Tourism Security Situation Awareness System using Multi-Source Remote Sensing and Adaptive Positioning Algorithms

This study proposes an efficient tourism safety situational awareness system based on multi-source remote sensing and an adaptive positioning algorithm. This system can improve the accuracy and timeliness of risk monitoring and emergency response in complex tourism environments. The designed system integrates data from optical remote sensing, synthetic aperture radar (SAR), laser ranging (LiDAR), and ground-based sensors at the data acquisition level. Theoretically, a three-stage imaging modeling approach (radiance-optical transfer-pixel integration) achieves spatiotemporally consistent CCD image fusion. Algorithmically, this study introduces sub-pixel disparity matching and a weighted boundary constraint optimization model. This improves target detection and localization accuracy. For situation assessment, this study uses the information entropy weighting method and the TOPSIS model to calculate a comprehensive safety index. This allows for multi-metric risk quantification. Experimental data was collected from typical tourist attractions in Hainan Province, covering three types of scenes, totaling 18,432 frames of time-series remote sensing images and 2,304 manually annotated risk samples. Comparative experimental results demonstrate that the proposed method achieves an F1 score of 0.9048 for dangerous target detection, a 6.57% improvement over the best comparable method. The positioning root mean square error (RMSE) was 5.1 $\mathbf{m}$, a $\mathbf{2 5. 0 \%}$ improvement over the suboptimal baseline. Frame latency was kept to 240 ms while maintaining real-time performance, a 22.58% improvement over the Optical+SAR method. The calculated comprehensive safety index outperformed the comparison methods in both reliability and stability.