A deep learning method to predict soil organic carbon content at a regional scale using satellite-based phenology variables

Yang, Lin; Cai, Yanyan; Zhang, Lei; Mao, Guo; Li, Anqi; Zhou, Chenghu

doi:10.1016/j.jag.2021.102428

Cited by 29 publications

(20 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Kenli District is located at the mouth of the Yellow River and forms a typical alluvial fan delta (Yang, Cai, et al, 2021; Yang, Lv, et al, 2021; Yang, Ma, et al, 2021; Yang, Yang, et al, 2021). It belongs to a continental monsoon climate zone, with obvious temperature differences in the four seasons, making it suitable for crop growing.…”

Section: Methodsmentioning

confidence: 99%

“…UAV technology is developing rapidly and is widely used, providing a new high‐precision remote sensing data sources with excellent properties (Ivushkin et al, 2017). Despite their inability to cover large areas, lightweight UAVs can collect multispectral data with centimeter‐to‐decimeter spatial resolution (Yang, Cai, et al, 2021; Yang, Lv, et al, 2021; Yang, Ma, et al, 2021; Yang, Yang, et al, 2021). They are more flexible than space‐borne sensors and can acquire data simultaneously with satellites (Alvarez et al, 2020).…”

Section: Introductionmentioning

confidence: 99%

“…The NMF method has strong advantages in fusing hyperspectral and multispectral data. It can properly select the dimension of feature space and obtain local and spatial details of original data to obtain fusion images with excellent spectral performance and detail retention, which has been widely applied in remote sensing data fusion (Chen et al, 2014; Yang, Cai, et al, 2021; Yang, Lv, et al, 2021; Yang, Ma, et al, 2021; Yang, Yang, et al, 2021). Differential coefficients can obtain different correction coefficients according to the differences between samples, so as to capture heterogeneous relations and achieve more accurate fusion (Madonsela et al, 2021).…”

Section: Introductionmentioning

confidence: 99%

“…Machine learning inversion models have been applied in several studies and were shown to be better than traditional statistical models (Lu et al, 2021; Zhou et al, 2020). Machine learning methods such as extreme gradient‐boosting (XGBoost) (Nguyen et al, 2021), extreme learning machine (ELM) (Guo et al, 2021), random forest (RF) (Tao et al, 2021), and CNN (Yang, Cai, et al, 2021; Yang, Lv, et al, 2021; Yang, Ma, et al, 2021; Yang, Yang, et al, 2021) show great potential in soil salinity monitoring. To improve the accuracy and reliability of the inversion model, the satellite‐UAV‐ground integrated method of monitoring soil salinity using machine learning models needs to be further explored.…”

Section: Introductionmentioning

confidence: 99%

“…UAV technology is developing rapidly and is widely used, providing a new high-precision remote sensing data sources with excellent properties (Ivushkin et al, 2017). Despite their inability to cover large areas, lightweight UAVs can collect multispectral data with centimeter-to-decimeter spatial resolution (Yang, Cai, et al, 2021;Yang, Lv, et al, 2021;.…”

mentioning

confidence: 99%

See 4 more Smart Citations

Soil salinity inversion in coastal cotton growing areas: An integration method using satellite‐ground spectral fusion and satellite‐UAV collaboration

Chang

Yang

et al. 2022

Land Degrad Dev

View full text Add to dashboard Cite

Soil salinity is a crucial factor in agriculture, rising salinity undermines cotton (Gossypium spp.) production in coastal areas of China and damages crops in other countries. In this study, we propose an effective integration method using satellite‐ground spectral fusion and satellite‐unmanned arial vehicle (UAV) collaboration for soil salinity monitoring in cotton growing areas. Firstly, an extreme learning machine (ELM), random forest (RF), and extreme gradient boosting (XGBoost) models were constructed based on UAV images from test areas. The optimal model was selected for soil salinity inversion. Meanwhile, ground imaging hyperspectrum and SENTINEL‐2A multispectral images were differentially fused by nonnegative matrix factorization (NMF). Then, taking the inversion results of UAV as the training sample to build convolutional neural network (CNN) model of the fused SENTINEL‐2A, the soil salinity distribution map of cotton fields in the study area was obtained by inversion, and the satellite‐UAV‐ground integrated inversion of soil salinity in coastal cotton fields was realized. The results showed that the spectrum after satellite‐ground fusion was closer to the original ground hyperspectrum, the fusion improved the correlation between spectrum and soil salinity, and UAV inversion data possessed great potential for the reference data of satellite inversion. The soil salinity obtained by satellite‐UAV‐ground integration approach was highly consistent with the measured salinity in the study area (R2 = 0.805), and the integration approach is suitable for soil salinity inversion in the cotton seedling stage in the coastal area. The satellite‐UAV‐ground integration approach proposed in this study fully tap the advantages of remote sensing data from different platforms and improved the ability to obtain soil salinisation information in large‐scale quantitatively, accurately, and quickly.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

mentioning

confidence: 99%

See 3 more Smart Citations

Soil salinity inversion in coastal cotton growing areas: An integration method using satellite‐ground spectral fusion and satellite‐UAV collaboration

