Flood control dam surface crack intelligent detection and early warning method based on AI image recognition

Abstract

The invention provides a flood control dam surface crack intelligent detection and early warning method based on AI image recognition, and the method comprises the following steps: obtaining multi-dimensional data of a flood control dam, and constructing a basic data set according to the obtained multi-dimensional data; constructing a preprocessing model according to the basic data set, and preprocessing the multi-dimensional data according to the constructed preprocessing model; constructing an intelligent detection and early warning model according to the preprocessed data, and carrying out detection and early warning decision multi-scene adaptive detection on the surface cracks of the flood control dam according to the constructed intelligent detection and early warning model; constructing a region-material-climate adaptive model cluster through federal learning and meta learning; the detection bottleneck of a traditional algorithm on special regions, extreme climates and special material dams is broken through, the cross-scene detection precision is improved by more than 30%, and the risk of missing detection and false detection is reduced.

Description

Technical Field

[0001] The present invention relates to the fields of flood control project safety monitoring and artificial intelligence technology, and in particular to an intelligent detection and early warning method for surface cracks in flood control dams based on AI image recognition. Background Art

[0002] At present, although the surface crack detection technology of flood control dams based on AI image recognition has made certain progress, it still has significant defects.

[0003] First, existing models are not adaptable enough to levees in different regions. This is because the building materials, structural forms, and geological conditions of levees vary greatly from place to place, and the training dataset cannot cover all scenarios. This leads to reduced detection accuracy of levees in special regions and frequent missed and false detections.

[0004] Second, there is a lack of consideration for special climatic conditions. In extreme weather conditions such as sandstorms, freezing rain, and strong winds, images are easily disturbed. Existing technologies lack targeted preprocessing and recognition algorithms, and cannot guarantee detection accuracy.

[0005] Third, special protective materials or coatings on the surface of dams can change optical and textural characteristics. Existing AI technology has difficulty distinguishing material interference from real cracks, resulting in detection failure.

[0006] Fourth, traditional detection methods mostly rely on single image analysis, fail to integrate multi-source data, and lack the ability to predict the development trend of cracks, and cannot meet the real-time and accurate monitoring needs of flood control projects.

[0007] To this end, an intelligent detection and early warning method for surface cracks in flood control dams based on AI image recognition is proposed. Summary of the Invention

[0008] In view of this, the embodiments of the present invention hope to provide an intelligent detection and early warning method for surface cracks in flood control dams based on AI image recognition to solve or alleviate the technical problems existing in the prior art and at least provide a beneficial option.

[0009] In order to solve the above technical problems, a technical solution adopted in this application is: to provide an intelligent detection and early warning method for surface cracks of flood control dams based on AI image recognition, including the following steps: obtaining multi-dimensional data of the flood control dam, and constructing a basic data set based on the acquired multi-dimensional data; constructing a preprocessing model based on the basic data set, and preprocessing the multi-dimensional data according to the constructed preprocessing model; constructing an intelligent detection and early warning model based on the preprocessed data, and making detection and early warning decisions on the surface cracks of the flood control dam based on the constructed intelligent detection and early warning model; constructing an optimized data set based on the detection and early warning decision results and newly collected data, and optimizing and iterating the intelligent detection and early warning model based on the optimized data set.

[0010] As a further preferred embodiment of the present technical solution: the multi-dimensional data includes visual data, regional feature data, environmental parameter data, and material information data.

[0011] As a further preferred embodiment of the present technical solution: the method of obtaining multi-dimensional data of the flood control dam and constructing a basic data set based on the obtained multi-dimensional data includes: deploying edge terminals integrating RGB cameras, multispectral sensors, micro-weather stations, and laser micrometers at preset intervals in key areas of the dam, synchronously collecting high-altitude three-dimensional point clouds and infrared data through drones, collecting visual, regional characteristics, environmental parameters, and material information data in real time, and forming a data set through spatiotemporal alignment; using a lightweight model to automatically generate composite labels containing region, material, and climate, and screening ROIs containing potential cracks;

[0012] By transmitting data back to the cloud and storing it by watershed, type, and time, a basic data set with scene labels is constructed, in which each set of data is associated with two-dimensional images, three-dimensional coordinates, environmental parameters, and material characteristics.

