基于无人机高光谱影像与机器学习算法的汤逊湖水质参数反演研究
PDF下载 (231)郭亚会,周 伟,李松泽,马瑞晨,高敬丹,张 璇,郝芳华.基于无人机高光谱影像与机器学习算法的汤逊湖水质参数反演研究[J].宁波大学学报(理工版),2025,38(5):20-30.DOI:10.20098/j.cnki.1001-5132.2025.0522
GUO Yahui,ZHOU Wei,LI Songze,MA Ruichen,GAO Jingdan,ZHANG Xuan,HAO Fanghua.Water quality parameter inversion in Tangxun Lake using UAV hyperspectral imagery and machine learning algorithms[J].Journal of Ningbo University(Natural Science & Engineering Edition),2025,38(5):20-30.DOI:10.20098/j.cnki.1001-5132.2025.0522
| Title: | Water quality parameter inversion in Tangxun Lake using UAV hyperspectral imagery and machine learning algorithms |
| 作者: | 郭亚会, 周 伟, 李松泽, 马瑞晨, 高敬丹, 张 璇, 郝芳华 |
| Author(s): | GUO Yahui, ZHOU Wei, LI Songze, MA Ruichen, GAO Jingdan, ZHANG Xuan, HAO Fanghua |
| 关键词: | 无人机遥感; 高光谱影像; 水质参数反演; 机器学习算法 |
| Keywords: | unmanned aerial vehicle remote sensing; hyperspectral imagery; water quality parameter inversion; machine learning algorithms |
| 分类号: | X824；TP79 |
| DOI: | 10.20098/j.cnki.1001-5132.2025.0522 |
| 文献标识码: | A |
| 摘要: | 及时准确地开展城市湖泊水质监测，对于保护城市生态环境、保障居民健康和促进可持续发展具有重要意义。结合无人机高光谱影像与机器学习算法，开展武汉市汤逊湖的水质参数反演研究，分别使用误差反向传播神经网络（Backpropagation Neural Network，BP）、随机森林（Random Forest，RF）、极限梯度提升（eXtreme Gradient Boosting，XGBoost）和支持向量机（Support Vector Machine，SVM）构建叶绿素a（Chla）、总磷（TP）、总氮（TN）和悬浮物（SS）的反演模型。结果表明：（1）Chla、TP、TN和SS反演的最优机器学习算法分别为BP（R²=0.949，RMSE=9.337μg/L）、RF（R²=0.948，RMSE=0.013mg/L）、XGBoost（R²=0.928，RMSE=0.095mg/L）、XGBoost（R²=0.875，RMSE=4.052mg/L）；（2）Chla最佳反演模型为BP神经网络结合高光谱全波段，TP最优反演模型为RF结合高光谱敏感波段，TN和SS的最优反演模型为XGBoost结合高光谱敏感波段。研究结果可为城市湖泊水质动态监测和管理提供技术支撑，有助于推动城市水环境的精细化管理和保护。 |
| Abstract: | Timely and accurately monitoring of urban lake water quality is crucial for protecting urban ecosystems, safeguarding public health, and promoting sustainable development. This article takes the water quality parameter inversion of a typical lake area in Tangxun Lake, Wuhan as the research object, and conducts research on water quality parameter inversion based on unmanned aerial vehicle (UAV) hyperspectral images and machine learning algorithms. Error backpropagation (BP) neural network, Random Forest (RF), eXtreme Gradient Boosting (XGBoost) and Support Vector Machine (SVM) are used to construct the inversion models for Chlorophyll-a (Chla), Total Phosphorus (TP), Total Nitrogen (TN), and Suspended Sediment (SS), respectively. The main conclusions are as follows: (1) The optimal models for Chla, TP, TN, and SS are BP (R²=0.949, RMSE=9.337μg/L), RF (R²=0.948, RMSE=0.013mg/L), XGBoost (R²=0.928, RMSE=0.095mg/L), and XGBoost (R²=0.875, RMSE=4.052mg/L), respectively; (2) Comparative analysis reveals that BP and RF performed best for Chla and TP inversion, while XGBoost demonstrated superior performance for TN and SS inversion. The research results of this article provide technical support for the dynamic monitoring and management of urban lake water quality, which help to promote the refined management and protection of urban water environment. |
| 参考文献 /References: | [1] PEKEL J F, COTTAM A, GORELICK N, et al. High-resolution mapping of global surface water and its long-term changes[J]. Nature, 2016, 540(7633):418-422. [2] HOEKSTRA A Y, BUURMAN J, VAN GINKEL K C H. Urban water security: a review[J]. Environmental research letters, 2018, 13(5):053002. [3] ZENG F X, SONG C Q, CAO Z G, et al. Monitoring inland water via Sentinel satellite constellation: a review and perspective[J]. ISPRS journal of photogrammetry and remote sensing, 2023, 204:340-361. [4] PAGE B P, OLMANSON L G, MISHRA D R. A harmonized image processing workflow using Sentinel- 2/MSI and Landsat-8/OLI for mapping water clarity in optically variable lake systems[J]. Remote sensing of environment, 2019, 231:12. [5] PAHLEVAN N, SARKAR S, FRANZ B A, et al. Sentinel-2 MultiSpectral Instrument (MSI) data processing for aquatic science applications: demonstra- tions and validations[J]. Remote sensing of environment, 2017, 201:47-56. [6] ZHU L D, CUI T W, RUNA A, et al. Robust remote sensing retrieval of key eutrophication indicators in coastal waters