The Application of a Water Depth Estimation Method to the Estimation of Seaweed Distribution in Shallow Water Using Satellite Remote Sensing

Abstract

To conserve seaweed beds in shallow waters, it is important to understand their time series change using accurate distribution data. Although remote sensing techniques can be used for regular observation in wide areas, the estimation of seaweed bed distribution using satellite image analysis requires manual correction of analysis results due to the noise generated on the water surface and the attenuation of light underwater. To improve the accuracy of this estimation, we used a depth estimation method that takes the seaweed flourishing season into consideration and applied it to seaweed bed distribution analysis. Our results show an overall accuracy of 78% and a tau coefficient of 0.53 under the application of satellite-derived bathymetry (SDB) estimated from the image taken during seaweed non-flourishing season to the analysis of the image taken during flourishing season. These results indicate that applying a water depth estimation method that takes seaweed flourishing season into consideration can improve the accuracy of seaweed bed distribution analysis.