Classification of Tree Species in Satoyama Coppice Forests Based on Change in Color Tone of Time-series Color Images

Abstract

In this study, we investigated the change in the color of leaves of typical component trees in Satoyama coppice forests in the suburbs of Fukuoka City, on the basis of the changing properties of RGB and HSI values of aerial color photo data taken time serially throughout the year. In addition, we extracted the conditions effective for identifying the tree species (characteristic volume of photographing time and colors of RGB and HSI), and studied the possibility to identify the tree species by the combination of these conditions. As a result, it has become clear that the variance was large in the values of R, H and I among the tree species in the middle of May and early of December, and it was possible to identify specific tree species by processing these values through addition and comparison operations and combining the processed results.