Intelligent Semantic Segmentation in 3D Reconstruction Using Enhanced Reverse Modeling

Three-dimensional (3D) reconstruction is a fundamental process in computer vision and geographic information systems, yet its practical applications are often limited by challenges in scene comprehension. This study introduces a novel method that integrates reverse modeling with artificial intelligence algorithms to enhance semantic segmentation in 3D reconstruction. We leverage 3D Gaussian Splatting (3DGS) to convert 2D video imagery of real-world scenes into comprehensive 3D reconstructions and employ a customized PointNet model for probabilistic prediction on point clouds segmented by CloudCompare. By incorporating these probabilistic predictions into the segmented point clouds, our approach not only improves the precision of 3D reconstructions but also significantly enhances object recognition and semantic interpretation. The re-rendered 3D models accurately depict segmented regions, providing a new perspective and robust data support for advancing 3D reconstruction research and applications.