Visual two-dimensional deflection detection system based on artificial intelligence

Abstract

The invention discloses a visual two-dimensional deflection detection system based on artificial intelligence, which relates to the technical field of engineering quality detection and comprises a visual image acquisition module, a deflection feature extraction module, an intelligent error compensation module, a multi-dimensional data fusion module, a dynamic evolution prediction module, a deflection fault early warning module and a self-learning and optimization module. According to the visual two-dimensional deflection detection system based on artificial intelligence, the defects of single data source, artificial feature extraction and poor robustness in a dynamic environment of a traditional system are overcome by adopting a self-adaptive feature extraction and deflection prediction model and a self-learning and optimization module; the system not only can correct the detection deviation in real time by fusing multi-dimensional vision, stress and displacement information and improve the detection precision and stability, but also has autonomous optimization and updating capabilities, and can keep efficient and accurate deflection prediction in a long-term monitoring process, thereby effectively guaranteeing the health and safety of the structure.

Description

Technical Field

[0001] The present invention relates to the technical field of engineering quality detection, and in particular to a visual dual-dimensional deflection detection system based on artificial intelligence. Background Art

[0002] In the health monitoring of modern infrastructure, deflection detection is one of the important means to evaluate the safety and stability of structures. Traditional deflection detection mainly relies on single sensors (such as displacement sensors, stress sensors, etc.) or simple visual monitoring methods. These methods are usually limited to single-dimensional data collection and analysis, resulting in low detection accuracy and failure to timely detect local and subtle structural deformations. In addition, traditional deflection detection systems often require manual intervention for data analysis and feature extraction, resulting in low efficiency and susceptibility to interference from human factors. In complex detection environments, the multi-dimensional characteristics of deflection often show the complexity of dynamic changes, and the existing technology lacks the ability to effectively extract and fuse these complex features, which greatly reduces the accuracy and robustness of the system in long-term monitoring.

[0003] The disadvantages of the prior art are as follows: Limitations of a single data source: Traditional deflection detection systems mostly rely on a single visual, displacement or stress sensor. This single-mode data acquisition method cannot fully capture the multi-dimensional deflection information of complex structures, resulting in low detection accuracy and instability, especially in a changing external environment, which is prone to errors.

[0004] Inefficiency of manual feature extraction: Most traditional deflection detection systems require manual setting of feature templates and rely on manual analysis of images or sensor data, which makes it difficult to cope with complex and variable deflection features. In addition, the efficiency and accuracy of feature extraction are easily affected by human factors.

[0005] Lack of multi-source data fusion capability: Existing deflection detection systems often only focus on a single type of sensor data and cannot effectively utilize the combined effects of stress, displacement and visual information. This lack of multi-dimensional data fusion makes it difficult for the system to perform accurate real-time correction and global prediction in practical applications.

[0006] Poor robustness in dynamic environments: Under changing environmental conditions, it is difficult for traditional systems to dynamically adjust models, resulting in detection results that are susceptible to noise interference and lack sufficient robustness and stability. Summary of the invention

[0007] In view of the deficiencies in the prior art, the present invention provides a visual two-dimensional deflection detection system based on artificial intelligence to solve the problems raised in the above background technology.

[0008] To achieve the above object, the present invention provides the following technical solutions: a visual two-dimensional deflection detection system based on artificial intelligence, comprising a visual image acquisition module, a deflection feature extraction module, an intelligent error compensation module, a multi-dimensional data fusion module, a dynamic evolution prediction module, a deflection fault warning module and a self-learning and optimization module; The visual image acquisition module acquires the two-dimensional image or point cloud data of the structure in real time based on a high-resolution camera or a laser radar; The deflection feature extraction module automatically extracts deflection features from the image through a deep learning model based on a convolutional neural network; The intelligent error compensation module, based on a dynamic error compensation algorithm, compensates for errors in real time through recognition and tracking of image feature points, time series analysis and Kalman filtering, to ensure the accuracy of deflection detection results; The multi-dimensional data fusion module deeply fuses the two-dimensional visual information with sensor data, such as stress sensors and displacement sensors, to form a multi-modal data analysis framework, and corrects the deviation in visual deflection detection in real time by fusing detection data of different dimensions; The dynamic evolution prediction module dynamically predicts the deflection change through time series analysis and long short-term memory network; The deflection fault warning module detects abnormal deflection and issues an alarm to the operator when the deflection exceeds a set safety threshold; The self-learning and optimization module is based on reinforcement learning and adaptive algorithms to continuously optimize the system's own performance according to actual detection data.

