Target-specific high-precision displacement monitoring method and related device

Abstract

The application discloses a high-precision displacement monitoring method based on a specific target and related equipment, and relates to the technical field of data processing, and comprises the following steps: acquiring an initial target image corresponding to a feature image on a specific target, continuously collecting the feature image on the specific target by using a camera based on a preset image collection frequency, obtaining a plurality of first target images, determining initial sub-pixel image coordinates of a plurality of feature points corresponding to the initial target image, and determining first sub-pixel image coordinates of the plurality of feature points corresponding to the first target image, and performing displacement monitoring based on the initial sub-pixel image coordinates and the first sub-pixel image coordinates. Based on the specific target, the long-term stability of the physical size thereof can be ensured in a complex outdoor environment, and then the sub-pixel coordinates of the feature image can be continuously and stably acquired by using the camera, and therefore, the deformation monitoring of underground gas pipelines, high slopes, foundation pits and other infrastructure projects can be efficiently and accurately performed.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a high-precision displacement monitoring method and related equipment based on a specific target. Background Technology

[0002] In related technologies, the stability of gas pipeline projects, high slope projects, and foundation pit projects is directly related to the safety of infrastructure. Existing automated monitoring methods are mainly divided into two categories: one is contact measurement, such as GNSS (Global Navigation Satellite System) and total stations, which require on-site deployment of sensors and cables, resulting in drawbacks such as complex installation, high cost, sparse monitoring points, and susceptibility to environmental influences; the other is traditional visual measurement methods, such as Digital Image Correlation (DIC), which, although enabling non-contact monitoring, faces fundamental challenges in the field: natural textures are greatly affected by lighting and seasonal changes, their features are unstable, and they are difficult to work effectively in areas without texture or with repetitive textures, requiring large amounts of computation and making it difficult to achieve high-frequency real-time monitoring.

[0003] In recent years, although deep learning-based visual methods have been proposed, existing solutions mostly transmit image data to the cloud or backend servers for processing. This results in problems such as high data transmission bandwidth pressure, high network latency, and poor real-time performance, making it difficult to meet the needs of rapid response safety monitoring. Therefore, there is an urgent need for a high slope deformation monitoring solution that can adapt to complex field environments, has high computational efficiency, is easy to deploy, and can achieve real-time intelligent analysis at the edge. Summary of the Invention

[0004] The main purpose of this application is to provide a high-precision displacement monitoring method and related equipment based on a specific target, aiming to solve the technical problem of how to efficiently and accurately monitor the deformation of high slopes in complex field environments.

[0005] To achieve the above objectives, this application proposes a high-precision displacement monitoring method based on a specific target, applicable to complex field environments. The high-precision displacement monitoring method based on a specific target includes: In response to a displacement monitoring command, an initial target image corresponding to a feature image on a specific target is acquired. Based on a preset image acquisition frequency, a camera continuously acquires feature images on the specific target to obtain multiple sets of first target images. The feature images are circular or reflective markers. The feature images are fabricated on Invar alloy, ceramic, or a rigid substrate covered with a protective film to ensure the long-term stability of their physical dimensions. The camera is fixed at a stable reference point, and the specific target is fixed at the monitoring point of the structure to be measured. The specific target has at least three feature images arranged in a non-collinear and known precise physical coordinate relationship. Determine the initial subpixel image coordinates of multiple feature points corresponding to the initial target image, and determine the first subpixel image coordinates of multiple feature points corresponding to the first target image; Displacement monitoring is performed based on the initial subpixel image coordinates and the first subpixel image coordinates.

[0006] In one embodiment, the step of performing displacement monitoring based on the initial subpixel image coordinates and the first subpixel image coordinates further includes: Obtain the physical world coordinates corresponding to each feature point; Based on the physical world coordinates, the initial subpixel image coordinates, and the first subpixel image coordinates, the pose of the specific target is calculated using a preset EPnP algorithm to obtain the initial pose and multiple sets of first poses corresponding to the specific target. Displacement monitoring is performed based on the initial pose and the multiple sets of first poses.

[0007] In one embodiment, the step of performing displacement monitoring based on the initial pose and the multiple sets of first poses further includes: Calculate the difference between the first pose and the initial pose at different acquisition times to obtain the three-dimensional translational displacement and three-dimensional rotation at different acquisition times. Displacement monitoring is performed based on the three-dimensional translational displacement and the three-dimensional rotational displacement.

[0008] In one embodiment, the step of determining the initial sub-pixel image coordinates of multiple feature points corresponding to the initial target image further includes: The camera is calibrated using a two-dimensional calibration board to obtain the camera's intrinsic parameter matrix and lens distortion coefficients. Based on the intrinsic parameter matrix and the lens distortion coefficient, the initial target image is corrected to obtain the corrected initial target image; Based on the corrected initial target image, the initial subpixel image coordinates of multiple feature points corresponding to the initial target image are determined.

[0009] In one embodiment, the step of determining the initial sub-pixel image coordinates of multiple feature points corresponding to the initial target image based on the corrected initial target image further includes: Based on the preset target detection model, feature point extraction is performed on the corrected initial target image to obtain multiple feature points corresponding to the corrected initial target image; Sub-pixel image coordinate extraction is performed on the multiple feature points to obtain the initial sub-pixel image coordinates of the multiple feature points corresponding to the initial target image.

[0010] In one embodiment, before the step of performing feature point extraction on the corrected initial target image based on a preset target detection model, the method further includes: Acquire sample data, the sample data corresponding to the first feature point extraction result; The sample data is processed using the current object detection model to obtain the second feature point extraction result; Determine whether the first feature point extraction result and the second feature point extraction result are consistent. If they are inconsistent, adjust the parameters of the current target detection model. Based on the current target detection model with adjusted parameters, return to the step of using the current target detection model to process the sample data and obtain the second feature point extraction result. Continue until the first feature point extraction result and the second feature point extraction result are consistent, and obtain the preset target detection model.

[0011] Furthermore, to achieve the above objectives, this application also proposes a high-precision displacement monitoring device based on a specific target, the high-precision displacement monitoring device based on a specific target comprising: The acquisition module is used to respond to displacement monitoring commands, acquire initial target images corresponding to feature images on a specific target, and continuously acquire feature images on the specific target using a camera based on a preset image acquisition frequency to obtain multiple sets of first target images. The feature images are circular or reflective markers. The feature images are fabricated on Invar alloy, ceramic, or a rigid substrate covered with a protective film to ensure the long-term stability of their physical dimensions. The camera is fixed at a stable reference point, and the specific target is fixed at the monitoring point of the structure to be measured. The specific target has at least three feature images arranged in a non-collinear and known precise physical coordinate relationship. The determining module is used to determine the initial subpixel image coordinates of multiple feature points corresponding to the initial target image, and to determine the first subpixel image coordinates of multiple feature points corresponding to the first target image. A displacement monitoring module is used to perform displacement monitoring based on the initial subpixel image coordinates and the first subpixel image coordinates.

