Deep Learning Enabled Smart Industrial Workers Precaution System Using Single Board Computer (SBC)

Abstract

The safety aspects in a factory cannot be compromised and periodic monitoring is mandatory. The Internet of Things (IoT) serving industry purposes is known as the Industrial Internet of Things (IIoT). This new proposed methodology incorporates the safety aspects verification of labor involving handling, controlling, welding, cutting, and grinding big critical machines. The proposed methodology works on the deep learning concept and verifies the presence of a helmet to ensure the safety (wellbeing) of the laborer. A camera is placed in front of the work-bench and then a camera connected with the Raspberry pi or any single board computer (SBC) verifies the presence of a helmet by using image processing. The TensorFlow deep learning method is used, and verification time is measured with different safety equipment and helmets using deep learning and image processing. This project deals with a low-cost safety mechanism that ensures the safety of workers inside industries. This device includes a sensor, Raspberry pi or any kind of single board computer (SBC) and locking actuators. The workers dealing with hazardous chemicals in the case of the chemical industry and heavy materials in the case of other industries are supposed to wear helmets and face masks for their protection. Most industrial accidents happen because of very low safety aspects. This module ensures the safety aspects before turning on the machine or opening a chemical chamber. The camera sensor is activated by an obstacle sensor, the image is processed through deep learning, and the necessary safety requirement is verified. Once the safety aspect is verified the locks are opened with the GPIO command. If the safety aspects are not satisfied, the worker image is acknowledged to the supervisor for further safety measures and corrective actions.