Bridge expansion joint width automatic measurement method and system based on vehicle-mounted camera

Abstract

The application discloses a bridge expansion joint width automatic measurement method and system based on a vehicle-mounted camera, and belongs to the technical field of bridge health monitoring and intelligent detection. The method comprises the following steps: installing at least one camera on a vehicle, wherein the camera is installed vertically downward or obliquely downward; collecting a road surface image containing an expansion joint when the vehicle passes through the expansion joint; inputting the image into a trained deep learning segmentation model to obtain a pixel-level segmentation result of the expansion joint region; and selecting a corresponding imaging geometric model according to the installation posture of the camera to calculate the actual physical width of the expansion joint. The application adopts the above-mentioned bridge expansion joint width automatic measurement method and system based on a vehicle-mounted camera, can automatically complete the expansion joint width measurement in daily driving, does not need to close the road, has high detection efficiency and low cost, supports two installation postures of vertical and oblique, is convenient for large-scale deployment on various vehicles, and realizes the normal intelligent inspection of the bridge expansion joint.

Description

Technical Field

[0001] This invention relates to the field of bridge health monitoring and intelligent detection technology, and in particular to an automatic measurement method and system for the width of bridge expansion joints based on vehicle-mounted cameras. Background Technology

[0002] Bridge expansion joints are important components designed to accommodate deformations in bridge structures caused by factors such as temperature changes, concrete shrinkage and creep, and vehicle loads. Abnormal changes in the width of expansion joints (such as excessive opening, compression, or misalignment) directly reflect the stress state and safety condition of the bridge structure. Therefore, regular inspection of their width is an important aspect of bridge health monitoring.

[0003] Currently, the detection of bridge expansion joint width mainly relies on three methods: manual inspection, fixed sensors, and dedicated inspection vehicles. Manual inspection requires personnel to walk to the expansion joint and use tools such as calipers for measurement. This method is time-consuming, inefficient, and the measurement results are greatly affected by subjective factors. Furthermore, it often requires closing part of the road, disrupting normal traffic. While fixed sensors (such as displacement gauges and strain gauges) can achieve real-time monitoring, their installation and maintenance costs are high, and they require supporting facilities such as power supply and data transmission, making large-scale deployment difficult on long bridges or urban road networks. Dedicated inspection vehicles (such as bridge inspection vehicles) are expensive, complex to operate, and require professional drivers and operators, making high-frequency, low-cost routine inspections impossible.

[0004] Therefore, the existing technology lacks a method and system that can quickly and automatically complete the detection of bridge expansion joint width without interfering with traffic, without the need for special equipment, and under normal driving conditions. This makes it difficult to meet the urgent needs of bridge maintenance departments for large-scale, high-frequency, and low-cost intelligent inspection of expansion joint conditions. Summary of the Invention

[0005] The purpose of this invention is to provide an automatic measurement method and system for the width of bridge expansion joints based on vehicle-mounted cameras. This aims to solve the problems of low efficiency, high cost, and road closures caused by existing detection methods that rely on manual inspections, fixed sensors, or dedicated inspection vehicles. The invention achieves rapid, accurate, and non-contact automatic measurement of the width of expansion joints during daily driving.

[0006] To achieve the above objectives, this invention provides an automatic measurement method for bridge expansion joint width based on a vehicle-mounted camera, comprising the following steps: S1. Install at least one camera on the vehicle, which is mounted on the top, bottom, bumper, or a special bracket of the vehicle, and the camera is mounted vertically downward or tilted downward; when the vehicle passes over the bridge expansion joint, collect road surface images including the expansion joint. S2. Input the road surface image into the trained deep learning segmentation model to obtain the pixel-level segmentation results of the expansion joint area; S3. Based on the camera's installation orientation, select the corresponding imaging geometry model and calculate the actual physical width of the expansion joint from the pixel-level segmentation results.

[0007] Preferably, when the camera is installed vertically downward, the optical axis of the camera is perpendicular to the road surface; when the camera is installed at an angle downward, the angle between the optical axis of the camera and the normal to the road surface is 30° to 60°.

[0008] Preferably, when the camera is mounted vertically downwards, the imaging geometry model is as follows: ; in, This refers to the actual physical width of the expansion joint. This represents the number of pixels along the width of the expansion joint area. For pixel size, The height of the camera above the ground. The focal length of the lens; When the camera is installed at an angle downwards, the angle between the camera's optical axis and the road surface normal is... The imaging geometric model is as follows: When the camera is installed at an angle downwards, the angle between the camera's optical axis and the road surface normal is... The imaging geometric model is as follows: .

