An intelligent inspection and analysis method and system for water conservancy projects

Abstract

The present invention discloses an intelligent inspection and analysis method and system for water conservancy projects, the method comprising: collecting water surface images within a region; partitioning the water surface images within the region to obtain water area partition results; traversing the water area partition results to perform secondary water area analysis to obtain water area partition sampling points; traversing the water area partition sampling points to perform sampling to obtain water area sampling results; performing water quality testing on the water area sampling results to obtain a pollutant type set; based on the pollutant type set, performing pollution backtracking in combination with the reservoir deployment location to generate suspicious pollution sources; and inspecting and screening suspicious pollution sources based on the pollutant type set. The present invention partitions and grayscales water surface images, and establishes a water surface reference grayscale model through machine learning, thereby improving the accuracy of processing the light intensity and detection time parameters of the reservoir or water area surface, so as to more accurately evaluate and understand the water quality status, improve the efficiency of water quality monitoring, and ensure the rationality of the intelligent inspection plan setting.

Description

Technical Field

[0001] The present invention relates to the technical field of water conservancy project pollution inspection, and in particular to an intelligent inspection and analysis method and system for water conservancy projects. Background Art

[0002] Water pollution has always been a significant threat to human health and the ecological environment. Current automated inspection systems for water conservancy projects rely on pre-configured routes for automated water sampling and monitoring, often performed by unmanned boats or buoys. These systems are typically equipped with various sensors, such as those for dissolved oxygen, conductivity, pH, turbidity, and specific pollutant concentrations. These sensors transmit data in real time or periodically, providing managers with essential information on the overall water quality of the reservoir.

[0003] However, this automated inspection approach primarily focuses on automating and regularizing monitoring tasks, and is insufficiently responsive to the refined management needs of complex water environments. It only uniformly samples the reservoir's waters to generate water quality analysis results, but fails to account for regional variations caused by reservoirs bordering different pollution sources. For example, water quality near industrial outfalls or agricultural irrigation drains may be more polluted. These localized pollution hotspots are often difficult to accurately capture using uniform sampling strategies, resulting in less representative water quality analysis results.

[0004] Therefore, there is an urgent need for a solution that can reduce regional differences in automated inspection methods and improve analysis accuracy. Summary of the Invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the present invention provides an intelligent inspection and analysis method and system for water conservancy projects to solve the technical problems in current technology that there are regional differences in automated inspection methods for water conservancy projects, the water quality analysis results are less representative, and it is difficult to improve the accuracy of inspection plan settings.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention provides an intelligent inspection and analysis method for a water conservancy project, comprising:

[0009] Collect water surface images within the area;

[0010] Performing partition processing on the water surface image in the area to obtain a water area partition result;

[0011] Traversing the water area partition results to perform secondary water area analysis and obtain water area partition sampling points;

[0012] Traversing the water area partition sampling points to perform sampling and obtain water area sampling results;

[0013] Performing water quality testing on the sampling results of the water area to obtain a set of pollutant types;

[0014] Based on the pollutant type set and combined with the reservoir deployment location, pollution backtracking is performed to generate suspected pollution sources;

[0015] Conduct inspections and investigations on suspected pollution sources based on the set of pollutant types.

[0016] As a preferred solution of the intelligent inspection and analysis method for water conservancy projects described in the present invention, the water surface image in the area is partitioned to obtain the water area partition result, including:

[0017] Performing grayscale processing on the water surface image in the region to obtain a water surface grayscale image;

[0018] Acquiring detection and monitoring information of the water surface image in the area, wherein the detection and monitoring information includes a light intensity parameter and a detection time parameter;

[0019] Processing the illumination intensity parameter and the detection time parameter through a grayscale calibration network to obtain a water surface reference grayscale;

[0020] Performing multi-value processing on the water surface grayscale image according to the water surface reference grayscale to generate water surface multi-value feature distribution data;

[0021] The water surface grayscale image is subjected to neighborhood hierarchical clustering according to the water surface multi-valued feature distribution data to generate the water area partition result.

[0022] As a preferred solution of the intelligent inspection and analysis method for water conservancy projects described in the present invention, wherein: the water surface grayscale image is multi-valued according to the water surface reference grayscale to generate water surface multi-valued feature distribution data, including:

[0023] Based on the water surface reference grayscale, configuring a multi-valued distribution interval set;

[0024] Subtracting the water surface reference grayscale from each pixel grayscale of the water surface grayscale image to obtain a corresponding pixel grayscale difference vector;

[0025] All obtained pixel grayscale difference vectors are distributed in the multi-valued distribution interval set to obtain the multi-valued feature distribution data of the water surface.

