基于卷积神经网络和高分辨率影像的湿地群落遥感分类——以洪河湿地为例
|
孟祥锐, 张树清, 臧淑英
|
Remote Sensing Classification of Wetland Communities Based on Convolutional Neural Networks and High Resolution Images: A Case Study of the Honghe Wetland
|
Xiangrui Meng, Shuqing Zhang, Shuying Zang
|
|
表3 CNN、TSP-SVM和SP-SVM方法在研究区A的分类精度对比(%) |
Table 3 Classification accuracy evaluation of CNN, TSP-SVN and SP-SVM in study site A (%) |
|
| 分类 | 水体 | 林地 | 旱生杂草 | a | b | c | d | CNN | 用户精度 | 92.16 | 99.83 | 72.97 | 99.32 | 71.28 | 81.20 | 99.05 | 制图精度 | 90.28 | 89.91 | 82.82 | 96.75 | 82.56 | 86.30 | 77.32 | 分类精度 | 91.22 | 94.87 | 77.89 | 98.04 | 76.92 | 83.75 | 88.19 | TSP-SVM | 用户精度 | 89.99 | 99.69 | 55.09 | 93.57 | 76.62 | 80.56 | 100 | 制图精度 | 92.52 | 88.66 | 78.20 | 96.33 | 74.33 | 84.72 | 61.91 | 分类精度 | 91.25 | 94.17 | 66.65 | 94.95 | 75.47 | 82.64 | 80.95 | SP-SVM | 用户精度 | 90.05 | 99.30 | 54.69 | 90.92 | 70.68 | 80.49 | 99.37 | 制图精度 | 89.87 | 88.39 | 74.92 | 95.85 | 73.53 | 83.72 | 62 | 分类精度 | 89.96 | 93.85 | 64.81 | 93.39 | 72.11 | 82.11 | 80.69 |
|
|
|