Abstract: Comparaion between higher spatial resolution image and lower spatial resolution image in time serial is very helpful for the changes of land cover. During the process of acquiring field survey data, higher spatial resolution image is always set as the reference image to select sampling plots. But sometimes the data of sampling plots need to be enchased to the lower spatial resolution image to get the simulation image for the further analysis. The paper sets wheat field survey data of He Bei province and TM image data as input parameters, and designs a pixel Unmixing model to acquire the 250 meter MODIS simulation image.

Key words: large-scale dimension, scaling transformation, data enchasing, image simulation

摘要： 在对时间序列土地覆盖变化的研究中，常常会遇到将高空间分辨率影像和低空间分辨率影像进行时间序列的比较的情况；同样在野外调查试验或采集时常用较高空间分辨率或大比例尺的地图作为选取采样点的参考图，回到实验室后又需要将实测点的数据镶嵌到低空间分辨率遥感图像上，获取更为直观的模拟影像进行研究分析。针对这一问题，在对河北省小麦种植区进行实地试验，获取一系列实测小麦参数后，提出一种基于高空间分辨率遥感影像的小麦实测数据在低空间分辨率影像上镶嵌的模拟方法。通过利用该方法建立的像元分解模型，结合种植区实验数据，获取了基于TM影像的小麦实测数据的MODIS模拟图像。

关键词: 大尺度空间, 尺度转换, 实测镶嵌, 模拟图像