Study on urban green space extraction from QUICKBIRD imagery based on decision tree

Abstract

The green space information extraction is significant for study on urban remote sensing because of its ecological and socioeconomic functions. And with the traditional methods the precision is very low, we educe decision tree of extraction to raise the accuracy rate. We take Shanghai as a case study area from QUICKBIRD imagery, analyzing spectral characteristics of 4 multi-spectral bands. And we also introduce accessorial information vegetation index (NDVI) to get rid of other land use types. The final result suggests that the extraction method of decision tree can accurately get the green space information. The accurate rate is over 95%. It can be widely used in the field of urban vegetation remote sensing analyses.