Signature research for crop classification by active and passive microwaves

Abstract

During the last six years we have carried out a series of ground-based microwave measurements on agricultural fields. Bare soils, corn (maize), wheat, sugar-beet, potatoes, grass, canola (rape), oat, and rye were observed under various conditions and phenological stages. Our calibrated radiometer-scatterometer system, RASAM, operates at frequencies between 3 and 11 GHz. The brightness temperature was measured at horizontal and vertical polarization, and the back scattering coefficient at HH, W, HV, and VH polarization, all at incidence angles from 0° to 70°. In addition we measured soil and vegetation parameters. The data are stored in our signature data base.

In this article the signature data base is reviewed with respect to crop classification. We find that between mid-June and mid-July it is possible to distinguish the dominant crops of Switzerland (cereals, corn, sugar-beet, potatoes, canola and grass) with two carefully chosen active or passive microwave parameters. The use of multi-temporal data for crop classification is discussed. Microwave remote sensing is able to collect multi-temporal data reliably due to its relative independence of weather conditions. Finally, the signature data base is used to illustrate the signatures of agricultural fields for the system configuration that will be used by the active microwave instrument on the ERS-1 satellite.

Notes

† Present address: JPL/CALTECH, MS 300-233, 4800 Oak Grove Drive, Pasadena, CA91109, U.S.A.