메뉴 건너뛰기
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)
고객센터 ENG
주제분류

논문 기본 정보

저자정보
(Yeungnam University) (Yeungnam University) (Yeungnam University) (Yeungnam University) (Yeungnam University)
저널정보
한국생산제조학회 한국생산제조학회지 한국생산제조학회지 Vol.34 No.6
오류 신고하기

피인용 1

검색

    초록·키워드

    In this study, a deep-learning-based defect classification system was developed for the surface inspection of automotive parts. A compact conveyor-based setup was constructed to simulate the actual production conditions and ensure consistent image acquisition. To minimize background noise from the patterned belt, HSV (hue, saturation, value) color space segmentation and morphological filtering were applied. A lightweight VGG16-based convolutional neural network (CNN) was trained using transfer learning and partial fine-tuning. Despite using a small and balanced dataset, the model achieved over 95% validation and 95-100% prediction accuracy on unobserved test images. Therefore, CNNs can effectively classify known and previously undefined defects and thereby overcome the limitations of rule-based machine vision systems. Future studies will explore the integration of lighting control and the application of advanced models for real-time defect localization and classification.

    최근 본 자료 전체보기

      UCI(KEPA) : I410-151-26-02-094698918