Binocular vision inertial navigation assisted pipeline defect automatic identification and report generation system

Abstract

The invention relates to the technical field of pipeline defect automatic identification and report generation systems, and discloses a binocular vision inertial navigation assisted pipeline defect automatic identification and report generation system, which comprises: a high dynamic range binocular camera module, which is used for collecting a high contrast stereo image in a pipeline under an extreme illumination condition; and the enhanced inertial measurement unit is combined with a magnetometer and a barometer and is used for providing more accurate position and attitude information of the mobile equipment in the pipeline. According to the pipeline defect automatic identification and report generation system assisted by binocular vision inertial navigation, a high-dynamic-range binocular camera module and an enhanced inertial measurement unit are integrated, so that a high-quality pipeline internal image can be obtained under an extreme illumination condition, and accurate position and attitude information is combined, so that the pipeline defect automatic identification and report generation system is realized. The accuracy and reliability of defect detection are improved, and the adaptive defect detection algorithm module uses a mixed architecture of a convolutional neural network and a recurrent neural network.

Description

Technical Field

[0001] The present invention relates to the technical field of pipeline defect automatic identification and report generation systems, and in particular to a pipeline defect automatic identification and report generation system assisted by binocular visual inertial navigation. Background Art

[0002] The automatic identification and report generation system for pipeline defects is an intelligent system that integrates multiple advanced technologies. It is designed to automatically detect and evaluate defects inside pipelines and generate detailed reports. The system is widely used in pipeline maintenance and inspection in the oil, natural gas, water supply, drainage and other industries. It can significantly improve detection efficiency and accuracy, reduce manual intervention, and reduce maintenance costs. However, the images collected by the existing technology under extreme lighting conditions often have problems of overexposure or underexposure, resulting in low image quality and affecting the accuracy of defect detection. In addition, the existing pipeline inspection system relies on a single sensor to obtain position and posture information, which is easily affected by noise and errors. At the same time, the existing technology often relies on a single image processing method for defect detection, which is difficult to deal with complex and diverse pipeline defects. Summary of the invention

[0003] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] The binocular visual inertial navigation-assisted pipeline defect automatic identification and report generation system consists of:

[0006] High dynamic range stereo camera module for capturing high contrast stereo images of the pipeline interior in extreme lighting conditions;

[0007] Enhanced inertial measurement unit, combined with magnetometer and barometer, to provide more accurate position and attitude information of mobile devices within pipelines;

[0008] An image processing module, connected to the high dynamic range binocular camera module, for creating an enhanced image based on the collected high contrast stereo image;

[0009] An adaptive defect detection algorithm module, connected to the image processing module and the enhanced inertial measurement unit, which uses a hybrid architecture of a convolutional neural network and a recurrent neural network to automatically identify defects in the pipeline based on the enhanced image and the position and posture data provided by the IMU;

[0010] The adaptive defect detection algorithm module also has self-learning capabilities and can adapt to different pipeline environments by updating model parameters online, thereby improving accuracy and robustness in long-term use.

[0011] As a further solution of the present invention: the system module also includes a multi-dimensional report generation module, which is connected to the adaptive defect detection algorithm module and is used to automatically generate a comprehensive report document containing information such as defect location, type, severity, historical change trend, etc. according to the results of the defect detection algorithm module, and supports 3D visualization and AR display, so that technicians can intuitively understand the pipeline defect situation;

[0012] The cloud collaborative processing unit is connected with the image processing module, the adaptive defect detection algorithm module and the multi-dimensional report generation module, and can upload the locally collected data to the cloud server, use powerful computing resources to perform complex data analysis and model training, and support multi-user access and data sharing to promote team collaboration;

[0013] Secure encrypted communication protocol is used to ensure that all transmitted data is encrypted to prevent leakage of sensitive information and protect user data security.

[0014] As a further solution of the present invention: the high dynamic range binocular camera module includes automatic exposure control, dynamically adjusting the exposure time according to the current lighting conditions to ensure the clarity and contrast of the image, post-image preprocessing, including denoising, color correction and image enhancement to improve the image quality, and stereo matching, using the two images captured by the binocular camera for stereo matching to generate a high-contrast stereo image.