[0012] In addition, to achieve the above objectives, this application also proposes a high-precision displacement monitoring device based on a specific target, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the high-precision displacement monitoring method based on a specific target as described above.

[0013] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the high-precision displacement monitoring method based on a specific target as described above.

[0014] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the high-precision displacement monitoring method based on a specific target as described above.

[0015] One or more technical solutions proposed in this application have at least the following technical effects: This application discloses a high-precision displacement monitoring method and related equipment based on a specific target, relating to the field of data processing technology. In contrast, existing automated monitoring methods are mainly divided into two categories: one is contact measurement, such as GNSS and total stations, which requires on-site deployment of sensors and cables, resulting in complex installation, high cost, sparse monitoring points, and susceptibility to environmental influences; the other is traditional visual measurement methods, such as digital image correlation (DIC), which, while enabling non-contact monitoring, faces fundamental challenges in field environments: natural textures are greatly affected by lighting and seasonal changes, their features are unstable, and they are difficult to work effectively in areas without texture or with repetitive textures, requiring large computational loads and making high-frequency real-time monitoring difficult. In this application, firstly, in response to a displacement monitoring command, an initial target image corresponding to the feature image on the specific target is acquired, and... Based on a preset image acquisition frequency, a camera continuously acquires feature images on the specific target to obtain multiple sets of first target images. The feature images are circular or reflective markers. The feature images are fabricated on Invar alloy, ceramic, or a rigid substrate covered with a protective film to ensure the long-term stability of their physical dimensions. The camera is fixed at a stable reference point, and the specific target is fixed at the monitoring point of the structure under test. The specific target has at least three feature images arranged in a non-collinear and known precise physical coordinate relationship. Then, the initial sub-pixel image coordinates of multiple feature points corresponding to the initial target image are determined, and the first sub-pixel image coordinates of multiple feature points corresponding to the first target image are determined. Finally, displacement monitoring is performed based on the initial sub-pixel image coordinates and the first sub-pixel image coordinates.

[0016] Understandably, this application, based on a specific target, can ensure the long-term stability of its physical dimensions in complex field environments. This allows for the continuous and stable acquisition of sub-pixel coordinates of feature images using a camera, thereby enabling efficient and accurate deformation monitoring of infrastructure projects such as underground gas pipelines, high slopes, and foundation pits. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating an embodiment of the high-precision displacement monitoring method based on a specific target provided in this application. Figure 2 This is a flowchart illustrating Embodiment 2 of the high-precision displacement monitoring method based on a specific target provided in this application; Figure 3 This is a flowchart illustrating Embodiment 3 of the high-precision displacement monitoring method based on a specific target provided in this application; Figure 4 This is a schematic diagram of the module structure of a high-precision displacement monitoring device based on a specific target, as described in an embodiment of this application. Figure 5 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the high-precision displacement monitoring method based on a specific target in the embodiments of this application.

[0020] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0022] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0023] The main solution in this application's embodiments is: In this embodiment, for ease of description, the following description uses a high-precision displacement monitoring device based on a specific target as the execution subject.

[0024] Due to the relevant technologies, existing automated monitoring methods are mainly divided into two categories: one is contact measurement, such as GNSS and total stations, which requires the on-site deployment of sensors and cables, resulting in drawbacks such as complex installation, high cost, sparse monitoring points, and susceptibility to environmental influences; the other is traditional visual measurement methods, such as digital image correlation (DIC), which can achieve non-contact monitoring, but faces fundamental challenges in the field environment: natural textures are greatly affected by light and seasonal changes, their features are unstable, and they are difficult to work effectively in areas without texture or with repetitive textures, requiring a large amount of computation and making it difficult to achieve high-frequency real-time monitoring.

[0025] This application provides a solution in which: First, in response to a displacement monitoring command, an initial target image corresponding to a feature image on a specific target is acquired. Based on a preset image acquisition frequency, a camera continuously acquires feature images on the specific target to obtain multiple sets of first target images. The feature images are circular or reflective markers. The feature images are fabricated on Invar alloy, ceramic, or a rigid substrate covered with a protective film to ensure the long-term stability of their physical dimensions. The camera is fixed at a stable reference point. The specific target is fixed at a monitoring point on the structure under test. The specific target has at least three feature images arranged in a non-collinear and known precise physical coordinate relationship. Then, the initial sub-pixel image coordinates of multiple feature points corresponding to the initial target image are determined, and the first sub-pixel image coordinates of multiple feature points corresponding to the first target image are determined. Finally, displacement monitoring is performed based on the initial sub-pixel image coordinates and the first sub-pixel image coordinates.

[0026] Understandably, this application, based on a specific target, can ensure the long-term stability of its physical dimensions in complex field environments. This allows for the continuous and stable acquisition of sub-pixel coordinates of feature images using a camera, thereby enabling efficient and accurate deformation monitoring of infrastructure projects such as underground gas pipelines, high slopes, and foundation pits.

[0027] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device capable of performing the above functions, or a high-precision displacement monitoring device based on a specific target. The following description uses a high-precision displacement monitoring device based on a specific target as an example to illustrate this embodiment and the subsequent embodiments.

[0028] Based on this, the embodiments of this application provide a high-precision displacement monitoring method based on a specific target, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the high-precision displacement monitoring method based on a specific target according to this application.

[0029] In this embodiment, the high-precision displacement monitoring method based on a specific target includes steps S10 to S30: Step S10: In response to the displacement monitoring command, acquire the initial target image corresponding to the feature image on the specific target, and continuously acquire the feature image on the specific target using a camera based on a preset image acquisition frequency to obtain multiple sets of first target images. The feature image is a circular or reflective marker. The feature image is made on an Invar alloy, ceramic, or rigid substrate covered with a protective film to ensure the long-term stability of its physical dimensions. The camera is fixed at a stable reference point, and the specific target is fixed at the monitoring point of the structure to be measured. The specific target has at least three feature images arranged in a non-collinear and known precise physical coordinate relationship. It should be noted that the displacement monitoring command is the start command that triggers the system to begin monitoring. It can be issued manually by the user or automatically triggered according to preset conditions.

[0030] It should be noted that a specific target refers to a rigid target with known and precise physical coordinates, which serves as the physical benchmark for monitoring.

[0031] It should be noted that the feature image is a feature image on the target used for visual positioning, such as a circular or reflective marker.

[0032] It should be noted that the initial target image refers to the image taken at the start of monitoring. The first image of the target was taken to establish the initial baseline.

[0033] It should be noted that the preset image acquisition frequency refers to the image capture interval set by the system (e.g., 1Hz), which controls the time resolution of the monitoring.

[0034] It should be noted that the first target image refers to the image at a subsequent time. Target images captured continuously at a set acquisition frequency.

