Pantograph thickness measuring method and device based on three-dimensional space calculation and visual reduction

Abstract

The invention discloses a pantograph thickness measuring method based on three-dimensional space calculation and visual reduction. The method comprises the following steps: step 1, deploying a device; 2, calibrating parameters; step 3, image acquisition; 4, solving a phase; 5, coordinate transformation; 6, generating a point cloud; step 7, point cloud splicing; 8, analyzing the information; and step 9, outputting and displaying. In addition, the invention also provides a pantograph thickness measuring device based on three-dimensional space calculation and visual reduction. Complete three-dimensional point clouds of the upper surface, the lower surface, the side face and the edge of the pantograph carbon slide plate are obtained through the multiple sets of area array structured light three-dimensional cameras and the phase encoding stripe structured light image collection mode, detection of local defects such as notches and peeling is finally achieved, and compared with single-line laser, the device has the higher environment light interference resisting capacity, and the detection precision is high. The measurement precision and reliability can be greatly improved; the method is high in practicability and has high popularization significance.

Description

Technical Field

[0001] This invention relates to the field of train track technology, and in particular to a method and apparatus for measuring pantograph thickness using three-dimensional spatial calculation and visual reconstruction. Background Technology

[0002] The carbon contact plate of the pantograph is a critical load-bearing and friction component of electric locomotives. Its thickness directly affects the conductivity and lifespan of the pantograph. Therefore, achieving high-precision, non-contact thickness detection is of great significance for train operation safety and maintenance efficiency. Current methods for pantograph carbon contact plate thickness detection mainly rely on manual measurement, two-dimensional image methods, and high-precision three-dimensional scanners. However, with the widespread use of high-speed trains and urban rail vehicles, traditional manual measurement methods and two-dimensional measurement methods are no longer sufficient to meet the demands for rapid, accurate, and automated on-site measurement.

[0003] Traditional manual measurement methods require static testing under conditions of power outage and train shutdown, which is not only inefficient and time-consuming, but also often fails to fully cover the thickness of the pantograph contactor, resulting in insufficient overall accuracy. Furthermore, this method involves working at heights and carries the risk of live electrical equipment, posing certain safety hazards. Two-dimensional imaging methods typically use industrial cameras to image the carbon contactor contactor. However, because they can only acquire two-dimensional planar information and lack depth information, the results are easily affected by the imaging angle and external lighting conditions (such as rain or strong light), leading to limited accuracy and poor stability.

[0004] High-precision 3D scanners typically employ a multi-set line laser structured light detection method. Existing patents, such as Chinese patents "CN202011567755.1 A Train Pantograph Integrated Geometric Parameter Online Detection Device" and "CN202111227612.0 A Pantograph Identification and Detection Device and Method Based on 3D Line Scanning," disclose a technical solution that involves setting multiple sets of structured light sensors on both sides of the pantograph detection area. Each set of structured light sensors consists of a camera and a line laser, used to acquire cross-sectional point cloud data of the upper and lower surfaces of the pantograph. This data is then stitched together to form a 3D model of the pantograph, thereby enabling the measurement of wear and offset.

[0005] Although the multi-group line laser structured light scheme can achieve three-dimensional point cloud acquisition and measurement of wear and offset on the upper and lower surfaces of the pantograph, this method has the following shortcomings: First, it only covers the upper and lower surfaces of the pantograph, and it is difficult to completely acquire the side and edge areas, making it impossible to fully detect local defects such as gaps or peeling that may exist in the pantograph; Second, line lasers are relatively sensitive to ambient light interference, and strong light, reflection, or rain and snow may affect the acquisition accuracy of laser stripes, thereby affecting the reliability of the measurement results. Summary of the Invention

[0006] Therefore, it is necessary to provide a method for measuring pantograph thickness using three-dimensional spatial calculation and visual reconstruction, addressing the shortcomings of existing technologies.

