论文部分内容阅读
—利用红外前视传感器输出,拟定了一种广义卡尔曼滤波器算法用于开环跟踪以解决跟踪点源目标问题。该滤波器分别估计由于实际目标运动和大气扰动引起的前视装置视场中目标位置的变化。应用蒙特卡罗分析决定滤波器的性能,它是信噪比、目标象元点大小、目标运动与大气扰动均方差之比、相关时间、目标实际象点与滤波器设定象点不匹配等因素的函数。在相同的条件下,比较了广义卡尔曼滤波器和现行的相关跟踪的性能。在信噪比为20:1和1:1的条件下,得到综合跟踪误差为0.2和0.8象元。当象点变小,目标相关时间在有限范围内增加时,滤波器参数作相应调整,性能没有降低。灵敏度分析表明;滤波器对目标光点次要变化是稳定的。
- Using infrared front-view sensor output, a generalized Kalman filter algorithm is developed for open-loop tracking to solve the tracking point source problem. The filter estimates changes in the target position in the FOV field of view due to actual target motion and atmospheric disturbances, respectively. The application of Monte Carlo analysis to determine the performance of the filter, which is the signal to noise ratio, the target pixel size, the target movement and atmospheric turbulence mean square error ratio, the relevant time, the target actual pixel and the filter set point does not match Function of factors. Under the same conditions, the performance of the generalized Kalman filter and the current correlation tracking is compared. Under the conditions of SNR 20: 1 and 1: 1, the integrated tracking error is 0.2 and 0.8 pixels. When the image point becomes smaller and the target related time increases within a limited range, the filter parameters are adjusted accordingly, and the performance is not reduced. Sensitivity analysis shows that the filter is stable to the secondary change of the target spot.