[0035] It should be noted that circular or reflective markers refer to the specific form of the feature image. Circular shapes facilitate sub-pixel positioning; reflective materials can enhance image contrast and adapt to low-light environments.

[0036] It should be noted that Invar alloy, ceramic, or rigid substrate with protective film refers to the material used to make the target substrate, which has an extremely low coefficient of thermal expansion and high mechanical stability, ensuring that the physical coordinates between feature points remain unchanged over a long period of time.

[0037] It should be noted that fixing the camera to a stable reference point means that the camera is installed at a reference position that will not shift (such as a riverbank monitoring pier) to ensure the stability of the observation geometry.

[0038] It should be noted that when a specific target is fixed to a monitoring point on the structure to be tested, the target is rigidly connected to the structure being tested (such as a bridge or slope) and deforms along with the structure.

[0039] It should be noted that at least three feature images arranged in a non-collinear and known precise physical coordinate relationship mean that non-collinearity guarantees a unique solution for the 3D pose; and known physical coordinates provide an absolute reference in the world coordinate system.

[0040] It should be noted that, in this embodiment, after receiving the displacement monitoring command, the system first collects the current time (considered as the initial time). The target image is used as the initial target image. Subsequently, the camera continuously captures images of the target at a preset acquisition frequency (e.g., once per second), generating a series of first target images. These images record the deformation process of the structure from the initial state to subsequent moments.

[0041] Understandably, this step establishes the monitoring timeline and a continuous image data stream. The specific target material and design ensure the long-term stability of the physical benchmark, eliminating target deformation caused by environmental changes (temperature, humidity). The camera is fixed to the benchmark point, ensuring long-term consistency of the imaging geometry and laying the foundation for subsequent high-precision displacement calculations. Circular or reflective markers help to accurately extract feature points in subsequent image processing, improving positioning accuracy and anti-interference capabilities.

[0042] Step S20: Determine the initial subpixel image coordinates of multiple feature points corresponding to the initial target image, and determine the first subpixel image coordinates of multiple feature points corresponding to the first target image; It should be noted that the initial subpixel image coordinates refer to the image coordinates of each feature point extracted in the initial target image using subpixel localization algorithms (such as ellipse fitting and gray-scale centroid method), with an accuracy of up to 0.01 pixels.

[0043] It should be noted that the first subpixel image coordinates refer to the subpixel image coordinates of the corresponding feature points in the first target image of each frame.

[0044] It should be noted that in this embodiment, the initial target image is corrected using the distortion coefficients obtained from camera intrinsic parameter calibration, and then the sub-pixel image coordinates of each feature point are extracted, denoted as ( , For each frame of the first target image, distortion correction is also performed, and the sub-pixel coordinates of the corresponding feature points are extracted. , During extraction, it is necessary to ensure the matching relationship of feature points in different images (usually achieved automatically through the geometric arrangement of the target image).

[0045] Understandably, improving the accuracy of feature point localization in an image to the sub-pixel level provides high-precision input data for subsequent geometric calculations. Feature point coordinates are the bridge connecting pixel space and physical space, and their extraction accuracy directly affects the final displacement measurement results. Distortion correction eliminates the influence of lens distortion, making the extracted coordinates closer to the ideal projection model.

[0046] Step S30: Displacement monitoring is performed based on the initial subpixel image coordinates and the first subpixel image coordinates.

[0047] It should be noted that displacement monitoring refers to calculating the three-dimensional translational displacement of the monitoring point relative to the initial moment. , , (and rotation angle) to achieve high-precision quantification of structural deformation.

[0048] It should be noted that, in this embodiment, based on the initial sub-pixel image coordinates and the first sub-pixel image coordinates obtained in step S20, combined with the pre-calibrated camera intrinsic parameter matrix and the known physical coordinates of the feature points, a PnP-type algorithm (such as EPnP) is used to calculate the three-dimensional pose of the target relative to the camera at each moment. The current pose is compared with the initial pose to calculate the three-dimensional translational displacement and rotation. Finally, the displacement data of the monitoring points is output, which can be further used for structural safety assessment.

[0049] Understandably, this step achieves a direct mapping from pixel coordinates to physical displacement, avoiding the accuracy loss caused by installation errors, lens distortion, and other factors in the traditional pixel equivalent method. It does not require the camera's optical axis to be perpendicular to the monitoring plane, nor does it require precise measurement of object distance, greatly reducing on-site installation difficulty. Three-dimensional displacement and rotation information can be directly obtained through pose calculation, meeting the monitoring needs of multi-dimensional deformations such as settlement, tilt, and torsion. Combined with AI feature tracking technology, it can maintain stable operation in complex environments such as changing lighting and partial occlusion, improving the system's long-term robustness. The entire process is automated, supporting unattended operation and remote monitoring, making it suitable for long-term structural health monitoring.

[0050] This application discloses a high-precision displacement monitoring method and related equipment based on a specific target, relating to the field of data processing technology. In contrast, existing automated monitoring methods are mainly divided into two categories: one is contact measurement, such as GNSS and total stations, which requires on-site deployment of sensors and cables, resulting in complex installation, high cost, sparse monitoring points, and susceptibility to environmental influences; the other is traditional visual measurement methods, such as digital image correlation (DIC), which, while enabling non-contact monitoring, faces fundamental challenges in field environments: natural textures are greatly affected by lighting and seasonal changes, their features are unstable, and they are difficult to work effectively in areas without texture or with repetitive textures, requiring large computational loads and making high-frequency real-time monitoring difficult. In this application, firstly, in response to a displacement monitoring command, an initial target image corresponding to the feature image on the specific target is acquired, and... Based on a preset image acquisition frequency, a camera continuously acquires feature images on the specific target to obtain multiple sets of first target images. The feature images are circular or reflective markers. The feature images are fabricated on Invar alloy, ceramic, or a rigid substrate covered with a protective film to ensure the long-term stability of their physical dimensions. The camera is fixed at a stable reference point, and the specific target is fixed at the monitoring point of the structure under test. The specific target has at least three feature images arranged in a non-collinear and known precise physical coordinate relationship. Then, the initial sub-pixel image coordinates of multiple feature points corresponding to the initial target image are determined, and the first sub-pixel image coordinates of multiple feature points corresponding to the first target image are determined. Finally, displacement monitoring is performed based on the initial sub-pixel image coordinates and the first sub-pixel image coordinates.

[0051] Understandably, this application, based on a specific target, can ensure the long-term stability of its physical dimensions in complex field environments. This allows for the continuous and stable acquisition of sub-pixel coordinates of feature images using a camera, thereby enabling efficient and accurate deformation monitoring of infrastructure projects such as underground gas pipelines, high slopes, and foundation pits.

