Automatic reinforcing steel bar binding detection method and system based on machine vision

Through automatic detection methods based on machine vision, image processing and neural network technology, the traditional steel bar binding detection methods are solved inefficient, poor accuracy and safety hazards, and efficient, accurate and safe steel bar binding quality inspection is achieved, and the construction of efficient, accurate and safe steel bar binding quality inspection is achieved. The project provides quality control assurance.

CN118941500AInactive Publication Date: 2024-11-12DEZHOU DINGWANG TIANCHENG CONSTR ENG CO LTD

Patent Information

Application Number
CN202410920258.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-10
Publication Date
2024-11-12
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Traditional steel bar binding detection methods are inefficient, have poor accuracy, have safety hazards and are costly, making it difficult to meet the needs of large-scale projects.

Method used

Using automatic detection method based on machine vision, through image acquisition, preprocessing, image segmentation and edge detection algorithms, combined with pre-trained neural network model, the characteristic analysis of steel bars and binding points is performed to determine whether preset conditions are met, and visualization is generated Test report.

Benefits of technology

Efficient, accurate and safe quality inspection of steel bar binding is achieved, reducing the influence of human factors, reducing the inspection cost, and improving the consistency and safety of the inspection results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118941500A_ABST
    Figure CN118941500A_ABST
Patent Text Reader

Abstract

The invention discloses a reinforcing steel bar binding automatic detection method and system based on machine vision, and the method comprises the steps: collecting an image of a to-be-detected target region, and carrying out the preprocessing; image features of the reinforcing steel bars and binding points are extracted from the preprocessed image through an image segmentation and edge detection algorithm; based on a pre-trained neural network recognition model, analyzing and processing the image features, and judging whether the binding points meet preset conditions or not; the detection result is displayed in a visual mode, and a detection report is generated. According to the automatic detection method, the detection time can be greatly shortened, and the requirements of large-scale engineering are met; by using machine learning and computer vision technologies, the influence of human factors is reduced, and the consistency and accuracy of detection results are improved. In addition, due to the application of the method, manual operation in high altitude and dangerous environments can be reduced, the safety risk is reduced, and the detection cost is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of machine vision and steel bar binding detection, and more particularly to a machine vision-based steel bar binding automatic detection method and system. Background Art

[0002] At present, steel bar binding is a key step in construction projects, and its quality is directly related to the stability and safety of the building structure. Traditional steel bar binding inspection methods mainly rely on manual inspection, which often results in loose binding or non-compliance with specifications, which may cause structural problems of the building and even endanger personnel safety. Generally, there are the following defects and disadvantages:

[0003] Inefficiency: Manual inspection is time-consuming and cannot meet the needs of large-scale projects.

[0004] Poor accuracy: Manual testing is easily affected by human factors, such as fatigue and lack of experience, which leads to inconsistent test results.

[0005] Safety hazards: Manual inspection at high altitudes or in dangerous environments presents significant safety risks.

[0006] High cost: It requires a lot of human resources, which increases the cost of the project.

[0007] In order to overcome the above problems, there is an urgent need for an automatic steel bar binding detection method and / or system to achieve efficient, accurate and safe steel bar binding quality detection. Summary of the invention

[0008] In view of this, the present invention provides a method and system for automatic detection of steel bar binding based on machine vision, which can effectively solve the defects and disadvantages of traditional steel bar binding detection methods, realize efficient, accurate and safe steel bar binding quality detection, and provide strong guarantee for quality control of construction projects.

[0009] In order to achieve the above object, the present invention adopts the following technical solution:

[0010] In a first aspect, an embodiment of the present invention provides a method for automatic detection of steel bar binding based on machine vision, comprising the following steps:

[0011] S1, collect images of the target area to be detected and perform preprocessing;

[0012] S2. Extracting the image features of the steel bars and binding points through image segmentation and edge detection algorithms for the preprocessed images;

[0013] S3, based on the pre-trained neural network recognition model, analyzing and processing the image features to determine whether the binding point meets the preset conditions;

[0014] S4. Display the test results in a visual way and generate a test report.