[0026] As a preferred solution of the intelligent inspection and analysis method for water conservancy projects described in the present invention, the light intensity parameter and the detection time parameter are processed through a grayscale calibration network to obtain a water surface reference grayscale, including:

[0027] Obtaining a dataset within the region, wherein the dataset within the region includes a light intensity dataset, a detection time dataset, and a reference grayscale identification dataset;

[0028] Taking the illumination intensity dataset and the detection time dataset as input and the reference grayscale identification dataset as output, a preset amount of regional datasets are retrieved to configure a backbone decision network, wherein the backbone decision network has a first fitting residual vector, which is equal to the mean of the deviation vectors of several identification values and output values;

[0029] When the modulus value of the first fitting residual vector is greater than or equal to the residual modulus threshold, constructing a first output fitting layer to update the output layer of the backbone decision network according to the first fitting residual vector, calling the reservoir data set of a preset data volume to configure the first updated backbone decision network, wherein the first updated backbone decision network has a second fitting residual vector;

[0030] When the modulus value of the second fitting residual vector is less than the residual modulus threshold, setting the first updated backbone decision network to the grayscale calibration network;

[0031] Otherwise, a second output fitting layer is constructed according to the second fitting residual vector to update the output layer of the first updated backbone decision network.

[0032] As a preferred solution of the intelligent inspection and analysis method for water conservancy projects described in the present invention, the method of traversing the water area partition results to perform secondary water area analysis and obtain water area partition sampling points includes:

[0033] Obtaining first water area partition grayscale distribution information of the first water area partition of the water area partition result;

[0034] Calculate the variance value of the grayscale distribution information of the first water area partition and set it as the discrete parameter of the first water area partition;

[0035] When the discrete parameter of the first water area partition is greater than or equal to the discrete parameter threshold, the sampling points of the first water area partition are evenly distributed according to the preset interval, and the sampling points of the first water area partition are added;

[0036] When the discrete parameter of the first water area partition is less than the discrete parameter threshold, randomly deploying a representative sampling point for the first water area partition and setting it as the first water area partition sampling point;

[0037] Add the first water area partition sampling points to the water area partition sampling points.

[0038] As a preferred solution of the intelligent inspection and analysis method for water conservancy projects described in the present invention, pollution backtracking is performed based on the pollutant type set and combined with the reservoir deployment location to generate suspicious pollution sources, including:

[0039] Taking the reservoir deployment location as a constraint, collect the pollution point set upstream of the water area, the pollution point set upstream of the wind direction, and the pollution point set upstream of the terrain within the preset range;

[0040] Traversing the pollution point set upstream of the water area, the pollution point set upstream of the wind direction, and the pollution point set upstream of the terrain, performing historical pollution analysis on the pollutant type set, and generating several sets of pollution trigger probability sets;

[0041] Traverse the set of pollution points upstream of the water area, the set of pollution points upstream of the wind direction, and the set of pollution points upstream of the terrain, perform historical pollution analysis on the set of pollutant types, and generate several sets of pollution trigger probability sets, including:

[0042] Obtaining a first pollution point historical data set of a preset time zone for the water upstream pollution point set, the wind upstream pollution point set, and the terrain upstream pollution point set;

[0043] extracting a first pollutant type from the pollutant type set;

[0044] Counting the trigger frequency ratio of the first pollutant type in the first pollution point historical data set, setting it as the first pollutant type trigger probability, and adding the first pollution point pollution trigger probability set;

[0045] Add the first pollution point pollution trigger probability set to the several groups of pollution trigger probability sets.

[0046] As a preferred embodiment of the intelligent inspection and analysis method for water conservancy projects of the present invention, the method further includes sorting the pollution point set upstream of the water area, the pollution point set upstream of the wind direction, and the pollution point set upstream of the terrain according to the plurality of pollution trigger probability sets to obtain the suspected pollution source, including:

[0047] Obtain a first pollution trigger probability set of a first pollutant type, and extract a first group of pollution points that are greater than or equal to a trigger probability threshold;

[0048] Until the Mth pollution trigger probability set of the Mth pollutant type is obtained, the Mth group of pollution points greater than or equal to the trigger probability threshold is extracted;

[0049] The first group of pollution points is set as the suspected pollution source of the first pollutant type until the Mth group of pollution points is set as the suspected pollution source of the Mth pollutant type.

[0050] In a second aspect, the present invention provides an intelligent inspection and analysis system for a water conservancy project, comprising: an acquisition module for acquiring images of a water surface in an area;

[0051] A partitioning module, used for performing partitioning processing on the water surface image in the area to obtain a water area partitioning result;

[0052] The first traversal module is used to traverse the water area partition results to perform secondary water area analysis and obtain water area partition sampling points;

[0053] The second traversal module is used to traverse the water area partition sampling points to perform sampling and obtain water area sampling results;

[0054] A detection module, configured to perform water quality detection on the water sampling results to obtain a set of pollutant types;

[0055] A tracing module is used to perform pollution tracing based on the pollutant type set and the reservoir deployment location to generate suspected pollution sources;

[0056] The inspection module is used to inspect and investigate suspicious pollution sources based on the pollutant type set.

[0057] In a third aspect, the present invention provides a computing device, comprising:

[0058] memory and processor;

[0059] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the intelligent inspection and analysis method of the water conservancy project are implemented.

