Multi-Scale Flow Field Mapping Method Based On Real-Time Feature Streamlines

Ocean current is an important physical phenomenon,which plays a restrictive role in climate,weather and other changing processes.The research and development of ocean flow field visualization methods have important theoretical and practical significance for analyzing the characteristics and rules of ocean current,guiding the design of ocean engineering,and improving the level of comprehensive management and public service of the ocean.Traditional static flow visualization methods have many problems,such as inability to express the continuity of the flow field,the messy visualization,time-consuming calculations and so on,which makes it difficult to visualize the flow field fast and intuitively.Besides,the dynamic expression method such as particle system is inconvenient to study the static distribution information of the flow field at a certain time.Therefore,from the perspective of data processing,streamline tracking,seed point layout and conflict detection,this thesis proposes a multi-scale flow field mapping method based on real-time feature streamlines,and the algorithm is as follows::To enhance the feature in small-scale flow field data,this thesis uses the mean filtering algorithm to perform filtering operation on this data,which can weaken the ocean current information with no obvious change in flow velocity and direction,and strengthen the feature information with obvious change.In this thesis,the first-order Euler method with high efficiency and less time consumption is used as the integral function to solve the problem of streamline tracing,and an adaptive step parameter calculation method is proposed to solve the problem of streamline step length calculation at different scales.Aiming at the seed point layout problem,this thesis adopts the "grid+feature" double seed point placement method to arrange the visual area.The global regular grid distribution method is used to evenly distribute seed points to the whole visual area,and the feature region random distribution method is used to place seed points in the ocean current feature area so as to solve the problems of uneven distribution of seed points and unclear feature expression.To solve the problem of streamline collision,this thesis puts forward a method to visualization after conflict processing,in this method,the space attribute information is used to rough testing the gathered gland of streamline,and a distance threshold parameter method is used to correct the coarse detection results of misjudgment.Finally,map load evaluation and comprehensive survey evaluation of this mapping method are carried out.And the experimental results show that the map load can reach the optimal interval of 18-28%when the basis sampling interval is 14-16 and the multi-scale sampling balance parameter is 4.Meanwhile,after modeling and analyzing the survey results of 100 random people in different fields,it can be found that the proportion of people who give five-star and four-star accounts for 71%.The results show that the majority of users think that the proposed method has good performance in multi-scale feature retention,flow field continuity expression,and drawing speed.