Concrete pole defect intelligent detection device and method based on machine vision

Abstract

The invention belongs to the technical field of concrete detection, and provides a concrete pole defect intelligent detection device and method based on machine vision, and the device comprises an annular supporting mechanism, and a walking mechanism and an image acquisition mechanism which are arranged on the annular supporting mechanism; the annular supporting mechanism is in an opening shape and comprises a first arc-shaped supporting element, a second arc-shaped supporting element and a rotary spring shaft pin used for connecting the first arc-shaped supporting element and the second arc-shaped supporting element. The walking mechanism comprises a vertical walking element, a transverse walking element and a horizontal monitoring element; an image acquisition telescopic arm is arranged in the image acquisition mechanism; the horizontal state of the walking mechanism is detected through the horizontal monitoring element, the vertical walking element and the image acquisition telescopic arm are adjusted in real time according to the detection result of the horizontal state of the walking mechanism, and the horizontal state of the intelligent detection device is kept; based on cooperation of the vertical walking element and the transverse walking element, intelligent obstacle avoidance of the intelligent detection device is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of concrete detection, and in particular relates to a device and method for intelligently detecting defects of concrete poles based on machine vision. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] Substations are a vital part of the power industry and an indispensable building in people's daily lives. Therefore, damage detection and maintenance of substation ring concrete poles are very important. Concrete poles are complex multiphase and porous systems composed of cement, aggregates, steel bars, etc. Changes in the surrounding environment and media as well as external forces will cause the strength and durability of concrete poles to decrease. Traditional detection methods require inspectors to climb up the poles to work. Not only is the detection efficiency and accuracy very low, but there are also great safety hazards. An intelligent detection system is needed to replace manual work.

[0004] The development of image processing technology has greatly enhanced the detection efficiency of surface damage defects of concrete poles. As an indispensable facility in the power transmission and distribution network, the structural integrity of concrete poles is directly related to the safe operation of transmission lines. As the concrete poles are used for a longer time, cracks and other defects will appear on the concrete poles, affecting the structural strength of the concrete poles. The bearing capacity of the concrete poles decreases, which seriously affects the safe operation of the distribution lines. Summary of the invention

[0005] To solve the above problems, the present invention proposes a device and method for intelligent detection of concrete pole defects based on machine vision. The improved intelligent detection device walks on the concrete pole to accurately obtain the surface image of the concrete pole in real time, thereby improving the accuracy of concrete pole defect detection.

[0006] According to some embodiments, a first solution of the present invention provides a machine vision-based intelligent detection device for concrete pole defects, which adopts the following technical solutions:

[0007] A machine vision-based intelligent detection device for concrete pole defects, comprising an annular support mechanism, and a walking mechanism and an image acquisition mechanism arranged on the annular support mechanism; wherein the annular support mechanism is open, comprising a first arc-shaped support element, a second arc-shaped support element, and a rotating spring axle pin for connecting the first arc-shaped support element and the second arc-shaped support element; the walking mechanism comprises a vertical walking element, a transverse walking element and a horizontal monitoring element, and the image acquisition mechanism is provided with an image acquisition telescopic arm; the horizontal state of the walking mechanism is detected by the horizontal monitoring element, and the vertical walking element and the image acquisition telescopic arm are adjusted in real time according to the detection result of the horizontal state of the walking mechanism to maintain the horizontal state of the intelligent detection device; based on the cooperation of the vertical walking element and the transverse walking element, intelligent obstacle avoidance of the intelligent detection device is realized.

[0008] As a further technical limitation, the vertical walking element and the transverse walking element do not work at the same time. When the vertical walking element climbs or descends, the transverse roller supporting telescopic arm in the transverse walking element contracts. When the vertical walking element encounters an obstacle during walking, the vertical roller supporting telescopic arm in the vertical walking element contracts, the vertical walking element stops walking, and the transverse roller supporting telescopic arm extends. The reaction force is transmitted to the force-bearing rotating rod in the transverse walking element through the support of the reaction force supporting rod in the transverse walking element. Under the action of the force-bearing rotating rod, the transverse walking element is The rotating shaft of the transverse roller device in the walking element rotates to make the transverse rollers in the transverse walking element contact and stick to the concrete pole until the transverse rollers in all the transverse walking elements are attached to the surface of the concrete pole; after the vertical walking element rotates away from the surface of the concrete pole, the transverse roller motor in the transverse walking element is controlled to rotate, driving the transverse walking element to rotate horizontally on the surface of the concrete pole. After the horizontal rotation is completed, after the transverse walking element rotates away from the surface of the concrete pole, the vertical walking element is controlled to continue walking, completing the intelligent obstacle avoidance of the intelligent detection device.

