Automated Helmet Detection

Developed For Reliance Petroleum

THE CLIENT

Reliance Petroleum Limited is a company owned by Reliance Industries Limited, one of India’s largest private sector companies. With a strategically located network of 1380+ retail outlets spread across the country, the retail business provides value to its customers through a 100% automated network, backed by superior technology.

THE BUSINESS REQUIREMENT

According to a recent survey, 1214 road crashes occur every day in India. Two-wheelers account for 25% of the total road crash deaths. Deaths of two-wheeler riders in crashes have more than doubled in a decade, between 2009 and 2019. Even though two-wheelers offer the least protective features for their occupants, they have become popular for being the most affordable and easiest mode of transport.

 

Our client, with a large network of retail outlets spread across the country, wanted to encourage two-wheeler drivers to wear helmets to ensure their safety. The company designed several incentive programs for its customers who regularly wear helmets in two-wheelers. Manually recording vehicle numbers of customers wearing helmets is not a feasible option in busy gas stations. Hence, the client wanted to automate the process of detecting helmet-wearing riders.

THE SOLUTION

Based on the business requirements, we built an AI-powered automated helmet detection system using computer vision technology.

 

 

The software is built with deep learning computer vision algorithms that have been trained on traffic camera footage to detect people wearing a helmet and those who are not. The algorithm is capable of detecting people, motorcycles, and helmets. This pipeline is designed to flag alerts when a violation is found. If a person riding a two-wheeler is identified as not wearing a helmet, then the image of the vehicle is captured and stored in a database. Daily or weekly status reports or violation logs can be generated from the database using a central dashboard. One key feature of the system is its near real-time performance achieved by a highly optimized software stack developed for Intel-based hardware.

Ready For Your Own Success Story?

Reach out to us today to discuss your project and avail a free a no-obligation consultation

Highlights

Helmet detection

Deep learning technology

Automated Number Plate Detection

Model trained on traffic camera footage

Technologies Used