A Particle Swarm Optimization based Approach to Pre-tune Programmable
Hyperspectral Sensors
Abstract
This article presents development of an innovative approach to identify
spectrally significant wavelength bands, for a given environment, to
tune hyperspectral sensor acquisition before UAV borne surveys. As
several programmable hyperspectral sensors are now available, it is
often a challenge to consider the suitable wavelengths of interest.
Researchers often conduct a thorough field survey to identify the
composition of target endmembers in an area to identify suitable
wavelengths before UAV survey, which is difficult and cumbersome.
Otherwise, the selection of wavelengths by trial-and-error is
error-prone.
To our knowledge, this is the first time a technique for optimal
hyperspectral band (or feature) selection has been proposed to pre-tune
UAV-hyperspectral sensors before the survey. A metaheuristic
evolutionary workflow using Particle Swarm Optimisation was used for
this. The method is easy in the field and efficient to identify optimal
bands before UAV-hyperspectral surveys.