| Title |
Defect Detection of Injection Modeled part based on Rigid transformation |
| Authors |
고정환(Jung-Hwan Ko) ; 이정석(Jung-Suk Lee) |
| DOI |
https://doi.org/10.5573/ieie.2020.57.7.85 |
| Keywords |
Injection modeled part; Rigid transformation; Machine vision |
| Abstract |
In this paper, the method for an effective defect detection of an injection modeled part acquired through the machine vision of the take-out robot is proposed. After processing the coordinate system transformation between camera and image, the distortion calibration and image alignment process of input image were calculated, and the rigid transformation and feature points for defect classification were extracted. Based on the extracted feature points, the adaptive defect pattern of injection molding parts is classified, and then the detection process is carried out from them. From some experiments on defect detectng with the injection molding parts images, it is analyzed that the proposed algorithm is capable of defect detection with high accuracy of less than 1mm in real-time on our own acquired data. From these good experimental results a possibility of implementing an defect detection using the proposed scheme is finally suggested. |