Precast beam reinforcement protection layer thickness detection system based on big data

Abstract

The invention relates to the technical field of construction detection, in particular to a precast beam reinforcement protection layer thickness detection system based on big data, which comprises a monitoring station, a detection module and a transmission module, the thickness detection software comprises an image processing unit, a feature extraction unit and a monitoring judgment unit. The problems that an existing prefabricated beam steel bar protection layer thickness detection system generally adopts drilling sampling detection, the detection method is low in intelligent degree and inconvenient to operate can be effectively solved.

Description

Technical Field

[0001] The present invention relates to the technical field of construction detection, and in particular to a prefabricated beam reinforcement protective layer thickness detection system based on big data. Background Art

[0002] The steel bar protective layer is the part of concrete in concrete components that protects the steel bars and prevents them from being directly exposed. The steel bar protective layer is to prevent the steel bars from being corroded, which affects the durability and safety of the concrete structure. If the thickness of the protective layer is insufficient, the steel bars will be exposed to harmful substances such as air, moisture, and salt, causing the steel bars to rust and expand, thereby destroying the concrete and reducing the bearing capacity and service life of the structure; ensure that there is sufficient bonding and anchoring force between the steel bars and the concrete so that they can bear the load together with the steel bars. If the thickness of the protective layer is too large or too small, it will affect the stress state and force transmission effect of the steel bars in the concrete. If the thickness of the protective layer is too large, the concrete will be in a reinforcement-free state, and cracks will easily occur on the road surface. If the thickness of the protective layer is too small, the stress on the steel bars will be uneven and slip will occur easily.

[0003] The existing precast beam reinforcement protection layer thickness detection system still has some shortcomings. Specifically, the existing precast beam reinforcement protection layer thickness detection system generally adopts drilling sampling detection, the detection method has a low degree of intelligence and is inconvenient to operate.

[0004] Therefore, a prefabricated beam steel bar protective layer thickness detection system based on big data is needed to solve the problems raised in the above background technology. Summary of the invention

[0005] The purpose of the present invention is to provide a prefabricated beam reinforcement protective layer thickness detection system based on big data to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A prefabricated beam steel bar protective layer thickness detection system based on big data includes a monitoring station, a detection module and a transmission module. Thickness detection software is installed inside the monitoring station. The thickness detection software includes an image processing unit, a feature extraction unit and a monitoring and judgment unit.

[0008] As a preferred solution of the present invention, the monitoring station is connected to the transmission module via a wireless data network, and the detection module is connected to the transmission module via a wire.

[0009] As a preferred solution of the present invention, the detection module includes various monitoring cameras, the transmission module includes various wireless data transmitters, and the detection module is provided in multiple groups.

[0010] As a preferred solution of the present invention, the specific analysis steps of the image processing unit are: the detection module detects the protective layer, the transmission module inputs the detection data into the thickness detection software in the monitoring station, the image processing unit divides the monitoring data image into a non-edge area and an edge area, and determines the discrete degree value of the pixel value in each preset direction within a preset neighborhood of each pixel point in the monitoring data image. If the pixel point is a pixel point in the non-edge area, the pixel point is subjected to image filtering processing according to the discrete degree value of the pixel point in each preset direction. If the pixel point is a pixel point in the edge area, the edge direction is determined according to the discrete degree value and gradient value of the pixel point in each preset direction, and directional smoothing processing and edge enhancement processing are performed according to the determined edge direction. After the processing is completed, the image processing unit inputs the processed image into the feature extraction unit.

[0011] As a preferred solution of the present invention, the specific analysis steps of the feature extraction unit are: the feature extraction unit constructs a convolutional neural network, the feature extraction unit uses the convolutional neural network to perform preliminary feature extraction on the processed image to obtain a first feature image, the first feature image is divided into multiple parts in the spatial dimension, each part is globally pooled to obtain multiple first spatial features, and then the number of channels of the first feature image is amplified by the convolutional neural network to obtain a first amplified feature image after the number of channels is amplified, and then the first amplified feature image is globally pooled, and the pooling result is divided into multiple first channel features in the channel dimension, and feature fusion is performed based on multiple first channel features, multiple first spatial features and multiple first channel features to obtain image features of the processed image, and input them into the monitoring and judgment unit.

[0012] As a preferred solution of the present invention, the specific analysis steps of the monitoring and judgment unit are: the monitoring and judgment unit constructs a sparse projection matrix for image feature discrimination, the monitoring and judgment unit inputs the image features of the processed image and the image features of the training samples into the sparse projection matrix, the sparse projection matrix processes the image features, and compares the processing results with a preset threshold. If the preset threshold is exceeded, it means that the thickness of the protective layer does not meet the standard.

