【作者】 李小涛；

【导师】 杜晓晴；

【作者基本信息】 重庆大学 ， 仪器科学与技术， 2011， 硕士

【摘要】 钢管在工业生产中扮演着极其重要的角色,其广泛应用于化工、冶金、机械等行业。直线度是衡量钢管质量指标的主要参数之一,国内外针对轴类、导轨等工件在给定平面内的直线度测量研究较多,但是针对钢管在轴向方向直线度的测量研究较少。研究钢管直线度的准确测量有着重要的科学意义及工业价值。针对钢管直线度的测量,常用的方法有位移传感器法、拉线法等,这些方法都是通过测量钢管表面母线直线度来估算其轴线方向直线度,所以测量准确度较低,且无法实现非接触测量,只适合于钢管直线度的抽样检测。机器视觉技术可以实现非接触实时测量,在直线度测量中也得到应用,其在测量方法上取得很大突破,但是在光学成像系统,边缘检测算法、直线度实时测量硬件电路设计等模块尚未展开深入研究。本文在前人研究基础之上,采用机器视觉方法测量钢管直线度,首先设计了应用于钢管直线度测量的光学成像系统;然后结合钢管形状,提出了新的边缘检测算法;同时设计了钢管直线度的计算算法,并将本系统测量方法和传统测量方法进行了实验比较;最后采用FPGA、DSP组合设计了直线度实时测量系统的硬件电路。本文主要研究内容如下:第一章对钢管直线度测量的国内外现状进行了分析,阐述了本文的研究目的和意义,并介绍了本文的主要研究工作。第二章设计了钢管直线度测量的光学成像系统。首先设计了成像系统的技术指标,并采用双高斯结构设计了光学成像系统基本结构,然后通过ZEMAX光学模拟及像质分析,对成像镜头的结构参数进行了优化,最后对设计结果进行了评估,该光学成像系统能够用于钢管直线度测量。第三章主要完成了钢管边缘检测算法设计。分析了传统边缘检测算子和数学方法边缘检测的原理及应用现状,在此基础之上,结合照明方式、钢管特殊形状,提出了一种新的边缘检测方法——相邻行灰度差异判定法,对钢管图像进行边缘检测。最后从运算量和检测效果两方面对本算法进行了评估。第四章完成了钢管直线度计算算法设计及实验。首先设计了中心轴线像素坐标计算算法,然后采用最小二乘法对直线度误差进行评定,并对测量系统进行标定。最后采用百分表测量钢管直线度的方法和本文采用的测量方法进行了实验比较,并对系统测量误差进行了分析。第五章设计了钢管直线度实时测量系统的硬件电路。首先给出了直线度测量系统的硬件整体方案设计,然后采用FPGA实现了“相邻行灰度差异判定法”算法,并给出了算法实现原理、进行了相关实验论证。最后采用DSP设计了钢管轴心线像素坐标计算算法和直线度误差评定算法,并进行了相关实验论证。第六章对本文的研究工作做了总结,指出了钢管直线度测量系统有待进一步解决的问题,并提出发展建议。更多还原

【Abstract】 Steel pipe does play a imprortant role in industry, it’s applied extensively to chemical industry, metallurgy, mechanics and so on. Straightness is one of the most important technology index for steel pipe. Lots of research is made on the straightness of axis and guidway in given plane between national and foreign, but there’s little research on steel pipe’s 3-D straightness. It has industry value and science significance to do research into straightness measurment with high precision.The displacement sensor and guy wire are often used to measure steel pipe’s straightness, which are used to evaluate steel pipe’s 3-D straightness by surface busbar. So it has low accuracy and could’t be realized non-contact measurement of steel pipe’s straightness, only can be applied for sampling the steel pipe’s straightness detection. Machine version is applied to realize non-contact measurment of the steel pipe’s straightness. It has made a breakthrough in measurment method. There’s no deep research on optical imaging sytem, edge detection algorithm or design of hardware circuit which are the modules of straightness real-time measurment. Based on others’research, machine version is adopted to measure steel pipe’s 3-D straightness in this paper. Firstly, optical imaging system is designed for steel pipe’s 3-D straightness measurment. Then, a new method of image edge detection is adopted to detect the steel pipe’s edge, which is associated with steel pipe’s special shape and illumination manner. Algorithm of the steel pipe’s 3-D straightness calculation is also designed, which is compared with traditional method by experiment. Finally,FPGA and DSP are adopted to design the hardware circuit of straightness measurment. The main contents of this paper are as follws.In the first chapter, the research status of steel pipe’s straightness measurment between national and foreign is analysed. The intention and significaion of this disseration are expatiated, and the main research work in this dissertation is introduced.In the second chapter, the optical imaging system which is applied to steel pipe’s straigtness measurment is designed. Firstly, technology index of the imaging system is designed. A structure of double-Gauss lens is adopted to design the basic structure of the optical imaging system. Then, ZEMAX is used to analyze image quality and also be used to simulate the optical system. The structure parameters of imaging lens are optimally designed. It is assured that the designed system can be applied to measure steel pipe’s straightness.In the third chapter, edge detection algorithm which is used for steel pipe’s straightness measurment have been designed. Principles of the traditional method edge detection operato and mathematical method are analysed, and also the status of them is analysed. A new method of image edge detection named judgment of different adjacent row’s gray is adopted to detect the steel pipe’s edge, which is associated with steel pipe’s sspecial shape and illumination manner. Finally, algorithm of edge detection is evaluated by quality and computation power.In the fourth chapter, algorithm of steel pipe’s straightness measurment has been designed with experiment. Firstly, algorithm of steel pipe’s pixel axis coordinate is designed. Then, Least-squares is adopted to evaluate straigtness error, and also the measurment system is calibrated. Finally, the method adopted in this paper is compared with the traditional method which is used dial indicator to measure steel pipe’s straightness. And also the measurment system’s error is analysed.In the fifth chapter, hardware circuit of steel pipe’s straightness real-time measurment is designed. Firstly, the overall design of the measurment system has been done. Then, FPGA is adopted to realize the algorithm named judgment of different adjacent row’s gray, and also the principle of the algorithm’s realization is analysed with experiment. Finally, algorithms of steel pipe’s pixel axis coordinate and straightness error evaluation are designed based on DSP, and also some experiments have been done.In the sixth chapter, the whole research work in this dissertation is summarized, and the issues of steel pipe’s straightness measurment for further research are pointed out and the suggestions are put forward .更多还原