基于连续小波系数的叶绿素a浓度估测模型 
下载： 956次
Estimation Model of Chlorophyll-a Concentration Based on Continuous Wavelet Coefficient
1 东华理工大学测绘工程学院, 江西 南昌 330013
2 流域生态与地理环境监测国家测绘地理信息局重点实验室, 江西 南昌 330013
3 广东省遥感与地理信息系统应用重点实验室, 广东省地理空间信息技术与应用公共实验室, 广东省遥感大数据应用工程技术研究中心, 广州地理研究所, 广东 广州 510070
图 & 表
图 1. 采样区域与采样点分布
Fig. 1. Sampling areas and sampling stations
下载图片 查看原文
图 2. 不同数据集中叶绿素a浓度的统计特征
Fig. 2. Statistical characteristics of chlorophyll-a in different datasets
下载图片 查看原文
图 3. 水体反射率曲线及其单波段相关性分析
Fig. 3. Analysis of water reflectance curve and its single-band correlation
下载图片 查看原文
图 4. RSI和NDCI相关系数矩阵图。(a) RSI;(b) NDCI
Fig. 4. Diagram of RSI and NDCI correlation coefficient matrix. (a) RSI; (b) NDCI
下载图片 查看原文
图 5. db4小波系数与叶绿素a浓度的相关系数图
Fig. 5. Correlation coefficient diagram of db4 wavelet coefficient and chlorophyll-a concentration
下载图片 查看原文
图 6. 基于sym6小波变换估测的叶绿素a浓度与实测值
Fig. 6. Predicted chlorophyll-a concentration based on sym6 wavelet transform and measured value
下载图片 查看原文
表 1连续小波变换的10个选定母小波基及其应用领域
Table1. Ten selected mother wavelet bases of continuous wavelet transform and their application fields
| Mother wavelet | Research object | Reference |
|---|
| rbio3.3bior3.3sym7db1 | Maize (chlorophyll, carotenoids) | Wang Z L, et al. (2020)[20] | | db5haarmexhmorl | Winter wheat (chlorophyll) | He R Y, et al. (2018)[21] | | mexh | Zizania caduciflora (chlorophyll) | Yu Z X, et al. (2018)[22] | | db4 | Broadleaf shrub,small tree (chlorophyll) | Fang S H, et al. (2015)[23] | | sym6 | Maize (chlorophyll) | Liu H J, et al. (2018)[24] |
|
查看原文
表 2采样点叶绿素a浓度的统计特征
Table2. Statistical characteristic of chlorophyll-a concentration at sampling points
| Sample set | Number /(μg·L-1) | Minimum /(μg·L-1) | Median /(μg·L-1) | Maximum /(μg·L-1) | Mean /(μg·L-1) | SD /(μg·L-1) | CV /(μg·L-1) |
|---|
| Calibration dataset | 48 | 0.434 | 13.100 | 50.338 | 18.510 | 14.500 | 0.783 | | Validation dataset | 21 | 0.482 | 9.400 | 44.873 | 17.706 | 16.771 | 0.947 | | All dataset | 69 | 0.434 | 12.700 | 50.338 | 18.265 | 15.106 | 0.827 |
|
查看原文
表 3叶绿素a浓度的常见反演模型及其建模精度
Table3. Common chlorophyll-a concentration inversion models and their modeling accuracy
| Spectral index | Model | Variable | r | R2 | RMSE/ (μg·L-1) |
|---|
| Reflectance | R1 | R1=712 nm | 0.328 | 0.153 | 15.063 | | RSI | | R1=654 nmR2=644 nm | 0.721 | 0.644 | 9.760 | | NDCI | | R1=658 nmR2=632 nm | 0.721 | 0.669 | 9.414 | | Three-band | (-)×R3 | R1=649 nmR2=653 nmR3=675 nm | 0.736 | 0.680 | 9.254 |
|
查看原文
表 4CWT-PLSR模型的精度评价
Table4. Accuracy evaluation of CWT-PLSR model
| Mother wavelet | PC | | RMSEC/ (μg·L-1) | | RMSEP/ (μg·L-1) | RPD |
|---|
| rbio3.3 | 3 | 0.700 | 7.864 | 0.635 | 10.830 | 1.551 | | bior3.3 | 4 | 0.551 | 9.613 | 0.662 | 9.545 | 1.759 | | sym7 | 4 | 0.593 | 9.156 | 0.717 | 8.725 | 1.924 | | db1 | 5 | 0.656 | 8.418 | 0.367 | 14.702 | 1.142 | | db5 | 5 | 0.727 | 7.491 | 0.725 | 9.530 | 1.762 | | haar | 9 | 0.692 | 7.966 | 0.399 | 14.472 | 1.160 | | mexh | 1 | 0.507 | 10.077 | 0.560 | 10.955 | 1.533 | | morl | 4 | 0.756 | 7.081 | 0.812 | 8.235 | 2.039 | | db4 | 7 | 0.779 | 6.751 | 0.668 | 10.974 | 1.530 | | sym6 | 3 | 0.732 | 7.431 | 0.732 | 6.457 | 2.600 |
|
查看原文
彭咏石, 陈水森, 陈金月, 赵晶, 王重洋, 官云兰. 基于连续小波系数的叶绿素a浓度估测模型[J]. 激光与光电子学进展, 2021, 58(8): 0828002. Yongshi Peng, Shuisen Chen, Jinyue Chen, Jing Zhao, Chongyang Wang, Yunlan Guan. Estimation Model of Chlorophyll-a Concentration Based on Continuous Wavelet Coefficient[J]. Laser & Optoelectronics Progress, 2021, 58(8): 0828002.
PDF全文