[0009] As a further technical limitation, the first arc-shaped support element or the second arc-shaped support element comprises two arc-shaped support arms with completely identical structures; the arc-shaped support arm in the first arc-shaped support element is connected with the arc-shaped support arm in the second arc-shaped support element through the rotating spring axle pin, and the clamping of the arc-shaped support arm on the concrete pole is achieved by adjusting the rotating spring axle pin.

[0010] As a further technical limitation, during the climbing or descending process of the intelligent detection device, the horizontal state of the walking mechanism is monitored in real time based on the horizontal electronic monitor, and the motion state of the vertical walking element is adjusted according to the horizontal state detection result. The horizontal state of the walking mechanism is adjusted by adjusting the motion state of different vertical rollers in the vertical walking element to keep the walking mechanism continuously level.

[0011] As a further technical limitation, during the process of static image acquisition by the intelligent detection device, the length of the image acquisition telescopic arm is adjusted according to the obtained horizontal state detection result of the walking mechanism, so that the concrete pole images acquired by the image acquisition mechanism are all in the same horizontal plane.

[0012] As a further technical limitation, the outer sides of the vertical rollers in the vertical walking element and the transverse rollers in the transverse walking element are both covered with patterned leather covers to increase the friction between the vertical rollers or the transverse rollers and the concrete poles during travel.

[0013] As a further technical limitation, a machine vision-based intelligent detection device for concrete pole defects also includes a control element that is communicatively connected to the annular support mechanism, the walking mechanism and the image acquisition mechanism respectively, and the control element is used to monitor the working status information of the annular support mechanism, the walking mechanism and the image acquisition mechanism in real time and perform feedback adjustment of the intelligent detection device according to the obtained working status information.

[0014] As a further technical limitation, a vertical pressure sensor and a lateral pressure sensor are respectively provided in the vertical walking element and the lateral walking element. When the walking mechanism is moving, the pressure state of the walking mechanism is monitored in real time by the vertical pressure sensor and the lateral pressure sensor, and the horizontal state of the walking mechanism is judged in combination with the horizontal monitoring element. Feedback adjustment of the horizontal state of the intelligent detection device is performed according to the detection result of the horizontal monitoring element.

[0015] According to some embodiments, the second solution of the present invention provides a method for intelligent detection of concrete pole defects based on machine vision, which adopts the intelligent detection device for concrete pole defects based on machine vision provided by the first solution, and adopts the following technical solutions:

[0016] A method for intelligent detection of concrete pole defects based on machine vision, comprising:

[0017] The intelligent detection device is installed on the concrete pole, and the intelligent detection device is controlled to move horizontally on the concrete pole through the control element and the horizontal monitoring element;

[0018] During the traveling process, the image of the concrete pole is acquired through the image acquisition mechanism and the horizontal monitoring element;

[0019] Processing the acquired concrete pole image to obtain a concrete pole image feature value;

[0020] Defect classification is performed based on the obtained concrete pole image feature values ​​to complete the defect detection of the concrete pole.

[0021] As a further technical limitation, the concrete pole image defects include at least no cracks, transverse cracks, longitudinal cracks and cracks.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] The present invention realizes intelligent obstacle avoidance of the intelligent detection device through the mutual cooperation of the horizontal walking element and the vertical walking element; a horizontal monitoring element is arranged to keep the intelligent detection device in a horizontal state during the traveling process and the image acquisition process, thereby increasing the service life of the traveling mechanism and improving the accuracy of the acquired concrete pole image, further improving the accuracy and efficiency of defect detection, and improving the stability of the intelligent detection device. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The drawings in the specification that constitute a part of this embodiment are used to provide a further understanding of this embodiment. The schematic embodiments of this embodiment and their descriptions are used to explain this embodiment and do not constitute improper limitations on this embodiment.

