2018 年 24 巻 p. 263-266
This study reports analyzing precision of grain size distribution of the riverbed surface by automatically analyzing UAV photographs. The grain size distribution curve and representative grain diameter varies according to the photographing condition; camera performance, flight height of UAV, wet riverbed and so on. Precision of analysis decreases by identifying the shadow, pattern on the grain surface or soil as a fine grain. This study examined the ability to improve the analyzing precision by monochrome image, higher brightness and population-break. Removing fine grains by population-break is effective in improving the analyzing precision, so the representative grain diameter is in error by less than 10 percent.