Elevator toe guard deformation quantity detection device and method based on machine vision

Abstract

The invention discloses an elevator toe guard deformation quantity detection device and method based on machine vision, and relates to the technical field of elevator safety detection, and the elevator toe guard deformation quantity detection device comprises a reference pattern layer which is arranged on the outer surface of a toe guard; an image acquisition end of the visual detection module corresponds to the reference image layer, and when the toe guard deforms, the visual detection module is used for acquiring a deformation image of the reference image layer and obtaining the deformation quantity of the toe guard according to the deformation image of the reference image layer; according to the technical scheme, the reference pattern layer is arranged on the toe guard, the reference pattern layer is formed by splicing the multiple square sub-pattern layers, the adjacent square sub-pattern layers are provided with the surface layers of different colors, a stable reference datum can be provided for visual inspection, and high-precision measurement of the deformation quantity of the toe guard under the stress condition is achieved by combining a camera and a sensor. Compared with a traditional mechanical measurement method, the method has the advantages that the computer vision technology is adopted, the detection precision is improved, and the influence of abrasion of a mechanical part or aging of an elastic element on a measurement result is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of elevator safety detection, and in particular to a device and method for detecting the deformation of an elevator footboard based on machine vision. Background Art

[0002] With the continuous advancement of science and technology, elevators, as an important tool in modern high-rise buildings, have been widely used in urban construction. As an important protective component of the elevator, the mechanical strength of the elevator toeboard directly affects the safety of the elevator.

[0003] At present, the testing method of the toe guard is usually to apply a standard pressure of 300N, and then measure the deformation of the toe guard through a measuring device to determine whether it meets the safety requirements. The measurement accuracy is usually not high.

[0004] Patent CN221528235U "An Elevator Footboard Strength Test Device" proposes a device for testing the strength of footboards using a linear drive and a pressure detection component. The device applies pressure through a linear drive and uses a displacement detection unit to detect the deformation of the footboard. Although this method improves the test accuracy, its complex mechanical structure and operation process make the maintenance cost high and the operation relatively cumbersome.

[0005] Patent CN220959966U "A device for measuring the elastic deformation of a car footboard" uses a compression spring and mechanical drive to measure the displacement of the compression spring to infer the elastic deformation of the footboard. This method simplifies the measurement process to a certain extent and is relatively simple to operate. However, due to the characteristics of the spring itself and the influence of external factors, there may be certain errors in the measurement results, and it still needs to rely on a complex mechanical structure for operation.

[0006] In summary, the existing detection devices still have deficiencies in measurement accuracy and ease of operation. In view of this, how to provide an elevator toe guard detection device that can ensure measurement accuracy, simplify the equipment structure and operation process, and reduce maintenance costs is a problem that technicians in this field need to solve urgently. Summary of the invention

[0007] The purpose of the present invention is to provide a device and method for detecting the deformation of an elevator toe plate based on machine vision, so as to solve the problems existing in the prior art.

[0008] To achieve the above object, the present invention provides the following solution: The present invention provides an elevator footboard deformation detection device based on machine vision, comprising:

[0009] A base layer is arranged on the outer surface of the toeboard;

[0010] The visual inspection module has an image acquisition end corresponding to the reference layer. When the foot guard is deformed, the image acquisition end is used to acquire a deformation image of the reference layer and obtain the deformation amount of the foot guard according to the deformation image of the reference layer.

[0011] Furthermore, the reference layer is formed by splicing a plurality of square sub-layers, and adjacent square sub-layers have surface layers of different colors.

[0012] Furthermore, the reference layer is mounted on the toeboard via a layer fixing frame.

[0013] Furthermore, the visual detection module includes:

[0014] A camera, corresponding to the reference layer, for collecting a deformation image of the reference layer;

[0015] Sensors to measure the distance and angle between the camera and the toe guard;

[0016] A tripod to support the camera and adjust the angle between the camera and the foot guard;

[0017] A display screen has a built-in image processing system for processing the deformation image of the reference layer and obtaining the deformation amount of the toe guard. The display screen can display the deformation amount of the toe guard, the distance and angle between the camera and the toe guard.