[0013] As a preferred solution of the present invention, the specific method for constructing a sparse projection matrix is: reconstructing the weight coefficients using the LLE-reconstruction weight method, reconstructing the projection matrix by minimizing the L2,1-norm regularized neighborhood, and performing L2,1-norm regularization on the projection matrix, and correspondingly obtaining a sparse projection matrix for directly extracting discriminant features of two-dimensional images.

[0014] As a preferred solution of the present invention, the specific method for the sparse projection matrix to process image features is: the sparse projection matrix performs normalization processing on the image features of the training samples to obtain a unified thickness image feature set, and the sparse projection matrix compares the image features of the processed image with the thickness image feature set to obtain feature similarity.

[0015] Compared with the prior art, the present invention has the following beneficial effects:

[0016] 1. In the present invention, the protective layer is detected by the detection module, the transmission module inputs the detection data into the thickness detection software in the monitoring station, the image processing unit divides the monitoring data image into a non-edge area and an edge area, determines the discrete degree value of the pixel value in each preset direction within the preset neighborhood of each pixel point in the monitoring data image, if the pixel point is a pixel point in the non-edge area, then according to the discrete degree value of the pixel point in each preset direction, the pixel point is subjected to image filtering processing, if the pixel point is a pixel point in the edge area, the edge direction is determined according to the discrete degree value and gradient value of the pixel point in each preset direction, and directional smoothing processing and edge enhancement processing are performed according to the determined edge direction, after the processing is completed, the image processing unit inputs the processed image into the feature extraction unit, the feature extraction unit constructs a convolutional neural network, the feature extraction unit uses the convolutional neural network to perform preliminary feature extraction on the processed image, obtains a first feature image, and divides the first feature image into multiple parts, globally pooling is performed on each part respectively to obtain multiple first spatial features, and then the number of channels of the first feature image is amplified by a convolutional neural network to obtain a first amplified feature image with amplified channels, and then the first amplified feature image is globally pooled, and the pooling result is divided into multiple first channel features in the channel dimension, and feature fusion is performed based on the multiple first channel features, the multiple first spatial features and the multiple first channel features to obtain image features of the processed image, and the image features are input into a monitoring and judgment unit, and the monitoring and judgment unit constructs a sparse projection matrix for image feature discrimination, and the monitoring and judgment unit inputs the image features of the processed image and the image features of the training samples into the sparse projection matrix, and the sparse projection matrix processes the image features, and compares the processing result with a preset threshold value. If the preset threshold value is exceeded, it means that the thickness of the protective layer does not meet the standard, and image feature extraction is used to detect whether the thickness of the protective layer is qualified. The detection method is relatively intelligent and the operation is convenient and fast. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the overall hardware structure of the present invention;

[0018] Figure 2 It is a schematic diagram of the system block structure of the present invention.

[0019] In the figure: 1. Monitoring station; 2. Detection module; 3. Transmission module; 4. Thickness detection software; 5. Image processing unit; 6. Feature extraction unit; 7. Monitoring and judgment unit. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0021] To facilitate understanding of the present invention, the present invention will be described more comprehensively below with reference to the relevant drawings. Several embodiments of the present invention are given. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

[0022] It should be noted that when an element is referred to as being "fixed on" another element, it may be directly on the other element or there may also be a central element. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may also be a central element. The terms "vertical", "horizontal", "left", "right" and similar expressions used in this article are for illustrative purposes only.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which the present invention belongs. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0024] For examples, see Figure 1-2 , the present invention provides a technical solution:

[0025] A precast beam steel bar protective layer thickness detection system based on big data includes a monitoring station 1, a detection module 2 and a transmission module 3. The monitoring station 1 is internally installed with thickness detection software 4, and the thickness detection software 4 includes an image processing unit 5, a feature extraction unit 6 and a monitoring judgment unit 7.

[0026] Furthermore, the monitoring station 1 is connected to the transmission module 3 via a wireless data network, and the detection module 2 is connected to the transmission module 3 via a wire.

[0027] Furthermore, the detection module 2 includes various surveillance cameras, the transmission module 3 includes various wireless data transmitters, and the detection module 2 is provided with multiple groups.

