Concrete surface crack detection method, device and equipment

Abstract

The invention relates to the technical field of image detection, in particular to a concrete surface crack detection method, device and equipment, and the method comprises the steps: obtaining the image data of a concrete surface; performing data preprocessing on the acquired concrete surface image data; after data preprocessing is completed, concrete surface image data are input into a convolutional neural network model based on a self-attention mechanism, and a concrete surface crack detection result is obtained and marked at a corresponding position in the image; detecting results are classified and evaluated through a classifier; and displaying classification and evaluation results through a user interface, and providing a crack analysis report. The concrete surface crack detection method adopts an image processing technology and an artificial intelligence technology to detect the cracks on the concrete surface, realizes automatic, high-efficiency and accurate concrete surface crack detection, does not need a large amount of manpower and material resources, and effectively solves the problems of high cost and low efficiency of a traditional concrete surface crack detection method.

Description

Technical Field

[0001] The present invention relates to the technical field of image detection, and in particular to a method, device and equipment for detecting cracks on a concrete surface. Background Art

[0002] In the field of construction engineering, cracks occur in concrete structures during construction and use due to various reasons, such as temperature changes, load effects, material aging, etc. Whether there are cracks on the concrete surface directly affects the safety and durability of the overall structure of the building. Therefore, timely detection and accurate assessment of the state of cracks are crucial to maintaining structural safety. Traditional methods for detecting cracks on concrete surfaces usually rely on manual visual inspection, ultrasonic testing, infrared thermal imaging, or the use of tools such as probes for detection. These methods are not only time-consuming and labor-intensive, but also easily affected by human factors, resulting in inaccurate detection results and inefficient detection processes. Summary of the invention

[0003] In view of the above-mentioned technical deficiencies, the present invention provides a method, device and equipment for detecting cracks on a concrete surface.

[0004] The present invention is achieved through the following technical solutions: A method for detecting cracks on a concrete surface is provided, the method comprising the following steps: Step S10: Acquire image data of the concrete surface; Step S20: performing data preprocessing on the acquired concrete surface image data; Step S30: After the data preprocessing is completed, the concrete surface image data is input into the convolutional neural network model based on the self-attention mechanism to obtain the concrete surface crack detection results and mark the corresponding positions in the image; Step S40: classifying and evaluating the detection results through a classifier; Step S50: displaying the classification and evaluation results through the user interface and providing a crack analysis report; Preferably, the step of acquiring concrete surface image data in step S10 includes photographing the concrete surface to be inspected using a camera with a resolution of 1k or above or an aerial drone under sufficient lighting and at different angles, and storing the photographed concrete surface image on a computer or server. The storage format is a common image format such as JPEG, PNG and PDF, etc., for subsequent processing.

[0005] Preferably, sufficient lighting means that enough light can be provided to make objects or details in the image clearly visible. The light intensity at the scene is measured by a illuminance meter. When the illuminance is between 150 and 2000 lux, the scene is considered to have sufficient lighting. Lux is the unit used to measure illuminance (i.e., luminous flux per unit area) in the International System of Units.

[0006] Preferably, the data preprocessing step in step S20 includes: Image denoising: Use bilateral filtering to remove noise from the concrete surface image, as shown in formula (1): (1) in, I ( x ) is the original concrete surface image position x The pixel value at I’ ( x ) is the position of the concrete surface image after filtering x The new pixel value at I ( y ) is the original concrete surface image position y The pixel value at is a spatial Gaussian kernel function used to measure position y With location x The spatial distance is the grayscale Gaussian kernel function, used to measure the position y With location x The pixel value difference is W So x The neighborhood window centered at .

[0007] Specifically, for each pixel y , will calculate its difference with the center pixel x Spatial distance and pixel value difference , spatial Gaussian kernel function It will be calculated based on y and x The spatial distance is given a weight, the grayscale Gaussian kernel function It will be calculated based on y and x The difference in pixel values ​​gives another weight, and the product of these two weights determines the pixel y For the new pixel value I’ ( x )’s contribution; Image segmentation: Sobel edge detection technology is used to separate the concrete surface area from the background in the concrete surface image after noise removal. The Sobel operator formula is shown in formula (2), and the edge strength formula is shown in formula (3): (2) (3) in, A is the original concrete surface image, Gx is the gradient value in the horizontal direction, G y is the gradient value in the vertical direction, G is the overall gradient strength; Image enhancement: After image segmentation is completed, gamma correction technology is used to control the brightness of the concrete surface image by adjusting the gamma value to enhance the details of the dark area or reduce the brightness of the overexposed area. The calculation formula is shown in formula (4): (4) in, L ( x , y ) is the pixel value of the original concrete surface image, L’ ( x , y ) is the pixel value of the adjusted concrete surface image, γ is the gamma value; when γ When >1, the dark details of the concrete surface image will be enhanced, and the bright details will be compressed, which is used to enhance the dark details in the image; when γ When <1, the bright details of the concrete surface image will be enhanced, and the dark details will be compressed, which is used to reduce the brightness of the overexposed area in the image; When γ = 1, no adjustment is made to the concrete surface image.

