Kumar et al. (2026) Cloudburst Detection from Satellite Images Using Haralick Features and Random Forest Classifier

Identification

• Journal: Lecture notes in networks and systems
• Year: 2026
• Date: 2026-01-01
• Authors: Shivam Kumar, Atulya Narayan, Anant Ram, Arun Mourya, Chinmoy Kar, Radha Tamal Goswami
• DOI: 10.1007/978-3-032-13544-5_5

Research Groups

• Techno International New Town, Kolkata, India
• Adamas University, Kolkata, West Bengal, India

Short Summary

This study develops a machine learning-based approach to detect cloudbursts from satellite imagery using Haralick texture features. It demonstrates that the Random Forest classifier outperforms other tree-based methods for this classification task, offering a potential for rapid early warning systems.

Objective

• To develop and evaluate a machine learning-based approach for identifying cloudburst events from satellite imagery using texture-based features to enable fast and efficient weather prediction for early warnings.

Study Configuration

• Spatial Scale: India, with a focus on mountainous and hilly areas prone to cloudbursts.
• Temporal Scale: Satellite images collected from cloudburst events and non-cloudburst clouds that occurred over India in recent years.

Methodology and Data

• Models used: Haralick features for texture-based image feature extraction; various tree-based machine learning classifiers, with a specific focus on the Random Forest classifier.
• Data sources: Satellite images of cloudburst events and non-cloudburst cloud formations over India.

Main Results

• The Random Forest classifier was observed to perform better than other tree-based classifiers in accurately predicting cloudburst events from satellite imagery.
• The proposed machine learning approach, leveraging Haralick features, effectively classifies cloudburst events.

Contributions

• Introduces a novel machine learning-based system for rapid and efficient cloudburst detection using texture features from satellite images.
• Provides a method for early warning of cloudbursts, which are significant natural hazards, particularly in mountainous and hilly regions.
• Demonstrates the effectiveness of Haralick features combined with Random Forest for cloudburst classification, contributing to disaster management and weather prediction systems.

Funding

• Not explicitly mentioned in the provided paper text.

Citation

@article{Kumar2026Cloudburst,
  author = {Kumar, Shivam and Narayan, Atulya and Ram, Anant and Mourya, Arun and Kar, Chinmoy and Goswami, Radha Tamal},
  title = {Cloudburst Detection from Satellite Images Using Haralick Features and Random Forest Classifier},
  journal = {Lecture notes in networks and systems},
  year = {2026},
  doi = {10.1007/978-3-032-13544-5_5},
  url = {https://doi.org/10.1007/978-3-032-13544-5_5}
}