油管扣件外形瑕疵視覺檢測技術開發

本系統旨在開發運用機器視覺檢測油管扣件，開發出一套高效率、低成本之檢測系統，因而避免人工檢測在長時間下，造成誤判影響產品品質使商譽受損。本系統可以有效降低生產成本與提高產品品質，以全自動化檢測為架構，主要分四個站別如下：供料、輸送、取像、分類四大站。供料使用儲料斗、震動供料機和振動平送機，不間斷地供應檢測物至輸送機上，當檢測物經過光電感測器時，觸發攝影機拍照取像，在經由自行開發的檢測軟體，判別檢測物件是否為良品或不良品，當檢測到不良品時，使用吹氣法是運用氣壓配合高速電磁閥，將此不良品吹進集料盒內，而良品則繼續往前移動至末端集料盒內，達成分類的目的。本開發軟體能一次性檢測外圓、內圓和環內瑕疵，非常的快速節省許多時間，軟體經實際測量驗證後，能準確判斷出瑕疵的油管扣件，內外圓尺寸量測重現性均在0.04mm以下。

關鍵字

油管扣件；機器視覺

並列摘要

A high efficiency and low cost inspection system was developed by utilizing machine vision to inspect oil pipeline fasteners, thus avoiding erroneous inspection results caused by manual inspection during extended work hours that can negatively affect quality of products and damage business reputation of the company. The system effectively lowers production cost and improves product quality. It uses an architecture of fully automated inspection and consists of four main stations: feeding, transportation, image capture, and classification. Feeding station utilizes hopper, vibration feeder, and vibration horizontal conveyor to continuously transport test objects to the conveyor. When test objects pass through a photoelectric sensor, they trigger a camera to take pictures. Then, the pictures are processed by self-developed inspection software to determine whether the test objects are good products or defective products. Classification is accomplished by blowing defective products into a collecting box when they are detected while moving good products forward into a terminal collecting box. The blowing method utilizes air pressure and high-speed electro-magnetic valves. The self-developed software was field tested and proved to be able to accurately identify defective pipeline fasteners. The reproducibility of measured inner and outer circle dimensions is less than 0.04mm.