High-precision geometric correction method of satellite remote sensing imagery by slightly correcting rpc parameters

Abstract

Provided is a method of realizing satellite remote sensing image high precision geometric correction through slightly modifying RPC parameters, which realizes image high precision geometric correction through adjusting 78 rational polynomial coefficients in RPC parameters of an RPC model after traditional RPC image space affine transformation parameter orientation, wherein a normal equation ill-conditioned problem utilizes precision information after image space affine transformation orientation to adaptively construct a regularizer and regularization parameters, thereby realizing robust resolution of 78 RPC parameters of satellite remote sensing images. Under the condition of sufficient ground control points, the method can obtain higher geometric correction precision than a traditional image space affine transformation RFM recursive model and a precise model, and has obvious advantages than traditional methods in the remote sensing image automatic matching and correction filed based on existing reference geographic data.

Description

technical field [0001] The invention relates to the technical field of geometric processing technology of satellite remote sensing images and the solution of ill-conditioned equations, in particular to a method for realizing high-precision geometric correction of satellite remote sensing images by slightly correcting RPC parameters. Background technique [0002] The form of the rational polynomial model RFM of the general processing model of satellite remote sensing images is as follows: [0003] [0004] In the formula: N umR (Y n ,X n ,Z n ),D enR (Y n ,X n ,Z n ), N umC (Y n ,X n ,Z n ),D enC (Y n ,X n ,Z n ) is the normalized object-space coordinate X n ,Y n ,Z n The ternary cubic polynomial, R n ,C n For normalized image coordinates, the form is: [0005] N umR (Y n ,X n ,Z n ) = a 1 +a 2 x n +a 3 Y n +a 4 Z n +a 5 x n Y n +a 6 x n Z n +a 7 Y n Z n +a 8 x n 2 +a 9 Y n 2 +a 10 Z n 2 +a 11 x n Y n Z n +a 12 x n 3...

AI Technical Summary

Problems solved by technology

Although the regularization theory has been developed and applied for nearly 40 years, the choice of a reasonable regularization parameter α and a reasonable stable functional Ω(Z) has always been the difficulty of the theory and method, and there is still no widely recognized The general optimal method for different practical problems brings uncertainty to the solution of the model

Images