Exploration Into the Development and Application of Image Localization: System of Systems Engineering Design Perspective

This paper introduces a practical all-weather multiscale image localization system that surpasses traditional radiobased methods. By integrating satellite, aerial, and street-view data with optimized multi-modal algorithms, it enhances localization precision and stability. Using a “data-model-service” architecture, the system creates a unified 3D data fusion system covering sky, ground, and air. Systematic analysis and iterative validation ensure all-weather operation and full-space coverage. Relying on a multi-modal intelligent processing engine, it achieves deep data fusion and efficient processing. The system provides allweather localization services for the ubiquitous PNT system, featuring a four-tier architecture with multi-modal fusion and a dynamic benchmark engine. The algorithm engine handles multisource info, the data storage layer manages data efficiently, the backend offers RESTful APIs and high-performance computing, and the frontend has a web-based visual interface. Its application in Yimeng Mountain tourist area shows value in scenic navigation, offering intelligent decision support with great promotion potential.