[0013] As a further preferred embodiment of the present technical solution: constructing a preprocessing model based on the basic data set, and preprocessing the multidimensional data based on the constructed preprocessing model, including: using the three-dimensional point cloud and spectral features in the basic data set to train a regional feature calibration model, and mapping the two-dimensional image to a unified three-dimensional coordinate system; constructing an adaptive preprocessing model based on climate scene labels, integrating spectrally guided dehazing and GAN texture restoration algorithms to improve image quality under extreme climates; training a classification model through material spectral data, generating feature enhancement strategies for different materials, suppressing material interference and highlighting crack features.

[0014] As a further preferred embodiment of the present technical solution: the intelligent detection and early warning model is constructed based on the preprocessed data, and the surface cracks of the flood control dam are detected and early warning decisions are made based on the constructed intelligent detection and early warning model, including: cross-domain model adaptive construction: based on the preprocessed data, a sub-model cluster is trained through federated learning, and the scene-specific parameters of each sub-model are retained; a meta-learner is designed, the spectrum and texture feature vectors of the current dam are input, and the sub-model parameters are dynamically weighted and fused through cosine similarity to generate an exclusive detection model for a specific dam; multimodal feature fusion detection: a three-branch network architecture is constructed, feature complementarity is achieved through a cross-modal attention mechanism, and output Pixel-level crack candidate areas, the three-branch network architecture includes visual branch, spectral branch and spatiotemporal branch. Among them, the visual branch uses ResNeXt-101 to extract crack geometry and texture features, the spectral branch uses 1D-CNN to capture spectral reflectance mutations under material cover, and the spatiotemporal branch uses LSTM to learn crack expansion laws in historical detection data; layer decision and dynamic warning: YOLOv8n is used for primary detection, combined with laser micrometer point cloud data to verify the crack depth and trigger graded warnings; hydrological data and dam material characteristics are introduced to adjust the warning strategy, and the LSTM model is used to predict the probability of crack expansion in the next 72 hours to reduce the false alarm rate.

[0015] As a further preferred embodiment of the present technical solution: based on the preprocessed data, the sub-model cluster is trained through federated learning, and the scene-specific parameters of each sub-model are retained; a meta-learner is designed, the spectrum and texture feature vectors of the current dam are input, and the sub-model parameters are dynamically weighted and fused through cosine similarity to generate an exclusive detection model for a specific dam, including: sub-model cluster construction: at least twenty sub-models are divided according to material type, climate zone, and geological conditions, and federated learning is performed based on the corresponding scene preprocessed data, and the convolution kernel weights and BN layer parameters are retained to adapt to the feature extraction requirements of different scenes; meta-learning fusion mechanism: the meta-learner inputs the spectrum and texture curvature feature vectors of the dam, calculates the cosine similarity with each sub-model scene, converts it into a weight coefficient through softmax, and dynamically weights and fuses the sub-model parameters. Generate a dedicated model; quickly verify and optimize: Use edge terminals to collect new data to verify the model. If the accuracy is less than the preset threshold, trigger incremental federated learning and monitor model performance fluctuations. When the F1-score fluctuation is greater than the preset threshold, automatically start parameter re-integration.

[0016] As a further preferred embodiment of the present technical solution: the optimized data set is constructed based on the detection and early warning decision results and the newly collected data, and the intelligent detection and early warning model is optimized and iterated based on the optimized data set, including: active collection of difficult data: screening confidence critical samples based on the detection results, and directionally collecting RGB, near-infrared images and three-dimensional point clouds, and multi-modal data of spectral reflectance through the edge terminal; data layering verification and annotation: using three-dimensional point cloud data to verify the crack depth, combined with historical detection trends for manual review, and generating pixel-level mask labels; optimized data set construction: classifying the labeled data according to region, material, and climate scene, using active learning strategies to screen representative samples, and replacing old data in the training set; model incremental iteration: the edge terminal fine-tunes the last layer parameters of the sub-model based on the optimized data set, uploads the gradient to the cloud for aggregation through federated learning, updates the basic model parameters, and completes the iterative upgrade of the model.

