流场可视化的最优视点选择方法

发布时间：2018-12-11 04:46

【摘要】：针对流场可视化当前视点下存在的遮挡与可能面临的非交互式环境的问题,提出一种最优视点选择方法.首先,根据当前视点下的流场行为和结构信息及可视投影的图像质量,利用信息理论概念进行最优视点综合性度量;其次,应用可视成像的相关性对视点间的相似性评估,以决定两视点间的最小重叠;最后,结合K均值聚类方法将相似视点归为一类,每一类中视点熵值最高的视点作为最具代表性的视点.应用各种流场数据集进行可视化最佳视点选择实验,给出了视点选择效果图、与其他方法的最优视点效果对比图以及选择过程的计算时间.实验结果表明,该方法能够自动、有效地进行最优视点选择.
[Abstract]:An optimal view selection method is proposed to solve the problems of occlusion and possible non-interactive environment under the current viewpoint of flow field visualization. Firstly, according to the flow field behavior and structure information of the current viewpoint and the image quality of the visual projection, the optimal viewpoint comprehensive measurement is carried out by using the concept of information theory. Secondly, the correlation of visual imaging is used to evaluate the similarity between viewpoints to determine the minimum overlap between the two viewpoints. Finally, the similar viewpoint is classified into a class with K-means clustering method, in which the viewpoint with the highest entropy value is regarded as the most representative viewpoint. The experiments of visual optimal view selection are carried out by using various flow field data sets, and the view selection effect diagram, the contrast diagram of the optimal view effect with other methods and the calculation time of the selection process are given. Experimental results show that the method can automatically and effectively select the optimal viewpoint.
【作者单位】：曲阜师范大学软件学院;中国科学院计算技术研究所虚拟现实实验室;
【基金】：国家自然科学基金(61379085) 山东省高等学校科技计划(J17KA062) 教育部产学合作协同育人项目(201602028014) 曲阜师范大学实验室开放基金项目(SK201723) 国家级大学生创新创业训练计划项目(201710446129)
【分类号】：TP391.41

【相似文献】