论文部分内容阅读
基于图像对目标的机动检测信息,提出了一种新的基于二维图像序列的飞行目标姿态角估计算法。算法结合飞行目标结构特点,利用图像对其进行快速机动检测,根据目标的机动方向和机动程度,建立了自适应动态模型库,并提出了启发式搜索法,最后采用动态模型匹配的原理进行目标姿态角估计。该算法避免了创建动态模型库的盲目性,减小了动态模板库的规模,提高了最佳匹配模型的搜索速度,仿真结果表明,在保证姿态估计精度的同时,极大的提高了实时性。
Based on the maneuvering detection information of the target, a new attitude estimation algorithm of flying target based on two-dimensional image sequence is proposed. The algorithm combined with the characteristics of the flight target structure, used the image to carry out rapid maneuver detection, set up an adaptive dynamic model base on the maneuvering direction and maneuvering degree of the target, put forward the heuristic search method, and finally adopted the principle of dynamic model matching Attitude estimation. The algorithm avoids the blindness of creating a dynamic model library, reduces the size of the dynamic template library and increases the search speed of the best matching model. The simulation results show that, while ensuring the accuracy of gesture estimation, the algorithm greatly improves the real-time performance .