鏟花品質檢測雲端化之研究

Toward Cloud-based Scraping Quality Identification

甘庭瑋(Ting-Wei Gan)

指導教授 ： 鄭宗明 ； 李英聯

朝陽科技大學/理工學院/工業工程與管理系/碩士(2017年)

摘要

工具機之平面鏟花乃仰賴人工以手持工具完成，而所得品質則隨執行者之經驗、技術、及穩定度而有相當之落差。鏟花除可用來增進平面之平整性外，主要目的仍在於產生均勻分佈之淺斑漥孔於滑動接觸之承重平面，做為微小之潤滑油儲油袋，以降低滑動面間之摩擦，增進移動之平穩性。由於其手工操作與微小深度之本質，藉由目視檢測其品質需要經驗且較難穩定，而檢測耗時又難以量化表述品質等級則是造成檢測成果難以推行後續補正操作之主要原因。除此之外，鏟花施工處之照明狀況亦不穩定，更增加了影像判讀之困難度。僅管如此，公認的鏟花品質仍以接觸率(POP)和支撐點數(PPI)兩項量化指標在做鑑定。因此，本計畫將以電腦視覺方法求得鏟花面影像之POP 與PPI，首先使用適應性閥值切割法將畫面切成較小單元各別處理，以克服光源不均造成之畫面亮度落差。再將小單元畫面數位化重組後做四分法切割，藉由小區域品質之呈現來表述整體鏟花面之品質及其分佈。實作上，整體品質鑑定之機能將整合建立於遠端伺服器中，並開發平板電腦操作介面，於工作現場以CMOS 相機連續拍攝鏟花面，藉由無線網路傳至伺服器作影像處理與品質計算，再將鑑定結果與補正建議傳回現場平板電腦，形成同步式之雲端製造管理機能。

關鍵字

鏟花品質鑑定 ； 電腦影像處理 ； 雲端製造管理

並列摘要

Surface scraping process is usually done by handwork, so the quality will rely heavily on worker’s experience, skill level and stability. In addition to flatness improvement, scraping also provides micro oil pockets to enhance the lubrication effect between the flat sliding surfaces. However, due to its handcrafted and trivial nature, inspection of scraped surfaces with bare eyes is prone to unreliable quality. Besides, in most cases, workplace illuminations are insufficient or uneven, and thus, adding more difficulties to image-based inspections. Nevertheless, two indices are still widely used in the industry to determining the quality of scraping: Percentage of points (POP) and Points per square inch (PPI). Therefore, this research will develop a web-based scrap inspection system that captures images of the scraped surface at the worksite, and sends the images to a remote server for image processing and inspection to draw the indices. At first, an adaptive threshold method is performed to cope with the light source problem, then find POP and PPI out of the scrapped image down to a quarter of a square inch. In practice, the images are captured with a CMOS camera via USB interface using a tablet PC, and the calculation are done at a web-linked workstation, then the overall quality and quality distributions are presented on the tablet to simulate a real-time cloud-ready manufacturing management process.

並列關鍵字

Scraped Surface Inspections ； Image Processing ； Cloud Manufacturing Management

參考文獻

1.Akkermans, H. A., Bogerd, P., Yücesan, E., & Van Wassenhove, L. N. The impact of ERP on supply chain management: Exploratory findings from a European Delphi study.European Journal of Operational Research,146(2), 284-301. 2003.

2.Bay, H., Ess, A., Tuytelaars, T., & Van Gool, L. Speeded-up robust features (SURF).Computer vision and image understanding,110(3), 346-359. 2008.

3.Fan, K. C., Torng, J., Jywe, W., Chou, R. C., & Ye, J. K. 3-D measurement and evaluation of surface texture produced by scraping process.Measurement,45(3), 384-392.2012.

4.Gibson, A., Gamble, M., Wolstencroft, K., Oinn, T., Goble, C., Belhajjame, K., & Missier, P. The data playground: An intuitive workflow specification environment.Future Generation Computer Systems,25(4), 453-459. 2009.

5.Helo, P., Suorsa, M., Hao, Y., & Anussornnitisarn, P.Toward a cloud-based manufacturing execution system for distributed manufacturing.Computers in Industry,65(4), 646-656.2014.