Rishikeshan et al. (2026) An Approach for Cyclone Tracking and Monitoring

Identification

• Journal: Lecture notes in networks and systems
• Year: 2026
• Date: 2026-01-01
• Authors: C. A. Rishikeshan, R. Jayanthi, Snehasis Ghosh, Navoneel Mondal, Srijanbroto Deb
• DOI: 10.1007/978-981-95-0701-6_9

Research Groups

• Vellore Institute of Technology (VIT), Chennai, India

Short Summary

This study presents a custom convolutional neural network (CNN) architecture for detecting and classifying cyclones from remote sensing images, achieving 92.5% accuracy in distinguishing cyclone activity from normal weather patterns.

Objective

• To develop and evaluate a deep learning approach using a custom convolutional neural network (CNN) to accurately detect and classify cyclones from remote sensing images, thereby enhancing storm monitoring and preparedness efforts.

Study Configuration

• Spatial Scale: Global (applicable to "vulnerable coastal regions" and "different regions" where remote sensing images are available).
• Temporal Scale: Not explicitly stated for the data used, but the objective of "monitoring and forecasting" implies an ongoing or near real-time application.

Methodology and Data

• Models used: Custom designed Convolutional Neural Network (CNN) architecture, comprising convolutional layers for feature extraction and a fully connected layer for final classification.
• Data sources: Remote sensing images.

Main Results

• The proposed CNN model effectively captures distinct features differentiating cyclone activity from normal weather conditions.
• The model demonstrated an improved detection accuracy of 92.5% when compared to traditional approaches.
• Performance evaluation was conducted through accuracy measurements and confusion matrix analysis.
• The model shows strong potential for improving storm detection across various regions, despite challenges such as limited labelled storm data.

Contributions

• Introduces a novel custom CNN architecture designed to retrieve meaningful contextual information (high-level features) from remote sensing images for cyclone detection.
• Provides an improved accuracy (92.5%) in distinguishing cyclone activity from normal weather patterns compared to existing traditional methods.
• Advances the application of deep learning in disaster risk management, offering a valuable tool for enhanced storm monitoring and preparedness.

Funding

• Not explicitly stated in the provided text.

Citation

@article{Rishikeshan2026Approach,
  author = {Rishikeshan, C. A. and Jayanthi, R. and Ghosh, Snehasis and Mondal, Navoneel and Deb, Srijanbroto},
  title = {An Approach for Cyclone Tracking and Monitoring},
  journal = {Lecture notes in networks and systems},
  year = {2026},
  doi = {10.1007/978-981-95-0701-6_9},
  url = {https://doi.org/10.1007/978-981-95-0701-6_9}
}