Miglino, Domenico (2024) River quality monitoring with Image analysis. [Tesi di dottorato]

Documento PDF PhD_Thesis_Domenico_Miglino_XXXVII_Ciclo.pdf Download (6MB)

Abstract

Fresh water is becoming a more and more valuable resource which deserves to be properly managed and wisely preserved, exploiting all available techniques and knowledges. In this context, water turbidity is an essential proxy for assessing river quality and protecting aquatic ecosystems, since it is linked to the presence of organic and inorganic suspended matter. Current monitoring practices are mainly limited by low spatial and temporal resolution, and high costs, preventing the achievement of extensive and timely water quality information at global scale. Digital cameras can overcome these constraints, providing significant advantages to existing techniques. This study proposes an image analysis procedure for river turbidity assessment under varying hydrological conditions. Several turbidity events were artificially re-created on site during the real-scale tests using clay tracers with different colors and concentrations, also varying camera types and installation setups, for the optimization of the monitoring system. The experimental validation of the data was performed by installing two turbidimeters within the river section. The field tests revealed that environmental and hydrological long-term factors, including the characteristics of suspended particles, water level, and light condition, influence the effectiveness of the method. Riverbed background reflectance should also be taken into account because it could strongly modify the total water upwelling light, especially for shallow waters. Considering all these variables, the results showed that for short-term experiments (minutes, hours) the camera single bands values, in particular red band, seems to be the most reliable indicator for estimating water turbidity, better than bands ratios. Instead, for long-term monitoring (days, seasons), the single bands reflectance tends to be more influenced by light and river flow variations, while bands ratios performances start to increase. The purpose of this work is to deepen the implementation of this camera system in real-world settings to support existing river monitoring practices with early warning networks and to aid in developing new solutions for water resources management.

Downloads

Actions (login required)

Modifica documento