[0008] Preferably, the training step of the convolutional neural network model based on the self-attention mechanism in step S30 includes: Data preparation and preprocessing: Prepare the crack image data of the concrete surface, and perform data cleaning and data partitioning. Data cleaning includes removing invalid values. Data partitioning is used to divide the data set into three sub-data sets: training set, validation set, and test set. Normalization: Perform Softmax normalization on the data in the dataset. The calculation formula is shown in formula (5): (5) Among them, x a is the ath element in the data set, and n is the total number of elements in the data set; Build a network model: This is achieved through a convolutional neural network based on a self-attention mechanism. The model includes convolutional layers, pooling layers, fully connected layers, and self-attention mechanism layers. ReLU is used as the activation function, and a multi-class cross entropy loss function is used as the loss function. The Adam optimizer is used to provide different learning rates for the model, and Dropout is used to prevent overfitting during model training. At the same time, the self-attention mechanism is used to allow the model to focus on areas with cracks during training; Model training: Use the training set divided in data preparation and preprocessing to train the model so that the model can learn the image features of concrete surface cracks; Model validation: Use the validation set divided in data preparation and preprocessing to validate the trained model to adjust hyperparameters and prevent overfitting data; Model testing: Use the test set divided in data preparation and preprocessing to test the trained model. When the model successfully identifies the image features of concrete surface cracks, the model test is successful and the model is uploaded to the host or cloud for concrete surface crack detection.

[0009] Preferably, the classifier in step S40 uses a support vector machine to classify the detected concrete surface cracks, and determines the severity of the cracks according to the length, width and area ratio of the cracks.

[0010] In addition, in order to achieve the above-mentioned purpose, the present invention also proposes a concrete surface crack detection device, the concrete surface crack detection device comprising: Image acquisition module: used to obtain image data of concrete surface; Image preprocessing module: used for preprocessing the acquired concrete surface image data; Crack detection module: After data preprocessing, the concrete surface image data is input into the convolutional neural network model based on the self-attention mechanism to obtain the concrete surface crack detection results and mark the corresponding positions in the image; Classification and evaluation module: used to classify and evaluate the detection results through the classifier; Result display module: used to display classification and evaluation results through the user interface and provide crack analysis reports.

[0011] In addition, to achieve the above-mentioned purpose, the present invention also proposes a concrete surface crack detection device, which includes: a memory, a processor, and programs such as data preprocessing and convolutional neural network model algorithm based on self-attention mechanism stored in the memory and run on the processor. The programs such as data preprocessing and convolutional neural network model algorithm based on self-attention mechanism are steps to implement the concrete surface crack detection method and system as described above.

[0012] Preferably, in order to achieve the above-mentioned purpose, the present invention also provides a computer program product, which includes programs such as data preprocessing and convolutional neural network model algorithm based on self-attention mechanism, and when the programs such as data preprocessing and convolutional neural network model algorithm based on self-attention mechanism are executed by a processor, the concrete surface crack detection method and system as described above are implemented.

[0013] The advantages and effects of the present invention are: The present invention proposes a method, device and equipment for detecting cracks on concrete surfaces, which use image processing technology and artificial intelligence technology to identify cracks on concrete surfaces, thereby realizing automated, efficient and accurate detection of cracks on concrete surfaces, effectively solving the problem of low efficiency of traditional methods for detecting cracks on concrete surfaces, and ensuring the work safety of personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] 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 some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0015] Figure 1 The present invention is a flow chart of a concrete surface crack detection method.

[0016] Figure 2 The present invention is a schematic structural diagram of a concrete surface crack detection device.

[0017] Figure 3 The present invention is a schematic block diagram of the structure of an electronic device for detecting cracks on a concrete surface. DETAILED DESCRIPTION

[0018] 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.

