Tree Cover for the Year 2010 of the Metropolitan Region of São Paulo, Brazil

Abstract

Mapping urban trees with images at a very high spatial resolution (≤1 m) is a particularly relevant recent challenge due to the need to assess the ecosystem services they provide. However, due to the effort needed to produce these maps from tree censuses or with remote sensing data, few cities in the world have a complete tree cover map. Here, we present the tree cover data at 1-m spatial resolution of the Metropolitan Region of São Paulo, Brazil, the fourth largest urban agglomeration in the world. This dataset, based on 71 orthorectified RGB aerial photographs taken in 2010 at 1-m spatial resolution, was produced using a deep learning method for image segmentation called U-net. The model was trained with 1286 images of size 64 × 64 pixels at 1-m spatial resolution, containing one or more trees or only background, and their labelled masks. The validation was based on 322 images of the same size not used in the training and their labelled masks. The map produced by the U-net algorithm showed an excellent level of accuracy, with an overall accuracy of 96.4% and an F1-score of 0.941 (precision = 0.945 and recall = 0.937). This dataset is a valuable input for the estimation of urban forest ecosystem services, and more broadly for urban studies or urban ecological modelling of the São Paulo Metropolitan Region.

LandCover.ai: Dataset for Automatic Mapping of Buildings, Woodlands, Water and Roads from Aerial Imagery

“…Datasets made for buildings or vegetation segmentation are useful but not entirely sufficient in our case. Those datasets include the Massachusetts Buildings Dataset [26], the Inria Aerial Image Labeling Dataset [25], the AIRS Automatic Mapping of Buildings Dataset [8], the Agriculture-Vision a Large Aerial Image Database for Agricultural Pattern Analysis [9] and the Tree Cover dataset for the year 2010 of the Metropolitan Region of So Paulo [33].…”

“…Table 1 presents the comparison of the statistics between the proposed dataset and closely related aerial datasets: Inria [25], AIRS [8], Massachusetts [26], Agriculture-Vision [9] and Tree Cover [33]. The older ones have a worse resolution (Tree Cover, Massachusetts).…”

Urban Tree Canopy Mapping Based on Double-Branch Convolutional Neural Network and Multi-Temporal High Spatial Resolution Satellite Imagery

“…Martins et al applied FCN, U-NET, Seg-Net, DeepLabV3+ and a dynamic dilated convolution network to the Campo Grande area for urban tree canopy mapping and obtained IOU ranging from 70.01% to 73.89% [27]. Wanger et al applied the U-NET model to São Paulo, Brazil, and obtained an overall accuracy of approximately 96% [57]. Existing studies have shown that U-NET is the fully convolutional neural network that currently performs the best for urban tree canopy mapping tasks after comprehensively considering indicators such as accuracy and the number of model parameters [27,58].…”