[0007] A method for measuring pantograph thickness using three-dimensional spatial calculation and visual reconstruction, comprising the following steps:

[0008] Step 1: Deploy the device. Install and arrange multiple sets of area array structured light 3D cameras on the train inspection shed. Cover the carbon slide plate with top, bottom and side views respectively. Set triggers in the inspection area to control the multiple sets of area array structured light 3D cameras to start automatically or standby. The area array structured light 3D camera includes a projection module and a camera module.

[0009] Step 2: Parameter calibration. Use a multi-dot array calibration board to accurately calibrate multiple sets of area array structured light 3D cameras, and obtain the intrinsic parameter matrix K of the camera module and the extrinsic parameter matrix [R|t] between the projection module and the camera module respectively.

[0010] Step 3: Image acquisition. When the trigger is triggered, multiple sets of area array structured light cameras are jointly started according to the preset scheduling strategy. The phase-encoded stripe projection method is adopted. The projection module of the area array structured light camera projects a set of sinusoidal stripe patterns with different phases onto the carbon slide surface, and the camera module captures the image sequence of these stripe patterns projected on the carbon slide surface.

[0011] Step 4: Phase solution. The encapsulated phase is obtained by solving a system of trigonometric equations constructed from multiple images. Then, the continuous phase distribution is recovered using a phase expansion algorithm. Specifically, the model of the trigonometric function equation system is as follows: in For the m-th stripe pattern in pixels grayscale value, Background brightness, In order to adjust the system, This is the initial phase of the pixel. The phase shift of the m-th image;

[0012] Step 5: Coordinate transformation, converting the phase value of each pixel. The corresponding imaging ray establishes a geometric relationship with the projection plane. Depth P is calculated based on the structured light triangulation model, and then the corresponding 3D point is obtained through back projection. Specifically, the back projection algorithm is as follows: in Here, P represents the 3D coordinates in the camera coordinate system, and P represents the projection depth. These are pixel coordinates;

[0013] Step 6: Point cloud generation. The three-dimensional reconstruction of all pixels is completed on multiple sets of structured light 3D cameras to obtain multiple sets of two-dimensional high-density point cloud maps of the carbon skateboard.

[0014] Step 7: Point cloud stitching to obtain the rotation matrix of each group of structured light 3D cameras. Translation vector Point clouds in the coordinate system of an arbitrary array structured light 3D camera The specific conversion algorithm for transforming to the world coordinate system is as follows: in The transformed global 3D coordinates This refers to the rotational extrinsic parameter between the projection module and the camera module. This refers to the translation extrinsic parameter between the projection module and the camera module.

[0015] Step 8: Analyze the information and extract the remaining thickness based on the geometric profile analysis method. That is, slice the point cloud model at equal intervals along the length direction (X-axis) of the carbon skateboard and divide the entire skateboard into several local profiles. For each profile area, extract the Y-axis coordinate value (i.e., height information) of all points in it, and combine noise filtering and statistical analysis to obtain the minimum Y value of the profile as its local remaining thickness index.

[0016] Step 9: Output and Display. Output all thickness and defect analysis results as structured data files, and generate visualization results such as thickness profile diagrams and notch distribution diagrams based on the structured data files, and output them to the web for display.

[0017] Furthermore, in step 1, the triggers are a magnetic steel sensor and an infrared beam sensor. When a train wheel approaches the magnetic steel sensor, it generates an electrical signal due to magnetic field disturbance, which serves as a train approach signal. Subsequently, the structured light 3D camera is controlled to start image acquisition only when the magnetic steel sensor and the infrared beam sensor are triggered synchronously.

[0018] Furthermore, in step 7, after converting and stitching the two-dimensional point cloud images into a three-dimensional point cloud image, the overlapping areas are further processed by voxel grid downsampling, weighted overlap fusion, or iterative nearest point (ICP) fine-tuning methods to improve the consistency and integrity of the point cloud.