[0009] To further optimize the technical solution, the visual image acquisition module avoids the influence of lighting or weather factors in complex environments through a camera or a laser radar, and obtains high-quality two-dimensional images or point cloud data; Cameras or lidars acquire 2D images or point cloud data from different angles by synchronizing multiple sensors to monitor the structure in all directions; The module also makes real-time adjustments based on changes in the external environment and automatically calibrates parameters to improve the accuracy and clarity of two-dimensional images or point cloud data.

[0010] To further optimize the technical solution, the deflection feature extraction module is trained using a large amount of annotated structural deformation data to automatically identify and segment the deflection area; This module is based on an adaptive method to cope with a variety of complex deformation forms. It does not require a preset feature template and dynamically updates the deflection feature library based on historical detection data for self-optimization to improve the accuracy of feature extraction.

[0011] To further optimize the technical solution, the adaptive method includes the following process: Adaptation based on multi-scale feature extraction: A multi-scale convolutional neural network is used to extract both subtle and large-scale deflection features in convolutional layers of different scales; Morphological adaptation and geometric invariants: By combining geometric invariants and morphological feature extraction methods, geometric invariant features such as curvature and angle are selected to ensure the robustness of feature extraction; based on morphological operations such as dilation and erosion, different types of boundary features are extracted to cope with nonlinear deformation of the structure; Adaptive transfer learning of deep learning: adapting the learned deflection features to new structures; Adaptive feature selection based on self-attention mechanism: Automatically identify the most important areas for deflection detection during feature extraction, ignore background noise, and adapt to different structural deflection characteristics by assigning different weights to different areas; The dynamic update of the deflection feature library includes the following process: Regularly extract new deflection features from historical test data, and continuously expand and update the feature library based on current test conditions; Automatically annotate data by detecting its internal correlation, and use this annotated data to train and optimize the model; The new data from each inspection is added to the model's training set in real time, the currently detected deflection features are compared with the historical feature library, and the model parameters are updated in real time when new changes or patterns are found.

[0012] To further optimize the technical solution, the multi-dimensional data fusion module uses a multi-sensor data fusion algorithm to find the best synergistic relationship between multi-source information and build a global deflection prediction model; Suppose we predict that a structure will be ) when the deflection ( ), using the output data of multiple sensors, such as visual data, stress, and displacement for fusion modeling; The deflection prediction model is as follows:

[0013] in, ( ) is the time ( )’s predicted deflection value; ( ) is the deflection of the two-dimensional visual information extracted from the visual inspection; ( ) is the stress sensor at time ( ) Stress data provided; ( ) is the displacement sensor at time ( ) provided displacement data; ( ) is the weighting coefficient, which indicates the contribution of different data sources to the model; ( ) is a time decay function, which is used to consider the impact of historical deflection on current deflection and reflect the temporal change of deflection; ( ) is a Gaussian noise term, representing measurement error or unpredictable randomness.

[0014] Further optimizing the technical solution, in the deflection prediction model, the change of deflection is time-dependent, and the deformation of the structure gradually increases with time accumulation. Based on the time attenuation function ( ) analyzes the impact of historical deflection on current deflection. The function adopts the following exponential decay form:

[0015] in,( ) is the time decay constant, indicating that the influence of historical deflection gradually weakens with the increase of time; The time decay function is used to capture the long-term deflection evolution of the structure to avoid relying solely on instantaneous data for prediction.

