Video surveillance is the key element of security control in manufacturing, logistics, and numerous other industries. It ensures protection from external threats, control over industrial and occupational safety measures, and reduces the incidence of injuries.
As many people are inclined to violate the existing safety procedures and regulations, safety instructions and training are not enough to make a change—the control must be autonomous, which can be achieved by digital transformation.
You are one in a million
Safety supervision on physical sites can be provided by artificial intelligence. VizorLabs H&S is a video surveillance solution with integrated response and notification subsystems, which uses a special algorithm to derive meaningful insights from camera footage. The solution is able to recognize transport and people using databases of employees and tenants, ensure occupational health and safety compliance, supervise operational personnel, and control the industrial process.
VizorLabs H&S operates on the Azure cloud platform, a multi-purpose toolkit to develop effective solutions for the special needs of businesses. VizorLabs H&S is one of 100+ Azure-based applications, which makes this platform particularly attractive and advantageous.
Enterprise-grade video content analysis has a strong focus on critical situations detected by the system. VizorLabs is able to recognize people, detect their limbs, and analyze if gloves, safety helmets, and other elements of personal protective equipment are on and done up. The solution tracks people on vehicle roads, in no-go areas or danger zones and stairs, detects instances of smoking, and monitors large vehicle passage.
The intuitive interface of the VizorLabs H&S includes icon indicators of different colors. The solution provides immediate access to the live footage and archive of every camera, which altogether enables quick visual industrial safety monitoring. So, it is no longer necessary for the operator to look at the screen all the time—violation events appear automatically in the event line indicating the date and time. Safety violation reports are generated automatically for selected periods and can be submitted to the person in charge via email or to their personal account.
This and similar video analytics solutions leave out-of-the-box detectors far behind, as the latter frequently fail to provide an appropriate assessment of the situation, are prone to false positives, and require a high-quality video stream.
Latest-generation video analytics detectors can be configured, combined, cascaded, and customized to the special needs of any business. They work even with a low-quality video stream and analog video surveillance systems, with a detection rate of 95% or higher.
A special image labeling tool and an integrated trainable neural network enable quick video stream labeling and the acquisition of the material to finish the training of the classifiers at a rate of 1,000+ images per hour per labeler.
Case study: an energy company
We have completed the task of monitoring the company personnel wearing personal protective equipment on industrial sites: our video surveillance system controls wearing safety helmets, protective shields, gloves, and done-up clothes. It also tracks the position of personnel relative to switchgears. The monitoring system prototype met all the client’s needs, and they decided to scale it.