[0025] Figure 1 It is a structural schematic diagram of a device for intelligent detection of concrete pole defects based on machine vision in Embodiment 1 of the present invention;

[0026] Figure 2 Schematic diagram of the structure of the transverse traveling element in the first embodiment of the present invention;

[0027] Figure 3 Schematic diagram of the structure of the vertical travel element in the first embodiment of the present invention;

[0028] Figure 4 Schematic diagram of the structure of the image acquisition mechanism in the first embodiment of the present invention;

[0029] Figure 5 This is a flow chart of a method for intelligently detecting defects in concrete poles based on machine vision in Embodiment 2 of the present invention;

[0030] Figure 6 This is a structural block diagram of image acquisition in Embodiment 2 of the present invention;

[0031] Figure 7 Flow chart of the U-Net method in Embodiment 2 of the present invention;

[0032] Among them, 1. vertical walking element; 11. vertical roller; 12. vertical roller rotating shaft; 13. vertical roller device rotating motor; 14. vertical roller telescopic arm; 15. vertical pressure sensor; 16. vertical roller motor; 2. balancing connection element; 3. horizontal walking element; 31. horizontal roller support telescopic arm; 32. reaction support rod; 33. horizontal roller; 34. horizontal roller motor; 35. horizontal roller connecting arm; 36. force-bearing rotating rod; 37. horizontal roller device rotating shaft; 38. horizontal pressure sensor; 4. rotating spring shaft pin; 5. control element; 6. horizontal monitoring element; 7. arc support arm; 8. image acquisition mechanism; 81. image acquisition support arm; 82. image acquisition telescopic arm; 83. image acquisition camera. DETAILED DESCRIPTION

[0033] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0034] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0035] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0036] In the present invention, terms such as "upper", "lower", "left", "right", "front", "back", "vertical", "horizontal", "side", "bottom", etc. indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings. They are relational words determined only for the convenience of describing the structural relationships of the various parts or elements of the present invention, and do not specifically refer to any part or element in the present invention and should not be understood as limitations on the present invention.

[0037] In the present invention, terms such as "fixed connection", "connected", "connection", etc. should be understood in a broad sense, indicating that it can be fixedly connected, integrally connected or detachably connected; it can be directly connected or indirectly connected through an intermediate medium. For relevant scientific research or technical personnel in this field, the specific meanings of the above terms in the present invention can be determined according to specific circumstances, and they cannot be understood as limitations on the present invention.

[0038] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0039] Embodiment 1

[0040] Embodiment 1 of the present invention introduces a machine vision-based intelligent detection device for concrete pole defects.

[0041] A machine vision-based intelligent detection device for concrete pole defects comprises an annular support mechanism, and a walking mechanism and an image acquisition mechanism 8 arranged on the annular support mechanism.

[0042] like Figure 1 As shown, the annular support mechanism in the intelligent detection device for concrete pole defects based on machine vision includes four arc-shaped support arms 7 and a rotating spring axle pin 4; the walking mechanism includes a vertical walking element 1 and a horizontal walking element 3; and also includes a balance; a balance connection element 2, a control element 5, a horizontal monitoring element 6 and an image acquisition mechanism 8.

[0043] like Figure 2 The lateral walking element 3 shown includes a lateral roller supporting telescopic arm 31, a reaction force supporting rod 32, a lateral roller 33, a lateral roller motor 34, a lateral roller connecting arm 35, a force-bearing rotating rod 36, a lateral roller device rotating shaft 37 and a lateral pressure sensor 38; Figure 3 The vertical walking element 1 shown includes a vertical roller 11 , a vertical roller rotating shaft 12 , a vertical roller device rotating motor 13 , a vertical roller telescopic arm 14 , a vertical pressure sensor 15 and a vertical roller motor 16 .

[0044] In this embodiment, if Figure 1 and Figure 3 As shown, the vertical walking element 1 is distributed on the upper and lower sides of the arc-shaped support arm 7, with a total of 12 vertical walking parts; the vertical roller 11 in the vertical walking element 1 is covered with a patterned sponge leather cover on the outer layer, which can ensure sufficient friction with the concrete pole during the travel; the vertical roller rotating shaft 12 is controlled by the vertical roller device rotating motor 13, ensuring that the vertical roller telescopic arm 14 can be rotated through remote operation; the vertical roller telescopic arm 14 can be controlled by the remote operation system to ensure that it can be extended and retracted during the travel or leveling process. The vertical pressure sensor 15 can monitor the pressure of each vertical roller 11 in real time, and transmit the monitored pressure data to the remote operation system; the vertical roller motor 16 can control the power of the vertical roller 11, and operate the travel and braking of the roller through the remote operation system through the wireless communication module.