[0016] Further optimizing the technical solution, the deflection prediction model is optimized and tuned based on Bayesian reasoning and information entropy, the Bayesian reasoning is used for weight optimization, and the information entropy is used for uncertainty analysis; The Bayesian inference formula is as follows:

[0017] in,( ) is the sensor ( )’s weight distribution probability; ( ) is the likelihood of the data given the weights; ( ) is the prior probability of the weight, ( ) is the total probability of the data; The information entropy calculation formula is as follows:

[0018] in,( ) represents the distribution of sensor data, ( ) is the probability distribution of the sensor state; by calculating the data entropy values ​​of different sensors, the weights are intelligently assigned to reduce the interference of high entropy data.

[0019] To further optimize the technical solution, the dynamic evolution prediction module predicts the future deflection in advance by learning the historical deflection change trend of the structure, so as to prevent the structure from reaching a harmful deformation state; It is updated online based on real-time detection data, and deflection evolution analysis is performed based on the stress-strain relationship based on the structure, and the prediction results are visualized as an intuitive trend chart.

[0020] To further optimize the technical solution, the deflection fault warning module has a built-in AI-based anomaly detection algorithm, which accurately identifies potential abnormal situations by comparing historical data of structural deflection with current data, and determines whether the structure is in a dangerous state based on the influence of environmental factors, and sends warning signals through various methods such as mobile phone apps and computer terminals.

[0021] To further optimize the present technical solution, the self-learning and optimization module also uploads the detection data to the cloud based on collaborative analysis between edge computing and the cloud, compares and learns with other detection cases in the cloud, optimizes model parameters, and automatically identifies repetitive errors in long-term detection for optimization and adjustment.

[0022] Compared with the prior art, the present invention provides a visual dual-dimensional deflection detection system based on artificial intelligence, which has the following beneficial effects: This artificial intelligence-based visual two-dimensional deflection detection system overcomes the shortcomings of traditional systems such as single data source, manual feature extraction and poor robustness in dynamic environments by adopting adaptive feature extraction, deflection prediction model and self-learning and optimization module. The system can not only correct detection deviations in real time by integrating multi-dimensional vision, stress and displacement information to improve detection accuracy and stability, but also has the ability of autonomous optimization and updating, and can maintain efficient and accurate deflection prediction during long-term monitoring, thereby effectively ensuring the health and safety of the structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a structural schematic diagram of a visual dual-dimensional deflection detection system based on artificial intelligence proposed by the present invention; Figure 2 A schematic flow chart of an adaptive method in a visual dual-dimensional deflection detection system based on artificial intelligence proposed by the present invention; Figure 3 A schematic diagram of a process for dynamically updating a deflection feature library in a visual dual-dimensional deflection detection system based on artificial intelligence proposed by the present invention; Figure 4 The present invention is a flowchart of a deflection prediction model in a visual dual-dimensional deflection detection system based on artificial intelligence. DETAILED DESCRIPTION

[0024] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Example

[0025] See also Figure 1 , a visual two-dimensional deflection detection system based on artificial intelligence, includes a visual image acquisition module, a deflection feature extraction module, an intelligent error compensation module, a multi-dimensional data fusion module, a dynamic evolution prediction module, a deflection fault warning module and a self-learning and optimization module.

[0026] The visual image acquisition module acquires the two-dimensional image or point cloud data of the structure in real time based on a high-resolution camera or a laser radar.

[0027] In this embodiment, the visual image acquisition module uses a camera or a laser radar to avoid the influence of lighting or weather factors in a complex environment and obtain high-quality two-dimensional images or point cloud data; Cameras or lidars acquire 2D images or point cloud data from different angles by synchronizing multiple sensors to monitor the structure in all directions; The module also makes real-time adjustments based on changes in the external environment and automatically calibrates parameters to improve the accuracy and clarity of two-dimensional images or point cloud data.

