Uso e aplicação de geotecnologias para estudo da arborização de vias públicas no município de São Pedro do Sul, RS

Abstract

The diagnosis of street afforestation is an essential element in urban forest planning. In this way, it becomes necessary to develop methods to obtain qualitative and quantitative variables from the trees present in cities. Therefore, the present study aims to evaluate the application of tools based on geotechnologies to characterize and obtain dendrometric variables of street afforestation in the municipality of São Pedro do Sul, RS. Initially, it started with a census of the trees on the municipality's public roads. Then, the collection of the geographic position and the popular name of all the municipality sidewalks trees occurred for this purpose. Finally, the data were tabulated in Microsoft Excel® to create a table containing the individual's family, scientific name, and geographic position. After that, the aerial images collection happened in randomly selected streets, using the Remotely Piloted Aircraft System (RPAS) model Phantom 4 Pro®. Pix4D Mapper® software generated the following data: Orthophotomosaic, Digital Surface Model (MDS). The computer application QGIS version 3.10 generated the Digital Elevation Model (DEM) and collected data on the trees' height and their respective canopy area through photo interpretation. The Meanshift algorithm present in the ArcGIS Pro® computer application performed the automatic segmentation of the crowns. The height and canopy data inventory comparison happened using the conventional method and image analysis. RStudio software version 4.1.0 performed the data comparison and descriptive analysis. Data were validated using linear regression analysis, and the models were compared using Root Mean Square Error (RMSE) and Mean Absolute Error (MAE). The results indicate that using geotechnologies and field data allowed municipal street trees' spatial and diversity characterization. Regarding the variables analyzed, the methods showed no significant difference for the Meliaceae family (p-value > 0.05) and for the species Handroanthus chrysotrichus and Melia azedarach, (pvalue > 0.05). In addition, Fabaceae and Bignoneaceae families had a higher average height than the other families. The canopy area values were higher for the conventional method (p-value < 0.05), differing from the other methods. Automatic segmentation showed no significant difference compared to photo interpretation (p-value > 0.05). The conclusion says that the use of geotechnologies helps diagnose the urban forest, enabling spatial characterization and obtaining essential quantitative variables for planning. In addition, the use of data obtained by RPAS and image processing techniques are rapidly and constantly expanding, demonstrating efficiency and potential for use in road afforestation management.