beeline launched a neural network that finds garbage in yards and warns public utilities

Beeline specialists have developed a solution that allows using CCTV cameras to analyze the house network online.

Artificial intelligence algorithms check the degree of filling of the container, the presence of garbage on the street and assign it the appropriate status “clean-dirty”. Shooting with the camera takes place at a future frequency.

A system for determining pollution parameters according to three criteria: garbage outside the container, the degree of detection of the container and the presence of the container itself on the site. Then she can send to the operator company, which is engaged in garbage collection or the supervisory authority, taking care of cleanliness in the city.

Notification comes to Telegram or email. Changes in status to “clean” or “dirty” are mandatory. Also, the user can check the frame, the neural network assessed the situation around the house.

We tried the pilots in the Moscow and Leningrad meetings. Our portfolio includes products for monitoring the roadway, finding illegal dumps, determining the number of “hares” in public transport, and so on.

Our system can help to quickly and quickly identify the destruction of the local area, solve them in strict accordance with the best way to improve the lives of citizens. Municipal services require time to process complaints, and operating companies reduce the number of fines for late garbage collection.

Another advantage of the Beeline solution is that it is boxed. That is, it can be installed at a customer who has video cameras with the corresponding technical characteristics. At the same time, the system has the necessary knowledge and can be modified to meet the specific needs of the customer. Our MSW monitoring system is easily scalable and allows us to analyze data from a different number of cameras.

— director of artificial intelligence and digital products of Beeline Konstantin Romanov

There are currently solutions on the market in which site cleanliness is usually assessed manually and is subjective and difficult to verify.

beeline is working on improvements to the neural network model. In the near future, operators have: a widespread extended personal account, advanced options for expanding models of work in difficult conditions and at night.