| In the remote sensing satellite ground application processing systems, remote sensing image processing system is the most important component. This system is based on satellite ground stations, whose main task is to timely receive earth observation information in available areas from remote sensing satellite ground receiving station, and remove all kinds of distortion affect to get useful image information. After it, transform the image information into visual image products, and distribute the satellite image products to users via product distribution system for applications or scientific research.Satellite remote sensing image processing system is based on the practical project. According to the characteristics and idea of software engineering, develop browse image and cloud mask generating modules in the satellite ground application system. The analysis of the two modules is accomplished in participation, and the design and implementation is completed independently.Browse image generating module is divided into two parts, which are scene browse image generating and long-strip browse image generating. The module is in the structure of "main program+base class dynamic library+configure dynamic library files+plug-ins", and it is implemented by the method of the registered plugs to get browse image. In the secene browse image generating part, the work is to parse satellite LO long-strip data through catalog information file into scenes, and extract the required scene data via the data format of LO image data(including special cases of reading multi-band image each band of whose is in one file, and combined multi-camera and so on), then resample and enhance the image. At last, get the LO scene browse image of satellite cameras, and count the cloud amount to get the image quality information. In the long-strip browse image generating part, parse the LO long-strip image data through the catalog information file, extract LO long-strip data, and ultimately generate the LO strip image. Above all, the two modules generate images and related data for subsequent archiving and query.Cloud mask generating module is to detect cloud data of L2product, extract binary cloud information, remove the noise and smooth the cloud image by closing and opening operation, and then get the clouds with points by seed region grow algorithm. And finally, convert the coordinate figure into geographical coordinates, and call shapelib library to generate shp file with the shape type of multipoint, which is the cloud mask vector file.This paper detailedly analyses the function and performance requirements of the two modules, take full account of the scalability and reusability of satellites application software. In short, both the design and implementation have met the demand, and now it has been tested and run throughout the system. |