[0019] Furthermore, in step 8, in addition to extracting the remaining thickness based on the geometric profile analysis method, the left and right edge regions of the point cloud model are scanned, and the possible peeling, damage or gap phenomena are automatically detected by combining the geometric features of the point cloud such as abrupt changes in normal, curvature anomalies and height drops.

[0020] Furthermore, in steps 2 and 7, calibration can be performed multiple times using a single camera rotation or line laser scanning, followed by registration using feature matching or Simultaneous Localization and Mapping (SLAM) methods to achieve point cloud stitching.

[0021] Furthermore, in step 3, multiple sets of line lasers plus camera scanning, or a time-of-flight (ToF) camera, or Gray code or binary encoding can be used to replace the phase-coded stripe projection method to complete image acquisition.

[0022] In addition, the present invention also provides a pantograph thickness measurement device for three-dimensional spatial calculation and visual reconstruction.

[0023] A pantograph thickness measurement device based on three-dimensional spatial calculation and visual reconstruction, used to perform the aforementioned pantograph thickness measurement method based on three-dimensional spatial calculation and visual reconstruction, includes multiple sets of area array structured light three-dimensional cameras, triggers, and a detection and analysis system. The multiple sets of area array structured light three-dimensional cameras are respectively installed above the detection shed, on the left and right sides of the detection shed, and between the carbon slide plate and the top of the train. The upper area array structured light three-dimensional camera is located above the central axis of the carbon slide plate, the left and right side area array structured light three-dimensional cameras are located on the left and right sides of the carbon slide plate, and the lower area array structured light three-dimensional camera is located in the area below the carbon slide plate. The triggers are installed on the train track and are signal-connected to the multiple sets of area array structured light three-dimensional cameras and the detection and analysis system. The detection and analysis system is signal-connected to the multiple sets of area array structured light three-dimensional cameras. Each area array structured light three-dimensional camera includes a projection module and a camera module.

[0024] Furthermore, the area array structured light 3D cameras on the left and right sides are symmetrically designed, and each side has two sets. The two sets of area array structured light 3D cameras on each side are arranged horizontally front and back, and the area array structured light 3D cameras in the front row are set to look backward at a 30° angle, while the area array structured light 3D cameras in the back row are set to look forward at a 30° angle.

[0025] In summary, the beneficial effects of the pantograph thickness measurement method and device based on three-dimensional spatial calculation and visual reconstruction of the present invention are as follows: By using multiple sets of area array structured light three-dimensional cameras and employing phase-coded fringe structured light image acquisition, complete three-dimensional point clouds of the upper and lower surfaces, sides, and edges of the pantograph carbon strip are obtained, ultimately achieving the detection of local defects such as notches and peeling. Compared with single-line lasers, it has stronger resistance to ambient light interference, significantly improving measurement accuracy and reliability. The multiple sets of area array structured light three-dimensional cameras are deployed inside the train inspection shed, enabling high-precision three-dimensional information acquisition of the upper surface, lower surface, and side areas of the carbon strip without interfering with the train's running structure or disassembling the carbon strip, providing comprehensive data support for carbon strip thickness measurement, notch identification, and other analyses. The design of multiple triggers for joint control, coupled with an intelligent system for processing and analysis, enables a fully automated closed-loop process from hardware acquisition to software analysis and visualization, improving inspection efficiency. The present invention is highly practical and has significant potential for widespread application. Attached Figure Description

[0026] Figure 1 This is a flowchart illustrating the pantograph thickness measurement method based on three-dimensional spatial calculation and visual reconstruction in this invention.

[0027] Figure 2 This is a top view of the pantograph thickness measurement device for three-dimensional spatial calculation and visual reconstruction in this invention. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of the invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention.

[0029] like Figure 1 As shown, the present invention provides a method for measuring pantograph thickness using three-dimensional spatial calculation and visual reconstruction, which includes the following steps:

[0030] Step 1: Deploy the device. Install and arrange multiple sets of area array structured light 3D cameras 10 on the train inspection shed. Cover the carbon slide plate with top, bottom and side views respectively. Set triggers 20 in the inspection area to control the multiple sets of area array structured light 3D cameras 10 to start automatically or standby. The area array structured light 3D camera 10 includes a projection module and a camera module.