[0015] Furthermore, the step S1 comprises:

[0016] Collect the image of the target area to be detected and perform the following processing in sequence:

[0017] 1) Use median filtering and Gaussian filtering to denoise the image and remove the noise in the image;

[0018] 2) Color space transformation is used to detect the shadow area of ​​the steel bars caused by illumination and remove the shadow area;

[0019] 3) grayscale the image;

[0020] 4) Perform image enhancement and geometric correction processing.

[0021] 3. According to the method of automatic detection of steel bar binding based on machine vision in claim 2, it is characterized in that the image is denoised using median filtering and Gaussian filtering to remove noise in the image, including:

[0022] Median filter formula:

[0023] I f1 (x, y)=median{I(x+i,y+j)|-k≤i,j≤k}

[0024] I(x, y) represents the pixel value of the input image at position (x, y); k represents half the size of the filter window, and the size of the filter window is (2k+1)*(2k+1); I f1 (x, y) represents the pixel value of the image at position (x, y) after median filtering; i, j represent the offset in the horizontal and vertical directions;

[0025] Gaussian filter formula:

[0026]

[0027] Among them, I f2 (x, y) represents the pixel value of the image at position (x, y) after Gaussian filtering; K represents half the size of the Gaussian kernel, the kernel size is (2K+1)*(2K+1), and the calculation formula of the Gaussian kernel G(i, j) is:

[0028]

[0029] σ represents the standard deviation of the Gaussian kernel and determines the degree of smoothing.

[0030] Furthermore, the color space is transformed to detect the steel bar shadow area caused by the illumination and remove the shadow area, including:

[0031] Convert the denoised image from RGB color space to HSV color space, separate the brightness channel, and use threshold processing to detect shadow areas;

[0032] V = max(R, G, B)

[0033]

[0034] In the formula, Shadow Mask represents the shadow mask, marking the shadow area; T represents the threshold, which is between 0.3 and 0.5; V represents the brightness channel in the HSV color space;

[0035] The CLAHE method is used to enhance the brightness channel V;

[0036] I enhanced (x, y) = CLAHE (I (x, y))

[0037] I enhanced (x, y) represents the pixel value of the enhanced image at position (x, y); I(x, y) represents the pixel value of the input denoised image at position (x, y);

[0038] Replace the enhanced brightness channel back into the original image and convert back to RGB color space;

[0039] The enhanced image is fused with the original image to obtain the final image with the shadow removed;

[0040] I final (x, y) = α·I enhanced (x, y) + (1-α) · I (x, y)

[0041] α represents the fusion factor, which is between 0.5 and 0.7.

[0042] Furthermore, the step S2 comprises:

[0043] S21, using the Canny edge detection algorithm to extract edge information of steel bars and binding points;

[0044] S22. Using Hough transform to detect the straight line characteristics of the steel bars, and determine the distribution and direction of the steel bars;

[0045] S23, performing preliminary segmentation of the image by using K-means clustering algorithm to separate the steel bars and binding points;

[0046] S24, and use an adaptive threshold segmentation algorithm to perform fine segmentation to extract the location, size and shape features of the binding points.

[0047] Furthermore, the step S3 comprises:

[0048] Use a pre-trained convolutional neural network model to classify and judge the extracted features;

[0049] The support vector machine is used for secondary classification, and the test results are compared to analyze whether the steel bar binding complies with the specifications.

[0050] Furthermore, the step S4 comprises:

[0051] The test results are displayed through a graphical interface, marking qualified and unqualified lashing points;

[0052] Generate a detailed inspection report including inspection time, inspection results, problem description, improvement suggestions and construction instructions.