[0060] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the intelligent inspection and analysis method for the water conservancy project.

[0061] Compared with the existing technology, the beneficial effects of the present invention are as follows: the present invention partitions and grayscales water surface images, and establishes a water surface reference grayscale model through machine learning, thereby improving the accuracy of processing the light intensity and detection time parameters of the reservoir or water surface, so as to more accurately evaluate and understand the water quality conditions and improve the efficiency of water quality monitoring. Moreover, through intelligent analysis, it provides strong technical support for pollution source tracking and water quality management, and ensures the rationality and accuracy of the setting of the intelligent inspection plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 any creative work.

[0063] Figure 1 This is a schematic diagram of the overall process of the intelligent inspection and analysis method for water conservancy projects described in the first embodiment of the present invention. DETAILED DESCRIPTION

[0064] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0065] Example 1

[0066] Reference Figure 1 , as an embodiment of the present invention, provides an intelligent inspection and analysis method for a water conservancy project, comprising:

[0067] S100: Acquire water surface images within the area;

[0068] Specifically, the collection method may be to send out a detection drone from the machine nest when a preset inspection cycle is met to collect water surface images in the area, such as reservoir water surface images.

[0069] Among them, the preset inspection cycle is a pre-set inspection cycle performed at a fixed period, for example, it can be 2 hours or any time set according to the situation; the machine nest contains multiple inspection and detection drones.

[0070] S200: performing partition processing on the water surface image in the area to obtain a water area partition result;

[0071] Specifically, the reservoir surface image can be processed by a water surface partitioning component deployed on a detection drone.

[0072] It should be noted that zoning the water surface can facilitate subsequent water quality sampling in different water areas.

[0073] In the embodiment of the present application, the water surface image in the area is partitioned to obtain the water area partition results, including:

[0074] S201: grayscale processing is performed on the water surface image in the area to obtain a water surface grayscale image;

[0075] S202: Acquire detection and monitoring information of the water surface image in the area, wherein the detection and monitoring information includes a light intensity parameter and a detection time parameter;

[0076] S203: Processing the light intensity parameter and the detection time parameter through a grayscale calibration network to obtain a water surface reference grayscale;

[0077] In the embodiment of the present application, the illumination intensity parameter and the detection time parameter are processed by a grayscale calibration network to obtain a water surface reference grayscale, including:

[0078] S203-1: Obtaining a dataset within the region, wherein the dataset within the region includes a light intensity dataset, a detection time dataset, and a reference grayscale identification dataset;

[0079] It should be noted that the reference grayscale identification data set is the reference grayscale identification corresponding to the light intensity data and the detection time data.

[0080] S203-2: Using the illumination intensity dataset and the detection time dataset as input and the reference grayscale identification dataset as output, a preset amount of regional datasets are retrieved to configure a backbone decision network, wherein the backbone decision network has a first fitting residual vector, which is equal to the mean of the deviation vectors of several identification values and output values;

[0081] S203-3: When the modulus value of the first fitting residual vector is greater than or equal to the residual modulus threshold, constructing an output layer of a first output fitting layer updating backbone decision network based on the first fitting residual vector, retrieving a reservoir data set of a preset data volume to configure the first updated backbone decision network, wherein the first updated backbone decision network has a second fitting residual vector;

[0082] It should be noted that when the modulus of the first fitting residual vector is greater than or equal to the residual modulus threshold, it indicates that the network's fitting effect is insufficient and requires further optimization. Therefore, based on the first fitting residual vector, a first output fitting layer is constructed to update the output layer of the backbone decision network. A preset amount of reservoir data sets are then retrieved to configure the first updated backbone decision network. Specifically, a preset amount of regional data sets are retrieved to continue configuring and training the updated backbone decision network.

[0083] S203-4: When the modulus value of the second fitting residual vector is less than the residual modulus threshold, setting the first updated backbone decision network to a grayscale calibration network;

[0084] Otherwise, a second output fitting layer is constructed according to the second fitting residual vector to update the output layer of the first updated backbone decision network.

[0085] It should be noted that when the modulus of the second fitting residual vector is less than the residual modulus threshold, it indicates that the updated network performance is good. Therefore, the first updated backbone decision network is set as the grayscale calibration network to ensure the accuracy of the grayscale calibration network output results. Otherwise, if the updated network performance still does not meet the requirements, a second output fitting layer is constructed based on the second fitting residual vector to update the output layer of the first updated backbone decision network.

[0086] S204: performing multi-value processing on the water surface grayscale image according to the water surface reference grayscale to generate water surface multi-value feature distribution data;

[0087] In the embodiment of the present application, the water surface grayscale image is multi-valued according to the water surface reference grayscale to generate water surface multi-valued feature distribution data, including:

[0088] S204-1: Based on the water surface reference grayscale, configure a multi-valued distribution interval set;

[0089] It should be noted that the multi-valued distribution interval is the distribution interval of the pixel grayscale difference vector.

[0090] For example, if the pixel grayscale difference vector is less than 0 and greater than -5, the eigenvalue is -1; if it is greater than 0 and less than 5, the eigenvalue is 1. By constructing the pixel grayscale difference vector, the pixel grayscale can be quickly divided.

[0091] S204-2: Subtracting the water surface reference grayscale from each pixel grayscale of the water surface grayscale image to obtain a corresponding pixel grayscale difference vector;

[0092] S204-3: Distribute all obtained pixel grayscale difference vectors into a multi-valued distribution interval set to obtain multi-valued feature distribution data of the water surface.

[0093] Specifically, the process is:

[0094] Obtain the first pixel grayscale of the water surface grayscale image and subtract the water surface reference grayscale to obtain a first pixel grayscale difference vector;

[0095] Until the Nth pixel grayscale of the water surface grayscale image is obtained minus the water surface reference grayscale, the Nth pixel grayscale difference vector is obtained;

[0096] The first pixel grayscale difference vector to the Nth pixel grayscale difference vector are distributed in a multi-valued distribution interval set to obtain water surface multi-valued feature distribution data.

[0097] It should be noted that this step quickly divides pixel features by setting a multi-valued distribution interval, thereby improving the efficiency of subsequent pixel processing.

[0098] S205: Performing neighborhood hierarchical clustering on the water surface grayscale image according to the multi-valued characteristic distribution data of the water surface to generate a water area partition result.

[0099] Specifically, neighborhood hierarchical clustering is performed, that is, adjacent pixels with the same eigenvalues are clustered into one type of area, and pixels with the same eigenvalues are repeatedly traversed until the eigenvalues of any two adjacent areas are not obtained at the same time, and the water area partition result is generated.

[0100] S300: traversing the water area partition results to perform secondary water area analysis and obtain water area partition sampling points;

[0101] It should be noted that the sampling schemes for different water zones are obtained based on the degree of pixel discreteness.

[0102] In the embodiment of the present application, the water area partition results are traversed to perform secondary water area analysis to obtain water area partition sampling points, including:

[0103] S301: Obtaining grayscale distribution information of a first water area partition of a water area partition result;

[0104] It should be noted that the grayscale distribution information of the first water area partition is the grayscale distribution interval of pixels in the first water area partition image.

[0105] S302: Calculate the variance value of the grayscale distribution information of the first water area partition and set it as the discrete parameter of the first water area partition;

[0106] S303: When the discrete parameter of the first water area partition is greater than or equal to the discrete parameter threshold, evenly distribute sampling points in the first water area partition according to a preset interval, and add the sampling points to the first water area partition;

[0107] It should be noted that when the discrete parameter of the first water area partition is greater than or equal to the discrete parameter threshold, the pixel grayscale feature distribution in the corresponding water area partition is relatively discrete, the corresponding image features change significantly, and there are differences in the secondary water areas, so multiple sampling is required.

[0108] S304: When the discrete parameter of the first water area partition is less than the discrete parameter threshold, randomly deploy a representative sampling point for the first water area partition and set it as the first water area partition sampling point;

[0109] It should be noted that when the discrete parameter of the first water partition is less than the discrete parameter threshold, the pixel grayscale feature distribution in the corresponding water partition is relatively concentrated, the corresponding image feature changes are not obvious, and the difference in secondary water areas is not obvious. In this case, a representative sampling point is randomly deployed for the first water partition and set as the first water partition sampling point.

[0110] S305: Add the first water area partition sampling point to the water area partition sampling point.

[0111] It should be noted that step S300 can achieve the acquisition of sampling points in different water area partitions, thereby improving the efficiency and accuracy of water area sampling.

[0112] S400: traverse the water area partition sampling points to perform sampling and obtain water area sampling results;

[0113] S500: Perform water quality testing on the water sampling results to obtain a pollutant type set;

[0114] Specifically, water quality testing can obtain a set of pollutant types through various water quality testing methods such as chemical testing and biological testing. The pollutant type set includes various pollutant categories such as crude oil, gasoline, lead, mercury, cadmium, chromium, and arsenic.

[0115] S600: Based on the pollutant type set and combined with the reservoir deployment location, pollution backtracking is performed to generate suspected pollution sources;

[0116] It should be noted that the acquired pollutant type set and suspected pollution sources are sent to the inspection management terminal, which then conducts inspections and investigations of suspected pollution sources based on the pollutant type set. This improves the efficiency of water quality monitoring and, through intelligent analysis, provides strong technical support for pollution source tracking and water quality management, ensuring the rationality and accuracy of the intelligent inspection plan.

[0117] In the embodiment of the present application, based on the pollutant type set and combined with the reservoir deployment location, pollution backtracking is performed to generate suspicious pollution sources, including:

[0118] S601: Using the reservoir deployment location as a constraint, collect the collection of pollution points upstream of the water area, upstream of the wind direction, and upstream of the terrain within a preset range;

[0119] S602: Traversing the pollution point set upstream of the water area, the pollution point set upstream of the wind direction, and the pollution point set upstream of the terrain, performing historical pollution analysis on the pollutant type set, and generating several sets of pollution trigger probability sets;

[0120] It should be noted that by obtaining several sets of pollution trigger probability sets, the correspondence between pollutant types, pollution sources, and trigger probabilities is established, facilitating subsequent pollution analysis. Based on these sets of pollution trigger probability sets, the collection of pollution points upstream of the water area, upstream of the wind direction, and upstream of the terrain is sorted to identify pollutant categories with higher pollution source trigger probabilities, thereby identifying suspected pollution sources.

[0121] Among them, include:

[0122] S602-1: Obtain the first pollution point historical data set of the preset time zone for the water upstream pollution point set, the wind upstream pollution point set, and the terrain upstream pollution point set;

[0123] It should be noted that the first pollution point historical data set records the pollutant emission record data generated by each pollution point set within a preset time interval, including the set of pollution points upstream of the water area, the set of pollution points upstream of the wind direction, and the set of pollution points upstream of the terrain. The record data of each pollution point includes the number of emissions and the type of pollutant discharged each time.

[0124] S602-2: Extracting the first pollutant type from the pollutant type set;

[0125] Specifically, the first pollutant type is a random pollutant type in the pollutant type set.

[0126] S602-3: Count the trigger frequency ratio of the first pollutant type in the historical data set of the first pollution point, set it as the trigger probability of the first pollutant type, and add the pollution trigger probability set of the first pollution point;

[0127] It should be noted that setting the trigger probability of the first pollutant type is to obtain the frequency ratio of the first pollutant type in the emission record data, and the first pollution point historical data set is the pollutant emission record data of a random pollution point in the pollution point set.

[0128] S602-4: Add the pollution trigger probability set of the first pollution point to several groups of pollution trigger probability sets.

[0129] In the embodiment of the present application, it also includes:

[0130] S603: Sorting the pollution point set upstream of the water area, the pollution point set upstream of the wind direction, and the pollution point set upstream of the terrain based on several pollution trigger probability sets to obtain suspected pollution sources, including:

[0131] S603-1: Obtain a first pollution trigger probability set for a first pollutant type, and extract a first group of pollution points that are greater than or equal to a trigger probability threshold;

[0132] It should be noted that the trigger probability threshold is a pre-set minimum trigger probability threshold. When it is greater than or equal to this threshold, the probability that the corresponding pollutant category originates from the pollution point is higher; otherwise, the probability that the pollutant category originates from the pollution point is lower.

[0133] S603-2: until the Mth pollution trigger probability set of the Mth pollutant type is obtained, extract the Mth group of pollution points that are greater than or equal to the trigger probability threshold;

[0134] Specifically, each group of pollution points corresponds to a type of pollutant.

[0135] S603-3: Set the first group of pollution points as suspected pollution sources of the first pollutant type, until the Mth group of pollution points is set as suspected pollution sources of the Mth pollutant type.

[0136] S700: Conduct inspections and investigations on suspected pollution sources based on the pollutant type set.

[0137] In summary, this invention addresses the technical issues in existing automated inspection methods for water conservancy projects, such as regional variations and poorly representative water quality analysis results, which hinder the accuracy of inspection plan setup. This improves the efficiency of water quality monitoring and, through intelligent analysis, provides strong technical support for pollution source tracking and water quality management, ensuring the rationality and accuracy of intelligent inspection plan setup.

[0138] The above is a schematic diagram of an intelligent inspection and analysis method for a water conservancy project according to this embodiment. It should be noted that the technical solution of the intelligent inspection and analysis system for a water conservancy project and the technical solution of the intelligent inspection and analysis method for a water conservancy project described above are based on the same concept. For details not described in detail in the technical solution of the intelligent inspection and analysis system for a water conservancy project in this embodiment, please refer to the description of the technical solution of the intelligent inspection and analysis method for a water conservancy project described above.

[0139] The intelligent inspection and analysis system for water conservancy projects in this embodiment includes:

[0140] An acquisition module, used for acquiring water surface images within the area;

[0141] A partitioning module is used to partition the water surface image in the area and obtain the water area partitioning result;

[0142] The first traversal module is used to traverse the water area partition results to perform secondary water area analysis and obtain water area partition sampling points;

[0143] The second traversal module is used to traverse the water area partition sampling points to perform sampling and obtain water area sampling results;

[0144] The detection module is used to perform water quality detection on the water sampling results and obtain a set of pollutant types;

[0145] The backtracking module is used to perform pollution backtracking based on the pollutant type set and the reservoir deployment location to generate suspected pollution sources;

[0146] The inspection module is used to inspect and investigate suspicious pollution sources based on the pollutant type set.

[0147] This embodiment further provides a computing device suitable for intelligent inspection and analysis of water conservancy projects, including:

[0148] Memory and processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the intelligent inspection and analysis method for water conservancy projects proposed in the above embodiment.