[0018] The present invention also provides a method for detecting the deformation amount of an elevator toeboard based on machine vision, using the device for detecting the deformation amount of an elevator toeboard based on machine vision, comprising the following steps:

[0019] S1: Arrange the reference layer on the outer surface of the toeboard;

[0020] S2: Arrange the camera on one side of the footboard through a tripod so that the camera corresponds to the reference layer, adjust the distance and angle between the camera and the footboard according to preset parameters, and confirm whether the preset parameters are reached through a sensor;

[0021] S3: When the footboard is not under stress, the deformation image of the reference layer is collected;

[0022] S4: applying a pressure of 300N to the foot guard, collecting the deformation image of the reference layer again, and obtaining the deformation amount of the foot guard;

[0023] S5: The deformation amount of the toe guard is displayed on the display screen, and the detection is completed.

[0024] Furthermore, in steps S3 and S4, distortion correction is performed after the base layer deformation image is collected. The distortion correction steps include: using Matlab's Camera_Calibrator toolbox to calibrate the camera, obtaining the camera's intrinsic parameter matrix and distortion coefficient, inputting the collected base layer deformation image together with the camera's intrinsic parameter matrix and distortion coefficient into OpenCV's cv2.remap function, performing a geometric correction operation, and removing the distortion in the base layer deformation image.

[0025] Furthermore, the step of obtaining the deformation amount of the toe guard includes:

[0026] The Shi-Tomasi corner detection algorithm is used to extract the corner positions of the base layer deformation image, and the pixel gradient changes in the local area of ​​the base layer deformation image are analyzed to identify the significant feature points in the image. Subsequently, multiple base layer deformation images are fused and the average position of multiple identification results is taken. The pixel distance between the corner points in the base layer deformation image is measured by the corner detection tool in OpenCV to confirm the pixel displacement ΔS corresponding to the base layer after applying 300N pressure to the foot guard.

[0027] Calculate the actual physical length P corresponding to each pixel:

[0028]

[0029] Where L is the side length of the base layer, and n is the number of pixels corresponding to a single side of the base layer. The deformation variable ΔZ of the footboard is calculated:

[0030]

[0031] Where θ is the angle between the camera and the toe guard.

[0032] The present invention discloses the following technical effects:

[0033] The reference layer is arranged on the toe guard, and the reference layer is formed by splicing a plurality of square sub-layers. Adjacent square sub-layers have surface layers of different colors, which can provide a stable reference benchmark for visual detection. Combined with the camera and the sensor, the high-precision measurement of the deformation amount of the toe guard under the force condition is realized. Compared with the traditional mechanical measurement method, the present invention adopts computer vision technology, which not only improves the detection accuracy, but also avoids the influence of the wear of mechanical parts or the aging of elastic elements on the measurement results.

[0034] In actual use, the camera can quickly capture the deformation image of the reference layer, and combined with the image processing system built into the display screen, analyze and display the deformation of the toeboard in real time. This measurement method has the characteristics of high precision and real-time performance, and is particularly suitable for dynamic monitoring and regular maintenance of elevator toeboards. Compared with the prior art, the present invention also has the advantages of simple operation, compact and portable equipment, and can be used in narrow installation environments, greatly improving the efficiency and convenience of detection work. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0036] Figure 1 It is a schematic diagram of the structure of the present invention;

[0037] Figure 2 Schematic diagram of camera and sensor;

[0038] Among them, 1. foot guard; 2. base layer; 3. layer fixing frame; 4. sensor; 5. camera; 6. tripod; 7. display screen. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0041] An embodiment of the present invention provides an elevator toe board deformation detection device based on machine vision, comprising: a reference layer 2, arranged on the outer surface of the toe board 1; a visual detection module, whose image acquisition end corresponds to the reference layer 2, and when the toe board 1 is deformed, it is used to collect a deformation image of the reference layer 2, and obtain the deformation amount of the toe board 1 according to the deformation image of the reference layer 2.