[0053] In a second aspect, an embodiment of the present invention further provides a steel bar binding automatic detection system based on machine vision, comprising:

[0054] An image acquisition and processing module is used to acquire images of the target area to be detected and perform preprocessing;

[0055] The feature extraction module is used to extract the image features of the steel bars and binding points from the preprocessed images through image segmentation and edge detection algorithms;

[0056] The lashing detection module analyzes and processes the image features based on a pre-trained neural network recognition model to determine whether the lashing point meets the preset conditions;

[0057] The report generation module is used to display the test results in a visual way and generate a test report.

[0058] In a third aspect, an embodiment of the present invention further provides a computer device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus;

[0059] Memory, used to store computer programs;

[0060] The processor, when used to execute the program stored in the memory, can implement a method for automatic detection of steel bar binding based on machine vision as described in any one of the first aspects.

[0061] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having computer program instructions stored thereon, and when the computer program instructions are executed by a processor, the processor executes a method for automatic detection of steel bar binding based on machine vision as described in any one of the first aspects.

[0062] It can be seen from the above technical solution that compared with the prior art:

[0063] 1. Improve detection efficiency: The automated detection method of the present invention can significantly shorten the detection time and meet the needs of large-scale projects.

[0064] 2. Improve detection accuracy: The present invention uses machine learning and computer vision technology to reduce the impact of human factors and improve the consistency and accuracy of detection results.

[0065] 3. Ensure operational safety: The application of the present invention can reduce manual work in high altitudes and dangerous environments and reduce safety risks.

[0066] 4. Reduce costs: The application of the present invention can reduce human resource investment, reduce detection costs, and has higher economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] 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 or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0068] Figure 1 A flow chart of the automatic detection method for steel bar binding based on machine vision provided by the present invention.

[0069] Figure 2 This is a flow chart of step S2 provided by the present invention.

[0070] Figure 3 A block diagram of the automatic detection system for steel bar binding based on machine vision provided by the present invention.

[0071] Figure 4 This is a structural block diagram of the computer device provided by the present invention. DETAILED DESCRIPTION

[0072] 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.

[0073] Embodiment 1:

[0074] The provisions on steel bar binding in Chapter 4 of the Construction Manual are as follows:

[0075] 4.1 Binding of steel mesh: The intersection points of the two rows of steel bars on all sides should be bound at each point, the middle part can be staggered and bound at intervals, and the steel mesh of the two-way main reinforcement must be bound point by point.

[0076] 4.2 When tying steel bars, the wire buckles at adjacent tying points should be in an eight-shape to prevent deformation of the steel mesh, and the hooks should face upwards; the hooks of the upper steel bars of the double-layer steel bars should face downwards.

[0077] 4.3 After the reinforcement is tied, a careful inspection should be carried out to ensure that the position and transverse spacing are within the preset range.

[0078] Based on the above specification, the embodiment of the present invention discloses a method for automatic detection of steel bar binding based on machine vision, referring to Figure 1 As shown, the following steps are included:

[0079] S1, collect images of the target area to be detected and perform preprocessing;

[0080] S2. Extracting the image features of the steel bars and binding points through image segmentation and edge detection algorithms for the preprocessed images;

[0081] S3, based on the pre-trained neural network recognition model, analyzing and processing the image features to determine whether the binding point meets the preset conditions;

[0082] S4. Display the test results in a visual way and generate a test report.

[0083] In specific implementation, the method can be developed into a mobile application (APP) and loaded on a smartphone. Supervisors can use a smartphone with the APP installed to take images of the steel bar binding area that needs to be inspected, and the system can automatically analyze the image and output the inspection results. In addition, the method can also be embedded and integrated in a dedicated inspection terminal, which is suitable for on-site automatic inspection equipment. It can be easily applied on actual construction sites, greatly facilitating the work of supervisors and improving the level of construction quality management.

[0084] In another embodiment, the image of the target area to be detected can be acquired in real time through the on-site monitoring probe, and then transmitted to the computer in the central control room for preprocessing and analysis, so as to finally output the detection result.

[0085] The following is a detailed description of each of the above steps:

[0086] In step S1, for example, an image of the target area to be detected is captured using a smart phone or an embedded detection terminal, or an image is acquired through real-time monitoring by a monitoring probe, and then image preprocessing is performed on the terminal or server side. When acquiring images through an on-site monitoring probe, for example, a high-resolution industrial camera (with a resolution of at least 5 million pixels) can be used in combination with a suitable light source to ensure that the acquired image has sufficient clarity and contrast.

[0087] When the resolution of a smartphone camera is low, a super-resolution reconstruction algorithm based on deep learning can also be used to improve the detail clarity of the low-resolution image to facilitate subsequent image detail processing.

[0088] For example, the deep learning network can be SRCNN or ESRGAN, which can realize low-resolution image input, extract image features through a series of convolutional layers and activation functions, and perform multi-scale feature fusion and enhancement, which can effectively improve the detail clarity of low-resolution images and provide higher quality image data for subsequent image processing.

[0089] The collected images are processed in the following order:

[0090] 1) Denoising: Use median filtering and Gaussian filtering to remove noise from the image.

[0091] Median filter formula:

[0092] I f1 (x, y)=median{I(x+i,y+j)|-k≤i,j≤k} (1)

[0093] I(x, y) represents the pixel value of the input image at position (x, y); k represents half the size of the filter window, and the size of the filter window is (2k+1)*(2k+1); I f1 (x, y) represents the pixel value of the image at position (x, y) after median filtering; i, j represent the offset in the horizontal and vertical directions;

[0094] Gaussian filter formula:

[0095]

[0096] Among them, I f2 (x, y) represents the pixel value of the image at position (x, y) after Gaussian filtering; K represents half the size of the Gaussian kernel, the kernel size is (2K+1)*(2K+1), and the calculation formula of the Gaussian kernel G(i, j) is:

[0097]

[0098] σ represents the standard deviation of the Gaussian kernel and determines the degree of smoothing.

[0099] Furthermore, in this embodiment, in addition to the above two denoising methods, bilateral filtering and / or non-local mean filtering may be used to remove high-frequency noise in the image and retain the image edge information.

[0100] 2) Shadow removal: Detect and remove the steel bar shadow area caused by lighting through color space transformation.

[0101] Convert the denoised image from RGB color space to HSV color space, separate the brightness channel, and use threshold processing to detect shadow areas;

[0102] V = max(R, G, B)

[0103]

[0104] Where Shadow_Mask represents the shadow mask, marking the shadow area; T represents the threshold, which is between 0.3 and 0.5; V represents the brightness channel in the HSV color space;

[0105] The CLAHE method is used to enhance the brightness channel V;

[0106] I enhanced (x, y) = CLAHE (I (x, y)) (5)

[0107] I enhanced (x, y) represents the pixel value of the enhanced image at position (x, y); I(x, y) represents the pixel value of the input denoised image at position (x, y);

[0108] Replace the enhanced brightness channel back into the original image and convert back to RGB color space;

[0109] The enhanced image is fused with the original image to obtain the final image with the shadow removed;

[0110] I final (x, y) = α·I enhanced (x,y)+(1-α)·I(x,y) (6)

[0111] α represents the fusion factor, which is between 0.5 and 0.7.

[0112] 3) Grayscale processing: Convert the image into a grayscale image to simplify subsequent processing steps.

[0113] I gray =0.299R+0.587G+0.114B

[0114] Among them, R, G and B are the pixel values ​​of the red, green and blue channels in the image respectively.

[0115] 4) Image enhancement and geometric correction:

[0116] At this time, the adaptive histogram equalization technology is used to enhance the image contrast, improve the visibility of the steel bars and binding points, and achieve image enhancement. The formula is as follows:

[0117]

[0118] I min and I max are the minimum and maximum pixel values ​​in the local window, respectively;

[0119] Then, geometric correction is performed to eliminate image distortion and ensure the accuracy of the image; for example, perspective transformation is used to eliminate distortion caused by shooting angle.

[0120] In step S2, refer to Figure 2 As shown, specifically including:

[0121] S21. Use the Canny edge detection algorithm to extract the edge information of steel bars and binding points; including: smoothing the image based on Gaussian filtering, calculating the image gradient, performing non-maximum suppression, and finally performing double threshold detection and edge connection.

[0122] S22, using Hough transform to detect the straight line characteristics of the steel bars and determine the distribution and direction of the steel bars;

[0123] S23. Perform preliminary segmentation of the image by using the K-means clustering algorithm to separate the steel bars and the binding points, so as to improve the separation effect of the steel bars and the binding points. The objective function of K-means is:

[0124]

[0125] Among them, C i is the i-th cluster, μ i is the center of the ith cluster.

[0126] S24, and use the adaptive threshold segmentation algorithm to perform fine segmentation to extract the location, size and shape characteristics of the binding points. The calculation formula of the adaptive threshold is:

[0127]

[0128] Among them, μ(x, y) and σ(x, y) are the mean and standard deviation of the local window respectively, R is the dynamic range of the image, and k is the adjustment parameter.

[0129] In step S3, it specifically includes:

[0130] Use a pre-trained convolutional neural network model to classify and judge the extracted features;

[0131] And combined with support vector machine for secondary classification to improve the detection accuracy; compare the test results to analyze whether the steel bar binding complies with the specifications.

[0132] For example, a convolutional neural network (CNN) is used to classify the extracted features and determine whether the binding points meet the specifications. The convolution layer and pooling layer are used to extract the deep features of the image. The convolution operation formula is:

[0133]

[0134] in, is the output of the kth convolution kernel at position (i, j), I (i+m)(j+n) is the pixel value of the input image at position (i+m,j+n), is the weight of the kth convolution kernel, b k is the bias term. Use the fully connected layer to map the convolutional features to the classification space, and finally output the classification results through the Softmax function. The formula is:

[0135]

[0136] Among them, z j is the output of the fully connected layer; locate the detected tying points, compare them with the standard tying points, and analyze their deviations. According to the output of the neural network, determine whether the tying points meet the preset conditions (based on the above specification settings, such as whether the tying point is missing, whether the shape of the tying point does not meet the set area, and whether the steel bar spacing is within the set range). For example, a threshold can be defined. When the classification probability is higher than the threshold, the tying point is considered qualified, otherwise it is unqualified.

[0137] In this step, the dual classification method of convolutional neural network combined with support vector machine improves the accuracy of classification results and ensures the accuracy of binding point judgment.

[0138] In step S4, it specifically includes:

[0139] S41. Display the test results through a graphical interface, and mark qualified and unqualified binding points; for example, mark the identified steel bars and binding points on the original image, and distinguish qualified and unqualified binding points with different colors. For example, use green to mark qualified points and red to mark unqualified points.

[0140] S42. Generate a detailed inspection report, including inspection time, inspection results, problem description, improvement suggestions and construction instructions.

[0141] Suppose that at a construction site, the quality of steel bar binding in a certain area needs to be tested. The operator uses a mobile phone to take an image of the area and automatically detects it using the above method. First, the image is preprocessed to remove noise and shadows, and then the features of the steel bars and binding points are extracted. Next, the pre-trained convolutional neural network and support vector machine are used for feature classification and judgment. Finally, the test results are displayed in the form of a graphical interface, and a detailed test report is generated, indicating which binding points are unqualified and giving improvement suggestions. This method can greatly improve the efficiency and accuracy of steel bar binding quality detection, reduce human errors, and ensure construction quality.

[0142] The automatic steel bar binding detection method based on machine vision provided by the present invention can effectively solve the defects and disadvantages of traditional steel bar binding detection methods through the implementation of the above technical scheme, realize efficient, accurate and safe steel bar binding quality detection, and provide strong guarantee for the quality control of construction projects.

[0143] Embodiment 2:

[0144] Reference Figure 3 As shown, an embodiment of the present invention further provides a steel bar binding automatic detection system based on machine vision, comprising:

[0145] An image acquisition and processing module is used to acquire images of the target area to be detected and perform preprocessing;

[0146] The feature extraction module is used to extract the image features of the steel bars and binding points from the preprocessed images through image segmentation and edge detection algorithms;

[0147] The lashing detection module analyzes and processes the image features based on a pre-trained neural network recognition model to determine whether the lashing point meets the preset conditions;

[0148] The report generation module is used to display the test results in a visual way and generate a test report.

[0149] In this embodiment, the automatic detection system automatically detects the steel bar binding points through machine vision technology, which greatly reduces the time and labor intensity of manual inspection and significantly improves the detection efficiency. In addition, advanced image processing technologies such as median filtering, Gaussian filtering, edge detection and image segmentation algorithms are used to ensure accurate extraction of image features, thereby improving the accuracy of detection. The use of pre-trained neural network models for image feature analysis and judgment can effectively identify and distinguish qualified and unqualified binding points, thereby improving the accuracy of detection.

[0150] The machine vision-based automatic steel bar binding detection system provided in the embodiment of the present invention improves the detection efficiency and accuracy while ensuring the real-time and robustness of the system, providing an effective technical means for the quality management of construction projects.

[0151] Embodiment 3:

[0152] The embodiment of the present invention further provides a computer device, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0153] Memory, used to store computer programs;

[0154] The processor, when used to execute the program stored in the memory, can implement an automatic detection method for steel bar binding based on machine vision as in Example 1.

[0155] like Figure 4 As shown, the electronic device may include: a processor 41, a communications interface 42, a memory 43 and a communication bus 44, wherein the processor 41, the communications interface 42 and the memory 43 communicate with each other via the communication bus 44. The processor 41 may call the logic instructions in the memory 43 to execute a method for automatic detection of steel bar binding based on machine vision, the method comprising:

[0156] S1, collect images of the target area to be detected and perform preprocessing;

[0157] S2. Extracting the image features of the steel bars and binding points through image segmentation and edge detection algorithms for the preprocessed images;

[0158] S3, based on the pre-trained neural network recognition model, analyzing and processing the image features to determine whether the binding point meets the preset conditions;

[0159] S4. Display the test results in a visual way and generate a test report.

[0160] In addition, the logic instructions in the above-mentioned memory 43 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0161] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0162] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An automatic detection method for steel bar binding based on machine vision, It is characterized in that The following steps are involved: S1, collect images of the target area to be detected and perform preprocessing; S2. Extracting the image features of the steel bars and binding points through image segmentation and edge detection algorithms for the preprocessed images; S3, based on the pre-trained neural network recognition model, analyzing and processing the image features to determine whether the binding point meets the preset conditions; S4. Display the test results in a visual way and generate a test report.

2. According to the method for automatic detection of steel bar binding based on machine vision according to claim 1, It is characterized in that The step S1 comprises: Collect the image of the target area to be detected and perform the following processing in sequence: 1) Use median filtering and Gaussian filtering to denoise the image and remove the noise in the image; 2) Color space transformation is used to detect the shadow area of ​​the steel bars caused by illumination and remove the shadow area; 3) grayscale the image; 4) Perform image enhancement and geometric correction processing.

3. According to claim 2, a method for automatic detection of steel bar binding based on machine vision, It is characterized in that Use median filtering and Gaussian filtering to denoise the image and remove the noise in the image, including: Median filter formula: I f1 (x,y)=median{I(x+i,y+j)|-k≤i,j≤k} I(x,y) represents the pixel value of the input image at position (x,y); k represents half the size of the filter window, and the size of the filter window is (2k+1)*(2k+1); I f1 (x, y) represents the pixel value of the image at position (x, y) after median filtering; i, j represent the offset in the horizontal and vertical directions; Gaussian filter formula: Among them, I f2 (x, y) represents the pixel value of the image at position (x, y) after Gaussian filtering; K represents half the size of the Gaussian kernel, the kernel size is (2K+1)*(2K+1), and the calculation formula of the Gaussian kernel G(i, j) is: σ represents the standard deviation of the Gaussian kernel and determines the degree of smoothing.

4. According to claim 2, a method for automatic detection of steel bar binding based on machine vision, It is characterized in that Color space transformation is used to detect the shadow area of ​​steel bars caused by lighting and remove the shadow area, including: Convert the denoised image from RGB color space to HSV color space, separate the brightness channel, and use threshold processing to detect shadow areas; V = max(R, G, B) In the formula, Shadow Mask represents the shadow mask, marking the shadow area; T represents the threshold, which is between 0.3 and 0.5; V represents the brightness channel in the HSV color space; The CLAHE method is used to enhance the brightness channel V; Yo enhanced (x, y) = CLAHE(I(x, y)) I enhanced (x, y) represents the pixel value of the enhanced image at position (x, y); I(x, y) represents the pixel value of the input denoised image at position (x, y); Replace the enhanced brightness channel back into the original image and convert back to RGB color space; The enhanced image is fused with the original image to obtain the final image with the shadow removed; I final (x,y)=α·I enhanced (x,y)+(1-α)·I(x,y) α represents the fusion factor, which is between 0.5 and 0.

7.

5. According to the method for automatic detection of steel bar binding based on machine vision according to claim 1, It is characterized in that The step S2 comprises: S21, using the Canny edge detection algorithm to extract edge information of steel bars and binding points; S22. Using Hough transform to detect the straight line characteristics of the steel bars, and determine the distribution and direction of the steel bars; S23, performing preliminary segmentation of the image by using K-means clustering algorithm to separate the steel bars and binding points; S24, and use an adaptive threshold segmentation algorithm to perform fine segmentation to extract the location, size and shape features of the binding points.

6. According to the method of automatic detection of steel bar binding based on machine vision in claim 1, It is characterized in that The step S3 comprises: Use a pre-trained convolutional neural network model to classify and judge the extracted features; The support vector machine is combined with secondary classification to compare the test results and analyze whether the steel bar binding complies with the specifications.

7. According to claim 1, a method for automatic detection of steel bar binding based on machine vision, It is characterized in that The step S4 comprises: The test results are displayed through a graphical interface, marking qualified and unqualified lashing points; Generate a detailed inspection report including inspection time, inspection results, problem description, improvement suggestions and construction instructions.

8. An automatic detection system for steel bar binding based on machine vision, It is characterized in that include: An image acquisition and processing module is used to acquire images of the target area to be detected and perform preprocessing; The feature extraction module is used to extract the image features of the steel bars and binding points from the preprocessed images through image segmentation and edge detection algorithms; The lashing detection module analyzes and processes the image features based on a pre-trained neural network recognition model to determine whether the lashing point meets the preset conditions; The report generation module is used to display the test results in a visual way and generate a test report.

9. A computer device, It is characterized in that It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus; Memory, used to store computer programs; The processor, when used to execute the program stored in the memory, can implement a method for automatic detection of steel bar binding based on machine vision as described in any one of claims 1 to 7.

10. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores computer program instructions, which, when executed by a processor, enable the processor to execute a method for automatic detection of steel bar binding based on machine vision as described in any one of claims 1 to 7.

Citation Information

Patent Citations

  • Detection method for quality of electric power pipe gallery reinforcing steel bar mesh based on machine vision and detection device

    CN105956942A

  • Multi-feature weight adaptive shadow elimination method

    CN106373096A

  • Finger vein recognition method and system based on convolutional neural network and SIFT algorithm

    CN112597812A

  • Snow pressing vehicle appearance defect detection method based on vision under complex background

    CN113160192A

  • Binding position identification method and system for automatic binding of reinforcing steel bars

    CN116524391A

Cited By

  • Base plate steel bar binding auxiliary monitoring method and device based on cushion block identification and medium

    CN119146898A

  • Bridge engineering construction quality detection method and system

    CN119722683A

  • A bridge engineering construction quality detection method and system

    CN119722683B

  • Automatic steel bar binding machine based on machine vision

    CN121187170A

  • Liquid level detection method and system based on image sensor

    CN121391653A