[0009] Preferably, when the camera is installed at an angle downwards, the road surface image can be first transformed by inverse perspective to obtain a top view, and then the actual physical width of the expansion joint can be calculated according to the imaging geometry model corresponding to the vertical downward installation posture. Inverse perspective transformation uses pre-calibrated camera intrinsic and extrinsic parameters, as well as road surface plane equations, to map image pixel coordinates to road surface world coordinates, resulting in a top-view image with consistent pixel equivalents; extrinsic parameters include camera installation height, pitch angle, roll angle, and yaw angle.

[0010] Preferably, in step S3, the imaging geometric model adopts the laser-assisted ruler method: a laser projector is installed on the same side as the camera to project a known physical distance onto the road surface. The laser pattern; the acquired image simultaneously contains both the expansion joint and the laser pattern; the pixel spacing of the laser pattern in the image. With known physical distance Determine the pixel equivalent of the current image: ; Then, based on the pixel width of the expansion joint area The actual physical width of the expansion joint is calculated using pixel equivalents. : .

[0011] Preferably, the laser projector projects parallel double lines or a fixed-distance dot matrix onto the road surface, and the known physical distance is the actual distance between the parallel double lines or the fixed-distance dot matrix.

[0012] Preferably, the deep learning segmentation model is a semantic segmentation network based on U-Net, DeepLab series, or Transformer architecture.

[0013] Preferably, before the vehicle passes through the bridge expansion joint, the system determines that the vehicle has reached the top of the bridge expansion joint and triggers image acquisition by using GPS positioning, RFID tag identification, accelerometer signal change detection, or image frame change detection.

[0014] The present invention also provides an automatic measurement system for the width of bridge expansion joints based on a vehicle-mounted camera, for performing the above-described method, including: The image acquisition module includes at least one camera installed on the vehicle for acquiring road surface images containing the expansion joint when the vehicle passes over a bridge expansion joint. The camera is mounted on the top, bottom, bumper, or a dedicated bracket of the vehicle. The image processing and recognition module has a built-in trained deep learning segmentation model, which is used to receive road surface images and obtain pixel-level segmentation results of the expansion joint area, as well as select the corresponding imaging geometry model according to the installation posture of the camera, and calculate the actual physical width of the expansion joint from the pixel-level segmentation results.

[0015] Preferred options also include: A laser projector used to project laser patterns with known physical spacing onto a road surface; The trigger module is used to determine that the vehicle has reached the top of the bridge expansion joint and trigger the shooting by detecting changes in GPS positioning, RFID tag identification, accelerometer signal changes or changes between image frames. A supplementary lighting device is used to provide synchronized lighting during shooting; The communication module is used to upload the measurement results to the cloud.

[0016] Therefore, the present invention employs the above-mentioned automatic measurement method and system for bridge expansion joint width based on vehicle-mounted cameras, which has the following beneficial effects: (1) High detection efficiency and no traffic disruption: The expansion joint width measurement can be completed automatically during the normal daily driving of vehicles, without the need for road closures or special inspection windows, which greatly improves the frequency and efficiency of inspection.

[0017] (2) High degree of automation and low cost: The entire process from image acquisition and expansion joint recognition to width calculation is automated and requires no manual intervention; it can be achieved with just a conventional vehicle-mounted camera and an embedded processing unit, and the equipment cost is far lower than that of a dedicated inspection vehicle.

[0018] (3) Flexible installation posture and large-scale deployment: It supports two installation methods, vertical downward or tilted downward, to adapt to different vehicle models and installation space requirements; it can be deployed on buses, sanitation vehicles, inspection vehicles and other vehicles that frequently cross bridges to form a normalized intelligent monitoring network covering the road network.

[0019] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0020] Figure 1 This is a flowchart of the automatic measurement method for bridge expansion joint width based on a vehicle-mounted camera according to the present invention; Figure 2 This is a schematic diagram of the camera being installed vertically downwards in the automatic measurement system for bridge expansion joint width based on a vehicle-mounted camera according to the present invention. Figure 3 This is a schematic diagram of the camera being installed at an angle downwards in the automatic measurement system for the width of bridge expansion joints based on a vehicle-mounted camera, as described in this invention. Detailed Implementation

[0021] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0022] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0023] Example 1 like Figure 1 As shown, this embodiment provides an automatic measurement method for the width of bridge expansion joints based on a vehicle-mounted camera. This embodiment uses common highway bridge expansion joints as the detection object; however, it should be understood that the method of this invention is also applicable to other types of bridge expansion joints.

