Estimation of botanical composition of forage crops using laboratory-based hyperspectral imaging and near-infrared spectrometer measurements

Harvested forage is the main raw feed for ruminant animals in Sweden, and is commonly cultivated in mixed stands of legume and grass species. The fraction of legume on a dry matter basis, known as botanical composition (BC) is a very important indicator of forage quality. In this study, hyperspectra...

Full description

Saved in:
Bibliographic Details
Main Authors: Junxiang Peng, Maryam Rahimi Jahangirlou, Julien Morel, Zhenjiang Zhou, David Parsons
Format: Article
Language:English
Published: Elsevier 2024-12-01
Series:Journal of Agriculture and Food Research
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2666154324003569
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846121219640786944
author Junxiang Peng
Maryam Rahimi Jahangirlou
Julien Morel
Zhenjiang Zhou
David Parsons
author_facet Junxiang Peng
Maryam Rahimi Jahangirlou
Julien Morel
Zhenjiang Zhou
David Parsons
author_sort Junxiang Peng
collection DOAJ
description Harvested forage is the main raw feed for ruminant animals in Sweden, and is commonly cultivated in mixed stands of legume and grass species. The fraction of legume on a dry matter basis, known as botanical composition (BC) is a very important indicator of forage quality. In this study, hyperspectral imaging and near-infrared spectrometer (NIRS) based methods were used to estimate BC, to overcome the shortcomings of hand separation, which is time and resource consuming. Timothy and red clover mix samples were collected from different harvests in 2017–2019 from multiple sites in Northern Sweden and hand separated. The samples were synthetically mixed to 11 different BC levels, i.e., 0–100 % clover content. Two different instruments (Specim shortwave infrared (SWIR) hyperspectral imaging system and Foss 6500 spectrometer) were used to collect spectral data of samples milled to two levels of coarseness. Three different regression analyses: partial least squares regression (PLSR), support vector regression (SVR) and random forest regression (RFR), were used to build BC estimation models. The effects of the milling particle sizes and the different instruments on the performances of the models were compared. The data from second harvest in 2019 were used for independent validation as evaluation, and the rest of data were randomly split for model calibration (75 %) and validation (25 %). The models were iteratively run 1000 times with different splits, to check the effect from the splitting of calibration and validation datasets. Among different regression analyses, PLSR performed best, with mean Nash-Sutcliffe efficiency (NSE) for model evaluation from 0.76 to 0.87, varying for different instruments and milling sizes. Finer milling made the model accuracies slightly higher. This study developed quick and robust methods to determine the BC of timothy grass and red clover mixtures, which can provide useful information for farmers or researchers.
format Article
id doaj-art-0f22f0beaebd4c0eacaed69ee6d36995
institution Kabale University
issn 2666-1543
language English
publishDate 2024-12-01
publisher Elsevier
record_format Article
series Journal of Agriculture and Food Research
spelling doaj-art-0f22f0beaebd4c0eacaed69ee6d369952024-12-16T05:37:21ZengElsevierJournal of Agriculture and Food Research2666-15432024-12-0118101319Estimation of botanical composition of forage crops using laboratory-based hyperspectral imaging and near-infrared spectrometer measurementsJunxiang Peng0Maryam Rahimi Jahangirlou1Julien Morel2Zhenjiang Zhou3David Parsons4Department of Crop Production Ecology, Swedish University of Agricultural Sciences, 90183 Umeå, Sweden; Corresponding author.Department of Crop Production Ecology, Swedish University of Agricultural Sciences, 90183 Umeå, Sweden; Department of Agroecology, Aarhus University, 8830 Tjele, DenmarkEuropean Commission, Joint Research Centre, 21027 Ispra, ItalyCollege of Biosystems Engineering and Food Science, Zhejiang University, 310058 Hangzhou, ChinaDepartment of Crop Production Ecology, Swedish University of Agricultural Sciences, 90183 Umeå, SwedenHarvested forage