A Prediction Method and System for Disease Types of Subway Tunnel Structures

Abstract

The present invention relates to a method and system for predicting the types of diseases in a subway tunnel structure, including: collecting infrared images of the subway tunnel; performing multi-scale enhancement on the infrared images of the subway tunnel to obtain infrared images of the subway tunnel with enhanced features; calibrating the types of diseases presented in each infrared image of the subway tunnel with enhanced features to obtain training samples; inputting the training samples into a MobileNet model for training to obtain a prediction model for the types of diseases in the subway tunnel; and using the prediction model for the types of diseases in the subway tunnel to monitor the target subway tunnel. By performing multi-scale enhancement on the infrared images, the present invention can improve the identifiability of disease features, making different types of diseases more obvious in the images and enhancing the effect of model training.

Description

Technical Field

[0001] The present invention relates to the technical field of disease identification, and particularly to a method and system for predicting the disease types of subway tunnel structures. Background Art

[0002] With the acceleration of the urbanization process, the subway, as an important public transportation means, has been widely used globally. However, due to the complex environment in which subway tunnels are located and the influence of long-term use, various structural diseases are likely to occur, seriously affecting the safe operation and service quality of the subway. Therefore, timely detection and accurate prediction of the disease types of tunnel structures are of great significance for improving the safety and operation efficiency of subway tunnels.

[0003] Existing tunnel disease detection technologies mostly rely on manual inspection, regular monitoring, and experience judgment. The methods are relatively single and it is difficult to comprehensively and timely identify and predict potential diseases in tunnels. Summary of the Invention

[0004] To solve the above problems, the purpose of the embodiments of the present invention is to provide a method and system for predicting the disease types of subway tunnel structures.

[0005] A method for predicting the disease types of subway tunnel structures includes:

[0006] Step 1: Collect infrared images of the subway tunnel;

[0007] Step 2: Perform multi-scale enhancement on the infrared image of the subway tunnel to obtain an infrared image of the subway tunnel with enhanced features;

[0008] Step 3: Calibrate the disease types presented in each infrared image of the subway tunnel with enhanced features to obtain training samples;

[0009] Step 4: Input the training samples into the MobileNet model for training to obtain a prediction model for the disease types of subway tunnels;

[0010] Step 5: Use the prediction model for the disease types of subway tunnels to monitor the target subway tunnel.

[0011] Preferably, the Step 2: Perform multi-scale enhancement on the infrared image of the subway tunnel to obtain an infrared image of the subway tunnel with enhanced features, includes:

[0012] Step 2.1: Perform Gaussian pyramid decomposition on the infrared image of the subway tunnel to obtain a Gaussian pyramid decomposition image;

[0013] Step 2.2: Perform edge monitoring processing on the infrared image of the subway tunnel to obtain an edge image;

[0014] Step 2.3: Upsample the Gaussian pyramid decomposed image to obtain an upsampled image;

[0015] Step 2.4: Construct pyramid image patches using the upsampled image and the edge image;

[0016] Step 2.5: Perform multi-scale feature enhancement on the pyramid image patches to obtain an infrared image of the subway tunnel with enhanced features.

[0017] Preferably, the Step 2.1: Perform Gaussian pyramid decomposition on the infrared image of the subway tunnel to obtain a Gaussian pyramid decomposed image, including:

[0018] Perform Gaussian pyramid decomposition on the infrared image of the subway tunnel using the Gaussian pyramid decomposition formula to obtain a Gaussian pyramid decomposed image; where the Gaussian pyramid decomposition formula is:

[0019]

[0020] Where, represents the infrared image of the subway tunnel, represents the first-layer Gaussian pyramid decomposed image, represents the i-th layer Gaussian pyramid decomposed image, represents the Gaussian filter function.

[0021] Preferably, the Step 2.2: Perform edge monitoring processing on the infrared image of the subway tunnel to obtain an edge image, including:

[0022] Step 2.2.1: Calculate the change amplitude and change direction of each pixel point in the infrared image of the subway tunnel; where the calculation formulas for the change amplitude and change direction of the pixel point are:

[0023]

[0024] Where, represents the change amplitude of the pixel point, represents the change direction of the pixel point, represents the amplitude in the x direction, represents the amplitude in the y direction;

[0025] Step 2.2.2: Take the pixel points with the change amplitude and change direction within a preset range as edge points, and set the remaining pixel points to zero to obtain an edge image.

