确匹配
实验编号 特征点精确匹配 实验一 实验二 表 6 配准算法实验数据对比
使用算法 对比指标 实验一 实验二 SIFT 配准精度/% 80.76 85.36 配准时间/s 5.3 6.1 改进的SIFT-Harris 配准精度/% 85.71 88.23 配准时间/s 4.6 5.7 表 7 配准元素标记后的图像
实验编号 标记后的图象 实验一 实验二 表 8 实验一检测数据
序号 元素 面积/mm2 周长/mm 离心率 1 外轮廓 1256.306 186.932 无 2 圆 193.861 23.407 0.450 3 圆 198.398 25.910 0.383 4 圆 196.532 24.541 0.139 5 圆 195.614 23.394 0.405 表 9 实验二检测数据
序号 元素 面积/mm2 周长/mm 离心率 1 外轮廓 1252.530 185.230 无 2 圆 193.161 24.205 0.395 3 圆 198.026 25.013 0.153 4 圆 198.528 25.114 0.106 5 圆 194.916 24..890 0.598 表 10 检测数据
$3\sigma $ 准则分析数据名称 实验一 实验二 面积/mm2 周长/mm 离心率 面积/mm2 周长/mm 离心率 平均值μ 196.101 24.313 0.344 196.158 24.777 0.313 均方差б 1.890 1.193 0.140 2.558 0.498 0.228 μ-3б 190.431 20.734 −0.075 188.484 23.283 −0.372 μ+3б 201.771 27.891 0.763 203.832 26.272 0.998 -
[1] 蔺小军, 单晨伟, 王增强, 等. 航空发动机叶片型面三坐标测量机测量技术[J]. 计算机集成制造系统, 2012, 18(1): 125-131LIN X J, SHAN C W, WANG Z Q, et al. Measurement techniques of coordinate measuring machine for blade surface of aero-engine[J]. Computer Integrated Manufacturing Systems, 2012, 18(1): 125-131 (in Chinese) [2] 张现东, 卜昆, 董一巍. 航空发动机叶片三坐标测量优化采样方法[J]. 航空动力学报, 2019, 34(1): 168-176ZHANG X D, BU K, DONG Y W. Optimal sampling strategy for aero-engine blade inspection with coordinate measuring machine[J]. Journal of Aerospace Power, 2019, 34(1): 168-176 (in Chinese) [3] 潘建栋. 长度计量中游标卡尺检定前常见故障的修理技巧初探[J]. 内燃机与配件, 2017(15): 82-83 doi: 10.3969/j.issn.1674-957X.2017.15.047PAN J D. Study on the repairing skills of common faults before vernier caliper verification in length measurement[J]. Internal Combustion Engine & Parts, 2017(15): 82-83 (in Chinese) doi: 10.3969/j.issn.1674-957X.2017.15.047 [4] XIAO G F, LI Y T, XIA Q X, et al. Research on the on-line dimensional accuracy measurement method of conical spun workpieces based on machine vision technology[J]. Measurement, 2019, 148: 106881 doi: 10.1016/j.measurement.2019.106881 [5] 刘俊博, 黄雅平, 王胜春, 等. 基于机器视觉的多线路钢轨扣件缺损检测方法[J]. 中国铁道科学, 2019, 40(4): 27-35 doi: 10.3969/j.issn.1001-4632.2019.04.04LIU J B, HUANG Y P, WANG S C, et al. Rail fastener defect detection method for multi railways based on machine vision[J]. China Railway Science, 2019, 40(4): 27-35 (in Chinese) doi: 10.3969/j.issn.1001-4632.2019.04.04 [6] 冯淼, 周传德, 孟明辉. 铝合金零件尺寸误差自动检测方法研究[J]. 机电工程, 2018, 35(4): 380-384 doi: 10.3969/j.issn.1001-4551.2018.04.009FENG M, ZHOU C D, MENG M H. Automatic detection method of dimension error for aluminum alloy parts[J]. Journal of Mechanical & Electrical Engineering, 2018, 35(4): 380-384 (in Chinese) doi: 10.3969/j.issn.1001-4551.2018.04.009 [7] 庄志惠, 施金鸿, 岑健, 等. 基于激光技术的光电无损检测系统[J]. 