RGB sensor integrated into unmanned aerial vehicle for monitoring cyanobacterial density in reservoirs

• Published in: Integrated environmental assessment and management 2025-01
• Main Authors: Silva, Will Jones Moura Soares da, Farias, Alex Bruno da Silva, Nery, Janiele França, Silva, Emanuel Araújo, Molica, Renato José Reis
• Format: Article
• Language: English
• ISSN: 1551-3777; 1551-3793
• DOI: 10.1093/inteam/vjae003

The proliferation of cyanobacteria has become a significant water management challenge due to the increasing eutrophication of water supply reservoirs. Cyanobacterial blooms thrive on elevated nutrient concentrations and form extensive green mats, disrupting the local ecosystem. Furthermore, many cyanobacterial species can produce toxins that are lethal to vertebrates called cyanotoxins. Traditional monitoring methods are inefficient for assessing water quality in reservoirs as a whole, given that sampling is only carried out in the catchment area for the public water supply, which exposes the population to the risk of contamination due to the multiple uses of these reservoirs. Therefore, novel monitoring methods supported by recent technological advances, such as the use of unmanned aerial vehicles (UAVs), are being tested for their effectiveness in monitoring cyanobacterial densities in aquatic ecosystems. This study analyzed UAV images of two water supply reservoirs to assess the effectiveness in monitoring cyanobacterial density. The UAVs were equipped with RGB sensors and flew over the study areas on the same day and at the same locations as water sampling performed for the determination of phytoplankton density, biovolume and chlorophyll-a. The phytoplankton community was dominated by cyanobacteria in both reservoirs. High coefficients of determination were obtained in the predictive models for chlorophyll-a concentration (r2 = 0.92), total phytoplankton and cyanobacterial densities (r2 = 0.89 and r2 = 0.97, respectively), and total phytoplankton and cyanobacterial biovolumes (r2 = 0.96 for both). Applying the predictive models to the orthomosaics generated from the UAV RGB images enabled the visualization of the spatial distribution of the phytoplankton and cyanobacterial biomass through distribution maps. This method has potential application in the management of water bodies that are crucial to the public water supply.