Using a Wavelet Based Method for High Resolution Satellite Image Fusion

Abstract

The number and quality of commercially available multispectral sensors and the
data they provide are continually improving, but there is always compromise
between achieving high spatial resolution, necessary for those applications that
require high degree of detail, and high spectral resolution when a better feature
discrimination level is needed. However, there are some situations that
simultaneously require high spatial and spectral resolutions in a single image. The
techniques of data fusion, or data merging, provide an alternative to that constraint,
being used to combine low-resolution multispectral satellite imagery with higher
resolution panchromatic or radar imagery, improving their visual quality and
interpretability. Many algorithms and software tools have been developed for fusing
panchromatic and multispectral datasets in remote sensing. Wavelet techniques are
increasingly being used for the processing of images. The algorithm used in this
paper was based on multiresolution wavelet decomposition. The image is
decomposed into multiple channels based on their local frequency content, obtaining
new images each one of them with different degree of resolution. A simple Wavelet
Transform is used, which is implemented in the ERDAS Imagine Software package
ver. 9.2. The procedure was based upon the wavelet transform is to improve the
spectral quality of high resolution image acquired by LANDSAT 7 ETM+.