Abstract
Selective sampling has been widely adopted for stand-volume estimation. However, the accuracy of this method is low, owing to the variability of the investigator’s experience and subjectivity. To address this limitation, we developed a new method for stand-volume estimation by correcting the selective sample method based on the data of the tree height and determining the spatial distribution of each tree in the forest stand by using unmanned aerial vehicle-structure from motion (UAV–SfM) images. We selected a Sugi plantation to test the effectiveness of the new method for stand-volume estimation. Furthermore, the stand-volume was calculated from the LiDAR data of the region. A 3D model was constructed based on UAV images; the mean tree height and stand-volume were also determined from these images. The stand-volume estimated using the selective sampling method was corrected using the data obtained from the UAV-SfM images. The improvement rate of the estimation accuracy was defined as the percentage of the interval of stand-volume estimation compared to the interval considered for the selective sample method; the estimations conducted using the selective sample method were corrected for the mean tree height and stand volume derived from UAV-SfM images. The new method improved the accuracy of the stand-volume estimations; notably, the accuracy of the estimations conducted using the data corrected based on the stand-volume was higher than that based on the mean tree height.
https://ror.org/05qg1fd04 