激光杂志, 2019, 40(7): 39-42ZHUANG Z H, SHI J H, CEN J, et al. Photoelectric nondestructive testing system based on laser technology[J]. Laser Journal, 2019, 40(7): 39-42 (in Chinese) [8] 花广如, 王进峰, 王洪运, 等. 激光位移传感器测量表面粗糙度的实验装置和方法[J]. 实验技术与管理, 2019, 36(8): 78-80HUA G R, WANG J F, WANG H Y, et al. Experimental device and method for measuring surface roughness with laser displacement sensor[J]. Experimental Technology and Management, 2019, 36(8): 78-80 (in Chinese) [9] 李鹏. 基于外标记点的心脏介入手术导航配准方法研究[D]. 武汉: 武汉理工大学, 2016.LI P. Research on external fiducially markers-based registration in cardiac interventional surgery navigation[D]. Wuhan: Wuhan University of Technology, 2016 (in Chinese). [10] 朱明, 姚强, 唐俊, 等. 超图约束和改进归一化互相关方法相结合的图像配准算法[J]. 国防科技大学学报, 2019, 41(3): 50-55 doi: 10.11887/j.cn.201903009ZHU M, YAO Q, TANG J, et al. Image registration algorithm with hypergraph constraint and improved normalized cross correlation method[J]. Journal of National University of Defense Technology, 2019, 41(3): 50-55 (in Chinese) doi: 10.11887/j.cn.201903009 [11] 何凯, 魏颖, 王阳, 等. 一种改进的非刚性图像配准算法[J]. 工程科学学报, 2019, 41(7): 955-960HE K, WEI Y, WANG Y, et al. An improved non-rigid image registration approach[J]. Chinese Journal of Engineering, 2019, 41(7): 955-960 (in Chinese) [12] 陈泽锋, 吴庆阳, 陈顺治, 等. 基于频谱和空域特征匹配的图像配准算法[J]. 激光与光电子学进展, 2020, 57(2): 021005CHEN Z F, WU Q Y, CHEN S Z, et al. Image registration based on spectral and spatial feature matching[J]. Laser & Optoelectronics Progress, 2020, 57(2): 021005 (in Chinese) [13] HOUGH P V C. Method and means for recognizing complex patterns[R]. NSA-17-008572, 1962. [14] 崔志勋, 张仁杰. 基于改进Hough变换的圆形防松垫片安装检测[J]. 电子测量技术, 2018, 41(22): 95-98CUI Z X, ZHANG R J. Detection of round lock washer mounting based on improved Hough transform[J]. Electronic Measurement Technology, 2018, 41(22): 95-98 (in Chinese) [15] 邓才龙, 刘焱雄, 麻德明, 等. 基于三维DLT方法的非量测数码相机检校[J]. 海岸工程, 2016, 35(2): 33-40 doi: 10.3969/j.issn.1002-3682.2016.02.005DENG C L, LIU Y X, MA D M, et al. Non-metric digital camera calibration based on 3D direct linear transformation[J]. Coastal Engineering, 2016, 35(2): 33-40 (in Chinese) doi: 10.3969/j.issn.1002-3682.2016.02.005 [16] LOWE D G. Distinctive image features from scale-invariant keypoints[J]. International Journal of Computer Vision, 2004, 60(2): 91-110 doi: 10.1023/B:VISI.0000029664.99615.94 [17] 张勇, 王志锋, 马文. 基于改进SIFT特征点匹配的图像拼接算法研究[J]. 微电子学与计算机, 2016, 33(3): 60-64ZHANG Y, WANG Z F, MA W. Research image mosaic algorithm based on improved SIFT feature matching[J]. Microelectronics & Computer, 2016, 33(3): 60-64 (in Chinese)