Estimating Grass Sward Quality and Quantity Parameters Using Drone Remote Sensing with Deep Neural Networks

The objective of this study is to investigate the potential of novel neural network architectures for measuring the quality and quantity parameters of silage grass swards, using drone RGB and hyperspectral images (HSI), and compare the results with the random forest (RF) method and handcrafted featu...

Full description

Bibliographic Details
Main Authors: Kirsi Karila, Raquel Alves Oliveira, Johannes Ek, Jere Kaivosoja, Niko Koivumäki, Panu Korhonen, Oiva Niemeläinen, Laura Nyholm, Roope Näsi, Ilkka Pölönen, Eija Honkavaara
Format: Article
Language:English
Published: MDPI AG 2022-06-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/14/11/2692
_version_ 1797491801626509312
author Kirsi Karila
Raquel Alves Oliveira
Johannes Ek
Jere Kaivosoja
Niko Koivumäki
Panu Korhonen
Oiva Niemeläinen
Laura Nyholm
Roope Näsi
Ilkka Pölönen
Eija Honkavaara
author_facet Kirsi Karila
Raquel Alves Oliveira
Johannes Ek
Jere Kaivosoja
Niko Koivumäki
Panu Korhonen
Oiva Niemeläinen
Laura Nyholm
Roope Näsi
Ilkka Pölönen
Eija Honkavaara
author_sort Kirsi Karila
collection DOAJ
description The objective of this study is to investigate the potential of novel neural network architectures for measuring the quality and quantity parameters of silage grass swards, using drone RGB and hyperspectral images (HSI), and compare the results with the random forest (RF) method and handcrafted features. The parameters included fresh and dry biomass (FY, DMY), the digestibility of organic matter in dry matter (D-value), neutral detergent fiber (NDF), indigestible neutral detergent fiber (iNDF), water-soluble carbohydrates (WSC), nitrogen concentration (Ncont) and nitrogen uptake (NU); datasets from spring and summer growth were used. Deep pre-trained neural network architectures, the VGG16 and the Vision Transformer (ViT), and simple 2D and 3D convolutional neural networks (CNN) were studied. In most cases, the neural networks outperformed RF. The normalized root-mean-square errors (NRMSE) of the best models were for FY 19% (2104 kg/ha), DMY 21% (512 kg DM/ha), D-value 1.2% (8.6 g/kg DM), iNDF 12% (5.1 g/kg DM), NDF 1.1% (6.2 g/kg DM), WSC 10% (10.5 g/kg DM), Ncont 9% (2 g N/kg DM), and NU 22% (11.9 N kg/ha) using independent test dataset. The RGB data provided good results, particularly for the FY, DMY, WSC and NU. The HSI datasets provided advantages for some parameters. The ViT and VGG provided the best results with the RGB data, whereas the simple 3D-CNN was the most consistent with the HSI data.
first_indexed 2024-03-10T00:54:29Z
format Article
id doaj.art-18e58c6e52b444fc9aafa571ece01702
institution Directory Open Access Journal
issn 2072-4292
language English
last_indexed 2024-03-10T00:54:29Z
publishDate 2022-06-01
publisher MDPI AG
record_format Article
series Remote Sensing
spelling doaj.art-18e58c6e52b444fc9aafa571ece017022023-11-23T14:45:53ZengMDPI AGRemote Sensing2072-42922022-06-011411269210.3390/rs14112692Estimating Grass Sward Quality and Quantity Parameters Using Drone Remote Sensing with Deep Neural NetworksKirsi Karila0Raquel Alves Oliveira1Johannes Ek2Jere Kaivosoja3Niko Koivumäki4Panu Korhonen5Oiva Niemeläinen6Laura Nyholm7Roope Näsi8Ilkka Pölönen9Eija Honkavaara10Finnish Geospatial Research Institute (FGI), National Land Survey of Finland, 02150 Espoo, FinlandFinnish Geospatial Research Institute (FGI), National Land Survey of Finland, 02150 Espoo, FinlandDepartment of Applied Physics, School of Science, Aalto University, 02150 Espoo, FinlandNatural Resources Institute Finland (Luke), 00790 Helsinki, FinlandFinnish Geospatial Research Institute (FGI), National Land Survey of Finland, 02150 Espoo, FinlandNatural Resources Institute Finland (Luke), 00790 Helsinki, FinlandNatural Resources Institute Finland (Luke), 00790 Helsinki, FinlandFarm Services, Valio Ltd., 00370 Helsinki, FinlandFinnish Geospatial Research Institute (FGI), National Land Survey of Finland, 02150 Espoo, FinlandFaculty of Information Technology, University