Rice Crop Detection Using LSTM, Bi-LSTM, and Machine Learning Models from Sentinel-1 Time Series

Filho, Hugo Crisóstomo de Castro; Júnior, Osmar Abı́lio de Carvalho; Carvalho, Osmar Luiz Ferreira de; de, Pablo Pozzobon; Moura, Rebeca dos Santos de; Albuquerque, Anesmar Olino de; Silva, Cristiano Rosa; Ferreira, Pedro Henrique Guimarães; Guimarães, Renato Fontes; Gomes, Roberto Arnaldo Trancoso

doi:10.3390/rs12162655

Cited by 117 publications

(56 citation statements)

References 127 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The best model occurs in S6, which means that the bi-directional architecture performs better than the single directional architecture for two-day-ahead forecasting after hyperparameter optimisation. The outperformance of the bi-directional architecture has been claimed by many studies [52,53,60,66] in other domains and is still supported by this research when applying to runoff forecasting.…”

Section: Overall Evaluationsupporting

confidence: 72%

See 1 more Smart Citation

Impact of Input Filtering and Architecture Selection Strategies on GRU Runoff Forecasting: A Case Study in the Wei River Basin, Shaanxi, China

Wang

Liu

Yue

et al. 2020

Water

View full text Add to dashboard Cite

A gated recurrent unit (GRU) network, which is a kind of artificial neural network (ANN), has been increasingly applied to runoff forecasting. However, knowledge about the impact of different input data filtering strategies and the implications of different architectures on the GRU runoff forecasting model’s performance is still insufficient. This study has selected the daily rainfall and runoff data from 2007 to 2014 in the Wei River basin in Shaanxi, China, and assessed six different scenarios to explore the patterns of that impact. In the scenarios, four manually-selected rainfall or runoff data combinations and principal component analysis (PCA) denoised input have been considered along with single directional and bi-directional GRU network architectures. The performance has been evaluated from the aspect of robustness to 48 various hypermeter combinations, also, optimized accuracy in one-day-ahead (T + 1) and two-day-ahead (T + 2) forecasting for the overall forecasting process and the flood peak forecasts. The results suggest that the rainfall data can enhance the robustness of the model, especially in T + 2 forecasting. Additionally, it slightly introduces noise and affects the optimized prediction accuracy in T + 1 forecasting, but significantly improves the accuracy in T + 2 forecasting. Though with relevance (R = 0.409~0.763, Grey correlation grade >0.99), the runoff data at the adjacent tributary has an adverse effect on the robustness, but can enhance the accuracy of the flood peak forecasts with a short lead time. The models with PCA denoised input has an equivalent, even better performance on the robustness and accuracy compared with the models with the well manually filtered data; though slightly reduces the time-step robustness, the bi-directional architecture can enhance the prediction accuracy. All the scenarios provide acceptable forecasting results (NSE of 0.927~0.951 for T + 1 forecasting and 0.745~0.836 for T + 2 forecasting) when the hyperparameters have already been optimized. Based on the results, recommendations have been provided for the construction of the GRU runoff forecasting model.

show abstract

Section: Overall Evaluationsupporting

confidence: 72%

“…The combination of several layers of the basic LSTM or GRU cells may enhance the model's ability when learning the nonlinear relationships and provide a better prediction accuracy [53,60]. The typical structure of a stacked GRU network with two layers is shown in Figure 3.…”

Section: Stacked and Bi-directional Gru (Bi-gru)mentioning

confidence: 99%

Impact of Input Filtering and Architecture Selection Strategies on GRU Runoff Forecasting: A Case Study in the Wei River Basin, Shaanxi, China

Wang

Liu

Yue

et al. 2020

Water

View full text Add to dashboard Cite

show abstract

“…DL enables pattern recognition in different data abstraction levels, varying from low-level information (corners and edges), up to high-level information (full objects) [4]. This approach achieves state-of-the-art results in different applications in remote sensing digital image processing [5]: pan-sharpening [6][7][8][9]; image registration [10][11][12][13], change detection [14][15][16][17], object detection [18][19][20][21], semantic segmentation [22][23][24][25], and time series analysis [26][27][28][29]. The classification algorithms applied in remote sensing imagery uses spatial, spectral, and temporal information to extract characteristics from the targets, where a wide variety of targets show significant results: clouds [30][31][32][33], dust-related air pollutant [34][35][36][37] land-cover/land-use [38][39][40][41], urban features [42][43][44][45], and ocean [46][47]…”