[0028] The deflection feature extraction module automatically extracts deflection features from images through a deep learning model based on a convolutional neural network.

[0029] In this embodiment, the deflection feature extraction module is trained using a large amount of labeled structural deformation data to automatically identify and segment the deflection area; This module is based on an adaptive method to cope with a variety of complex deformation forms. It does not require a preset feature template and dynamically updates the deflection feature library based on historical detection data for self-optimization to improve the accuracy of feature extraction.

[0030] like Figure 2 As shown, the adaptive method includes the following process: Adaptation based on multi-scale feature extraction: A multi-scale convolutional neural network is used to extract both subtle and large-scale deflection features in convolutional layers of different scales; Morphological adaptation and geometric invariants: By combining geometric invariants and morphological feature extraction methods, geometric invariant features such as curvature and angle are selected to ensure the robustness of feature extraction; based on morphological operations such as dilation and erosion, different types of boundary features are extracted to cope with nonlinear deformation of the structure; Adaptive transfer learning of deep learning: adapting the learned deflection features to new structures; Adaptive feature selection based on self-attention mechanism: In the feature extraction process, the most important areas for deflection detection are automatically identified, and background noise is ignored. Different weights are assigned to different areas to adapt to different structural deflection characteristics.

[0031] like Figure 3 As shown, the dynamic update of the deflection feature library includes the following process: Regularly extract new deflection features from historical test data, and continuously expand and update the feature library based on current test conditions; Automatically annotate data by detecting its internal correlation, and use this annotated data to train and optimize the model; The new data from each inspection is added to the model's training set in real time, the currently detected deflection features are compared with the historical feature library, and the model parameters are updated in real time when new changes or patterns are found.

[0032] At the same time, it has the ability to detect anomalies and can identify characteristic anomalies that occur during the detection process, such as false deflection signals caused by equipment jitter. When an anomaly is detected, the system automatically corrects the parameters of the model and uses a feedback mechanism to exclude these anomalies from future detections to reduce false alarms.

[0033] The intelligent error compensation module may cause errors in data collection during deflection detection due to external factors such as camera shaking, light changes, etc. Based on the dynamic error compensation algorithm, the above effects are eliminated by identifying and tracking image feature points.

[0034] In this embodiment, the error is compensated in real time through time series analysis and Kalman filtering to ensure the accuracy of the deflection detection result.

[0035] The multi-dimensional data fusion module deeply fuses the two-dimensional visual information with sensor data, such as stress sensors and displacement sensors, to form a multi-modal data analysis framework. By fusing detection data of different dimensions, the deviation in visual deflection detection can be corrected in real time.

[0036] In this embodiment, the multi-dimensional data fusion module uses a multi-sensor data fusion algorithm to find the best synergistic relationship between multi-source information and build a global deflection prediction model; Suppose we predict that a structure will be ) when the deflection ( ), using the output data of multiple sensors, such as visual data, stress, and displacement for fusion modeling; The deflection prediction model is as follows:

[0037] in, ( ) is the time ( )’s predicted deflection value; ( ) is the deflection of the two-dimensional visual information extracted from the visual inspection; ( ) is the stress sensor at time ( ) Stress data provided; ( ) is the displacement sensor at time ( ) provided displacement data; ( ) is the weighting coefficient, which indicates the contribution of different data sources to the model; ( ) is a time decay function, which is used to consider the impact of historical deflection on current deflection and reflect the temporal change of deflection; ( ) is a Gaussian noise term, representing measurement error or unpredictable randomness.

[0038] Visual inspection data( ) comes from two-dimensional visual analysis, which uses cameras or lidar to obtain structural surface images and extract deflection features. Stress sensor data ( ) evaluates the potential impact of deflection by detecting changes in stress within the structure. Displacement sensors ( ) provides information on changes in the structural position, which is often an important signal during large or gradually accumulated structural deformation.

