Seed purity assessment by means of spectral imaging

In this work, we propose a technique for identifying impurity grains from spectral images using neural networks that is able to analyze a heap of seeds, grouping grains with similar spectral and morphological characteristics and optimizing the main stages of forming a training sample of a neural net...

Full description

Saved in:
Bibliographic Details
Main Authors: G.V. Nesterov, A.V. Guryleva, A.A. Zolotukhina, D.S. Fomin, Y.K. Shashko, A.S. Machikhin
Format: Article
Language:English
Published: Samara National Research University 2025-06-01
Series:Компьютерная оптика
Subjects:
Online Access:https://computeroptics.ru/eng/KO/Annot/KO49-3/490312e.html
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849230221798014976
author G.V. Nesterov
A.V. Guryleva
A.A. Zolotukhina
D.S. Fomin
D.S. Fomin
Y.K. Shashko
A.S. Machikhin
author_facet G.V. Nesterov
A.V. Guryleva
A.A. Zolotukhina
D.S. Fomin
D.S. Fomin
Y.K. Shashko
A.S. Machikhin
author_sort G.V. Nesterov
collection DOAJ
description In this work, we propose a technique for identifying impurity grains from spectral images using neural networks that is able to analyze a heap of seeds, grouping grains with similar spectral and morphological characteristics and optimizing the main stages of forming a training sample of a neural network model, recording and processing data. An architecture of the neural network model is proposed based on sequentially running LSTM layers and fully connected layers of neurons. Approaches are proposed for choosing the training sample size, the number and position of central wavelengths of video spectrometer channels used in analysis, and a method for segmenting spectral images to form a training sample. The developed methodology is distinguished by the ability to analyze a heap of seeds and the ease of replenishing the database of distinguished crops and impurities. Testing of the method on wheat and barley seeds showed high classification accuracy (over 99 %) even for grains with very similar spectral and morphological characteristics. The proposed approach increases the accuracy, productivity and objectivity of assessing the purity of seed material, does not require the involvement of experienced personnel and, thus, may be expected to facilitate the introduction of video spectrometers when addressing research and production problems of the agro-industrial complex.
format Article
id doaj-art-4b9bfb2bb550431ab3fd8aa07782cb1c
institution Kabale University
issn 0134-2452
2412-6179
language English
publishDate 2025-06-01
publisher Samara National Research University
record_format Article
series Компьютерная оптика
spelling doaj-art-4b9bfb2bb550431ab3fd8aa07782cb1c2025-08-21T06:58:28ZengSamara National Research UniversityКомпьютерная оптика0134-24522412-61792025-06-0149346146910.18287/2412-6179-CO-1512Seed purity assessment by means of spectral imagingG.V. Nesterov0A.V. Guryleva1A.A. Zolotukhina2D.S. Fomin3D.S. Fomin4Y.K. Shashko5A.S. Machikhin6Scientific and Technological Centre of Unique Instrumentation of the Russian Academy of Sciences; PREDURALIE LtdScientific and Technological Centre of Unique Instrumentation of the Russian Academy of SciencesScientific and Technological Centre of Unique Instrumentation of the Russian Academy of Sciences; PREDURALIE LtdPREDURALIE Ltd; Perm Federal Research Center, Ural Branch of the Russian Academy of SciencesPREDURALIE Ltd; Perm Federal Research Center, Ural Branch of the Russian Academy of SciencesRepublican Scientific Subsidiary Unitary Enterprise "The Institute for Soil Science and Agrochemistry"Scientific and Technological Centre of Unique Instrumentation of the Russian Academy of Sciences; PREDURALIE LtdIn this work, we propose a technique for identifying impurity grains from spectral images using neural networks that is able to analyze a heap of seeds, grouping grains with similar spectral and morphological characteristics and optimizing the main stages of forming a training sample of a neural network model, recording and processing data. An architecture of the neural network model is proposed based on sequentially running LSTM layers and fully connected layers of neurons. Approaches are proposed for choosing the training sample size, the number and position of central wavelengths of video spectrometer channels used in analysis, and a method for segmenting spectral images to form a training sample. The developed methodology is distinguished by the ability to analyze a heap of seeds and the ease of replenishing the database of distinguished crops and impurities. Testing of the method on wheat and barley seeds showed high classification accuracy (over 99 %) even for grains with very similar spectral and morphological characteristics. The proposed approach increases the accuracy, productivity and objectivity of assessing the purity of seed material, does not require the involvement of experienced personnel and, thus, may be expected to facilitate the introduction of video spectrometers when addressing research and production problems of the agro-industrial complex.https://computeroptics.ru/eng/KO/Annot/KO49-3/490312e.htmlvideospectrometryhyperspectral imagingdigital image processingspectral characteristicsmachine learninglstm neural networkseed materialagriculture
spellingShingle G.V. Nesterov
A.V. Guryleva
A.A. Zolotukhina
D.S. Fomin
D.S. Fomin
Y.K. Shashko
A.S. Machikhin
Seed purity assessment by means of spectral imaging
Компьютерная оптика
videospectrometry
hyperspectral imaging
digital image processing
spectral characteristics
machine learning
lstm neural network
seed material
agriculture
title Seed purity assessment by means of spectral imaging
title_full Seed purity assessment by means of spectral imaging
title_fullStr Seed purity assessment by means of spectral imaging
title_full_unstemmed Seed purity assessment by means of spectral imaging
title_short Seed purity assessment by means of spectral imaging
title_sort seed purity assessment by means of spectral imaging
topic videospectrometry
hyperspectral imaging
digital image processing
spectral characteristics
machine learning
lstm neural network
seed material
agriculture
url https://computeroptics.ru/eng/KO/Annot/KO49-3/490312e.html
work_keys_str_mv AT gvnesterov seedpurityassessmentbymeansofspectralimaging
AT avguryleva seedpurityassessmentbymeansofspectralimaging
AT aazolotukhina seedpurityassessmentbymeansofspectralimaging
AT dsfomin seedpurityassessmentbymeansofspectralimaging
AT dsfomin seedpurityassessmentbymeansofspectralimaging
AT ykshashko seedpurityassessmentbymeansofspectralimaging
AT asmachikhin seedpurityassessmentbymeansofspectralimaging