Details
| Original language | English |
|---|---|
| Article number | 109494 |
| Number of pages | 8 |
| Journal | Computers and Electronics in Agriculture |
| Volume | 227 |
| Issue number | 1 |
| Early online date | 28 Sept 2024 |
| Publication status | Published - Dec 2024 |
Abstract
The main terrestrial carbon (C) fraction is soil organic carbon (SOC), which has a considerable effect on climate change and greenhouse gas emissions through the absorption and sequestration of carbon dioxide (CO2). This has made SOC assessment very important from both economic and environmental viewpoints. The growing count of soil spectral libraries (SSLs) from regional to global scales has brought a tremendous opportunity for the quantification of SOC through developing spectral-based prediction models. Hence, there is a need to take advantage of big data analytics for spectral data processing. The unique ability of deep learning (DL) techniques to leverage important features of high-dimensional large-scale SSLs has made them top-demanding for more sophisticated modeling. The core objective of the present study was to assess the ability of two different DL algorithms, i.e., one-dimensional convolutional neural network (1DCNN) and fully connected neural network (FCNN) coupled with stacked autoencoder (SAE) feature extraction for SOC prediction based on the data from the land use/cover area frame statistical survey (LUCAS) database. SAE extracted the high-level deep features from the visible–near-infrared–shortwave infrared (Vis–NIR–SWIR) spectra of 11441 soil samples, which were then considered as inputs to the 1DCNN and FCNN models for predicting the SOC content. Both SAE-DL feature-selected models yielded higher accuracy than those the DL developed on the entire spectra and a random forest (RF) model was constructed for comparison. The best prediction was achieved by SAE-1DCNN (R2= 0.78, RMSE = 3.94%, RPD = 4.88, RPIQ = 3.91) followed by 1DCNN (R2= 0.73, RMSE = 5.43%, RPD = 3.67, RPIQ = 2.84) proving the superiority of 1DCNN over FCNN in this study. These results supported the applicability of combined deep features extraction and regression methods for predicting SOC using high dimensional large-scale SSLs.
Keywords
- Convolutional neural network, Deep feature extraction, Fully connected neural network, Large-scale, Soil organic carbon, Soil spectral library, Stacked autoencoder
ASJC Scopus subject areas
- Agricultural and Biological Sciences(all)
- Forestry
- Agricultural and Biological Sciences(all)
- Agronomy and Crop Science
- Computer Science(all)
- Computer Science Applications
- Agricultural and Biological Sciences(all)
- Horticulture
Sustainable Development Goals
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
In: Computers and Electronics in Agriculture, Vol. 227, No. 1, 109494, 12.2024.
Research output: Contribution to journal › Article › Research › peer review
}
TY - JOUR
T1 - Enhancing soil organic carbon prediction of LUCAS soil database using deep learning and deep feature selection
AU - Saberioon, Mohammadmehdi
AU - Gholizadeh, Asa
AU - Ghaznavi, Ali
AU - Chabrillat, Sabine
AU - Khosravi, Vahid
N1 - Publisher Copyright: © 2024 The Authors
PY - 2024/12
Y1 - 2024/12
N2 - The main terrestrial carbon (C) fraction is soil organic carbon (SOC), which has a considerable effect on climate change and greenhouse gas emissions through the absorption and sequestration of carbon dioxide (CO2). This has made SOC assessment very important from both economic and environmental viewpoints. The growing count of soil spectral libraries (SSLs) from regional to global scales has brought a tremendous opportunity for the quantification of SOC through developing spectral-based prediction models. Hence, there is a need to take advantage of big data analytics for spectral data processing. The unique ability of deep learning (DL) techniques to leverage important features of high-dimensional large-scale SSLs has made them top-demanding for more sophisticated modeling. The core objective of the present study was to assess the ability of two different DL algorithms, i.e., one-dimensional convolutional neural network (1DCNN) and fully connected neural network (FCNN) coupled with stacked autoencoder (SAE) feature extraction for SOC prediction based on the data from the land use/cover area frame statistical survey (LUCAS) database. SAE extracted the high-level deep features from the visible–near-infrared–shortwave infrared (Vis–NIR–SWIR) spectra of 11441 soil samples, which were then considered as inputs to the 1DCNN and FCNN models for predicting the SOC content. Both SAE-DL feature-selected models yielded higher accuracy than those the DL developed on the entire spectra and a random forest (RF) model was constructed for comparison. The best prediction was achieved by SAE-1DCNN (R2= 0.78, RMSE = 3.94%, RPD = 4.88, RPIQ = 3.91) followed by 1DCNN (R2= 0.73, RMSE = 5.43%, RPD = 3.67, RPIQ = 2.84) proving the superiority of 1DCNN over FCNN in this study. These results supported the applicability of combined deep features extraction and regression methods for predicting SOC using high dimensional large-scale SSLs.
AB - The main terrestrial carbon (C) fraction is soil organic carbon (SOC), which has a considerable effect on climate change and greenhouse gas emissions through the absorption and sequestration of carbon dioxide (CO2). This has made SOC assessment very important from both economic and environmental viewpoints. The growing count of soil spectral libraries (SSLs) from regional to global scales has brought a tremendous opportunity for the quantification of SOC through developing spectral-based prediction models. Hence, there is a need to take advantage of big data analytics for spectral data processing. The unique ability of deep learning (DL) techniques to leverage important features of high-dimensional large-scale SSLs has made them top-demanding for more sophisticated modeling. The core objective of the present study was to assess the ability of two different DL algorithms, i.e., one-dimensional convolutional neural network (1DCNN) and fully connected neural network (FCNN) coupled with stacked autoencoder (SAE) feature extraction for SOC prediction based on the data from the land use/cover area frame statistical survey (LUCAS) database. SAE extracted the high-level deep features from the visible–near-infrared–shortwave infrared (Vis–NIR–SWIR) spectra of 11441 soil samples, which were then considered as inputs to the 1DCNN and FCNN models for predicting the SOC content. Both SAE-DL feature-selected models yielded higher accuracy than those the DL developed on the entire spectra and a random forest (RF) model was constructed for comparison. The best prediction was achieved by SAE-1DCNN (R2= 0.78, RMSE = 3.94%, RPD = 4.88, RPIQ = 3.91) followed by 1DCNN (R2= 0.73, RMSE = 5.43%, RPD = 3.67, RPIQ = 2.84) proving the superiority of 1DCNN over FCNN in this study. These results supported the applicability of combined deep features extraction and regression methods for predicting SOC using high dimensional large-scale SSLs.
KW - Convolutional neural network
KW - Deep feature extraction
KW - Fully connected neural network
KW - Large-scale
KW - Soil organic carbon
KW - Soil spectral library
KW - Stacked autoencoder
UR - http://www.scopus.com/inward/record.url?scp=85204972149&partnerID=8YFLogxK
U2 - 10.1016/j.compag.2024.109494
DO - 10.1016/j.compag.2024.109494
M3 - Article
AN - SCOPUS:85204972149
VL - 227
JO - Computers and Electronics in Agriculture
JF - Computers and Electronics in Agriculture
SN - 0168-1699
IS - 1
M1 - 109494
ER -