[0019] The present invention provides a method for detecting cracks on a concrete surface. Figure 1 As shown, the following steps are included: Step S10: Acquire image data of the concrete surface; Among them, obtaining image data of the concrete surface includes taking multiple images of the concrete surface under sufficient lighting conditions and at different angles to ensure that the captured images cover the entire concrete surface area; using a high-resolution camera with a resolution of more than 1k or an aerial photography drone to photograph the concrete surface to ensure that the image clarity and resolution are high enough to capture small defect features; storing the captured images on a computer or server in common image formats such as JPEG, PNG, and PDF for subsequent processing; sufficient lighting means that enough light can be provided to make the objects or details in the image clearly visible, and the light intensity on site is measured by a illuminance meter. When the illuminance is between 150 and 2000 lux, the on-site environment is considered to have sufficient lighting. Lux is the unit used in the International System of Units to measure illuminance (i.e., luminous flux per unit area)

[0020] Step S20: performing data preprocessing on the acquired concrete surface image data.

[0021] Specifically, the data preprocessing step in step S20 includes: Image denoising: Use bilateral filtering to remove noise from the concrete surface image, as shown in formula (1): (1) in, I ( x ) is the original concrete surface image position x The pixel value at I’ ( x ) is the position of the concrete surface image after filtering x The new pixel value at I ( y ) is the original concrete surface image position y The pixel value at is a spatial Gaussian kernel function used to measure position y With location x The spatial distance is the grayscale Gaussian kernel function, used to measure the position y With location x The pixel value difference is W So x The neighborhood window centered at ; Specifically, for each pixel y , will calculate its difference with the center pixel x Spatial distance and pixel value difference , spatial Gaussian kernel function It will be calculated based on y and x The spatial distance is given a weight, the grayscale Gaussian kernel function It will be calculated based on y and x The difference in pixel values ​​gives another weight, and the product of these two weights determines the pixel y For the new pixel value I’ ( x )’s contribution; Image segmentation: Sobel edge detection technology is used to separate the concrete surface area from the background in the concrete surface image after noise removal. The Sobel operator formula is shown in formula (2), and the edge strength formula is shown in formula (3): (2) (3) in, A is the original concrete surface image, G x is the gradient value in the horizontal direction, G y is the gradient value in the vertical direction, G is the overall gradient strength; Image enhancement: After image segmentation is completed, gamma correction technology is used to control the brightness of the concrete surface image by adjusting the gamma value to enhance the details of the dark area or reduce the brightness of the overexposed area. The calculation formula is shown in formula (4): (4) in, L ( x , y ) is the pixel value of the original concrete surface image, L’ ( x , y ) is the pixel value of the adjusted concrete surface image, γ is the gamma value; when γ When >1, the dark details of the concrete surface image will be enhanced, and the bright details will be compressed, which is used to enhance the dark details in the image; when γ When <1, the bright details of the concrete surface image will be enhanced, and the dark details will be compressed, which is used to reduce the brightness of the overexposed area in the image; When γ = 1, no adjustment is made to the concrete surface image.

[0022] Step S30: After the data preprocessing is completed, the concrete surface image data is input into the convolutional neural network model based on the self-attention mechanism to obtain the concrete surface crack detection results and mark the corresponding positions in the image.

[0023] Among them, the training steps of the convolutional neural network model based on the self-attention mechanism include: Data preparation and preprocessing: Prepare the crack image data of the concrete surface, and perform data cleaning and data partitioning. Data cleaning includes removing invalid values. Data partitioning is used to divide the data set into three sub-data sets: training set, validation set, and test set. Normalization: Perform Softmax normalization on the data in the dataset. The calculation formula is shown in formula (5): (5) Among them, x a is the ath element in the data set, and n is the total number of elements in the data set; Build a network model: This is achieved through a convolutional neural network based on a self-attention mechanism. The model includes convolutional layers, pooling layers, fully connected layers, and self-attention mechanism layers. ReLU is used as the activation function, and a multi-class cross entropy loss function is used as the loss function. The Adam optimizer is used to provide different learning rates for the model, and Dropout is used to prevent overfitting during model training. At the same time, the self-attention mechanism is used to allow the model to focus on areas with cracks during training; Model training: Use the training set divided in data preparation and preprocessing to train the model so that the model can learn the image features of concrete surface cracks; Model validation: Use the validation set divided in data preparation and preprocessing to validate the trained model to adjust hyperparameters and prevent overfitting data; Model testing: Use the test set divided in data preparation and preprocessing to test the trained model. When the model successfully identifies the image features of concrete surface cracks, the model test is successful and the model is uploaded to the host or cloud for concrete surface crack detection.

