Method for automatically optimizing and solving parameters of rational function model based on embedded regression

Abstract

The invention aims to convert a problem for solving rational polynomial coefficients (RPCs) into a multi-element linear regression problem with multi-collinearity, and provides an embedded regression method for automatically optimizing parameters of a rational function model by taking a determination coefficient as an evaluation rule. Coefficients with obvious influence in the rational function model are gradually introduced, and redundant coefficients with low influence are eliminated, so that the RPCs can be automatically optimized and solved, and the brief rational function model is obtained. By the method, the rational function model can be automatically simplified, the using amount of control points is reduced, and the image correction and collimation precision is improved.

Description

technical field [0001] The invention relates to an automatic optimization and solution method of rational function model parameters, which can simplify the rational function model, reduce the number of control points used, and improve image correction accuracy. It can be applied to remote sensing, photogrammetry, surveying and mapping, image processing and other fields. Background technique [0002] The rational function model (Rational Function Model, RFM) is a complete mathematical model. Its essence is to use a large number of control points to fit the strict imaging relationship. It does not require the actual characteristics of the sensor and the imaging process. It is suitable for different types of sensor, and can hide various physical parameters of the sensor in the imaging process, so it has been widely used in the field of photogrammetry and remote sensing. The key to using the rational function model to process satellite remote sensing images is to accurately sol...

AI Technical Summary

Problems solved by technology

However, because the form of multiple correlations between variables is very complicated, and there is still a lack of very reliable testing methods, the practice of deleting some related variables often leads to an increase in the interpretation error of the model, and discards the information that should be retained [Wang Huiwen. Partial least squares Regression method and its application. Beijing: National Defense Industry Press, 1999.]

Images