An Interannual Transfer Learning Approach for Crop Classification in the Hetao Irrigation District, China

Hu, Yueran; Zeng, Hongwei; Tian, Fuyou; Zhang, Miao; Wu, Bingfang; Gilliams, Sven; Li, Sen; Li, Yuanchao; Lu, Y.; Yang, Honghai

doi:10.3390/rs14051208

Cited by 31 publications

(23 citation statements)

References 76 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently, different researchers have addressed different problems related to transfer learning in agriculture. The authors of [32,33] proposed a knowledge-transfer model to classify different crops and reduce the retraining and labelling time. In this work, authors reduced 20% of the time compared to the normal time.…”

Section: Transfer Learning For Agriculture and Irrigation Systemmentioning

confidence: 99%

Sustainable Irrigation Requirement Prediction Using Internet of Things and Transfer Learning

Blessy

Kumar

Prabagaran³

et al. 2023

Sustainability

View full text Add to dashboard Cite

Irrigation systems are a crucial research area because it is essential to conserve fresh water and utilize it wisely. As a part of this study, the reliability of predicting the usage of water in the present and future is investigated in order to develop an effective prediction model to communicate demand. In order to improve prediction, we develop a prediction model and share the updated model with nearby farmers. In order to forecast the irrigation requirements, the recommended model utilizes the Internet of Things (IoT), k-nearest neighbours (KNN), cloud storage, long short-term memory (LSTM), and adaptive network fuzzy inference system (ANFIS) techniques. By collecting real-time environmental data, KNN identifies the closest water requirement from the roots and its surrounding. In order to predict short-term requirements, ANFIS is used. To transfer the new requirements for better prediction, transfer learning is used. Time-series-data updates are predicted using LSTM for future forecasting, and the integrated model is shared with other farmers using cloud environments to enhance forecasting and analysis. For implementation, a period of nine to ten months of data was collected from February to December 2021, and banana tree was used to implement the planned strategy. Four farms, with measurements, were considered at varying intervals to determine the minimum and maximum irrigation needs. The requirements of farms were collected over time and compared to the predictions. Future requirements at 8, 16, 24, 32, and 48 h were also anticipated. The results indicated were compared to manual water pouring, and, thus, the entire crop used less water, making our prediction model a real-world option for irrigation. The prediction model was evaluated using R2, MSLE and the average initial prediction value of R2 was 0.945. After using transfer learning, the prediction of the model of Farm-2, 3 and 4 were 0.951, 0.958 and 0.967, respectively.

show abstract

Section: Transfer Learning For Agriculture and Irrigation Systemmentioning

confidence: 99%

Sustainable Irrigation Requirement Prediction Using Internet of Things and Transfer Learning

Blessy

Kumar

Prabagaran³

et al. 2023

Sustainability

View full text Add to dashboard Cite

show abstract

“…There were also obvious signs of increase in vegetation in farmlands in the sub-basin III, which mainly located at the Hetao Plain and Ningxia Plain. These plains depend on the Yellow River to nurture developed irrigated agriculture (Hu et al, 2022), and the widespread planting of crops promotes…”

Section: Spatial Heterogeneity Of the Effects Of Driving Factors On Fvcmentioning

confidence: 99%

Spatiotemporal evolution of growing‐season vegetation coverage and its natural‐human drivers in the Yellow River Basin, China

