基于机器视觉的零件尺寸测量方法研究

Study on Part Size Measurement Method Based on Machine Vision

为了提高视觉测量软件平台图像处理的效率与实用性,在进行基于机器视觉零件尺寸测量软件开发前必须要规划出一套行之有效的测量方案。通过对比现有的图像处理技术,分析了相机标定、滤波、阈值化处理、形态学处理、边界检测、轮廓提取等算法,给出了实用性更强、测量效率更高的零件尺寸测量方案。在新提出的零件测量流程中,明确了特征值获取的算法,同时在零件模型构建算法中,给出了改进后的零件位姿匹配算法,通过整合各个已改进的算法,极大地提高了本套零件尺寸测量方法的实用性。

In order to improve the efficiency and practicability of image processing in measuring software,it is necessary to propose a scheme in how to do measurement before designing the software for part dimension measurement.Compared with the existing image process technologies,the algorithm about camera calibration,smoothing,thresholding processing,morphological processing,edge detection and contour extraction is analyzed.Finally a novel part dimension measurement system is proposed,which is of more practicability and higher efficiency.In this novel part measurement flow,the algorithm in eigenvalue acquisition is confirmed,at the same time,in part model structure algorithm,an enhanced part pose matching algorithm is proposed.By integrating all enhanced methods,the efficiency of the new proposed part dimension measurement system is greatly improved.