VEGETATION DENSITY ESTIMATION FOR RIVERINE BAMBOO FORESTS USING UAV IMAGES AND TREETOP DETECTION

Abstract

 An estimation method of vegetation density in the forests of three bamboo species was proposed to detect treetops using local maximum filtering (LMF) with a circular window under a high-resolution digital surface model via UAV images. At the highest case of estimation accuracy, the bamboo forest density results revealed that the method underestimated only about 10% in any species. The estimation accuracy depended on the window size (WS), a parameter of LMF. Thus, the application of optimal WS according to the canopy size of bamboo was paramount to obtain highly accurate results.