[0015] Furthermore, the automatic exposure control process is as follows:

[0016] Collect the histogram of the current image;

[0017] Calculate the cumulative distribution function of the histogram;

[0018] Adjust exposure time according to the cumulative distribution function;

[0019] Histogram H:

[0020] Where P(j) is the frequency of occurrence of pixel value j;

[0021] Cumulative distribution function CDF:

[0022] Adjust the exposure time t:

[0023] Among them, t0 is the initial exposure time, target_value is the target pixel value;

[0024] The denoising uses bilateral filtering, and the formula is as follows:

[0025]

[0026] Among them, I(p) is the grayscale value of pixel p, S is the neighborhood window, gσs and gσr are the spatial Gaussian kernel and the range Gaussian kernel, respectively.

[0027] As a further solution of the present invention: the enhanced inertial measurement unit specifically includes:

[0028] Data collection: real-time collection of the position and attitude information of mobile devices in the pipeline, including acceleration, angular velocity, magnetic field strength and air pressure data;

[0029] Data correction: Use Kalman filter to correct the collected data and eliminate noise and errors;

[0030] Data fusion, which fuses the corrected data with the image data collected by the binocular camera module to provide more accurate position and posture information;

[0031] Furthermore, the data correction adopts Kalman filter, which is as follows:

[0032] Prediction steps:

[0033]

[0034] Update steps:

[0035]

[0036] P k|k =(IK k H k ) k|k-1 ;

[0037] Where x^ is the state estimate, P is the estimation error covariance, F is the state transfer matrix, B is the control input matrix, u is the control input, Q is the process noise covariance, H is the observation matrix, R is the observation noise covariance, K is the Kalman gain, and z is the observation.

[0038] As a further solution of the present invention: image alignment is required in the image processing module, the two images captured by the binocular camera module are aligned to eliminate parallax, and multi-scale image fusion technology is used to fuse images of different resolutions to improve the detail expression of the image, and image enhancement algorithm, histogram equalization and contrast stretching are applied to improve the contrast and clarity of the image.

[0039] As a further solution of the present invention: the adaptive defect detection algorithm module includes:

[0040] Feature extraction: using convolutional neural networks to extract feature information from enhanced images and generate feature data sets;

[0041] Feature fusion, fusing the feature data set with the position and attitude data provided by the IMU to generate a multimodal feature vector;

[0042] Defect classification, using recurrent neural network to perform regression prediction on multimodal feature vectors to determine the key point information and category of defects;

[0043] Online learning, by updating model parameters online, adapts to different pipeline environments and improves accuracy and robustness for long-term use.

[0044] As a further solution of the present invention: the adaptive defect detection algorithm module also includes:

[0045] Anomaly detection, identifying and filtering out false alarms caused by external interference, improving the system's anti-interference ability;

[0046] Adaptive threshold adjustment: dynamically adjust the threshold of defect detection according to environmental changes to improve detection accuracy;

[0047] Real-time feedback: The test results are fed back to the user in real time so that timely measures can be taken.

[0048] As a further solution of the present invention: the multi-dimensional report generation module includes result analysis, which analyzes the detection results generated by the adaptive defect detection algorithm module, extracts information such as defect location, type, severity, etc., and requires report generation at the same time. A comprehensive report document containing information such as defect location, type, severity, historical change trend, etc. is generated according to the analysis results, and a visual display is performed, supporting 3D visualization and AR display, generating an intuitive report format, and then outputting the report. The generated report document is output in multiple formats, including PDF, Excel table and 3D visualization report.

[0049] As a further solution of the present invention: the secure encrypted communication protocol comprises the following steps:

[0050] Data encryption: end-to-end encryption of all transmitted data to ensure data security;

[0051] Key management: adopt a secure key management mechanism, update keys regularly, and prevent key leakage;

[0052] Data verification: Verify the integrity of the transmitted data to ensure that the data has not been tampered with;

[0053] Access control: Implement strict access control policies so that only authorized users can access and download data.

[0054] As a further solution of the present invention: the multi-dimensional report generation module further includes:

[0055] Historical data analysis, generating a time series analysis report containing historical inspection records, helping users track the changing trends of pipeline defects;

[0056] Trend forecasting, using time series analysis models to predict possible future defect developments;

[0057] The early warning mechanism generates early warning information based on the prediction results to remind users to take preventive measures in advance.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] By integrating a high dynamic range binocular camera module and an enhanced inertial measurement unit, it is possible to obtain high-quality images of the interior of the pipeline under extreme lighting conditions, and combined with precise position and attitude information, the accuracy and reliability of defect detection are improved. The adaptive defect detection algorithm module uses a hybrid architecture of convolutional neural networks and recurrent neural networks, which can not only automatically identify defects in the pipeline, but also continuously optimize the model through online learning to adapt to pipeline detection needs in different environments, thereby enhancing the intelligence and robustness of the system. The Kalman filter is used to correct the data collected by the IMU to eliminate noise and errors, and then the corrected data is fused with the image data of the binocular camera to provide more accurate position and attitude information. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 This is a system connection diagram of the binocular visual inertial navigation-assisted pipeline defect automatic identification and report generation system. DETAILED DESCRIPTION

[0061] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and understandable, the specific implementation modes of the present invention are described in detail below in conjunction with the accompanying drawings.

[0062] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0063] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0064] Example

[0065] See also Figure 1 , is an embodiment of the present invention, which provides a binocular visual inertial navigation-assisted pipeline defect automatic identification and report generation system, including:

[0066] High dynamic range binocular camera module;

[0067] Enhanced Inertial Measurement Unit;

[0068] Image processing module;

[0069] Adaptive defect detection algorithm module;

[0070] Multi-dimensional report generation module;

[0071] Cloud collaborative processing unit;

[0072] Secure encrypted communication protocol;

[0073] Hardware configuration:

[0074] Camera model: XYZ-1234;

[0075] Resolution: 1920x1080;

[0076] Dynamic range: 120dB;

[0077] Automatic exposure control:

[0078] Collect the histogram of the current image: Use the camera to collect the image in the current environment and calculate the frequency P(j) of each pixel value;

[0079] Compute the cumulative distribution function of a histogram:

[0080]

[0081] Adjust the exposure time t:

[0082] Among them, t0 is the initial exposure time, which is set to 10ms, target_value is the target pixel value, which is set to 128;

[0083] The denoising uses bilateral filtering, and the formula is as follows:

[0084]

[0085] Where I(p) is the grayscale value of pixel p, S is the neighborhood window, which is set to 5x5, gσs and gσr are the spatial Gaussian kernel and the range Gaussian kernel, which are set to 1.5 and 0.1 respectively;

[0086] Color correction and image enhancement: Perform color correction and histogram equalization to improve the contrast and clarity of the image;

[0087] Stereo matching: Use two images captured by a binocular camera to perform stereo matching to generate a high-contrast stereo image;

[0088] Enhanced inertial measurement unit, hardware configuration:

[0089] I MU model: ABC-5678;

[0090] Sampling rate: 100Hz;

[0091] Sensor type: three-axis accelerometer, three-axis gyroscope, three-axis magnetometer, and barometer;

[0092] Data collection:

[0093] Collect the position and attitude information of mobile devices in the pipeline in real time, including acceleration, angular velocity, magnetic field strength and air pressure data;

[0094] Data correction:

[0095] The Kalman filter is used to correct the collected data to eliminate noise and errors. The specific steps are as follows:

[0096] Prediction steps:

[0097]

[0098] Update steps:

[0099]

[0100] P k|k =(IK k H k ) k|k-1 ;

[0101] Where x^ is the state estimate, P is the estimation error covariance, F is the state transfer matrix, B is the control input matrix, u is the control input, Q is the process noise covariance, H is the observation matrix, R is the observation noise covariance, K is the Kalman gain, and z is the observation value;

[0102] Data Fusion:

[0103] Data preparation:

[0104] I MU data: including acceleration, angular velocity, magnetic field strength and air pressure data;

[0105] Image data: high-contrast stereo images collected by binocular cameras;

[0106] Data preprocessing:

[0107] IMU data correction: Use Kalman filter to correct IMU data to eliminate noise and errors;

[0108] Image preprocessing: denoising, color correction and image enhancement;

[0109] Time Synchronization:

[0110] Synchronize IMU and image data: Ensure that the timestamps of IMU data and image data are consistent. Interpolation methods can be used for time alignment.

[0111] Feature extraction:

[0112] IMU features: Extract acceleration, angular velocity, magnetic field strength and air pressure features from IMU data;

[0113] Image features: Use the ORB feature matching algorithm to extract key points and descriptors in the image;

[0114] Feature Fusion:

[0115] Multilayer Perceptron:

[0116] Input layer: contains IMU features and image features;

[0117] Hidden layer: multi-layer neural network using ReLU activation function;

[0118] Output layer: outputs the fused multimodal feature vector;

[0119] Fusion results:

[0120] Position information: Combine the acceleration and angular velocity data of the IMU to calculate the precise position of the device;

[0121] Attitude information: Combine the magnetic field strength and air pressure data of IMU to calculate the attitude information of the device;

[0122] Image information: Combine image features to provide richer visual information;

[0123] Image processing module:

[0124] Image Alignment:

[0125] Specific steps:

[0126] Feature Matching:

[0127] ORB feature matching: Use the ORB feature matching algorithm to find the corresponding points between two images;

[0128] Feature extraction: extract ORB feature points and descriptors in two images;

[0129] Matching point pairs: Use Hamming distance to match feature point pairs;

[0130] Homography matrix estimation:

[0131] RANSAC algorithm: randomly select four pairs of points from the matching point pairs and calculate the homography matrix;

[0132] Model verification: Use the remaining matching point pairs to verify the validity of the homography matrix and select the best model;

[0133] Disparity map generation:

[0134] Semi-global matching: Use the SGM algorithm to generate a disparity map and optimize the alignment effect;

[0135] Cost calculation: calculate the matching cost of each pixel;

[0136] Path aggregation: Aggregate path costs in multiple directions to generate a disparity map;

[0137] Multi-scale image fusion:

[0138] Specific steps:

[0139] Image decomposition:

[0140] Laplacian pyramid: decomposes an image into sub-band images of multiple scales;

[0141] Gaussian Pyramid: Use Gaussian filter to downsample the image to generate images of different resolutions;

[0142] Fusion processing:

[0143] Weighted average: At each scale, the two images are weighted averaged;

[0144] Gradient-based method: select the appropriate fusion strategy based on the gradient information of the image;

[0145] Image reconstruction:

[0146] Reconstructed image: Recombining the fused sub-band images to form the final multi-scale fused image;

[0147] Inverse pyramid: Use the inverse pyramid method to restore images of different scales layer by layer;

[0148] Image Enhancement

[0149] Specific steps:

[0150] Histogram Equalization:

[0151] Local histogram equalization: Divide the image into multiple small blocks, perform histogram equalization on each small block to improve local contrast;

[0152] Contrast stretching:

[0153] Adaptive contrast stretching: automatically adjust the contrast range according to the local statistical characteristics of the image;

[0154] Compute local statistics: compute the mean and standard deviation around each pixel;

[0155] Adjust contrast: adjust pixel values ​​according to mean and standard deviation to improve the overall visual effect of the image;

[0156] Adaptive defect detection algorithm module:

[0157] Feature extraction:

[0158] Specific steps:

[0159] Convolutional Neural Networks:

[0160] Pre-trained model: Use a pre-trained deep convolutional neural network as a feature extractor;

[0161] Feature map: extract deep feature map from input image;

[0162] Pooling operation:

[0163] Max pooling: Use the max pooling operation to reduce the size of the feature map and retain important information;

[0164] Average pooling: Use average pooling operation to smooth feature maps and reduce noise;

[0165] Feature Fusion:

[0166] Specific steps:

[0167] Multilayer Perceptron:

[0168] Input layer: contains image features extracted by CNN and position and posture data provided by IMU;

[0169] Hidden layer: multi-layer neural network using ReLU activation function;

[0170] Output layer: outputs the fused multimodal feature vector;

[0171] Feature stitching:

[0172] Feature splicing: splice image features and IMU data into a vector;

[0173] Normalization: Normalize the concatenated feature vectors to ensure that the scales of different features are consistent;

[0174] Multi-dimensional report generation module: result analysis, analyze the detection results generated by the adaptive defect detection algorithm module, extract defect location, type, severity and other information, and generate a report. According to the analysis results, a comprehensive report document containing defect location, type, severity, historical change trend and other information is generated, and visualized. It supports 3D visualization and AR display, generates intuitive report forms, and then outputs the report. The generated report document is output in multiple formats, including PDF, Excel table and 3D visualization report;

[0175] Historical data analysis, generating a time series analysis report containing historical inspection records, helping users track the changing trends of pipeline defects;

[0176] Trend forecasting, using time series analysis models to predict possible future defect developments;

[0177] Early warning mechanism, which generates early warning information based on the prediction results to remind users to take preventive measures in advance;

[0178] Secure encrypted communication protocol:

[0179] Data encryption: End-to-end encryption of all transmitted data to ensure data security;

[0180] Key management: adopt a secure key management mechanism, update keys regularly, and prevent key leakage;

[0181] Data verification: Verify the integrity of the transmitted data to ensure that the data has not been tampered with;

[0182] Access control: Implement strict access control policies so that only authorized users can access and download data.