[0024] Overall process overview The overall process of this embodiment is as follows: First, at least one camera is installed on the vehicle, and the installation posture is selected as vertically downward or tilted downward according to the vehicle structure and installation space; when the vehicle passes through the bridge expansion joint, the road surface image containing the expansion joint is acquired; then, the road surface image is input into a trained deep learning segmentation model to obtain the pixel-level segmentation result of the expansion joint area; finally, according to the installation posture of the camera, the corresponding imaging geometry model is selected, and the actual physical width of the expansion joint is calculated from the pixel-level segmentation result.

[0025] The specific implementation methods for each of the above steps will be described in detail below.

[0026] Step S1: Image Acquisition At least one camera is installed on the vehicle, mounted on the top, bottom, bumper, or a dedicated bracket, with the camera positioned vertically downwards or tilted downwards. When the camera is vertically downwards, its optical axis is perpendicular to the road surface, and the lens plane is parallel to the road surface. In this case, image distortion is minimal, and the width can be directly calculated using a pinhole camera model. When the camera is tilted downwards, the angle between the camera's optical axis and the road surface normal is... , The angle range is 30° to 60°, which allows for easy installation of the camera in locations other than directly below the vehicle, such as on the bumper. The camera can be fixed to the bottom of the vehicle, the bumper, or a dedicated bracket.

[0027] When a vehicle passes over a bridge expansion joint, images of the road surface including the expansion joint are captured. Acquisition methods include single-frame or multi-frame capture, or extracting keyframes from a video stream. To ensure the expansion joint is fully visible within the field of view, when vertically installed, the image is captured when the vehicle passes directly above the expansion joint; when installed at an angle, the image is captured earlier or later.

[0028] Optionally, a supplementary lighting device can be installed on the same side as the camera to provide synchronized strobe lighting during shooting, ensuring image brightness. A filter can also be optionally installed to reduce ambient light interference.

[0029] Triggering data collection: Before a vehicle passes through a bridge expansion joint, a trigger module senses the vehicle's approach and triggers image acquisition. The trigger module can be implemented in several ways: using GPS positioning to determine if the vehicle has entered the bridge area; using RFID tags to identify pre-set electronic tags on the bridge surface; using accelerometers to detect changes in characteristic vibration signals as the vehicle passes the expansion joint; or using image frame change detection to automatically trigger image capture when the expansion joint's characteristic texture appears in the image.

[0030] Step S2: Automatic Recognition of Expansion Joints The road surface image acquired in step S1 is input into the trained deep learning segmentation model to obtain pixel-level segmentation results for the expansion joint area.

[0031] The deep learning segmentation model is a semantic segmentation network that classifies each pixel in an image, with categories including at least background and expansion joints. Specifically, it can employ U-Net, the DeepLab series (such as DeepLabV3+), or semantic segmentation networks based on the Transformer architecture (such as SegFormer). The model's training data includes bridge expansion joint images under various conditions, such as bright lighting, dim lighting, flooded roads, and stained roads, ensuring stable recognition under different environments.

[0032] After obtaining the pixel-level segmentation results, further morphological post-processing can be performed: morphological closing operations are performed on the segmentation mask to fill small holes; connected regions are extracted, and noise regions with too small an area are filtered out, finally obtaining the accurate contour of the expansion joint region.

[0033] Step S3: Width Calculation Based on the camera's installation orientation, the corresponding imaging geometry model is selected, and the actual physical width of the expansion joint is calculated from the pixel-level segmentation results. The calculation methods for the two installation orientations and the laser-assisted method are explained below.

[0034] (a) Vertical downward installation posture (e.g.) Figure 2 ) When the camera is installed vertically downwards, its optical axis is perpendicular to the road surface, and the imaging geometry model uses a pinhole imaging model. The formula for calculating the actual physical width W of the expansion joint is: ; in, This refers to the actual physical width of the expansion joint, in millimeters. This represents the number of pixels along the width direction of the expansion joint area in the image. The pixel size is expressed in millimeters per pixel. The vertical distance from the optical center of the camera to the road surface is measured in millimeters and can be obtained in real time through timed calibration or an ultrasonic sensor. This refers to the lens focal length, expressed in millimeters.

