For cattle, stabilization of feeding behavior plays one of the main roles. Highly productive animals need high-quality feed, full-fledged diets, and compliance with feeding technology. An important task, in addition to providing animals with feed, is to achieve high efficiency of feed use.

It is always necessary to monitor the quantity and quality of feed for cows, as well as the condition of the animals themselves for the productive work of the farm and getting enough milk. The feed can be stored for no more than a day, so it can not be put with a reserve. Therefore, you have to put food at least twice a day. If this is not done, the cows will starve and give less milk.

Vision Systems has developed computer vision-based software that evaluates whether there is enough fresh feed for cows, and also tracks their location and movement around the farm. Based on these data, farmers can track the main metrics that can be used to predict the amount of milk. The effective functioning of dairy cattle breeding in the new economic conditions involves the use of new technologies for feeding, maintenance and breeding.

Using the existing customer surveillance infrastructure, our system allows to determine the availability and quantity of feed and sends signals to farmers when a new supply is required. Thus, it is possible to optimize feed consumption and prevent overfeed or underfeed. Consequently, the feed stopped spoiling, which also had a favorable effect on the expenses of the owners of the farm.

The core of AI the system is a software that uses machine-learning algorithms. We used a large amount of historical data in combination with artificially synthesized samples to train models. Since it was part of a large project, we made the design in the form of a separate module that can be integrated into the server either locally or in the cloud. The module is optimized enough to work in any form.