Fig. 1. Study area and Sentinel-2A acquired on November 4, 2020 (a), photos of cornresidue cover areas (b)
Fig. 2. Time-Sequence NDVI and NDRI change curves of corn planting areas in the study area in 2020
(a): Temporal NDVI variation curve; (b): NDRI variation curve during H-S2
Fig. 3. Flowchart of identification technology of corn residue cover area
Fig. 4. Spectral reflectance characteristics of different ground types (Sentinel-2A images on 4 November 2020)
(a): Corn stalk residue; (b): Rice stalk residue; (c): Water; (d): Woods
Fig. 5. Spectral reflectance characteristics of different crops (Sentinel-2A image on 22 July 2020)
(a): Corn; (b): Rice
Fig. 6. QT feature construction schematic diagram
(a): Blue wave heat map; (b): Single pixel time series data; (c): Time series data selection; (d): Quantile dataset
Fig. 7. Connected domain calibration schematic diagram
(a): Numerical diagram of calssification area;(b): Schematic diagram of connected domain of pixels
Fig. 8. Visualization results of classification based on different feature dataset
(a): The original sub-image 1_w; (b): The result of M1; (c): The result of M2; (d): The result of M3; (e): The result of M4; (f): The result of M5
Fig. 9. Combined with the connected domain calibration classification result 1
(a): The original sub-image 1_S; (b): The result of M5_1; (c): The result of M5; (d): The result of M6
Fig. 10. Combined with the classification result of connected domain calibration 2
(a): The original sub-image 2_S; (b): The original sub-image 2_W; (c): The result of M5_1; (d): The result of M5; (e): The result of M6
Fig. 11. Identification results of corn residuecover area
(a): The sub-image 1 of calssification result;(b): The sub-image 2 of calssification result
决策树数量 | Kappa/% | OA总体分类精度/% |
---|
5 | 93.97 | 95.07 | 10 | 94.54 | 95.61 | 20 | 96.65 | 95.76 | 30 | 97.41 | 97.91 | 40 | 96.98 | 97.54 |
|
Table 1. Quantitative evaluation of classification models for different decision trees
模型 | 光谱特征集 | 指数特征集 | QT特征集 | Kappa/% | OA/% |
---|
M1 | + | - | - | 91.57 | 93.22 | M2 | + | + | - | 92.89 | 94.27 | M3 | - | + | + | 94.17 | 95.31 | M4 | + | - | + | 96.12 | 96.87 | M5 | + | + | + | 97.41 | 97.91 |
|
Table 2. Classification quantitative evaluation results based on different feature dataset
| 5月—6月 | 7月 | 8月 | 9月 | 10月 | 11月 | Kappa/% | OA/% |
---|
M5_1 | - | - | - | - | - | + | 93.51 | 94.79 | M5_2 | - | - | - | - | + | + | 92.54 | 94.02 | M5_3 | - | - | - | + | + | + | 92.15 | 93.71 | M5_4 | - | - | + | + | + | + | 93.89 | 95.17 | M5_5 | - | + | + | + | + | + | 95.46 | 96.35 | M5_6 | + | + | + | + | + | + | 97.41 | 97.91 |
|
Table 3. Quantitative evaluation results based on different time scale feature classification