The port of the mask is mandatory in some places and the practice could be extended. But how to ensure that everyone is doing well, without hindering private life of a person? In the Face of this need, the tools of detection port of the mask without resorting to facial recognition - are multiplying. After the start-up Outsight, which has deployed a kind of radar that is able to distinguish and follow in a crowd of people without a mask, it is the turn of Drone Volt to unveil his solution.
The toulouse-based company, specialized in the artificial intelligence onboard, and the construction of drones, civilian professionals, has developed a video application detection port of the protective mask, in the framework of the fight against Covid-19. This software "allows you to detect the port of the mask via a network of IP cameras, existing. The markets are diverse: transportation services, or public health, sports facilities, shopping centres, etc," explains Olivier Gualdoni, CEO of Drone Volt.
in Concrete terms, the camera is in link with the software that detects if the mask is worn, and how well positioned it is. In the opposite case, a warning is sent. A neural network is trained to detect "anomalies": the absence of a mask or possibly a helmet on a construction site. The geographical location of the person who is not in compliance is simply reported by the camera. There is no recognition of the person, or recording of such information in any database. "We use the flow of images captured by cameras in place to detect if a person wears a mask or not. It is just a management tool, additional post processing of images", says Olivier Gualdoni.Improve the prevention of risks
This software can be an additional management tool for manufacturers in places where wearing a mask is essential. Thus, when a "no mask" is sent, an audible warning is given, where the site manager wants. For the purpose of improving the systems of risk prevention.
This software has been developed by Aérialtronics, a subsidiary of Drone Volt and will be sold on a subscription basis (SaaS, Software as a Service). "Beyond the immediate concern of public health, this solution could be deployed for a wide range of applications: recognition of the personal protective equipment (PPE) on a construction site, detection of gatherings or movements of the crowd,...", adds Olivier Gualdoni. The aim being to better protect the people. To develop this solution, the group has started from scratch to build a system of motor learning that can not only detect that the person is wearing a mask or not, but also whether the mask is properly positioned. Under the chin, and there is nothing!The editorial team conseilleLa face recognition in the era of the port general masqueSujetsCovid-19MasquesAucun comment
there are currently no comments on this article.
Be the first to give your opinion !