[0149] This embodiment further provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for realizing intelligent inspection and analysis of water conservancy projects as proposed in the above embodiment is implemented.

[0150] The storage medium proposed in this embodiment and the intelligent inspection and analysis method for water conservancy projects proposed in the above embodiment belong to the same inventive concept. Technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0151] Through the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented with the help of software and necessary general-purpose hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as a computer's floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of various embodiments of the present invention.

[0152] Example 2

[0153] With reference to Tables 1 and 2, based on the previous embodiment, this embodiment provides an application comparison case of an intelligent inspection and analysis method for a water conservancy project to illustrate the feasibility and beneficial effects of our solution.

[0154] UAV water surface image acquisition and water area zoning: Time: June 15, 2023, Location: A large reservoir, preset inspection cycle: 7 days, UAV model: DJI Phantom 4RTK, light intensity: 1500lx, detection time: 10:00AM, number of water surface images collected: 50.

[0155] The water surface divisions obtained after processing the reservoir water surface image include 5 main water area divisions: Area 1: area 2.5km 2 , the variance value of grayscale distribution information is 35.6, Region 2: area 3.0km 2 , Variance value of grayscale distribution information 25.3, Region 3: Area 1.8km 2 , Variance value of grayscale distribution information 40.1, Region 4: Area 2.2km 2 , Variance value of grayscale distribution information 28.7, Region 5: Area 1.5km 2The variance value of the grayscale distribution information is 32.4. Among them, the discrete parameter threshold is 30, then the sampling method of area 1, area 3, and area 5 is uniform distribution sampling, and area 2 and area 4 are representative sampling.

[0156] Area 1: Sampling point 1 data: water temperature 22.3℃, pH value 7.2, dissolved oxygen (mg / L) 8.0, crude oil (mg / L) 0.06, lead (mg / L) 0.012, mercury (mg / L) 0.005, cadmium (mg / L) 0.002, chromium (mg / L) 0.014, arsenic (mg / L) 0.003.

[0157] Data of sampling point 2: water temperature 22.1℃, pH value 7.3, dissolved oxygen (mg / L) 7.8, crude oil (mg / L) 0.05, lead (mg / L), mercury (mg / L), cadmium (mg / L), chromium (mg / L), arsenic (mg / L).

[0158] Data of sampling point 3: water temperature 22.4℃, pH value 7.1, dissolved oxygen (mg / L) 8.2, crude oil (mg / L) 0.07, lead (mg / L), mercury (mg / L), cadmium (mg / L), chromium (mg / L), arsenic (mg / L).

[0159] Data of sampling point 4: water temperature 22.2℃, pH value 7.2, dissolved oxygen (mg / L) 8.1, crude oil (mg / L) 0.05, lead (mg / L), mercury (mg / L), cadmium (mg / L), chromium (mg / L), arsenic (mg / L).

[0160] Data of sampling point 5: water temperature 22.3℃, pH value 7.2, dissolved oxygen (mg / L) 8.0, crude oil (mg / L) 0.06, lead (mg / L), mercury (mg / L), cadmium (mg / L), chromium (mg / L), arsenic (mg / L).

[0161] Area 2: Sampling point data: water temperature 22.0℃, pH value 7.3, dissolved oxygen (mg / L) 7.7, crude oil (mg / L) 0.04, lead (mg / L) 0.007, mercury (mg / L) 0.004, cadmium (mg / L) 0.001, chromium (mg / L) 0.013, arsenic (mg / L) 0.002.

[0162] Area 3: Sampling point 1 data: water temperature 22.5℃, pH value 7.2, dissolved oxygen (mg / L) 8.3, crude oil (mg / L) 0.06, lead (mg / L) 0.012, mercury (mg / L) 0.005, cadmium (mg / L) 0.002, chromium (mg / L) 0.014, arsenic (mg / L) 0.003.

[0163] Data from sampling point 2: water temperature 22.4℃, pH value 7.1, dissolved oxygen (mg / L) 8.2, crude oil (mg / L) 0.07, lead (mg / L) 0.013, mercury (mg / L) 0.006, cadmium (mg / L) 0.002, chromium (mg / L) 0.015, and arsenic (mg / L) 0.003.

[0164] Data of sampling point 3: water temperature 22.5℃, pH value 7.2, dissolved oxygen (mg / L) 8.3, crude oil (mg / L) 0.06, lead (mg / L) 0.012, mercury (mg / L) 0.005, cadmium (mg / L) 0.002, chromium (mg / L) 0.015, arsenic (mg / L) 0.003.

[0165] Data of sampling point 4: water temperature 22.3℃, pH value 7.3, dissolved oxygen (mg / L) 8.0, crude oil (mg / L) 0.05, lead (mg / L) 0.010, mercury (mg / L) 0.004, cadmium (mg / L) 0.001, chromium (mg / L) 0.014, arsenic (mg / L) 0.002.

