A New Wavelet Fitting Method for Geometric Correction

Since the trend is to integrate remote sensing images and other spatial data in geographic information systems, geometric correction is an indispensable preprocessing in various applications. The purpose of geometric correction is to compensate for the distortions caused by the earth’s curvature and rotation, the spacecraft’s speed, altitude and attitude, and the scan skew etc. Today polynomial fitting (PF) is a commonly used method for geometric correction. However, the correction accuracy in some cases is not very satisfactory. A new wavelet fitting (WF) method is presented in this paper. By introducing a rotation factor to the wavelet function, we get an improved wavelet basis which is more appropriate for describing the geometric distortion. The experimental results show that the presented method is very efficient and can greatly improve the correction results.