Pine wilt disease detection in high-resolution UAV images using object-oriented classification

Abstract Pine wilt disease (PWD) is currently one of the main causes large-scale forest destruction. To control spread PWD, it essential to detect affected pine trees quickly. This study investigated feasibility using object-oriented multi-scale segmentation algorithm identify discolored by PWD. We used an unmanned aerial vehicle (UAV) platform equipped with RGB digital camera obtain high spatial resolution images, and was applied delineate tree crown, coupling use classification classify Then, optimal scale implemented estimation parameter (ESP2) plug-in. The feature space results optimized, appropriate features were selected for classification. showed that scale, shape, compactness values crown 56, 0.5, 0.8, respectively. producer’s accuracy (PA), user’s (UA), F1 score 0.722, 0.605, 0.658, There no significant errors in final results, low attributed number objects count caused incorrect segmentation. method could accurately PWD a straightforward rapid processing. provides technical monitoring occurrence identifying UAV-based high-resolution images.