Unmanned aerial vehicle-mounted hyperspectral water quality monitoring system and method with automatic calibration

Abstract

The unmanned aerial vehicle-mounted hyperspectral water quality monitoring system comprises an unmanned aerial vehicle, an airborne hyperspectral camera, a radiation acquisition and water quality calibration analysis module, a main control module and an unmanned aerial vehicle airport, and the unmanned aerial vehicle airport is a command station and a logistics supply station of the unmanned aerial vehicle; the main control module is respectively communicated with the unmanned aerial vehicle, the airborne hyperspectral camera and the radiation acquisition and water quality calibration analysis module, and is used for starting and stopping work of the three modules based on an instruction of an airport of the unmanned aerial vehicle, and calculating and analyzing data acquired by each module to output a final water quality remote measurement value; the radiation acquisition and water quality calibration analysis module and the airborne hyperspectral camera are used for performing spectrum acquisition according to an instruction of the main control module, the acquired spectrum is used for calculating a telemetering value and calibrating the telemetering value, and data is provided for the main control module to calculate a final water quality telemetering value; the system can realize full-automatic calibration of water quality parameters by the unmanned aerial vehicle, and can correct and calibrate to obtain a more accurate water quality monitoring result.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicles, and in particular to an unmanned aerial vehicle-borne hyperspectral water quality monitoring system and method with automatic calibration. Background Art

[0002] The pollution of urban water bodies in my country mainly comes from eutrophication of water bodies and the discharge of a large amount of external sewage. Water quality monitoring plays a vital role in water environment protection, water pollution control and maintaining healthy water quality. For a long time, the widely used water quality monitoring method includes collecting and preserving water samples on site after calibrating the monitoring points in the water area to be tested, and then sending the well-preserved water samples to the laboratory for analysis. Commonly used detection methods include laboratory chemical detection technologies such as water quality detection instrument analysis, redox and electrochemical methods, heating and oxidant decomposition. These methods can provide water quality parameter analysis of the water area to be tested and propose corresponding solutions. Although laboratory chemical analysis can obtain more accurate water quality parameters, it requires a lot of manpower and time, and has strict requirements on the preservation conditions of water samples. In addition, this method is usually limited to fixed-point water collection, and it is difficult to achieve real-time monitoring of large areas and long water areas, resulting in the inability to comprehensively and specifically formulate water quality maintenance and treatment plans, and it is even more difficult to effectively predict water quality changes.

[0003] In recent years, remote sensing technology has been applied to water quality monitoring, especially the hyperspectral water quality remote sensing technology based on drones. The image data obtained by remote sensing technology can be used to monitor water pollution in real time for waters with large areas to be measured and long monitoring lengths. Water quality hyperspectral remote sensing technology uses the characteristics of large-scale, continuous observation, multi-platform, multi-band, and multi-resolution of hyperspectral remote sensing technology to periodically measure the physical and chemical properties of water to evaluate and determine the quality of water bodies. The basis of water quality remote sensing technology lies in the spectral characteristics of water body components. The changes in the spectral curves of these components reflect the changes in the total composition of the water body. Remote sensing technology monitors water quality by capturing these spectral changes. When sunlight shines on the water surface, part of the light enters the atmosphere directly due to mirror reflection, while the other part of the light is refracted into the water, and after reflection, scattering, and reflection from the bottom of the water, it returns to the atmosphere. The sensors in the hyperspectral remote sensing equipment capture these lights to form monitoring data. By performing regression analysis and deep learning on these spectral data, an inversion model for water quality parameters is established to invert the concentration, content or distribution of pollutants in the water, and then evaluate the degree of pollution and water quality level of the water area.

[0004] However, hyperspectral water quality parameters are affected by weather and environment, and often need to be combined with traditional sampling and detection technology to improve accuracy. In addition, the flight control of drones also requires professionally trained and licensed personnel. Therefore, accurately obtaining a wide range of water quality parameters still requires a lot of manpower and time investment. In addition, the water quality parameters obtained by hyperspectral remote sensing are easily affected by various conditions such as the environment, meteorology, and solar altitude. There is a problem that the trends are close but the values ​​fluctuate greatly. Accurate hyperspectral remote sensing data needs to be compared and corrected with the laboratory data of sampling analysis to further improve accuracy. This undoubtedly further increases the implementation cost and difficulty of hyperspectral remote sensing, and to a certain extent limits the application of hyperspectral remote sensing in water quality monitoring. Summary of the invention

[0005] In view of the problems in the related art, the present invention proposes an unmanned aerial vehicle-mounted hyperspectral water quality monitoring system and method with automatic calibration to overcome the above-mentioned technical problems existing in the existing related technology.

