Font Size: a A A

Study Of Precision Measurement Technology Based On Optical Images

Posted on:2011-10-25Degree:MasterType:Thesis
Country:ChinaCandidate:Y S ZhuFull Text:PDF
GTID:2178330338480106Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
With the development of computer technology and photoelectricity, a new inspection technology based on computer vision has been presented.This technolog- y is a non-contacted method for the dimensional measure using image processing and CCD as the picture sensor, which is now widely applied to the precision measure of the part dimension.This thesis project originated from the demand of an automobile manufacturer -- the First Automobile Works, who places high emphasis on product quality, and imposes rigorous standard on the reliability of electrical connection components and its precision.Based on an analysis of such application level requirement and according to the actual shape of electrical connector,aiming at measuring the size of the electrical connector, this thesis focuses on the theory and method of improving the measurement accuracy, including camera calibration technique, digital image processing technologies and sub-pixel edge detection techniques.In understanding the basis of the camera imaging model, using the pinhole imaging model, and with the lens distortion factors, we can solve the parameters of the camera. By comparing the advantages and disadvantages of various camera calibration methods, this thesis utilizes two-dimensional plane camera calibration method, which is simple, and with high precision, fully meets the requirements of measurement accuracy. In the parts of digital image processing technology, this thesis mainly analyzes digital image filtering, binarization using method of Otsu, the traditional edge detection methods and Hough transform to detect circular. And for the shortcomings of traditional Hough Transform, this thesis application Point Hough Transform to improve the speed of detection. In the parts of sub-pixel edge detection, this thesis introduces Sub-pixel measurements using moment-based edge operator and its error correction, and also introduces orthogonal moment operators (Zernike) for subpixel edge detection and the fast algorithm of the orthogonal moment operators (Zernike).Finally this thesis introduces the least square method to fit circle, through this method it can solve the circle parameters in the level of sub-pixel.In the end of this thesis, we using Visual C++ as development platform,and modular design thinking, divide the entire software system into the four modules.The software system integrates the technology introduced before, realizing the potential of non-contact messurement. So it can provide software support to the follow-up overall platform .We have deliberately made extensibility a key property of the system, and thus it can be easily adapted according to changes in application demands.
Keywords/Search Tags:computer vision, digital image processing, camera calibration, sub- pixel edge detection
PDF Full Text Request
Related items