[0052] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 2 The step of performing displacement monitoring based on the initial sub-pixel image coordinates and the first sub-pixel image coordinates further includes steps A10 to A30: Step A10: Obtain the physical world coordinates corresponding to each feature point; It should be noted that physical world coordinates refer to the precise coordinate values ​​of each feature image on a specific target in three-dimensional space. These coordinates are usually defined in the target's own local coordinate system (e.g., with the target center as the origin and the plane as the XY plane), and their values ​​are obtained through precision machining or high-precision measurement, possessing known and stable geometric relationships (e.g., the spacing and angles between feature points). In visual displacement monitoring, physical world coordinates serve as three-dimensional reference points, used to establish a correspondence with image coordinates.

[0053] It should be noted that, in this embodiment, before system deployment or during the initial calibration phase, the three-dimensional physical world coordinates of each feature point on the target are obtained using the target's design drawings or precision measurement methods. These coordinates are stored in the system for subsequent pose calculation.

[0054] Understandably, this step provides an absolute geometric reference in three-dimensional space, enabling the association of two-dimensional points in the image with points in three-dimensional space, thereby calculating the target's six-degree-of-freedom pose. The accuracy of the physical world coordinates directly determines the precision of the final displacement measurement. Therefore, using a highly stable substrate (such as Invar alloy) and precise manufacturing processes ensures that the coordinates remain unchanged over a long period, laying the foundation for high-precision monitoring.

[0055] Step A20: Based on the physical world coordinates, the initial sub-pixel image coordinates, and the first sub-pixel image coordinates, the pose of the specific target is calculated using a preset EPnP algorithm to obtain the initial pose and multiple sets of first poses corresponding to the specific target. It should be noted that the preset EPnP algorithm refers to Efficient Perspective-n-Point (EPnP), a highly efficient non-iterative PnP solution algorithm that can quickly solve the 3D pose of an object in the camera coordinate system from n (n≥3) 3D points and their corresponding 2D image points. This algorithm uses a weighted sum of 3D points to represent virtual control points, transforming the problem into solving for the coordinates of these control points in the camera coordinate system, offering advantages such as high computational speed and high accuracy.

[0056] It should be noted that pose refers to the position and orientation of an object in three-dimensional space. Position is represented by a translation vector, describing the coordinates of the origin of the target coordinate system in the camera coordinate system; orientation is represented by a rotation matrix, describing the orientation of the target coordinate system relative to the camera coordinate system. Together, they determine the six-degree-of-freedom spatial state of the target.

[0057] It should be noted that the initial pose refers to the target pose calculated using the EPnP algorithm at the start of monitoring, based on the sub-pixel coordinates of feature points extracted from the initial target image and their corresponding physical world coordinates. It serves as the benchmark for all subsequent displacement calculations.

[0058] It should be noted that the first pose refers to the current pose obtained again at a subsequent monitoring time by using the sub-pixel coordinates of feature points extracted from the first target image and the same physical world coordinates, and is denoted as . Multiple sets of first poses correspond to the structural states at different times.

[0059] It should be noted that in this embodiment, after obtaining the physical world coordinates of the feature points, the system inputs the sub-pixel image coordinates at the initial moment and the physical coordinates into the EPnP algorithm to calculate the initial pose. Subsequently, for each frame of the first target image, the EPnP algorithm is repeatedly executed, substituting the current sub-pixel image coordinates and physical world coordinates to obtain the current pose. The pose calculation results at all moments (including the initial pose) are stored for subsequent analysis.

[0060] Understandably, the EPnP algorithm directly establishes the correspondence between 2D image points and 3D physical points, avoiding the intermediate step of calculating pixel equivalents and then displacement in traditional methods. This eliminates the indirect influence of installation errors, lens distortion, and other factors on the measurement results. It does not require the camera's optical axis to be perpendicular to the monitoring plane, nor does it require precise measurement of the object distance, significantly reducing the difficulty and cost of on-site installation. The pose calculation results contain complete six-degree-of-freedom information of the target (3D translation and 3D rotation), providing a data foundation for subsequent multi-dimensional deformation monitoring. The efficiency of the EPnP algorithm ensures the feasibility of real-time monitoring and is suitable for high-frequency acquisition scenarios.

[0061] Step A30: Based on the initial pose and the multiple sets of first poses, perform displacement monitoring.

[0062] It should be noted that displacement monitoring refers to calculating the three-dimensional translational displacement and rotation angle changes (such as Euler angles or rotation matrix differences) of the monitoring point relative to the initial state by comparing the pose of the target at different times, thereby quantifying the structural deformation.

[0063] It should be noted that in this embodiment, the multiple sets of first poses obtained in step A20 are differentially calculated with the initial pose to obtain the translational displacement, which directly reflects the positional change of the monitoring point in three-dimensional space. By calculating the relative rotation matrix and decomposing it into rotation angles (such as roll, pitch, and yaw) around each coordinate axis, it is used to monitor the tilt or torsion of the structure. Finally, these displacements and rotations are output for engineers to assess the structural safety status.

[0064] Understandably, this step directly outputs three-dimensional displacement and rotation information, meeting the monitoring needs of various deformations such as bridge deflection, slope slippage, and building tilt, offering a more comprehensive approach than traditional two-dimensional measurements. Differential calculations eliminate fixed biases in the initial pose, retaining only the structural deformation itself, thus improving the intuitiveness and accuracy of the measurement. The entire calculation process is automated, requiring no manual intervention, and supports long-term, continuous, and remote structural health monitoring. Combined with the robustness of AI feature extraction, it can stably output displacement data even under environmental changes (lighting, shading), ensuring the system's reliability under complex outdoor conditions.

[0065] Specifically, the step of performing displacement monitoring based on the initial pose and the multiple sets of first poses further includes steps A31 to A32: Step A31: Calculate the difference between the first pose and the initial pose at different acquisition times to obtain the three-dimensional translational displacement and three-dimensional rotation at different acquisition times. It should be noted that the first pose corresponding to different acquisition times refers to the pose calculated for each frame of the first target image in step A20, corresponding to each acquisition time point.

[0066] It should be noted that the three-dimensional translation displacement refers to the change in position of the monitoring point in three-dimensional space from the initial moment to the current moment, which is obtained by directly subtracting the translation vectors.

[0067] It should be noted that the three-dimensional rotation is the rotational change of the target (i.e., the monitoring point) from its initial attitude to its current attitude, and can be represented in various ways, such as Euler angles (roll, pitch, yaw), rotation matrices, or quaternions. It is obtained by calculating and decomposing the relative rotation matrix.

[0068] It should be noted that in this embodiment, the system iterates through all monitoring times and performs the following operations for each time: Retrieve the first pose and initial pose at this moment from storage. Calculate the translational displacement to obtain the three-dimensional translational displacement. Calculate the rotational change and, if necessary, decompose it into rotation angles about the X, Y, and Z axes (e.g., using the Euler angle formula) to obtain the three-dimensional rotation. Store the calculation results along with the corresponding timestamp or output them in real time.