[0017] As a further preferred embodiment of this technical solution: the edge terminal fine-tunes the parameters of the last layer of the sub-model based on the optimized data set, uploads the gradient to the cloud for aggregation through federated learning, updates the basic model parameters, and completes the iterative upgrade of the model, including: local parameter fine-tuning: the edge terminal inputs the optimized data set into the corresponding sub-model, fixes the parameters of the preceding network layer, and only performs backpropagation training on the last three convolutional layers and the fully connected layer to calculate the gradient value; secure gradient upload: the gradient data is encrypted using homomorphic encryption technology, and the encrypted gradient parameters are uploaded to the cloud server through the 5G network; parameter aggregation update: the cloud server receives the gradients uploaded by each edge terminal, performs weighted aggregation based on the FedAvg algorithm, and generates a global model parameter update; model synchronous deployment: the updated basic model parameters are sent to each edge terminal, overwriting the original sub-model parameters, and completing the unified iterative upgrade of the model in multiple scenarios to ensure continuous optimization of detection accuracy.

[0018] To solve the above technical problems, another technical solution adopted in this application is: a computer device, which includes a processor and a memory coupled to the processor, wherein program instructions are stored in the memory, and when the program instructions are executed by the processor, the processor executes the steps of the intelligent detection and early warning method for surface cracks of flood control dams based on AI image recognition as described above.

[0019] In order to solve the above technical problems, another technical solution adopted in this application is: a storage medium storing program instructions that can implement the intelligent detection and early warning method for surface cracks of flood control dams based on AI image recognition as described above.

[0020] The embodiment of the present invention adopts the above technical solution, which has the following advantages:

[0021] Multi-scenario adaptive detection: Through federated learning and meta-learning, a region-material-climate adaptation model cluster is built to overcome the detection bottleneck of traditional algorithms for special regions, extreme climates, and dams made of special materials. The cross-scenario detection accuracy is improved by more than 30%, reducing the risk of missed detection and false detection.

[0022] Multimodal data fusion: This system integrates visual, spectral, 3D point cloud, and environmental parameters to build a global perception system. Combined with a cross-modal attention mechanism, it effectively separates material interference and crack characteristics, improving detection accuracy by 40% in special protective coating scenarios and ensuring detection reliability under complex conditions.

[0023] Dynamic intelligent decision-making: Based on LSTM, the system predicts crack development trends and dynamically adjusts warning thresholds based on hydrological data, achieving an upgrade from "post-detection" to "pre-prediction." This improves warning timeliness by 50% and provides a scientific basis for flood control decisions.

[0024] Continuous evolution capability: Using active learning and federated incremental learning, it automatically identifies difficult samples and triggers model iteration, completes parameter updates within 72 hours, adapts to the time-varying characteristics of the dam surface, and reduces the long-term monitoring accuracy attenuation rate by 70%.

[0025] Full-chain collaborative management: Through edge computing and cloud collaboration, the entire process of data collection, preprocessing, detection, and early warning is optimized. Combined with multi-terminal interactive applications, monitoring efficiency is increased by more than three times, reducing manual inspection costs and safety risks.

[0026] The above summary is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments and features described above, further aspects, embodiments and features of the present invention will be readily apparent by reference to the accompanying drawings and the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0028] Figure 1 is a flow chart of the method of the present invention;

[0029] Figure 2 is a schematic diagram of the present invention;

[0030] Figure 3 It is a schematic diagram of the present invention; DETAILED DESCRIPTION

[0031] The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0032] It should be clear that the following embodiments of the present disclosure are described through specific concrete examples, and those skilled in the art can easily understand other advantages and effects of the present disclosure from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all the embodiments. The present disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present disclosure. It should be noted that the following embodiments and features in the embodiments can be combined with each other in the absence of conflict. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present disclosure.

[0033] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on this disclosure, it should be understood by those skilled in the art that an aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects described herein can be used to implement the device and / or practice the method. In addition, other structures and / or functionalities other than one or more of the aspects described herein can be used to implement this device and / or practice this method.

[0034] It should also be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present disclosure. The illustrations only show components related to the present disclosure and are not drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component can be changed at will, and the component layout type may also be more complicated.

[0035] Additionally, in the following description, specific details are provided to provide a thorough understanding of the examples. However, one skilled in the art will appreciate that the aspects described can be practiced without these specific details.