[0035] In practice, the number of pixels The method for obtaining the pixel width is as follows: On the segmented expansion joint mask, calculate the pixel width of multiple horizontal cross-sections along the longitudinal direction of the expansion joint, and take the average pixel width of each cross-section as the mean. This method of averaging effectively reduces random errors caused by single-point measurements and improves the robustness of the measurement results.

[0036] (ii) Inclined downward installation posture (e.g.) Figure 3 ) When the camera is mounted at an angle downwards, the angle between the camera's optical axis and the road surface normal is α. Since the image plane is not parallel to the road surface, the scale factor differs at different locations in the image, resulting in perspective distortion. There are two calculation methods for this.

[0037] Method 1 uses an approximate analytical calculation. This applies when the expansion joint is located near the image center and the tilt angle is... When the value is not too large, the relationship between the actual width and the pixel width can be approximated as follows: ; The meanings of each parameter are the same as those in the vertical installation formula. It is the angle between the camera's optical axis and the road surface normal.

[0038] Method two employs the inverse perspective transformation (IPM) method. First, an inverse perspective transformation is performed on the road image to correct the tilted image into a top-down view. The inverse perspective transformation uses pre-calibrated camera intrinsic parameters (such as focal length, pixel size, and image principal point), extrinsic parameters, and the road plane equation (Z=0) to map the image pixel coordinates to the road surface world coordinates, resulting in a top-down image with consistent pixel equivalents. Extrinsic parameters include the camera installation height. Pitch angle Roll angle and yaw angle. After transformation, the pixel width of the expansion joint is measured in the top view image, and then the actual physical width of the expansion joint is calculated according to the imaging geometry model corresponding to the vertically downward installation posture.

[0039] The core idea of ​​inverse perspective transformation is to assume the road surface is an ideal plane and, using the camera's intrinsic and extrinsic parameters obtained from the calibration, establish a mathematical mapping relationship from the original image pixel coordinate system to the road surface world coordinate system. This mapping relationship is usually described by a 3x3 homography matrix.

[0040] In practice, the following sub-steps are included: Construct the mapping: Calculate the homography matrix based on the calibration parameters.

[0041] Coordinate transformation: Traverse each pixel in the original image and use the homography matrix to transform its coordinates. Physical coordinates transformed into the road surface world coordinate system For non-integer coordinates, methods such as bilinear interpolation are used to determine their pixel values.

[0042] Generate a top-down view: Rearrange all transformed pixels according to the set physical resolution (e.g., each pixel represents 1 millimeter) to generate a new, perspective-free top-down image of the road surface. In this top-down image, every point on the road surface has a consistent physical scale.

[0043] Finally, the original expansion joint region segmented by the deep learning model is also transformed into the top view image using the same homography matrix transformation, and its pixel count in the width direction is directly measured. Then multiply by the preset physical resolution of the top-view image. (mm / pixel) can be used to obtain the actual physical width of the expansion joint. .

[0044] (III) Laser-assisted scale method As an enhancement to the aforementioned purely visual method, the imaging geometry model in step S3 can also employ a laser-assisted scale method. This method is applicable to both vertically downward and tilted downward installation orientations, and does not require prior knowledge of the camera's installation height and angle parameters.

[0045] Simultaneously with the installation of the camera, a laser projector is installed on the same side as the camera. The laser projector projects a known physical distance onto the road surface. The laser pattern can be a parallel double line or a fixed-distance dot matrix. The known physical spacing is the actual spacing of the parallel double line or the fixed-distance dot matrix, for example, 10 centimeters.

[0046] When a vehicle passes through an expansion joint, the captured road surface image simultaneously contains both the expansion joint and a laser pattern. During image processing, the laser pattern is identified, and its pixel spacing within the image is measured. This determines the pixel equivalent of the current image. : ; Its unit is millimeters per pixel, which represents the actual physical size of each pixel in the image under the current shooting conditions.

[0047] Then, based on the pixel width of the expansion joint area With pixel equivalent Calculate the actual physical width W of the expansion joint: ; The advantages of the laser-assisted ruler method are: the pixel equivalent of the current image is calibrated in real time by the pixel spacing of the laser pattern in the image, which can automatically adapt to changes in camera height and angle caused by vehicle bumps and load changes, without needing to know the camera installation height and focal length, and is suitable for both vertical and tilted installation postures, and has higher accuracy.