[0166] Area 4: Sampling point data: water temperature 22.0℃, pH value 7.3, dissolved oxygen (mg / L) 7.6, crude oil (mg / L) 0.05, lead (mg / L) 0.010, mercury (mg / L) 0.005, cadmium (mg / L) 0.002, chromium (mg / L) 0.013, arsenic (mg / L) 0.003.

[0167] Area 5: Sampling point 1 data: water temperature 22.3℃, pH value 7.2, dissolved oxygen (mg / L) 8.0, crude oil (mg / L) 0.05, lead (mg / L) 0.012, mercury (mg / L) 0.003, cadmium (mg / L) 0.003, chromium (mg / L) 0.014, arsenic (mg / L) 0.005.

[0168] Data from sampling point 2: water temperature 22.4℃, pH value 7.1, dissolved oxygen (mg / L) 8.2, crude oil (mg / L) 0.07, lead (mg / L) 0.013, mercury (mg / L) 0.007, cadmium (mg / L) 0.003, chromium (mg / L) 0.010, and arsenic (mg / L) 0.005.

[0169] Data of sampling point 3: water temperature 22.5℃, pH value 7.3, dissolved oxygen (mg / L) 8.0, crude oil (mg / L) 0.04, lead (mg / L) 0.014, mercury (mg / L) 0.005, cadmium (mg / L) 0.002, chromium (mg / L) 0.014, arsenic (mg / L) 0.004.

[0170] Table 1: Part of the recorded data of the pollution point historical data set:

[0171]

[0172] Table 2: Pollutant trigger probability set for the pollution point historical data set:

[0173]

[0174] Crude oil pollution sources: The trigger probability threshold is 60%, screening out pollution points with a trigger probability greater than or equal to 60%. Pollution points 1 and 4 are the main crude oil pollution sources and are both industrial wastewater discharge points.

[0175] Chromium pollution sources: The trigger probability threshold is 50%, and pollution points with a trigger probability greater than or equal to 50% are screened. Pollution points 1 and 4 are the main sources of chromium pollution and are both industrial wastewater discharge points.

[0176] Lead pollution sources: The trigger probability threshold is 50%, screening out pollution points with a trigger probability greater than or equal to 50%. Pollution points 1, 4, and 5 are the main lead pollution sources, and Pollution point 5 is an agricultural wastewater discharge point.

[0177] Mercury pollution sources: The trigger probability threshold is 30%, and pollution points with a trigger probability greater than or equal to 30% are screened. Pollution points 2 and 5 are the main mercury pollution sources, both of which are agricultural wastewater discharge points.

[0178] Cadmium pollution sources: The trigger probability threshold is 20%, and pollution sites with a trigger probability greater than or equal to 20% are screened. Pollution site 2 is the main cadmium pollution source and is an agricultural wastewater discharge point.

[0179] Arsenic pollution source: The trigger probability threshold is 10%, and pollution points with a trigger probability greater than or equal to 10% are screened. Pollution point 3 is the main arsenic pollution source and is a domestic sewage discharge point.

[0180] By analyzing pollutant types and their trigger probability sets, the suspected pollution sources of each pollutant can be effectively identified, providing a scientific basis for pollution source tracing and water quality management. This solves the technical problem of regional differences in existing automated inspection methods for water conservancy projects, resulting in poor representativeness of water quality analysis results and difficulty in improving the accuracy of inspection plan settings. This improves the efficiency of water quality monitoring and, through intelligent analysis, provides strong technical support for pollution source tracing and water quality management, ensuring the rationality and accuracy of intelligent inspection plan settings.

[0181] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An intelligent inspection and analysis method for water conservancy projects, characterized in that: include: Collect water surface images within the area; Performing partition processing on the water surface image in the area to obtain a water area partition result; Traversing the water area partition results to perform secondary water area analysis and obtain water area partition sampling points; Traversing the water area partition sampling points to perform sampling and obtain water area sampling results; Performing water quality testing on the sampling results of the water area to obtain a set of pollutant types; Based on the pollutant type set and combined with the reservoir deployment location, pollution backtracking is performed to generate suspicious pollution sources; Conduct inspections and investigations on suspected pollution sources based on the pollutant type set; Performing partition processing on the water surface image in the area to obtain water area partition results, including: Performing grayscale processing on the water surface image in the region to obtain a water surface grayscale image; Acquiring detection and monitoring information of the water surface image in the area, wherein the detection and monitoring information includes a light intensity parameter and a detection time parameter; The illumination intensity parameter and the detection time parameter are processed through a grayscale calibration network to obtain a water surface reference grayscale, including: Obtaining a dataset within the region, wherein the dataset within the region includes a light intensity dataset, a detection time dataset, and a reference grayscale identification dataset; Taking the illumination intensity dataset and the detection time dataset as input and the reference grayscale identification dataset as output, a preset amount of regional datasets are retrieved to configure a backbone decision network, wherein the backbone decision network has a first fitting residual vector, which is equal to the mean of the deviation vectors of several identification values and output values; When the modulus value of the first fitting residual vector is greater than or equal to the residual modulus threshold, constructing a first output fitting layer to update the output layer of the backbone decision network based on the first fitting residual vector, retrieving a reservoir data set of a preset data volume to configure a first updated backbone decision network, wherein the first updated backbone decision network has a second fitting residual vector; When the modulus value of the second fitting residual vector is less than the residual modulus threshold, setting the first updated backbone decision network to the grayscale calibration network; Otherwise, constructing a second output fitting layer according to the second fitting residual vector to update the output layer of the first updated backbone decision network; Performing multi-value processing on the water surface grayscale image according to the water surface reference grayscale to generate water surface multi-value feature distribution data; Performing neighborhood hierarchical clustering on the water surface grayscale image according to the water surface multi-valued feature distribution data to generate the water area partition result; The water surface grayscale image is subjected to multi-value processing according to the water surface reference grayscale to generate water surface multi-value feature distribution data, including: Based on the water surface reference grayscale, configuring a multi-valued distribution interval set; Subtracting the water surface reference grayscale from each pixel grayscale of the water surface grayscale image to obtain a corresponding pixel grayscale difference vector; All obtained pixel grayscale difference vectors are distributed in the multi-valued distribution interval set to obtain the multi-valued feature distribution data of the water surface.

