Study On Key Technologies Of All-weather And Wide-coverage Water Environment Monitoring System Based On Optical Analysis

Water is the source of life.Human production and life are closely related to water resources.Water area degradation and water quality deterioration have become important issues affecting the sustainable development of society in the process of industrialization and urbanization.The water environment monitoring technology based on optical analysis has the advantages of good environmental adaptability,high-resolution large-range characteristics and easy networking.It has strong progressiveness in the application of water environment monitoring with complex environment.Distributed water level monitoring for regional water environment,widely testing and evaluation to water quality information and all-weather wide-coverage working platform for water environmental monitoring are the mainly demands in practice.In order to meet these requirements of the actual water environment monitoring,we studied the water level monitoring method using distributed optical fiber sensing technology,the hyperspectral classification algorithm for water quality remote sensing and UAV precise landing technology based on optical guidance.The relevant mechanism,algorithm and experimental research are completed,as well as the water environment operation platform.The main research contents of this dissertation include the following parts:(1)A Phase Optical Time Domain Reflectometry(φ-OTDR)based water level monitoring method is proposed to meet the requirement of distributed water level monitoring in regional water environment.The modulation mechanism of temperature difference on optical signal is studied.A water level sensor based on Scattering Enhanced Optical Fiber(SEOF)is prepared,and a corresponding distributed phase sensing demodulation algorithm is developed to obtain the laser phase information propagating along the fiber in real time.Large range and high precision water level detection are realized.The scattering enhanced fiber has scattering enhancement points with equal spacing distribution,which can effectively suppress the random space-time phase noise in φ-OTDR and improve the phase measurement accuracy of the system.At the same time,through the special sensitization package of SEOF,the resolution of the distributed liquid level measurement system can be improved to 142 μm,the measurement range can reach 320 m,and the measurement resolution is not affected by the measuring range.(2)For the needs of large-scale water quality detection and pollution classification and evaluation of regional water environment,a hyperspectral classification algorithm based on attention mechanism is proposed.The proposed algorithm can enhance the differences of hyperspectral data between different categories,suppress the similarity of spectral data between categories,and achieve the purpose of distinguishing and locating the exact differences between different categories of hyperspectral data.The accuracy of the classification algorithm is verified by equivalent tests on hyperspectral data sets with different spatial resolution,spectral resolution and from different imaging sensors,which are three classic hyperspectral data sets,Indian Pines,Pavia and Salinas.Through the analysis of the overall accuracy(OA),average accuracy(AA)and kappa coefficient,the classification algorithm proposed in this paper has the best comprehensive performance and better classification separability.(3)For the operation requirements of the all-weather and wide-coverage monitoring platform of the regional water environment,the precise landing technology of UAV based on optical guidance is studied,and a rapid solution of positioning information and attitude adjustment method based on multiple markers is proposed.Through the identification and calculation of Apriltag multiple markers,the position and angle deviation of the UAV relative to the apron are dynamically adjusted to achieve the precise landing of the UAV.Experiments verify that the precision landing technology proposed in this paper achieves a position deviation of less than ± 12 cm and an angle deviation of less than ± 5 °,which provides a technical basis for the automatic extended range endurance operation of UAV.(4)Aiming at the practical application of water environment monitoring,the multi unmanned unit network control technology based on cloud management is studied.The data link between UAV,UAV hive and cloud server is realized by using heterogeneous network protocols.The "End-Station-Cloud" integrated water environment monitoring platform is built,which greatly improves the operation efficiency.By carrying the liquid level sensor on the UAV,the monitoring of the position change of the 0～420 mm liquid level relative to the sensor is realized in the range of 0～50 s.While the fixed deployment of the liquid level sensor realizes the continuous monitoring of the 0～80 mm liquid level change in the range of 0～4.5h.Through the collected spectral data,the suspended solids and particles data set is selected to verify the accuracy of the classification algorithm,which affects the water quality conditions.The experimental results show that the algorithm has achieved the classification effect with the best comprehensive performance.Through two kinds of application tests,the overall verification of the water environment monitoring system has the ability of all-weather and wide-coverage.