Processing technology of multispectral remote sensing images

This article is devoted to technology of geometric and spectral correction of a multispectral satellite image. The aim was the satellite image resolution enhancement technology based on wavelet packet transforms. The proposed technology has been tested on satellite images. Comparison of quantitative indicators (Shannon entropy, signal entropy, mean deviation, Shannon conditional entropy, conditional signal entropy and peak signal-to-noise ratio) as well as the visual results show the advantage of using the proposed technology. The results can be used for further object recognition and thematic processing of scanner images.

I. Introduction

Optoelectronic space images of high spatial resolution have a crucial importance for a large number of subject areas (physical state of remote sensing material object diagnostics, environmental monitoring, land use, Geographic Information Systems etc). The WorldView-2 and the WorldView-3 are among the most advanced satellites of optoelectronic complexes of ultrahigh spatial resolution. They are the core of the Earth remote sensing orbital grouping. WorldView-3 data, due to the additional shortwave 8-channel infrared sensor SWIR, help make capture through fog, smog, and dust with output spatial resolution of 3.72 m. It is especially sensitive to moisture and to the properties of electromagnetic radiation absorption by minerals and soils [1].