2019年年底,世界蔓延起了嚴重特殊傳染性肺炎(Coronavirus disease 2019, COVID-19),人們的生活被迫改變,只有維持好自身的衛生安全,才能讓自己免於感染疾病,而口罩就是現在非常必須的衛生用品。口罩能防止各式會透過飛沫傳染之流行疾病,每個公共場所都被要求必須佩戴口罩、保持安全距離以避免群聚,現有的應對方案皆須人員在出入口進行管控,需要耗費不少人力資源,缺點也十分明顯,例如捷運車廂內要做到每一節皆配置人員進行管控就不太可能實行,因此本論文希望能將此改善成更便利的系統,透過將本技術整合在車廂內的監視器中,即可達到自動辨識空間內所有人的狀況,當有不符合規定的狀況發生時發出警示,提醒空間內的乘客將口罩佩戴完全,保護自己也保護他人。 為因應目前正在世界大流行的嚴重特殊傳染性肺炎,同時在可能出現群聚的場所能更便利的辨識大眾是否有佩戴口罩,本論文將嘗試使用YOLO加入人有佩戴口罩及未佩戴口罩的圖片進行訓練口罩辨識的模型,之後在NVIDIA Jetson nano上結合鏡頭進行口罩辨識,其電量的需求較低且兼顧效能,體積小設置在出入口也比較不會造成阻礙。
At the end of 2019, COVID-19 spread across the world, and people’s lives were forced to change. Every public place is required to wear masks and keep a safe distance to avoid crowds. Existing response solutions require personnel to control at the entrance, which requires a lot of human resources. But it is impossible to implement the management and control of each section of the MRT carriage. Therefore, this paper hopes to improve this into a more convenient system. By integrating image recognition technology into the monitor in the car, it can automatically identify the situation of everyone in the space, and issue a warning when a non-compliant situation occurs. This paper tried to use YOLO to train the mask recognition model using pictures of people wearing masks and not wearing masks, and then combined the webcam to perform mask recognition on NVIDIA Jetson nano.