Abishek Poudel

Monitoring forest regeneration is vital for sustainable management, yet traditional ground surveys and early aerial imagery methods face cost, labor, and resolution limitations. Unmanned Aerial Vehicles (UAVs) offer cost-effective, flexible platforms for acquiring ultra-high-resolution data to address these gaps. This paper reviews recent research on UAV applications in forest regeneration surveys (FRS), tracing the evolution from field-based surveys and conventional aerial approaches to current UAV practices, and synthesizing developments in data acquisition, processing workflows, and analysis. It contrasts established Canopy Height Model (CHM) and point-cloud approaches with the growing use of deep learning, particularly Convolutional Neural Networks (CNNs) for seedling detection, crown delineation, density, height estimation, and species classification. Particular attention is given to accuracy assessment, examining sampling design, reference data, prediction-to-reference matching, and evaluation metrics, and highlighting the disconnect between traditional map-validation principles and standard deep-learning metrics that often neglect background classes. The case study applying Mask R-CNN to red pine seedlings in an Adirondack Park plantation achieved stand-level recall of 70.3% and precision of 98.7%, while plot-level DL detections represented only 36.8% of the field-observed seedling count. These results demonstrate the potential of DL for reliably identifying visible red pine seedlings while highlighting its limitations for complete regeneration inventories, particularly when seedlings have small crown sizes.

This study demonstrates the power of combining unmanned aerial vehicle (UAV) imagery and deep learning (DL) for monitoring forest regeneration, specifically focusing on young white pine (Pinus strobus). Using high-resolution three-band RGB and five-band multispectral orthomosaics derived from UAV flights, 20 DL object-detection models were evaluated within ArcGIS Pro 3.4 software (Esri Inc., Redlands, CA, USA). The models were tested across study sites in St. Lawrence County, NY, to assess performance on three distinct size classes of white pine, each stratified into low, medium, and high density areas. The Faster R-CNN (F-RCNN) model, particularly when trained with image rotation and no augmentation, significantly outperformed others, achieving an average precision of 0.88 across both imagery types. Subsequent confusion matrix analysis yielded 91% and 90% overall accuracy in medium and high-density white pine blocks, respectively. These findings validate the use of UAV-DL systems as an accurate and efficient tool for operational white pine regeneration assessment, reducing the need for labor-intensive fieldwork.

The identification, delineation, and mapping of landcover is integral for resource management and planning as it establishes a baseline for thematic mapping and change detection analysis. The availability of high-resolution satellite imagery and the development of machine learning algorithms have significantly improved the prediction and accuracy of landcover classification. In this study, landcover classification is performed on seven-band Landsat 9 imagery and eight-band PlanetScope imagery for the village of Tully, NY, with an area of 900 square kilometers. The resolution of Landsat imagery is 30 meters, whereas the resolution of PlanetScope imagery is 3 meters. Classification schema is developed in ArcGIS Pro with five classification levels: conifer forest, hardwood forest, agriculture, developed, and water. Pixel-based supervised classification is performed using Support Vector Machine (SVM), Random Tress (RT), K-Nearest Neighbor (K-NN), and Maximum Likelihood Classifier (MLC). The reference dataset is acquired by an image interpreter using high-resolution imagery for map accuracy assessment. All the classification methods for Landsat imagery have more than 78% accuracy, but SVM performed best with 82% accuracy. For PlanetScope imagery, SVM performed best with 85% accuracy, whereas MLC had the lowest accuracy of 77%.