基于特征提取与跟踪的时变流体可视化系统的实现

Visualization system for time-varying flow volume based on feature extraction and tracking

针对流体数据存在特征多样化、定义复杂的问题,提出并实现了一种基于特征提取与跟踪的时变流体可视化系统。该系统采用体绘制技术对流体数据进行可视化,并基于预测和校准的方法对用户指定的流体特征进行提取和跟踪。对于已提取出的时变特征,系统应用淡入效果对其在连续时间内的运动轨迹进行重建,并利用骨骼化算法对其形状的变化进行分析。实验表明,系统在有效节省空间和时间的基础上,可以达到辅助用户深入理解时变流体内部变化规律的目的。

Due to the diversity and complex definition issue of the flow features, a feature extraction and tracking based visualization system is proposed and implemented in this paper. Volume rendering technique is utilized to render the flow data while the tracking purpose of the user specified features can be achieved by adopting the prediction-correction method. For the extracted time-varying features, the fade-in effect of the application system is used to reconstruct its motion trajectory, and the skeletonization algorithm for the shape changes is analyzed. According to the results of the experiments, the comprehension of the entire time-varying flow volume could be obtained through the system with the advantage of efficient spatial and temporal consumption.