[0048] This embodiment also provides an automatic bridge expansion joint width measurement system based on a vehicle-mounted camera for performing the above method. The system includes: The image acquisition module includes at least one camera mounted on the vehicle for acquiring road surface images containing the expansion joint when the vehicle passes over a bridge expansion joint. The camera is mounted on the top, bottom, bumper, or a dedicated bracket of the vehicle. The camera is mounted vertically downward or tilted downward.

[0049] The image processing and recognition module incorporates a trained deep learning segmentation model. It receives road surface images and obtains pixel-level segmentation results for the expansion joint area. Based on the camera's installation orientation, it selects the corresponding imaging geometry model and calculates the actual physical width of the expansion joint from the pixel-level segmentation results. This module is deployed in an embedded processor or industrial computer, allowing all calculations to be performed at the edge.

[0050] Optionally, the system also includes the following modules: A laser projector, installed on the same side as the camera, projects laser patterns with known physical spacing onto the road surface. The image processing and recognition module determines the pixel equivalent of the current image by recognizing the laser pattern and calculates the actual physical width of the expansion joint accordingly.

[0051] The trigger module is used to determine that the vehicle has reached the top of the bridge expansion joint and trigger the image acquisition module to take pictures by detecting changes in GPS positioning, RFID tag identification, accelerometer signal changes or changes between image frames.

[0052] The supplementary lighting device is used to provide synchronous strobe lighting during image acquisition, ensuring that clear road surface images can still be obtained in low-light environments such as at night or in tunnels.

[0053] The communication module is used to upload the measured actual physical width of the expansion joint, along with the corresponding timestamp and GPS location information, to a cloud server for remote viewing and analysis by bridge maintenance departments.

[0054] Experimental results verification To verify the effectiveness of the method of this invention, tests were conducted on an actual highway bridge. The test bridge was a three-span continuous beam bridge with three expansion joints on the bridge deck, with designed widths of 20mm, 30mm, and 40mm, respectively. The test vehicle was a regular SUV, with the camera mounted on the front bumper at a downward tilt angle. The camera is approximately 45° off the ground. Approximately 400mm, lens focal length 8mm, pixel size The resolution is 0.0055 mm / pixel. The laser projector projects two lines of light onto the road surface at a distance of 0.0055 mm. It is a parallel laser line of 100mm.

[0055] Thirty sets of data were collected for each expansion joint under three vehicle speeds (20km / h, 40km / h, 60km / h) and three lighting conditions (sunny noon, cloudy, and night).

[0056] The measurement results obtained by the method of this invention were compared with those obtained by manual vernier caliper measurement. Under conditions of 40 km / h vehicle speed and clear weather, the average measurement results for the three expansion joints were 19.86 mm (design value 20 mm), 30.25 mm (design value 30 mm), and 39.71 mm (design value 40 mm), respectively. The measurement repeatability was good under the three lighting conditions, with the standard deviation of each measurement less than 0.5 mm. The measurement results under the three vehicle speed conditions showed good consistency, with the maximum deviation between different vehicle speeds less than 1.0 mm.

[0057] The above experimental results show that the method of the present invention can quickly and accurately measure the width of expansion joints automatically without closing the road or contacting the bridge structure. The measurement accuracy meets the requirements of engineering inspection and has good robustness to changes in vehicle speed and illumination.

[0058] Therefore, the present invention adopts the above-mentioned automatic measurement method and system for bridge expansion joint width based on vehicle-mounted cameras, which combines deep learning image segmentation with imaging geometric models. It can automatically select the corresponding calculation method according to the camera installation posture, and can complete the rapid and accurate measurement of expansion joint width during daily driving without road closure. It has the advantages of high detection efficiency, high degree of automation, low cost, flexible installation posture, and large-scale deployment, providing an effective technical means for the routine intelligent inspection of bridge expansion joints.

[0059] Therefore, the present invention employs the aforementioned automatic measurement method and system for bridge expansion joint width based on vehicle-mounted cameras, which can automatically complete the measurement of expansion joint width while vehicles are in normal operation, without the need for road closures, resulting in high detection efficiency. The system is low-cost, flexible in installation, and can be deployed on vehicles such as buses, forming a routine intelligent inspection network and providing reliable data support for bridge maintenance.