[0069] Understandably, by using pose difference analysis, the pure deformation of the structure itself can be directly extracted, eliminating the fixed deviations caused by the initial installation state, making the measurement results more intuitive and with clear physical meaning. Simultaneously obtaining three-dimensional translation and rotation information can comprehensively reflect the spatial motion state of the structure (such as settlement, horizontal displacement, tilt, and torsion), meeting the needs of complex deformation monitoring. It provides core data for subsequent safety assessments (such as threshold early warning and trend analysis). The calculation process is simple and fast, suitable for real-time monitoring systems.

[0070] Step A32: Based on the three-dimensional translational displacement and the three-dimensional rotational displacement, perform displacement monitoring.

[0071] It should be noted that, in this embodiment, the system organizes the three-dimensional displacement and rotation data output in step A31 at each moment in chronological order and performs one or more of the following operations: Data storage: Displacement, rotation and corresponding timestamps are stored in a database or file to form historical records.

[0072] Real-time display: The deformation trend is displayed on the monitoring interface in the form of curves, numerical values ​​or 3D animation.

[0073] Threshold judgment: The current displacement is compared with the preset safety threshold (such as deflection limit, tilt angle limit). If the limit is exceeded, an alarm signal (audio-visual, SMS, email, etc.) is automatically issued.

[0074] Data analysis: Perform statistical analysis and spectrum analysis on the cumulative displacement to extract deformation patterns.

[0075] Report generation: Regularly generate monitoring reports for engineers to review.

[0076] Understandably, the goal is to transform technical measurement results into engineering-usable information to truly achieve the monitoring objective. Real-time early warning functions can promptly detect structural anomalies and prevent safety accidents. Accumulated historical data helps analyze the long-term performance evolution of the structure and guide maintenance decisions. Automated data processing and display reduce the burden of manual analysis and improve monitoring efficiency. Combining three-dimensional displacement and rotation measurements allows for a more comprehensive assessment of structural health, avoiding the risk of missing information based on a single indicator.

[0077] Based on the first and second embodiments of this application, in the third embodiment of this application, the content that is the same as or similar to that in embodiments one and two above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 3 The step of determining the initial sub-pixel image coordinates of multiple feature points corresponding to the initial target image further includes steps B10 to B30: Step B10: Based on the two-dimensional calibration board, calibrate the camera to obtain the camera's intrinsic parameter matrix and lens distortion coefficients; It should be noted that a two-dimensional calibration plate is a planar plate with a high-precision known geometric image, commonly a checkerboard or an array of dots. The physical coordinates of its feature points (such as corner points or center points) in the world coordinate system are precisely known (usually with the calibration plate plane as the Z=0 plane). This is used to establish the correspondence between three-dimensional world points and two-dimensional image points, thereby solving for camera intrinsic parameters and distortion coefficients.

[0078] It should be noted that camera calibration refers to the process of calculating the camera's internal geometric parameters by taking images of a calibration board in different poses. These parameters describe the camera's optical imaging characteristics, including focal length, principal point coordinates, etc.

[0079] It should be noted that the intrinsic parameter matrix is ​​a 3×3 matrix that describes the ideal projection transformation from the camera coordinate system to the image coordinate system.

[0080] It should be noted that the lens distortion coefficient describes the imaging distortion caused by the actual lens manufacturing process, mainly including radial distortion (caused by the lens shape) and tangential distortion (caused by the lens not being parallel to the imaging plane).

[0081] It should be noted that, in this embodiment, before deployment in the laboratory or field, a two-dimensional calibration board is placed within the camera's field of view, and multiple images (typically 10-20) are captured from different angles and distances. For each image, the system automatically detects feature points on the calibration board (such as checkerboard corner points) to obtain their sub-pixel image coordinates. Using these image coordinates and their corresponding physical world coordinates (based on the calibration board's own coordinate system), the intrinsic parameter matrix of the camera and the lens distortion coefficients are obtained through optimization using algorithms such as the Zhang Zhengyou calibration method. These parameters are saved for use in all subsequent monitoring steps.

[0082] Understandably, the intrinsic parameter matrix and distortion coefficients are the foundation for all subsequent image correction and 3D calculations, and their accuracy directly affects the overall precision of the monitoring system. After a single calibration, the imaging model of the camera and lens combination is accurately modeled, eliminating the need for repeated calibration in subsequent uses (unless the lens or camera is adjusted). Calibration transforms the systematic error of lens distortion into a known, correctable parameter, laying the foundation for high-precision measurements. The calibration process is standardized, automated, simple to operate, and yields reliable results.

[0083] Step B20: Based on the intrinsic parameter matrix and the lens distortion coefficient, the initial target image is corrected to obtain the corrected initial target image; It should be noted that correction refers to using known lens distortion coefficients to remap the original image, eliminating the effects of radial and tangential distortion, so that the corrected image conforms to the ideal pinhole imaging model.

[0084] The corrected initial target image refers to the image obtained after distortion correction processing, in which the pixel positions have been rearranged so that straight lines in the real world remain straight lines in the corrected image, eliminating geometric distortions caused by lens distortion.

[0085] In this embodiment, the system reads the initial target image (original image) saved in step S10, and simultaneously calls the lens distortion coefficients and intrinsic parameter matrix obtained in step B10. For each pixel position in the original image, its new position in the corrected image is calculated according to the distortion model, and the corrected image pixel value is generated through an interpolation algorithm (such as bilinear interpolation). Finally, a corrected image with lens distortion eliminated is obtained, which is the corrected initial target image.

[0086] Understandably, eliminating the systematic offset of feature point positions caused by lens distortion makes the subsequently extracted image coordinates closer to those under the ideal projection model. In the corrected image, the geometric relationships (such as collinearity and equidistance) of the target feature points are restored, providing reliable input for accurate pose calculation. Without correction, lens distortion will directly cause deviations in the feature point image coordinates, leading to pose calculation errors and affecting the accuracy of displacement monitoring.

[0087] Step B30: Based on the corrected initial target image, determine the initial sub-pixel image coordinates of multiple feature points corresponding to the initial target image.

[0088] It should be noted that the corrected initial target image refers to the image obtained in step B20, which has eliminated the effects of lens distortion.

[0089] It should be noted that the initial subpixel image coordinates refer to the image coordinates of each feature point extracted using a subpixel localization algorithm in the corrected initial target image. These coordinates have an accuracy of 0.01 pixels or higher and are one of the input data for subsequent pose calculation.

[0090] In this embodiment, after obtaining the corrected initial target image, the system processes the image to extract the sub-pixel coordinates of each feature image (such as circular markers). The specific method varies depending on the type of feature image: For circular markers, edge detection, ellipse fitting, and other methods can be used to obtain the sub-pixel coordinates of the center position.

[0091] For reflective markers, their high contrast characteristics can be utilized to accurately locate them using the gray-scale centroid method or Gaussian surface fitting method.