[0024] Step S40: Classify and evaluate the detection results through a classifier.

[0025] Specifically, the classifier in step S40 uses a support vector machine to classify the detected concrete surface cracks, and determines the severity of the cracks according to the length, width and area ratio of the cracks.

[0026] Step S50: Display the classification and evaluation results through the user interface and provide a crack analysis report.

[0027] Specifically, the classification and evaluation results are displayed on the user interface, which can be displayed on a computer screen, a tablet computer, or a mobile phone; the classification and evaluation results include the types of cracks on the concrete surface, including plastic collapse cracks, plastic shrinkage (drying shrinkage) cracks, temperature cracks, hydration thermal cracks, structural cracks, etc.; the crack analysis report includes the characteristics and types of cracks, analysis of the causes of cracks, the impact and evaluation of cracks, and the treatment and prevention measures of cracks.

[0028] In addition, the present invention also provides a concrete surface crack detection device, please refer to Figure 2 , the concrete surface crack detection device comprises: Image acquisition module: used to obtain image data of concrete surface; Image preprocessing module: used for preprocessing the acquired concrete surface image data; Crack detection module: After data preprocessing, the concrete surface image data is input into the convolutional neural network model based on the self-attention mechanism to obtain the concrete surface crack detection results and mark the corresponding positions in the image; Classification and evaluation module: used to classify and evaluate the detection results through the classifier; Result display module: used to display classification and evaluation results through the user interface and provide crack analysis reports.

[0029] The concrete surface crack detection device provided by the present application adopts the concrete surface crack detection method in the above embodiment, which can solve the technical problem of low efficiency of the traditional concrete surface crack detection method. Compared with the prior art, the beneficial effects of the concrete surface crack detection device provided by the present application are the same as the beneficial effects of the concrete surface crack detection method provided by the above embodiment, and the other technical features of the concrete surface crack detection device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0030] The present application provides a concrete surface crack detection device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the concrete surface crack detection method in the above-mentioned embodiment 1.

[0031] Reference below Figure 3, which shows a schematic diagram of the structure of a concrete surface crack detection device suitable for implementing the embodiment of the present application. The concrete surface crack detection device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The concrete surface crack detection device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0032] Figure 3 The concrete surface crack detection device shown may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM) 1004. In the RAM 1004, various programs and data required for the operation of the concrete surface crack detection device are also stored. The processing device 1001, the ROM 1002, and the RAM 1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the concrete surface crack detection device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a concrete surface crack detection device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0033] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0034] The concrete surface crack detection device provided by the present application adopts the concrete surface crack detection method in the above embodiment, which can solve the technical problem of low efficiency of the traditional concrete surface crack detection method. Compared with the prior art, the beneficial effects of the concrete surface crack detection device provided by the present application are the same as the beneficial effects of the concrete surface crack detection method provided by the above embodiment, and the other technical features of the concrete surface crack detection device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0035] The various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0036] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned concrete surface crack detection method when executed by a processor.

[0037] The computer program product provided by the present application can solve the technical problem of low efficiency of the traditional concrete surface crack detection method. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as the beneficial effects of the concrete surface crack detection method provided by the above embodiment, which will not be repeated here.

[0038] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for detecting cracks on a concrete surface, characterized in that: The method comprises the following steps: Step S10: Acquire image data of the concrete surface; Step S20: performing data preprocessing on the acquired concrete surface image data; Step S30: After the data preprocessing is completed, the concrete surface image data is input into the convolutional neural network model based on the self-attention mechanism to obtain the concrete surface crack detection results and mark the corresponding positions in the image; Step S40: classifying and evaluating the detection results through a classifier; Step S50: displaying the classification and evaluation results through the user interface and providing a crack analysis report; The step of obtaining concrete surface image data in step S10 includes photographing the concrete surface to be inspected using a camera with a resolution of 1k or above or an aerial drone under sufficient lighting and at different angles, and storing the photographed concrete surface image on a computer or server.

