A Critical Comparison Among Pansharpening Algorithms

Abstract

Pansharpening aims at fusing a multispectral and a panchromatic image, featuring the result of the processing with the spectral resolution of the former and the spatial resolution of the latter. In the last decades, many algorithms addressing this task have been presented in the literature. However, the lack of universally recognized evaluation criteria, available image data sets for benchmarking, and standardized implementations of the algorithms makes a thorough evaluation and comparison of the different pansharpening techniques difficult to achieve. In this paper, the authors attempt to fill this gap by providing a critical description and extensive comparisons of some of the main state-of-the-art pansharpening methods. In greater details, several pansharpening algorithms belonging to the component substi- tution or multiresolution analysis families are considered. Such techniques are evaluated through the two main protocols for the assessment of pansharpening results, i.e., based on the full- and reduced-resolution validations. Five data sets acquired by different satellites allow for a detailed comparison of the algorithms, characterization of their performances with respect to the different instruments, and consistency of the two validation procedures. In addition, the implementation of all the pansharpening techniques considered in this paper and the framework used for running the simulations, comprising the two validation procedures and the main assessment indexes, are collected in a MATLAB toolbox that is made available to the community

Pansharpening by Convolutional Neural Networks

“…Clearly, the whole procedure rests upon the hypothesis that performance does not depend critically on scale, which is not always the case, as discussed in the Introduction. Therefore, to reduce the possible mismatches, we smooth data before downsampling, following [2,22], using a filter that matches the modulation transfer function of the sensor. Likewise, we upsample the MS component using the interpolation kernel proposed in [13].…”

“…We tested all the techniques analyzed in the recent review paper [2], whose implementation has been made available online [50]. However, in the following tables, we report results only for the most competitive ones, listed below, according to the quality indices adopted.…”

“…Pansharpening is very important for remote sensing scene interpretation, and is also used as a pre-processing step for several image processing tasks, like feature extraction, segmentation and classification. Consequently, it has been the object of intense research, and many methods have been proposed in the last decades, as reported for example in [2].…”

Scale-Aware Pansharpening Algorithm for Agricultural Fragmented Landscapes

“…Many PS algorithms are available today [7][8][9][10][11][12]; most of them follow two approaches [8,9]: component substitution (CS) and multi-resolution analysis (MRA). The former includes widely-used PS algorithms, such as: intensity-hue-saturation (IHS) [13], principal component analysis (PCA) [14], Grand-Schmidt (GS) [15] and Brovey transform (BT) [16].…”

“…The main contribution of this new methodology is to produce artifact-free high spatial resolution pansharpened images, improving the MS edges by becoming aware of the scale (scale-aware) in images with a predominantly agricultural landscape. The performance of this method was compared with the WATFRAC, BT, GS, IHS and MTF-GLP (MTF using generalized Laplacian pyramid) [28] algorithms and evaluated using a set of spectral, spatial and overall quality indices [7,8]. The set of quality indices was summarized and compared using the Borda count (BC) method [29,30].…”