[0006] To this end, the specific technical solution adopted by the present invention is as follows: A drone-mounted hyperspectral water quality monitoring system and method with automatic calibration, comprising: a drone, an airborne hyperspectral camera, a radiation collection and water quality calibration and analysis module, a main control module, and a drone airport. The onboard hyperspectral camera, radiation collection and water quality calibration and analysis module, and main control module are all integrated on the drone; The drone airport is a command station and logistics supply station for drones, used to command drone flights, perform data collection and calibration analysis tasks, provide drone parking aprons, drone charging, data download, processing and transmission, and provide logistics services for drones; The main control module communicates with the drone, airborne hyperspectral camera, and radiation collection and water quality calibration analysis module respectively, and is used to start and stop the three based on the instructions of the drone airport. At the same time, it calculates and analyzes the data collected by each module to output the final water quality telemetry value; The radiation collection and water quality calibration and analysis module and the airborne hyperspectral camera are used to collect spectra according to the instructions of the main control module. The collected spectra are used to calculate telemetry values ​​and calibrate telemetry values, providing data for the main control module to calculate the final water quality telemetry values.

[0007] Furthermore, the radiation collection and water quality calibration analysis module includes an optical path and a flow path. The optical path includes a cuvette, a switching reflector, a cosine corrector, a spectrometer, a first collimating lens, a second collimating lens, a third collimating lens and a light source. The cuvette is connected to the flow path. The switching reflector and the first collimating lens are sequentially arranged on the left side of the cuvette. The first collimating lens is connected to the spectrometer through an optical fiber. The second collimating lens is located above the switching reflector. The second collimating lens is connected to the cosine corrector through an optical fiber. The third collimating lens is arranged on the right side of the cuvette. The light source is arranged on the right side of the third collimating lens. The light source, the spectrometer and the switching reflector are all electrically connected to the main control module. The switching reflector receives the instruction of the main control module to switch the reflection angle for spectrum collection. The spectra collected at different reflection angles are respectively used to calculate the telemetry value and calibrate the telemetry value. The top of the cosine corrector is higher than the top of the drone to collect sunlight. The flow path is a plurality of pipes connected to the cuvette, and is used to provide different water samples to the cuvette according to the instructions of the main control module.

[0008] Furthermore, the light source adopts a pulse xenon lamp.

[0009] Furthermore, the spectral range of the spectrometer covers 200-800 nm, and the FWHH resolution is less than 2 nm.

[0010] Furthermore, the cuvette is made of fused quartz.

[0011] Furthermore, when the radiation collection and water quality calibration analysis module performs spectrum collection in water, the reflector is switched to a horizontal state. When the radiation collection and water quality calibration analysis module performs spectrum collection under solar radiation, the reflector is switched to a state inclined 45° relative to the horizontal.

