Research On Remote Sensing Image Correction Model In Region Of Lack Control Points

The orientation of remote sensing (RS) image is one of the key techniques in RS applications. The traditional RS image orientation methods must depend on certain number of ground control points (GCP). Precise orientation of RS image cannot be done in the regions with few GCPs. However, there are a lot of regions without or/with less maps and with difficult to find GCPs in our country. Wherefore, it is significant both academically and practically to study the precise correction methods of RS image in the GCP scarce regions.The main contents of this thesis are:Establishes a model for the solution of satellite image elements of exteriororientation and its precision analysis. The key technique for precise orientation of RS image in GCP scarce region is to obtain the high-precision elements of exterior orientation on the imaging moment. The assistant information file of SPOTS (Metadata.dim) involves abundant Position and Orientation System (POS) data related with positions and attitudes. Based on these data, Lagrange interpolation, polynomial fitting, cubic spline fitting and statistics methods are used to carry out high-precision parameter equations of each POS data in different time. The elements of exterior orientation of SPOTS at random imaging moment can be calculated with these equations and precision can be estimated.A collinearity equation based space resection adjustment model is presented under frame imaging pattern. This model takes the initial values of elements of exterior orientation as the virtual observations and collinearity equation as conditions. Condition equations with a few GCPs (less than 3) are set up. The corrections for the elements of exterior orientation can be calculated through adjustment and the maximum likelihood values of each element are obtained. The precise correction of images cab be done with these results in the cooperation of DEM data. The space resection adjustment model is inducted to CCD imaging pattern and a new dynamic scanning imaging pattern based space resection condition adjustment model is build to solve the precise correction problem of SPOTS images in the GCP scarce regions.