[0042] In this embodiment, the reference layer 2 is formed by splicing a plurality of 5 cm×5 cm square sub-layers, and adjacent square sub-layers have surface layers of different colors. The reference layer 2 is mounted on the footboard 1 through a layer fixing frame 3 .

[0043] In this embodiment, the visual detection module includes: a camera 5, corresponding to the reference layer 2, for collecting the deformation image of the reference layer 2; a sensor 4, for measuring the distance and angle between the camera 5 and the toe guard 1, the sensor 4 is a distance sensor 4, and the angle between the camera 5 and the toe guard 1 can be obtained by measuring the distance between the camera 5 and the toe guard 1; a tripod 6, for supporting the camera 5 and being able to adjust the angle between the camera 5 and the toe guard 1; a display screen 7, with a built-in image processing system, for processing the deformation image of the reference layer 2 and obtaining the deformation amount of the toe guard 1, and the display screen 7 can display the deformation amount of the toe guard 1, the distance and angle between the camera 5 and the toe guard 1.

[0044] The following describes in detail the elevator footboard deformation detection method based on machine vision in combination with the above device, including the following steps:

[0045] S1: Arrange the reference layer 2 on the outer surface of the toe guard 1;

[0046] S2: Arrange the camera 5 on one side of the toe guard 1 through the tripod 6, so that the camera 5 corresponds to the reference layer 2, adjust the distance and angle between the camera 5 and the toe guard 1 according to the preset parameters, and confirm whether the preset parameters are reached through the sensor 4. In this embodiment, the camera 5 needs to be separated from the toe guard 1 by an appropriate distance, and the image acquisition end of the camera 5 is facing the toe guard 1; after adjusting the position of the camera 5, fix the tripod 6 through a fixing device such as a lock to limit the displacement of the tripod 6 and the camera 5, so as to avoid offset in subsequent operations.

[0047] S3: When the footboard 1 is not subjected to force, a deformation image of the reference layer 2 is collected;

[0048] S4: Apply 300N pressure to the foot guard 1 at the position of Figure 1 At point A in the middle, the deformation image of the reference layer 2 is collected again to obtain the deformation of the foot guard 1;

[0049] S5: The deformation amount of the foot guard 1 is displayed on the display screen 7, and the detection is completed.

[0050] In steps S3 and S4, distortion correction is performed after the deformed image of the reference layer 2 is collected. The steps of distortion correction include: using the Camera_Calibrator toolbox of Matlab to calibrate the camera 5, obtaining the intrinsic parameter matrix and distortion coefficient of the camera 5, inputting the collected deformed image of the reference layer 2 together with the intrinsic parameter matrix and distortion coefficient of the camera 5 into the cv2.remap function of OpenCV, performing geometric correction operations, and removing the distortion in the deformed image of the reference layer 2. Distortion correction ensures the accurate position of each pixel in the image, thereby providing reliable basic data for subsequent deformation calculation. Specifically, first collect multiple calibration plate images, use the Matlab toolbox or OpenCV to perform camera calibration, and obtain the intrinsic parameters and distortion coefficients of the camera 5. Then use these parameters to correct the image to ensure that the image will not cause errors due to the distortion of the camera 5.

[0051] Furthermore, the step of obtaining the deformation amount of the toe guard 1 includes:

[0052] The Shi-Tomasi corner detection algorithm is used to extract the corner positions of the deformed image of the base layer 2, and the pixel gradient changes in the local area of ​​the deformed image of the base layer 2 are analyzed to identify the significant feature points in the image. Then, multiple deformed images of the base layer 2 are fused, and the average position of multiple identification results is taken. The pixel distance between the corners in the deformed image of the base layer 2 is measured by the corner detection tool in OpenCV to confirm the pixel displacement ΔS corresponding to the base layer 2 after applying 300N pressure to the foot guard 1. To ensure accuracy, the pixel distance between multiple corners needs to be measured multiple times and the average value is taken. In addition, due to light, equipment jitter or other factors, the pixel displacement measurement may be inaccurate, so it is necessary to use a multi-frame image fusion method to improve the accuracy of displacement measurement by averaging the results of multiple consecutive images.

