Winter wheat planting area extraction using multi-temporal remote sensing data based on field parcel characteristic

Winter wheat area estimation by mean of remote sensing is one of important research content in agriculture information monitoring field. It is very important for yield estimation to obtain planting area timely and accurately. This paper proposed a new method for extraction winter wheat area based on object oriented technology used multi-sources and multi-temporal remote sensing data. Field parcel information was extracted from SPOT5 panchromatic and multi-temporal fusion data. Multi-temporal ETM+ images were segmented assistant by the thematic layer of field parcel at the same segmentation scale, and then, the segmentation results with the same shape feature were acquired to construct winter wheat extraction model based on dynamic spectral feature rule set. Through the cross validation, the winter wheat area was extracted. The experiment results show that the method can rapidly recognize winter wheat area. The overall accuracy of the result is over 90%，which can meet the calculation precision requirements of agriculture information remote sensing monitoring. Therefore, the method presented in the paper is practicable and can be widely used for winter wheat area quick extraction in future.