4:30 PM - 4:45 PM
[AOS17-10] Estimation of the velocity field of a river front captured over a wide range
Keywords:a river front, estimation of the velocity field, drone
An aerial drone was used to observe a river front in Akkeshi bay, Hokkaido to investigate the mixing process between river water and seawater. A river front is a boundary between river water and seawater, where salinity changes abruptly, and the location of this front changes in minutes and meters. We need to take images at a frequency of a few seconds and a spatial resolution of a meter or less, thus, capturing the detailed spatial structure and evolution of the front is difficult from ship observation. So far, we used images that were taken from an aerial drone by pointing the camera directly below to achieve uniform resolution and convenience for analysis. However, this method has a narrow shooting range and cannot be used to grasp the large spatial structure of a river front.
In this study, we tilted the camera to capture the velocity field along a river front with high resolution nearby while also capturing a wide range at the far area. We used the spatial location optimization method to efficiently track a river front’s contours and accommodate deformations between images. The velocity field was successfully estimated along the river front after the effects of an aperture eclipse and illumination in the images were filtered out. Estimated values were close to those observed at the site using ADCP.
In the future, we plan to estimate the velocity field with higher resolution to capture the eddies forming within a river front.
In this study, we tilted the camera to capture the velocity field along a river front with high resolution nearby while also capturing a wide range at the far area. We used the spatial location optimization method to efficiently track a river front’s contours and accommodate deformations between images. The velocity field was successfully estimated along the river front after the effects of an aperture eclipse and illumination in the images were filtered out. Estimated values were close to those observed at the site using ADCP.
In the future, we plan to estimate the velocity field with higher resolution to capture the eddies forming within a river front.