一种用于敏捷卫星自主任务规划的云顶高探测方法

A Cloud-Top Height Detection Method for Autonomous Planning of Satellites

高分辨率光学对地观测敏捷卫星自主任务规划需要对成像目标区域上空云层进行在轨提前探测,以修正成像时间窗,提高卫星使用效能。其中云高度是必不可少的信息,提出了一种用于敏捷卫星自主任务规划的云顶高探测方法。根据不同角度的遥感图像对之间云层以及地面匹配关系的差异,以地面为基准,求出图像间云层对地面的投影偏差,再利用几何关系,计算云层云顶高度。仿真模拟实验的结果表明本文算法探测云顶高与实际云顶高差距小于1 pixel,并且在图像间发生旋转、仿射等变换的情况下依然有稳健的表现。误差分析给出了算法相对误差随成像角度的变化趋势。多角度成像分光辐射度计图像处理结果表明该算法在实际应用中性能良好。

Autonomous planning of high-resolution optical earth-observing agile satellites requires on-orbit clouds detection over observation zones ahead, in order to reschedule the time window to enhance the operational effectiveness of satellites. Cloud heights are essential information, a cloud heights detection method for autonomous planning of satellites is proposed. On the basis of the difference in matching relationship between cloud and ground with different points remote sensing images, projecting deviation of cloud to ground is computed on the reference of ground and then cloud-top height is calculated by geometric relation. The simulation experimental result demonstrates that the disparity of cloud-top heights between detected and actual is less than 1 pixel and the method performs stably even rotation and affine transformation among images. Error analysis shows the trend of relative error along with different imaging angles. The experimental results of muti-angle imaging spectroradiometer images suggest that the proposed method has convincing performance over the real appilication.