论文部分内容阅读
所研究的序列假设测试适用于数字图象序列中有关低可视机动点源目标的多帧检测和跟踪。所提出的多重多阶假设测试跟踪算法采用多阶假设跟踪策略，从序列检测目标轨迹那部分延伸所形成的轨迹。ＭＭＨＴＴ算法不需概率目标机动模型。利用截断序列概率测试求出计算的有效系数，以便删除密集候选目标轨迹树和评价被检测的那部分轨迹。文中利用光学卫星跟踪试验结果，介绍和证实了分析的性能评定。
The proposed sequence hypothesis test is suitable for multi-frame detection and tracking of low-visual, moving-point source targets in digital image sequences. The proposed multiple multi-stage hypothesis testing and tracking algorithm uses a multi-stage hypothetical tracking strategy to extend the formed trajectory from the portion of the target detection trajectory. The MMHTT algorithm does not require a probabilistic target maneuver model. The truncated sequence probability test is used to find the calculated significant coefficients in order to delete the dense candidate target trajectory tree and to evaluate the part of the trajectory being detected. In this paper, optical satellite tracking test results, introduced and confirmed the performance evaluation of the analysis.