DocumentCode
2500284
Title
Automatic measurement of the mechanical part based on binocular stereo vision
Author
Guo, Jin ; Liu, Xianyong ; Chen, Xiaoning
Author_Institution
Sch. of Inf. Eng., Southwest Univ. of Sci. & Technol., Mianyang
fYear
2008
fDate
25-27 June 2008
Firstpage
8487
Lastpage
8491
Abstract
Based on binocular stereo vision, a method which can automatically measure the distance between two holes on the mechanical part was proposed in order to improve the precision of the traditional measurement. Binocular vision sensor was calibrated; the edge of mechanical part was measured by using canny operator, the region of the hole was recognised by the function, the targets center was detected by using least square fitting method; the automatic stereo matching between two images was come true by the epipolar-slope constraint; the 3D coordinate of each hole was reconstructed and the distance was calculated. The experimental results proves that the measurement of the distance is high accurate, the system error is less than 0.06 mm, which can reach the requirement of online measurement precision.
Keywords
computerised instrumentation; distance measurement; image matching; image sensors; least squares approximations; stereo image processing; automatic stereo matching; binocular stereo vision; binocular vision sensor; canny operator; epipolar-slope constraint; least square fitting method; mechanical part automatic measurement; Coordinate measuring machines; Image edge detection; Image recognition; Image sensors; Least squares methods; Mechanical sensors; Mechanical variables measurement; Stereo image processing; Stereo vision; Target recognition; binocular stereo vision; calibration; canny; epipolar-slope; least square fitting;
fLanguage
English
Publisher
ieee
Conference_Titel
Intelligent Control and Automation, 2008. WCICA 2008. 7th World Congress on
Conference_Location
Chongqing
Print_ISBN
978-1-4244-2113-8
Electronic_ISBN
978-1-4244-2114-5
Type
conf
DOI
10.1109/WCICA.2008.4594260
Filename
4594260
Link To Document