Concrete defect detection method, detection device, electronic equipment and storage medium

Abstract

The application discloses a concrete defect detection method, a detection device, an electronic device and a storage medium. The concrete defect detection method comprises the following steps: in response to a received detection instruction, acquiring a current concrete image of concrete, wherein the detection instruction is used for instructing to perform defect detection on the concrete; inputting the current concrete image into a target defect detection model to obtain a corresponding defect detection result, wherein the target defect detection model is obtained based on a single-step multi-frame target detection model and a texture enhancement model, and the texture enhancement model is used for enhancing and extracting texture features of the current concrete image. The method realizes defect detection on the concrete by using the target defect detection model constructed based on the single-step multi-frame target detection model and the texture enhancement model, the detection time is short, and the detection efficiency and detection accuracy of the defect detection on the concrete are improved.

Description

Technical Field

[0001] This application belongs to the field of concrete testing technology, and in particular relates to a concrete defect detection method, detection device, electronic equipment and storage medium. Background Technology

[0002] Concrete possesses excellent integrity, moldability, durability, and fire resistance, making it widely used in the construction of infrastructure such as buildings, bridges, tunnels, and dams. However, under the combined effects of material aging, external temperature changes, and long-term loads, concrete is prone to defects during construction and operation, such as cracks and damage. These defects reduce the concrete's load-bearing capacity and increase safety risks. Therefore, defect detection in concrete is of paramount importance.

[0003] Currently, the detection of defects in concrete mainly relies on manual inspection and observation. Manual inspection and observation is time-consuming, and there may be omissions or misjudgments of defects. Summary of the Invention

[0004] In view of the above, embodiments of this application provide a concrete defect detection method, detection device, electronic device and storage medium to overcome or at least partially solve the problems of the prior art.

[0005] In a first aspect, embodiments of this application provide a concrete defect detection method, comprising: in response to a received detection instruction, acquiring a current concrete image, wherein the detection instruction is used to instruct the concrete to perform defect detection; inputting the current concrete image to a target defect detection model to obtain a corresponding defect detection result, wherein the target defect detection model is obtained based on a single-step multi-frame target detection model and a texture enhancement model, wherein the texture enhancement model is used to enhance and extract the texture features of the current concrete image.

[0006] Secondly, embodiments of this application provide a concrete defect detection device, which includes a first acquisition module and a first input module. The first acquisition module is used to acquire a current concrete image in response to a received detection command, the detection command being used to instruct the concrete to perform defect detection; the first input module is used to input the current concrete image into a target defect detection model to obtain the corresponding defect detection result, the target defect detection model being obtained based on a single-step multi-frame target detection model and a texture enhancement model, the texture enhancement model being used to enhance and extract the texture features of the current concrete image.

[0007] Thirdly, embodiments of this application provide an electronic device, including a memory; one or more processors coupled to the memory; and one or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the one or more processors, and the one or more application programs are configured to perform the concrete defect detection method as provided in the first aspect above.

[0008] Fourthly, embodiments of this application provide a computer-readable storage medium storing program code, which can be called by a processor to execute the concrete defect detection method provided in the first aspect above.

[0009] Fifthly, embodiments of this application provide a computer program product that, when run on a computer device, causes the computer device to execute the concrete defect detection method provided in the first aspect above.

[0010] The solution provided in this application, in response to a received detection command, acquires the current concrete image. The detection command instructs the concrete to be inspected for defects, and inputs the current concrete image into the target defect detection model to obtain the corresponding defect detection result. The target defect detection model is based on a single-step multi-frame target detection model and a texture enhancement model. The texture enhancement model is used to enhance and extract the texture features of the current concrete image. This solution enables the target defect detection model based on the single-step multi-frame target detection model and the texture enhancement model to perform defect detection on concrete. The detection time is short, which improves the detection efficiency and accuracy of concrete defect detection.

[0011] Furthermore, constructing a target defect detection model based on a single-step multi-frame target detection model and a texture enhancement model is beneficial to improving the target defect detection model's ability to learn the texture features of the current concrete image, making concrete defects easier to detect and further improving the accuracy of concrete defect detection. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This illustration shows a scenario diagram of the concrete defect detection system provided in an embodiment of this application.

