Land cover information extraction based on multi-temporal HJ-1 satellite images

We proposed a remote sensing image classification method based on the object´s phenological characteristics as auxiliary information: obtained the NDVI time series data after processing the CCD images, then, got the object´s NDVI time series curve as the phonological information, based on this, selected classification training area of various objects so as to complete image classification, afterwards, used random sampling method to evaluate the precision of classification results. The evaluation demonstrated that this method could distinguish various land cover types exactly, besides, had high classification accuracy. The method was a viable remote sensing image classification and can be applied in large scale land cover classification.