In order to photograph old giant trees that are tens of meters tall living outdoors from the ground, appropriate settings such as distance from the camera, light quantity, exposure time and direction are essential. In addition, the vegetation index us...
In order to photograph old giant trees that are tens of meters tall living outdoors from the ground, appropriate settings such as distance from the camera, light quantity, exposure time and direction are essential. In addition, the vegetation index used to diagnose tree strength using high-resolution hyperspectral images varies sensitively depending on the selected wavelength bands, so a review of its appropriateness according to tree species is required.
The purpose of this study was to select a hyperspectral image vegetation indices and wavelength bands suitable for diagnosing the vigor of old trees. To this end, we measured the luminous efficiency used for diagnosing the growth status of Zelkova, Ginkgo and Pine trees, which account for a large proportion of natural monuments, protected trees, street trees, and park trees.
In the research process, considering the shooting specifications of the camera, acquisition of leaf spectral information, and noise generation, it was confirmed that a separation distance of 20m was the most efficient.
The vegetation indices were calculated by sequentially selecting the spectral reflectance in the range of ±50nm centered on the wavelength bands suggested in the previous studies. Correlation analysis was attempted for hundreds of vegetation indices and luminous efficiency indices. In the process, vegetation indices that showed high correlation with the luminous efficiency indices were identified.
Among the vegetation indices, NDVI, GNDVI, GCI, PRI, and MCARI, RENDVI and MRENDVI which are sensitive to stress, aging, and chlorophyll content, were found to be able to replace the luminous efficiency indices.
Through the analysis, it was revealed that Y(Ⅱ) can be used for Zelkova, NPQ for Pine, and Y(NO) for Ginkgo to diagnose tree Vigor. In addition, the wavelengths to be used for calculating the selected vegetation indices were also presented. In particular, it was revealed that the wavelength bands that were found to be effective for PRI was very effective in replacing light efficiency.
If the results of this study are used for diagnosis and periodic monitoring of old giant trees, it will increase the utility of hyperspectral imaging in maintaining the biological health of old trees and reducing the socioeconomic costs required for it.