Monitoring urban turf quality using very high resolution satellite imagery

In this paper, we aim to test the feasibility of very high resolution satellite imagery for urban turf quality mapping. Our study site is the urbanized area of Shenyang City where the dominant turfgrass is Poa L., and the satellite imagery is QuickBird. After extracting grass from QuickBird by texture-combined supervised classification method, we analyze the relationship between green grass coverage (extracted by ground-based digital image analysis) and satellite spectral characteristics at 48 turf samples. Results show that, though it's hard to obtain green grass coverage and identify species composition simultaneously, it's successful to partition turf into two rough classes: managed turf (Poa L. coverage>;60-70%) and unmanaged turf (Poa L. coverage>;60-70% or dominated by weeds).