[0012] A method for unmanned aerial vehicle-mounted hyperspectral water quality monitoring with automatic calibration is implemented by using any of the above-mentioned unmanned aerial vehicle-mounted hyperspectral water quality monitoring systems with automatic calibration, and comprises the following steps: S1: Use the radiation collection and water quality calibration analysis module to collect the background spectrum of pure water samples; S2: The drone flies to any sampling point A within the monitoring route, and uses the radiation collection and water quality calibration analysis module to collect the spectrum of the water sample at sampling point A. The main control module calculates the water quality parameter P of sampling point A based on the background spectrum of the pure water sample and the spectrum of the water sample at sampling point A. A ; S3: The UAV rises straight along sampling point A to the first height, and the radiation collection and water quality calibration analysis module is far away from the water surface. At the same time, the radiation collection and water quality calibration analysis module and the airborne hyperspectral camera are used to collect the module radiation spectrum S at the first height of sampling point A. A1 and the camera radiation spectrum H A1 ; S4: The main control module generates a spectrum S at the first height. A1 and spectrum HA1 Calculate the reflectance spectrum R of sampling point A A , R A =S A1 / H A1 ; S5: The main control module uses the reflectance spectrum R A The inversion is performed to obtain the first height hyperspectral water quality parameter P of sampling point A A ’ , and then use the water quality parameter P of sampling point A A Perform calibration, obtain the calibration coefficient and output it, the calibration coefficient k=P A / P A ’ ; S6: The UAV continues to rise to the second altitude, and uses the radiation collection and water quality calibration analysis module and the airborne hyperspectral camera to collect the spectrum S at the second altitude respectively. Ah and spectrum H Ah ; S7: The drone flies to each monitoring point within the monitoring route to collect spectra. The collection steps for each monitoring point are the same. When the drone flies to monitoring point B, the radiation collection and water quality calibration analysis module and the airborne hyperspectral camera are used to collect the spectrum S of monitoring point B. Bh and spectrum H Bh ; S8: The main control module generates the module radiation spectrum S at the first height of sampling point A. A1 and the camera radiation spectrum H A1 , the second highest spectrum S Ah and spectrum H Ah And the spectrum S of monitoring point B Bh and spectrum H Bh Calculate the reflectance spectrum R(H) of monitoring point B Bh , S Bh ), , where A and B are constants, which are fixed parameters determined by optics and circuits; S9: The main control module adopts R(H Bh , S Bh ) Invert the module radiation water quality parameters of monitoring point B to obtain the water quality parameter remote sensing value P B ’ , and then use the calibration coefficient k to calibrate the inversion result to obtain the calibrated remote sensing value P of monitoring point B B , P B =kP B ’ ; S10: Repeat steps S7-S9 to obtain the calibrated remote sensing value of each monitoring point and output it.

[0013] The beneficial effects of the present invention are: 1. This application can realize the automatic calibration of water quality parameters by drones and can correct the calibration to obtain more accurate water quality monitoring results; 2. This application can automatically calibrate the water quality parameters collected and inverted by the UAV hyperspectral data through the design of optical path and flow path, thereby improving the accuracy of the water quality data monitored by the hyperspectral data; 3. This application improves the measurement accuracy of reflectivity by making real-time corrections to the atmospheric transmission of solar radiation, thereby improving the quality of remote sensing data of water quality parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0015] Figure 1 is a structural schematic diagram of a drone-mounted hyperspectral water quality monitoring system with automatic calibration according to Embodiment 1 of the present invention; Figure 2 is a schematic diagram of the structure of a drone of a drone-mounted hyperspectral water quality monitoring system with automatic calibration according to Embodiment 1 of the present invention; Figure 3 It is a schematic diagram of the optical path and flow path of a radiation collection and water quality calibration analysis module of a drone-mounted hyperspectral water quality monitoring system with automatic calibration according to an embodiment of the present invention when performing spectrum collection in water; Figure 4 The present invention is a schematic diagram of the optical path and flow path of a radiation collection and water quality calibration analysis module of an unmanned aerial vehicle-mounted hyperspectral water quality monitoring system with automatic calibration when performing spectrum collection under solar radiation according to an embodiment of the present invention.

[0016] In the figure: 1. UAV; 2. Airborne hyperspectral camera; 3. UAV airport; 4. Radiation collection and water quality calibration analysis module; 41. Cuvette; 42. Switching reflector; 43. Cosine corrector; 44. Spectrometer; 45. First collimating lens; 46. Second collimating lens; 47. Third collimating lens; 48. Light source. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] According to an embodiment of the present invention, a drone-mounted hyperspectral water quality monitoring system and method with automatic calibration are provided. Embodiment 1

