AUTS: A Novel Approach to Mapping Winter Wheat by Automatically Updating Training Samples Based on NDVI Time Series

Wang, Chunyang; Zhang, Huan; Wu, Xuan Hui; Yang, Wei; Shen, Yanjun; Lu, Bao-Liang; Wang, Jianlong

doi:10.3390/agriculture12060817

Cited by 5 publications

(1 citation statement)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, it adopts window self-attention to reduce the computational complexity. This model can produce satisfactory results in tasks such as instance segmentation, semantic segmentation and object detection [74].…”

Section: Deeplabv3+ and Improvementmentioning

confidence: 96%

Identifying Winter Wheat Using Landsat Data Based on Deep Learning Algorithms in the North China Plain

Zhang,

Wang,

Wang

et al. 2023

Remote Sensing

View full text Add to dashboard Cite

The North China Plain (NCP) represents a significant agricultural production region in China, with winter wheat serving as one of its main grain crops. Accurate identification of winter wheat through remote sensing technology holds significant importance in ensuring food security in the NCP. In this study, we have utilized Landsat 8 and Landsat 9 imagery to identify winter wheat in the NCP. Multiple convolutional neural networks (CNNs) and transformer networks, including ResNet, HRNet, MobileNet, Xception, Swin Transformer and SegFormer, are used in order to understand their uncertainties in identifying winter wheat. At the same time, these deep learning (DL) methods are also compared to the traditional random forest (RF) method. The results indicated that SegFormer outperformed all methods, of which the accuracy is 0.9252, the mean intersection over union (mIoU) is 0.8194 and the F1 score (F1) is 0.8459. These DL methods were then applied to monitor the winter wheat planting areas in the NCP from 2013 to 2022, and the results showed a decreasing trend.

show abstract

Section: Deeplabv3+ and Improvementmentioning

confidence: 96%

Identifying Winter Wheat Using Landsat Data Based on Deep Learning Algorithms in the North China Plain

Zhang,

Wang,

Wang

et al. 2023

Remote Sensing

View full text Add to dashboard Cite

show abstract

Refinement of Cropland Data Layer with Effective Confidence Layer Interval and Image Filtering

Maleki,

Wu,

Oubara

et al. 2024

Agriculture

View full text Add to dashboard Cite

Various systems have been developed to process agricultural land data for better management of crop production. One such system is Cropland Data Layer (CDL), produced by the National Agricultural Statistics Service of the United States Department of Agriculture (USDA). The CDL has been widely used for training deep learning (DL) segmentation models. However, it contains various errors, such as salt-and-pepper noise, and must be refined before being used in DL training. In this study, we used two approaches to refine the CDL for DL segmentation of major crops from a time series of Sentinel-2 monthly composite images. Firstly, different confidence intervals of the confidence layer were used to refine the CDL. Secondly, several image filters were employed to improve data quality. The refined CDLs were then used as the ground-truth in DL segmentation training and evaluation. The results demonstrate that the CDL with +45% and +55% confidence intervals produced the best results, improving the accuracy of DL segmentation by approximately 1% compared to non-refined data. Additionally, filtering the CDL using the majority and expand–shrink filters yielded the best performance, enhancing the evaluation metrics by about 1.5%. The findings suggest that pre-filtering the CDL and selecting an effective confidence interval can significantly improve DL segmentation performance, contributing to more accurate and reliable agricultural monitoring.

show abstract

Assessment of Forest Fire Severity for a Management Conceptual Model: Case Study in Vilcabamba, Ecuador

González,

Morante-Carballo,

González

et al. 2024

Forests

View full text Add to dashboard Cite

Wildfires are affecting natural ecosystems worldwide, causing economic and human losses and exacerbated by climate change. Models of fire severity and fire susceptibility are crucial tools for fire monitoring. This case study analyses a fire event on 3 September 2019 in Vilcabamba parish, Loja province, Ecuador. This article aims to assess the severity and susceptibility of a fire through spectral indices and multi-criteria methods for establishing a fire action plan proposal. The methodology comprises the following: (i) the acquisition of Sentinel-2A products for the calculation of spectral indices; (ii) a fire severity model using differentiated indices (dNBR and dNDVI) and a fire susceptibility model using the Analytic Hierarchy Process (AHP) method; (iii) model validation using Logistic Regression (LR) and Non-metric Multidimensional Scaling (NMDS) algorithms; (iv) the proposal of an action plan for fire management. The Normalised Burn Ratio (NBR) index revealed that 10.98% of the fire perimeter has burned areas with moderate-high severity in post-fire scenes (2019) and decreased to 0.01% for post-fire scenes in 2021. The Normalised Difference Vegetation Index (NDVI) identified 67.28% of the fire perimeter with null photosynthetic activity in the post-fire scene (2019) and 5.88% in the post-fire scene (2021). The Normalised Difference Moisture Index (NDMI) applied in the pre-fire scene identified that 52.62% has low and dry vegetation (northeast), and 8.27% has high vegetation cover (southwest). The dNDVI identified 10.11% of unburned areas and 7.91% using the dNBR. The fire susceptibility model identified 11.44% of the fire perimeter with null fire susceptibility. These results evidence the vegetation recovery after two years of the fire event. The models demonstrated excellent performance for fire severity models and were a good fit for the AHP model. We used the Root Mean Square Error (RMSE) and area under the curve (AUC); dNBR and dNDVI have an RMSE of 0.006, and the AHP model has an RMSE of 0.032. The AUC = 1.0 for fire severity models and AUC = 0.6 for fire susceptibility. This study represents a holistic approach by combining Google Earth Engine (GEE), Geographic Information System (GIS), and remote sensing tools for proposing a fire action plan that supports decision making. This study provides escape routes that considered the most significant fire triggers, the AHP, and fire severity approaches for monitoring wildfires in Andean regions.

show abstract

AUTS: A Novel Approach to Mapping Winter Wheat by Automatically Updating Training Samples Based on NDVI Time Series

Cited by 5 publications

References 43 publications

Identifying Winter Wheat Using Landsat Data Based on Deep Learning Algorithms in the North China Plain

Identifying Winter Wheat Using Landsat Data Based on Deep Learning Algorithms in the North China Plain

Refinement of Cropland Data Layer with Effective Confidence Layer Interval and Image Filtering

Assessment of Forest Fire Severity for a Management Conceptual Model: Case Study in Vilcabamba, Ecuador

Contact Info

Product

Resources

About