[0060] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. An automatic measurement method for bridge expansion joint width based on a vehicle-mounted camera, characterized in that, Includes the following steps: S1. Install at least one camera on the vehicle, which is mounted on the top, bottom, bumper, or a special bracket of the vehicle, and the camera is mounted vertically downward or tilted downward; when the vehicle passes over the bridge expansion joint, collect road surface images including the expansion joint. S2. Input the road surface image into the trained deep learning segmentation model to obtain the pixel-level segmentation results of the expansion joint area; S3. Based on the camera's installation orientation, select the corresponding imaging geometry model and calculate the actual physical width of the expansion joint from the pixel-level segmentation results.

2. The automatic measurement method for bridge expansion joint width based on a vehicle-mounted camera according to claim 1, characterized in that, When the camera is installed vertically downwards, its optical axis is perpendicular to the road surface; when the camera is installed at an angle downwards, the angle between the camera's optical axis and the road surface normal is 30° to 60°.

3. The automatic measurement method for bridge expansion joint width based on a vehicle-mounted camera according to claim 2, characterized in that, When the camera is mounted vertically downwards, the imaging geometry model is: ； in, This refers to the actual physical width of the expansion joint. This represents the number of pixels along the width of the expansion joint area. For pixel size, The height of the camera above the ground. The focal length of the lens; When the camera is installed at an angle downwards, the angle between the camera's optical axis and the road surface normal is... The imaging geometric model is as follows: 。 4. The automatic measurement method for bridge expansion joint width based on a vehicle-mounted camera according to claim 3, characterized in that, When the camera is installed at an angle downwards, the road surface image can be transformed by inverse perspective to obtain a top view, and then the actual physical width of the expansion joint can be calculated according to the imaging geometry model corresponding to the vertical downward installation posture. Inverse perspective transformation uses pre-calibrated camera intrinsic and extrinsic parameters, as well as road surface plane equations, to map image pixel coordinates to road surface world coordinates, resulting in a top-view image with consistent pixel equivalents; extrinsic parameters include camera installation height, pitch angle, roll angle, and yaw angle.

5. The automatic measurement method for bridge expansion joint width based on a vehicle-mounted camera according to claim 1, characterized in that, In step S3, the imaging geometry model adopts the laser-assisted ruler method: a laser projector is installed on the same side as the camera to project a known physical distance onto the road surface. The laser pattern; the acquired image simultaneously contains the expansion joint and the laser pattern; the pixel spacing of the laser pattern in the image. With known physical distance Determine the pixel equivalent of the current image: ； Then, based on the pixel width of the expansion joint area The actual physical width of the expansion joint is calculated using pixel equivalents. : 。 6. The automatic measurement method for bridge expansion joint width based on a vehicle-mounted camera according to claim 5, characterized in that, A laser projector projects parallel double lines or a fixed-distance dot matrix onto the road surface. The physical distance is known to be the actual distance between the parallel double lines or the fixed-distance dot matrix.

7. The automatic measurement method for bridge expansion joint width based on a vehicle-mounted camera according to claim 1, characterized in that, Deep learning segmentation models are semantic segmentation networks based on the U-Net, DeepLab series, or Transformer architecture.

8. The automatic measurement method for bridge expansion joint width based on a vehicle-mounted camera according to claim 1, characterized in that, Before a vehicle passes over a bridge expansion joint, GPS positioning, RFID tag identification, accelerometer signal changes, or image frame changes are used to determine that the vehicle has reached above the bridge expansion joint and trigger image acquisition.

9. An automatic bridge expansion joint width measurement system based on a vehicle-mounted camera, used to perform the method described in any one of claims 1-8, characterized in that, include: The image acquisition module includes at least one camera installed on the vehicle for acquiring road surface images containing the expansion joint when the vehicle passes over a bridge expansion joint. The camera is mounted on the top, bottom, bumper, or a dedicated bracket of the vehicle. The image processing and recognition module has a built-in trained deep learning segmentation model, which is used to receive road surface images and obtain pixel-level segmentation results of the expansion joint area, as well as select the corresponding imaging geometry model according to the installation posture of the camera, and calculate the actual physical width of the expansion joint from the pixel-level segmentation results.

10. The automatic bridge expansion joint width measurement system based on a vehicle-mounted camera according to claim 9, characterized in that, Also includes: A laser projector used to project laser patterns with known physical spacing onto a road surface; The trigger module is used to determine that the vehicle has reached the top of the bridge expansion joint and trigger the shooting by detecting changes in GPS positioning, RFID tag identification, accelerometer signal changes or changes between image frames. A supplementary lighting device is used to provide synchronized lighting during shooting; The communication module is used to upload the measurement results to the cloud.

Images