[0045] In this embodiment, if Figure 1 and Figure 2As shown, the lateral walking element 3 is distributed in the middle of the four arc-shaped support arms 7, with a total of four horizontal walking parts; similarly, the outer layer of the lateral roller 33 in the lateral walking element 3 is covered with a patterned sponge leather cover to ensure the friction between the concrete pole; the lateral roller support telescopic arm 31 is controlled by the remote control system and can be telescoped as required; the reaction force support rod 32 is arranged in the middle of the arc-shaped support arm 7, which can not only support the lateral walking element 3, but also ensure the stability of the overall structure of the intelligent detection device. The reaction force support rod 32 and the force-bearing rotating rod 36 jointly complete the travel support of the lateral walking element 3; the lateral roller motor 34 drives the lateral roller 33 to travel and brake, which is controlled by the remote operation system.

[0046] like Figure 1 As shown, the balancing connection element 2 is arranged on both sides of the four arc-shaped support arms 7, and is used to connect the support arms to form a whole, and has a certain mass, which is used to balance the balance of the entire shooting device so that it remains horizontal on the initial horizontal plane; the control element 5 is equipped with programs and batteries, and is used to perform wireless transmission operations and remote system operations, which can ensure that each motor and pressure sensor operates independently, outputs information and feeds back to the remote operation platform. The remote operation system can operate and control each motor and each telescopic rod through the control element 5. The electronic level monitoring element 6 detects the horizontal state of the entire process of detection by the intelligent detection device, and feeds back the detection result data to the remote operation platform through the control element 5; it is equipped with three image acquisition mechanisms 8 to acquire images, and can shoot the same cross section of the concrete pole, and can transmit the acquired photos to the remote operation system in real time through the control element 5 to perform the next step of image processing.

[0047] In this embodiment, if Figure 1 and Figure 4 As shown, the image acquisition mechanism 8 includes an image acquisition support arm 81, an image acquisition telescopic arm 82 and an image acquisition camera 83; the image acquisition mechanism 8 is fixed on the arc-shaped support arm 7 by the image acquisition support arm 81, and the position movement of the image acquisition camera 83 is achieved by the extension and contraction of the image acquisition telescopic arm 82.

[0048] It should be noted that a motor is arranged inside the rotating spring shaft pin 4, and the arc support arm 7 is adjusted based on the rotating spring shaft pin 4 through the internal motor, so that the arc support arm 7 is clamped on the concrete pole, ensuring that the intelligent detection device crawls stably on the concrete pole.

[0049] In this embodiment, the horizontal monitoring element 6 is used to monitor the horizontal level, and the data is transmitted to the dynamic balancing system for feedback. The relevant motors in the walking mechanism are controlled according to the horizontal data to adjust the relevant telescopic arms and the relevant rotating shafts to keep them horizontal. When the pressure sensor data in the walking mechanism increases or decreases suddenly, the horizontal monitoring element 6 will shake, and the telescopic arms and rotating shafts will be adjusted in the same way. After each adjustment, the horizontal monitoring element 6 will continue to detect the horizontal level of the device. If it is still not horizontal, it will continue to control the telescopic arms and rotating shafts for leveling operations; if it is horizontal, it will wait for the next horizontal detection.

[0050] In this embodiment, the traveling mechanism can control the traveling and stopping of the intelligent detection device, which is divided into vertical climbing and descending and horizontal rotation. The vertical traveling element 1 and the horizontal traveling element 3 do not work at the same time. When the vertical traveling element 1 is climbing and descending, the horizontal roller supporting telescopic arm 31 in the horizontal traveling element 3 is retracted to ensure the normal operation of the vertical traveling element 1. When the lateral walking element 3 rotates, the lateral roller supporting telescopic arm 31 extends, and the force is transmitted to the force-bearing rotating rod 36 through the support of the reaction force support rod 32. The force-bearing rotating rod 36 rotates around the lateral roller device rotating axis 37, so that the lateral roller 33 contacts and sticks to the concrete pole. When all the lateral walking elements 3 support the entire shooting device, the vertical walking element 1 rotates upward and downward respectively, so that the vertical walking element 1 is separated from the surface of the concrete pole; the lateral roller motor 34 is controlled to rotate 60o through remote control operation; when the rotation is completed, the vertical walking element 1 is restored to the vertical moving state through the vertical roller motor 16 through the remote control system, and the lateral walking element is retracted at the same time; the obstacle avoidance function of the intelligent detection device can be realized through this process.