[0014] Figure 2A schematic flowchart of a concrete defect detection method provided in an embodiment of this application is shown.

[0015] Figure 3 This paper illustrates another flowchart of the concrete defect detection method provided in an embodiment of this application.

[0016] Figure 4 A schematic diagram of a texture enhancement model in the concrete defect detection method provided in this application is shown.

[0017] Figure 5 This paper illustrates a structural schematic diagram of an initial defect detection model in the concrete defect detection method provided in an embodiment of this application.

[0018] Figure 6 This illustration shows another schematic flowchart of the concrete defect detection method provided in the embodiments of this application.

[0019] Figure 7 This illustration shows another flowchart of the concrete defect detection method provided in the embodiments of this application.

[0020] Figure 8 A structural block diagram of a concrete defect detection device provided in an embodiment of this application is shown.

[0021] Figure 9 A functional block diagram of an electronic device provided in an embodiment of this application is shown.

[0022] Figure 10 This application illustrates a computer-readable storage medium for storing or carrying program code that implements a concrete defect detection method according to an embodiment of this application.

[0023] Figure 11 This application illustrates a computer program product for storing or carrying program code that implements the concrete defect detection method provided in the embodiments of this application. Detailed Implementation

[0024] To make the inventive objectives, features, and advantages of this application more apparent and understandable, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0025] It should be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0026] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0027] It should also be further understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0028] Furthermore, in the description of this application, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0029] Concrete possesses excellent integrity, moldability, durability, and fire resistance, making it widely used in the construction of infrastructure such as buildings, bridges, tunnels, and dams. However, under the combined effects of material aging, external temperature changes, and long-term loads, concrete is prone to defects during construction and operation, such as cracks and damage. These defects reduce the concrete's load-bearing capacity and increase safety risks. Therefore, defect detection in concrete is of paramount importance.

[0030] Currently, the detection of defects in concrete mainly relies on manual inspection and observation. Manual inspection and observation is time-consuming, and there may be omissions or misjudgments of defects.

[0031] To address the aforementioned issues, the concrete defect detection method, detection device, electronic device, and storage medium provided in this application embodiment, in response to a received detection command, acquire a current concrete image. The detection command instructs the concrete to undergo defect detection, and inputs the current concrete image into a target defect detection model to obtain the corresponding defect detection result. The target defect detection model is based on a single-step multi-frame target detection model and a texture enhancement model. The texture enhancement model is used to enhance and extract the texture features of the current concrete image. This enables the target defect detection model constructed based on the single-step multi-frame target detection model and the texture enhancement model to perform defect detection on concrete. The detection time is short, improving the detection efficiency and accuracy of concrete defect detection.

[0032] Furthermore, constructing a target defect detection model based on a single-step multi-frame target detection model and a texture enhancement model is beneficial to improving the target defect detection model's ability to learn the texture features of the current concrete image, making concrete defects easier to detect and further improving the accuracy of concrete defect detection.

[0033] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0034] Please see Figure 1 This illustration shows an application scenario diagram of the concrete defect detection system provided in the embodiments of this application, which may include concrete 100, camera 200 and processing equipment 300. The camera 200 can be installed relative to the concrete 100. The camera 200 can be connected to the processing equipment 300 through a network and interact with the processing equipment 300 through the network.

[0035] Among them, concrete 100 can be any of cement concrete, silicate concrete, gypsum concrete, water glass concrete, asphalt concrete or polymer concrete, etc. The type of concrete 100 is not limited here, and can be set according to actual needs.

[0036] The camera 200 can be used to acquire images of the concrete 100, obtain concrete images, and send the concrete images to the processing device 300 via the network.

[0037] Camera 200 can be any of the following: wide-angle camera, macro camera, ultra-wide-angle camera, or panoramic camera. The type of camera 200 is not limited here, and can be set according to actual needs.

