Table 4.

Details of the OBIA-based LULC classification for the Sydney use case.

Satellite images	Resolution of 1.6 m
Segmentation parameters	Scale of 30, shape index of 0.8 and compactness of 0.5
LULC classes	Grass, Trees, Algae, Roads, Water body, Built up area, Bare soil
Features and algorithms	Shape indexes, GLCM textural parameters, normalized difference vegetation index (0.24> and <0.3), ratio of green (<0.3), length/width (0.9>), rectangular fit indexes (1.3–1.6 and 0.3–0.05), shape indexes, GLCM textural parameters, normalized difference vegetation index (0.3> and <0.8), ratio of green (0.4>), brightness (135>), length/width (0.9>), rectangular fit (1.2–1.5), mean (1.6>)
Classification algorithm	Sample-based supervised classification based on nearest neighbor
Accuracy assessment	Control points for the error matrix and to calculate the Kappa and QADI