在高解析度地理影像的三維可視化應用中，傳統的投影方法，如麥卡托投影，面臨畫面延遲、幀率降低等問題，嚴重限制高精度三維地球模型的生成和應用。為了克服這些困難，本研究利用旋轉矩陣和插值計算方法，將二維影像轉換為六面立方體貼圖，再利用旋轉變換投影方法實現六面立方體貼圖到三維球體模型的高效投影，有效解決傳統方法在投影上的困難，並且更進一步展現高品質的三維地球。研究中，首先通過地理影像定位技術，將來自美國太空總署的高解析度地圖圖資與地球表面的真實位置進行精確對位。然後，利用旋轉變換投影、麥卡托投影、四邊形球面立方體投影和三維點雲等方法，將二維地圖分別投影到三維球體模型。在效率評估實驗中，旋轉變換投影展現出最優的性能表現，在高解析度(5400*2700)下的投影效率相比麥卡托投影提升了23.14%。此外，通過結構相似性和峰值訊噪比的量化評估，首先分析旋轉變換投影在模型精度上的可信度，再利用實際距離測量驗證旋轉變換投影的真實性。本研究提出的方法不僅為高解析度地理影像的三維可視化提供了高效、高精度與高分辨率的解決方案，同時也為地理資訊系統、遙測製圖等領域的技術創新和應用方法提供了新的概念。;In the three-dimensional visualization of high-resolution geographic imagery, traditional projection methods such as Mercator projection face issues, including frame delays and frame rate reduction, which severely limit the creation and application of high-precision 3D earth models. To overcome these challenges, this study utilizes rotation matrices and interpolation methods to convert 2D images into cube maps. Subsequently, using a rotation transformation projection method, these maps are efficiently projected onto a 3D earth model, effectively solving the projection difficulties associated with traditional methods and further presenting high-quality 3D models.Initially, the research accurately aligns high-resolution map data from NASA with the real-world positions on the Earth′s surface using geographic image registration techniques. Subsequently, the study employs rotational transformation projection, Mercator projection, and 3D point cloud methods to project the 2D maps onto 3D spherical models. Efficiency assessment experiments demonstrate that the rotational transformation projection method exhibits superior performance, enhancing projection efficiency by 23.14% compared to Mercator projection at a high resolution of 5400*2700. Furthermore, through quantitative evaluations using structural similarity and peak signal-to-noise ratio, the reliability of the rotation transformation projection in terms of model accuracy was analyzed, followed by the verification of the projection’s authenticity using actual distance measurements. The proposed method not only provides an efficient,high-precision, and high-resolution solution for the 3D visualization of high-resolution geographic imagery but also offers new concepts for technological innovation and application methods in geographic information systems and remote sensing cartography.
