Estimation of Water Depth of Shallow Rivers by Analyzing Optical Remote Sensing Images Captured with a Drone

• Published in: Sensors and materials Vol. 35; no. 9; p. 3363
• Main Authors: Choi, Byoung Gil, Kwon, Yong Hee, Lee, Jun Hee
• Format: Journal Article
• Language: English
• Published: Tokyo MYU Scientific Publishing Division 01.01.2023
• Subjects: Data processing; Deep water; Estimation; Image acquisition; Image resolution; Pixels; Regression analysis; Remote sensing; Rivers; Shadows; Water depth
• ISSN: 0914-4935; 2435-0869
• DOI: 10.18494/SAM4428

In this study, we developed a method for estimating the water depth of shallow rivers by analyzing images captured with a drone, using optical remote sensing techniques. In an attempt to compensate for the shortcomings of existing surveying methods, optical-remote-sensing-based methods are being actively developed, but environmental conditions and data processing methods for application to rivers have not yet been sufficiently optimized. Here, we present an equation for estimating the water depth of shallow rivers from drone images and field survey results acquired under various conditions, and we aimed to verify accuracy using checkpoints. We found that estimating the water depth by calculating the parameters using multiple linear regression analysis based on the pixel values ​​of each band of the image and the field-surveyed water depth is more efficient than the conventional field survey method. In addition, the use of high-resolution images taken at noon without shadows and the removal of reflected light using a polarizing filter proved to be effective approaches in that nearly 88% of the images were within the acceptable range for bathymetry and about 94% were within the acceptable range when converted to low resolution. Finally, estimation of the water depth using the optical remote sensing technique indicated that the accuracy was low for deep water and that pixel values could be distorted by water plants or shadows.