메뉴 건너뛰기
.. 내서재 .. 알림
소속 기관/학교 인증
인증하면 논문, 학술자료 등을  무료로 열람할 수 있어요.
한국대학교, 누리자동차, 시립도서관 등 나의 기관을 확인해보세요
(국내 대학 90% 이상 구독 중)
로그인 회원가입 고객센터 ENG
주제분류

추천
검색
질문

논문 기본 정보

자료유형
학술대회자료
저자정보
Jaemin Lee (Samsung Heavy Industries) Yonguk Kim (Samsung Heavy Industries) Doojin Choi (Samsung Heavy Industries) Hyungjin Kim (Samsung Heavy Industries)
저널정보
제어로봇시스템학회 제어로봇시스템학회 국제학술대회 논문집 ICCAS 2024
발행연도
2024.10
수록면
1,516 - 1,520 (5page)

이용수

표지
📌
연구주제
📖
연구배경
🔬
연구방법
🏆
연구결과
AI에게 요청하기
추천
검색
질문

초록· 키워드

오류제보하기
Anomaly detection in manufacturing systems is challenging, because anomalies are rarely generated. Additionally, Radiography Testing (RT) data for welds are difficult to share publicly because it requires a high level of security to determine the quality of the products. The performance of the defect detection model required by industrial fields is expected to be guaranteed in terms of recall, even if the precision is somewhat compromised. Unsupervised learning based on normal data can be applied to address these two challenges. In this study, we propose an RT defect-detection model for pipe welds. This model first performs image preprocessing based on the expertise of NDT(Non-Destructive Testing) experts. Subsequently, the weld line is recognized using an image segmentation model, and then weld defects are detected based on the reconstruction based model. Finally, the performance of the segmentation model and the reconstruction-based anomaly detection model are discussed. As a result, more than 99% of the defective recalls were achieved compared with the reference model.

목차

Abstract
1. INTRODUCTION
2. METHODS
3. PERFORMANCE EVALUATION
4. CONCLUSION
REFERENCES

참고문헌 (0)

참고문헌 신청

함께 읽어보면 좋을 논문

논문 유사도에 따라 DBpia 가 추천하는 논문입니다. 함께 보면 좋을 연관 논문을 확인해보세요!

최근 본 자료

전체보기

댓글(0)

0