[0026] Preferably, the Step 2.4: Construct pyramid image patches using the upsampled image and the edge image, including:

[0027] The upsampled image and the Gaussian pyramid decomposed image are subjected to a difference operation to obtain a pyramid image block; wherein, the construction formula of the pyramid image block is:

[0028]

[0029] Wherein, represents the i-th pyramid image block, represents the upsampled image corresponding to the Gaussian pyramid decomposed image of the i-th layer, represents the edge image.

[0030] Preferably, the step 2.5: performing multi-scale feature enhancement on the pyramid image block to obtain an infrared image of the subway tunnel with enhanced features includes:

[0031] Step 2.5.1: Extract the first pyramid image block as the base image, and use the remaining pyramid image blocks as detail images;

[0032] Step 2.5.2: Perform Gamma correction on the base image to obtain a corrected base image;

[0033] Step 2.5.3: Perform enhancement processing on the detail images to obtain enhanced detail images;

[0034] Step 2.5.4: Fuse the corrected base image and the enhanced detail images to obtain an infrared image of the subway tunnel with enhanced features.

[0035] Preferably, the step 2.5.3: performing enhancement processing on the detail images to obtain enhanced detail images includes:

[0036] Using a sharpening enhancement operator to perform enhancement processing on the detail images to obtain enhanced detail images; wherein, the enhancement processing formula is:

[0037]

[0038] Wherein, represents the enhanced detail image, represents the pixel value of the detail image at the position, represents the sharpening enhancement operator, represents a preset threshold.

[0039] The present invention also provides a prediction system for subway tunnel structure disease types, including:

[0040] An infrared image acquisition module for acquiring infrared images of the subway tunnel;

[0041] An enhancement module for performing multi-scale enhancement on the infrared image of the subway tunnel to obtain an infrared image of the subway tunnel with enhanced features;

[0042] A calibration module for calibrating the disease types presented in each infrared image of the subway tunnel with enhanced features to obtain training samples;

[0043] A training module for inputting the training samples into the MobileNet model for training to obtain a prediction model for the disease types of the subway tunnel;

[0044] A monitoring module for using the prediction model for the disease types of the subway tunnel to monitor the target subway tunnel.

[0045] The present invention also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored on the memory and executable on the processor. The transceiver, the memory, and the processor are connected through the bus. It is characterized in that when the computer program is executed by the processor, the steps in the above-mentioned method for predicting the disease types of the subway tunnel structure are implemented.

[0046] The present invention also provides a computer-readable storage medium, on which a computer program is stored. It is characterized in that when the computer program is executed by a processor, the steps in the above-mentioned method for predicting the disease types of the subway tunnel structure are implemented.

[0047] According to the specific embodiments provided by the present invention, the following technical effects are disclosed:

[0048] The present invention relates to a method for predicting the disease types of the subway tunnel structure. Compared with the prior art, by performing multi-scale enhancement on the infrared image, the present invention can improve the identifiability of the disease features, make different types of diseases more obvious in the image, and improve the effect of model training.

[0049] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following preferred embodiments are specifically exemplified and described in detail in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0051] Figure 1 It is a flowchart of a method for predicting the disease types of the subway tunnel structure provided by the present invention;

[0052] Figure 2 The infrared image of the subway tunnel provided for the present invention that has not been processed;

[0053] Figure 3 The infrared image of the subway tunnel with enhanced features provided for the present invention. Specific embodiments

[0054] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", etc. indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be understood as a limitation to the present invention.

[0055] In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, "a plurality" means two or more unless otherwise specifically defined.

[0056] In the present invention, unless otherwise clearly specified and defined, the terms "install", "connect", "connect", "fix", etc. should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0057] Please refer to Figure 1 , a prediction method for subway tunnel structure disease types, including:

[0058] Step 1: Collect the infrared image of the subway tunnel;

[0059] Step 2: Perform multi-scale enhancement on the infrared image of the subway tunnel to obtain an infrared image of the subway tunnel with enhanced features;

[0060] Furthermore, the step 2 includes:

[0061] Step 2.1: Perform Gaussian pyramid decomposition on the infrared image of the subway tunnel to obtain a Gaussian pyramid decomposition image;

[0062] Among them, step 2.1 includes:

[0063] Performing Gaussian pyramid decomposition on the infrared image of the subway tunnel using the Gaussian pyramid decomposition formula to obtain a Gaussian pyramid decomposition image; where the Gaussian pyramid decomposition formula is:

[0064]

[0065] Among them, represents the infrared image of the subway tunnel, represents the first-layer Gaussian pyramid decomposition image, represents the i-th layer Gaussian pyramid decomposition image, represents the Gaussian filter function.

