Detection of the stem-boring damage by pine shoot beetle ( spp.) to Yunan pine ( Franch.) using UAV hyperspectral data.

INTRODUCTION

The stem-boring damage caused by pine shoot beetle (PSB, spp.) cuts off the transmission of water and nutrients. The aggregation of beetles during the stem-boring stage results in the rapid mortality of Yunnan pines ( Franch.). Timely identification and precise localization of stem-boring damage caused by PSB are crucial for removing infected wood and preventing further spread of the infestation. Unmanned airborne vehicle (UAV) hyperspectral data demonstrate great potential in assessing pest outbreaks in forested landscapes. However, there is a lack of studies investigating the application and accuracy of UAV hyperspectral data for detecting PSB stem-boring damage.

METHODS

In this study, we compared the differences in spectral features of healthy pines (H level), three levels of shoot-feeding damage (E, M and S levels), and the stem-boring damage (T level), and then used the Random Forest (RF) algorithm for detecting stem-boring damage by PSB.

RESULTS

The specific canopy spectral features, including red edge (such as Dr, SDr, and D711), blue edge (such as Db and SDb), and chlorophyll-related spectral indices (e.g., MCARI) were sensitive to PSB stem-boring damage. The results of RF models showed that the spectral features of first-order derivative (FD) and spectral indices (SIs) played an important role in the PSB stem-boring damage detection. Models incorporating FD bands, SIs and a combination of all variables proved more effective in detecting PSB stem-boring damage.

DISCUSSION

These findings demonstrate the potential of canopy spectral features in detecting PSB stem-boring damage, which significantly contributed to the prevention and management of PSB infestations.