[0019] like Figure 1-Figure 4 As shown, the unmanned aerial vehicle-mounted hyperspectral water quality monitoring system with automatic calibration according to an embodiment of the present invention comprises an unmanned aerial vehicle 1, an airborne hyperspectral camera 2, a radiation collection and water quality calibration and analysis module 4, a main control module and an unmanned aerial vehicle airport 3. The airborne hyperspectral camera 2, the radiation collection and water quality calibration and analysis module, and the main control module are all integrated on the UAV 1; UAV Airport 3 is the command station and logistics supply station of UAV 1, which is used to command the flight of UAV, perform data collection and calibration analysis tasks, provide UAV parking apron, UAV charging, data download, processing and transmission, and provide logistics services for UAV; The main control module communicates with the drone 1, the airborne hyperspectral camera 2 and the radiation collection and water quality calibration analysis module 4 respectively, and is used to start and stop the work of the three based on the instructions of the drone airport 3, and at the same time calculates and analyzes the data collected by each module to output the final water quality telemetry value; The radiation collection and water quality calibration analysis module 4 and the airborne hyperspectral camera 2 are used to collect spectra according to the instructions of the main control module. The collected spectra are used to calculate telemetry values ​​and calibrate telemetry values, providing data for the main control module to calculate the final water quality telemetry values.

[0020] It should be noted that the main control module includes necessary control circuits and data processing computers. The data processing computer can be integrated with the airborne hyperspectral camera 2, or installed on the fuselage of the drone 1 together with the radiation collection and water quality calibration analysis module 4. The airborne hyperspectral camera 2 is connected to the drone 1 through the gimbal and the gimbal bracket. The gimbal is equipped with a wiring harness to enable the airborne hyperspectral camera 2 to draw power from the drone 1 and communicate with the drone. The radiation collection and water quality calibration analysis module 4 can be installed on the top of the drone, or integrated into the drone fuselage, depending on the structural design of the drone.

[0021] Drone Airport 3 is connected to the power supply and network. Users can pre-set the route and the mission points in the route to realize the inspection of drone 1. During the mission, drone airport 3 can command drone 1 to fly, perform data collection and calibration tasks, provide drone parking aprons, drone charging, and download, process and transmit data.

[0022] The radiation collection and water quality calibration analysis module 4 includes an optical path and a flow path. The optical path includes a cuvette 41, a switching reflector 42, a cosine corrector 43, a spectrometer 44, a first collimating lens 45, a second collimating lens 46, a third collimating lens 47 and a light source 48. The light source 48 adopts a pulsed xenon lamp. The spectral range of the spectrometer 44 covers 200-800nm, and the FWHH resolution is less than 2nm. The cuvette 41 is made of fused quartz. The cuvette 41 is connected to the flow path, the switching reflector 42 and the first collimating lens 45 are sequentially arranged on the left side of the cuvette 41, the first collimating lens 45 is connected to the spectrometer 44 through an optical fiber, the second collimating lens 46 is located above the switching reflector 42, the second collimating lens 46 is connected to the cosine corrector 43 through an optical fiber, the third collimating lens 47 is arranged on the right side of the cuvette 41, the light source 48 is arranged on the right side of the third collimating lens 47, the light source 48, the spectrometer 44, and the switching reflector 42 are all electrically connected to the main control module, the switching reflector 42 receives the instruction of the main control module to switch the reflection angle for spectrum collection, and the spectra collected at different reflection angles are respectively used to calculate the telemetry value and calibrate the telemetry value, and the top of the cosine corrector 43 is higher than the top of the drone 1 to collect sunlight; The flow path is a plurality of pipes connected to the cuvette 41, which are used to provide different water samples to the cuvette 41 according to the instructions of the main control module. The flow path part includes a circuit-controlled three-way switching valve, a pipeline, and a sampling pump connected to the cuvette and the pipeline. In addition, there is also a necessary control drive circuit to realize the control of optical path, flow path switching, etc.

[0023] The radiation collection and water quality calibration analysis module 4 needs to collect spectra in the process of calculating the telemetry value and calibrating the telemetry value, and needs to collect spectra in water and under solar radiation respectively, such as Figure 3 As shown, when collecting spectrum in water, the switch reflector 42 is rotated to a horizontal state, as shown in FIG. Figure 4 As shown, when the radiation collection and water quality calibration analysis module 4 performs spectrum collection under solar radiation, the switching reflector 42 rotates to a state of 45° relative horizontal inclination. In addition, in order to cooperate with the realization of spectrum collection in water, pure water, cleaning liquid, containers and pipelines need to be set up in the airport. When the drone 1 is recovered to the airport, the sampling tube of the drone 1 is in a retracted state. The main control board of the radiation collection and water quality calibration analysis module 4 communicates with the airport, and then allows the airport to control which container the sampling tube is connected to.

