Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
METHOD FOR DETECTING DEFECT IN CARBIDE INSERT BY USING PRODUCT SYMMETRY AND IMAGE SIMILARITY
Document Type and Number:
WIPO Patent Application WO/2018/070710
Kind Code:
A1
Abstract:
The present invention relates to a method for detecting a defect in a carbide insert by using product symmetry and image similarity, whereby a defect in a carbide insert is detected by acquiring an image of a carbide insert through a high-resolution camera and comparing product symmetry and image similarity on the basis of the image, the method comprising: a step of preparing a carbide insert (a specimen); a step of performing optical processing on the carbide insert and acquiring an image thereof; a step of estimating the shape of the specimen and classifying regions according to symmetry, on the basis of the image acquired in the previous step and through mathematical interpretation; a step of acquiring, based on the image acquired in the image-acquiring step, partial images of an inspection region in which a defect is to be detected and of a corresponding region classified in the classifying step; and a step of detecting a defect by comprising similarity between the partial images acquired in the previous step. Accordingly, defects having various shapes and sizes can be detected by comparing similarities with corresponding images by using the symmetry of products in which the defects are to be inspected.

Inventors:
SONG IL JAE (KR)
KOO HYUNG IL (KR)
KIM JUN HYOUNG (KR)
KIM SHIN GON (KR)
JEONG SANG KYO (KR)
Application Number:
PCT/KR2017/010796
Publication Date:
April 19, 2018
Filing Date:
September 28, 2017
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
WINTECH AUTOMATION CO LTD (KR)
UNIV AJOU IND ACADEMIC COOP FOUND (KR)
International Classes:
G01N21/88; G01B11/30; G06T7/00
Foreign References:
KR101575691B12015-12-08
KR100512869B12005-09-06
KR20110043032A2011-04-27
KR101078404B12011-11-01
KR20070098189A2007-10-05
Attorney, Agent or Firm:
KIM, Jae Wang (KR)
Download PDF: