Multispectral remote sensing water quality monitoring method and system, electronic equipment and medium

Abstract

The invention provides a multispectral remote sensing water quality monitoring method and system, electronic equipment and a medium, and the method comprises the steps that multispectral remote sensing equipment fixedly installed on the basis of a shore base obtains an original water body image of a target water area, and the multispectral remote sensing equipment and the horizontal plane of the shore base are arranged at a preset angle; image splicing and spatial correction are carried out on the basis of the original water body images, corrected images are obtained, and the spatial correction comprises image distortion correction based on camera parameters and oblique image correction based on a perspective transformation matrix; performing water quality parameter inversion based on the corrected image to obtain a water quality monitoring parameter image; performing abnormal water mass identification based on the water quality monitoring parameter image to obtain abnormal water mass parameter information; the abnormal water mass parameter information at least comprises the area, the position and the drifting direction of the abnormal water mass; and performing water quality monitoring based on the abnormal water mass parameter information to obtain a water quality state.

Description

Technical Field

[0001] This application belongs to the technical field of water quality testing, and relates to a multispectral remote sensing water quality monitoring method, system, electronic equipment and medium. Background Technology

[0002] With socio-economic development and increased environmental awareness, the need for efficient, accurate, and comprehensive monitoring of water quality is becoming increasingly urgent. Water quality monitoring is a crucial foundation for assessing the health status of water bodies, providing early warnings of pollution incidents, and supporting management decisions. Currently, mainstream water quality monitoring technologies are mainly divided into three categories: in-situ water sampling, real-time monitoring at water quality stations, and UAV / satellite remote sensing monitoring, but each has significant limitations.

[0003] In-situ water quality sampling is highly dependent on manual labor, consuming significant manpower and resources, resulting in high costs. It also suffers from poor spatiotemporal coverage, with a limited number of sparsely distributed sampling points, making it difficult to reflect the spatial heterogeneity of water bodies. Furthermore, the monitoring frequency is low, typically involving discrete, periodic sampling, which fails to capture rapid dynamic changes in water quality parameters and sudden pollution events. Therefore, in-situ sampling cannot meet the demands of modern water quality monitoring that requires large-scale, high-frequency, and real-time dynamic monitoring.

[0004] Real-time water quality monitoring stations enable dynamic and continuous water quality monitoring, making them a relatively efficient monitoring method. However, their core problem lies in the severely limited monitoring spatial range. Water quality stations are costly to construct and maintain, have a limited number of stations with fixed locations, and can only acquire point-like or very small-scale data. Their ability to acquire spatial information such as the overall water quality status and pollutant migration and diffusion paths of vast water bodies (such as large lakes, rivers, and nearshore areas) is extremely limited, making true area-wide monitoring impossible.

[0005] Unmanned aerial vehicle (UAV) / satellite remote sensing monitoring is a viable alternative. Satellite remote sensing boasts an extremely wide coverage area, theoretically enabling global monitoring. However, its application is limited by the satellite revisit cycle (temporal resolution), typically requiring several days or even longer to acquire images of the same area, making it difficult to meet the demands of continuous, high-frequency monitoring. Atmospheric interference (especially cloud cover) is severe, making it difficult to obtain effective data in rainy or cloudy weather, resulting in poor data continuity and incomplete time series. Spatial resolution limitations hinder the identification of small-scale water bodies or detailed features. Compared to satellites, UAV remote sensing platforms offer greater flexibility, maneuverability, and spatial resolution, allowing for on-demand, refined regional monitoring. However, its application also faces numerous limitations: flight regulations restrict operations in no-fly zones (such as around airports, military facilities, densely populated areas, and ecological reserves); flight approval processes are complex and time-consuming, especially in sensitive areas; single-flight endurance is limited, resulting in a much smaller coverage area compared to satellites; it is also affected by weather conditions (such as strong winds and rain); and strict airspace management restricts long-term, frequent flights in the same area, making uninterrupted continuous monitoring impossible. Furthermore, the accuracy and universality of remote sensing inversion models still need to be further improved.

[0006] The aforementioned mainstream water quality monitoring technologies all suffer from insurmountable bottlenecks:

[0007] To address the issues of spatial image distortion, inaccurate calculation of the area and location of anomalous water masses, and insufficient stability of anomaly identification in continuous monitoring scenarios for shore-based fixed multispectral remote sensing equipment under tilted observation conditions; satellites and drones can cover a wide area but have weak continuous monitoring capabilities (limited by cycle, weather, and airspace).

[0008] The contradiction between timeliness and cost and regulations: high-frequency, real-time monitoring (such as water quality stations) is costly and space-constrained; large-scale monitoring (such as satellites / drones) is difficult to meet the high-frequency real-time requirements, and drones are easily subject to regulations.

[0009] The contradiction between accuracy and coverage: high-precision methods (such as in-situ sampling) have a small coverage area; large-scale monitoring methods (remote sensing) usually have lower accuracy than direct measurement. Summary of the Invention

[0010] This application provides a multispectral remote sensing water quality monitoring method, system, electronic device, and medium to solve the problems of insufficient spatial coverage, poor temporal continuity, and severe weather-related constraints in existing water quality monitoring technologies.

[0011] In a first aspect, this application provides a multispectral remote sensing water quality monitoring method, comprising: acquiring original water images of a target water area based on a shore-based fixed multispectral remote sensing device, wherein the multispectral remote sensing device is positioned at a preset angle to the horizontal plane of the shore base; the original water images are tilted images; performing image stitching and spatial correction based on each of the original water images to obtain a corrected image, wherein the spatial correction includes image distortion correction based on camera parameters and tilt image correction based on a perspective transformation matrix; performing water quality parameter inversion based on the corrected image to obtain a water quality monitoring parameter image; identifying abnormal water masses based on the water quality monitoring parameter image to obtain abnormal water mass parameter information; wherein the abnormal water mass parameter information includes at least the area, location, and drift direction of the abnormal water mass, and the abnormal water mass identification includes image registration, pixel difference extraction, and abnormal region extraction; and performing water quality monitoring based on the abnormal water mass parameter information to obtain the water quality status, wherein the water quality status includes a normal status and an abnormal pollution status.

[0012] In one implementation of the first aspect, the process of performing image stitching and spatial correction based on the original water body images to obtain a corrected image includes: performing image stitching based on the original water body images to obtain a stitched water body image; performing image distortion correction based on the stitched water body image using camera parameters to obtain a first corrected image; the camera parameters include a camera intrinsic parameter matrix, distortion coefficients, and an extrinsic parameter matrix; and performing tilt image correction based on the first corrected image to obtain a corrected image.

[0013] In one implementation of the first aspect, performing tilt image correction based on the first corrected image to obtain the corrected image includes: obtaining a perspective transformation matrix; and performing tilt image correction based on the first corrected image using the perspective transformation matrix to obtain the corrected image.

[0014] In one implementation of the first aspect, the process of inverting water quality parameters based on the corrected image to obtain a water quality monitoring parameter image includes: performing quantitative inversion based on the corrected image using a water quality inversion algorithm model to obtain a water quality monitoring parameter image; the water quality monitoring parameter image includes at least an image of chlorophyll a concentration, an image of potassium permanganate index, an image of total nitrogen concentration, an image of total phosphorus concentration, and an image of turbidity.

[0015] In one implementation of the first aspect, the water quality inversion algorithm model is a multiple regression model based on selected key channels, and its expression is: ,in, This represents the i-th original water body image data. , … This indicates the reflectivity of the selected key channel. , , … Represents the regression coefficient. The term indicates the error term. The selected key channel bands include, but are not limited to, 416.0nm, 449.4nm, 490.9nm, 554.7nm, 658.4nm, 676.6nm, 716.0nm, and 839.7nm.

[0016] In one implementation of the first aspect, identifying abnormal water masses based on the water quality monitoring parameter image and obtaining abnormal water mass parameter information includes: acquiring the water quality monitoring parameter image at the current moment and a preset standard image; performing image registration on the water quality monitoring parameter image at the current moment and the preset standard image to obtain a registered image; extracting pixel differences based on the registered image, extracting abnormal regions based on the extracted image pixel difference data, and obtaining the area and location of the abnormal water mass; determining the drift direction of the abnormal water mass based on the positional changes of the abnormal regions at consecutive moments, wherein the positional changes of the abnormal regions at consecutive moments are obtained from the original water body images at consecutive moments.

[0017] In one implementation of the first aspect, pixel difference extraction is performed based on the registered image, and abnormal regions are extracted based on the extracted image pixel difference data, including: performing grayscale conversion on the registered image to obtain a grayscale image of the corrected image; performing grayscale conversion on a preset standard image to obtain a grayscale image of the standard image; obtaining image pixel difference data based on the grayscale images of the corrected image and the standard image; obtaining a comparison result based on the image pixel difference data and a preset threshold; if the image pixel difference data is greater than the preset threshold, the comparison result indicates that the corrected image contains abnormal water masses, and abnormal regions are extracted based on the image pixel difference data to obtain the area and location of the abnormal water masses; otherwise, the comparison result indicates that the corrected image does not contain abnormal water masses.

[0018] Secondly, this application provides a multispectral remote sensing water quality monitoring system, the system comprising: an acquisition module configured to acquire original water images of a target water area based on a shore-based fixed multispectral remote sensing device, wherein the multispectral remote sensing device is positioned at a preset angle to the horizontal plane of the shore; a correction module configured to perform image stitching and spatial correction based on each of the original water images to acquire a corrected image, wherein the spatial correction includes image distortion correction based on camera parameters and tilt image correction based on a perspective transformation matrix; a water quality parameter inversion module configured to perform water quality parameter inversion based on the corrected image to acquire a water quality monitoring parameter image; an abnormal water mass identification module configured to identify abnormal water masses based on the water quality monitoring parameter image to acquire abnormal water mass parameter information; wherein the abnormal water mass parameter information includes at least the area, location, and drift direction of the abnormal water mass, and the abnormal water mass identification includes image registration, pixel difference extraction, and abnormal region extraction; and a water quality monitoring module configured to monitor water quality based on the abnormal water mass parameter information to acquire the water quality status, wherein the water quality status includes a normal status and an abnormal pollution status.

[0019] Thirdly, this application provides an electronic device, including: a processor and a memory; wherein the memory is used to store a computer program; and the processor is used to execute the computer program stored in the memory to enable the electronic device to perform the multispectral remote sensing water quality monitoring method as described above.

[0020] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the multispectral remote sensing water quality monitoring method described above.

[0021] As described above, the multispectral remote sensing water quality monitoring method, system, electronic device, and medium of this application have the following beneficial effects:

[0022] This application provides a multispectral remote sensing water quality monitoring method. It acquires tilted original water images of a target water area using a shore-based, tilted multispectral remote sensing device fixed at a certain angle. The multispectral remote sensing device is positioned at a preset angle to the horizontal plane of the shore. The original water images are tilted. Based on the original water images, image stitching and spatial correction are performed to obtain corrected images. The spatial correction includes image distortion correction based on camera parameters and tilt image correction based on a perspective transformation matrix. Water quality parameters are inverted based on the corrected images to obtain water quality monitoring parameter images. Abnormal water masses are identified based on the water quality monitoring parameter images to obtain abnormal water mass parameter information. The abnormal water mass parameter information includes at least the area, location, and drift direction of the abnormal water mass. The abnormal water mass identification includes image registration, pixel difference extraction, and abnormal region extraction. Water quality monitoring is performed based on the abnormal water mass parameter information to obtain the water quality status, which includes a normal state and an abnormally polluted state. This application utilizes a multispectral remote sensing device deployed at a certain angle to capture raw water images. Compared to vertical imaging, this expands the field of view for a single monitoring session, resulting in a larger monitoring range than vertical multispectral camera imaging. This enables continuous, large-scale water quality monitoring. Spatial correction of the tilted images improves the accuracy of anomalous water mass area calculation and location determination. Furthermore, based on the area, location, and drift direction of anomalous water masses, it can determine whether pollution anomalies exist in the target water area and analyze the degree of pollution and suspected source direction. This achieves continuous water quality anomaly identification and monitoring in shore-based scenarios, assisting in the determination of pollutant sources at monitoring sections, and is not limited by time, location, or weather. Attached Figure Description

[0023] Figure 1 The diagram shown is a hardware scene schematic of the multispectral remote sensing water quality monitoring device described in the embodiments of this application.

[0024] Figure 2 The diagram shown is a flowchart of the multispectral remote sensing water quality monitoring method described in the embodiments of this application.

[0025] Figure 3 The image shown is a single original water body image as described in the embodiments of this application.

[0026] Figure 4 The image shown is a schematic diagram of the image result after stitching and spatial correction of 12 single original water body images as described in the embodiment of this application.

[0027] Figure 5 The example shown here refers to the embodiments described in this application. Figure 3 Remote sensing inversion map of chlorophyll a concentration from a single original water body image.

[0028] Figure 6The example shown here refers to the embodiments described in this application. Figure 4 Remote sensing inversion map of chlorophyll a concentration after splicing and spatial correction.

[0029] Figure 7 This is a schematic diagram of the structure of the multispectral remote sensing water quality monitoring system described in the embodiments of this application.

[0030] Figure 8 The diagram shown is a structural schematic of the electronic device described in an embodiment of this application.

[0031] Figure 9 The image shown is a physical front view of the electronic device described in the embodiments of this application.

[0032] Figure 10 The image shown is a rear view of the electronic device described in the embodiments of this application.

[0033] Component designation explanation

[0034] 1 Multispectral remote sensing water quality monitoring device 4 electronic devices 11 Processing unit 41 processor 12 storage unit 42 memory 13 Input devices 421 RAM 14 Output devices 422 Cache memory 2 Multispectral remote sensing equipment 423 Storage System 3 Multispectral remote sensing water quality monitoring system 424 Utility tools 31 Get Module 4241 Program Module 32 Calibration module 43 bus 33 Water quality parameter inversion module 44 I / O interface 34 Abnormal water mass identification module 45 Network adapter 35 Water quality monitoring module 6 Stretch Radiation Calibration Plate 5 gimbal 8 support 7 Surveillance cameras S11~S15 step Detailed Implementation

[0035] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0036] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic technical solutions of this application. The drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0037] The following embodiments of this application provide a multispectral remote sensing water quality monitoring method, system, electronic device, and medium, which solves the problems of insufficient spatial coverage, poor temporal continuity, and severe weather-related constraints in existing water quality monitoring technologies.

[0038] The multispectral remote sensing water quality monitoring method provided in this application embodiment can be operated in a multispectral remote sensing water quality monitoring device. Figure 1 For example, Figure 1This is a hardware block diagram of a multispectral remote sensing water quality monitoring device for running the aforementioned multispectral remote sensing water quality monitoring method. The multispectral remote sensing water quality monitoring device includes, but is not limited to, a processing unit 11 and a storage unit 12. The processing unit 11 and the storage unit 12 are connected via a bus.

[0039] Storage unit 12 is the non-transitory computer-readable storage medium provided in this application. The storage unit stores instructions executable by at least one processing unit to cause the at least one processing unit 11 to execute the multispectral remote sensing water quality monitoring method provided in this application. The non-transitory computer-readable storage medium of this application stores computer instructions for causing a computer to execute the multispectral remote sensing water quality monitoring method provided in this application.

[0040] Storage unit 12 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function. The data storage area may store image data required by the multispectral remote sensing water quality monitoring device and data created based on the use of electronic equipment determined by multispectral remote sensing water quality monitoring. Furthermore, storage unit 12 may include high-speed random access memory and non-transient memory, such as at least one disk storage device, flash memory device, or other non-transient solid-state storage device. In some embodiments, storage unit 12 may optionally include memory remotely located relative to processing unit 11. This remote memory can be connected to the multispectral remote sensing water quality monitoring device 1 determined by multispectral remote sensing water quality monitoring via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0041] The multispectral remote sensing water quality monitoring device 1 also includes an input device 13 and an output device 14. The input device 13 can receive image data input from the multispectral remote sensing device 2, such as original RGB true-color images of water bodies and multispectral remote sensing data. These original RGB true-color images of water bodies and multispectral remote sensing data can be stored in the storage unit 12, so that the processing unit 11 can perform image correction, abnormal water mass identification and water quality monitoring based on these original RGB true-color images of water bodies and multispectral remote sensing data, and output the monitored water quality status through the output device 14.

[0042] The input device 13 may include, but is not limited to, a multispectral camera and a multispectral display. The output device 14 may include, but is not limited to, a liquid crystal display (LCD), a light-emitting diode (LED) display, a plasma display, and a touch screen. This application does not limit the scope of the embodiments.

[0043] In the embodiments of this application, the above-mentioned components of the multispectral remote sensing water quality monitoring device 1 and Figure 1Other components, not shown, can also be connected to each other, for example, via a bus. It should be understood that... Figure 1 The block diagram of the computing device shown is for illustrative purposes only and is not intended to limit the scope of this application. Those skilled in the art can add or replace other components as needed.

[0044] The multispectral remote sensing water quality monitoring device 1 can be any type of stationary or mobile computing device, including mobile computers or mobile computing devices (e.g., tablet computers, personal digital assistants, laptop computers, notebook computers, netbooks, etc.), mobile phones (e.g., smartphones), wearable computing devices (e.g., smartwatches, smart glasses, etc.) or other types of mobile devices, or stationary computing devices such as desktop computers or PCs. The multispectral remote sensing water quality monitoring device 1 can also be a mobile or stationary server. This application does not limit the scope of the embodiments.

[0045] The technical solutions in the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0046] like Figure 2 As shown in the figure, this embodiment provides a multispectral remote sensing water quality monitoring method, which includes the following steps S11 to S14.

[0047] Step S11: Acquire raw water images of the target water area using a shore-based, fixedly installed multispectral remote sensing device. The multispectral remote sensing device is positioned at a preset angle to the horizontal plane of the shore base; the raw water image is an oblique image. The observation axis of the multispectral remote sensing device is positioned at a 25° angle to the horizontal plane and points downwards towards the target water area. The installation height is 2m to 20m, and it has 8 spectral channels covering the visible to near-infrared bands, with a band range of 410nm-900nm. The overall spatial resolution is better than 0.05m, thus achieving continuous, large-scale water quality monitoring without being limited by time, location, or weather.

[0048] In some embodiments, this application utilizes remote sensing equipment, such as a multispectral camera equipped with the optimal sensitive wavelength for water quality, installed on shore. The multispectral camera is positioned at a certain angle to obtain a larger monitoring range than when the multispectral camera is vertically positioned, thereby achieving continuous, large-scale water quality monitoring and assisting in the determination of pollutant sources at monitoring sections. This application uses a multispectral camera positioned at a certain angle to capture raw water images of rivers and lakes. For example, Figure 3 The image shown is a schematic diagram of the original water body containing abnormal water masses as described in the embodiments of this application.

[0049] In some embodiments, the observation axis of the multispectral remote sensing device changes with the gimbal. When observing the target water area below, the observation axis of the multispectral remote sensing device forms a 25° angle downwards with the horizontal plane; when the multispectral remote sensing device performs radiometric calibration, the telescopic radiometric calibration plate extends, and the observation axis of the multispectral remote sensing device forms a 71° angle downwards with the horizontal plane; when the multispectral remote sensing device photographs the sky, the observation axis of the multispectral remote sensing device forms a 45° angle upwards with the horizontal plane. The installation height is 2m to 20m, the number of spectral channels is 8, the band range is 410-900nm, and the overall spatial resolution within the target monitoring area is better than 0.05m.

[0050] Step S12: Perform image stitching and spatial correction based on the original water body images to obtain the corrected image. The original water body images are tilted images of river and lake water bodies captured by multispectral cameras deployed at a certain angle. The spatial correction includes image distortion correction based on camera parameters and tilt image correction based on perspective transformation matrix. The spatial correction is a dedicated correction process suitable for tilted observation scenarios, used to reduce geometric distortion and perspective deformation caused by tilted imaging.

[0051] In some embodiments, this application employs a tilted arrangement of multispectral sensors, positioning them at a certain angle to the water surface for large-area water quality parameter monitoring. However, the tilted images cannot directly capture the actual water area. This application addresses this by constructing a tilted image spatial correction algorithm. By sequentially performing camera calibration, distortion correction, and perspective transformation on the tilted images acquired by the multispectral camera, a corrected image corresponding to the actual spatial location of the water surface is obtained, thereby improving the accuracy of calculating the area, location, and spatial distribution of abnormal water masses and water quality parameters. Specifically, the image spatial correction process includes: acquiring the camera intrinsic parameter matrix, extrinsic parameter matrix, and distortion coefficients; performing distortion correction on the tilted original water image; selecting corresponding points between the corrected image and the target water surface; solving the perspective transformation matrix; and performing spatial transformation on the image based on the perspective transformation matrix to obtain the corrected image.

[0052] In one embodiment of this application, image stitching and spatial correction are performed based on the original water body images to obtain the corrected image, including the following steps S121 to S122.

[0053] Step S121: Perform image stitching based on each of the original water body images to obtain the stitched water body image.

[0054] Step S122: Based on the stitched water image, perform image distortion correction using camera parameters to obtain a first corrected image; the camera parameters include the camera intrinsic parameter matrix, distortion coefficients, and extrinsic parameter matrix.

[0055] Step S123: Perform tilt image correction based on the first corrected image to obtain the corrected image.

[0056] Figure 4 The image shown is a schematic diagram illustrating the image result after stitching and spatial correction of the original water body image as described in the embodiments of this application. Figure 4 As shown, in some embodiments, this application utilizes camera calibration to correct image distortion in the original water body image, obtaining a distortion-corrected image, i.e., a first corrected image. Then, perspective transformation is used to perform tilt image correction on the first corrected image to obtain a spatially corrected image (see [reference]). Figure 4 ).

[0057] In one embodiment of this application, image distortion correction is performed using camera parameters based on the stitched water image to obtain a first corrected image, which includes the following steps S1221 to S1222.

[0058] Step S1221: Obtain camera parameters; the camera parameters include the camera intrinsic matrix and distortion coefficients; the camera intrinsic matrix includes the camera intrinsic parameter matrix and the extrinsic parameter matrix.

[0059] Step S1222: Based on the camera's intrinsic matrix and the distortion coefficients, perform image distortion correction on the stitched water image to obtain a first corrected image.

[0060] In one embodiment of this application, tilt image correction is performed based on the first corrected image to obtain the corrected image, which includes the following steps S1231 to S1232.

[0061] Step S1231: Obtain the perspective transformation matrix.

[0062] Step S1232: Based on the first corrected image, perform tilt image correction using the perspective transformation matrix to obtain the corrected image.

[0063] In some embodiments, camera calibration methods can be broadly categorized into three types: conventional calibration, self-calibration, and active vision calibration. Conventional calibration methods require the use of a known high-precision calibration target, providing high accuracy at the cost of increased operational complexity. These methods are suitable for scenarios demanding stringent accuracy. Self-calibration techniques abandon the use of physical calibration workpieces, relying instead on the inherent constraints of camera parameters for calculation. However, while offering simplicity, they exhibit lower robustness and are best suited for applications where lower accuracy is acceptable. Active vision calibration involves manipulating the camera through specific motions and calculating camera parameters by correlating images with motion data. This method is independent of the calibration target, offering algorithmic simplicity and higher robustness. However, it requires expensive equipment and is less suitable for unstructured environments. This application achieves spatial calibration of tilted images by combining camera calibration and perspective transformation.

[0064] First, this application corrects image distortion of tilted images through camera calibration. Calculating the camera's intrinsic matrix and radial distortion coefficients is the primary objective of the calibration process. The camera's intrinsic matrix is ​​a 3×3 matrix, and five variables are used... Let's define the 3×3 camera intrinsic matrix. In this case, the pixel size of the focal length in the x and y directions is represented as... When the x-axis and y-axis are perfectly perpendicular to each other, the tilt parameter... It equals 0. Furthermore, the pixel coordinates of the optical center are determined by... This indicates that the radial distortion coefficient is also a key element of the calibration procedure. and The following formula (1) is the expression for the camera's intrinsic matrix:

[0065] Formula (1)

[0066] The actual steps of the calibration process include the following steps S12101 to S12107.

[0067] Step S12101: Prepare a calibration board. Print a checkerboard pattern image with 12 columns and 9 rows as the calibration board for the calibration image, and paste it onto a flat surface. A black and white checkerboard pattern image can be used because its corners are easy to detect. Print the calibration board and fix it on a rigid flat surface to avoid deformation.

[0068] Step S12102: Capture calibration images. Fix the camera in place, manipulate the position of the plane, and capture three or more checkerboard pattern images from different angles, that is, capture the calibration board from different angles and distances to cover various areas of the camera's field of view.

[0069] Step S12103: Corner detection, identifying feature points (checkerboard corners) of the checkerboard pattern in each image. This application can detect checkerboard corners in each image using computer vision tools (such as OpenCV).

[0070] Step S12104: Calculate camera parameters. Calculate the intrinsic and extrinsic parameters of the camera based on the detected feature points.

[0071] Step S12105: Determine the radial distortion coefficient using the least squares method.

[0072] Step S12106: Optimize the results using the maximum likelihood method to improve the accuracy of the estimation.

[0073] Step S12107: Image distortion can be corrected through camera calibration.

[0074] In some embodiments, perspective transformation is a fundamental technique in computer vision and graphics that helps project visual information from one plane to another while preserving the linearity of lines within an image. This process is crucial for applications such as image correction, depth-of-field simulation, and 3D reconstruction.

[0075] The implementation of perspective transformation typically involves four key points. These points define a rectangular region in the source image and are correspondingly mapped to specified points on the target plane. Depending on the application, these points might be corners of the source image (i.e., the original water body image), where matching points cannot be selected for pure water bodies, and where the original water body image is not an image of pure water. The perspective transformation matrix is ​​a 3×3 matrix calculated based on these four points, providing a transformation from pixel coordinates in the source image to corresponding points on the target plane (i.e., the true plane coordinates). Every pixel in the source image undergoes this transformation to accurately remap onto the target plane.

[0076] Mathematically, perspective transformation is represented by the following formula (2):

[0077] Formula (2)

[0078] in The perspective transformation matrix is ​​specifically represented by the following formula (3):

[0079] Formula (3)

[0080] With matrix ( After multiplication, the original image coordinates After transformation, new image coordinates are obtained. Since the focus of this application is on two-dimensional images, it is necessary to change the source coordinates from... Convert to target image coordinates Thus, the following relationship is obtained (see formula (4)):

[0081] Formula (4)

[0082] Expanding the equation, we separate the terms involving (X) and (Y) as shown in formula (5):

[0083] Formula (5)

[0084] The two equations in formula (5) above have eight unknown coefficients that need to be solved. Therefore, to successfully calculate the perspective transformation matrix, four pairs of corresponding points must be selected between the source image and the target image. After calculating the perspective transformation matrix, the perspective transformation of formula (2) is used for tilt correction. By substituting the point coordinates of the original water body image into formula (2), the tilt-corrected image can be obtained.

[0085] This application constructs a tilt image spatial correction algorithm, which realizes streamlined and efficient correction of images under various on-site environments. The correction efficiency is high and the applicable scenarios are wide.

[0086] Step S13: Based on the corrected image, perform water quality parameter inversion to obtain a water quality monitoring parameter image.

[0087] In one embodiment of this application, the process of inverting water quality parameters based on the corrected image to obtain a water quality monitoring parameter image includes the following steps: S131, quantitative inversion is performed based on the corrected image using a water quality inversion algorithm model to obtain a water quality monitoring parameter image; the water quality monitoring parameter image includes at least a chlorophyll a concentration image (see...). Figure 6 Images of potassium permanganate index, total nitrogen concentration, total phosphorus concentration, and turbidity. Figure 5 The image shown is a schematic diagram of chlorophyll a concentration from the original water body image described in the embodiments of this application. Figure 6 The image shown is a schematic diagram of the chlorophyll a concentration image retrieved by remote sensing according to an embodiment of this application.

[0088] In some embodiments, the water quality monitoring parameter images include at least colored dissolved organic matter images and chemical oxygen demand images.

[0089] In some embodiments, the water quality inversion algorithm model is a multiple regression model based on selected key channels, and its expression is:

[0090]

[0091] in, This represents the i-th original water body image data. , … This indicates the reflectivity of the selected key channel. , , … Represents the regression coefficient. The term indicates the error term. The selected key channel bands include, but are not limited to, 416.0nm, 449.4nm, 490.9nm, 554.7nm, 658.4nm, 676.6nm, 716.0nm, and 839.7nm.

[0092] In some embodiments, this application acquires multispectral remote sensing image data of river and lake water quality parameters by installing remote sensing equipment such as multispectral cameras equipped with the optimal sensitive wavelengths for water quality on a shore-based platform. The multispectral remote sensing image data includes several original water body images from different angles and at different times.

[0093] In this embodiment, multi-source sensors (video probes, multispectral sensors) are uniformly managed, decoded, and preprocessed; multi-source data are analyzed to enable early warning of water quality anomalies.

[0094] First, this application performs data decoding on the captured multispectral remote sensing image data. This application establishes a multi-protocol parsing library to address the differentiated decoding needs of video probes and multispectral sensors.

[0095] The raw image data output by the sensors (such as dissolved oxygen, pH, conductivity, etc.) undergoes unit unification and encoding conversion. The standard score normalization (Z-Score) method is used to map multi-source data to a unified numerical range, solving the problem of inconsistent dimensions.

[0096] Secondly, this application performs data preprocessing on the decoded multispectral remote sensing image data. Geometric correction and atmospheric correction are used to eliminate the influence of environmental interference (such as changes in illumination, temperature and humidity fluctuations) on the optical sensor. Outliers are smoothed using sliding window mean filtering or Kalman filtering algorithms to improve the signal-to-noise ratio.

[0097] Based on the multispectral remote sensing data, a water quality inversion algorithm model is used to perform quantitative inversion and obtain water quality monitoring parameter images. These images include at least images of chlorophyll a concentration, potassium permanganate index, total nitrogen concentration, total phosphorus concentration, and turbidity.

[0098] In some embodiments, the water quality inversion algorithm model is a multiple regression model based on selected key channels, and the specific formula is as follows (6):

[0099] Formula (6)

[0100] in, Indicates the first One original water body image data, , … This indicates the reflectivity of the selected key channel. , , … Represents the regression coefficient. The term "error" is indicated. The key channels are eight discrete bands set by the multispectral remote sensing equipment within the range of 410nm-900nm. Based on principles such as priority of sensitive bands and band independence, the optimal key channel bands are selected as: 416.0nm, 449.4nm, 490.9nm, 554.7nm, 658.4nm, 676.6nm, 716.0nm, and 839.7nm. Chlorophyll a shows absorption valleys in the 449.4nm blue light band and the 676.6nm red light band, a secondary absorption peak in the 490.9nm band, and a reflection peak in the 716.0nm near-infrared band; turbidity is relatively sensitive in the 554.7nm band; dissolved organic matter is relatively sensitive in the 416.0nm band. The 658.4nm peak corresponds to the scattering characteristic peak of suspended matter and is close to the shoulder of the chlorophyll a absorption valley. It is not easily saturated in water bodies with high turbidity and high algae content and is stable in response to particulate pollutants and nutrient carriers.

[0101] In some embodiments, this application, based on a constructed water quality-optics database, employs empirical, semi-empirical, and machine learning methods to study the optimal sensitive bands for different water quality parameters and their relationship with spectral information. By constructing an inversion algorithm model, spectral information is transformed into specific water quality parameters.

[0102] In some embodiments, this application constructs a water quality-optics database containing various water quality parameters and spectral information by collecting water quality parameters, spectral reflectance, and environmental data of rivers and lakes. This database will be used to support subsequent data mining and model building. Correlation analysis is used to screen the bands most relevant to each water quality parameter and select key channels, as shown in the following formula (7):

[0103] Formula (7)

[0104] in, For the first Correlation between each band and water quality parameters For the first The first band The spectral reflectance value of each sample, For the first The average spectral reflectance of each band, The first water quality parameter Each sample value This represents the average value of the water quality parameters.

[0105] First, based on the selected key channels, an inversion algorithm is constructed using methods such as multiple regression. The relationship between water quality parameters and spectral data is established, as shown in the following formula (8):

[0106] Formula (8)

[0107] in, For water quality parameters, For the reflectivity of the selected key channel, , , … All are regression coefficients. This represents the error term. Based on principles such as prioritizing sensitive bands and band independence, the optimal key channel bands are selected as follows: 416.0nm, 449.4nm, 490.9nm, 554.7nm, 658.4nm, 676.6nm, 716.0nm, and 839.7nm.

[0108] This application utilizes a water quality inversion algorithm to convert acquired multispectral remote sensing data into specific water quality parameters, such as chlorophyll a concentration, potassium permanganate index, total nitrogen concentration, total phosphorus concentration, and turbidity. Water quality monitoring is then conducted based on these parameters to determine the current water quality status. Furthermore, this application's water quality inversion algorithm can filter out the optimal sensitive bands for different water quality elements across different water body types, thereby rapidly determining water quality and obtaining its status.

[0109] Step S14: Based on the water quality monitoring parameter image, identify abnormal water masses and obtain abnormal water mass parameter information; the abnormal water mass parameter information includes at least the area, location and drift direction of the abnormal water mass, and the abnormal water mass identification includes image registration, pixel difference extraction and abnormal region extraction.

[0110] In one embodiment of this application, the process of identifying abnormal water masses based on the water quality monitoring parameter image and obtaining abnormal water mass parameter information includes the following steps S141 to S144.

[0111] Step S141: Obtain the water quality monitoring parameter image and preset standard image at the current moment.

[0112] Step S142: Perform image registration on the water quality monitoring parameter image at the current moment and the preset standard image to obtain the registered image.

[0113] Step S143: Extract pixel differences based on the registered image, extract abnormal regions based on the extracted image pixel difference data, and obtain the area and location of abnormal water masses.

[0114] Step S144: Determine the drift direction of the abnormal water mass based on the positional changes of the abnormal region at continuous time intervals. The positional changes of the abnormal region at continuous time intervals are obtained from the original water body images at continuous time intervals.

[0115] In one embodiment of this application, pixel difference extraction is performed based on the registered image, and abnormal regions are extracted based on the extracted image pixel difference data, including the following steps S1431 to S1436.

[0116] Step S1431: Perform grayscale conversion based on the registered image to obtain the grayscale image of the corrected image.

[0117] Step S1432: Perform grayscale conversion based on a preset standard image to obtain a grayscale image of the standard image.

[0118] Step S1433: Obtain image pixel difference data based on the grayscale image of the corrected image and the grayscale image of the standard image.

[0119] Step S1434: Obtain the comparison result based on the image pixel difference data and the preset threshold.

[0120] Step S1445: If the image pixel difference data is greater than the preset threshold, the comparison result indicates that there is an abnormal water mass in the corrected image, and the abnormal region is extracted based on the image pixel difference data to obtain the area and location of the abnormal water mass.

[0121] Step S1446: Otherwise, the comparison result is that the corrected image does not contain abnormal water masses.

[0122] In some embodiments, this application defines abnormal water masses as those whose color differs significantly from normal water bodies, or as debris appearing in the water. Because their shape and color are not fixed, they cannot be detected using machine learning algorithms. However, the appearance of abnormal water masses is often dynamic, and they appear, move, and disappear in surveillance videos. To address this characteristic, difference detection can be used to identify abnormal water masses. The main identification principle is summarized as follows: An image is captured at regular time intervals, and the pixel changes between two consecutive images are compared to determine if there is any moving object in the scene.

[0123] In one implementation, this application first converts both the corrected image and the preset standard image into grayscale images. The two grayscale images are then subtracted, and the average difference is calculated. A threshold is then set. If this average difference is greater than the threshold, it is considered that an object is moving in the image; otherwise, it is considered that no object is moving. The purpose of setting the threshold is to prevent minor changes in the image from reducing the recognition effect.

[0124] If two similar images are in the same overall position and differ only in some places, then simply subtracting them will reveal the differences. However, real-world situations are often much more complex. Due to differences in shooting time and angle, two images may have differences in perspective, grayscale, rotation, scale, etc. Therefore, finding the transformation relationship between the two images is a prerequisite for implementing this function.

[0125] First, this application obtains the homography transformation matrix through feature matching and uses the homography transformation matrix to represent the transformation relationship between two images. It is determined that at least four matching pairs are required for homography transformation. The SIFT descriptor is used in the image feature matching part because SIFT possesses good gray-level invariance, rotation invariance, and scale invariance. SIFT feature points are found in the two images and matched to obtain a pre-matching set. Then, some matching pairs are eliminated based on their distance ratio with adjacent matching pairs. The remaining matching pairs are considered relatively good and used to find the homography transformation.

[0126] Secondly, this application establishes a homography transformation and determines the detection area. The homography transformation requires at least four point pairs to calculate. The homography transformation matrix is ​​obtained using the findHomography function in OpenCV, and the detection points in the reference image (a video screenshot from a certain time ago) are mapped to the corresponding positions in the image to be detected (a video screenshot at the current time). The range of the monitoring point positions in the reference image is determined by the location of the region where a correct matching pair exists (the wider the coverage of the matching pair, the wider the coverage of the monitoring points). Then, point pairs with differences greater than a threshold are output.

[0127] Finally, abnormal water mass detection is achieved by clustering pairs of points with differences greater than a threshold. After clustering the detected difference points, the difference locations are obtained, thus enabling the detection of abnormal water masses.

[0128] Step S15: Based on the abnormal water mass parameter information, perform water quality monitoring to obtain the water quality status.

[0129] This application identifies abnormal water masses to determine whether they exist in the water. If they are present, the water is contaminated; otherwise, it is uncontaminated. Simultaneously, during the identification of abnormal water masses, parameters such as surface velocity, water mass area, and drift direction are measured to determine the direction of the water pollution source.

[0130] Table 1 below compares the technical solution of this application with existing technologies. Compared with existing technologies, this application has significantly improved in terms of correction error, abnormal water mass identification rate, single effective monitoring range, and accuracy in determining the direction of pollution sources. This application uses a shore-based, fixed-installation multispectral remote sensing device to capture original water body images. The observation axis of the multispectral remote sensing device is positioned at a 25° angle to the horizontal plane and facing the target water area downwards. The installation height is 2m to 20m, and the number of spectral channels is 8, covering the visible to near-infrared bands. The band range is 410nm-900nm, and the overall spatial resolution is better than 0.05m, thereby achieving the purpose of continuous large-scale water quality monitoring without being limited by time, location, or weather. At the same time, by performing image stitching and spatial correction on the original water body images, the acquired tilted images are corrected to images under orthophoto projection, which can improve the accuracy of area measurement. This application obtains water quality monitoring parameter products by performing water quality parameter inversion on the corrected images, and then analyzes the water quality monitoring parameter products to identify abnormal water masses, thereby assessing water quality changes and pollution status.

[0131] Table 1

[0132] index Traditional vertical perspective monitoring methods This application method Increase Correction error 11.5% 6.4% Reduced by 44.3% Abnormal water mass recognition rate 86.1% 91.8% Increased by 5.7 percentage points Single effective monitoring range 350 m² 500 m² Increased by 42.9% Accuracy of determining the direction of pollution source 80.2% 87.6% An increase of 7.4 percentage points

[0133] The scope of protection of the multispectral remote sensing water quality monitoring method described in this application is not limited to the order of steps listed in this embodiment. Any solution implemented by adding, subtracting, or replacing steps in the prior art based on the principles of this application is included within the scope of protection of this application.

[0134] This application also provides a multispectral remote sensing water quality monitoring system, which can implement the multispectral remote sensing water quality monitoring method described in this application. However, the implementation device of the multispectral remote sensing water quality monitoring method described in this application includes, but is not limited to, the structure of the multispectral remote sensing water quality monitoring system listed in this embodiment. All structural modifications and substitutions of the prior art made based on the principles of this application are included within the protection scope of this application.

[0135] like Figure 7 As shown, this embodiment provides a multispectral remote sensing water quality monitoring system. The multispectral remote sensing water quality monitoring system 3 includes: an acquisition module 31, a correction module 32, a water quality parameter inversion module 33, an abnormal water mass identification module 34, and a water quality monitoring module 35.

[0136] The acquisition module 31 is configured to acquire the original water body image of the target water area based on a multispectral remote sensing device fixedly installed on the shore, wherein the multispectral remote sensing device is arranged at a preset angle to the horizontal plane of the shore.

[0137] The correction module 32 is configured to perform image stitching and spatial correction based on each of the original water body images to obtain the corrected image. The spatial correction includes image distortion correction based on camera parameters and tilt image correction based on perspective transformation matrix.

[0138] The water quality parameter inversion module 33 is configured to perform water quality parameter inversion based on the corrected image to obtain a water quality monitoring parameter image.

[0139] The abnormal water mass identification module 34 is configured to identify abnormal water masses based on the water quality monitoring parameter image and obtain abnormal water mass parameter information; the abnormal water mass parameter information includes at least the area, location and drift direction of the abnormal water mass, and the abnormal water mass identification includes image registration, pixel difference extraction and abnormal region extraction.

[0140] The water quality monitoring module 35 is configured to monitor water quality based on the abnormal water mass parameter information and obtain the water quality status, which includes normal status and abnormal pollution status.

[0141] It should be noted that the functions or operations of the acquisition module 31, correction module 32, water quality parameter inversion module 33, abnormal water mass identification module 34, and water quality monitoring module 35 described in this embodiment correspond one-to-one with the steps in the multispectral remote sensing water quality monitoring method described above, and therefore will not be repeated here.

[0142] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, or methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or modules or units may be electrical, mechanical, or other forms.

[0143] The modules / units described as separate components may or may not be physically separate. The components shown as modules / units may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules / units can be selected to achieve the objectives of the embodiments of this application, depending on actual needs. For example, the functional modules / units in the various embodiments of this application may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0144] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0145] like Figure 8 As shown, this embodiment provides an electronic device, which includes: a processor and a memory; wherein,

[0146] The memory is used to store computer programs;

[0147] The memory includes various media capable of storing program code, such as read-only memory (ROM), random access memory (RAM), magnetic disk, USB flash drive, memory card, or optical disk.

[0148] The processor is used to execute the computer program stored in the memory, so that the electronic device performs the multispectral remote sensing water quality monitoring method as described above.

[0149] Preferably, the processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0150] like Figure 8 As shown, the electronic device 4 of this application is manifested in the form of a general-purpose computing device, and its specific implementation products are diverse, including smart mobile devices such as mobile phones, laptops, tablets, smart wearable devices, in-vehicle systems, etc. The components of the electronic device may include, but are not limited to: one or more processors or processor 41, memory 42, and bus 43 connecting different system components (including memory 42 and processor 41).

[0151] Bus 43 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MCA) bus, the Extended Industry Standard Architecture (EISA) bus, the Video Electronics Standards Association Local Bus (VESA LocalBus), and the Peripheral Component Interconnect (PCI) bus.

[0152] Electronic device 4 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by a control terminal, including volatile and non-volatile media, removable and non-removable media.

[0153] Memory 42 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 421 and / or cache memory 422. The control terminal may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, memory 42 may be used to read and write non-removable, non-volatile magnetic media (…). Figure 8 Not shown; usually referred to as a "hard drive"). Although Figure 8 Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 43 via one or more data media interfaces. Memory 42 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of this disclosure.

[0154] A program / utility 424 having a set (at least one) of program modules 4241 may be stored, for example, in memory 42. Such program modules 4241 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 4241 typically perform the functions and / or methods described in the embodiments of this disclosure.

[0155] Electronic device 4 can also communicate with one or more external devices (e.g., keyboard, pointing device, display, etc.), and with one or more devices that enable the user to interact with the control terminal, and / or with any device that enables the control terminal to communicate with one or more other computing devices (e.g., network card, modem, etc.). This communication can be performed through input / output (I / O) interface 44. Furthermore, the control terminal can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 45. Figure 8 As shown, network adapter 45 communicates with other modules of the control terminal via bus 43. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with the control terminal, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, Redundant Array of Independent Disks (RAID) systems, tape drives, and data backup storage systems.

[0156] Figure 9 The image shown is a physical front view of the electronic device described in the embodiments of this application. Figure 10 The image shown is a rear view of the electronic device described in an embodiment of this application. Figures 9-10 As shown, the multispectral remote sensing device 2, electronic device 4, pan-tilt unit 5, telescopic radiometric calibration plate 6, and monitoring camera 7 are communicatively connected and mounted on the bracket 8. The multispectral remote sensing device 2 is mounted on the pan-tilt unit. In some embodiments, the observation axis of the multispectral remote sensing device changes with the pan-tilt unit. When observing the target water area below, the observation axis of the multispectral remote sensing device forms a 25° angle downwards with the horizontal plane; when the multispectral remote sensing device performs radiometric calibration, the telescopic radiometric calibration plate extends, and the observation axis of the multispectral remote sensing device forms a 71° angle downwards with the horizontal plane; when the multispectral remote sensing device photographs the sky, the observation axis of the multispectral remote sensing device forms a 45° angle upwards with the horizontal plane. The installation height is 2m to 20m, the number of spectral channels is 8, the band range is 410-900nm, and the overall spatial resolution within the target monitoring area is better than 0.05m.

[0157] This application also provides a computer-readable storage medium. Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing a processor. The program can be stored in a computer-readable storage medium, which is a non-transitory medium, such as random access memory, read-only memory, flash memory, hard disk, solid-state drive, magnetic tape, floppy disk, optical disc, and any combination thereof. The storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disc (DVD)), or a semiconductor medium (e.g., solid-state drive (SSD)).

[0158] This application embodiment may also provide a computer program product comprising one or more computer instructions. When the computer instructions are loaded and executed on a computing device, all or part of the processes or functions described in this application embodiment are generated. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, or data center to another via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means.

[0159] When the computer program product is executed by a computer, the computer performs the method described in the foregoing method embodiments. The computer program product can be a software installation package; when the foregoing method is required, the computer program product can be downloaded and executed on the computer.