[0028] Furthermore, the specific analysis steps of the image processing unit 5 are as follows: the detection module 2 detects the protective layer, the transmission module 3 inputs the detection data into the thickness detection software 4 in the monitoring station 1, the image processing unit 5 divides the monitoring data image into a non-edge area and an edge area, and determines the discrete degree value of the pixel value in each preset direction within the preset neighborhood of each pixel point in the monitoring data image; if the pixel point is a pixel point in the non-edge area, the pixel point is subjected to image filtering processing according to the discrete degree value of the pixel point in each preset direction; if the pixel point is a pixel point in the edge area, the edge direction is determined according to the discrete degree value and gradient value of the pixel point in each preset direction, and directional smoothing processing and edge enhancement processing are performed according to the determined edge direction; after the processing is completed, the image processing unit 5 inputs the processed image into the feature extraction unit 6.

[0029] Furthermore, the specific analysis steps of the feature extraction unit 6 are as follows: the feature extraction unit 6 constructs a convolutional neural network, the feature extraction unit 6 uses the convolutional neural network to perform preliminary feature extraction on the processed image to obtain a first feature image, the first feature image is divided into multiple parts in the spatial dimension, each part is globally pooled to obtain multiple first spatial features, and then the convolutional neural network is used to amplify the number of channels of the first feature image to obtain a first amplified feature image after the number of channels is amplified, and then the first amplified feature image is globally pooled, and the pooling result is divided into multiple first channel features in the channel dimension, and feature fusion is performed based on multiple first channel features, multiple first spatial features and multiple first channel features to obtain image features of the processed image, and input them into the monitoring and judgment unit 7.

[0030] Furthermore, the specific analysis steps of the monitoring and judgment unit 7 are: the monitoring and judgment unit 7 constructs a sparse projection matrix for image feature discrimination, the monitoring and judgment unit 7 inputs the image features of the processed image and the image features of the training samples into the sparse projection matrix, the sparse projection matrix processes the image features, and compares the processing results with a preset threshold. If the preset threshold is exceeded, it means that the thickness of the protective layer does not meet the standard.

[0031] Furthermore, the specific method of constructing a sparse projection matrix is ​​as follows: reconstruct the weight coefficients using the LLE-reconstruction weight method, reconstruct the projection matrix by minimizing the neighborhood of L2,1-norm regularization, and perform L2,1-norm regularization on the projection matrix, and accordingly obtain a sparse projection matrix for directly extracting discriminant features of two-dimensional images.

[0032] Furthermore, the specific method of processing image features by the sparse projection matrix is ​​as follows: the sparse projection matrix normalizes the image features of the training samples to obtain a unified thickness image feature set, and the sparse projection matrix compares the image features of the processed image with the thickness image feature set to obtain feature similarity.

[0033] The workflow of the present invention is as follows: when a prefabricated beam steel bar protective layer thickness detection system based on big data designed by the present invention is in operation, the detection module 2 detects the protective layer, the transmission module 3 inputs the detection data into the thickness detection software 4 in the monitoring station 1, and the image processing unit 5 divides the monitoring data image into a non-edge area and an edge area, and determines the discrete degree value of the pixel value in each preset direction within the preset neighborhood of each pixel point in the monitoring data image. If the pixel point is a pixel point in the non-edge area, the pixel point is subjected to image filtering processing according to the discrete degree value of the pixel point in each preset direction. If the pixel point is a pixel point in the edge area, the edge direction is determined according to the discrete degree value and gradient value of the pixel point in each preset direction, and directional smoothing processing and edge enhancement processing are performed according to the determined edge direction. After the processing is completed, the image processing unit 5 inputs the processed image into the feature extraction unit 6;

[0034] The feature extraction unit 6 constructs a convolutional neural network. The feature extraction unit 6 uses the convolutional neural network to perform preliminary feature extraction on the processed image to obtain a first feature image, divide the first feature image into multiple parts in the spatial dimension, perform global pooling on each part, obtain multiple first spatial features, and then use the convolutional neural network to amplify the number of channels of the first feature image to obtain a first amplified feature image after the number of channels is amplified. Then, the first amplified feature image is globally pooled, and the pooling result is divided into multiple first channel features in the channel dimension. Feature fusion is performed based on the multiple first channel features, the multiple first spatial features, and the multiple first channel features to obtain image features of the processed image, and the image features are input into the monitoring and judgment unit 7;

[0035] The monitoring and judgment unit 7 reconstructs the weight coefficient using the LLE-reconstruction weight method, reconstructs the projection matrix by minimizing the neighborhood of L2,1-norm regularization, and performs L2,1-norm regularization on the projection matrix, and accordingly obtains a sparse projection matrix for directly extracting the discriminant features of the two-dimensional image. The monitoring and judgment unit 7 inputs the image features of the processed image and the image features of the training sample into the sparse projection matrix. The sparse projection matrix performs normalization processing on the image features of the training sample to obtain a unified thickness image feature set. The sparse projection matrix compares the image features of the processed image with the thickness image feature set to obtain feature similarity, and compares the processing result with a preset threshold. If it exceeds the preset threshold, it means that the thickness of the protective layer does not meet the standard.

