Three-Dimensional Reconstruction System Based on Point Cloud and NeRF
3D reconstruction can help people better view the details of the reconstructed object and is widely used in object display,virtual scene simulation,cultural relic protection and other scenes.Traditional multi view 3D reconstruction is prone to a large amount of noise,voids,and other issues,and the display effect is poor.The emerging NeRF algorithm that utilizes neural networks for 3D reconstruction can achieve near real results,but the rendering speed is slow and cannot meet the requirements of interaction.The article combines two methods to obtain a 3D reconstruction system that balances interaction and rendering effects.The results show that the system in this article achieved good interaction and rendering effects on the sparse image dataset collected..
NETL
NSTL
万方数据
