Assessing the Potential of Multi-Seasonal High Resolution Pléiades Satellite Imagery for Mapping Urban Tree Species

Abstract

The multi-season Pleiades image data were compared and analyzed for their capabilities of classifying and mapping the seven urban forest tree species in the City of Tampa, FL, USA to understand the seasonal effect on tree species classification accuracy. The seven species and groups included sand live oak (Quercus geminata), laurel oak (Q. laurifolia), live oak (Q. virginiana), pine (species group), palm (species group), camphor (Cinnamomum camphora), and magnolia (Magnolia grandiflora). A multi-level classification system was adopted to classify image objects of the tree species. Shade image objects (IOs) were spectrally normalized to similar sunlit IOs, and the tree species fractions were extracted from the seasonal images using a spectral unmixing approach and used as additional features. Using selected features extracted from the five individual season and the two dry-wet season combined Pleiades images, tree species were identified and mapped using a random Forest, support vector machines and a linear discriminant analysis classifiers. The experimental results indicate significantly improved tree species mapping accuracies using late spring season (April) image compared to all other seasonal images (p<0.01), and combined dry-wet season images performed even better. Results suggest a significant seasonal effect on tree species classification. The novel significance of this study was to demonstrate the potential of multi-season Pleiades imagery in improving urban tree species mapping. Therefore, in practice, it is important to choose appropriate seasonal remote sensing data for mapping tree species.