中文摘要：

摘要：首先通过换算视场坐标确定灰度矩阵中每个元素对应的采样点在地球上的经纬度，从而将灰度矩阵转化为卫星云图，并添加海岸线．在此基础上，使用相关匹配法对具有一定时间问隔的两幅相关卫星云图进行模板匹配生成风矢场（云导风）矢量．然后，借助于近年来发展起来的数值微分方法，从图像灰度中提取出图像梯度信息，再利用正则化方法，实现了云导风的反演．对云图中加入灰度梯度信息和未加入灰度梯度信息的风场反演结果进行比较．结果表明，加入图像灰度梯度信息后所实施的新反演方法可有效减小图像干扰的影响，同时也大大提高了风矢量反演的精度，为卫星云图反演云导风探索出一条新路．

英文摘要：

Firstly, find out the latitude and longitude of sampling points correspond to each element in the gray-scale matrix, create a satellite cloud image and add the coastline on it. On this basis, generate a cloud-derived wind vector through using the correlation matching method from two satellite cloud images with a certain time interval. And then, achieve the inversion of the cloud-derived wind by the grayscale gradient information extracted from the image. Comparison of cloud images added grayscale gradient information wind retrieval results and not. The results show that the new inversion method can effectively reduce the impact of image interference, but also greatly enhance the accuracy of wind vector retrieval, and explore a new way.