[0024] The entire system can automatically inspect the target water area without human intervention. And its automatic calibration capability can obtain water quality parameter analysis values ​​at the beginning of each mission to calibrate the telemetry values, without the need for personnel to go to the site for calibration operations or perform subsequent data processing, so as to ensure data quality. After the inspection, the system automatically processes the data, uploads the data, and automatically cleans the water quality analysis module. Therefore, fully automated and reliable water quality parameter remote sensing operations can be achieved. In addition to regular maintenance, battery inspection, and pure water replenishment, there is no other work in order.

[0025] Hyperspectral remote sensing of water quality is based on the change in light reflectivity caused by changes in components in the water body. The change in reflectivity is essentially still caused by the absorption of light by specific molecules in the water. In this process, the reflectivity spectrum curve is used to measure specific components, and to achieve a quantitative level, it is necessary to obtain the reflectivity curve more accurately. In addition, even if a more accurate reflectivity curve is obtained, it is still difficult to obtain accurate water quality parameters. Because the components in the water are not the only factors that affect the reflectivity curve, other factors include the intensity of solar radiation, solar altitude angle, water depth, water turbidity and other meteorological conditions. Based on this complex actual situation, the present invention designs a set of optical paths and flow paths to achieve real-time reflectivity detection and water quality parameter analysis and calibration. Since the water quality parameters inverted by the hyperspectral method have a good correlation with the water quality parameters obtained by the transmission analysis method, the calibration can be completed by polynomial fitting and other methods.

[0026] Specifically, Figure 3 As shown, when the spectrum is collected in water, the main control module controls the switching reflector 42 to rotate to the horizontal state in the figure. The complex light emitted by the xenon lamp of the light source 48 is collimated as parallel light by the third collimating lens 47 and incident on the cuvette 41. The material of the cuvette 41 is preferably fused quartz to achieve high transmittance in the ultraviolet band. The light beam enters the lens of the spectrometer 44 through the cuvette 41 and is converged, and is received and transmitted to the spectrometer 44 by the optical fiber for collection. The workflow is as follows: the three-way valve is switched to the pure water path, the sampling pump is started for a period of time so that the cuvette 41 is filled with pure water, the pulse xenon lamp is triggered to flash, and the spectrometer 44 is also triggered to collect the spectrum. At this time, the background spectrum of pure water is obtained, and then the three-way valve is switched to the sample water / cleaning liquid path. If the drone 1 is collecting above the water body, then this path is connected to the sample water body, the sampling pump is started for a period of time so that the cuvette 41 is filled with pure water, the xenon lamp is triggered to flash, and the spectrometer 44 is also triggered to collect the spectrum. At this time, the sample water spectrum is obtained, and then the absorbance spectrum of the sample water is obtained, and then the water quality parameters are calculated based on the absorbance spectrum of the sample water and the pre-built model.

[0027] Specifically, Figure 4As shown, when spectrum collection is performed under solar radiation, the switching reflector 42 is rotated to the state of tilting 45° in the figure. At this time, the xenon lamp no longer works synchronously with the spectrometer 44, and the light path of the cuvette 41 is blocked. The solar radiation passes through the cosine corrector 43 and is transmitted by the optical fiber, then collimated by the second collimating lens 46, reflected by the switching reflector 42, converged by the first collimating lens 45 and transmitted by the optical fiber, and finally received by the spectrometer 44. Embodiment 2