[0160] In summary, the multispectral remote sensing water quality monitoring method, system, electronic device, and medium described in this application have the following beneficial effects:

[0161] This application acquires raw water body images captured by multispectral remote sensing equipment; performs image stitching and spatial correction based on the raw water body images to obtain corrected images; performs water quality parameter inversion based on the corrected images to obtain water quality monitoring parameter products, then analyzes the water quality monitoring parameter products to identify abnormal water masses and obtain abnormal water mass parameter information; and performs water quality monitoring based on the abnormal water mass parameter information to obtain water quality status. This application utilizes multispectral remote sensing equipment deployed at a certain angle to capture raw water images, achieving a larger monitoring range than vertical multispectral camera imaging. This enables continuous, large-scale water quality monitoring, assisting in the identification of pollutant sources at monitoring sections, reducing weather impacts, and supporting continuous monitoring around the clock. Simultaneously, by stitching and spatially correcting the raw water images, the acquired tilted images are transformed into orthophoto images, improving the accuracy of area calculation and providing more accurate data support for subsequent abnormal water body identification and water quality monitoring. Furthermore, this application identifies abnormal water masses in the corrected images and monitors water quality based on abnormal water mass parameters to obtain water quality status, thereby assessing water quality changes and pollution levels.

[0162] The descriptions of the processes or structures corresponding to the above figures each have their own emphasis. For parts of a process or structure that are not described in detail, please refer to the relevant descriptions of other processes or structures.

[0163] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A multispectral remote sensing method for water quality monitoring, characterized in that, include: The original water body image of the target water area is acquired by a multispectral remote sensing device that is fixedly installed on the shore. The multispectral remote sensing device is set at a preset angle to the horizontal plane of the shore. The original water body image is an inclined image. Image stitching and spatial correction are performed on the original water body images to obtain the corrected image. The spatial correction includes image distortion correction based on camera parameters and tilt image correction based on perspective transformation matrix. Water quality parameters are inverted based on the corrected image to obtain water quality monitoring parameter images; Based on the water quality monitoring parameter image, abnormal water masses are identified to obtain abnormal water mass parameter information; the abnormal water mass parameter information includes at least the area, location, and drift direction of the abnormal water mass, and the abnormal water mass identification includes image registration, pixel difference extraction, and abnormal region extraction; Water quality monitoring is performed based on the abnormal water mass parameter information to obtain the water quality status, which includes normal status and abnormal pollution status.

2. The multispectral remote sensing water quality monitoring method according to claim 1, characterized in that, Based on the original water body images, image stitching and spatial correction are performed to obtain the corrected images, including: Based on the original water body images, image stitching is performed to obtain the stitched water body image; Based on the stitched water image, image distortion correction is performed using camera parameters to obtain a first corrected image; the camera parameters include the camera intrinsic parameter matrix, distortion coefficients, and extrinsic parameter matrix. Based on the first corrected image, tilt image correction is performed to obtain the corrected image.

3. The multispectral remote sensing water quality monitoring method according to claim 2, characterized in that, Based on the first corrected image, tilt image correction is performed to obtain the corrected image, including: Obtain the perspective transformation matrix; Based on the first corrected image, the perspective transformation matrix is ​​used to perform tilt image correction to obtain the corrected image.

4. The multispectral remote sensing water quality monitoring method according to claim 1, characterized in that, Based on the corrected image, water quality parameters are inverted to obtain water quality monitoring parameter images, including: Based on the corrected image, a water quality inversion algorithm model is used to perform quantitative inversion to obtain water quality monitoring parameter images; the water quality monitoring parameter images include at least chlorophyll a concentration image, potassium permanganate index image, total nitrogen concentration image, total phosphorus concentration image, and turbidity image.

5. The multispectral remote sensing water quality monitoring method according to claim 4, characterized in that, The water quality inversion algorithm model is a multiple regression model based on selected key channels, and its expression is: ,in, This represents the i-th original water body image data. , … This indicates the reflectivity of the selected key channel. , , … Represents the regression coefficient. The term indicates the error term. The selected key channel bands include, but are not limited to, 416.0nm, 449.4nm, 490.9nm, 554.7nm, 658.4nm, 676.6nm, 716.0nm, and 839.7nm.

6. The multispectral remote sensing water quality monitoring method according to claim 1, characterized in that, Based on the water quality monitoring parameter images, abnormal water masses are identified, and abnormal water mass parameter information is obtained, including: Acquire the current water quality monitoring parameter image and the preset standard image; The water quality monitoring parameter image at the current moment and the preset standard image are image registered to obtain the registered image; Pixel differences are extracted from the registered image, and abnormal regions are extracted based on the extracted image pixel difference data to obtain the area and location of abnormal water masses. The drift direction of the abnormal water mass is determined by the positional changes of the abnormal region at consecutive time intervals, which are obtained from the original water body images at consecutive time intervals.

7. The multispectral remote sensing water quality monitoring method according to claim 6, characterized in that, Pixel differences are extracted from the registered image, and abnormal regions are extracted based on the extracted pixel difference data, including: Based on the registered image, perform grayscale conversion to obtain the grayscale image of the corrected image; The grayscale image of the standard image is obtained by performing grayscale conversion based on a preset standard image. Image pixel difference data is obtained based on the grayscale images of the corrected image and the standard image. The comparison results are obtained based on the image pixel difference data and a preset threshold. If the image pixel difference data is greater than the preset threshold, the comparison result indicates that there is an abnormal water mass in the corrected image, and the abnormal region is extracted based on the image pixel difference data to obtain the area and location of the abnormal water mass; Otherwise, the comparison result indicates that the corrected image does not contain abnormal water masses.

8. A multispectral remote sensing water quality monitoring system, characterized in that, include: The acquisition module is configured to acquire raw water images of the target water area based on a multispectral remote sensing device fixedly installed on the shore, wherein the multispectral remote sensing device is arranged at a preset angle to the horizontal plane of the shore. The correction module is configured to perform image stitching and spatial correction based on each of the original water body images to obtain a corrected image. The spatial correction includes image distortion correction based on camera parameters and tilt image correction based on perspective transformation matrix. The water quality parameter inversion module is configured to perform water quality parameter inversion based on the corrected image to obtain a water quality monitoring parameter image. An abnormal water mass identification module is configured to identify abnormal water masses based on the water quality monitoring parameter image and obtain abnormal water mass parameter information; the abnormal water mass parameter information includes at least the area, location and drift direction of the abnormal water mass, and the abnormal water mass identification includes image registration, pixel difference extraction and abnormal region extraction; The water quality monitoring module is configured to monitor water quality based on the abnormal water mass parameter information and obtain the water quality status, which includes normal status and abnormal pollution status.

9. An electronic device, characterized in that, include: Processor and memory; among which, The memory is used to store computer programs; The processor is used to execute the computer program stored in the memory to cause the electronic device to perform the multispectral remote sensing water quality monitoring method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the multispectral remote sensing water quality monitoring method according to any one of claims 1 to 7.

Images