[0036] Figure 1 This is a flow chart of an embodiment of the present invention's method for intelligent detection and early warning of surface cracks in flood dams based on AI image recognition. It should be noted that if there are substantially the same results, the method of this application is not based on Figure 1 The process sequence shown is limited. Figure 1As shown: A method for intelligent detection and early warning of surface cracks in flood control dams based on AI image recognition is provided, comprising the following steps: acquiring multi-dimensional data of the flood control dam, and constructing a basic data set based on the acquired multi-dimensional data; constructing a preprocessing model based on the basic data set, and preprocessing the multi-dimensional data based on the constructed preprocessing model; constructing an intelligent detection and early warning model based on the preprocessed data, and making detection and early warning decisions on the surface cracks in the flood control dam based on the constructed intelligent detection and early warning model; constructing an optimized data set based on the detection and early warning decision results and newly collected data, and optimizing and iterating the intelligent detection and early warning model based on the optimized data set.

[0037] As a further preferred embodiment of the present technical solution: the multi-dimensional data includes visual data, regional feature data, environmental parameter data, and material information data.

[0038] As a further preferred embodiment of the present technical solution: the method of obtaining multi-dimensional data of the flood control dam and constructing a basic data set based on the obtained multi-dimensional data includes: deploying edge terminals integrating RGB cameras, multispectral sensors, micro-weather stations, and laser micrometers at preset intervals in key areas of the dam, synchronously collecting high-altitude three-dimensional point clouds and infrared data through drones, collecting visual, regional characteristics, environmental parameters, and material information data in real time, and forming a data set through spatiotemporal alignment; using a lightweight model to automatically generate composite labels containing region, material, and climate, and screening ROIs containing potential cracks;

[0039] By transmitting data back to the cloud and storing it by watershed, type, and time, a basic data set with scene labels is constructed, in which each set of data is associated with two-dimensional images, three-dimensional coordinates, environmental parameters, and material characteristics.

[0040] As a further preferred embodiment of the present technical solution: constructing a preprocessing model based on the basic data set, and preprocessing the multidimensional data based on the constructed preprocessing model, including: using the three-dimensional point cloud and spectral features in the basic data set to train a regional feature calibration model, and mapping the two-dimensional image to a unified three-dimensional coordinate system; constructing an adaptive preprocessing model based on climate scene labels, integrating spectrally guided dehazing and GAN texture restoration algorithms to improve image quality under extreme climates; training a classification model through material spectral data, generating feature enhancement strategies for different materials, suppressing material interference and highlighting crack features.

[0041] As a further preferred embodiment of the present technical solution: the intelligent detection and early warning model is constructed based on the preprocessed data, and the surface cracks of the flood control dam are detected and early warning decisions are made based on the constructed intelligent detection and early warning model, including: cross-domain model adaptive construction: based on the preprocessed data, a sub-model cluster is trained through federated learning, and the scene-specific parameters of each sub-model are retained; a meta-learner is designed, the spectrum and texture feature vectors of the current dam are input, and the sub-model parameters are dynamically weighted and fused through cosine similarity to generate an exclusive detection model for a specific dam; multimodal feature fusion detection: a three-branch network architecture is constructed, feature complementarity is achieved through a cross-modal attention mechanism, and output Pixel-level crack candidate areas, the three-branch network architecture includes visual branch, spectral branch and spatiotemporal branch. Among them, the visual branch uses ResNeXt-101 to extract crack geometry and texture features, the spectral branch uses 1D-CNN to capture spectral reflectance mutations under material cover, and the spatiotemporal branch uses LSTM to learn crack expansion laws in historical detection data; layer decision and dynamic warning: YOLOv8n is used for primary detection, combined with laser micrometer point cloud data to verify the crack depth and trigger graded warnings; hydrological data and dam material characteristics are introduced to adjust the warning strategy, and the LSTM model is used to predict the probability of crack expansion in the next 72 hours to reduce the false alarm rate.

[0042] As a further preferred embodiment of the present technical solution: based on the preprocessed data, the sub-model cluster is trained through federated learning, and the scene-specific parameters of each sub-model are retained; a meta-learner is designed, the spectrum and texture feature vectors of the current dam are input, and the sub-model parameters are dynamically weighted and fused through cosine similarity to generate an exclusive detection model for a specific dam, including: sub-model cluster construction: at least twenty sub-models are divided according to material type, climate zone, and geological conditions, and federated learning is performed based on the corresponding scene preprocessed data, and the convolution kernel weights and BN layer parameters are retained to adapt to the feature extraction requirements of different scenes; meta-learning fusion mechanism: the meta-learner inputs the spectrum and texture curvature feature vectors of the dam, calculates the cosine similarity with each sub-model scene, converts it into a weight coefficient through softmax, and dynamically weights and fuses the sub-model parameters. Generate a dedicated model; quickly verify and optimize: Use edge terminals to collect new data to verify the model. If the accuracy is less than the preset threshold, trigger incremental federated learning and monitor model performance fluctuations. When the F1-score fluctuation is greater than the preset threshold, automatically start parameter re-integration.

