Multi-View 3D Reconstruction of Vessel Hull with Spatial Density Weighted NeRF

A bstract- The 3D reconstruction of vessel hulls is crucial for enhancing safety, efficiency, and knowledge in the maritime industry. Neural Radiance Fields (NeRFs) is an alternative for 3D reconstruction and rendering from multi-view images, particularly tensor-based methods have proven effective in improving efficiency. However, existing tensor-based methods typically suffer from a lack of spatial coherence, resulting in gaps in the reconstruction of fine-grained geometric structures. This paper proposes a spatial density weighted NeRF (DW-N eRF) for accurate and efficient surface reconstruction of vessel hulls. The proposed method develops a weighted feature space decomposition mechanism, which utilizes tensorial radiance fields and weighted algorithms to extract feature vectors of sampled points. To enhance training efficiency, a coarse-fine strategy is employed for the density-weighted algorithm during the training process. Extensive experiments conducted on a synthetic vessel hull (SVH) dataset demonstrate that the effectiveness of our method in achieving remarkable rendering quality, high reconstruction accuracy and timeliness.