A processing method for rice crop inventory using multi-date ENVISAT-1 ASAR data

摘要

Synthetic Aperture Radar (SAR) is anticipated to be the dominant high-resolution remote sensing data source for agricultural applications in tropical and subtropical regions due to its independent from cloud cover. ENVISAT-1 ASAR is the most advanced satellite radar-imaging instrument, its capabilities include beam steering for acquiring images with different incidence angles, duel polarization and wide swath coverage. Agricultural crop inventory based on remote sensed data will be improved greatly by ASAR's new capabilities. In this paper, a procedure has been developed using multi-date ASAR data for rice crop inventory. The procedure comprises two parts: data preprocessing and classification of multi-date data for rice field. In order to carry out the research, 6 scenes of ASAR images covering Fuzhou area year round of 2004 were used. PCI 9.1 is used for data preprocessing which includes data calibration, image co-registration, speckle suppression, orthorecitification and amplitude-to-dB conversion. Some novel methods are applied in this procedure such as correlation matching for image co-registration and multi-channel filtering for speckle suppression. Object-oriented classifier was used, compared with K-means supervised classifier and maximum likelihood classifier, and higher classification accuracy was achieved. By adopting the procedure presented in this paper, more than 90% classification accuracy for rice was achieved in Fuzhou city with multi-date Envisat ASAR data. This indicates that the procedure is feasible for rice crop inventory using multi-date ASAR data.