[0066] Step 2.2: Performing edge monitoring processing on the infrared image of the subway tunnel to obtain an edge image;

[0067] In the said step 2.2, it includes:

[0068] Step 2.2.1: Calculating the change amplitude and change direction of each pixel point of the infrared image of the subway tunnel; where the calculation formulas for the change amplitude and change direction of the pixel point are:

[0069]

[0070] Among them, represents the change amplitude of the pixel point, represents the change direction of the pixel point, represents the amplitude in the x direction, represents the amplitude in the y direction;

[0071] Step 2.2.2: Taking the pixel points with the change amplitude and change direction within a preset range as edge points, and setting the remaining pixel points to zero to obtain an edge image.

[0072] Step 2.3: Performing upsampling on the Gaussian pyramid decomposition image to obtain an upsampled image;

[0073] Step 2.4: Using the upsampled image and the edge image to construct a pyramid image block;

[0074] Among them, the said step 2.4 includes:

[0075] Performing a difference operation on the upsampled image and the Gaussian pyramid decomposition image to obtain a pyramid image block; where the construction formula for the pyramid image block is:

[0076]

[0077] Among them, Denote the i-th pyramid image block, Denote the upsampled image corresponding to the image obtained by Gaussian pyramid decomposition at the i-th layer, Denote the edge image.

[0078] By introducing edge information, the present invention can highlight important features in each layer of the pyramid, enhance the visualization effect of information, and provide a clearer basis for subsequent image analysis tasks.

[0079] Step 2.5: Perform multi-scale feature enhancement on the pyramid image block to obtain an infrared image of the subway tunnel with enhanced features;

[0080] Among them, Step 2.5 includes:

[0081] Step 2.5.1: Extract the first pyramid image block as the base image, and use the remaining pyramid image blocks as detail images;

[0082] Step 2.5.2: Perform Gamma correction on the base image to obtain the corrected base image;

[0083] The base image contains the global information of the entire image. By performing Gamma correction on the base image, the present invention can avoid the problem of excessive enhancement of local contrast in the image.

[0084] Step 2.5.3: Use a sharpening enhancement operator to perform enhancement processing on the detail image to obtain an enhanced detail image; among them, the enhancement processing formula is:

[0085]

[0086] Among them, Denote the enhanced detail image, Denote the pixel value of the detail image at the position, Denote the sharpening enhancement operator, Denote the preset threshold.

[0087] By using the sharpening enhancement operator, the present invention can extract high-frequency information in the image, which is particularly important for image texture analysis and can make the image look clearer.

[0088] Step 2.5.4: Fuse the corrected base image and the enhanced detail image to obtain an infrared image of the subway tunnel with enhanced features;

[0089] Through the linear weighted fusion method, the integrity of global information can be maintained while enhancing details, thereby avoiding information loss. The fusion formula is:

[0090]

[0091] Among them, represents the infrared image of the subway tunnel after feature enhancement, represents the corrected basic image, represents the enhanced detail image, represents the weight.

[0092] As Figures 2 - 3 shown, through the fusion processing of the infrared image of the subway tunnel, the present invention can achieve significant contrast effects and rich texture detail displays.

[0093] Step 3: Calibrate the disease types presented in each infrared image of the subway tunnel after feature enhancement to obtain training samples;

[0094] In practical applications, the present invention cooperates with professional subway maintenance engineers or tunnel disease detection experts to identify and classify various disease types in infrared images, and uses image annotation tools (such as Labelme, VGG ImageAnnotator, etc.) to annotate the infrared images.

[0095] Step 4: Input the training samples into the MobileNet model for training to obtain a subway tunnel disease type prediction model;

[0096] Step 5: Use the subway tunnel disease type prediction model to monitor the target subway tunnel.

[0097] Through multi-scale enhancement of the infrared image, the present invention can improve the identifiability of disease features, make different types of diseases more obvious in the image, and make it easier for machine learning models to identify and classify targets, thereby improving the training effect of the model.

