项目名称: 面向异态检出的运动轨迹计算分析及其可视化关键技术研究
项目编号: No.61471261
项目类型: 面上项目
立项/批准年度: 2015
项目学科: 无线电电子学、电信技术
项目作者: 徐庆
作者单位: 天津大学
项目金额: 82万元
中文摘要: 本项目以运动轨迹数据为处理对象,有机地结合计算分析和可视化技术,深入研究运动轨迹的表示模型、运动轨迹聚类的个数选择、运动轨迹异态检出等亟需解决的关键性问题。主要包括:研究运动轨迹采样点的非参数化概率统计模型及其可视化,研究运动轨迹特殊局部特征的定量化度量,研究单个运动轨迹的规范表示及其可视化,基于互信息和熵散度研究轨迹数据的最优聚类个数选择,研究运动轨迹之间的距离度量,基于信息熵研究运动轨迹的异态检出及其可视化,研究运动轨迹数据聚类、异态检出等总体情况可视化,研究单个运动轨迹属性信息的二维及三维可视化,研究单个运动轨迹在其发生场景的三维可视化。本项目适应当前相关学术领域的研究趋势,立足解决学术前沿的基础性问题,对相关科学研究具有积极贡献。同时,本项目符合我国经济高速发展以及国家重大安全等对于应用基础性科研工作的迫切需求,研究成果具备潜在应用价值。
中文关键词: 运动轨迹;异态检出;可视化
英文摘要: The research on the analytics of trajectory data for anomaly detection is proposed in this project. Based on the computational analytics and visualization, this research basically includes the modeling of trajectory, the selection of the optimal number of clusters for trajectory data, and the anomaly detection. The research activities involved are as follows. Non-parametric modeling of trajectory points and its visualization, the quantitative measure of regional features for trajectory, the formal representation of a single trajectory and its visualization, the selection of the optimal number of clusters for trajectory data using mutual information and entropy based divergences, the quantitative measure for the difference between trajectories, Shannon entropy based anomaly detection and its visualization, the visual exploration of the trajectory data. The achievements by this project can not only help the abnormality detection and the computational analytics of trajectory data, but also do the potential contributions to national economics and safety.
英文关键词: trajectory;anomaly detection;visualization