[0039] Each data source is weighted by a weighting factor ( ) to balance the contribution of data from different sources to the overall deflection prediction. The setting of weights can be adjusted by the relevance of features in the historical data. For example, in some scenarios, visual information may be more informative, while in other scenarios, stress or displacement data may dominate the deflection prediction.

[0040] In the deflection prediction model, the change of deflection is time-dependent, and the deformation of the structure gradually increases with time accumulation. Based on the time decay function ( ) analyzes the impact of historical deflection on current deflection. The function adopts the following exponential decay form:

[0041] in,( ) is the time decay constant, indicating that the influence of historical deflection gradually weakens with the increase of time; The time decay function is used to capture the long-term deflection evolution of the structure to avoid relying solely on instantaneous data for prediction.

[0042] The noise term in the model ( ) is assumed to be a Gaussian distribution, simulating various measurement errors in reality and randomness in the system. The existence of the noise term makes the model more robust and prevents the prediction from overfitting to a specific set of data. Through Bayesian reasoning, the noise term ( ) can be dynamically adjusted by updating the probability distribution to ensure the accuracy of the prediction results.

[0043] The deflection prediction model performs parameter optimization and tuning based on Bayesian reasoning and information entropy, wherein the Bayesian reasoning is used for weight optimization and the information entropy is used for uncertainty analysis; The Bayesian inference formula is as follows:

[0044] in,( ) is the sensor ( )’s weight distribution probability; ( ) is the likelihood of the data given the weights; ( ) is the prior probability of the weight, ( ) is the total probability of the data.

[0045] Through the above-mentioned Bayesian reasoning, the system can dynamically adjust the weights during real-time detection so that the synergy between the sensor data is optimized.

[0046] The information entropy calculation formula is as follows:

[0047] in,( ) represents the distribution of sensor data, ( ) is the probability distribution of the sensor state; by calculating the data entropy values ​​of different sensors, the weights are intelligently assigned to reduce the interference of high entropy data.

[0048] like Figure 4 As shown, in actual use, the deflection prediction model can be applied through the following steps: Real-time data acquisition and fusion: The system obtains real-time data from various sensors (such as cameras, stress sensors, displacement sensors) and inputs it into the deflection prediction model.

[0049] Multi-dimensional data weighted calculation: The system dynamically adjusts the weight of each sensor based on the Bayesian reasoning and information entropy calculation results ( ) and fuse these data.

[0050] Deflection evolution prediction: Using the time decay function ( ), the system considers current and historical deflection data and predicts future deflection changes.

[0051] Model updating and optimization: Through continuous detection feedback and data accumulation, the system dynamically adjusts model parameters to ensure the accuracy of the prediction model in long-term operation.

[0052] The dynamic evolution prediction module dynamically predicts the deflection change through time series analysis and long short-term memory network.

[0053] In this embodiment, the dynamic evolution prediction module predicts the future deflection in advance by learning the historical deflection change trend of the structure, so as to prevent the structure from reaching a harmful deformation state; It is updated online based on real-time detection data, and deflection evolution analysis is performed based on the stress-strain relationship based on the structure, and the prediction results are visualized as an intuitive trend chart.

[0054] The deflection fault warning module detects abnormal deflection and sends an alarm to the operator when the deflection exceeds a set safety threshold.

[0055] In this embodiment, the deflection fault warning module has a built-in AI-based anomaly detection algorithm, which accurately identifies potential abnormal situations by comparing the historical data of the structural deflection with the current data, and determines whether the structure is in a dangerous state based on the influence of environmental factors, and sends warning signals through various methods such as mobile phone apps and computer terminals.

[0056] The self-learning and optimization module is based on reinforcement learning and adaptive algorithms to continuously optimize the system's own performance according to actual detection data.