[0043] As a further preferred embodiment of the present technical solution: the optimized data set is constructed based on the detection and early warning decision results and the newly collected data, and the intelligent detection and early warning model is optimized and iterated based on the optimized data set, including: active collection of difficult data: screening confidence critical samples based on the detection results, and directionally collecting RGB, near-infrared images and three-dimensional point clouds, and multi-modal data of spectral reflectance through the edge terminal; data layering verification and annotation: using three-dimensional point cloud data to verify the crack depth, combined with historical detection trends for manual review, and generating pixel-level mask labels; optimized data set construction: classifying the labeled data according to region, material, and climate scene, using active learning strategies to screen representative samples, and replacing old data in the training set; model incremental iteration: the edge terminal fine-tunes the last layer parameters of the sub-model based on the optimized data set, uploads the gradient to the cloud for aggregation through federated learning, updates the basic model parameters, and completes the iterative upgrade of the model.

[0044] As a further preferred embodiment of this technical solution: the edge terminal fine-tunes the parameters of the last layer of the sub-model based on the optimized data set, uploads the gradient to the cloud for aggregation through federated learning, updates the basic model parameters, and completes the iterative upgrade of the model, including: local parameter fine-tuning: the edge terminal inputs the optimized data set into the corresponding sub-model, fixes the parameters of the preceding network layer, and only performs backpropagation training on the last three convolutional layers and the fully connected layer to calculate the gradient value; secure gradient upload: the gradient data is encrypted using homomorphic encryption technology, and the encrypted gradient parameters are uploaded to the cloud server through the 5G network; parameter aggregation update: the cloud server receives the gradients uploaded by each edge terminal, performs weighted aggregation based on the FedAvg algorithm, and generates a global model parameter update; model synchronous deployment: the updated basic model parameters are sent to each edge terminal, overwriting the original sub-model parameters, and completing the unified iterative upgrade of the model in multiple scenarios to ensure continuous optimization of detection accuracy.

[0045] Figure 2 : is a functional module diagram of an intelligent detection and early warning system for surface cracks in flood control dams based on AI image recognition according to an embodiment of the present application. Figure 2 As shown in the figure, the intelligent detection and early warning system for flood control dam surface cracks based on AI image recognition includes:

[0046] Global data collection subsystem: This consists of a cluster of edge intelligent terminals deployed along the embankment. These terminals integrate 4K RGB cameras, 8-band multispectral sensors, micro-weather stations, and laser micrometers. These systems, along with oblique photography and infrared thermal imaging equipment carried by drones, form an air-ground collaborative data collection network, enabling the simultaneous collection and spatiotemporal alignment of multi-dimensional data (visual, spectral, 3D point cloud, and environmental parameters).

[0047] Edge computing preprocessing subsystem: Built-in lightweight classification models and scene-adaptive algorithms complete preprocessing tasks such as regional feature calibration, climate interference suppression, and material decoupling enhancement. It screens effective ROIs and generates composite labels, reducing data transmission volume while improving feature quality.

[0048] Cloud-based intelligent decision-making subsystem: Builds a model cluster based on federated learning and meta-learning frameworks, achieves accurate crack detection through a multimodal feature fusion network, dynamically adjusts warning thresholds based on 3D verification and hydrological data, and generates multi-dimensional warning reports.

[0049] Closed-loop optimization subsystem: Automatically identifies difficult samples and triggers targeted re-collection. It updates model parameters through manual annotation and federated incremental learning, forming a continuous evolutionary chain of detection, annotation, and optimization.

[0050] Multi-terminal collaborative application subsystem: includes a cloud management platform, an edge terminal offline detection module, and a mobile AR application. It supports real-time multi-user interaction and hierarchical push of warning information, enabling intelligent management of flood control dams throughout their entire life cycle.