Section: Introductionmentioning

confidence: 99%

Instance Segmentation for Large, Multi-Channel Remote Sensing Imagery Using Mask-RCNN and a Mosaicking Approach

et al. 2020

Self Cite

View full text Add to dashboard Cite

Instance segmentation is the state-of-the-art in object detection, and there are numerous applications in remote sensing data where these algorithms can produce significant results. Nevertheless, one of the main problems is that most algorithms use Red, Green, and Blue (RGB) images, whereas Satellite images often present more channels that can be crucial to improve performance. Therefore, the present work brings three contributions: (a) conversion system from ground truth polygon data into the Creating Common Object in Context (COCO) annotation format; (b) Detectron2 software source code adaptation and application on multi-channel imagery; and (c) large scene image mosaicking. We applied the procedure in a Center Pivot Irrigation System (CPIS) dataset with ground truth produced by the Brazilian National Water Agency (ANA) and Landsat-8 Operational Land Imager (OLI) imagery (7 channels with 30-m resolution). Center pivots are a modern irrigation system technique with massive growth potential in Brazil and other world areas. The round shapes with different textures, colors, and spectral behaviors make it appropriate to use Deep Learning instance segmentation. We trained the model using 512 × 512-pixel sized patches using seven different backbone structures (ResNet50- Feature Pyramid Network (FPN), Resnet50-DC5, ResNet50-C4, Resnet101-FPN, Resnet101-DC5, ResNet101-FPN, and ResNeXt101-FPN). The model evaluation used standard COCO metrics (Average Precision (AP), AP50, AP75, APsmall, APmedium, and AR100). ResNeXt101-FPN had the best results, with a 3% advantage over the second-best model (ResNet101-FPN). We also compared the ResNeXt101-FPN model in the seven-channel and RGB imagery, where the multi-channel model had a 3% advantage, demonstrating great improvement using a larger number of channels. This research is also the first with a mosaicking algorithm using instance segmentation models, where we tested in a 1536 × 1536-pixel image using a non-max suppression sorted by area method. The proposed methodology is innovative and suitable for many other remote sensing problems and medical imagery that often present more channels.

show abstract

“…The first category uses supervised classification algorithms which require a large number of training samples of the same rice season. Machine learning techniques such as Random Forest or Support-Vector Machines have been used for rice mapping in India [8], in Myanmar [9], in Vietnam [10], in France [11], in China [12], and in Brazil [13]). More classical methods such as K-means have also been used, for example in India [14] and in China [15].…”

mentioning

confidence: 99%

Understanding Dense Time Series of Sentinel-1 Backscatter from Rice Fields: Case Study in a Province of the Mekong Delta, Vietnam

2021

View full text Add to dashboard Cite

Rice is the primary staple food of more than half of the world’s population and plays an especially important role in global economy, food security, water use, and climate change. The usefulness of Synthetic Aperture Radars (SAR) for rice mapping and monitoring has been demonstrated locally in many studies, in particular in the last five years with the availability of an unprecedented amount of free Sentinel-1 data within the Copernicus program. However, although earlier studies from the 1990s have laid the foundations of the physical understanding of the SAR response of rice fields, the more recent studies tend to overlook this aspect and to favor instead approaches driven by supervised learning which provide accurate results locally but cannot necessarily be extended to wide areas. The objective of this study is to analyze in detail the backscatter temporal variation of rice fields, using Sentinel-1 from 2015 to 2020 and in-situ data for the 5 rice seasons over 2 years 2017–2018, in order to derive robust SAR-based indicators useful for rice monitoring applications, which are essential for planning, monitoring and food security applications. The test region is the An Giang province, in the Mekong River Delta, Vietnam, one of the world’s major rice regions which presents a diversity in rice cultivation practices, in cropping density, and in crop calendar. The SAR data have been analyzed as a function of rice parameters, and the temporal and polarization behaviors of the radar backscatter of different rice varieties have been interpreted physically. New backscatter indicators for the detection of rice paddy area, the estimation of the sowing date, phenological stage and the mapping of the short cycle and long cycle rice varieties have been developed and discussed regarding the generality of the methods with respect to the rice cultural practices and the SAR data characteristics.

show abstract

Rice Crop Detection Using LSTM, Bi-LSTM, and Machine Learning Models from Sentinel-1 Time Series

Cited by 117 publications

References 127 publications

Impact of Input Filtering and Architecture Selection Strategies on GRU Runoff Forecasting: A Case Study in the Wei River Basin, Shaanxi, China

Impact of Input Filtering and Architecture Selection Strategies on GRU Runoff Forecasting: A Case Study in the Wei River Basin, Shaanxi, China

Instance Segmentation for Large, Multi-Channel Remote Sensing Imagery Using Mask-RCNN and a Mosaicking Approach

Understanding Dense Time Series of Sentinel-1 Backscatter from Rice Fields: Case Study in a Province of the Mekong Delta, Vietnam

Contact Info

Product

Resources

About