[0053] Calculate the actual physical length P corresponding to each pixel:

[0054]

[0055] Where L is the side length of the reference layer 2, and n is the number of pixels corresponding to a single side of the reference layer 2. The deformation variable ΔZ of the footboard 1 is calculated:

[0056]

[0057] Wherein, θ is the angle between the camera 5 and the foot guard 1 .

[0058] In the actual measurement process, there may be some sources of error, including slight deviation between the camera 5 and the footboard 1, changes in ambient light, etc. To reduce these errors, the system introduces angle correction and illumination correction mechanisms:

[0059] Angle correction is that the system measures the distance between the camera 5 and the toe board 1 in real time through the sensor 4, and calculates the angle θ between the camera 5 and the toe board 1 based on the distance. If the system detects that the angle between the camera 5 and the toe board 1 is deviated, the system will provide a real-time prompt for angle adjustment on the integrated display screen 7. The measurement personnel can manually adjust the angle of the camera 5 according to the angle adjustment prompt on the display screen 7 to ensure that the camera 5 is facing the toe board 1 (the image acquisition path of the camera 5 is perpendicular to the toe board 1), reduce the measurement error caused by the angle deviation, and make the deformation calculation result more accurate.

[0060] Light correction is performed by the system through the automatic exposure adjustment function of the camera 5 in an environment with insufficient light or large light changes. The camera 5 automatically adjusts the exposure parameters (such as exposure time and gain) according to the brightness of the image captured in real time to ensure that the brightness of the image is moderate. After the adjustment is completed, the system will display a prompt message of the exposure status on the display screen 7, and the surveyor can decide whether further manual intervention is required based on the prompt. If the image brightness is still not ideal, the surveyor can adjust the external light source or manually fine-tune the exposure parameters of the camera 5 to further optimize the image quality. After the surveyor makes manual adjustments, the system will optimize the exposure parameters again and display feedback prompts on the integrated display screen 7. When the image reaches the ideal exposure state, the system will prompt the surveyor to confirm to ensure that the image has reached the optimal brightness state, and then the detection task can be continued. After completing the exposure adjustment and image acquisition, the system will post-process the image and use the histogram equalization algorithm to improve the contrast and clarity of the image. The algorithm redistributes the brightness values ​​of the image pixels so that the deformed image of the reference layer 2 in the image can maintain sufficient visibility and contrast even in the case of uneven light, thereby ensuring the accuracy of corner detection. In an environment with poor lighting conditions, noise may appear in the image. The system uses median filtering or Gaussian filtering technology to reduce random noise in the image, ensure the purity of the image, and improve detection accuracy. The system also processes multiple frames of images through image fusion technology and takes the average result of multiple images to reduce noise and errors in a single frame. By fusing multiple frames of images, the system can further improve the stability and accuracy of deformation calculation.

[0061] After completing the above steps, the system displays the calculated deformation variable ΔZ of the toeboard 1 on the display screen 7 in real time, and automatically saves the measurement data to the system database. The data is saved with the timestamp, the distance and angle between the camera 5 and the toeboard 1 and other measurement conditions to ensure that the measurement environment can be accurately traced back in subsequent analysis. This operation ensures the traceability of each measurement and provides complete data support for elevator maintenance.

