| The progress of remote sensing develops rapidly and the application is widely used in many fields, especially in agriculture. The acquisition of agricultural information by remote sensing has become the main mean. Previous crop information is mostly planting concentration, block structured and easy to distinguish with the surrounding crops, for example, winter wheat, cotton etc. There is less study on the planting area dispersion, extremely heterogeneous spatial distribution. Due to plant dispersion and confused with other crops of Cassava, the acquisition of planting information and spatial distribution has become the technical bottleneck for the correct assessment of cassava ethanol biofuel.High spatial resolution image has higher resolution and more detailed textures than low resolution images. Therefore the traditional classification method based on pixel has been unable to meet the need of the application. It will take a large amount of time and labor to extract information for massive data. Given all of the above, the extraction of Cassava planting information is finished by combing the high resolution image with moderate resolution image based on object oriented method in the paper. Object oriented classification method cannot only overcome the defects on the basis of traditional pixel classification, but also get higher classification accuracy by simulating human thinking mode, the integration of multi-source data and the use of special data make the result more practical.Taking Wu Ming County in Guangxi province as research area,object oriented classification method is used to extract cassava planting area and spatial distribution in 2012 and 2009, high resolution image(RapidEye) and moderate resolution image (TM) as data source based on the statistical data, field survey data and land use data. The results show that the accuracy of classification method based on object oriented in high resolution image is much higher than the results based on pixel, it can also get ideal classification accuracy in moderate resolution image. In accordance with classification results of high resolution image, by using the object-oriented classification method, the medium resolution image information is extracted, it provide an effective reference method in the extraction of crop planting area. |
