Helmet detection for large industrial sites

Kapernikov developed a Proof of Concept (PoC) for a worldwide energy company, based on deep learning techniques, that enables to see whether people wear their safety helmets at work in designated areas.

Challenge

Monitoring people for safety reasons on large construction sites or at large factories is almost impossible with human operators. However, cameras equipped with the right artificial intelligence are able to do this automatically. A worldwide energy company asked Kapernikov to find a way to see whether people were wearing their safety helmets on site.

Role

Kapernikov made use of existing, robust people detection technology, but developed in-house algorithms to classify the detected people into those who are wearing a helmet and those who aren’t.

For both the people detection and helmet classification, Kapernikov used open source neural network architectures. For the training of the classification model, we used a training and a validation set consisting of examples of people wearing helmets and a group not wearing helmets.

Results

The proof of concept outlines camera requirements for application on actual sites. The results will be integrated into future camera surveillance projects.

Applied solutions

Machine learning
Computer vision

Industrial Localization