[0028] A method for unmanned aerial vehicle-mounted hyperspectral water quality monitoring with automatic calibration is implemented by using the unmanned aerial vehicle-mounted hyperspectral water quality monitoring system with automatic calibration of the first embodiment, and comprises the following steps: S1: using the radiation collection and water quality calibration analysis module 4 to collect the background spectrum of the pure water sample; S2: UAV 1 flies to any sampling point A within the monitoring route, and uses the radiation collection and water quality calibration analysis module 4 to collect the spectrum of the water sample at sampling point A. The main control module calculates the water quality parameter P of sampling point A based on the background spectrum of the pure water sample and the spectrum of the water sample at sampling point A. A ; S3: The drone 1 rises straight along the sampling point A to a first height of about 10 meters. The radiation collection and water quality calibration analysis module 4 is far away from the water surface. At the same time, the radiation collection and water quality calibration analysis module 4 and the airborne hyperspectral camera 2 are used to collect the module radiation spectrum S at the first height of the sampling point A. A1 and the camera radiation spectrum H A1 ; S4: The main control module generates a spectrum S at the first height. A1 and spectrum H A1 Calculate the reflectance spectrum R of sampling point A A , R A =S A1 / H A1 ; S5: The main control module uses the reflectance spectrum R A The inversion is performed to obtain the first height hyperspectral water quality parameter P of sampling point A A ’ , and then use the water quality parameter P of sampling point A A Perform calibration, obtain the calibration coefficient and output it, the calibration coefficient k=P A / P A ’ ; S6: The UAV 1 continues to rise to a second altitude of about 100 meters, and uses the radiation collection and water quality calibration analysis module 4 and the airborne hyperspectral camera 2 to collect the spectrum S at the second altitude. Ah and spectrum H Ah ; S7: UAV 1 flies to each monitoring point in the monitoring route to collect spectra. The collection steps for each monitoring point are the same. When UAV 1 flies to monitoring point B, it uses the radiation collection and water quality calibration analysis module 4 and the airborne hyperspectral camera 2 to collect the spectrum S of monitoring point B. Bh and spectrum H Bh ; S8: The main control module generates the module radiation spectrum S at the first height of sampling point A. A1 and the camera radiation spectrum H A1 , the second highest spectrum S Ah and spectrum H Ah And the spectrum S of monitoring point B Bh and spectrum H Bh Calculate the reflectance spectrum R(H) of monitoring point B Bh , S Bh ), , where A and B are constants, which are fixed parameters determined by optics and circuits; S9: The main control module adopts R(H Bh , S Bh ) Invert the module radiation water quality parameters of monitoring point B to obtain the water quality parameter remote sensing value P B ’ , and then use the calibration coefficient k to calibrate the inversion result to obtain the calibrated remote sensing value P of monitoring point B B , P B =kP B ’ ; S10: Repeat steps S7-S9 to obtain the calibrated remote sensing value of each monitoring point and output it.

[0029] It should be noted that the hyperspectral remote sensing inversion of water quality parameters is mainly obtained by analyzing the reflectance of water bodies. Therefore, accurate acquisition of water body reflectance is the prerequisite and key to the correct inversion of water quality parameters. Since UAV hyperspectral remote sensing is in an open environment, the states of the hyperspectral camera, light source (sun) and target sample (flowing water body) cannot remain static, so it is not feasible to obtain absolutely accurate reflectance, and the only way is to eliminate the main system errors as much as possible.

[0030] The most common method is to use a standard reflectance plate, which is generally made of metal or Teflon material to form a diffuse reflective surface with known reflectivity. By collecting the spectra of the standard reflectance plate and the sample and comparing the two, a more accurate sample reflectance spectrum can be obtained. This method is simple, effective, and highly accurate. The measurement conditions of the sample and the standard reflectance plate are almost exactly the same, and the only variable is the reflectivity of the sample itself. Solar radiation, atmospheric transmission, and environmental changes are all taken into account. However, this method also has obvious disadvantages - lack of timeliness. As a light source that changes and moves, the altitude angle and cloud cover of sunlight are constantly changing. Sometimes the effective radiation from the sun to the sample surface can change dramatically in just a few minutes. In actual monitoring, the task may often last up to half an hour, and the measurement of the standard reflectance plate cannot be carried out throughout the entire time. Therefore, the reflectivity collected during the task actually fluctuates with the change of solar radiation. In addition, the carrying and deployment of the reflectance plate must be done manually, which is not conducive to automated operations.

[0031] Another solution is to use solar radiation to measure solar radiation in real time. In the early stage, standard reflectance plates or standard light sources are still used to calibrate hyperspectral cameras and solar radiation, so that the measurement of radiation intensity is relatively consistent. In this way, a real-time reflectance spectrum can be obtained by using solar radiation data and hyperspectral camera data. This solution solves the problem that standard reflectance plates cannot provide data in real time. However, the accuracy of using solar radiation to correct reflectance is limited. The main reason is that the installation position of solar radiation is on the top of the drone, about 100 meters from the ground. Therefore, the light intensity here cannot fully represent the light intensity on the surface of the water body. Air transmission is not considered.

