- 學位論文
應用機器視覺於建構木紋掃描與木材缺陷檢測系統
Automated Inspection of Application to the Construction of Wood Grain Scanning and Wood Surface Defect Inspection Systems
摘要
木材因為取得容易，廣泛應用於日常用品中。木材的好壞主要由樹木的品種和缺陷所決定，而樹木容易受到季節、氣候、生物和生長環境等因素而產生缺陷，主要可分為三個部分：天然缺陷，如木節、斜紋理以及因生長應力或自然損傷而形成的缺陷；生物為害的缺陷，主要有腐朽、變色和蟲蛀等因素；另外，乾燥及機械加工引起的缺陷也是引起木材缺陷的主要原因之一。木材缺陷檢測通常是利用人眼檢視，人眼檢視常因個人主觀或內外在環境而影響檢測結果。 本研究之目的在開發自動化木紋掃描與檢測系統，並利用支持向量機於木節良莠分類上。於系統上主要可分為兩部分:木材輸送帶系統(Wood Conveying System)和木材缺陷檢測系統(Wood defect inspection system)。木材輸送系統主要的功能是使木材在檢測和輸送的流程中能順利進行，透過感應器與譯碼器的控制使工業相機能夠順利取像；木材缺陷檢測系統則透過電腦視覺將取得之部份影像，運用影像縫合技術將數張影像縫合成木材完整影像，並藉由影像處理之技術取得缺陷之型態位置與大小，作為後製程切割參考之依據。在木節分類上，因本身有良莠之區分，故以支持向量機分類木節之好壞。
並列摘要
Wood is easy to get and also widely used in our daily lives. The quality of the wood depends on the tree species and its defects. Trees may be damaged because of many causes such as season, climate, biology and environment. Wood defects can be divided into three categories: natural defects such as knots, grain, or caused by the stress of developing new xylem or natural damage; biological defects such as rot, discoloration, and decay by insects and so on. Besides, defects caused by arid and machine are another factors resulting in the defects. Currently, wood defect inspection is usually conducted by human eyes, but their inspection results are seriously subject to personal subjective view, the internal, and external environmental factors. Therefore, the purpose of this study is to develop an effective wood scanning system and a precise inspection system, using Support Vector Machines for the classification of the knots. The content of this project can be divided into two parts: wood conveying and wood defect inspection systems. The main function of wood conveying system is to make wood testing and delivery process proceed smoothly; the industrial camera is controlled by the sensor and the encoder for the image capture. After getting the wood images, the system combines those images into a complete image of wood by the stitching techniques. Thus, the system applies image process to wood characteristics and its defect analysis as the base for cutting. As for the classification of knots, with the distinction between good and bad, Support Vector Machines are used for the classification.