STUDY ON THE APPLICATION OF THE MACHINE LEARNING MODEL 'GRAINet' TO A GRAVEL BEACH AND CHANGES IN GRAIN SIZE

Abstract

 The surveying method using UAV-SfM/MVS produces high-resolution ortho mosaic images that provide information on both the elevation and changes in gravel beach components. In addition, gravel size can be useful to understand the process of topographic change on gravel beaches. This study focuses on the machine learning model 'GRAINet' and its ability to estimate the size of gravel grains on Ida beach. A map of the distribution of gravel grain sizes was created, and changes in the spatio-temporal distribution of topography and gravel grain size were investigated. The results suggest that the demonstration data should be verified by comparison with the sieving test at the target site.