Mc. Dobson et al., KNOWLEDGE-BASED LAND-COVER CLASSIFICATION USING ERS-1 JERS-1 SAR COMPOSITES/, IEEE transactions on geoscience and remote sensing, 34(1), 1996, pp. 83-99
Land-cover classification of an ERS-1/JERS-1 composite is explored in
the context of regional- to global-scale applicability, Each of these
orbiting synthetic aperture radars provide somewhat complementary info
rmation since data is collected using significantly different frequenc
ies, polarizations, and look angles (ERS-1: C-band, VV polarization, 2
3 degrees; JERS-1: L-band, HH polarization, 35 degrees), This results
in a classification procedure for the composite image (a co-registered
pair from the same season) that is superior to that obtained from eit
her of the two sensors alone. A conceptual model is presented to show
how simple structural attributes of terrain surfaces and vegetation co
ver relate to the data from these two sensors, The conceptual model is
knowledge-based; and it is supported by both theoretical consideratio
ns and experimental observations, The knowledge-based, conceptual mode
l is incorporated into a classifier that uses hierarchical decision ru
les to differentiate land-cover classes, The land-cover classes are de
fined on the basis of generalized structural properties of widespread
applicability, The classifier operates sequentially and produces two l
evels of classification, At Level-1, terrain is structurally different
iated into man-made features (urban), surfaces, short vegetation, and
tall vegetation, At Level-2, the tall vegetation class is differentiat
ed on the basis of plant architectural properties of the woody stems a
nd foliage, Growth forms of woody stems include excurrent (i.e., pines
), decurrent (i.e., oaks), and columnar (i.e., palm) architecture, Two
classes of leaves are considered: broadleaf and needle-leaf, The comp
osite classifier yields overall accuracies in excess of 90% for a test
site in northern Michigan located along the southern ecotone of the b
oreal forest, For the area examined, the SAR-based classification is s
uperior to unsupervised classification of multitemporal AVHRR data sup
plemented with a priori information on elevation, climate, and ecoregi
on.