The particle swarm optimisation for spectral matching applied to inland water quality remote sensing

Abstract

Spectrum matching is one approach to estimating water quality parameter concentrations (chlorophyll a, tripton and coloured dissolved organic matter (CDOM)) from remotely sensed images of inland waters. The Particle Swarm Optimisation (PSO) is a stochastic search technique that can be used to search the solution space for the closest match and deliver the best estimate of the water quality parameter concentrations. For Burdekin Falls Dam, a tropical freshwater impoundment, the PSO and four different matching criteria were applied to MERIS images to retrieve the water quality parameters. The mean retrieval error of the best performed similarity measure was 2.0 μgl-1, 2.45 mgl-1 and 0.3 m-1 for chlorophyll a, tripton and CDOM respectively. The paper found that the PSO, as implemented in this case, did not offer improvements in accuracy and precision sufficient enough to justify the increased computational overhead in the inversion.