Author(s): Yosei Yamasaki; Ryuya Takeda; Shoji Okada
Linked Author(s):
Keywords: Turbidity; Spectral Reflectance; Satellite Data; Multispectral Camer
Abstract: Turbidity monitoring in river sediment monitoring is generally conducted by acquiring point data, and measuring spatially and temporally varying turbidity requires a great deal of labor and time. To solve this problem, there is a method to estimate turbidity from the reflectance of turbid water by remote sensing. The turbidity estimation equation used in this method is constructed based on the relationship between turbidity and reflectance measured in the field or experimentally1) 2). However, because turbidity varies depending on the composition and particle size of the suspended solids that make up the turbid water, it is difficult to use a specific formula to predict the turbidity that is spread over an area due to multiple sources of turbid water. Therefore, it is necessary to understand the reflectance characteristics of turbid water for each type of suspended solids and to select an appropriate turbidity estimation equation. In this study, we conducted experiments1) using riverbed materials collected near the mouth of a river in Japan to clarify the characteristics of changes in reflectance as turbidity increases. Experiments were conducted using a small multispectral camera to observe reflectance at red, green, blue, and near-infrared wavelengths for several turbid waters. The results of this experiment showed that each sand had different reflectance characteristics. Cluster analysis was performed using the Ward method. As a result, it was confirmed that some degree of classification was possible when sand with characteristic coloration was used. Although further clustering with additional sand types and calibration of the turbidity estimation equation are needed, it is thought that calibration of the turbidity estimation equation for each river can be omitted if the reflectance of the red, green, and blue wavelength bands of turbid water in the river is measured.
Year: 2024