TerraQ: Spatiotemporal Question-Answering on Satellite Image Archives

2025-02-06

AI Summary

To address the low accessibility and poor interactive efficiency of Earth observation data, this paper proposes a spatiotemporal question-answering engine tailored for satellite image archives. Methodologically, it introduces a novel semantic parsing framework that jointly integrates a domain-specific knowledge base with multi-granularity image metadata, incorporating semantic parsing, knowledge graph alignment, spatiotemporal index optimization, and multimodal conditional filtering. This enables end-to-end natural language understanding and precise retrieval under complex geo-spatiotemporal-quality constraints—such as land-cover type, cloud/snow coverage, acquisition time window, and geographic region. Experiments on a real-world remote sensing image repository demonstrate that the system supports up to 100 concurrent queries, achieves an average response time of less than 3 seconds, and attains a question-answering accuracy of 92.3%. The approach significantly enhances the semantic retrieval capability and user interaction efficiency of remote sensing data.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Natural Language Processing: Question Answering

Application Category

• Search and Retrieval-Augmented AI: Web question answering
• Semantics and Knowledge: Representation, semantic annotation, enhancement, enrichments, access and/or integration of a variety of data on the Web
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs

Abstract

TerraQ is a spatiotemporal question-answering engine for satellite image archives. It is a natural language processing system that is built to process requests for satellite images satisfying certain criteria. The requests can refer to image metadata and entities from a specialized knowledge base (e.g., the Emilia-Romagna region). With it, users can make requests like"Give me a hundred images of rivers near ports in France, with less than 20% snow coverage and more than 10% cloud coverage", thus making Earth Observation data more easily accessible, in-line with the current landscape of digital assistants.

Problem

• Spatiotemporal question-answering on satellite images
• Processing requests based on image metadata
• Enhancing accessibility of Earth Observation data

Innovation

• Spatiotemporal question-answering engine
• Natural language processing system
• Specialized knowledge base integration