[0032] The above-mentioned influence of solar radiation and atmospheric transmission on reflectivity not only makes the inversion results of water quality parameters inaccurate, but also requires more data processing and correction work to improve the accuracy of the data. In order to realize the fully automatic operation of hyperspectral water quality monitoring, the present invention not only provides the hardware design, but also provides the work and calculation method to improve the accuracy of reflectivity.

[0033] The 400-1000nm band used for water quality hyperspectral remote sensing is an atmospheric window, but there is still absorption by water and carbon dioxide. In addition, particulate matter and aerosols in the air will also scatter sunlight. Since drones equipped with hyperspectral cameras generally work at an altitude of about 100 meters, the water body reflected light collected by the hyperspectral camera under the same solar radiation is affected by atmospheric absorption and scattering. This is particularly significant in areas with severe air pollution or fog. The water body reflectance obtained in this way has a large error, which causes errors in the inversion of water quality parameters.

[0034] Based on the above, the present invention proposes that solar radiation and hyperspectral cameras need to be calibrated using the same light source before leaving the factory, and the following relationship is obtained: s =AI hs +B, where I s is the spectrum collected from solar radiation, I hs is the spectrum collected by the hyperspectral camera, A and B are constants. s and I hs Both can be experimentally measured at the same time, so A and B can be calculated. A and B are fixed parameters determined by optics and circuits.

[0035] Therefore, before executing each flight mission, the system first performs an automatic calibration, and the calibration method is as described in the above steps S3-S8.

[0036] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An unmanned aerial vehicle hyperspectral water quality monitoring system with automatic calibration, characterized in that: include: UAV (1), airborne hyperspectral camera (2), radiation collection and water quality calibration analysis module (4), main control module and UAV airport (3), An airborne hyperspectral camera (2), a radiation collection and water quality calibration and analysis module, and a main control module are all integrated on the drone (1); The drone airport (3) is a command post and logistics supply post for the drone (1), used to command the flight of the drone, perform data collection and calibration analysis tasks, provide a drone landing pad, charge the drone, download, process and transmit data, and provide logistics services for the drone; The main control module communicates with the drone (1), the airborne hyperspectral camera (2) and the radiation collection and water quality calibration analysis module (4) respectively, and is used to start and stop the work of the three based on the instructions of the drone airport (3), and at the same time calculates and analyzes the data collected by each module to output the final water quality telemetry value; The radiation collection and water quality calibration analysis module (4) and the airborne hyperspectral camera (2) are used to collect spectra according to the instructions of the main control module. The collected spectra are used to calculate telemetry values ​​and perform telemetry value calibration, providing data for the main control module to calculate the final water quality telemetry value.

2. The unmanned aerial vehicle-mounted hyperspectral water quality monitoring system with automatic calibration according to claim 1, characterized in that: The radiation collection and water quality calibration analysis module (4) includes an optical path and a flow path. The optical path includes a cuvette (41), a switching reflector (42), a cosine corrector (43), a spectrometer (44), a first collimating lens (45), a second collimating lens (46), a third collimating lens (47) and a light source (48). The cuvette (41) is connected to the flow path. The switching reflector (42) and the first collimating lens (45) are sequentially arranged on the left side of the cuvette (41). The first collimating lens (45) is connected to the spectrometer (44) via an optical fiber. The second collimating lens (46) is located above the switching reflector (42). ) is connected to a cosine corrector (43) via an optical fiber, a third collimating lens (47) is arranged on the right side of the cuvette (41), a light source (48) is arranged on the right side of the third collimating lens (47), the light source (48), the spectrometer (44), and the switching reflector (42) are all electrically connected to a main control module, the switching reflector (42) receives an instruction from the main control module to switch the reflection angle for spectrum collection, and the spectra collected at different reflection angles are respectively used to calculate telemetry values ​​and perform telemetry value calibration, and the top of the cosine corrector (43) is higher than the top of the drone (1) to collect sunlight; The flow paths are a plurality of pipes connected to the cuvette (41), and are used to provide different water samples to the cuvette (41) according to instructions from the main control module.

