Sitronics Video Analytics Platform for Recognizing Human Behavior

The main articles are:

• Computer vision (machine vision)
• Video analytics (terms, applications, technologies)

2023: Developing a video analytics platform to recognise human behaviour

Sitronics Group on June 14, 2023 announced that it is developing a video analytics platform for recognizing human behavior.

A set of technologies is needed to control complex manual technological operations, safety of technological processes, monitoring public safety, process control and sports analytics.

Sitronics Group is working to develop a video analytics platform using a variety of computer vision techniques, including skeletal model building algorithms. The complex of technologies is combined into a single platform designed to solve various applied problems in such areas as education and sports, medicine, construction, industry, retail, transport and logistics.

Computer vision technology makes it possible to monitor personnel in real time, monitor absence from the workplace, perform tasks and processes, transport flows, and determine the occurrence of emergency situations in a given territory. At the production facilities, the video analytics platform can detect the lack of personal protective equipment on employees, their incorrect use. Also, technologies allow detecting disturbance of order, dangerous inactivity, intersection of control lines.

Our computer vision modules can cover 80% of tasks in the direction of personnel control and safety. Technologies allow you to react much faster, receive a signal about some non-standard, dangerous situations in real time, which can reduce the reaction time to prevent and help. Various scenarios of human behavior are laid down in the video analytics platform, including if the object of observation became bad and, for example, he is unconscious. At the same time, the customization of analytics algorithms can be carried out for a specific task, "said Viktor Konokotin, Deputy Director of the Sitronics Group Product Promotion Department.

Complex tasks for recognizing human actions are solved by algorithms for building a skeletal model of a person. The system tracks the slightest movements and position of body parts, determining more than 20 points on the body of the observation object. This makes it possible to recognize its normal physical behavior and detect deviations in the event of a sharp appearance of movement dynamics.

Also among the basic modules of computer vision of the Sitronics Group video analytics platform is a queue detector, a visitor count, an object heat map, a smoke and open fire detector, as well as, among other things, a module for detecting the impact on the surveillance camera itself (turn, close or damage).