Under the Background of Intelligent Navigation and Positioning Technology: 3D Environmental Modeling for Low-Altitude UAVs

In low-altitude scenarios, intelligent navigation and positioning technology is the core support for UAV swarm cooperative control and path planning. However, in complex low-altitude environments, it often faces issues like positioning deviations and insufficient path planning safety due to the lack of accurate 3D environmental data, urgently requiring resolution via high-precision 3D environmental modeling. To this end, this study proposes an improved multi-stage interpolation technique to enhance the accuracy of UAV 3D environmental modeling under limited data points. The method integrates an enhanced cubic spline interpolation algorithm and Kriging interpolation. Specifically, the enhanced cubic spline interpolation algorithm introduces a node determination function to optimize spline coefficients, enabling more accurate capture of terrain features. The Kriging interpolation, on the other hand, performs refined processing based on high-resolution UAV perception data collected in designated airspace. For the multi-stage process, an initial environmental surface is first generated via the enhanced cubic spline interpolation, and then refined using Kriging interpolation. Experimental results show the improved cubic spline interpolation achieves a root mean square error (RMSE) of 15.43, 56.81% lower than the original cubic spline, and 29.11% and 32.54% lower than Gaussian process and LOESS, respectively. This two-stage method performs well under sparse data conditions, generates high-precision UAV 3D environmental models, is practically applicable in low-altitude navigation, emergency mapping, and scene perception, and outperforms traditional smoothing and regression techniques.