Chang

Yang

et al. 2022

Land Degrad Dev

View full text Add to dashboard Cite

show abstract

Monitoring soil organic carbon under coastal restoration using time series Sentinel‐1

Yang,

Huang

2024

Land Degrad Dev

View full text Add to dashboard Cite

Salt marsh soils are key reservoirs of organic carbon. Although the potential for using Sentinel‐1 data to estimate organic carbon in marsh soils has been recognized, there is still a need to find suitable proxies and effective algorithms that enhance predictions. In this study, we assessed the effectiveness of Sentinel‐1 data for predicting soil organic carbon (SOC) stocks in salt marshes located on the coast of eastern‐central China. We defined the relations between SOC stocks and remotely sensed data using a knowledge‐based approach (i.e., structural equation modeling (SEM)), and the results were compared with those obtained using a common data‐driven approach (i.e., random forest (RF)). Predictive models were developed using (1) refined images associated with phenological stages (phenological variables), (2) refined images that were classified as statistically important using Monte Carlo feature selection (MCFS‐based variables), and (3) images from the combination of these refined datasets. The predictions were validated using a leave‐one‐out cross‐validation approach. The results showed that SEM was more accurate than RF when predicting SOC stocks using the same type of predictor variables. Models using phenological variables alone yielded the least accurate predictions. Adding phenological variables to the structural equation model with MCFS‐based variables increased the prediction accuracy by 36% in terms of the R2. The results of the case study suggest that images related to phenological stages are good predictors for mapping SOC stocks in marshes. This study highlights the superiority of SEM over RF for developing effective remote sensing‐based models to quantify and map SOC stocks in salt marshes.

show abstract

Predictive soil mapping based on the similarity of environmental covariates using a spatial convolutional autoencoder

Zhang

Wang

2023

Soil Science Soc of Amer J

View full text Add to dashboard Cite

Individual predictive soil mapping (iPSM) can predict the soil properties and quantify the prediction uncertainty by using a limited quantity of soil samples. This method assumes that the locations with similar environmental conditions have similar soil properties. The similarity between the locations in the iPSM is calculated based on the environment covariates corresponding to each location. However, this method does not consider the spatial structure information of covariates at the locations. To address this shortcoming, this study proposes a convolutional autoencoder (CAE) model to obtain new covariates called the CAE covariates. These covariates represent the spatial structure information of environment covariates at a certain location. The iPSM based on the similarity calculated using the CAE covariates is applied to predict clay, silt, and sand. A total of 60 sets of compressed CAE covariates are generated based on different spatial neighborhoods and embedded layer sizes. Compared to the original covariates, the iPSM based on the similarity calculated by the CAE covariates improves the prediction accuracy and significantly reduces the prediction uncertainty for clay, silt, and sand. For embedded 20 of neighborhood 9, the improvements in clay, silt, and sand are the highest, and they are 20.2%, 15.8%, and 18.5%, respectively. The largest improvement in the average uncertainty is 96.4% for embedded 5 of neighborhood 13. By applying a range of user‐defined thresholds, the proposed method improves the prediction accuracy for clay, silt, and sand, reduces the prediction uncertainty, and expands the prediction area simultaneously. The trade‐off between the prediction accuracy and the areas that can be predicted by the iPSM is alleviated. The CAE covariates are promising for other machine learning algorithms for digital soil mapping and models that require using spatial environment variables as an input.

show abstract

A deep learning method to predict soil organic carbon content at a regional scale using satellite-based phenology variables

Cited by 29 publications

References 49 publications

Soil salinity inversion in coastal cotton growing areas: An integration method using satellite‐ground spectral fusion and satellite‐UAV collaboration

Soil salinity inversion in coastal cotton growing areas: An integration method using satellite‐ground spectral fusion and satellite‐UAV collaboration

Monitoring soil organic carbon under coastal restoration using time series Sentinel‐1

Predictive soil mapping based on the similarity of environmental covariates using a spatial convolutional autoencoder

Contact Info

Product

Resources

About