高光谱遥感影像纹理特征提取的对比分析

邵文静,孙伟伟,杨刚

Comparison of Texture Feature Extraction Methods for Hyperspectral Imagery Classification

Wenjing Shao,Weiwei Sun,Gang Yang

表2 帕维亚大学分类精度 (%)

Table 2 Classification accuracy of Pavia University data

地物类别 纹理特征提取方法/% riLBP SLIC EMP DMP AP 3D-Gabor JBF GF SVM 柏油路 91.45 95.24 96.63 92.18 97.96 91.18 97.82 96.73 88.69 草地 99.25 99.51 99.85 99.45 99.68 95.92 99.81 100 96.05 沙砾 94.63 77.31 98.44 88.47 94.52 64.99 73.63 65.42 47.16 树木 60.16 90.51 97.30 95.30 93.81 90.35 89.84 83.68 75.35 金属板 93.18 98.35 99.86 96.53 98.63 98.65 100 100 99.78 裸地 98.69 99.16 99.93 98.43 98.49 74.85 79.15 71.67 53.72 沥青屋顶 87.56 90.87 95.08 94.76 99.71 71.16 75.63 77.71 59.18 地砖 95.41 88.37 95.06 92.76 98.01 83.26 97.78 98.44 79.34 阴影 45.18 90.30 89.91 82.03 96.65 95.48 98.62 99.77 80.91 OA 92.87 95.60 98.33 96.20 98.36 89.01 92.88 93.04 83.30 AA 85.06 92.18 96.90 93.32 97.50 85.09 90.25 88.16 75.76 KC 90.49 94.16 97.78 94.94 97.82 85.31 81.55 80.77 77.45