標準作業流程順應性之電腦檢驗模型

Abstract

標準作業流程可以讓公司內部溝通、工廠作業等過程更明確且有效率，也能維持作業環境的安全，但仍有許多意外的發生是源自於人為問題，因為未遵守標準作業流程導致許多悲劇的發生，因此標準作業流程的檢測除了人工確認以外，應再發展出電腦檢測的系統，讓確認的過程多一道防護措施，以降低員工的安全風險及公司的損失。 影像辨識的技術日趨成熟，本研究期望能將影像辨識與標準作業流程檢測做結合，發展一個標準流程自動監控的檢驗方法，讓管理者能夠輸入影像辨識資料之後，系統就能判斷每個影像畫面對應的流程階段為何，系統記錄完再依照各個階段出現的順序判斷，最後輸出檢驗結果是否通過。 本研究的模型考慮了流程進行中會產生的不確定性，以及常見的迴圈流程、平行活動等結構。

Standard operation procedure(SOP) makes inner-communication and operation more clear and efficient in a company, it also maintains safety in workplace. However, there are still a lot of accidents that happen due to violations of SOP by workers. In addition to manual-checking, an automatic-checking system should be developed to double check if the workers follow the SOP. Therefore, the risk in the whole company can be reduced. This research provides a solution by combining the technology of image recognition and process checking. The model is designed to check process compliance. Operators only need to input the data of the image recognition’s results, then the model can interpret the data at each step of the process. By applying the state-transition-based algorithm, this model can check if each step is executed sequentially. Finally, it can output the result of process compliance checking. This paper includes 11 cases showing the fact that this model is able to consider the uncertainty during process execution, and common process structures such as loops, parallel activities, and exclusive activities.