[0036] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A prefabricated beam reinforcement cover thickness detection system based on big data, comprising a monitoring station (1), a detection module (2) and a transmission module (3), characterized in that: The monitoring platform (1) is internally installed with thickness detection software (4), and the thickness detection software (4) comprises an image processing unit (5), a feature extraction unit (6) and a monitoring judgment unit (7).

2. According to the big data-based precast beam reinforcement cover thickness detection system of claim 1, it is characterized by: The monitoring station (1) is connected to the transmission module (3) via a wireless data network, and the detection module (2) is connected to the transmission module (3) via a wire.

3. According to the big data-based prefabricated beam reinforcement cover thickness detection system of claim 1, it is characterized by: The detection module (2) includes various monitoring cameras, the transmission module (3) includes various wireless data transmitters, and the detection module (2) is provided in multiple groups.

4. According to the big data-based precast beam reinforcement cover thickness detection system of claim 1, it is characterized by: The specific analysis steps of the image processing unit (5) are as follows: the detection module (2) detects the protective layer, the transmission module (3) inputs the detection data into the thickness detection software (4) in the monitoring station (1), the image processing unit (5) divides the monitoring data image into a non-edge area and an edge area, determines the discrete degree value of the pixel value in each preset direction within the preset neighborhood of each pixel point in the monitoring data image, if the pixel point is a pixel point in the non-edge area, then according to the discrete degree value of the pixel point in each preset direction, the pixel point is subjected to image filtering processing; if the pixel point is a pixel point in the edge area, the edge direction is determined according to the discrete degree value and gradient value of the pixel point in each preset direction, and directional smoothing processing and edge enhancement processing are performed according to the determined edge direction. After the processing is completed, the image processing unit (5) inputs the processed image into the feature extraction unit (6).

5. According to the big data-based prefabricated beam reinforcement cover thickness detection system of claim 1, it is characterized by: The specific analysis steps of the feature extraction unit (6) are as follows: the feature extraction unit (6) constructs a convolutional neural network, the feature extraction unit (6) uses the convolutional neural network to perform preliminary feature extraction on the processed image to obtain a first feature image, divides the first feature image into multiple parts in the spatial dimension, performs global pooling on each part to obtain multiple first spatial features, then uses the convolutional neural network to amplify the number of channels of the first feature image to obtain a first amplified feature image after the number of channels is amplified, then performs global pooling on the first amplified feature image, divides the pooling result into multiple first channel features in the channel dimension, performs feature fusion based on the multiple first channel features, the multiple first spatial features and the multiple first channel features, obtains image features of the processed image, and inputs the features into the monitoring judgment unit (7).

6. The precast beam reinforcement cover thickness detection system based on big data according to claim 1 is characterized by: The specific analysis steps of the monitoring and judging unit (7) are as follows: the monitoring and judging unit (7) constructs a sparse projection matrix for image feature discrimination; the monitoring and judging unit (7) inputs the image features of the processed image and the image features of the training sample into the sparse projection matrix; the sparse projection matrix processes the image features; and the processing result is compared with a preset threshold value. If the preset threshold value is exceeded, it means that the thickness of the protective layer does not meet the standard.

7. The precast beam reinforcement cover thickness detection system based on big data according to claim 6 is characterized by: The specific method for constructing the sparse projection matrix is: reconstructing the weight coefficients using the LLE-reconstruction weight method, reconstructing the projection matrix by minimizing the neighborhood of L2,1-norm regularization, and performing L2,1-norm regularization on the projection matrix, and correspondingly obtaining a sparse projection matrix for directly extracting discriminant features of a two-dimensional image.

8. The precast beam reinforcement cover thickness detection system based on big data according to claim 6 is characterized by: The specific method of processing the image features by the sparse projection matrix is ​​as follows: the sparse projection matrix performs normalization processing on the image features of the training samples to obtain a unified thickness image feature set, and the sparse projection matrix compares the image features of the processed image with the thickness image feature set to obtain feature similarity.

Images