UAV-acquired multispectral imagery for detecting quantitative parameters of urban tree lines

Monitoring of urban greenery is an important aspect in the planning, construction, management and maintenance of green areas. Challenges such as climate change, intensive urbanization and increased traffic concentration emphasize the need for effective methods of monitoring the state of greenery. Multispectral imaging ensures the quantification of light reflection from the surface of trees, bushes or grassy surfaces, defining it as a multispectral print where each pixel carries a certain value of reflection of part of the spectrum of solar radiation. This paper investigates the possibility of applying multispectral technology for precise mapping and analysis of urban green areas, as a more advanced alternative to low-resolution satellite images (0.5 m/px and 3 m/px). The aim of this work is the analysis of multispectral recordings in order to obtain a characteristic multispectral imprint of different types of trees in urban areas. Through the calculation of the relevant vegetation indices, the current physiological condition of the city's tree rows is assessed, which enables the detection of stress factors and degradation processes. Research has shown that different tree species exhibit characteristic reflection patterns in the visible and near-infrared part of the spectrum, with vegetation indices such as TVI, CVI and NIRR enabling reliable differentiation of species such as horse chestnut, linden, nettle and birch. It was also determined that recording conditions, especially the presence of shadows and light intensity, significantly affect the accuracy of the data, whereby recording in diffuse light (eg at sunset) proved to be optimal. It was concluded that multispectral data is an effective tool for detecting the physiological state of vegetation and classifying species in urban environments, which opens the possibility for more precise planning, maintenance and interventions in urban ecosystems.