based on explainable machine learning[J]. ISPRS journal of photogrammetry and remote sensing, 2024, 211:262-280. [7] KUHN C, VALERIO A D, WARD N, et al. Performance of Landsat-8 and Sentinel-2 surface reflectance products for river remote sensing retrievals of chlorophyll-a and turbidity[J]. Remote sensing of environment, 2019, 224: 104-118. [8] XIAO Y, CHEN J H, XU Y, et al. Monitoring of chlorophyll-a and suspended sediment concentrations in optically complex inland rivers using multisource remote sensing measurements[J]. Ecological indicators, 2023, 155:111041. [9] MAMUN M, KIM J J, ALAM M A, et al. Prediction of algal chlorophyll-a and water clarity in monsoon-region reservoir using machine learning approaches[J]. Water, 2020, 12(1):30. [10] MCCRACKIN M L, MULLER-KARULIS B, GUSTA- FSSON B G, et al. A century of legacy phosphorus dynamics in a large drainage basin[J]. Global biogeo- chemical cycles, 2018, 32(7):1107-1122. [11] WANG X, WEN Z D, LIU G, et al. Remote estimates of total suspended matter in China’s main estuaries using Landsat images and a weight random forest model[J]. ISPRS journal of photogrammetry and remote sensing, 2022, 183:94-110. [12] 刘曼, 付波霖, 何宏昌, 等. 基于多时相主被动遥感的漓江水面监测与水质参数反演(2016—2020年)[J]. 湖泊科学, 2021, 33(3):687-706. [13] GUO H W, TIAN S, HUANG J J, et al. Performance of deep learning in mapping water quality of Lake Simcoe with long-term Landsat archive[J]. ISPRS journal of photogrammetry and remote sensing, 2022, 183:451-469. [14] ZENG C Q, RICHARDSON M, KING D J. The impacts of environmental variables on water reflectance measured using a lightweight unmanned aerial vehicle (UAV)-based spectrometer system[J]. ISPRS journal of photo- grammetry and remote sensing, 2017, 130:217-230. [15] CHEN B T, MU X, CHEN P, et al. Machine learning-based inversion of water quality parameters in typical reach of the urban river by UAV multispectral data[J]. Ecological indicators, 2021, 133:108434. [16] 刘彦君. 基于无人机多光谱影像的小微水域水质要素反演[D]. 杭州: 浙江农林大学, 2018. [17] 赵琳. 基于多光谱遥感技术的白洋淀水质演变监测方法研究[J]. 现代信息科技, 2023(3):106-109. [18] OLMANSON L G, BREZONIK P L, BAUER M E. Airborne hyperspectral remote sensing to assess spatial distribution of water quality characteristics in large rivers: the Mississippi River and its tributaries in Minnesota[J]. Remote sensing of environment, 2013, 130:254-265. [19] ABD-ELRAHMAN A, CROXTON M, PANDE- CHETTRI R, et al. In situ estimation of water quality parameters in freshwater aquaculture ponds using hyperspectral imaging system[J]. ISPRS journal of photogrammetry and remote sensing, 2011, 66(4):463- 472. [20] MACIEL D A, BARBOSA C C F, NOVO E, et al. Water clarity in Brazilian water assessed using Sentinel-2 and machine learning methods[J]. ISPRS journal of photogrammetry and remote sensing, 2021, 182:134-152. [21] GUO Y H, XIAO Y, HAO F H, et al. Comparison of different machine learning algorithms for predicting maize grain yield using UAV-based hyperspectral images [J]. International journal of applied earth observation and geoinformation, 2023, 124:103528. [22] LI J, XIAO Z Q, SUN R, et al. A method to estimate leaf area index from VIIRS surface reflectance using deep transfer learning[J]. ISPRS journal of photogrammetry and remote sensing, 2023, 202:512-527. [23] HAFEEZ S, WONG M S, HO H C, et al. Comparison of machine learning algorithms for retrieval of water quality indicators in case-ii waters: a case study of Hong Kong[J]. Remote sensing, 2019, 11(6):617. [24] NEGRI R G, DUTRA L V, SANT’ANNA S J S. An innovative support vector machine based method for contextual image classification[J]. ISPRS journal of photogrammetry and remote sensing, 2014, 87:241-248. [25] XIAO Y, LI X R, CHEN J H, et al. A novel approach for tracking the diurnal dynamics of chlorophyll-a in lakes using UAV hyperspectral images[J]. Journal of hydrology, 2025, 660:133403. |
| 备注/Memo: | 收稿日期：2025−05−23. 宁波大学学报（理工版）网址：http://journallg.nbu.edu.cn/ 基金项目：国家社科基金重大项目（23&ZD105） 第一作者：郭亚会，硕导/副教授，主要研究方向为无人机遥感。E-mail: guoyh@ccnu.edu.cn 宁波大学学报（理工版）网址：http://journallg.nbu.edu.cn/ |
