Research on semantic segmentation algorithm of high latitude urban river ice based on deep transfer learning DOI
Wangyuan Zhao, Yanzhuo Xue,

Fenglei Han

et al.

International Journal of Remote Sensing, Journal Year: 2024, Volume and Issue: 45(13), P. 4278 - 4299

Published: June 14, 2024

Automated observation methods for monitoring river ice in high-latitude urban areas are crucial resource utilization, risk assessment, and navigation. However, current research lacks actual-scale classification, such as low-altitude surveys. This study established a dataset of the Songhua River near Harbin, Northeast China, using UAV aerial photography applied RININet semantic segmentation algorithm precise classification different types remote sensing images. To address environmental challenges, feature extraction method integrating channel spatial attention mechanisms was adopted, along with an improved pyramid pool structure to enhance recognition. Additionally, two-stage transfer learning recognition database, overcoming issues like small data volume high annotation costs. Comparative evaluation metrics demonstrated accuracy framework. Furthermore, estimating blockage proposed, applicable various management tasks, practical significance.

Language: Английский

Regional Flood Risk Assessment and Prediction Based on Environmental Attributes and Pipe Operational Characteristics DOI Open Access
Jinping Zhang,

Yirong Yang,

Lixin Zhang

et al.

Water, Journal Year: 2025, Volume and Issue: 17(10), P. 1477 - 1477

Published: May 14, 2025

Urban flood risk assessments play a crucial role in urban resilience and disaster management. This paper proposes comprehensive method for assessment prediction that is based on environmental attributes the operational characteristics of pipe networks. Using central area Zhengzhou as case study, an integrated evaluation index system was developed, entropy weight applied to quantify indicators. A loosely coupled RF-XGBoost model constructed predict different rainfall scenarios. The results indicate (1) overall study exhibits increasing trend from northeast southwest, with medium- high-risk zones being predominant; (2) spatial distribution pattern closely aligns but shows slight variations under influence network risks; (3) demonstrates superior predictive accuracy multi-factor coupling When characteristics, attributes, risks are comprehensively considered, Nash–Sutcliffe Efficiency (NSE) predictions improves 0.85 (when using only characteristics) 0.94. provides valuable insights technical support mitigating risks.

Language: Английский

Citations

0
Forecasting the River Ice Break-Up Date in the Upper Reaches of the Heilongjiang River Based on Machine Learning DOI Open Access
Zhi Liu, Hongwei Han, Yu Li

et al.

Water, Journal Year: 2025, Volume and Issue: 17(3), P. 434 - 434

Published: Feb. 4, 2025

Ice-jam floods (IJFs) are a significant hydrological phenomenon in the upper reaches of Heilongjiang River, posing substantial threats to public safety and property. This study employed various feature selection techniques, including Pearson correlation coefficient (PCC), Grey Relational Analysis (GRA), mutual information (MI), stepwise regression (SR), identify key predictors river ice break-up dates. Based on this, we constructed machine learning models, Extreme Gradient Boosting (XGBoost), Backpropagation Neural Network (BPNN), Random Forest (RF), Support Vector Regression (SVR). The results indicate that reserves Oupu Heihe section have most impact date section. Additionally, accumulated temperature during period average before identified as features closely related river’s opening all four methods. choice method notably impacts performance models predicting Among tested, XGBoost with PCC-based achieved highest accuracy (RMSE = 2.074, MAE 1.571, R2 0.784, NSE 0.756, TSS 0.950). provides more accurate effective for dates, offering scientific basis preventing managing IJF disasters.

Language: Английский

Citations

0
The Analysis of Present and Future Use of Non-Conventional Water Resources in Heilongjiang Province, China DOI Open Access
Hongcong Guo,

Yingna Sun,

Tienan Li

et al.

Sustainability, Journal Year: 2024, Volume and Issue: 16(9), P. 3727 - 3727

Published: April 29, 2024

Analyzing the development trend of non-conventional water resources and identifying main influencing factors is initial step toward rapidly increasing utilization allocation these in a rational scientific manner. This will help relieve pressure on improve ecological environment. study introduces concept comparison testing employs advanced Dematel Random Forest models to identify two sets optimal indicators from pool nine. Based best indicator sets, three prediction models—BP neural network, Particle Swarm Optimization-optimized BP Genetic network—were used forecast future potential resource use Heilongjiang Province. The findings reveal that economic are most significant Province’s resources. this us understand extent utilizing

Language: Английский

Citations

1
Exploring the Influence of Stakeholders' Opinions on the Selection and Weighting of Social Vulnerability Variables in Flood Risk Management DOI
Md. Munjurul Haque, Wanyun Shao, Hemal Dey

et al.

Published: Jan. 1, 2024

Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI

Language: Английский

Citations

0
Research on semantic segmentation algorithm of high latitude urban river ice based on deep transfer learning DOI
Wangyuan Zhao, Yanzhuo Xue,

Fenglei Han

et al.

International Journal of Remote Sensing, Journal Year: 2024, Volume and Issue: 45(13), P. 4278 - 4299

Published: June 14, 2024

Automated observation methods for monitoring river ice in high-latitude urban areas are crucial resource utilization, risk assessment, and navigation. However, current research lacks actual-scale classification, such as low-altitude surveys. This study established a dataset of the Songhua River near Harbin, Northeast China, using UAV aerial photography applied RININet semantic segmentation algorithm precise classification different types remote sensing images. To address environmental challenges, feature extraction method integrating channel spatial attention mechanisms was adopted, along with an improved pyramid pool structure to enhance recognition. Additionally, two-stage transfer learning recognition database, overcoming issues like small data volume high annotation costs. Comparative evaluation metrics demonstrated accuracy framework. Furthermore, estimating blockage proposed, applicable various management tasks, practical significance.

Language: Английский

Citations

0