[0183] In summary, by integrating a high dynamic range binocular camera and an enhanced inertial measurement unit, and through advanced image processing technology, adaptive defect detection algorithms, and multi-dimensional report generation technology, accurate identification of internal defects in pipelines and efficient report generation are achieved. In particular, by introducing innovative technologies such as adaptive exposure control, bilateral filtering denoising, and multi-layer perceptron feature fusion, image quality and defect detection accuracy are significantly improved.

[0184] At the same time, the Kalman filter is used to correct the IMU data and combined with the image data for multimodal fusion, which further enhances the robustness and reliability of the system. In addition, the system also has powerful cloud-based collaborative processing capabilities, strict security encryption communication protocols, and multi-dimensional report generation functions that support 3D visualization and AR display, providing users with a comprehensive and intelligent pipeline inspection solution.

[0185] Importantly, it should be noted that the construction and arrangement of the present application shown in a plurality of different exemplary embodiments are only exemplary. Although only a few embodiments are described in detail in this disclosure, it should be readily understood by those who refer to this disclosure that many modifications are possible (e.g., the size, scale, structure, shape and proportion of various elements, and parameter values ​​(e.g., temperature, pressure, etc.), installation arrangement, use of materials, color, directional changes, etc.) without substantially departing from the novel teachings and advantages of the subject matter described in the application. For example, the element shown as integrally formed can be composed of multiple parts or elements, the position of the element can be inverted or otherwise changed, and the nature or number or position of the discrete element can be changed or changed. Therefore, all such modifications are intended to be included in the scope of the present invention. The order or sequence of any process or method steps can be changed or reordered according to alternative embodiments. In the claims, any "device plus function" clause is intended to cover the structure of performing the function described herein, and is not only structurally equivalent but also equivalent structure. Without departing from the scope of the present invention, other replacements, modifications, changes and omissions can be made in the design, operating conditions and arrangement of the exemplary embodiments. Therefore, the invention is not limited to a specific embodiment, but extends to several modifications still falling within the scope of the appended claims.

[0186] Additionally, in order to provide a concise description of exemplary embodiments, all features of an actual embodiment (ie, those features that are not relevant to the best mode presently contemplated for carrying out the invention or those that are not relevant to implementing the invention) may not be described.

[0187] It will be appreciated that in the development of any actual implementation, as in any engineering or design project, numerous implementation-specific decisions may be made. Such a development effort may be complex and time-consuming, but will be a routine task of design, fabrication, and production for those of ordinary skill having the benefit of this disclosure without undue experimentation.

[0188] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the 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. The binocular visual inertial navigation-assisted pipeline defect automatic identification and report generation system is characterized by: The system consists of: High dynamic range stereo camera module for capturing high contrast stereo images of the pipeline interior in extreme lighting conditions; Enhanced inertial measurement unit, combined with magnetometer and barometer, to provide more accurate position and attitude information of mobile devices within pipelines; An image processing module, connected to the high dynamic range binocular camera module, for creating an enhanced image based on the collected high contrast stereo image; An adaptive defect detection algorithm module, connected to the image processing module and the enhanced inertial measurement unit, which uses a hybrid architecture of a convolutional neural network and a recurrent neural network to automatically identify defects in the pipeline based on the enhanced image and the position and posture data provided by the IMU; The adaptive defect detection algorithm module also has self-learning capabilities and can adapt to different pipeline environments by updating model parameters online, thereby improving accuracy and robustness in long-term use.

