结合地籍数据的高密度城区面向对象遥感分类 

利用高分辨率遥感影像和GIS辅助数据,对高密度城区进行面向对象的土地利用覆被分类研究。使用NAIP高分辨率航空遥感影像,在多尺度影像分割的基础上,针对特定地物选择合适的影像分割参数。采用决策树方法建立高密度城市地区的分类规则,并结合该地区地籍图数据作为辅助数据,逐步进行高密度城市地区地物信息提取。利用辅助数据进行面向对象的遥感分类效果优于单纯依靠遥感影像进行的分类,且有效提取了道路和复杂的房屋等信息,得到了理想的分类结果,其总分类精度从常规面向对象方法的84.08%提高到89.79%。利用辅助数据进行遥感分类提高了高分辨率遥感影像的分类精度,说明了利用辅助数据进行遥感分类方法的有效性。

High-resolution Remote Sensing Classification Aided by the Auxiliary Data in High\|density Urban Area

High-resolution remote sensing images and GIS ancillary datasets such as parcels are combined to perform land use/cover change mapping in the urban\|built areas.NAIP datasets,a novel high\|resolution aerial remote sensing images in The National Agriculture Imagery Program,are used in these works.After trial and error image segmentation pursuing for good processing results,an objcet\|oriented image classification framework based on decision tree rules,combined with the cadastral datasets as a secondary data,was built to improve high\|resolution remote sensing image classification on the high\|density urban areas.The classification accuracy  of object\|oriented remote sensing image classification combined with geographic auxiliary data are better than only using the remote sensing images.Experiments studies showed that roads,building and others are excellently extracted.Comparing with the conventional object\|oriented classification,the overall classification accuracy of this novel methodology increased from 84.08% to 89.79%.Such a result reveals that auxiliary data can effectively improve the accuracy of high\|resolution remote sensing image classification.