[0031] Specifically, in this embodiment, the trigger 20 is a magnetic steel sensor and an infrared beam sensor. When the train wheel approaches the magnetic steel sensor, it generates an electrical signal due to magnetic field disturbance as a train approach signal. Thereafter, only when the magnetic steel sensor and the infrared beam sensor are triggered synchronously will the area array structured light 3D camera 10 be controlled to start image acquisition. This dual trigger 20 design can effectively avoid the problem of false triggering and improve resource utilization efficiency.

[0032] Step 2: Parameter calibration. Use a multi-dot array calibration board to accurately calibrate multiple sets of area array structured light 3D cameras 10, and obtain the intrinsic parameter matrix K of the camera module and the extrinsic parameter matrix [R|t] between the projection module and the camera module respectively.

[0033] Step 3: Image acquisition. When trigger 20 is triggered, multiple sets of area array structured light cameras are jointly started according to the preset scheduling strategy. The phase-encoded stripe projection method is adopted. The projection module of the area array structured light camera projects a set of sinusoidal stripe patterns with different phases onto the carbon slide surface. The camera module then captures and acquires the image sequence of these stripe patterns projected on the carbon slide surface.

[0034] Step 4: Phase solution. The encapsulated phase is obtained by solving a system of trigonometric equations constructed from multiple images. Then, the continuous phase distribution is recovered using a phase expansion algorithm. Specifically, the model of the trigonometric function equation system is as follows: in For the m-th stripe pattern in pixels grayscale value, Background brightness, In order to adjust the system, This is the initial phase of the pixel. The phase shift of the m-th image;

[0035] Step 5: Coordinate transformation, converting the phase value of each pixel. The corresponding imaging ray establishes a geometric relationship with the projection plane. Depth P is calculated based on the structured light triangulation model, and then the corresponding 3D point is obtained through back projection. Specifically, the back projection algorithm is as follows: in Here, P represents the 3D coordinates in the camera coordinate system, and P represents the projection depth. These are pixel coordinates;

[0036] Step 6: Point cloud generation. The three-dimensional reconstruction of all pixels is completed on multiple sets of structured light 3D cameras 10 to obtain multiple sets of two-dimensional high-density point cloud maps of the carbon skateboard.

[0037] Step 7: Point cloud stitching to obtain the rotation matrix of each group of structured light 3D cameras 10. Translation vector Point clouds in the 10-coordinate system of an arbitrary array structured light 3D camera The specific conversion algorithm for transforming to the world coordinate system is as follows: in The transformed global 3D coordinates This refers to the rotational extrinsic parameter between the projection module and the camera module. This refers to the translation extrinsic parameter between the projection module and the camera module.

[0038] After converting and stitching the various two-dimensional point cloud images into a three-dimensional point cloud image, voxel grid downsampling, weighted overlap fusion, or iterative nearest point (ICP) fine-tuning methods are used to further perform edge fusion, noise filtering, and resolution equalization on the overlapping areas to improve the consistency and integrity of the point cloud.

[0039] Step 8: Analyze the information and extract the remaining thickness based on the geometric profile analysis method. That is, slice the point cloud model at equal intervals along the length direction (X-axis) of the carbon skateboard and divide the entire skateboard into several local profiles. For each profile area, extract the Y-axis coordinate value (i.e., height information) of all points in it, and combine noise filtering and statistical analysis to obtain the minimum Y value of the profile as its local remaining thickness index.

[0040] In addition to extracting the remaining thickness based on the geometric profile analysis method, the left and right edge regions of the point cloud model are also scanned. Combined with the geometric features of the point cloud such as abrupt changes in normal, curvature anomalies and height drops, possible peeling, damage or gap phenomena are automatically detected.

