基于二次分类的葡萄干晾房遥感信息挖掘及反演

以鄯善绿洲为研究靶区,应用遥感技术对葡萄干晾房遥感影像特征进行光谱和纹理分析,使用二次分类方法提取葡萄干加工晾房信息。结果表明,晾房光谱特征由晾房顶部和阴影组成,纹理特征因具有阴影而增强;葡萄干晾房斑块共有701块,占地面积3 120hm2,最大斑块面积178hm2,最小斑块面积338.76m2,平均面积4.34hm2;建筑单体18 819栋,建筑面积261hm2。经验证,一次分类精度为84.06%,Kappa系数为0.80;二次分类精度为88.66%,Kappa系数为0.84。

Data Mining and Retrieval of Remote Sensing Information in Raisin Room Based on Secondary Classification

Taking oasis in Shanshan area as research target,remote sensing images of rooms for drying raisin were analyzed in spectrum and texture,information about the rooms for trying raisins was extracted by secondary classification method during raisin processing.Result showed that the spectral characteristics of drying rooms were composed of hanged roof and shadow,and the texture feature was enhanced by the shadow; plaques of raisin drying rooms were 701 pieces together,which covered 3 120 hm²,the patch area ranged from 338.76 m² to 178 hm²,and the average area was 4.34 hm²; single buildings were 18 819 with building area of 261 hm². Verification indicated that the first classification accuracy was 84.06% with the Kappa coefficient 0.80,and the secondary classification accuracy was 88.66% with the Kappa coefficient 0.84.