A new lake algae detection method supported by a drone-based multispectral camera

Abstract

Algal detection and quantification are essential steps needed to maintain the appropriate ecological status of freshwater bodies. Although there are still some technical issues to be addressed, remote sensing technologies possess benefits over traditional testing methods. To overcome these difficulties, algal concentrations at selected locations in Lake Balaton, Hungary, were determined with the use of a multispectral camera, mounted by a 3D printed tool on a drone. The algae concentration was defined from three different camera output variables, including light level, irradiance and reflectance. The determination was based on blue/green and also NIR/red indices. To validate the method, results from drone measurements were compared to laboratory measurements of collected water samples from the same 29 sites at which the drone camera took images. Pearson's correlation was applied to test the agreement of the measured and method-derived values. The blue/green ratio proved to be a more adequate input than NIR/RED, with the highest correlation being produced by the light level, blue/green ratio-based data that exhibited a highly significant Pearson correlation coefficient (r = .96). This newly developed drone-based method was shown to provide notably better spatial resolution than the satellites. Accordingly, the newly developed, quick-process measurements obtained in the present study can be done as frequently as required with a markedly lower budget.