[0041] Step 9: Output and Display. Output all thickness and defect analysis results as structured data files, and generate visualization results such as thickness profile diagrams and notch distribution diagrams based on the structured data files, and output them to the web for display.

[0042] Understandably, in other embodiments, calibration can be performed multiple times using a single camera rotation or line laser scanning, followed by point cloud stitching through feature matching or Simultaneous Localization and Mapping (SLAM) methods. In step 3, multiple sets of line lasers plus camera scanning, or a time-of-flight (ToF) camera, or Gray code or binary encoding can be used to replace phase-coded fringe projection for image acquisition.

[0043] In addition, such as Figure 2As shown, the present invention also provides a pantograph thickness measurement device 100 for three-dimensional spatial calculation and visual reconstruction, which is used to perform the above-mentioned pantograph thickness measurement method for three-dimensional spatial calculation and visual reconstruction, including multiple array structured light three-dimensional cameras 10, triggers 20 and detection and analysis system (not shown).

[0044] Multiple sets of structured light 3D cameras 10 are respectively installed above the inspection shed, on the left and right sides of the inspection shed, and between the carbon slide plate and the top of the train. The upper structured light 3D camera 10 is located above the central axis of the carbon slide plate, while the lower structured light 3D camera 10 is located below the carbon slide plate. The top camera acquires complete structural information of the upper surface of the slide plate through a vertically downward-facing imaging method, mainly used for contact surface wear morphology analysis and thickness upper limit determination. The bottom camera acquires the bottom contour information of the carbon slide plate in an upward-facing manner, completing the acquisition of lower surface data without contacting the train.

[0045] The area array structured light 3D cameras 10 on the left and right sides are located on the left and right sides of the carbon slide plate. Specifically, in this embodiment, the area array structured light 3D cameras 10 on the left and right sides are symmetrically designed, and each side is provided with two sets. The two sets of area array structured light 3D cameras 10 on each side are arranged horizontally front and back, and the area array structured light 3D cameras 10 in the front row are set to look backward at a 30° angle, while the area array structured light 3D cameras 10 in the back row are set to look forward at a 30° angle.

[0046] The side-facing camera, with its staggered front-to-back and oblique incident structure, can effectively avoid the obstruction of the central touch screen structure and fully cover the left and right sides, edges and lower contour areas of the two carbon slide plates, providing multi-view support for the 3D reconstruction of point cloud data and edge anomaly detection.

[0047] The trigger 20 is mounted on the train track and is connected to multiple sets of area array structured light 3D cameras 10. The detection and analysis system is also connected to multiple sets of area array structured light 3D cameras 10. Each area array structured light 3D camera 10 includes a projection module and a camera module. Specifically, in this embodiment, the detection and analysis system is an industrial PC, which is conventional technology in the field and will not be described in detail. It should be emphasized that the physical spatial installation relationship between the detection and analysis system and other structures is not restricted; it is only necessary to ensure signal connection between the system and other structures.

[0048] When the device is in operation, the trigger 20 set at the train track first detects the approach and passage of the train. When the trigger 20 is triggered, multiple sets of area array structured light 3D cameras 10 deployed in the detection shed are controlled by the detection and analysis system and jointly started according to the preset scheduling strategy. They collect structured light images of the upper surface, lower surface and side structure of the carbon slide from the top, bottom and left and right oblique angles respectively.

[0049] The inspection and analysis system then automatically invokes the multi-view registration and 3D reconstruction algorithm module to perform joint modeling based on the structured light projection model and geometric constraints, generating complete 3D point cloud data of the carbon skateboard. Based on the point cloud, the system further executes calculations such as carbon skateboard region localization, attitude correction, thickness extraction, and edge defect identification, ultimately outputting stable and accurate thickness and defect parameters. Finally, the system converts the parameter results into visualized data and outputs it to a display terminal for presentation to inspection personnel.