[0092] During the extraction process, it is necessary to ensure that the recognition order of feature points corresponds one-to-one with their physical world coordinates (usually through automatic matching using the geometric arrangement of the target image). This ultimately yields a set of sub-pixel image coordinates, each corresponding one-to-one with the physical world coordinates of the feature points.

[0093] Understandably, this step provides the two-dimensional input points needed for mapping from image space to physical space, and its accuracy directly determines the accuracy of subsequent pose calculations. Subpixel localization technology improves the accuracy of feature point localization to the subpixel level, which is more precise than integer pixel localization and can capture minute structural displacements. Localization is performed on the distortion-corrected image, avoiding interference from lens distortion and further ensuring measurement accuracy. These initial coordinates serve as a reference and will be used together with coordinates extracted at subsequent time steps for displacement calculation; therefore, their accuracy is crucial.

[0094] Specifically, the step of determining the initial sub-pixel image coordinates of multiple feature points corresponding to the initial target image based on the corrected initial target image further includes steps B31 to B32: Step B31: Based on the preset target detection model, perform feature point extraction on the corrected initial target image to obtain multiple feature points corresponding to the corrected initial target image; It should be noted that the preset target detection model is an AI model based on deep learning (such as convolutional neural networks). After being trained on a large number of labeled images, it can automatically identify the location of specific targets in images (such as circular or reflective markers on a target). This model is capable of resisting changes in lighting, partial occlusion, and weather interference, and is used to robustly detect feature images in complex environments.

[0095] It should be noted that feature points are the corresponding locations of feature images (such as circular markers) on the target in the image, usually represented by bounding boxes, center points, or regions. In this step, feature points refer to the approximate area or rough location of the detected markers.

[0096] In this embodiment, the system inputs the corrected initial target image into a preset target detection model. This model, through forward inference, identifies all targets belonging to the target feature image in the image and outputs the bounding box or mask for each target, thereby determining the approximate location of each feature point in the image. These detection results (i.e., feature point regions) are recorded for use in the next step of precise localization.

[0097] Understandably, leveraging the powerful feature extraction capabilities of deep learning models enables stable detection of target features in complex environments such as varying lighting conditions, shadows, rain, fog, and partial occlusion, significantly improving the system's environmental adaptability. Compared to traditional image processing algorithms (such as thresholding and edge detection), deep learning models are more robust to noise and interference, reducing the probability of feature extraction failures and ensuring the reliability of long-term monitoring. The detection results provide an accurate initial region for subsequent sub-pixel localization, narrowing the search range and improving localization speed and accuracy.

[0098] It should also be noted that, before the step of performing feature point extraction on the corrected initial target image based on the preset target detection model, steps B311 to B313 are also included: Step B311: Obtain sample data, which corresponds to the first feature point extraction result; It should be noted that sample data refers to the image dataset used to train the target detection model. Each image contains feature images of a specific target and covers various environmental conditions that may be encountered in actual monitoring (such as different lighting, weather, occlusion, etc.). Sample data is the material for model learning.

[0099] It should be noted that the first feature point extraction result refers to the true location annotation of the feature points in the sample data, which is usually accurately labeled manually or semi-automatically and serves as the ground truth for model training. The annotation content may include the bounding box or center coordinates of the feature points.

[0100] In this embodiment, the system collects a large number of images containing specific targets, covering various lighting conditions (day / night, strong / weak light), weather scenes (rain, fog, snow), and partial occlusion. The feature images in each image are precisely labeled (e.g., marking the center position of circular markers), forming the "first feature point extraction result." This labeled data, together with the images, constitutes the training sample set.

[0101] Understandably, this step provides the model with a reliable reference standard, forming the basis of supervised learning. By covering samples in complex environments, the model learns the invariant features of feature points under different disturbances, improving robustness in subsequent practical applications. The accuracy of annotation directly affects the performance of the final model; therefore, high-precision annotation methods (such as sub-pixel level annotation) must be used.

[0102] Step B312: Process the sample data using the current target detection model to obtain the second feature point extraction result; It should be noted that the current object detection model refers to a deep learning model that has not yet converged during the training process, and its parameters are continuously updated and optimized in each iteration.

[0103] It should be noted that the second feature point extraction result refers to the feature point prediction result output by the current target detection model after inference on the input sample image, including the predicted position coordinates and confidence level.

[0104] In this embodiment, the sample image is input into the current target detection model, the model performs forward propagation calculation, and outputs the predicted feature point positions (i.e., the second feature point extraction result). This result will be compared with the ground truth (the first feature point extraction result) from step B311 in the next step.

[0105] Step B313: Determine whether the first feature point extraction result and the second feature point extraction result are consistent. If they are inconsistent, adjust the parameters of the current target detection model. Based on the current target detection model with adjusted parameters, return to the step of using the current target detection model to process the sample data and obtain the second feature point extraction result, until the first feature point extraction result and the second feature point extraction result are consistent, and obtain the preset target detection model.

[0106] It should be noted that consistency means that the difference between the second feature point extraction result output by the model and the first feature point extraction result manually labeled is less than a preset threshold. It is usually quantified by a loss function (such as mean square error). When the loss value converges or reaches a specified accuracy, it is considered consistent.

[0107] It should be noted that adjusting the parameters of the current object detection model means calculating the gradient through the backpropagation algorithm based on the difference (loss) between the predicted result and the true value, and using the optimizer to update the model weights to reduce the error.

[0108] In this embodiment, the loss value between the second feature point extraction result and the first feature point extraction result is calculated. If the loss value does not meet the preset stopping condition (e.g., the loss no longer decreases or falls below the accuracy requirement), the model parameters are adjusted through backpropagation, and then the process returns to step B312. The updated model is used to process the sample data again, and the iteration is repeated. When the loss value meets the requirement (the model prediction is consistent with the true value) or the maximum number of training rounds is reached, the training terminates, and the final preset target detection model is obtained.

[0109] Understandably, through iterative optimization, the model gradually approximates the true mapping, ultimately yielding a target detection model capable of accurately and robustly detecting target feature points. The convergence of the training process ensures the model's reliability in practical applications. A fully trained model can withstand environmental interference (lighting, weather, occlusion), guaranteeing the success rate of feature extraction in long-term monitoring. The automated training process reduces the workload of manual parameter tuning and allows for fine-tuning of the model for different scenarios, improving adaptability.

[0110] Step B32: Perform sub-pixel image coordinate extraction operation on the multiple feature points to obtain the initial sub-pixel image coordinates of the multiple feature points corresponding to the initial target image.

[0111] It should be noted that subpixel image coordinate extraction is an image processing technique that, based on a known approximate region of feature points, calculates the precise coordinates of the center point of the feature image using algorithms (such as ellipse fitting, gray-level centroid method, and Gaussian surface fitting), achieving an accuracy of 0.01 pixels or even higher. This operation utilizes the geometric or gray-level distribution characteristics of the feature image to improve the positioning accuracy to the subpixel level.