[0098] The present invention also provides a subway tunnel structure disease type prediction system, including:

[0099] An infrared image acquisition module for acquiring infrared images of the subway tunnel;

[0100] An enhancement module for performing multi-scale enhancement on the infrared image of the subway tunnel to obtain an infrared image of the subway tunnel after feature enhancement;

[0101] A calibration module for calibrating the disease types presented in each infrared image of the subway tunnel after feature enhancement to obtain training samples;

[0102] A training module for inputting the training samples into the MobileNet model for training to obtain a subway tunnel disease type prediction model;

[0103] A monitoring module for using the subway tunnel disease type prediction model to monitor the target subway tunnel.

[0104] The present invention also provides an electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored on the memory and executable on the processor. The transceiver, the memory, and the processor are connected via the bus. It is characterized in that when the computer program is executed by the processor, the steps in the above-mentioned method for predicting the types of subway tunnel structure diseases are realized. Compared with the prior art, the beneficial effects of the electronic device provided by the present invention are the same as those of the method for predicting the types of subway tunnel structure diseases described in the above technical solution, and will not be elaborated here.

[0105] The present invention also provides a computer-readable storage medium, on which a computer program is stored. It is characterized in that when the computer program is executed by a processor, the steps in the above-mentioned method for predicting the types of subway tunnel structure diseases are realized. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present invention are the same as those of the method for predicting the types of subway tunnel structure diseases described in the above technical solution, and will not be elaborated here.

[0106] As described above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of technical solutions of changes or substitutions, which should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for predicting the type of structural damage in a subway tunnel, characterized in that: include: Step 1: Collect infrared images of subway tunnels; Step 2: performing multi-scale enhancement on the subway tunnel infrared image to obtain a feature-enhanced subway tunnel infrared image; The step 2: performing multi-scale enhancement on the subway tunnel infrared image to obtain a feature-enhanced subway tunnel infrared image, comprises: Step 2.1: performing Gaussian pyramid decomposition on the subway tunnel infrared image to obtain a Gaussian pyramid decomposition image; The step 2.1: performing Gaussian pyramid decomposition on the subway tunnel infrared image to obtain a Gaussian pyramid decomposition image, comprises: The subway tunnel infrared image is subjected to Gaussian pyramid decomposition using a Gaussian pyramid decomposition formula to obtain a Gaussian pyramid decomposition image; wherein the Gaussian pyramid decomposition formula is: Among them, I(x,y) represents the infrared image of the subway tunnel, G1(x,y) represents the first-level Gaussian pyramid decomposition image, G i (x, y) represents the i-th level Gaussian pyramid decomposition image, G σ (x,y) represents the Gaussian filter function; Step 2.2: performing edge monitoring processing on the subway tunnel infrared image to obtain an edge image; The step 2.2: performing edge monitoring processing on the subway tunnel infrared image to obtain an edge image, comprises: Step 2.2.1: Calculate the change amplitude and change direction of each pixel point of the subway tunnel infrared image; wherein the calculation formula for the change amplitude and change direction of the pixel point is: Among them, G represents the change amplitude of the pixel point, θ represents the change direction of the pixel point, and G X represents the magnitude in the x direction, G Y represents the magnitude in the y direction; Step 2.2.2: Pixels whose change amplitude and change direction are within a preset range are regarded as edge points, and the remaining pixels are set to zero to obtain an edge image; Step 2.3: Upsample the Gaussian pyramid decomposition image to obtain an upsampled image; Step 2.4: constructing a pyramid image block using the upsampled image and the edge image; The step 2.4: constructing a pyramid image block using the up-sampled image and the edge image, comprises: The upsampled image is differentially calculated with the Gaussian pyramid decomposition image to obtain a pyramid image block; wherein the pyramid image block construction formula is: Among them, L i represents the i-th pyramid image block, represents the upsampled image corresponding to the i-th layer Gaussian pyramid decomposition image, C i represents an edge image; Step 2.5: performing multi-scale feature enhancement on the pyramid image block to obtain a feature-enhanced subway tunnel infrared image; The step 2.5: performing multi-scale feature enhancement on the pyramid image block to obtain a feature-enhanced subway tunnel infrared image, comprises: Step 2.5.1: Extract the first pyramid image block as the base image and the remaining pyramid image blocks as detail images; Step 2.5.2: performing gamma correction on the basic image to obtain a corrected basic image; Step 2.5.3: performing enhancement processing on the detail image to obtain an enhanced detail image; Step 2.5.4: Fuse the corrected basic image and the enhanced detail image to obtain the feature-enhanced subway tunnel infrared image; Step 3: Calibrate the damage type presented by each feature-enhanced subway tunnel infrared image to obtain training samples; Step 4: Input the training samples into the MobileNet model for training to obtain a subway tunnel disease type prediction model; Step 5: Use the subway tunnel disease type prediction model to monitor the target subway tunnel.