[0050] In summary, the beneficial effects of the pantograph thickness measurement method and device based on three-dimensional spatial calculation and visual reconstruction of the present invention are as follows: By using multiple sets of area array structured light three-dimensional cameras 10 and employing phase-coded fringe structured light image acquisition, complete three-dimensional point clouds of the upper and lower surfaces, sides, and edges of the pantograph carbon strip are obtained, ultimately achieving the detection of local defects such as notches and peeling. Compared with single-line lasers, it has stronger resistance to ambient light interference, significantly improving measurement accuracy and reliability. The multiple sets of area array structured light three-dimensional cameras 10 are deployed inside the train inspection shed, enabling high-precision three-dimensional information acquisition of the upper surface, lower surface, and side areas of the carbon strip without interfering with the train's running structure or disassembling the carbon strip, providing comprehensive data support for carbon strip thickness measurement, notch identification, and other analyses. Through the design of multiple triggers 20 for joint control and the integration with an intelligent system for processing and analysis, a fully automated closed-loop process from hardware acquisition to software analysis and visualization can be achieved, improving inspection efficiency. The present invention is highly practical and has significant potential for widespread application.

[0051] The embodiments described above illustrate only one implementation of the invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept, and these all fall within the scope of protection of the invention. Therefore, the scope of protection of the invention patent should be determined by the appended claims.

Claims

1. A method for measuring pantograph thickness using three-dimensional spatial calculation and visual reconstruction, characterized in that: Includes the following steps: Step 1: Deploy the device. Install and arrange multiple sets of area array structured light 3D cameras on the train inspection shed. Cover the carbon slide plate with top, bottom and side views respectively. Set triggers in the inspection area to control the multiple sets of area array structured light 3D cameras to start automatically or standby. The area array structured light 3D camera includes a projection module and a camera module. Step 2: Parameter calibration. Use a multi-dot array calibration board to accurately calibrate multiple sets of area array structured light 3D cameras, and obtain the intrinsic parameter matrix K of the camera module and the extrinsic parameter matrix [R|t] between the projection module and the camera module respectively. Step 3: Image acquisition. When the trigger is triggered, multiple sets of area array structured light cameras are jointly started according to the preset scheduling strategy. The phase-encoded stripe projection method is adopted. The projection module of the area array structured light camera projects a set of sinusoidal stripe patterns with different phases onto the carbon slide surface, and the camera module captures the image sequence of these stripe patterns projected on the carbon slide surface. Step 4: Phase solution. The encapsulated phase is obtained by solving a system of trigonometric equations constructed from multiple images. Then, the continuous phase distribution is recovered using a phase expansion algorithm. Specifically, the model of the trigonometric function equation system is as follows: in For the m-th stripe pattern in pixels grayscale value, Background brightness, In order to adjust the system, This is the initial phase of the pixel. The phase shift of the m-th image; Step 5: Coordinate transformation, converting the phase value of each pixel. The corresponding imaging ray establishes a geometric relationship with the projection plane. Depth P is calculated based on the structured light triangulation model, and then the corresponding 3D point is obtained through back projection. Specifically, the back projection algorithm is as follows: in Here, P represents the 3D coordinates in the camera coordinate system, and P represents the projection depth. These are pixel coordinates; Step 6: Point cloud generation. The three-dimensional reconstruction of all pixels is completed on multiple sets of structured light 3D cameras to obtain multiple sets of two-dimensional high-density point cloud maps of the carbon skateboard. Step 7: Point cloud stitching to obtain the rotation matrix of each group of structured light 3D cameras. Translation vector Point clouds in the coordinate system of an arbitrary array structured light 3D camera The specific conversion algorithm for transforming to the world coordinate system is as follows: in The transformed global 3D coordinates This refers to the rotational extrinsic parameter between the projection module and the camera module. This refers to the translation extrinsic parameter between the projection module and the camera module. Step 8: Analyze the information and extract the remaining thickness based on the geometric profile analysis method. That is, slice the point cloud model at equal intervals along the length direction (X-axis) of the carbon skateboard and divide the entire skateboard into several local profiles. For each profile area, extract the Y-axis coordinate value (i.e., height information) of all points in it, and combine noise filtering and statistical analysis to obtain the minimum Y value of the profile as its local remaining thickness index. Step 9: Output and Display. Output all thickness and defect analysis results as structured data files, and generate visualization results such as thickness profile diagrams and notch distribution diagrams based on the structured data files, and output them to the web for display.