[0112] In this embodiment, within each feature point region obtained in step B31, the system executes a sub-pixel coordinate extraction algorithm. The coordinates output by the algorithm are the sub-pixel image coordinates of that feature point. The same operation is performed on all detected feature points to obtain a set of initial sub-pixel image coordinates, and a correspondence is established between these coordinates and the physical world coordinates of the feature points.

[0113] Understandably, improving feature point localization accuracy from pixel-level to sub-pixel-level enables the capture of minute displacements (such as 0.1 mm or even smaller), meeting the demands of high-precision engineering monitoring. Sub-pixel localization eliminates quantization errors caused by pixel discretization, providing more accurate input for subsequent EPNP calculations, thereby improving displacement measurement accuracy. Combined with robust detection in step B31, fully automated extraction of high-precision feature points in complex environments is achieved, balancing stability and accuracy. These initial coordinates serve as the benchmark for all subsequent displacement calculations, and their accuracy directly impacts the performance of the entire monitoring system.

[0114] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the high-precision displacement monitoring method based on a specific target. Any simple modifications based on this technical concept are within the protection scope of this application.

[0115] This application also provides a high-precision displacement monitoring device based on a specific target; please refer to [reference needed]. Figure 4 The high-precision displacement monitoring device based on a specific target includes: The acquisition module 10 is used to respond to the displacement monitoring command, acquire the initial target image corresponding to the feature image on the specific target, and continuously acquire the feature image on the specific target using a camera based on a preset image acquisition frequency to obtain multiple sets of first target images. The feature image is a circular or reflective marker point. The feature image is made on an Invar alloy, ceramic or rigid substrate covered with a protective film to ensure the long-term stability of its physical dimensions. The camera is fixed at a stable reference point, and the specific target is fixed on the monitoring point of the structure to be measured. The specific target has at least three feature images arranged in a non-collinear and known precise physical coordinate relationship. The determining module 20 is used to determine the initial subpixel image coordinates of multiple feature points corresponding to the initial target image, and to determine the first subpixel image coordinates of multiple feature points corresponding to the first target image. The displacement monitoring module 30 is used to perform displacement monitoring based on the initial subpixel image coordinates and the first subpixel image coordinates.

[0116] In one embodiment, the displacement monitoring module further includes: The first acquisition unit is used to acquire the physical world coordinates corresponding to each feature point; The first calculation unit is used to calculate the pose of the specific target based on the physical world coordinates, the initial sub-pixel image coordinates and the first sub-pixel image coordinates using a preset EPnP algorithm, so as to obtain the initial pose and multiple sets of first poses corresponding to the specific target. The first displacement monitoring unit is used to perform displacement monitoring based on the initial pose and the multiple sets of first poses.

[0117] In one embodiment, the displacement monitoring module further includes: The second calculation unit is used to calculate the difference between the first pose and the initial pose at different acquisition times, and to obtain the three-dimensional translational displacement and three-dimensional rotation at different acquisition times. The second displacement monitoring unit is used to perform displacement monitoring based on the three-dimensional translational displacement and the three-dimensional rotational displacement.

[0118] In one embodiment, the determining module further includes: The calibration unit is used to calibrate the camera based on a two-dimensional calibration board to obtain the camera's intrinsic parameter matrix and lens distortion coefficients. The correction unit is used to correct the initial target image based on the intrinsic parameter matrix and the lens distortion coefficient to obtain the corrected initial target image. The determining unit is used to determine the initial subpixel image coordinates of multiple feature points corresponding to the initial target image based on the corrected initial target image.

[0119] In one embodiment, the determining module further includes: The feature point extraction unit is used to perform feature point extraction on the corrected initial target image based on a preset target detection model, so as to obtain multiple feature points corresponding to the corrected initial target image. The subpixel image coordinate extraction unit is used to perform subpixel image coordinate extraction operation on the multiple feature points to obtain the initial subpixel image coordinates of the multiple feature points corresponding to the initial target image.

[0120] In one embodiment, the determining module further includes: The second acquisition unit is used to acquire sample data, which corresponds to the first feature point extraction result. The data processing unit is used to process the sample data using the current target detection model to obtain the second feature point extraction result; The iterative training unit is used to determine whether the first feature point extraction result and the second feature point extraction result are consistent. If they are inconsistent, the parameters of the current target detection model are adjusted. Based on the current target detection model with adjusted parameters, the step of using the current target detection model to process the sample data and obtain the second feature point extraction result is returned until the first feature point extraction result and the second feature point extraction result are consistent, and a preset target detection model is obtained.

[0121] The high-precision displacement monitoring device based on a specific target provided in this application, employing the high-precision displacement monitoring method based on a specific target in the above embodiments, can solve the technical problem of high-precision displacement monitoring based on a specific target. Compared with related technologies, the beneficial effects of the high-precision displacement monitoring device based on a specific target provided in this application are the same as those of the high-precision displacement monitoring method based on a specific target provided in the above embodiments, and other technical features in the high-precision displacement monitoring device based on a specific target are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0122] This application provides a high-precision displacement monitoring device based on a specific target. The high-precision displacement monitoring device based on a specific target includes: 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 to enable the at least one processor to execute the high-precision displacement monitoring method based on a specific target in the first embodiment described above.

[0123] The following is for reference. Figure 5 This document illustrates a structural schematic diagram of a high-precision displacement monitoring device based on a specific target, suitable for implementing embodiments of this application. The high-precision displacement monitoring device based on a specific target in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), and in-vehicle terminals (e.g., in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 5 The high-precision displacement monitoring device based on a specific target shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0124] like Figure 5As shown, a high-precision displacement monitoring device based on a specific target may include a processing unit 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the high-precision displacement monitoring device based on the specific target. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the high-precision displacement monitoring equipment based on a specific target to communicate wirelessly or wiredly with other devices to exchange data. Although the figure shows a high-precision displacement monitoring equipment based on a specific target with various systems, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0125] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0126] The high-precision displacement monitoring device based on a specific target provided in this application, employing the high-precision displacement monitoring method based on a specific target in the above embodiments, can solve the technical problem of high-precision displacement monitoring based on a specific target. Compared with related technologies, the beneficial effects of the high-precision displacement monitoring device based on a specific target provided in this application are the same as the beneficial effects of the high-precision displacement monitoring method based on a specific target provided in the above embodiments, and other technical features in this high-precision displacement monitoring device based on a specific target are the same as those disclosed in the method of the previous embodiment, and will not be repeated here.

[0127] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0128] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0129] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, which are used to execute the high-precision displacement monitoring method based on a specific target in the above embodiments.

[0130] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0131] The aforementioned computer-readable storage medium may be included in a high-precision displacement monitoring device based on a specific target; or it may exist independently and not be assembled into a high-precision displacement monitoring device based on a specific target.

