【摘要】 针对数字孪生点云数据不适合直接浏览，并且难以直接用于三维网格模型重建的问题，提出了一种将场景稠密点云转换到多组全景照片和全景深度图，并合并到现有三维重建流程的方法。算法可以将现有激光扫描设备输出的点云数据采样和处理为连续的RGB颜色和深度图像数据，并与照片的重建流程、及连续深度图的重建流程相结合。经过实测表明，方法对重建效果有明显的提升，并且对具体的三维重建流程和算法没有限制，可以广泛应用到各种三维重建相关的具体应用中。更多还原

【Abstract】 This paper proposes a method to convert dense point clouds which are difficult to be used directly, to multiple sets of panoramic RGB and depth maps, and merge them into existing 3 D reconstruction algorithms. Our method processes point cloud data output from existing laser scanning devices into continuous RGB color and depth image data, and combine it with the photo-based reconstruction algorithms, as well as depth map-based reconstruction algorithms like the TSDF. Our method is proved to significantly improve the reconstruction result and has no restrictions on specific reconstruction algorithms, which can be widely used in related applications.更多还原