A team of scientists from the Biological Faculty of Moscow State University, the Federal Scientific Center (FNTs) named after Michurin, Tambov State University, the VIM Agroengineering Center, together with Yandex, have created a system for monitoring agricultural plants based on artificial intelligence, the press service of the search engine told Vedomosti. The new tool will allow farmers to predict the yields of certain crops, assess their condition at different stages of ripening, and, after harvesting, plan the chain of logistics and distribution of agricultural products, his representative says.
According to the developers, the farmer will be able to shoot land (for example, an orchard) on a smartphone camera and upload a photo to Yandex.Cloud – there it will be processed by a specially trained neural network that compares the image with reference images of fruits or inflorescences. The learning process is painstaking – in order to fill the network with data, each plant species needs to be photographed 3000 times. “The algorithm in real time generates a forecast in the form of graphs and tables, created in the cloud data visualization service Yandex DataLens. With its help, you can show on the map even individual trees with estimates of the yield for each ”, – assures the representative of“ Yandex ”.
.