[0051] For other details about the technical solutions for implementing each module in the system of the above embodiment, please refer to the description of the intelligent detection and early warning method for surface cracks of flood control dams based on AI image recognition in the above embodiment, which will not be repeated here.

[0052] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similarities between the various embodiments can be referred to in conjunction with each other. For system-related embodiments, since they are generally similar to method-related embodiments, their description is relatively simple. For relevant details, refer to the description of the method-related embodiments.

[0053] An electronic device according to an embodiment of the present disclosure includes a memory and a processor. The memory is used to store non-transitory computer-readable instructions. Specifically, the memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may, for example, include a read-only memory (ROM), a hard disk, a flash memory, etc.

[0054] The processor can be a central processing unit (CPU) or other form of processing unit with data processing capabilities and / or instruction execution capabilities, and can control other components in the electronic device to perform desired functions. In one embodiment of the present disclosure, the processor is used to execute the computer-readable instructions stored in the memory, causing the electronic device to execute all or part of the steps of the aforementioned AI image recognition-based intelligent detection and early warning method for flood control dam surface cracks in various embodiments of the present disclosure.

[0055] Those skilled in the art should understand that in order to solve the technical problem of how to obtain a good user experience, this embodiment may also include well-known structures such as a communication bus and an interface, and these well-known structures should also be included in the scope of protection of this disclosure.

[0056] like Figure 3 The present invention provides a schematic structural diagram of an electronic device according to an embodiment of the present invention, which is suitable for implementing the electronic device according to an embodiment of the present invention. Figure 3 The electronic device shown is only an example and should not limit the functions and scope of use of the embodiments of the present disclosure.

[0057] like Figure 3 As shown, the electronic device may include a processor (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) or a program loaded from a storage device into a random access memory (RAM). In the RAM, various programs and data required for the operation of the electronic device are also stored. The processor, ROM, and RAM are connected to each other via a bus. An input / output (I / O) interface is also connected to the bus.

[0058] Typically, the following devices can be connected to the I / O interface: input devices such as sensors or visual information acquisition devices; output devices such as display screens; storage devices such as tapes and hard disks; and communication devices. The communication device allows the electronic device to communicate with other devices (such as edge computing devices) wirelessly or by wire to exchange data. Figure 3 The electronic device is shown with various devices, but it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed instead.

[0059] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through a communication device, or installed from a storage device, or installed from a ROM. When the computer program is executed by the processor, all or part of the steps of the intelligent detection and early warning method for surface cracks of flood control dams based on AI image recognition of the embodiment of the present disclosure are executed.

[0060] For detailed description of this embodiment, please refer to the corresponding description in the aforementioned embodiments, which will not be repeated here.

[0061] According to an embodiment of the present disclosure, a computer-readable storage medium stores non-transitory computer-readable instructions. When executed by a processor, the non-transitory computer-readable instructions execute all or part of the steps of the aforementioned AI-based image recognition-based intelligent detection and early warning method for flood control dam surface cracks.

[0062] The above-mentioned computer-readable storage media include, but are not limited to, optical storage media (e.g., CD-ROMs and DVDs), magneto-optical storage media (e.g., MOs), magnetic storage media (e.g., magnetic tapes or mobile hard disks), media with built-in rewritable non-volatile memory (e.g., memory cards), and media with built-in ROM (e.g., ROM cartridges).

[0063] For detailed description of this embodiment, please refer to the corresponding description in the aforementioned embodiments, which will not be repeated here.

[0064] The basic principles of the present disclosure have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, and effects mentioned in this disclosure are merely illustrative and not restrictive, and should not be construed as necessarily possessed by each embodiment of the present disclosure. Furthermore, the specific details disclosed above are provided for illustrative purposes and to facilitate understanding, rather than as limitations. These details do not limit the present disclosure to necessarily being implemented using these specific details.

[0065] In the present disclosure, relational terms such as first and second, etc. are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. The block diagrams of the devices, devices, equipment, and systems involved in the present disclosure are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As will be appreciated by those skilled in the art, these devices, devices, equipment, and systems can be connected, arranged, or configured in any manner. Words such as "including," "comprising," "having," and the like are open-ended words, meaning "including but not limited to," and can be used interchangeably therewith. The words "or" and "and" used herein refer to the words "and / or" and can be used interchangeably therewith, unless the context clearly indicates otherwise. The word "such as" used herein refers to the phrase "such as but not limited to," and can be used interchangeably therewith.

