Neural networks have learned to predict plant growth

Scientists from Skoltech trained neural networks to evaluate and predict the nature of plant growth, taking into account all external factors. After analysis, artificial intelligence offers the optimal balance between nutrient requirements and other growth-determining parameters. The study is published in the journal EurekAlert!

Over the past few years, many attempts have been made to use artificial intelligence in almost all areas of life. The use of artificial intelligence for growing plants, including for the needs of agriculture, is no exception.

A group of Skoltech scientists forced the neural networks to work in tandem with computer vision algorithms to predict plant growth and analyze their current state.

The system is based on a single-board prototype Raspberry Pi computer with an external Intel Movidius graphics card. The device uses a compact and powerful Myriad 2 graphics processor, operating at a speed of 150 gigaflops with a power of only 1 W, which is comparable to supercomputers in the mid-1990s.

In the future, scientists intend to bring these computers to a wide market. According to them, the device can become the basis for creating agricultural monitoring networks.

hightech.fm