Author Affiliations
Key Laboratory of Virtual Geographic Environment, Ministry of Education, Nanjing Normal University, Nanjing 210023, Chinashow less
Fig. 1. The geographical location and surrounding environment of the study area
Fig. 2. Example of sampling point collection
Fig. 3. Spectral characteristics of non-shadow water, shadow and vegetation on water surface
Fig. 4. The scatter diagram of reflectance at 666 nm/791 nm and 492 nm
Fig. 5. Umbra, penumbra, non-shadow water image and reflectance spectral comparison
Fig. 6. The process of surface shadow detection
Fig. 7. The shadow detection result of test scenario 1
Fig. 8. The shadow detection result of test scenario 2
Fig. 9. The shadow detection result of test scenario 3
Fig. 10. The shadow detection result of test scenario 4
Fig. 11. Building shadow and tree shadow recognition effect
Fig. 12. Analysis on the effect of floating objects on water shadow detection
Fig. 13. The results of shadow detection when divided into 2 and 3 categories
Fig. 14. Black and odor water recognition result in test scenario 3 before and after shadow mask
序号 | 仪器设备名称 | 主要技术指标 |
---|
1 | 大疆M600 pro多旋翼无人机 | 旋翼数量:6个轴距:1133 mm起飞重量:15.5 kg最大载荷:6 kg 最大水平飞行速度:65 km/h(无风环境)最大续航时间:60 min(空载)18 min(5.5 kg载重)实用升限:4500 m | 2 | ZK-VNIR-FPG480机载高光谱成像仪 | 光谱范围:400~1000 nm光谱分辨率:2.8 nm光谱通道数:270个空间通道数:480个空间分辨率:0.08 m(120 m高度)(35 mm镜头)视场角:26°A/D转换:12 bits最大帧频:100 fps数据接口:GigE最大功耗:20 W外形尺寸:310×87×87 mm重量:2.2 kg成像方式:采用外置推扫连续成像,采集画幅无限制,扫描路线一次成图,影像无畸变 |
|
Table 1. Main technical parameters of the instrument and equipment
区域 | 生产者精度 | | 用户精度 | | 总体精度 | Kappa系数 |
---|
| | | | |
---|
测试场景1 | 85.27 | 82.89 | | 93.25 | 67.00 | | 86.89 | 0.83 | 测试场景2 | 89.57 | 82.26 | 89.05 | 82.98 | 87.30 | 0.88 | 测试场景3 | 85.70 | 95.22 | 96.74 | 91.78 | 90.38 | 0.95 | 测试场景4 | 90.04 | 96.49 | 84.56 | 91.15 | 88.61 | 0.92 |
|
Table 2. Accuracy evaluation of each test scenario
测试场景 | 阴影提取前 | | 阴影提取后 |
---|
黑臭水体像元数/个 | 正常水体像元数/个 | 黑臭像元占比/% | 正常水体像元占比/% | 黑臭水体像元数/个 | 正常水体像元数/个 | 黑臭像元占比/% | 正常水体像元占比/% |
---|
测试场景1 | 16 859 | 5622 | 74.99 | 25.01 | | 7235 | 140 | 98.10 | 1.90 | 测试场景2 | 27 202 | 7450 | 78.50 | 21.50 | 13 040 | 112 | 99.15 | 0.85 | 测试场景3 | 13 612 | 12 885 | 51.37 | 48.63 | 10 445 | 533 | 95.14 | 4.86 | 测试场景4 | 17 827 | 6106 | 74.49 | 25.51 | 12 130 | 928 | 92.89 | 7.11 |
|
Table 3. The pixel proportion of black and odor water identified before and after shadow extraction