2. A method for detecting cracks on a concrete surface according to claim 1, characterized in that: The data preprocessing step in step S20 includes: Image denoising: Use bilateral filtering to remove noise from the concrete surface image, as shown in formula (1): （1） in, I ( x ) is the concrete surface image position x The pixel value at I’ ( x ) is the position of the concrete surface image after filtering x The new pixel value at I ( y ) is the original concrete surface image position y The pixel value at is the spatial Gaussian kernel function, is the grayscale Gaussian kernel function, W So x The neighborhood window centered at ; Image segmentation: Sobel edge detection technology is used to separate the concrete surface area from the background in the concrete surface image after noise removal. The Sobel operator formula is shown in formula (2), and the edge strength formula is shown in formula (3): （2） （3） in, A is the original concrete surface image, G x is the gradient value in the horizontal direction, G y is the gradient value in the vertical direction, G is the overall gradient strength; Image enhancement: After image segmentation is completed, gamma correction technology is used to control the brightness of the concrete surface image by adjusting the gamma value to enhance the details of the dark area or reduce the brightness of the overexposed area. The calculation formula is shown in formula (4): （4） in, L ( x , y ) is the pixel value of the original concrete surface image, L’ ( x , y ) is the pixel value of the adjusted concrete surface image, γ is the gamma value; when γ When >1, the dark details of the concrete surface image will be enhanced, and the bright details will be compressed, which is used to enhance the dark details in the image; when γ When <1, the bright details of the concrete surface image will be enhanced, and the dark details will be compressed, which is used to reduce the brightness of the overexposed area in the image; When γ = 1, no adjustment is made to the concrete surface image.

3. A method for detecting cracks on a concrete surface according to claim 1, characterized in that: The training steps of the convolutional neural network model based on the self-attention mechanism in step S30 include: Data preparation and preprocessing: Download the concrete surface crack image dataset from the Internet, and perform data cleaning and data partitioning. Data cleaning includes removing invalid values. Data partitioning is used to divide the dataset into three sub-datasets: training set, validation set, and test set. Normalization: Perform Softmax normalization on the data in the dataset. The calculation formula is shown in formula (7): （7） Among them, x a is the ath element in the data set, and n is the total number of elements in the data set; Build a network model: This is achieved through a convolutional neural network based on a self-attention mechanism. The model includes convolutional layers, pooling layers, fully connected layers, and self-attention mechanism layers. ReLU is used as the activation function, and a multi-class cross entropy loss function is used as the loss function. The Adam optimizer is used to provide different learning rates for the model, and the Dropout function is used to prevent overfitting during model training. At the same time, the self-attention mechanism is used to allow the model to focus on areas with cracks during training; Model training: Use the training set divided in data preparation and preprocessing to train the model so that the model can learn the image features of concrete surface cracks; Model validation: Use the validation set divided in data preparation and preprocessing to validate the trained model to adjust hyperparameters and prevent overfitting data; Model testing: Use the test set divided in data preparation and preprocessing to test the trained model. When the model successfully identifies the image features of concrete surface cracks, the model test is successful and the model is uploaded to the host or cloud for concrete surface crack detection.

4. A method for detecting cracks on a concrete surface according to claim 1, characterized in that: The classifier in step S40 uses a support vector machine to classify the detected concrete surface cracks, and determines the severity of the cracks according to the length, width and area ratio of the cracks.

5. A concrete surface crack detection device, characterized in that: The concrete surface crack detection device comprises: Image acquisition module: used to obtain image data of concrete surface; Image preprocessing module: used for preprocessing the acquired concrete surface image data; Crack detection module: After data preprocessing, the concrete surface image data is input into the convolutional neural network model based on the self-attention mechanism to obtain the concrete surface crack detection results and mark the corresponding positions in the image; Classification and evaluation module: used to classify and evaluate the detection results through the classifier; Result display module: used to display classification and evaluation results through the user interface and provide crack analysis reports; Among them, the step of obtaining concrete surface image data in the image acquisition module includes using a camera with a resolution of 1k or above or an aerial drone to shoot the concrete surface to be inspected under sufficient lighting and at different angles, and storing the shot concrete surface image on a computer or server.

6. A concrete surface crack detection device, characterized in that: The concrete surface crack detection device comprises: A memory, a processor, and a concrete surface crack detection program stored in the memory and executable on the processor, wherein the concrete surface crack detection program, when executed by the processor, implements the concrete surface crack detection method according to any one of claims 1 to 4.

7. A computer program product, characterized in that The computer program product comprises a concrete surface crack detection program, and when the concrete surface crack detection program is executed by a processor, the concrete surface crack detection method according to any one of claims 1 to 4 is implemented.

Images