针对目前在混合现实(MR)环境中高效率建立高质量三维(3D)模型的需求,基于神经辐射场算法(NeRF)的三维重建技术,提出了一种基于Laplacian算子的数据集优化算法。首先,围绕某线切割设备录制了一段1 min 51 s的视频,并采取等距提取视频帧...针对目前在混合现实(MR)环境中高效率建立高质量三维(3D)模型的需求,基于神经辐射场算法(NeRF)的三维重建技术,提出了一种基于Laplacian算子的数据集优化算法。首先,围绕某线切割设备录制了一段1 min 51 s的视频,并采取等距提取视频帧的方式,获取了训练数据集;然后,使用Laplacian算子对数据集进行了优化,同时保留了原始数据集作为对比,使用了基于NeRF算法的重建方式与传统的基于COLMAP的稠密点云重建方式,分别对两组数据集进行了三维重建;最后,在重建精度与重建速度方面,对不同重建方式、不同重建数据集的重建结果进行了比较。研究结果表明:COLMAP稠密点云重建耗时是基于NeRF重建耗时的9.98倍,而相较于COLMAP稠密点云重建,使用NeRF重建方式的模型表面缺陷较少;此外,使用Laplacian算子优化的数据集的NeRF重建在峰值信噪比(PSNR)和结构相似性(SSIM)指标上分别提升了2.43%、0.72%,有利于提升重建模型的质量。研究结果支持混合现实技术在制造业数字化转型中的应用,可为其提供有益的参考。展开更多
The resistance loss of transportation was studied and the influences of buoyancy layout,mineral content and elastic modulus of flexible hose were investigated based on three-dimensional finite element model of fluid-s...The resistance loss of transportation was studied and the influences of buoyancy layout,mineral content and elastic modulus of flexible hose were investigated based on three-dimensional finite element model of fluid-solid interaction by MSC.MARC/MENTAT software.The numerical results show that the resistance losses increase with the increase of mineral content Cv and velocity of internal fluid v and decrease with the increase of elastic modulus E of flexible hose.The buoyancy layout and the velocity of internal fluid have greater impacts on the resistance losses than the elastic modulus of flexible hose.In order to reduce the resistance losses and improve the efficiency of the deep-ocean mining,Cv and v must be restricted in a suitable range (e.g.10%-25% and 2.5-4 m/s).Effective buoyancy layout (such as Scheme C and D) should be adopted and the suitable material of moderate E should be used for the flexible hose in deep-ocean mining.展开更多
文摘针对目前在混合现实(MR)环境中高效率建立高质量三维(3D)模型的需求,基于神经辐射场算法(NeRF)的三维重建技术,提出了一种基于Laplacian算子的数据集优化算法。首先,围绕某线切割设备录制了一段1 min 51 s的视频,并采取等距提取视频帧的方式,获取了训练数据集;然后,使用Laplacian算子对数据集进行了优化,同时保留了原始数据集作为对比,使用了基于NeRF算法的重建方式与传统的基于COLMAP的稠密点云重建方式,分别对两组数据集进行了三维重建;最后,在重建精度与重建速度方面,对不同重建方式、不同重建数据集的重建结果进行了比较。研究结果表明:COLMAP稠密点云重建耗时是基于NeRF重建耗时的9.98倍,而相较于COLMAP稠密点云重建,使用NeRF重建方式的模型表面缺陷较少;此外,使用Laplacian算子优化的数据集的NeRF重建在峰值信噪比(PSNR)和结构相似性(SSIM)指标上分别提升了2.43%、0.72%,有利于提升重建模型的质量。研究结果支持混合现实技术在制造业数字化转型中的应用,可为其提供有益的参考。
基金Project(2006AA09Z240)supported by the National High Technology Research and Development Program of China
文摘The resistance loss of transportation was studied and the influences of buoyancy layout,mineral content and elastic modulus of flexible hose were investigated based on three-dimensional finite element model of fluid-solid interaction by MSC.MARC/MENTAT software.The numerical results show that the resistance losses increase with the increase of mineral content Cv and velocity of internal fluid v and decrease with the increase of elastic modulus E of flexible hose.The buoyancy layout and the velocity of internal fluid have greater impacts on the resistance losses than the elastic modulus of flexible hose.In order to reduce the resistance losses and improve the efficiency of the deep-ocean mining,Cv and v must be restricted in a suitable range (e.g.10%-25% and 2.5-4 m/s).Effective buoyancy layout (such as Scheme C and D) should be adopted and the suitable material of moderate E should be used for the flexible hose in deep-ocean mining.