[0062] In the description of the present invention, it should be understood that the terms "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside" and "outside" etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0063] The embodiments described above are only descriptions of the preferred modes of the present invention, and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A device for detecting the deformation of elevator footboards based on machine vision, characterized in that: include: A reference layer (2) is arranged on the outer surface of the footboard (1); A visual inspection module, whose image acquisition end corresponds to the reference layer (2), is used to acquire a deformation image of the reference layer (2) when the foot guard (1) is deformed, and to obtain the deformation amount of the foot guard (1) based on the deformation image of the reference layer (2).

2. The device for detecting the deformation of elevator toeboards based on machine vision according to claim 1, characterized in that: The reference layer (2) is formed by splicing a plurality of square sub-layers, and adjacent square sub-layers have surface layers of different colors.

3. The device for detecting the deformation of elevator toeboards based on machine vision according to claim 2, characterized in that: The reference layer (2) is mounted on the footboard (1) via a layer fixing frame (3).

4. The device for detecting the deformation of elevator toeboards based on machine vision according to claim 2, characterized in that: The visual inspection module comprises: A camera (5), corresponding to the reference layer (2), and used for collecting a deformation image of the reference layer (2); A sensor (4) for measuring the distance and angle between the camera (5) and the foot guard (1); A tripod (6) for supporting the camera (5) and capable of adjusting the angle between the camera (5) and the foot guard (1); A display screen (7) has a built-in image processing system for processing the deformation image of the reference layer (2) and obtaining the deformation amount of the foot guard (1). The display screen (7) can display the deformation amount of the foot guard (1), the distance and angle between the camera (5) and the foot guard (1).

5. A method for detecting the deformation of elevator toeboards based on machine vision, characterized in that: The device for detecting the deformation of an elevator toeboard (1) based on machine vision according to claim 4 comprises the following steps: S1: Arranging a reference layer (2) on the outer surface of the toe guard (1); S2: placing a camera (5) on one side of the footboard (1) via a tripod (6) so that the camera (5) corresponds to the reference layer (2), adjusting the distance and angle between the camera (5) and the footboard (1) according to preset parameters, and confirming whether the preset parameters are reached via a sensor (4); S3: When the footboard (1) is not subjected to force, a deformation image of the reference layer (2) is collected; S4: applying a pressure of 300 N to the foot guard (1), collecting the deformation image of the reference layer (2) again, and obtaining the deformation amount of the foot guard (1); S5: The deformation amount of the foot guard (1) is displayed on the display screen (7), and the detection is completed.

6. The method for detecting the deformation of elevator toeboards based on machine vision according to claim 5, characterized in that: In steps S3 and S4, distortion correction is performed after the deformed image of the reference layer (2) is collected. The distortion correction steps include: using the Camera_Calibrator toolbox of Matlab to calibrate the camera (5), obtaining the intrinsic parameter matrix and distortion coefficient of the camera (5), inputting the collected deformed image of the reference layer (2) together with the intrinsic parameter matrix and distortion coefficient of the camera (5) into the cv2.remap function of OpenCV, performing a geometric correction operation, and removing the distortion in the deformed image of the reference layer (2).

7. The method for detecting the deformation of elevator toeboards based on machine vision according to claim 5, characterized in that: The step of obtaining the deformation amount of the foot guard (1) comprises: The Shi-Tomasi corner detection algorithm is used to extract the corner positions of the deformed image of the reference layer (2), and the pixel gradient changes in the local area of ​​the deformed image of the reference layer (2) are analyzed to identify the significant feature points in the image. Subsequently, multiple deformed images of the reference layer (2) are fused and the average position of multiple identification results is taken. The pixel distance between the corner points in the deformed image of the reference layer (2) is measured by the corner detection tool in OpenCV to confirm the pixel displacement ΔS corresponding to the reference layer (2) after a pressure of 300N is applied to the foot guard (1); Calculate the actual physical length P corresponding to each pixel: Wherein, L is the side length of the reference layer (2), and n is the number of pixels corresponding to a single side of the reference layer (2); Calculate the deformation ΔZ of the toe guard (1): Wherein, θ is the angle between the camera (5) and the foot guard (1).

Images