[0051] The intelligent detection device for concrete pole defects based on machine vision introduced in this embodiment can realize the following functions:

[0052] (1) When concrete spalling occurs in a certain area, the intelligent detection device will suddenly become unstable, and the data of the pressure sensor related to the walking mechanism will suddenly change. After the horizontal monitoring element 6 detects the data change, it will perform intelligent adjustments on the device to ensure the horizontal state of the intelligent detection device;

[0053] (2) The wear of the horizontal roller 33 and the vertical roller 11 during the travel process may cause the intelligent detection device to deviate; therefore, when a large deviation is detected, the intelligent detection device will adjust the posture in time; at the same time, the intelligent detection device will also be checked regularly to ensure the safety and reliability of the shooting work and to ensure that the image is not distorted due to shaking;

[0054] (3) When there is an obstacle on the concrete pole, the intelligent detection device is rotated by the lateral walking element 3 to achieve intelligent obstacle avoidance.

[0055] In order to ensure the normal operation of concrete poles, this embodiment detects and identifies defects in concrete poles; the surface image of the concrete pole can be obtained through an intelligent detection device, and accurate images can be obtained efficiently to improve detection accuracy and reliability.

[0056] Embodiment 2

[0057] The second embodiment of the present invention introduces a method for intelligently detecting defects of concrete poles based on machine vision. The method adopts the intelligent device for intelligently detecting defects of concrete poles based on machine vision introduced in the first embodiment.

[0058] A method for intelligent detection of concrete pole defects based on machine vision, comprising:

[0059] The intelligent detection device is installed on the concrete pole, and the intelligent detection device is controlled to move horizontally on the concrete pole through the control element and the horizontal monitoring element;

[0060] During the traveling process, the image of the concrete pole is acquired through the image acquisition mechanism and the horizontal monitoring element;

[0061] Processing the acquired concrete pole image to obtain a concrete pole image feature value;

[0062] Defect classification is performed based on the obtained concrete pole image feature values ​​to complete the defect detection of the concrete pole.

[0063] The process of the intelligent detection method for concrete pole defects based on machine vision in this embodiment is as follows: Figure 5 As shown, it includes: image acquisition—image collection—image preprocessing—feature value extraction—crack classification.

[0064] In the process of image acquisition in this embodiment, the concrete pole to be inspected is first determined, and the intelligent detection device introduced in the first embodiment is stably placed at the bottom of the concrete pole in a horizontal state. After the intelligent detection device is started, the image of the concrete pole is acquired in combination with the image acquisition mechanism. In the process of image acquisition, the intelligent detection device first moves upward for a fixed distance, and stops after the fixed distance, that is, the intelligent detection device is in a stationary state, and then the image is acquired based on the image acquisition mechanism to prevent image distortion caused by the jitter during movement; when it moves to the top of the pole, the horizontal roller is started, the vertical roller is retracted, and the horizontal roller is rotated 60o horizontally. After the rotation, the horizontal roller is retracted, and the vertical roller is started to descend. In the same descent process, shooting is performed at fixed distances, and the image acquisition mechanism always maintains dynamic balance during the upward or downward movement.

[0065] like Figure 6As shown, the image acquisition mechanism requires the cooperation of the wireless communication module and the motor drive module during the image acquisition process; the wireless communication module can transmit images and feedback information, including a camera module, a horizontal module and a pressure sensor module; the motor drive module can be remotely controlled to move the position of the device and the state of the rod, including a leveling module, a telescopic arm module and a travel module; the horizontal module, the pressure sensor module, the leveling module and the telescopic arm module together constitute a dynamic balance system.

[0066] This embodiment adopts Figure 7 The U-Net encoder-decoder network structure shown in the figure performs image preprocessing. The encoder extracts feature values ​​through convolution, pooling, etc., and gradually reduces the input dimension. According to the information provided by the encoder, the decoder repairs detailed features through multi-scale feature fusion, upsampling, etc., and obtains higher accuracy.

