Object-based Classification of Urban Plant Species from Very High-Resolution Satellite Imagery

Urban vegetation, in particular trees, are essential as they provide services in terms of air pollution mitigation, freshness, biodiversity, and citizens’ well-being. Accurate data on location, species, and structural characteristics are essential for quantifying their benefits. However, the cost of measuring thousands of individual trees through field campaigns can be prohibitive and reliable information on domestic gardens is lacking due to difficulties in acquiring systematic data.

The main objective of this study was to investigate the suitability of very-high resolution satellite imagery, e.g., WorldView-2, for detecting, delineating, and classifying urban vegetation in both public and private areas. The characterization of urban vegetation is difficult due to the complexity of the urban environment (buildings, shadows, open courtyards), the diversity of species and the spatial proximity between trees.

To overcome these constraints, an object-based classification was developed with the selection of new relevant spectral and texture-based features for each plant species. Four spectral bands (blue, green, yellow, red) and four texture features (i.e., energy, entropy, inverse difference moment, Haralick correlation) were found to be the most efficient attributes for object-based classification from WV-2 images. Then, a classification of plant species, by using a Random Forest classifier, and ground validation were performed.

In the two study areas, Aix-en-Provence (France) and Florence (Italy), 22 and 20 dominant tree species, and grassland, were identified and classified with an overall accuracy of 84% and 83%, respectively.