2. The binocular visual inertial navigation-assisted pipeline defect automatic identification and report generation system according to claim 1 is characterized in that: The system module also includes a multi-dimensional report generation module, which is connected to the adaptive defect detection algorithm module and is used to automatically generate a comprehensive report document including defect location, type, severity and historical change trend according to the results of the defect detection algorithm module, and supports 3D visualization and AR display, so that technicians can intuitively understand the pipeline defect situation; The cloud collaborative processing unit is connected to the image processing module, the adaptive defect detection algorithm module and the multi-dimensional report generation module, and can upload the locally collected data to the cloud server, use powerful computing resources to perform complex data analysis and model training, and support multi-user access and data sharing to promote team collaboration; Secure encrypted communication protocol is used to ensure that all transmitted data is encrypted to prevent leakage of sensitive information and protect user data security.

3. The binocular visual inertial navigation-assisted pipeline defect automatic identification and report generation system according to claim 1 is characterized in that: The high dynamic range binocular camera module includes automatic exposure control, dynamically adjusting the exposure time according to the current lighting conditions to ensure the clarity and contrast of the image, post-image preprocessing, including denoising, color correction and image enhancement to improve the quality of the image, and stereo matching, using two images captured by the binocular camera for stereo matching to generate a high-contrast stereo image.

4. The binocular visual inertial navigation-assisted pipeline defect automatic identification and report generation system according to claim 1 is characterized in that: The enhanced inertial measurement unit specifically includes: Data collection: real-time collection of the position and attitude information of mobile devices in the pipeline, including acceleration, angular velocity, magnetic field strength and air pressure data; Data correction: Use Kalman filter to correct the collected data and eliminate noise and errors; Data fusion, which fuses the corrected data with the image data collected by the binocular camera module to provide more accurate position and posture information.

5. The binocular visual inertial navigation-assisted pipeline defect automatic identification and report generation system according to claim 1 is characterized in that: The image processing module requires image alignment to align the two images collected by the binocular camera module to eliminate parallax, and use multi-scale image fusion technology to fuse images of different resolutions to improve the detail expression of the image. Image enhancement algorithm, histogram equalization and contrast stretching are applied to improve the contrast and clarity of the image.

6. The binocular visual inertial navigation-assisted pipeline defect automatic identification and report generation system according to claim 1 is characterized by: The adaptive defect detection algorithm module includes: Feature extraction: using convolutional neural networks to extract feature information from enhanced images and generate feature data sets; Feature fusion, fusing the feature data set with the position and attitude data provided by the IMU to generate a multimodal feature vector; Defect classification, using recurrent neural network to perform regression prediction on multimodal feature vectors to determine the key point information and category of defects; Online learning, by updating model parameters online, adapts to different pipeline environments and improves accuracy and robustness for long-term use.

7. The binocular visual inertial navigation-assisted pipeline defect automatic identification and report generation system according to claim 1 is characterized in that: The adaptive defect detection algorithm module also includes: Anomaly detection, identifying and filtering out false alarms caused by external interference, improving the system's anti-interference ability; Adaptive threshold adjustment: dynamically adjust the threshold of defect detection according to environmental changes to improve detection accuracy; Real-time feedback: The test results are fed back to the user in real time so that timely measures can be taken.

8. The binocular visual inertial navigation-assisted pipeline defect automatic identification and report generation system according to claim 2 is characterized by: The multi-dimensional report generation module includes result analysis, which analyzes the detection results generated by the adaptive defect detection algorithm module, extracts the defect location, type and severity, and requires report generation. A comprehensive report document containing the defect location, type, severity and historical change trend is generated according to the analysis results, and a visual display is performed. 3D visualization and AR display are supported, and an intuitive report format is generated. After the report is output, the generated report document is output in multiple formats, including PDF, Excel tables and 3D visualization reports.

9. The binocular visual inertial navigation-assisted pipeline defect automatic identification and report generation system according to claim 2 is characterized by: The secure encrypted communication protocol includes the following steps: Data encryption: End-to-end encryption of all transmitted data to ensure data security; Key management: adopt a secure key management mechanism, update keys regularly, and prevent key leakage; Data verification: Verify the integrity of the transmitted data to ensure that the data has not been tampered with; Access control: Implement strict access control policies so that only authorized users can access and download data.

10. The binocular visual inertial navigation-assisted pipeline defect automatic identification and report generation system according to claim 2, characterized in that: The multi-dimensional report generation module also includes: Historical data analysis, generating a time series analysis report containing historical inspection records, helping users track the changing trends of pipeline defects; Trend forecasting, using time series analysis models to predict possible future defect developments; The early warning mechanism generates early warning information based on the prediction results to remind users to take preventive measures in advance.

Images