Research On Water Quality Monitoring For Urban Rivers By UAV Hyperspectral Remote Sensing

In recent years,with the development of Wuhan’s urban social economy,the population of Wuhan has continued to grow,its economic scale has continued to expand,the process of industrialization has accelerated,and the pressure on water resources has gradually increased.As an important part of the urban ecosystem,urban rivers not only have the functions of providing water sources,flood control and drainage,but also have ecological environmental effects such as climate regulation and pollution reduction.Real-time and accurate monitoring of urban river water bodies and water quality,and timely detection of abnormal spatial and temporal changes in river water quality are of great significance to strengthening the long-term management of urban rivers.The water quality of urban rivers is reflected in the combined effect of multiple water quality parameters,such as suspended solids,chlorophyll a,ammonia nitrogen,and dissolved oxygen.Some water quality parameters in the water cause changes in the surface optical properties of the water.The remote sensing spectral signal can detect this change keenly,so the water quality parameters can be monitored through remote sensing.Urban rivers are usually narrow,and common remote sensing data cannot meet the needs of water quality monitoring.In the process of monitoring river water quality by remote sensing,satellite images cannot meet the requirements of spectral and spatial resolution at the same time,and it is difficult to carry out fine pollution monitoring of urban rivers.The high spatial and hyperspectral characteristics of unmanned aerial vehicle(UAV)hyperspectral remote sensing images can effectively make up for the lack of satellite remote sensing monitoring.In this paper,the Shahu Port and Xunsi River in Wuhan City are used as the experimental area,combined with machine learning methods,to monitor the water quality parameters of the experimental area,and obtain the water quality distribution map of the research area.Combined with the surrounding environment of the experimental area and the results of field monitoring,the rationality of the transparency results was analyzed and demonstrated.The main research results of the paper are as follows:(1)Achieve the UAV hyperspectral remote sensing image preprocessing process,which mainly includes sensor radiation calibration,geometric correction,site absolute radiation correction,noise removal,field measured spectrum and laboratory measured spectrum data conversion,water extraction and spectrum extraction.(2)Based on the obtained pre-processed observation data,comprehensively using UAV hyperspectral image and spectral curve information,compared with the traditional empirical semi-empirical model,the XGBoost algorithm has the highest monitoring accuracy for water quality parameters in the experimental area.The established inversion model was used to generate water quality distribution maps of the two study areas,and the results showed that the water quality distribution results were consistent with the results of field monitoring.