Emerald ash borer (Agrilus planipennis) is an invasive pest that has killed millions of ash trees (Fraxinus spp.) in the USA since its first detection in 2002. Although the current methods for trapping emerald ash borers (e.g., sticky traps and trap trees) and visual ground and aerial surveys are generally effective, they are inefficient for precisely locating and assessing the declining and dead ash trees in large or hard-to-access areas. This study was conducted to develop and evaluate a new tool for safe, efficient, and precise detection and assessment of ash decline and death caused by emerald ash borer by using aerial surveys with unmanned aerial systems (a.k.a., drones) and a deep learning model. Aerial surveys with drones were conducted to obtain 6174 aerial images including ash decline in the deciduous forests in West Virginia and Pennsylvania, USA. The ash trees in each image were manually annotated for training and validating deep learning models. The models were evaluated using the object recognition metrics: mean average precisions (mAP) and two average precisions (AP50 and AP75). Our comprehensive analyses with instance segmentation models showed that Mask2former was the most effective model for detecting declining and dead ash trees with 0.789, 0.617, and 0.542 for AP50, AP75, and mAP, respectively, on the validation dataset. A follow-up in-situ field study conducted in nine locations with various levels of ash decline and death demonstrated that deep learning along with aerial survey using drones could be an innovative tool for rapid, safe, and efficient detection and assessment of ash decline and death in large or hard-to-access areas.

Invasive knotweeds are rhizomatous and herbaceous perennial plants that pose significant ecological threats due to their aggressive growth and ability to outcompete native plants. Although detecting and identifying knotweeds is crucial for effective management, current ground-based survey methods are labor-intensive and limited to cover large and hard-to-access areas. This study was conducted to determine the optimum flight height of drones for aerial detection of knotweeds at different phenological stages and to develop automated detection of knotweeds on aerial images using the state-of-the-art Swin Transformer. The results of this study found that, at the vegetative stage, Japanese knotweed and giant knotweed were detectable at ≤35 m and ≤25 m, respectively, above the canopy using an RGB sensor. The flowers of the knotweeds were detectable at ≤20 m. Thermal and multispectral sensors were not able to detect any knotweed species. Swin Transformer achieved higher precision, recall, and accuracy in knotweed detection on aerial images acquired with drones and RGB sensors than conventional convolutional neural networks (CNNs). This study demonstrated the use of drones, sensors, and deep learning in revolutionizing invasive knotweed detection.

The invasive morning glory, Ipomoea purpurea (Convolvulaceae), poses a mounting challenge in vineyards by hindering grape harvest and as a secondary host of disease pathogens, necessitating advanced detection and control strategies. This study introduces a novel automated image analysis framework using aerial images obtained from a small fixed-wing unmanned aircraft system (UAS) and an RGB camera for the large-scale detection of I. purpurea flowers. This study aimed to assess the sampling fidelity of aerial detection in comparison with the actual infestation measured by ground validation surveys. The UAS was systematically operated over 16 vineyard plots infested with I. purpurea and another 16 plots without I. purpurea infestation. We used a semi-supervised segmentation model incorporating a Gaussian Mixture Model (GMM) with the Expectation-Maximization algorithm to detect and count I. purpurea flowers. The flower detectability of the GMM was compared with that of conventional K-means methods. The results of this study showed that the GMM detected the presence of I. purpurea flowers in all 16 infested plots with 0% for both type I and type II errors, while the K-means method had 0% and 6.3% for type I and type II errors, respectively. The GMM and K-means methods detected 76% and 65% of the flowers, respectively. These results underscore the effectiveness of the GMM-based segmentation model in accurately detecting and quantifying I. purpurea flowers compared with a conventional approach. This study demonstrated the efficiency of a fixed-wing UAS coupled with automated image analysis for I. purpurea flower detection in vineyards, achieving success without relying on data-driven deep-learning models.

Japanese knotweed (Fallopia japonica) is a noxious invasive plant species that requires scalable and precise monitoring methods. Current visually based ground surveys are resource-intensive and inefficient for detecting Japanese knotweed in landscapes. This study presents a transformer-based semantic segmentation framework for the automated detection of Japanese knotweed patches using high-resolution RGB imagery acquired with unmanned aerial vehicles (UAVs). We used the Twins Spatially Separable Vision Transformer (Twins-SVT), which utilizes a hierarchical architecture with spatially separable self-attention to effectively model long-range dependencies and multiscale contextual features. The model was trained on 6945 annotated aerial images collected in three sites infested with Japanese knotweed in West Virginia, USA. The results of this study showed that the proposed framework achieved superior performance compared to other transformer-based baselines. The Twins-SVT model achieved a mean Intersection over Union (mIoU) of 94.94% and an Average Accuracy (AAcc) of 97.50%, outperforming SegFormer, Swin-T, and ViT. These findings highlight the model’s ability to accurately distinguish Japanese knotweed patches from surrounding vegetation. The method and protocol presented in this research provide a robust, scalable solution for mapping Japanese knotweed through aerial imagery and highlight the successful use of advanced vision transformers in ecological and geospatial information analysis.