A Deep-Learning Framework for Land-Sliding Classification from Remote Sensing Image

2025-07-17

AI Summary

To address class imbalance and model overfitting in landslide detection from remote sensing imagery, this paper proposes a synergistic framework integrating online/offline data augmentation, EfficientNet-Large as the backbone for feature extraction, and SVM-based post-processing classification. Its key contributions are: (1) a staged data augmentation strategy to mitigate representational bias in the minority landslide class; (2) leveraging EfficientNet-Large’s strong robustness to extract discriminative multi-scale features; and (3) replacing the conventional fully connected layer with an SVM classifier to enhance decision-boundary accuracy and generalization. Evaluated on the Zindi public test set, the method achieves an F1 score of 0.8938—significantly outperforming baseline approaches—and demonstrates both effectiveness and practicality for disaster identification tasks.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Computer Vision: Object Detection & Categorization
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

The use of satellite imagery combined with deep learning to support automatic landslide detection is becoming increasingly widespread. However, selecting an appropriate deep learning architecture to optimize performance while avoiding overfitting remains a critical challenge. To address these issues, we propose a deep-learning based framework for landslide detection from remote sensing image in this paper. The proposed framework presents an effective combination of the online an offline data augmentation to tackle the imbalanced data, a backbone EfficientNet_Large deep learning model for extracting robust embedding features, and a post-processing SVM classifier to balance and enhance the classification performance. The proposed model achieved an F1-score of 0.8938 on the public test set of the Zindi challenge.

Problem

• Optimizing deep learning for landslide detection from satellite images
• Addressing data imbalance with online-offline augmentation techniques
• Enhancing classification using EfficientNet and SVM post-processing

Innovation

• Combines online and offline data augmentation
• Uses EfficientNet_Large for feature extraction
• Integrates SVM for enhanced classification