[0067] In the U-Net network structure, the convolution layer contains a set of learnable convolution kernels, and uses a full zero padding method to retain the image boundary information to control the size of the output image, achieve weight sparsity and weight sharing, and the pooling layer selects the maximum value of the 2×2 non-overlapping area in each channel and outputs it to the next layer. The effect of the transposed convolution is opposite to that of the convolution. When the step size of the transposed convolution is greater than 1, padding is required in the middle of the image.

[0068] In this embodiment, the batch normalization layer normalizes the current batch data and limits the output of each layer to a fixed distribution range, which speeds up the network training and improves the adaptability of the network. For a set of training batches B[x1,x 2,… x m ]. Batch normalization is: Among them, μ B Represents the mean of the current training batch; in, Represents the variance of the current training batch; in, is the normalized result after subtracting the mean and variance of the current training batch; Among them, y i is the result of scaling and translation, and the scaling factor γ and translation coefficient β are the parameters that need to be learned.

[0069] The activation function in this embodiment uses a ReLU function with unilateral suppression output and sparse activation of the underlying network, which has a faster convergence speed than other activation functions, that is, ReLU=max(0,x).

[0070] In this embodiment, the projection integral method is used to extract the crack feature value. Each feature of the image is the arithmetic mean of pixels in a specific row or column. The image with a perfect longitudinal crack has a constant projection integral along a specific direction: the horizontal axis has a constant value, while there is an intensity peak on the vertical axis; for images with perfect transverse cracks, this phenomenon is the opposite; for images with crazing cracks and images without cracks, the projection integrals of the two axes tend to be stable. However, the response average value in the case of crazing cracks is higher than the response average value in the case of no cracks.

[0071] The features used in this embodiment include 8 features that characterize projection integrals and 3 crack object features: the minimum value of the projection integral on the horizontal axis and vertical axis of the image; the maximum value of the projection integral on the horizontal axis and vertical axis of the image; the average value of the projection integral on the horizontal axis and vertical axis of the image; the standard deviation of the projection integral on the horizontal axis and vertical axis of the image; the number of crack objects; the average value of the directions of crack objects; and the standard deviation of the directions of crack objects.

[0072] The vector machine (SVM) method used in this embodiment requires setting a penalty constant (C) and a kernel function parameter (r). A grid search process can be used to appropriately select the values ​​of these hyperparameters of the SVM.

[0073] The images are classified into no cracks, transverse cracks, longitudinal cracks and cracks by the SVM method. Based on the classified images, the images without cracks are removed, and the remaining images are all crack images; based on the location and degree of damage of the cracks, maintenance and repair work on the concrete poles can be carried out, which greatly improves work efficiency.

[0074] The detailed steps are the same as the working principles of the intelligent detection device for concrete pole defects based on machine vision provided in Example 1, and will not be repeated here.

[0075] The above description is only a preferred embodiment of the present embodiment and is not intended to limit the present embodiment. For those skilled in the art, the present embodiment may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present embodiment shall be included in the protection scope of the present embodiment.

Claims

1. An intelligent detection device for concrete pole defects based on machine vision, characterized in that: It includes an annular support mechanism and a walking mechanism and an image acquisition mechanism arranged on the annular support mechanism; wherein the annular support mechanism is open, including a first arc-shaped support element, a second arc-shaped support element and a rotating spring axle pin for connecting the first arc-shaped support element and the second arc-shaped support element; the walking mechanism includes a vertical walking element, a horizontal walking element and a horizontal monitoring element, and the image acquisition mechanism is provided with an image acquisition telescopic arm; the horizontal state of the walking mechanism is detected by the horizontal monitoring element, and the vertical walking element and the image acquisition telescopic arm are adjusted in real time according to the detection result of the horizontal state of the walking mechanism to maintain the horizontal state of the intelligent detection device; based on the cooperation of the vertical walking element and the horizontal walking element, the intelligent obstacle avoidance of the intelligent detection device is realized.

