Uso de veículo aéreo não tripulado (Vant) na estimativa da turbidez das águas do Canal São Gonçalo, Pelotas - RS, Brasil.

Abstract

Water is an integrating and fundamental element for human life and development. To ensure the availability of water with adequate quality for different uses, effective and aligned management and control tools are needed as the speed of that the modern world presents. Conventional sampling techniques for monitoring water resources have spatial and temporal limitations and in this point remote sensing through unmanned aerial vehicles (UAVs) is one of the technologies capable of adding speed and expanding the monitoring areas. The objective of this dissertation was to estimate the turbidity of water through reflectance values obtained from the calibration of orthomosaics produced by UAVs in the visible spectrum (RGB). The methodology developed in this study included field activities of turbidity sampling, aerial image capture by UAVs and Sentinel 2 satellite image acquisition. In situ turbidity sampling was performed using a multiparametric probe and aerial imagery was captured using an RGB image sensor embedded in a Phantom 4 drone (DJI) in almost sync with the passage of the MSI sensor (Sentinel 2). For spectral calibration, reference plates with known reflectance were used. The relationship between the variables investigated allowed the identification of 7 main factors that affected the results achieved, in addition to several recommendations for application in research involving similar studies in small aquatic ecosystems.