2. The intelligent inspection and analysis method for water conservancy projects according to claim 1, characterized in that: Traverse the water area division results to perform secondary water area analysis and obtain water area division sampling points, including: Obtaining grayscale distribution information of a first water area partition of the water area partition result; Calculate the variance value of the grayscale distribution information of the first water area partition and set it as the discrete parameter of the first water area partition; When the discrete parameter of the first water area partition is greater than or equal to the discrete parameter threshold, sampling points are evenly distributed in the first water area partition according to a preset interval, and sampling points of the first water area partition are added; When the discrete parameter of the first water area partition is less than the discrete parameter threshold, randomly deploying a representative sampling point for the first water area partition and setting it as the first water area partition sampling point; Add the first water area partition sampling points to the water area partition sampling points.

3. The intelligent inspection and analysis method for water conservancy projects according to claim 2, characterized in that: Based on the pollutant type set and combined with the reservoir deployment location, pollution backtracking is performed to generate suspected pollution sources, including: Taking the reservoir deployment location as a constraint, collect the pollution point set upstream of the water area, the pollution point set upstream of the wind direction, and the pollution point set upstream of the terrain within the preset range; Traversing the pollution point set upstream of the water area, the pollution point set upstream of the wind direction, and the pollution point set upstream of the terrain, performing historical pollution analysis on the pollutant type set, and generating several sets of pollution trigger probability sets; Traverse the set of pollution points upstream of the water area, the set of pollution points upstream of the wind direction, and the set of pollution points upstream of the terrain, perform historical pollution analysis on the set of pollutant types, and generate several sets of pollution trigger probability sets, including: Obtaining a first pollution point historical data set of a preset time zone for the water upstream pollution point set, the wind upstream pollution point set, and the terrain upstream pollution point set; extracting a first pollutant type from the pollutant type set; Counting the trigger frequency ratio of the first pollutant type in the first pollution point historical data set, setting it as the first pollutant type trigger probability, and adding the first pollution point pollution trigger probability set; Add the first pollution point pollution trigger probability set to the several groups of pollution trigger probability sets.

4. The intelligent inspection and analysis method for water conservancy projects according to claim 3, characterized in that: The method further includes sorting the pollution point set upstream of the water area, the pollution point set upstream of the wind direction, and the pollution point set upstream of the terrain according to the plurality of pollution trigger probability sets to obtain the suspected pollution source, including: Obtain a first pollution trigger probability set of a first pollutant type, and extract a first group of pollution points that are greater than or equal to a trigger probability threshold; Until the Mth pollution trigger probability set of the Mth pollutant type is obtained, the Mth group of pollution points greater than or equal to the trigger probability threshold is extracted; The first group of pollution points is set as the suspected pollution source of the first pollutant type until the Mth group of pollution points is set as the suspected pollution source of the Mth pollutant type.

5. An intelligent inspection and analysis system for water conservancy projects, the system being used to execute the method according to any one of claims 1 to 4, characterized in that: include: An acquisition module, used for acquiring water surface images within the area; A partitioning module, used for performing partitioning processing on the water surface image in the area to obtain a water area partitioning result; The first traversal module is used to traverse the water area partition results to perform secondary water area analysis and obtain water area partition sampling points; The second traversal module is used to traverse the water area partition sampling points to perform sampling and obtain water area sampling results; A detection module, configured to perform water quality detection on the water sampling results to obtain a set of pollutant types; A tracing module is used to perform pollution tracing based on the pollutant type set and the reservoir deployment location to generate suspected pollution sources; The inspection module is used to inspect and investigate suspicious pollution sources based on the pollutant type set.

6. An electronic device comprising: memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the intelligent inspection and analysis method for water conservancy projects described in any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the intelligent inspection and analysis method for water conservancy projects as described in any one of claims 1 to 4.

Images