透過您的圖書館登入
IP:18.218.234.83
  • 期刊

基於U型神經網路應用於紡織品數位之自動光學檢測檢驗

U-Net-based Automated Fabric Inspection

摘要


現今自動光學檢測(Automated optical inspection, AOI)主要目標為取代人工檢測,本研究所對花檢測的主要物件為針織品表面,其外表為不規則紋理,該針織品物件若有缺陷會穿插在不規則紋理間,其布料表面的缺陷大部分隱藏於不規則紋理之中,人工檢測不易發現,使用一般機器視覺演算法也很難將該缺陷歸類出來。本研究透過深度學習之UNET對布料表面做缺陷檢測,達成針織瑕疵之有效辨識。

並列摘要


Nowadays, the main objective of automated optical inspection (AOI) is to replace manual inspection. The main target in this research is the registration inspection on the surface of knitted fabrics, whose appearance is irregular texture. In this study, the UNET of deep learning was used to detect defects on the surface of the fabric to achieve effective identification of knitting defects.

延伸閱讀