[0132] The aforementioned computer-readable storage medium carries one or more programs that, when executed by a high-precision displacement monitoring device based on a specific target, cause the high-precision displacement monitoring device based on the specific target to: In response to a displacement monitoring command, an initial target image corresponding to a feature image on a specific target is acquired. Based on a preset image acquisition frequency, a camera continuously acquires feature images on the specific target to obtain multiple sets of first target images. The feature images are circular or reflective markers. The feature images are fabricated on Invar alloy, ceramic, or a rigid substrate covered with a protective film to ensure the long-term stability of their physical dimensions. The camera is fixed at a stable reference point, and the specific target is fixed at the monitoring point of the structure to be measured. The specific target has at least three feature images arranged in a non-collinear and known precise physical coordinate relationship. Determine the initial subpixel image coordinates of multiple feature points corresponding to the initial target image, and determine the first subpixel image coordinates of multiple feature points corresponding to the first target image; Displacement monitoring is performed based on the initial subpixel image coordinates and the first subpixel image coordinates.

[0133] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0134] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0135] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0136] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described high-precision displacement monitoring method based on a specific target, thereby solving the technical problem of high-precision displacement monitoring based on a specific target. Compared with related technologies, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the high-precision displacement monitoring method based on a specific target provided in the above embodiments, and will not be repeated here.

[0137] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the high-precision displacement monitoring method based on a specific target as described above.

[0138] The computer program product provided in this application can solve the technical problem of high-precision displacement monitoring based on a specific target. Compared with related technologies, the beneficial effects of the computer program product provided in this application are the same as those of the high-precision displacement monitoring method based on a specific target provided in the above embodiments, and will not be repeated here.

[0139] The above description is only a part of the embodiments of this application and does not limit the scope of protection of this application. All equivalent structural transformations made under the technical concept of this application and using the content of this application specification and drawings, or direct / indirect applications in other related technical fields, are included in the scope of protection of this application.

Claims

1. A high-precision displacement monitoring method based on a specific target, characterized in that, Applied to complex field environments, the high-precision displacement monitoring method based on a specific target includes: In response to a displacement monitoring command, an initial target image corresponding to a feature image on a specific target is acquired. Based on a preset image acquisition frequency, a camera continuously acquires feature images on the specific target to obtain multiple sets of first target images. The feature images are circular or reflective markers. The feature images are fabricated on Invar alloy, ceramic, or a rigid substrate covered with a protective film to ensure the long-term stability of their physical dimensions. The camera is fixed at a stable reference point, and the specific target is fixed at the monitoring point of the structure to be measured. The specific target has at least three feature images arranged in a non-collinear and known precise physical coordinate relationship. Determine the initial subpixel image coordinates of multiple feature points corresponding to the initial target image, and determine the first subpixel image coordinates of multiple feature points corresponding to the first target image; Displacement monitoring is performed based on the initial subpixel image coordinates and the first subpixel image coordinates.

2. The high-precision displacement monitoring method based on a specific target as described in claim 1, characterized in that, The step of performing displacement monitoring based on the initial sub-pixel image coordinates and the first sub-pixel image coordinates further includes: Obtain the physical world coordinates corresponding to each feature point; Based on the physical world coordinates, the initial subpixel image coordinates, and the first subpixel image coordinates, the pose of the specific target is calculated using a preset EPnP algorithm to obtain the initial pose and multiple sets of first poses corresponding to the specific target. Displacement monitoring is performed based on the initial pose and the multiple sets of first poses.

3. The high-precision displacement monitoring method based on a specific target as described in claim 2, characterized in that, The step of performing displacement monitoring based on the initial pose and the multiple sets of first poses further includes: Calculate the difference between the first pose and the initial pose at different acquisition times to obtain the three-dimensional translational displacement and three-dimensional rotation at different acquisition times. Displacement monitoring is performed based on the three-dimensional translational displacement and the three-dimensional rotational displacement.

4. The high-precision displacement monitoring method based on a specific target as described in claim 1, characterized in that, The step of determining the initial sub-pixel image coordinates of multiple feature points corresponding to the initial target image further includes: The camera is calibrated using a two-dimensional calibration board to obtain the camera's intrinsic parameter matrix and lens distortion coefficients. Based on the intrinsic parameter matrix and the lens distortion coefficient, the initial target image is corrected to obtain the corrected initial target image; Based on the corrected initial target image, the initial subpixel image coordinates of multiple feature points corresponding to the initial target image are determined.

5. The high-precision displacement monitoring method based on a specific target as described in claim 4, characterized in that, The step of determining the initial sub-pixel image coordinates of multiple feature points corresponding to the initial target image based on the corrected initial target image further includes: Based on the preset target detection model, feature point extraction is performed on the corrected initial target image to obtain multiple feature points corresponding to the corrected initial target image; Sub-pixel image coordinate extraction is performed on the multiple feature points to obtain the initial sub-pixel image coordinates of the multiple feature points corresponding to the initial target image.

6. The high-precision displacement monitoring method based on a specific target as described in claim 5, characterized in that, Before the step of performing feature point extraction on the corrected initial target image based on a preset target detection model, the method further includes: Acquire sample data, the sample data corresponding to the first feature point extraction result; The sample data is processed using the current object detection model to obtain the second feature point extraction result; Determine whether the first feature point extraction result and the second feature point extraction result are consistent. If they are inconsistent, adjust the parameters of the current target detection model. Based on the current target detection model with adjusted parameters, return to the step of using the current target detection model to process the sample data and obtain the second feature point extraction result. Continue until the first feature point extraction result and the second feature point extraction result are consistent, and obtain the preset target detection model.

7. A high-precision displacement monitoring device based on a specific target, characterized in that, The high-precision displacement monitoring device based on a specific target includes: The acquisition module is used to respond to displacement monitoring commands, acquire initial target images corresponding to feature images on a specific target, and continuously acquire feature images on the specific target using a camera based on a preset image acquisition frequency to obtain multiple sets of first target images. The feature images are circular or reflective markers. The feature images are fabricated on Invar alloy, ceramic, or a rigid substrate covered with a protective film to ensure the long-term stability of their physical dimensions. The camera is fixed at a stable reference point, and the specific target is fixed at the monitoring point of the structure to be measured. The specific target has at least three feature images arranged in a non-collinear and known precise physical coordinate relationship. The determining module is used to determine the initial subpixel image coordinates of multiple feature points corresponding to the initial target image, and to determine the first subpixel image coordinates of multiple feature points corresponding to the first target image. A displacement monitoring module is used to perform displacement monitoring based on the initial subpixel image coordinates and the first subpixel image coordinates.

8. A high-precision displacement monitoring device based on a specific target, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the high-precision displacement monitoring method based on a specific target as described in any one of claims 1 to 6.

9. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the high-precision displacement monitoring method based on a specific target as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the steps of the high-precision displacement monitoring method based on a specific target as described in any one of claims 1 to 6.

Images