is the main raw feed for ruminant animals in Sweden, and is commonly cultivated in mixed stands of legume and grass species. The fraction of legume on a dry matter basis, known as botanical composition (BC) is a very important indicator of forage quality. In this study, hyperspectral imaging and near-infrared spectrometer (NIRS) based methods were used to estimate BC, to overcome the shortcomings of hand separation, which is time and resource consuming. Timothy and red clover mix samples were collected from different harvests in 2017–2019 from multiple sites in Northern Sweden and hand separated. The samples were synthetically mixed to 11 different BC levels, i.e., 0–100 % clover content. Two different instruments (Specim shortwave infrared (SWIR) hyperspectral imaging system and Foss 6500 spectrometer) were used to collect spectral data of samples milled to two levels of coarseness. Three different regression analyses: partial least squares regression (PLSR), support vector regression (SVR) and random forest regression (RFR), were used to build BC estimation models. The effects of the milling particle sizes and the different instruments on the performances of the models were compared. The data from second harvest in 2019 were used for independent validation as evaluation, and the rest of data were randomly split for model calibration (75 %) and validation (25 %). The models were iteratively run 1000 times with different splits, to check the effect from the splitting of calibration and validation datasets. Among different regression analyses, PLSR performed best, with mean Nash-Sutcliffe efficiency (NSE) for model evaluation from 0.76 to 0.87, varying for different instruments and milling sizes. Finer milling made the model accuracies slightly higher. This study developed quick and robust methods to determine the BC of timothy grass and red clover mixtures, which can provide useful information for farmers or researchers.http://www.sciencedirect.com/science/article/pii/S2666154324003569ForageBotanical compositionHyperspectral imagingNIRSParticle size
spellingShingle Junxiang Peng
Maryam Rahimi Jahangirlou
Julien Morel
Zhenjiang Zhou
David Parsons
Estimation of botanical composition of forage crops using laboratory-based hyperspectral imaging and near-infrared spectrometer measurements
Journal of Agriculture and Food Research
Forage
Botanical composition
Hyperspectral imaging
NIRS
Particle size
title Estimation of botanical composition of forage crops using laboratory-based hyperspectral imaging and near-infrared spectrometer measurements
title_full Estimation of botanical composition of forage crops using laboratory-based hyperspectral imaging and near-infrared spectrometer measurements
title_fullStr Estimation of botanical composition of forage crops using laboratory-based hyperspectral imaging and near-infrared spectrometer measurements
title_full_unstemmed Estimation of botanical composition of forage crops using laboratory-based hyperspectral imaging and near-infrared spectrometer measurements
title_short Estimation of botanical composition of forage crops using laboratory-based hyperspectral imaging and near-infrared spectrometer measurements
title_sort estimation of botanical composition of forage crops using laboratory based hyperspectral imaging and near infrared spectrometer measurements
topic Forage
Botanical composition
Hyperspectral imaging
NIRS
Particle size
url http://www.sciencedirect.com/science/article/pii/S2666154324003569
work_keys_str_mv AT junxiangpeng estimationofbotanicalcompositionofforagecropsusinglaboratorybasedhyperspectralimagingandnearinfraredspectrometermeasurements
AT maryamrahimijahangirlou estimationofbotanicalcompositionofforagecropsusinglaboratorybasedhyperspectralimagingandnearinfraredspectrometermeasurements
AT julienmorel estimationofbotanicalcompositionofforagecropsusinglaboratorybasedhyperspectralimagingandnearinfraredspectrometermeasurements
AT zhenjiangzhou estimationofbotanicalcompositionofforagecropsusinglaboratorybasedhyperspectralimagingandnearinfraredspectrometermeasurements
AT davidparsons estimationofbotanicalcompositionofforagecropsusinglaboratorybasedhyperspectralimagingandnearinfraredspectrometermeasurements