The onboard acquisition instruments of MS and Pan, sharing the same optics, cannot have parallel axes, but there is a small angle in the orbit plane. The accuracy of the geometric correction depends by the DEM used, and if the latter is too coarse with respect to the pixel size, a residual misalignment survives. A multivariate linear regression between the low-pass filtered Pan and the interpolated MS bands can be exploited to model the spectral and spatial relationships among the different channels. The pixel varying residue is a measure of the local mismatch and will be used to correct the interpolated MS bands after being multiplied by the contrast-based injection gains.
Automatic fine alignment of multispectral and panchromatic images
Arienzo A;Aiazzi B;
2020
Abstract
The onboard acquisition instruments of MS and Pan, sharing the same optics, cannot have parallel axes, but there is a small angle in the orbit plane. The accuracy of the geometric correction depends by the DEM used, and if the latter is too coarse with respect to the pixel size, a residual misalignment survives. A multivariate linear regression between the low-pass filtered Pan and the interpolated MS bands can be exploited to model the spectral and spatial relationships among the different channels. The pixel varying residue is a measure of the local mismatch and will be used to correct the interpolated MS bands after being multiplied by the contrast-based injection gains.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.