Remote Diagnosis of Tree Vigor in National Natural Heritage Using Digital Hyperspectral Image Analysis

Abstract. This study aims to non-destructively diagnose the growth condition of old-giant-trees designated as natural heritage by identifying vegetation indices and wavelength ranges using hyperspectral images. The main findings are as follows. First, the study established a method for acquiring hyperspectral images of trees in outdoor environments. It recommends a 20-meter standoff distance, which enables full coverage of the canopy region while ensuring stable acquisition of spectral data from the leaves. Second, vegetation indices suitable for diagnosing tree vigor were derived for zelkova, ginkgo, and pine, which are tree species that account for a high proportion of old-giant-trees designated as natural heritage. The results show that vegetation indices can effectively replace conventional light efficiency indicators. In the case of zelkova, the specific bands used to calculate indices with high correlation to light efficiency were identified. Third, a regression equation for the light efficiency indicator was developed and applied to canopy-level hyperspectral images, demonstrating that vegetation indices derived from selected wavelength ranges can be used to diagnose tree growth condition. This study is significant in that it proposes a method for quantitatively diagnosing the growth condition of old- giant-trees across their entire canopy using hyperspectral images. The findings can be applied as a scientific, non-destructive management technique for conserving the physiological characteristics and historical value of old-giant-trees designated as natural heritage.