of Jyväskylä, 40014 Jyväskylä, FinlandFinnish Geospatial Research Institute (FGI), National Land Survey of Finland, 02150 Espoo, FinlandThe objective of this study is to investigate the potential of novel neural network architectures for measuring the quality and quantity parameters of silage grass swards, using drone RGB and hyperspectral images (HSI), and compare the results with the random forest (RF) method and handcrafted features. The parameters included fresh and dry biomass (FY, DMY), the digestibility of organic matter in dry matter (D-value), neutral detergent fiber (NDF), indigestible neutral detergent fiber (iNDF), water-soluble carbohydrates (WSC), nitrogen concentration (Ncont) and nitrogen uptake (NU); datasets from spring and summer growth were used. Deep pre-trained neural network architectures, the VGG16 and the Vision Transformer (ViT), and simple 2D and 3D convolutional neural networks (CNN) were studied. In most cases, the neural networks outperformed RF. The normalized root-mean-square errors (NRMSE) of the best models were for FY 19% (2104 kg/ha), DMY 21% (512 kg DM/ha), D-value 1.2% (8.6 g/kg DM), iNDF 12% (5.1 g/kg DM), NDF 1.1% (6.2 g/kg DM), WSC 10% (10.5 g/kg DM), Ncont 9% (2 g N/kg DM), and NU 22% (11.9 N kg/ha) using independent test dataset. The RGB data provided good results, particularly for the FY, DMY, WSC and NU. The HSI datasets provided advantages for some parameters. The ViT and VGG provided the best results with the RGB data, whereas the simple 3D-CNN was the most consistent with the HSI data.https://www.mdpi.com/2072-4292/14/11/2692droneremote sensinghyperspectralRGBCNNimage transformer
spellingShingle Kirsi Karila
Raquel Alves Oliveira
Johannes Ek
Jere Kaivosoja
Niko Koivumäki
Panu Korhonen
Oiva Niemeläinen
Laura Nyholm
Roope Näsi
Ilkka Pölönen
Eija Honkavaara
Estimating Grass Sward Quality and Quantity Parameters Using Drone Remote Sensing with Deep Neural Networks
Remote Sensing
drone
remote sensing
hyperspectral
RGB
CNN
image transformer
title Estimating Grass Sward Quality and Quantity Parameters Using Drone Remote Sensing with Deep Neural Networks
title_full Estimating Grass Sward Quality and Quantity Parameters Using Drone Remote Sensing with Deep Neural Networks
title_fullStr Estimating Grass Sward Quality and Quantity Parameters Using Drone Remote Sensing with Deep Neural Networks
title_full_unstemmed Estimating Grass Sward Quality and Quantity Parameters Using Drone Remote Sensing with Deep Neural Networks
title_short Estimating Grass Sward Quality and Quantity Parameters Using Drone Remote Sensing with Deep Neural Networks
title_sort estimating grass sward quality and quantity parameters using drone remote sensing with deep neural networks
topic drone
remote sensing
hyperspectral
RGB
CNN
image transformer
url https://www.mdpi.com/2072-4292/14/11/2692
work_keys_str_mv AT kirsikarila estimatinggrassswardqualityandquantityparametersusingdroneremotesensingwithdeepneuralnetworks
AT raquelalvesoliveira estimatinggrassswardqualityandquantityparametersusingdroneremotesensingwithdeepneuralnetworks
AT johannesek estimatinggrassswardqualityandquantityparametersusingdroneremotesensingwithdeepneuralnetworks
AT jerekaivosoja estimatinggrassswardqualityandquantityparametersusingdroneremotesensingwithdeepneuralnetworks
AT nikokoivumaki estimatinggrassswardqualityandquantityparametersusingdroneremotesensingwithdeepneuralnetworks
AT panukorhonen estimatinggrassswardqualityandquantityparametersusingdroneremotesensingwithdeepneuralnetworks
AT oivaniemelainen estimatinggrassswardqualityandquantityparametersusingdroneremotesensingwithdeepneuralnetworks
AT lauranyholm estimatinggrassswardqualityandquantityparametersusingdroneremotesensingwithdeepneuralnetworks
AT roopenasi estimatinggrassswardqualityandquantityparametersusingdroneremotesensingwithdeepneuralnetworks
AT ilkkapolonen estimatinggrassswardqualityandquantityparametersusingdroneremotesensingwithdeepneuralnetworks
AT eijahonkavaara estimatinggrassswardqualityandquantityparametersusingdroneremotesensingwithdeepneuralnetworks