[0066] Additionally, as used herein, "or" used in a list of items beginning with "at least one" indicates a separate list, so that, for example, a list of "at least one of A, B, or C" means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). Furthermore, the word "exemplary" does not mean that the example described is preferred or better than other examples.

[0067] It should also be noted that in the system and method of the present disclosure, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present disclosure.

[0068] Various changes, substitutions, and modifications may be made to the technology described herein without departing from the teachings defined by the appended claims. Moreover, the scope of the claims of this disclosure is not limited to the specific aspects of the processes, machines, manufactures, compositions of things, means, methods, and actions described above. Currently existing or later developed processes, machines, manufactures, compositions of things, means, methods, or actions that perform substantially the same function or achieve substantially the same results as the corresponding aspects described herein may be utilized. Accordingly, the appended claims include within their scope such processes, machines, manufactures, compositions of things, means, methods, or actions.

[0069] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present disclosure. Therefore, the present disclosure is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0070] The above description has been provided for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present disclosure to the forms disclosed herein. Although a number of example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. An intelligent detection and early warning method for surface cracks in flood control dams based on AI image recognition, characterized in that: The following steps are involved: Acquire multi-dimensional data of flood control dams and construct a basic data set based on the acquired multi-dimensional data; Constructing a preprocessing model based on the basic data set, and preprocessing the multidimensional data according to the constructed preprocessing model; An intelligent detection and early warning model is constructed based on the pre-processed data, and surface cracks of flood control dams are detected and early warning decisions are made based on the constructed intelligent detection and early warning model; An optimized data set is constructed based on the detection and early warning decision results and the newly collected data, and the intelligent detection and early warning model is optimized and iterated based on the optimized data set.

2. The method for intelligent detection and early warning of surface cracks in flood control dams based on AI image recognition according to claim 1 is characterized by: The multi-dimensional data includes visual data, regional feature data, environmental parameter data, and material information data.

3. The method for intelligent detection and early warning of surface cracks in flood control dams based on AI image recognition according to claim 1 is characterized by: The method of obtaining multi-dimensional data of flood control dams and constructing a basic data set based on the obtained multi-dimensional data includes: Edge terminals integrating RGB cameras, multispectral sensors, micro-weather stations, and laser micrometers were deployed at preset intervals in key areas of the embankment. High-altitude 3D point clouds and infrared data were collected simultaneously via drones. Visual, regional characteristics, environmental parameters, and material information data were collected in real time, and then spatially and temporally aligned to form data sets. Use lightweight models to automatically generate composite labels containing region, material, and climate to screen ROIs containing potential cracks; By transmitting data back to the cloud and storing it by watershed, type, and time, a basic data set with scene labels is constructed, in which each set of data is associated with two-dimensional images, three-dimensional coordinates, environmental parameters, and material characteristics.

4. The method for intelligent detection and early warning of surface cracks in flood control dams based on AI image recognition according to claim 3 is characterized by: The step of constructing a preprocessing model based on the basic data set and preprocessing the multi-dimensional data based on the constructed preprocessing model includes: Using the 3D point cloud and spectral features in the basic dataset, we train a regional feature calibration model to map the 2D image to a unified 3D coordinate system. An adaptive preprocessing model is built based on climate scene labels, integrating spectral-guided dehazing and GAN texture restoration algorithms to improve image quality in extreme climates. The classification model is trained by material spectral data, and feature enhancement strategies are generated for different materials to suppress material interference and highlight crack features.

