基于匹配法的卫星云图自动定位微调算法

Algorithm based on match in automatic inching orientation in satellite imagery

利用交叉相关系数法,通过设计晴空合成底图模板,并与当前云图进行匹配,以计算由于卫星飘移等问题出现的云图定位偏移量△x、△y,从而实现卫星云图的精确定位。结果表明:对于自动校正卫星云图中较小的定位误差效果(偏移量小于20个像素)的这种方法,可以用于静止卫星接收处理系统中在初定位后的精确校正。该算法从原理上也可用在定位误差较小的极轨卫星云图的准确定位上。匹配区域选择在海陆交界处、岛屿、湖泊和起伏大的山地等地表特征差异明显的区域效果明显。

To fulfill the accurate orientation of the satellite image, with the help of the method of cross correction, the cloudless sky synthesized bottom image template and designed was matched with the current satellite image to calculate the image offsets Δx and Δy caused by the excursion of the satellite. The outcome shows that this method has a good effect in the correction of the error in cloud motion wind retrieving caused by the excursion of the satellite; in addition, the algorithm can be used precisely in the orientation of the polar orbit satellite image whose error is small. It can be found the match effect is apparent when the selected matching area is close to the coastline, islands and islets, lake and fluctuant mountains.