Extraction of urban vegetation from high resolution remote sensing image

The extraction of urban vegetation information is a focal study point of the city remote sensing. To address the limitations of urban regional scale and the features of extraction of urban vegetation from high resolution satellite image based on object-oriented approach, this paper presented a new approach to use segmentation of high-resolution remote sensing image and the fuzzy classification technique based on multi-thresholds method, and then forests, thin grassland, thick grassland were extracted accurately. The new object-based method performances were assessed using Kappa coefficients and overall accuracy. High accuracy (93.72%) and overall Kappa coefficient (0.8236) were achieved by this new method using Quickbird image; the experimental results demonstrate the new approach is simple for computation in urban regional scale.