GeoIRIS: Geospatial Information Retrieval and Indexing System-Content Mining, Semantics Modeling, and Complex Queries

Area of Science:

• Geoinformatics
• Computer Science
• Remote Sensing

Background:

• Geospatial data retrieval from heterogeneous sources is challenging due to semantic complexity and the need for efficient analysis.
• Current methods often require significant human intervention for data collection and analysis.

Purpose of the Study:

• To present the Geospatial Information Retrieval and Indexing System (GeoIRIS) for content-based multimodal geospatial information retrieval.
• To enable faster identification of relevant imagery for image analysts through an integrated system.

Main Methods:

• Developed automatic feature extraction and visual content mining from large-scale image databases.
• Implemented high-dimensional database indexing for rapid retrieval.
• Created techniques for complex queries merging heterogeneous geospatial databases, object retrieval by shape/visuals, and multiobject spatial relationship analysis.

Main Results:

• GeoIRIS integrates diverse technologies for coherent geospatial information retrieval.
• The system supports complex queries, including semantic linking of image features to high-level descriptors.
• GeoIRIS can answer complex analytical questions in seconds, such as finding similar objects and spatial relationships near a landmark.

Conclusions:

• GeoIRIS significantly enhances the speed and efficiency of searching and analyzing heterogeneous geospatial databases.
• The system empowers image analysts to rapidly identify relevant imagery and extract complex spatial-temporal information.
• GeoIRIS represents a significant advancement in multimodal geospatial information retrieval systems.