Digital twin(DT)modelling is a prerequisite for the successful application of DT technology in the power industry.However,traditional scene modelling methods are costly,time-consuming,focus on overall features and lac...Digital twin(DT)modelling is a prerequisite for the successful application of DT technology in the power industry.However,traditional scene modelling methods are costly,time-consuming,focus on overall features and lack real-time updates,hindering the interaction between DT models and physical power equipment scenes.Therefore,a scene DT modelling technique focusing on local features in risk areas and real-time updates is urgently needed.Herein,real-time modelling of the±800 kV converter transformer is achieved by improving the neural radiation field based on a hybrid attention mechanism and multiresolution hash encoding.Compared to traditional methods,modelling time is reduced from hours to 1 min without professional equipment or manual intervention.The model quality is more concerned with local features of risk areas in transformers while ensuring the overall scene,and the accuracy is improved by about 6%,realising the real-time modelling of transformers and the DT of scenes.展开更多
Recently, neural implicit function-basedrepresentation has attracted more and more attention,and has been widely used to represent surfacesusing differentiable neural networks. However, surfacereconstruction from poin...Recently, neural implicit function-basedrepresentation has attracted more and more attention,and has been widely used to represent surfacesusing differentiable neural networks. However, surfacereconstruction from point clouds or multi-view imagesusing existing neural geometry representations stillsuffer from slow computation and poor accuracy. Toalleviate these issues, we propose a multi-scale hashencoding-based neural geometry representation whicheffectively and efficiently represents the surface asa signed distance field. Our novel neural networkstructure carefully combines low-frequency Fourierposition encoding with multi-scale hash encoding. Theinitialization of the geometry network and geometryfeatures of the rendering module are accordinglyredesigned. Our experiments demonstrate that theproposed representation is at least 10 times faster forreconstructing point clouds with millions of points.It also significantly improves speed and accuracyof multi-view reconstruction. Our code and modelsare available at https://github.com/Dengzhi-USTC/Neural-Geometry-Reconstruction.展开更多
基金National Key Research and Development Program of China,Grant/Award Number:2021YFB2401700。
文摘Digital twin(DT)modelling is a prerequisite for the successful application of DT technology in the power industry.However,traditional scene modelling methods are costly,time-consuming,focus on overall features and lack real-time updates,hindering the interaction between DT models and physical power equipment scenes.Therefore,a scene DT modelling technique focusing on local features in risk areas and real-time updates is urgently needed.Herein,real-time modelling of the±800 kV converter transformer is achieved by improving the neural radiation field based on a hybrid attention mechanism and multiresolution hash encoding.Compared to traditional methods,modelling time is reduced from hours to 1 min without professional equipment or manual intervention.The model quality is more concerned with local features of risk areas in transformers while ensuring the overall scene,and the accuracy is improved by about 6%,realising the real-time modelling of transformers and the DT of scenes.
基金supported by the National Natural Science Foundation of China(Nos.62122071 and 62272433)the Fundamental Research Funds for the Central Universities(No.WK3470000021)the Alibaba Innovation Research Program(AIR).
文摘Recently, neural implicit function-basedrepresentation has attracted more and more attention,and has been widely used to represent surfacesusing differentiable neural networks. However, surfacereconstruction from point clouds or multi-view imagesusing existing neural geometry representations stillsuffer from slow computation and poor accuracy. Toalleviate these issues, we propose a multi-scale hashencoding-based neural geometry representation whicheffectively and efficiently represents the surface asa signed distance field. Our novel neural networkstructure carefully combines low-frequency Fourierposition encoding with multi-scale hash encoding. Theinitialization of the geometry network and geometryfeatures of the rendering module are accordinglyredesigned. Our experiments demonstrate that theproposed representation is at least 10 times faster forreconstructing point clouds with millions of points.It also significantly improves speed and accuracyof multi-view reconstruction. Our code and modelsare available at https://github.com/Dengzhi-USTC/Neural-Geometry-Reconstruction.
