Construction of a drought monitoring model using deep learning based on multi-source remote sensing data

Shen, Runping; Huang, Anqi; Li, Bolun; Guo, Jia

doi:10.1016/j.jag.2019.03.006

Cited by 79 publications

(34 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In recent years, random forest has gained popularity as an effective classification method in the remote sensing domain [84][85][86]. Results from our study additionally confirm that the random forest ensemble is a robust and accurate method for regression type applications as well.…”

Section: Discussionsupporting

confidence: 73%

Uncertainty Analysis of Remote Sensing Pretreatment for Biomass Estimation on Landsat OLI and Landsat ETM+

Zhang

et al. 2020

IJGI

View full text Add to dashboard Cite

The accurate quantification of biomass helps to understand forest productivity and carbon cycling dynamics. Research on uncertainty during pretreatment is still lacking despite it being one of the major sources of uncertainty and an essential step in biomass estimation. In this study, we investigated pretreatment uncertainty and conducted a comparative study on the uncertainty of three optical imagery preprocessing stages (radiometric calibration, atmospheric and terrain correction) in biomass estimation. A combination of statistical models (random forest) and multisource data (Landsat enhanced thematic mapper plus (ETM+), Landsat operational land imager (OLI), national forest inventory (NFI)) was used to estimate forest biomass. Particularly, mean absolute error (MAE) and relative error (RE) were used to assess and quantify the uncertainty of each pretreatment, while the coefficient of determination (R2) was employed to evaluate the accuracy of the model. The results obtained show that random forest (RF) and 10-fold cross validation algorithms provided reliable accuracy for biomass estimation to better understand the uncertainty in pretreatments. In this study, there was a considerable uncertainty in biomass estimation using original OLI and ETM+ images from. Uncertainty was lower after data processing, emphasizing the importance of pretreatments for improving accuracy in biomass estimation. Further, the effects of three pretreatments on uncertainty of biomass estimation were objectively quantified. In this study (results of test sample), a 33.70% uncertainty was found in biomass estimation using original images from the OLI, and a 34.28% uncertainty in ETM+. Radiometric calibration slightly increased the uncertainty of biomass estimation (OLI increased by 1.38%, ETM+ increased by 2.08%). Moreover, atmospheric correction (5.56% for OLI, 4.41% for ETM+) and terrain correction (1.00% for OLI, 1.67% for ETM+) significantly reduced uncertainty for OLI and ETM+, respectively. This is an important development in the field of improving the accuracy of biomass estimation by remote sensing. Notably, the three pretreatments presented the same trend in uncertainty during biomass estimation using OLI and ETM+. This may exhibit the same effects in other optical images. This article aims to quantify uncertainty in pretreatment and to analyze the resultant effects to provide a theoretical basis for improving the accuracy of biomass estimation.

show abstract

Section: Discussionsupporting

confidence: 73%

Uncertainty Analysis of Remote Sensing Pretreatment for Biomass Estimation on Landsat OLI and Landsat ETM+

Zhang

et al. 2020

IJGI

View full text Add to dashboard Cite

show abstract

“…In addition, if the seeds are not severely wet (seed color does not change and seeds do not swell), the relative humidity has no influence on the method of hyperspectral imaging combined deep learning. At present, deep networks have been successfully applied to plant disease identification [36][37][38], drought monitoring [39], land type classification [40], weed detection [41], and other areas of agriculture. To date, there are few reports on the identification of soybean seed varieties by deep learning, and whether it has advantages that is also unknown.…”

mentioning

confidence: 99%

A Rapid and Highly Efficient Method for the Identification of Soybean Seed Varieties: Hyperspectral Images Combined with Transfer Learning

et al. 2019

View full text Add to dashboard Cite

Convolutional neural network (CNN) can be used to quickly identify crop seed varieties. 1200 seeds of ten soybean varieties were selected, hyperspectral images of both the front and the back of the seeds were collected, and the reflectance of soybean was derived from the hyperspectral images. A total of 9600 images were obtained after data augmentation, and the images were divided into a training set, validation set, and test set with a 3:1:1 ratio. Pretrained models (AlexNet, ResNet18, Xception, InceptionV3, DenseNet201, and NASNetLarge) after fine-tuning were used for transfer training. The optimal CNN model for soybean seed variety identification was selected. Furthermore, the traditional machine learning models for soybean seed variety identification were established by using reflectance as input. The results show that the six models all achieved 91% accuracy in the validation set and achieved accuracy values of 90.6%, 94.5%, 95.4%, 95.6%, 96.8%, and 97.2%, respectively, in the test set. This method is better than the identification of soybean seed varieties based on hyperspectral reflectance. The experimental results support a novel method for identifying soybean seeds rapidly and accurately, and this method also provides a good reference for the identification of other crop seeds.

show abstract

“…These comprehensive approaches for monitoring drought include different data-driven models such as the time series model, probabilistic model and traditional regression model, which all have the limitation of dealing with non-linear properties of the remote sensing data set [21][22][23]. In this regard, a deep neural network is more flexible and robust in the case of drought characterization and forecasting skill, by extracting the non-linear relationship between different drought factors compared with traditional models [24,25].…”

Section: Introductionmentioning

confidence: 99%

Deep Learning for Monitoring Agricultural Drought in South Asia Using Remote Sensing Data

et al. 2021

View full text Add to dashboard Cite

Drought, a climate-related disaster impacting a variety of sectors, poses challenges for millions of people in South Asia. Accurate and complete drought information with a proper monitoring system is very important in revealing the complex nature of drought and its associated factors. In this regard, deep learning is a very promising approach for delineating the non-linear characteristics of drought factors. Therefore, this study aims to monitor drought by employing a deep learning approach with remote sensing data over South Asia from 2001–2016. We considered the precipitation, vegetation, and soil factors for the deep forwarded neural network (DFNN) as model input parameters. The study evaluated agricultural drought using the soil moisture deficit index (SMDI) as a response variable during three crop phenology stages. For a better comparison of deep learning model performance, we adopted two machine learning models, distributed random forest (DRF) and gradient boosting machine (GBM). Results show that the DFNN model outperformed the other two models for SMDI prediction. Furthermore, the results indicated that DFNN captured the drought pattern with high spatial variability across three penology stages. Additionally, the DFNN model showed good stability with its cross-validated data in the training phase, and the estimated SMDI had high correlation coefficient R2 ranges from 0.57~0.90, 0.52~0.94, and 0.49~0.82 during the start of the season (SOS), length of the season (LOS), and end of the season (EOS) respectively. The comparison between inter-annual variability of estimated SMDI and in-situ SPEI (standardized precipitation evapotranspiration index) showed that the estimated SMDI was almost similar to in-situ SPEI. The DFNN model provides comprehensive drought information by producing a consistent spatial distribution of SMDI which establishes the applicability of the DFNN model for drought monitoring.

show abstract

Construction of a drought monitoring model using deep learning based on multi-source remote sensing data

Cited by 79 publications

References 23 publications

Uncertainty Analysis of Remote Sensing Pretreatment for Biomass Estimation on Landsat OLI and Landsat ETM+

Uncertainty Analysis of Remote Sensing Pretreatment for Biomass Estimation on Landsat OLI and Landsat ETM+

A Rapid and Highly Efficient Method for the Identification of Soybean Seed Varieties: Hyperspectral Images Combined with Transfer Learning

Deep Learning for Monitoring Agricultural Drought in South Asia Using Remote Sensing Data

Contact Info

Product

Resources

About