Zang,

Yu,

et al. 2023

Land Degrad Dev

View full text Add to dashboard Cite

The Yellow River Basin (YRB) is an important ecological security barrier in northern China, where vegetation has undergone rapid changes under the high‐intensity anthropological activities and ongoing climate change in recent decades. This study focused on the spatiotemporal evolution of fractional vegetation cover (FVC) and the driving factors during the growing season in different sub‐basins of the YRB between 2000 and 2020. The trend analysis was employed to examine the FVC spatiotemporal changes, and the geographical detector model (GDM) and the geographically weighted regression (GWR) model were utilized to explore spatial heterogeneity of natural‐human drivers on FVC. The 21‐year average FVC was 0.5780 in the growing season in the YRB, and the southeast region was greater than the northwest region. The FVC exhibited a fluctuating and rising trend of 0.44% yr−1 in the whole YRB from 2000 to 2020, and areas with improved FVC were 36.40% more than those with degraded FVC. The dominant factors influencing the growing‐season FVC were precipitation and temperature, followed by elevation and human activity intensity. The influence of natural factors on FVC showed a decreasing trend, whereas human factors exhibited an increasing trend. Natural factors promoted vegetation growth in most area of the YRB, especially in the northwest, whereas human factors had more negative effects, except for the middle parts of the basin. The spatial differentiation analyses of FVC and driving factors provide scientific guidance for ecological restoration and sustainable development in the YRB.

show abstract

“…Although few studies on the identification of forage grass have been carried out, research on the classification of land types and crops is mature. Hu et al [54] had an OA of 89% for crop classification in the Hetao Irrigation District in 2020 based on Sentinel-1/2 and an RF classifier with 300 trees. Kushal, KC et al [55] showed that the OA of the four types of cover crops in the Maumee River was 75%, and the Kappa coefficient was 0.63 based on Landsat and an RF classifier for 2008-2019.…”

Section: Classification Accuracy Evaluationmentioning

confidence: 99%

Identification and Area Information Extraction of Oat Pasture Based on GEE—A Case Study in the Shandan Racecourse (China)

et al. 2022

View full text Add to dashboard Cite

Forage grass is very important for food security. The development of artificial grassland is the key to solving the shortage of forage grass. Understanding the spatial distribution of forage grass in alpine regions is of great importance for guiding animal husbandry and the rational selection of forage grass management measures. With its powerful computing power and complete image data storage, Google Earth Engine (GEE) has become a new method to address remote sensing data collection difficulties and low processing efficiency. High-resolution mapping of pasture distributions on the Tibetan Plateau (China) is still a difficult problem due to cloud disturbance and mixed planting of forage grass. Based on the GEE platform, Sentinel-2 data and three classifiers, this study successfully mapped the oat pasture area of the Shandan Racecourse (China) on the eastern Tibetan Plateau over 3 years from 2019 to 2021 at a resolution of 10 m based on cultivated land identification. In this study, the key phenology windows were determined by analysing the time series differences in vegetation indices between oat pasture and other forage grasses in the Shandan Racecourse, and monthly scale features were selected as features for oat pasture identification. The results show that the mean Overall Accuracy (OA) of Random Forest (RF) classifier, Support Vector Machine (SVM) classifier, and Classification and Regression Trees (CART) classifier are 0.80, 0.69, and 0.72 in cultivated land identification, respectively, with corresponding the Kappa coefficients of 0.74, 0.58, and 0.62. The RF classifier far outperforms the other two classifiers. In oat pasture identification, the RF, SVM and CART classifiers have high OAs of 0.98, 0.97, and 0.97 and high Kappa values of 0.95, 0.94, and 0.95, respectively. Overall, the RF classifier is more suitable for our research. The oat pasture areas in 2019, 2020 and 2021 were 347.77 km2 (15.87%), 306.19 km2 (13.97%) and 318.94 km2 (14.55%), respectively, with little change (1.9%) from year to year. The purpose of this study was to explore the identification model of forage grass area in alpine regions with a high spatial resolution, and to provide technical and methodological support for information extraction of the forage grass distribution status on the Tibetan Plateau.

show abstract

An Interannual Transfer Learning Approach for Crop Classification in the Hetao Irrigation District, China

Cited by 31 publications

References 76 publications

Sustainable Irrigation Requirement Prediction Using Internet of Things and Transfer Learning

Sustainable Irrigation Requirement Prediction Using Internet of Things and Transfer Learning

Spatiotemporal evolution of growing‐season vegetation coverage and its natural‐human drivers in the Yellow River Basin, China

Identification and Area Information Extraction of Oat Pasture Based on GEE—A Case Study in the Shandan Racecourse (China)

Contact Info

Product

Resources

About