Mapping the distribution of mistletoe in Uludağ fir stands using uavs integrated with multispectral cameras (kökez forest management chiefdom)

This master's thesis aims to determine the spatial distribution of mistletoe (Viscum album) in Uludağ Fir (Abies nordmanniana subsp. bornmülleriana) forests within the Kökez Forest Management Chiefdom, located in the Bolu Aladağ Forest Enterprise Directorate. Mistletoe is a parasitic plant that extracts water and nutrients from its host trees, adversely affecting their physiological processes and limiting their growth and development. The precise mapping of such parasitic species is crucial for the sustainable management of forest ecosystems. However, traditional field-based methods for detecting and monitoring mistletoe are labor-intensive, time-consuming, and challenging to implement over large areas. In this context, remote sensing technologies offer an innovative and effective approach for monitoring and managing invasive plant species in forest ecosystems. This study employs Unmanned Aerial Vehicles (UAVs) equipped with multispectral cameras to detect and map the distribution of mistletoe. A DJI Mavic 3M UAV was selected, capturing RGB imagery alongside multispectral data in red (R), green (G), red edge (RE), and near-infrared (NIR) bands. The UAV flights were conducted at an altitude of 100 meters, and the acquired data were processed and analyzed using Agisoft Metashape Professional, ArcGIS, and eCognition software. The image processing workflow included orthophoto generation, vegetation index computation, and spatial analysis. Normalized Difference Vegetation Index (NDVI) and other spectral indices were utilized to assess the presence of mistletoe; however, no statistically significant results were obtained. The study compares pixel-based and object-based classification methods. Pixel-based classification assigns each pixel to specific classes based on spectral properties, but it was found to be less effective due to noise and misclassification, particularly in high-resolution imagery. Conversely, object-based classification, which groups neighboring pixels using segmentation algorithms, yielded more accurate results, particularly for high-resolution datasets. The findings indicate that the Kappa coefficient for pixel-based classification ranged between 0.28 and 0.56, while object-based classification achieved a Kappa coefficient between 0.69 and 0.89, demonstrating its superior reliability and accuracy in mistletoe detection.This research underscores the effectiveness of UAVs and multispectral imaging as a valuable tool for detecting and mapping parasitic plants in forest ecosystems. The results highlight the advantages of object-based classification in providing higher accuracy and greater efficiency compared to traditional methods. The findings contribute valuable insights for forest management and conservation, offering a framework for future ecological monitoring and remote sensing applications in the detection and assessment of parasitic species