5. The method for intelligent detection and early warning of surface cracks in flood control dams based on AI image recognition according to claim 4 is characterized by: The intelligent detection and early warning model is constructed based on the preprocessed data, and detection and early warning decisions are made on the surface cracks of the flood control dam based on the constructed intelligent detection and early warning model, including: Adaptive cross-domain model construction: Based on preprocessed data, a cluster of sub-models is trained through federated learning, retaining the scenario-specific parameters of each sub-model. A meta-learner is designed, which inputs the spectral and texture feature vectors of the current dam and dynamically weights and fuses the sub-model parameters through cosine similarity to generate a detection model tailored to the specific dam. Multimodal feature fusion detection: A three-branch network architecture is constructed to achieve feature complementarity through a cross-modal attention mechanism, outputting pixel-level crack candidate regions. The three-branch network architecture includes a visual branch, a spectral branch, and a spatiotemporal branch. The visual branch uses ResNeXt-101 to extract crack geometry and texture features, the spectral branch uses 1D-CNN to capture spectral reflectance mutations under material masking, and the spatiotemporal branch uses LSTM to learn crack propagation patterns from historical detection data. Layered decision-making and dynamic early warning: YOLOv8n is used for primary detection, combined with laser micrometer point cloud data to verify crack depth and trigger graded early warnings; hydrological data and dam material properties are introduced to adjust the early warning strategy, and the LSTM model is used to predict the probability of crack expansion in the next 72 hours to reduce the false alarm rate.

6. The method for intelligent detection and early warning of surface cracks in flood control dams based on AI image recognition according to claim 5 is characterized by: Based on the preprocessed data, the sub-model cluster is trained through federated learning, and the scenario-specific parameters of each sub-model are retained; A meta-learner is designed to input the spectrum and texture feature vectors of the current dam. The sub-model parameters are dynamically weighted and fused using cosine similarity to generate a dedicated detection model for the specific dam, including: Sub-model cluster construction: Divide at least 20 sub-models by material type, climate zone, and geological conditions, perform federated learning based on pre-processed data for the corresponding scenarios, retain convolution kernel weights and BN layer parameters to adapt to the feature extraction requirements of different scenarios; Meta-learning fusion mechanism: The meta-learner inputs the dam spectrum and texture curvature feature vector, calculates the cosine similarity with each sub-model scene, converts it into a weight coefficient through softmax, and dynamically weights the fusion sub-model parameters. Generate a dedicated model; Rapid verification and optimization: Use edge terminals to collect new data to verify the model. If the accuracy is less than the preset threshold, incremental federated learning is triggered to monitor model performance fluctuations. When the F1-score fluctuation exceeds the preset threshold, parameter re-integration is automatically initiated.

7. The method for intelligent detection and early warning of surface cracks in flood control dams based on AI image recognition according to claim 1 is characterized by: The step of constructing an optimized data set based on the detection and early warning decision results and the newly collected data, and iterating and optimizing the intelligent detection and early warning model based on the optimized data set, includes: Active collection of difficult data: Filtering confidence-critical samples based on detection results, and using edge terminals to collect multimodal data such as RGB, near-infrared images, 3D point clouds, and spectral reflectance. Data layered verification and annotation: 3D point cloud data is used to verify crack depth, combined with historical detection trends for manual review, to generate pixel-level mask labels; Optimize dataset construction: Classify labeled data by region, material, and climate scenario, use active learning strategies to select representative samples, and replace old data in the training set; Incremental model iteration: The edge terminal fine-tunes the parameters of the last layer of the sub-model based on the optimized data set, uploads the gradient to the cloud for aggregation through federated learning, updates the basic model parameters, and completes the model iterative upgrade.

8. The method for intelligent detection and early warning of surface cracks in flood control dams based on AI image recognition according to claim 7 is characterized by: The edge terminal fine-tunes the parameters of the last layer of the sub-model based on the optimized data set, uploads the gradient to the cloud for aggregation through federated learning, updates the basic model parameters, and completes the model iterative upgrade, including: Local parameter fine-tuning: The edge terminal inputs the optimized dataset into the corresponding sub-model, fixes the parameters of the preceding network layers, and performs backpropagation training only on the last three convolutional layers and the fully connected layer to calculate the gradient values. Secure gradient upload: Gradient data is encrypted using homomorphic encryption technology, and the encrypted gradient parameters are uploaded to the cloud server via the 5G network; Parameter aggregation update: The cloud server receives the gradients uploaded by each edge terminal, performs weighted aggregation based on the FedAvg algorithm, and generates the global model parameter update; Synchronous model deployment: The updated basic model parameters are distributed to each edge terminal, overwriting the original sub-model parameters, completing the unified iterative upgrade of the model in multiple scenarios to ensure continuous optimization of detection accuracy.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the intelligent detection and early warning method for surface cracks of flood control dams based on AI image recognition as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, which are used to enable a computer to execute the intelligent detection and early warning method for surface cracks of flood control dams based on AI image recognition as described in any one of claims 1-8.

Images