| 摘要: |
| 针对目标个数的不确定性和目标航迹的初始化问题,提出一种综合的检测算法.首先将序列图像在时域上进行组合,消除背景影响.在此基础上对组合帧图像进行网格划分,估计出目标存在区域,从而缩短了目标搜索时间,然后利用帧间目标的互关联特性初始化各目标的航迹.仿真试验结果表明,该算法具有有效性与实时性的特点. |
| 关键词: 红外图像 运动目标检测 概率加权 双向互关联 检测前跟踪算法 数据融合 概率加权 帧间 互关联 多目标 初值估算 Correlation Bidirectional Weighted Probabilistic Based Estimation Initialization 实时性 有效性 检测算法 结果 仿真试验 目标航迹 初始化问题 特性 |
| DOI：10.3969/j.issn.1001-893X. |
| Revised:November 16, 2007 |
| 基金项目:国家自然科学基金
,
新疆维吾尔自治区教育厅高校科研计划科学研究重点项目 |
|
| Multi-target Trajectory Initialization Estimation Based on Probabilistic Weighted and Bidirectional Inter-frame Correlation |
|
| () |
| Abstract: |
| For solving the problem of the uncertainty of target number and the initialization of target tracks,a comprehensive detection algorithm is presented.First,the image sequences are combined in the time domain and the background clutter is eliminated.Then on this basis,the combined frame images are segmented to estimate the area existing target,and the search time is greatly shortened.Each target track can be initialized through using the correlation of inter-frame targets.The simulation result shows that this algorithm has effective and real-time characteristics. |
| Key words: infrared image,MTD,probability weighted,bidirectional correlation,TBD algorithm,data fusion |
