Title :
Auto-detect of machine vision and its application in assembling inspection
Author :
Wang, Jing ; Yang, Xiaoyi
Author_Institution :
Dept. of Comput. Applic., Chongqing Coll. of Electron. Eng., Chongqing, China
Abstract :
Aimed at the inspection of non-defect work-piece, the paper proposed a sort of image processing method to make scathe-less detect for work-piece by industrial CT. Based on the auto-detect of image edge by image segment algorithm extracted geometry size of CT image, firstly non-defect work-piece was chose out to implement accurate location of CT image for the same work-piece section by phase correlation method, then the location of possible template image in the waited detect image could be judged by modified Hausdorff distance. In this paper, it took the recognition inspection of precision assembly of metal parts as an example that judged the location information in the original image by means of template image. The judged results viewed that the correct recognition rate was 100 for part assembly being right and whether leak assembly. The instance inspection results show that it is high in precision and repetitiveness.
Keywords :
assembling; computer vision; computerised tomography; correlation methods; edge detection; feature extraction; geometry; image segmentation; inspection; object recognition; production engineering computing; CT image location; Hausdorff distance; assembling inspection; geometry size extraction; image edge auto-detect; image processing method; image segment algorithm; industrial CT; machine vision auto-detect; nondefect work-piece inspection; object recognition; phase correlation method; Assembly; Computed tomography; Image edge detection; Image registration; Inspection; Machine vision; Pixel; CT image; image auto-measurement; objection recognition; phase correlation method;
Conference_Titel :
Intelligent Control and Automation (WCICA), 2011 9th World Congress on
Conference_Location :
Taipei
Print_ISBN :
978-1-61284-698-9
DOI :
10.1109/WCICA.2011.5970537
