Abstract
Vegetation distribution in urban areas has previously been investigated through image classification and interpretation techniques using satellite images. However, vegetation distribution and density within cities have not been extensively studied using the imagery from satellites WorldView-3 and WorldView-2 having a high resolution of 30 cm. In this study, to evaluate the distribution of trees and lawns in a city park in Tokyo, the object-oriented technique for the classification of images with different spatial resolutions was compared with the pixel-based technique. The results of comparative study showed the KNN method for the images obtained by the WorldView-3 satellite to be the most effective. In addition, the possibility of creating vegetation distribution maps for urban parks using these features was considered.
