Project Overview

The project aimed to develop an application for a prominent telecommunications company in India. The application utilized Vision AI technology to detect safety gear worn by workers, such as helmets/hardhats and safety vests. It enabled the analysis of images and real-time video captured through mobile app cameras or CCTV feeds, ensuring safety compliance during attendance marking.

The Problem

Our client identified a need to enhance safety measures and compliance in various work environments. Ensuring that workers wear appropriate safety gear such as helmets/hardhats and safety vests is critical for preventing accidents and maintaining a safe work environment. However, manually monitoring safety gear compliance during attendance marking was an unfeasible, time-consuming, and error-prone process.

The Goal

The goal was to develop a POC to automatically detect safety gear, including helmets/hardhats and safety vests, through image and video analysis. The objectives included:

1. Develop an application using Vision AI technology to detect safety gear in real-time.
2. Enable the application to process images and video captured through mobile app cameras or CCTV feeds.
3. Improve safety compliance monitoring during attendance marking.

User Research

• Interviews with supervisors to gather insights on current safety practices and challenges.
• Observations to identify the most common safety gear used and potential obstacles faced during compliance monitoring.
• Analysis of existing attendance marking processes and systems to identify areas for improvement.

Solution Brief

The solution involved developing an application that utilized Vision AI technology for safety gear detection. The key components and features of the solution were:

• Safety Gear Detection: The application utilized a pre-trained model to detect a person, helmet/hardhat, and safety vest in images and real-time video streams.
• Mobile App Integration: The application was designed in a way that it seamlessly integrates with mobile app cameras, allowing workers to capture images and record videos for safety gear analysis.
• CCTV Feed Integration: The application supported the integration of CCTV feeds, enabling real-time safety gear monitoring across different work environments.
• Safety Compliance Assessment: The application processed captured images and video to determine safety compliance during attendance marking, providing immediate feedback on gear usage.

Impact and Benefits

The implementation of the Safety Gear Detection application had several significant impacts. The key outcomes of the project were:

• Improved Safety Compliance: The automated detection of safety gear reduced the chances of human error and enhanced safety compliance during attendance marking.
• Real-time Monitoring: The application provided real-time monitoring of safety gear usage through mobile app cameras or CCTV feeds, allowing immediate intervention in case of non-compliance.
• Efficient Attendance Marking: The automated assessment of safety compliance simplified attendance marking processes, saving time and effort for workers and supervisors.
• Enhanced Safety Culture: The application contributed to fostering a safety-conscious culture within the client company by prioritizing and promoting the use of safety gear.

Challenges and Lessons Learned

Throughout the project, several challenges and valuable lessons were encountered:

• Model Accuracy: Ensuring high accuracy in safety gear detection required continuous fine-tuning and training of the Vision AI model to account for various lighting conditions, angles, and worker positions.
• Integration Complexity: Integrating the application with diverse mobile app cameras and CCTV systems presented technical challenges that required careful coordination and testing.
• Ease of User Adoption: Encouraging worker adoption and compliance with the application required effective communication, training, and ongoing support to address any concerns or resistance.