3. The unmanned aerial vehicle-mounted hyperspectral water quality monitoring system with automatic calibration according to claim 2 is characterized in that: The light source (48) is a pulsed xenon lamp.

4. The unmanned aerial vehicle-mounted hyperspectral water quality monitoring system with automatic calibration according to claim 2 or 3, characterized in that: The spectrometer (44) covers a spectral range of 200-800 nm and a FWHH resolution of less than 2 nm.

5. The unmanned aerial vehicle-mounted hyperspectral water quality monitoring system with automatic calibration according to claim 2 or 3, characterized in that: The cuvette (41) is made of fused quartz.

6. The unmanned aerial vehicle-mounted hyperspectral water quality monitoring system with automatic calibration according to claim 2 or 3, characterized in that: When the radiation collection and water quality calibration analysis module (4) collects spectra in water, the switching reflector (42) is rotated to a horizontal state; when the radiation collection and water quality calibration analysis module (4) collects spectra under solar radiation, the switching reflector (42) is rotated to a state inclined at 45° relative to the horizontal.

7. A method for unmanned aerial vehicle-mounted hyperspectral water quality monitoring with automatic calibration, implemented by using the unmanned aerial vehicle-mounted hyperspectral water quality monitoring system with automatic calibration as claimed in any one of claims 1 to 6, characterized in that: The following steps are involved: S1: Use the radiation collection and water quality calibration analysis module (4) to collect the background spectrum of the pure water sample; S2: The drone (1) flies to any sampling point A within the monitoring route, and uses the radiation collection and water quality calibration analysis module (4) to collect the spectrum of the water sample at sampling point A. The main control module calculates the water quality parameter P of sampling point A based on the background spectrum of the pure water sample and the spectrum of the water sample at sampling point A. A ; S3: The drone (1) rises to a first height along the sampling point A in a straight line, the radiation collection and water quality calibration analysis module (4) is away from the water surface, and the radiation collection and water quality calibration analysis module (4) and the airborne hyperspectral camera (2) are used to collect the module radiation spectrum S at the first height of the sampling point A respectively. A1 and the camera radiation spectrum H A1 ; S4: The main control module generates a spectrum S at the first height. A1 and spectrum H A1 Calculate the reflectance spectrum R of sampling point A A , R A =S A1 / H A1 ; S5: The main control module uses the reflectance spectrum R A The inversion is performed to obtain the first height hyperspectral water quality parameter P of sampling point A A ’ , and then use the water quality parameter P of sampling point A A Perform calibration, obtain the calibration coefficient and output it, the calibration coefficient k=P A / P A ’ ; S6: The drone (1) continues to rise to the second altitude, and simultaneously uses the radiation collection and water quality calibration analysis module (4) and the airborne hyperspectral camera (2) to collect the spectrum S at the second altitude. Ah and spectrum H Ah ; S7: The drone (1) flies to each monitoring point within the monitoring route to collect spectra. The collection steps for each monitoring point are the same. When the drone (1) flies to monitoring point B, the radiation collection and water quality calibration analysis module (4) and the airborne hyperspectral camera (2) are used to collect the spectrum S of monitoring point B. Bh and spectrum H Bh ; S8: The main control module generates the module radiation spectrum S at the first height of sampling point A. A1 and the camera radiation spectrum H A1 , the second highest spectrum S Ah and spectrum H Ah And the spectrum S of monitoring point B Bh and spectrum H Bh Calculate the reflectance spectrum R(H) of monitoring point B Bh , S Bh ), , where A and B are constants, which are fixed parameters determined by optics and circuits; S9: The main control module adopts R(H Bh , S Bh ) Invert the module radiation water quality parameters of monitoring point B to obtain the water quality parameter remote sensing value P B ’ , and then use the calibration coefficient k to calibrate the inversion result to obtain the calibrated remote sensing value P of monitoring point B B , P B =kP B ’ ; S10: Repeat steps S7-S9 to obtain the calibrated remote sensing value of each monitoring point and output it.

Images