[0057] In this embodiment, the self-learning and optimization module also uploads the detection data to the cloud based on edge computing and cloud collaborative analysis, compares and learns with other detection cases in the cloud, optimizes model parameters, and automatically identifies repetitive errors in long-term detection and performs optimization adjustments.

[0058] The beneficial effects of the present invention are: This artificial intelligence-based visual two-dimensional deflection detection system overcomes the shortcomings of traditional systems such as single data source, manual feature extraction and poor robustness in dynamic environments by adopting adaptive feature extraction, deflection prediction model and self-learning and optimization module. The system can not only correct detection deviations in real time by integrating multi-dimensional vision, stress and displacement information to improve detection accuracy and stability, but also has the ability of autonomous optimization and updating, and can maintain efficient and accurate deflection prediction during long-term monitoring, thereby effectively ensuring the health and safety of the structure.

[0059] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0060] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A visual dual-dimensional deflection detection system based on artificial intelligence, characterized in that: It includes visual image acquisition module, deflection feature extraction module, intelligent error compensation module, multi-dimensional data fusion module, dynamic evolution prediction module, deflection fault warning module and self-learning and optimization module; The visual image acquisition module acquires the two-dimensional image or point cloud data of the structure in real time based on a high-resolution camera or a laser radar; The deflection feature extraction module automatically extracts deflection features from the image through a deep learning model based on a convolutional neural network; The intelligent error compensation module, based on a dynamic error compensation algorithm, compensates for errors in real time through recognition and tracking of image feature points, time series analysis and Kalman filtering, to ensure the accuracy of deflection detection results; The multi-dimensional data fusion module deeply fuses the two-dimensional visual information with sensor data, such as stress sensors and displacement sensors, to form a multi-modal data analysis framework, and corrects the deviation in visual deflection detection in real time by fusing detection data of different dimensions; The dynamic evolution prediction module dynamically predicts the deflection change through time series analysis and long short-term memory network; The deflection fault warning module detects abnormal deflection and issues an alarm to the operator when the deflection exceeds a set safety threshold; The self-learning and optimization module is based on reinforcement learning and adaptive algorithms to continuously optimize the system's own performance according to actual detection data.

2. According to the artificial intelligence-based visual dual-dimensional deflection detection system of claim 1, it is characterized in that: The visual image acquisition module uses a camera or a laser radar to avoid the influence of lighting or weather factors in complex environments and obtain high-quality two-dimensional images or point cloud data; Cameras or lidars acquire 2D images or point cloud data from different angles by synchronizing multiple sensors to monitor the structure in all directions; The module also makes real-time adjustments based on changes in the external environment and automatically calibrates parameters to improve the accuracy and clarity of two-dimensional images or point cloud data.

3. The visual dual-dimensional deflection detection system based on artificial intelligence according to claim 1 is characterized in that: The deflection feature extraction module is trained using a large amount of labeled structural deformation data to automatically identify and segment the deflection area; This module is based on an adaptive method to cope with a variety of complex deformation forms. It does not require a preset feature template and dynamically updates the deflection feature library based on historical detection data for self-optimization to improve the accuracy of feature extraction.

4. The artificial intelligence-based visual dual-dimensional deflection detection system according to claim 3 is characterized in that: The adaptive method The following processes are included: Adaptation based on multi-scale feature extraction: A multi-scale convolutional neural network is used to extract both subtle and large-scale deflection features in convolutional layers of different scales; Morphological adaptation and geometric invariants: By combining geometric invariants and morphological feature extraction methods, geometric invariant features such as curvature and angle are selected to ensure the robustness of feature extraction; based on morphological operations such as dilation and erosion, different types of boundary features are extracted to cope with nonlinear deformation of the structure; Adaptive transfer learning of deep learning: adapting the learned deflection features to new structures; Adaptive feature selection based on self-attention mechanism: Automatically identify the most important areas for deflection detection during feature extraction, ignore background noise, and adapt to different structural deflection characteristics by assigning different weights to different areas; The dynamic update of the deflection feature library includes the following process: Regularly extract new deflection features from historical test data, and continuously expand and update the feature library based on current test conditions; Automatically annotate data by detecting its internal correlation, and use this annotated data to train and optimize the model; The new data from each inspection is added to the model's training set in real time, the currently detected deflection features are compared with the historical feature library, and the model parameters are updated in real time when new changes or patterns are found.