2. According to claim 1, a method for predicting the type of structural damage in a subway tunnel is characterized in that: The step 2.5.3: performing enhancement processing on the detail image to obtain an enhanced detail image comprises: The detail image is enhanced using a sharpening enhancement operator to obtain an enhanced detail image; wherein the enhancement processing formula is: Among them, g(x,y) represents the enhanced detail image, f(x,y) represents the pixel value of the detail image at the position (x,y), represents the sharpening enhancement operator, and β represents the preset threshold.

3. A subway tunnel structural damage type prediction system, characterized in that: include: Infrared image acquisition module, used to collect infrared images of subway tunnels; An enhancement module, used for performing multi-scale enhancement on the subway tunnel infrared image to obtain a feature-enhanced subway tunnel infrared image; The enhancement module includes: Step 2.1: performing Gaussian pyramid decomposition on the subway tunnel infrared image to obtain a Gaussian pyramid decomposition image; The step 2.1: performing Gaussian pyramid decomposition on the subway tunnel infrared image to obtain a Gaussian pyramid decomposition image, comprises: The subway tunnel infrared image is subjected to Gaussian pyramid decomposition using a Gaussian pyramid decomposition formula to obtain a Gaussian pyramid decomposition image; wherein the Gaussian pyramid decomposition formula is: Among them, I(x,y) represents the infrared image of the subway tunnel, G1(x,y) represents the first-level Gaussian pyramid decomposition image, G i (x, y) represents the i-th level Gaussian pyramid decomposition image, G σ (x,y) represents the Gaussian filter function; Step 2.2: performing edge monitoring processing on the subway tunnel infrared image to obtain an edge image; The step 2.2: performing edge monitoring processing on the subway tunnel infrared image to obtain an edge image, comprises: Step 2.2.1: Calculate the change amplitude and change direction of each pixel point of the subway tunnel infrared image; wherein the calculation formula for the change amplitude and change direction of the pixel point is: Among them, G represents the change amplitude of the pixel point, θ represents the change direction of the pixel point, and G X represents the magnitude in the x direction, G Y represents the magnitude in the y direction; Step 2.2.2: Pixels whose change amplitude and change direction are within a preset range are regarded as edge points, and the remaining pixels are set to zero to obtain an edge image; Step 2.3: Upsample the Gaussian pyramid decomposition image to obtain an upsampled image; Step 2.4: constructing a pyramid image block using the upsampled image and the edge image; The step 2.4: constructing a pyramid image block using the up-sampled image and the edge image, comprises: The upsampled image is differentially calculated with the Gaussian pyramid decomposition image to obtain a pyramid image block; wherein the pyramid image block construction formula is: Among them, L i represents the i-th pyramid image block, represents the upsampled image corresponding to the i-th layer Gaussian pyramid decomposition image, C i represents an edge image; Step 2.5: performing multi-scale feature enhancement on the pyramid image block to obtain a feature-enhanced subway tunnel infrared image; The step 2.5: performing multi-scale feature enhancement on the pyramid image block to obtain a feature-enhanced subway tunnel infrared image, comprises: Step 2.5.1: Extract the first pyramid image block as the base image and the remaining pyramid image blocks as detail images; Step 2.5.2: performing gamma correction on the basic image to obtain a corrected basic image; Step 2.5.3: performing enhancement processing on the detail image to obtain an enhanced detail image; Step 2.5.4: Fuse the corrected basic image and the enhanced detail image to obtain the feature-enhanced subway tunnel infrared image; A calibration module is used to calibrate the type of damage presented by each feature-enhanced subway tunnel infrared image to obtain training samples; The training module is used to input the training samples into the MobileNet model for training to obtain a subway tunnel disease type prediction model; A monitoring module is used to monitor the target subway tunnel using the subway tunnel disease type prediction model.

4. An electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, wherein: When the computer program is executed by the processor, the steps of the method for predicting the type of structural damage in a subway tunnel as described in any one of claims 1-2 are implemented.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a method for predicting structural damage types of a subway tunnel as described in any one of claims 1-2 are implemented.

Images