November 16. /TASS/. Scientists at Samara University will develop a system of "smart farming" based on compact unmanned aerial vehicles. The project, which is scheduled to be completed next year, will create a global cloud service for storing and processing hyperspectral images, the university's press service said.
"Scientists at Samara National Research University will develop a system of 'smart farming' based on compact unmanned aerial vehicles. The system will include drones equipped with an ultra-compact hyperspectrometer and a cloud platform for collecting, storing and processing hyperspectral images — the so-called 'hypercubes'. Artificial intelligence will be engaged in image analysis. Work on the project is being carried out jointly with Byte-Samara company within the framework of a grant from the Fund for the Promotion of Innovations," the message reads.
The university plans to complete the project by June 2021. "At present, several experimental samples of the hyperspectrometer have been manufactured, the university specialists are optimizing the design of the device and testing the technical process for mass production of diffractive optical elements. Flight tests with trial processing of hyperspectral imagery data are scheduled for spring," the press service reported.
"The compact modular airborne hyperspectrometer will be installed on unmanned aerial vehicles to obtain various important information for an agricultural producer, for example, on the presence of weeds or areas of crops damaged by pests in the field. The scientific novelty of the development lies in the use of unique diffractive optical elements, which allow to significantly reduce and cut the cost of the design of the hyperspectrometer. The microrelief of such elements simultaneously contains the structures of a harmonic lens and a diffraction grating, which makes it possible to combine in one element the functions of imaging and decomposition into a spectrum," the press service quoted Pavel Yakimov, Associate Professor of the Department of Supercomputers and General Informatics at Samara University.
The device is planned to be installed on Zhuzha drone, developed by the Byte-Samara company, its dimensions do not exceed 30 cm. "In the future, the device can be used on other types of domestic drones capable of lifting more than 300 g. In addition to the hyperspectrometer itself, the weight of which will amount to approximately 100 g, a visible range camera and a single-board mini-computer will also be installed on board the drone," the university's press service added.
The drone will be able to operate on a given route either under ground stations control or in automatic mode using machine vision. Depending on the speed and height of a shoot, one such device can shoot up to 1,500 hectares per day. The developers expect that the project should significantly reduce the cost of hyperspectrometers used for monitoring in agriculture. "Due to the high cost of such equipment, there are now very few public libraries of hyperspectral data in the world. The appearance of available devices will make it possible to multiply the array of publicly available hypercubes and form an array of data sufficient to teach neural networks to analyze and qualify such images qualitatively," the statement reads.