The Research Of The Validation Of Remote Sensing Land Surface Products And System Implementation

In the domain of remote sensing, the validation of the land surface products produced from the satellite data is important. This thesis would research the validation process and find the method to solve the key problem. Also a validation system would be developed for validation. The main research contents are as below:First, we research how to read the remote-sensing image file using the GDAL, and get the basic information of the image the same time such as the projection information and the geographic coordinate information. HDF is a data format used popular in remote-sensing image, but GDAL can not get the subDataset. We researched this and solved this problem. We can also transform the HDF file to other image format. The Feng Yun satellite data does not contain geographic information. We develop a method to get its coordinate fast by analyzing the Feng Yun tile data format.Scale transformation is an important part in validation. We use image resampling method to complete scale transformation. First, research how to get the corresponding pixel between two images in different scale. Then we use some typical resampling methods to transform the image, including nearest neighbor, bilinear, cubic convolution and pixel aggregate. As the land surface products do not have value in cloud area and water area, when we use ENVI to achieve transformation, the result would have some error pixel value. We solve this problem by abandon the void pixel during resampling.The thesis uses the longitude and latitude as the intermediate variable to get the transformational relation between two different projections. There are several steps to transform the projection of an image.First, calculate affine transformation parameters and size of the output image. Then create the output image and add geographic information to the image. And resample the image at last.At last we developed the validation of land surface satellite products system. The software can read and display remote-sensing image, and provide the function to transform the image scale and projection. We can complete the whole validation process in the software system. The system will give a evaluate result after validation, including evaluation parameters, differential image and RGB image. The thesis chooses some typical Feng Yun satel ite surface products to research and validate at last.In the thesis, we research the image format and develop a method to get the coordinates of Feng Yun data. The thesis analyzes the scale transformation and projection transformation then. And a validation system is developed, which provides the function of validation. We use MODIS data to validate FengYun data, and from the validation result we get that difference exists between the two products but not big.