ESD Safety Wear Detection And Voice Alert Using Deep Learning And Embedded System

Electro Static Discharge (ESD) is one of the prime causes of failure of electronic products. To reduce the level of failures caused by ESD, all employees in a manufacturing industry are instructed to wear an ESD safety wear in an Electrostatic Discharge Protected Area (EPA). Thus, this paper proposes a system that can automatically keep a track of the ESD safety wear worn by workers. Using machine learning and deep learning algorithms, a system is developed which can detect ESD safety wear in a real time environment. In this paper, we applied deep learning to multi-class object detection. To implement the object detection module, we used the Single Shot Multibox Detector (SSD) Inception Common Objects in Context (COCO) model for fast and efficient object detection. We have trained the model for 5 object classes (head cap, coat, safety shoe, shoe cover and mask) with 10,000 dataset images. This system can reiterate a warning (voice alert) if some workers are not wearing the above mentioned ESD safety wear appositely. The ability of deep learning to detect the ESD safety wear is studied by conducting experiments using Closed-Circuit Television (CCTV) camera.