2. The pantograph thickness measurement method based on three-dimensional spatial calculation and visual reconstruction as described in claim 1, characterized in that: In step 1, the triggers are a magnetic steel sensor and an infrared beam sensor. When a train wheel approaches the magnetic steel sensor, it generates an electrical signal due to magnetic field disturbance, which serves as a train approach signal. Subsequently, the structured light 3D camera is controlled to start image acquisition only when the magnetic steel sensor and the infrared beam sensor are triggered synchronously.

3. The pantograph thickness measurement method based on three-dimensional spatial calculation and visual reconstruction as described in claim 1, characterized in that: In step 7, after converting and stitching the two-dimensional point cloud images into a three-dimensional point cloud image, the overlapping areas are further processed by voxel grid downsampling, weighted overlap fusion, or iterative nearest point (ICP) fine-tuning methods to improve the consistency and integrity of the point cloud.

4. The pantograph thickness measurement method based on three-dimensional spatial calculation and visual reconstruction as described in claim 1, characterized in that: In step 8, in addition to extracting the remaining thickness based on the geometric profile analysis method, the left and right edge regions of the point cloud model are also scanned. Combined with the geometric features of the point cloud, such as abrupt changes in normal, curvature anomalies, and sudden drops in height, possible peeling, damage, or gaps are automatically detected.

5. The pantograph thickness measurement method based on three-dimensional spatial calculation and visual reconstruction as described in claim 1, characterized in that: In steps 2 and 7, calibration can be performed multiple times using a single camera rotation or line laser scanning, and then point cloud stitching can be achieved by registration using feature matching or Simultaneous Localization and Mapping (SLAM) methods.

6. The pantograph thickness measurement method based on three-dimensional spatial calculation and visual reconstruction as described in claim 1, characterized in that: In step 3, multiple sets of line lasers plus camera scanning, or a time-of-flight (ToF) camera, or Gray code or binary encoding can be used to replace the phase-coded stripe projection method to complete image acquisition.

7. A pantograph thickness measurement device based on three-dimensional spatial calculation and visual reconstruction, used to perform the pantograph thickness measurement method based on three-dimensional spatial calculation and visual reconstruction as described in any one of claims 1 to 4, characterized in that: The system includes multiple sets of area-array structured light 3D cameras, triggers, and a detection and analysis system. The multiple sets of area-array structured light 3D cameras are respectively installed above the detection shed, on the left and right sides of the detection shed, and between the carbon slide plate and the top of the train. The upper area-array structured light 3D camera is located above the central axis of the carbon slide plate, the left and right side area-array structured light 3D cameras are located on the left and right sides of the carbon slide plate, and the lower area-array structured light 3D camera is located in the area below the carbon slide plate. The triggers are installed on the train track and are connected to the multiple sets of area-array structured light 3D cameras and the detection and analysis system. The detection and analysis system is also connected to the multiple sets of area-array structured light 3D cameras. Each area-array structured light 3D camera includes a projection module and a camera module.

8. The pantograph thickness measuring device for three-dimensional spatial calculation and visual reconstruction as described in claim 1, characterized in that: The area array structured light 3D cameras on the left and right sides are symmetrically designed, and there are two sets on each side. The two sets of area array structured light 3D cameras on each side are arranged horizontally front and back, and the area array structured light 3D cameras in the front row are set to look backward at a 30° angle, while the area array structured light 3D cameras in the back row are set to look forward at a 30° angle.

Images