Classifying multisensor images by support vector machine in Chongming Dongtan

Optical remote sensing (ORS) technology has been extensively used for the investigation of the environment and resources. Considering it is heavily constrained by the weather conditions, especially in the coastal zone, the round-the-clock SAR (Synthetic Aperture Radar) data are chosen to compensate for the shortcomings of optical data. In this paper, we will use the fusion image of ASAR and TM to identify five land cover types in Chongming Dongtan. And the SVM algorithm is adopted because of its capability to take numerous and heterogeneous parameters into account. Results have been shown that the fusion data of SAR and ORS is particularly suited to account for the rainy and cloudy weather in costal zone. And the SVM algorithm has attained a high level of classification performance with the overall accuracy 90.83%.