Satellite imagery fusion with an equalized trade-off between spectral and spatial quality

A methodology for obtaining fused images with an equalized trade-off between the spectral and spatial quality has been proposed. This methodology is based on a joint MultiDirection-MultiResolution representation (MDMR), defined through a Directional Low Pass Filter Bank (DLPFB) and complemented with a strategy of search DLPFB's parameters. This strategy uses a stochastic method, which optimizes an objective function associated to a spectral and spatial quality measure of the fused image. The results obtained in this work, show that a low number of iterations are needed for getting values of the DLPFB's parameters which provide fused images with a higher spectral quality than the images fused by other methods investigated here, conserving its spatial quality. It allows an equalized trade-off between the two considered qualities, against the other methods.