5. The visual dual-dimensional deflection detection system based on artificial intelligence according to claim 1 is characterized in that: The multi-dimensional data fusion module uses a multi-sensor data fusion algorithm to find the best synergistic relationship between multi-source information and build a global deflection prediction model; Suppose we predict that a structure will be ) when the deflection ( ), using the output data of multiple sensors, such as visual data, stress, and displacement for fusion modeling; The deflection prediction model is as follows: ； in, ( ) is the time ( )’s predicted deflection value; ( ) is the deflection of the two-dimensional visual information extracted from the visual inspection; ( ) is the stress sensor at time ( ) Stress data provided; ( ) is the displacement sensor at time ( ) provided displacement data; ( ) is the weighting coefficient, which indicates the contribution of different data sources to the model; ( ) is a time decay function, which is used to consider the impact of historical deflection on current deflection and reflect the temporal change of deflection; ( ) is a Gaussian noise term, representing measurement error or unpredictable randomness.

6. The artificial intelligence-based visual dual-dimensional deflection detection system according to claim 5, characterized in that: In the deflection prediction model, the change of deflection is time-dependent, and the deformation of the structure gradually increases with time accumulation. Based on the time decay function ( ) analyzes the impact of historical deflection on current deflection. The function adopts the following exponential decay form: ； in,( ) is the time decay constant, indicating that the influence of historical deflection gradually weakens with the increase of time; The time decay function is used to capture the long-term deflection evolution of the structure to avoid relying solely on instantaneous data for prediction.

7. The artificial intelligence-based visual dual-dimensional deflection detection system according to claim 5, characterized in that: The deflection prediction model performs parameter optimization and tuning based on Bayesian reasoning and information entropy, wherein the Bayesian reasoning is used for weight optimization and the information entropy is used for uncertainty analysis; The Bayesian inference formula is as follows: ； in,( ) is the sensor ( )’s weight distribution probability; ( ) is the likelihood of the data given the weights; ( ) is the prior probability of the weight, ( ) is the total probability of the data; The information entropy calculation formula is as follows: ； in,( ) represents the distribution of sensor data, ( ) is the probability distribution of the sensor state; by calculating the data entropy values ​​of different sensors, the weights are intelligently assigned to reduce the interference of high entropy data.

8. The artificial intelligence-based visual dual-dimensional deflection detection system according to claim 1, characterized in that: The dynamic evolution prediction module predicts the future deflection in advance by learning the historical deflection change trend of the structure, so as to prevent the structure from reaching a harmful deformation state; It is updated online based on real-time detection data, and deflection evolution analysis is performed based on the stress-strain relationship based on the structure, and the prediction results are visualized as an intuitive trend chart.

9. The visual dual-dimensional deflection detection system based on artificial intelligence according to claim 1 is characterized in that: The deflection fault warning module has a built-in AI-based anomaly detection algorithm. By comparing the historical data of the structural deflection with the current data, it can accurately identify potential abnormal situations, and combine the influence of environmental factors to determine whether the structure is in a dangerous state. It also sends warning signals through various methods such as mobile phone apps and computer terminals.

10. The artificial intelligence-based visual dual-dimensional deflection detection system according to claim 1, characterized in that: The self-learning and optimization module is also based on edge computing and cloud collaborative analysis, uploads the detection data to the cloud, compares and learns with other detection cases in the cloud, optimizes model parameters, and automatically identifies repetitive errors in long-term detection and performs optimization adjustments.

Images