REMOTE SENSING AND SENSOR TECHNOLOGIES APPLICATION FOR DATA COLLECTION AND PROCESSING IN PRECISION AGRICULTURE
Rubrics: AGRICULTURAL ECONOMICS AND POLICIES
Authors:
1.
2.
Federal State Budgetary Educational Institution of Higher Education "Saint Petersburg State University"
Type:
Article
Pages:
from
to
Status:
Published
Received:
06.12.2024
Accepted:
15.12.2024
Published:
25.12.2024
Language:
Russian
Keywords:
distancionnoe zondirovanie, aerofotos'emka, besprovodnye sensornye seti, tochnoe zemledelie, datasety, mul'tispektral'naya s'emka, giperspektral'naya s'emka
Abstract (English):
Currently, artificial intelligence methods are increasingly being used to solve precision agriculture problems, while, as before, there is a shortage of high-quality labeled source information. Researchers often have to synthesize data by artificially expanding datasets due to limited resources for training models. At the same time, the effectiveness and accuracy of scientific computational experiments directly depend on the initial data. Therefore, there is a need to develop a set of approaches and tools for the rapid collection and pre-processing of data in precision agriculture. In this study, two directions were chosen: the use of remote sensing and sensor technologies. The objects of the study are experimental agricultural fields located in the Leningrad region, where ground-based measurements of agroecological parameters are collected annually, as well as the laying of test sites. Two unmanned aerial systems were used for aerial photography: based on Geoscan-401 and DJI Matrice. In the course of the study, approaches were developed for the collection and preprocessing of multispectral and hyperspectral aerophotos in precision agriculture, the creation of multi-layered specialized datasets. At the same time, in addition to preprocessed geo-linked orthophotomaps, algorithms for creating additional vector layers with appropriate markings (based on ground measurements) were worked out. As a second direction of collecting information reflecting the state of the agricultural field, a prototype of a wireless sensor network was developed: the architecture of the sensor node, as well as the base station, was proposed. Prototype solutions have been implemented and pre-tested. The main tasks have been identified as areas of work development.
Currently, artificial intelligence methods are increasingly being used to solve precision agriculture problems, while, as before, there is a shortage of high-quality labeled source information. Researchers often have to synthesize data by artificially expanding datasets due to limited resources for training models. At the same time, the effectiveness and accuracy of scientific computational experiments directly depend on the initial data. Therefore, there is a need to develop a set of approaches and tools for the rapid collection and pre-processing of data in precision agriculture. In this study, two directions were chosen: the use of remote sensing and sensor technologies. The objects of the study are experimental agricultural fields located in the Leningrad region, where ground-based measurements of agroecological parameters are collected annually, as well as the laying of test sites. Two unmanned aerial systems were used for aerial photography: based on Geoscan-401 and DJI Matrice. In the course of the study, approaches were developed for the collection and preprocessing of multispectral and hyperspectral aerophotos in precision agriculture, the creation of multi-layered specialized datasets. At the same time, in addition to preprocessed geo-linked orthophotomaps, algorithms for creating additional vector layers with appropriate markings (based on ground measurements) were worked out. As a second direction of collecting information reflecting the state of the agricultural field, a prototype of a wireless sensor network was developed: the architecture of the sensor node, as well as the base station, was proposed. Prototype solutions have been implemented and pre-tested. The main tasks have been identified as areas of work development.
Keywords:
distancionnoe zondirovanie, aerofotos'emka, besprovodnye sensornye seti, tochnoe zemledelie, datasety, mul'tispektral'naya s'emka, giperspektral'naya s'emka
distancionnoe zondirovanie, aerofotos'emka, besprovodnye sensornye seti, tochnoe zemledelie, datasety, mul'tispektral'naya s'emka, giperspektral'naya s'emka