2. The intelligent detection device for concrete pole defects based on machine vision as claimed in claim 1, characterized in that: The vertical walking element and the transverse walking element do not work at the same time. When the vertical walking element climbs or descends, the transverse roller support telescopic arm in the transverse walking element contracts. When the vertical walking element encounters an obstacle during walking, the vertical roller telescopic arm in the vertical walking element contracts, the vertical walking element stops walking, and the transverse roller support telescopic arm extends. The force is transmitted to the force-bearing rotating rod in the transverse walking element through the support of the reaction force support rod in the transverse walking element. Under the action of the force-bearing rotating rod, the rotating shaft of the transverse roller device in the transverse walking element rotates, so that the transverse rollers in the transverse walking element contact and adhere to the concrete pole until all the transverse rollers in the transverse walking elements are attached to the surface of the concrete pole; after the vertical walking element rotates and separates from the surface of the concrete pole, the transverse roller motor in the transverse walking element is controlled to rotate, driving the transverse walking element to rotate horizontally on the surface of the concrete pole. After the horizontal rotation is completed, after the transverse walking element rotates and separates from the surface of the concrete pole, the vertical walking element is controlled to continue walking, completing the intelligent obstacle avoidance of the intelligent detection device.

3. The intelligent detection device for concrete pole defects based on machine vision as claimed in claim 1, characterized in that: The first arc-shaped support element or the second arc-shaped support element comprises two arc-shaped support arms with completely identical structures; the arc-shaped support arm in the first arc-shaped support element and the arc-shaped support arm in the second arc-shaped support element are connected via the rotating spring axle pin, and the clamping of the arc-shaped support arm on the concrete pole is achieved by adjusting the rotating spring axle pin.

4. The intelligent detection device for concrete pole defects based on machine vision as claimed in claim 1, characterized in that: During the process of the intelligent detection device climbing or descending, the horizontal state of the walking mechanism is monitored in real time based on the horizontal electronic monitor, and the motion state of the vertical walking element is adjusted according to the horizontal state detection result. The horizontal state of the walking mechanism is adjusted by adjusting the motion state of different vertical rollers in the vertical walking element to keep the walking mechanism continuously level.

5. The intelligent detection device for concrete pole defects based on machine vision as claimed in claim 1, characterized in that: During the process of static image acquisition by the intelligent detection device, the length of the image acquisition telescopic arm is adjusted according to the obtained horizontal state detection result of the walking mechanism, so that the concrete pole images acquired by the image acquisition mechanism are all in the same horizontal plane.

6. The intelligent detection device for concrete pole defects based on machine vision as claimed in claim 1, characterized in that: The outer sides of the vertical rollers in the vertical traveling element and the transverse rollers in the transverse traveling element are both covered with patterned leather covers, which increase the friction between the vertical rollers or the transverse rollers and the concrete poles during travel.

7. The intelligent detection device for concrete pole defects based on machine vision as claimed in claim 1, characterized in that: It also includes a control element that is respectively connected to the annular support mechanism, the walking mechanism and the image acquisition mechanism for communicating with each other. The control element is used to monitor the working status information of the annular support mechanism, the walking mechanism and the image acquisition mechanism in real time and perform feedback adjustment of the intelligent detection device according to the obtained working status information.

8. The intelligent detection device for concrete pole defects based on machine vision as claimed in claim 1, characterized in that: A vertical pressure sensor and a lateral pressure sensor are respectively provided in the vertical walking element and the lateral walking element. When the walking mechanism is moving, the pressure state of the walking mechanism is monitored in real time by the vertical pressure sensor and the lateral pressure sensor, and the horizontal state of the walking mechanism is judged in combination with the horizontal monitoring element. Feedback adjustment of the horizontal state of the intelligent detection device is performed according to the detection result of the horizontal monitoring element.

9. A method for intelligent detection of concrete pole defects based on machine vision, using the intelligent detection device for concrete pole defects based on machine vision as claimed in any one of claims 1 to 8, characterized in that: include: The intelligent detection device is installed on the concrete pole, and the intelligent detection device is controlled to move horizontally on the concrete pole through the control element and the horizontal monitoring element; During the traveling process, the image of the concrete pole is acquired through the image acquisition mechanism and the horizontal monitoring element; Processing the acquired concrete pole image to obtain a concrete pole image feature value; Defect classification is performed based on the obtained concrete pole image feature values ​​to complete the defect detection of the concrete pole.

10. The method for intelligent detection of concrete pole defects based on machine vision as claimed in claim 9, characterized in that: The image defects of the concrete poles at least include no cracks, transverse cracks, longitudinal cracks and cracks.

Images