摘要
为了实现对大批量生产工件尺寸的精密测量,该文基于双目立体视觉视差原理,设计了一种工件尺寸实时测量系统;系统通过两个工业数字摄像头实时采集工件不同角度的两幅图像,并基于机器视觉软件HALCON的HDevelop开发环境,对两幅图像进行相关预处理,获取感兴趣区域;提出一种基于canny的亚像素边缘检测、边缘细化、基于atukey权重函数的最小二乘法边缘拟合及边缘均匀取点算法获取精确的边缘特征点,并通过双目系统标定、立体校正、极线约束与NCC相结合的立体匹配、世界坐标转换、几何计算相结合的非接触式双目视觉方法进行尺寸测量;通过大量实验证明,该方法可有效地获取工件特征点的三维坐标值,不借助外部测量仪器实现工件关键尺寸的高精度实时检测,检测精度达±0.2mm以上,简单快捷,具有较好的准确性和实时性。
In order to achieve the precision measuring for the mass production of parts dimensions.A real-time measurement system for parts dimensions based on the principle of binocular stereo vision parallax is designed.The system acquires two real-time images of parts from different angles through two industrial digital cameras,and then acquire the regions of interest of two images after a series of pretreatments based on HDevelop development environment of machine vision software HALCON.A sub-pixel edge detection based on Canny,edge thinning,least square method based on atukey weight function to fit edge and edge uniform picking algorithm are proposed to obtain accurate edge feature points,and an non-contact binocular vision measurement method combining binocular measurement system calibration,stereo correction,stereo matching of polar constraints combining with gray correlation,world coordinate transformation and geometric calculation for size measurement.Through a large number of experiments show that this method can effectively obtain the 3D coordinate value of parts feature points,high precision real time inspection of key parts of parts without external measuring instruments,the measuring accuracy is above plus or minus 0.2 mm.Simple and fast,the proposed method has a good accuracy and real-time.
出处
《计算机测量与控制》
2018年第1期59-63,共5页
Computer Measurement &Control
基金
山东省自然科学基金(ZR2011EL038)
关键词
双目立体视觉测量
系统标定
特征点提取
立体匹配
HALCON
binocular stereo vision measurement
system calibration
feature point extraction
stereo matching
Halcon
