Deep learning approaches for dynamic object understanding and defect detection

Yuan Tsung Chang, W. K.T.M. Gunarathne, Timothy K. Shih

研究成果: 雜誌貢獻期刊論文同行評審

6 引文 斯高帕斯(Scopus)

摘要

Industrial product defect detection has been known for a while to make sure the released products meet the expected requirements. Earlier, product defect detection was commonly done manually by humans; they have detected whether the products consist of defects or not by using their human senses based on the standard. In this industrial era, product defect detection is expected to be faster and more accurate, while humans could be exhausted and become slower and less reliable. Deep learning technology is very famous in the field of image processing, such as image classification, object detection, object tracking, and of course the defect detection. In this study, we propose a novel automated solution system to identify the good and defective products on a production line using deep learning technology. In the experiment, we have compared several algorithms of defect detections using a data set, which comprises 20 categories of objects and 50 images in each category. The experimental results demonstrated that the proposed system had produced effective results within a short time.

原文???core.languages.en_GB???
頁(從 - 到)783-790
頁數8
期刊Journal of Internet Technology
21
發行號3
DOIs
出版狀態已出版 - 2020

指紋

深入研究「Deep learning approaches for dynamic object understanding and defect detection」主題。共同形成了獨特的指紋。

引用此