基于机器视觉的二维精密测量系统
2019-12-16
作者：粟序明1，方成刚1，洪荣晶1，吴伟伟2，朱浪1
单位：1南京工业大学；2扬州大学
摘要：受环境及自身畸变影响，机器视觉测量系统的测量精度不高且不稳定，是视觉精密测量领域亟需解决的问题。为此，在结合相对法标定以及超分辨重构算法的基础上，提出一种改进的畸变校正模型；在MATLAB软件平台上开发专用测量算法，实现亚像素边缘检测；配备专用光源、工业视觉传感器和FA镜头，最终完成了二维精密测量系统。该视觉测量系统致力于提高测量精度和鲁棒性，降低设备成本。通过工业实际加工管零件作为试验对象，验证了方法的有效性。试验结果表明，重复测量精度可达0.02mm，相对误差达到0.09%，满足实际加工要求。
关键词：畸变校正；MATLAB；精密测量；超分辨率重构；亚像素边缘检测
中图分类号：TG806；TH741文献标志码：ADOI：10.3969/j.issn.1000-7008.2019.12.022
Twodimensional Precision Measuring System Based on Machine Vision
Su Xuming，Fang Chenggang，Hong Rongjing，Wu Weiwei，Zhu Lang
Abstract:Machine vision measurement is widely used in industry.However，the measurement accuracy of visual measurement system is low and unstable affected by the environment and its own distortion，which has become an urgent problem to be solved in the field of visual precision measurement.To solve this problem，an improved distortion correction model is proposed based on relative calibration and superresolution reconstruction algorithm.A special measurement algorithm is developed to realize subpixel edge detection on MATLAB software platform.Equipped with a special light source，industrial vision sensor and FA lens，the twodimensional precision measurement system is finally completed.The vision measurement system is dedicated to improve the measurement accuracy and robustness and reduce the equipment cost.The validity of the method is verified by the experiment of pipe parts processed in industry.The experimental results show that the repeated measurement accuracy can reach 0.02mm and the relative error can reach 0.09%，which meets the actual processing requirements.
Keywords:distortion correction；MATLAB；precision measurement；super resolution reconstruction；subpixel edge detection