Research on construction method of digital twin scene of coastal channel
A construction method of digital twin scenes of coastal channel was proposed aiming at the difficulties of insufficient ability of channel element collection,data fusion organization and rapid construction of multi-dimensional mapping model in the process of constructing digital twin scenes of coastal channel.Based on density clustering and DENCLUE algorithm,the representative feature points such as ship position,turning point and intersection point were extracted to construct the channel traffic flow network,providing a navigation area for unmanned boat cruise and data collection.According to computer vision technology and neural network algorithm,the semantic segmentation of water environment based on monocular camera carried by autonomous cruise unmanned boat was realized.The situational awareness of navigation environment was realized.The marine environment elements were integrated,and the collected information such as channel space elements and surface ship navigation was applied to the construction of twin scenes.A multi-source heterogeneous data fusion model of channel was developed,which realized the rapid creation of channel element data index under the condition of large concurrency.A digital twin of all elements of coastal channel with"geometry-physics-behavior-rule"was constructed,and the multi-dimensional mapping from physical channel to digital channel was realized.The verification results in the experimental sea area show that the method realizes the dynamic perception of unmanned boat environment,autonomous cruise,efficient modeling and enhanced display of channel elements,thereby providing an effective way for the rapid construction of digital twin scenes in coastal channel.
digital twinautonomous cruise controlunmanned vesselchannel elementsmulti-source isomerismdynamic perceptiondensity clustering
万方数据
维普
