Satellite image classification with neural quantum kernels

2024-09-30

Machine Learning: Science and Technology

AI Summary

This work addresses the practical remote sensing classification challenge of solar panel detection in satellite imagery for Earth observation. We propose an end-to-end quantum machine learning framework integrating classical preprocessing with a neural quantum kernel (NQK). To our knowledge, this is the first application of an NQK—derived from a trained quantum neural network (QNN)—to remote sensing image classification. Evaluated on real-world satellite data, the approach demonstrates robustness and scalability up to 8 qubits. By synergistically combining classical dimensionality reduction with quantum kernel methods, the model achieves accuracy and generalization performance comparable to state-of-the-art classical models, while maintaining stable performance as qubit count increases. This study provides critical empirical validation and a methodological blueprint for deploying quantum machine learning in large-scale Earth observation tasks.

Technology Category

• Machine Learning: Quantum Machine Learning
• Computer Vision: Remote Sensing / Geospatial AI
• Humans and AI: Human-in-the-loop Machine Learning

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Security and privacy of machine learning and AI applications

Abstract

Achieving practical applications of quantum machine learning for real-world scenarios remains challenging despite significant theoretical progress. This paper proposes a novel approach for classifying satellite images, a task of particular relevance to the earth observation (EO) industry, using quantum machine learning techniques. Specifically, we focus on classifying images that contain solar panels, addressing a complex real-world classification problem. Our approach begins with classical pre-processing to reduce the dimensionality of the satellite image dataset. We then apply neural quantum kernels (NQKs)—quantum kernels derived from trained quantum neural networks (QNNs)—for classification. We evaluate several strategies within this framework, demonstrating results that are competitive with the best classical methods. Key findings include the robustness of our results and their scalability, with successful performance achieved up to $8$ qubits.

Problem

• Classifying satellite images using quantum machine learning
• Focusing on solar panel detection in complex scenarios
• Evaluating neural quantum kernels for scalable classification

Innovation

• Classical pre-processing reduces satellite image dimensionality.
• Neural quantum kernels classify images using quantum neural networks.
• Scalable performance demonstrated up to 8 qubits.