[0038] The processing device 300 can be used to receive concrete images sent by the camera 200 and perform defect detection on the concrete 100 based on the concrete images. The processing device 300 can be any of the following: a server or a terminal device, etc., without limitation, and can be configured according to actual needs.

[0039] A server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), big data, and artificial intelligence platforms.

[0040] Terminal devices can be mobile terminal devices (e.g., in-vehicle terminals, PDAs, tablet PCs, laptops, etc.) or fixed terminal devices (desktop computers, smart panels, etc.).

[0041] The network can be any of the following: ZigBee network, Bluetooth (BT) network, Wireless Fidelity (Wi-Fi) network, Thread network, Long Range Radio (LoRa) network, Low-Power Wide-Area Network (LPWAN), Infrared network, Narrow Band Internet of Things (NB-IoT), Controller Area Network (CAN), Digital Living Network Alliance (DLNA) network, Wide Area Network (WAN), Local Area Network (LAN), Metropolitan Area Network (MAN), or Wireless Personal Area Network (WPAN), etc., without limitation.

[0042] Please see Figure 2 This document illustrates a flowchart of a concrete defect detection method according to an embodiment of this application. In a specific embodiment, the concrete defect detection method can be applied to, for example... Figure 1 The concrete defect detection system shown includes a processing device 300. The following section uses processing device 300 as an example to discuss... Figure 2 The process shown is described in detail. The concrete defect detection method may include the following steps S110 to S120.

[0043] Step S110: In response to the received detection command, acquire the current concrete image.

[0044] In this embodiment of the application, when the inspector needs to inspect the concrete for defects, he / she can send an inspection command to the processing device. The processing device receives and responds to the inspection command, and sends a first acquisition command to the camera through the network. The camera receives and responds to the first acquisition command, performs image acquisition on the concrete, obtains the current concrete image, and sends the current concrete image to the processing device through the network. The processing device receives the current concrete image returned by the camera.

[0045] Among them, the inspection command can be used to instruct the processing equipment to perform defect inspection on the concrete.

[0046] In some implementations, the processing equipment can detect the operations of the inspector. When it is determined from the detected operations that the inspector has input a test instruction to detect defects in the concrete, the equipment receives the test instruction to detect defects in the concrete.

[0047] For example, when inspectors need to perform defect detection on concrete, they can perform touch operations on the operation panel of the processing equipment. The processing equipment responds to the inspector's touch operation, generates a corresponding touch signal, and analyzes the touch signal. When it is determined that the touch signal is a preset detection signal used to characterize defect detection of concrete, it is determined that a detection instruction to perform defect detection on concrete has been received.

[0048] In some implementations, the processing device may be equipped with a voice recognition module. When an inspector needs to perform defect detection on concrete, the inspector can send voice information within the voice acquisition range of the voice recognition module. The voice recognition module collects the voice information sent by the inspector, performs voice recognition on the collected voice information, and determines, based on the recognition result, that the recognition result contains keywords used to indicate defect detection on concrete, such as "concrete defect detection," or, for example, "concrete" and "defect detection," etc., then determines that a detection instruction to perform defect detection on concrete has been received.

[0049] As an example, if the inspector issues a voice message stating "Concrete defect detection," and the voice recognition result contains the keywords "concrete" and "defect detection," then it is confirmed that an inspection instruction to perform defect detection on the concrete has been received.

[0050] In some implementations, the concrete defect detection system may also include a client associated with the inspector, which is connected to the processing equipment via a network and interacts with the processing equipment via the network.

[0051] When inspectors need to inspect concrete for defects, they can send an inspection command to the client. The client receives and responds to the inspection command, and forwards it to the processing equipment via the network. The processing equipment then receives the inspection command forwarded by the client.

[0052] The client can be any of the following: a mobile client (e.g., a mobile phone client, a PDA client, a Tablet PC client, a laptop client, a smartwatch client, a smart bracelet client, or a wearable client) or a fixed client (e.g., a desktop computer client, a smart panel client). The type of client is not limited here and can be set according to actual needs.

[0053] Step S120: Input the current concrete image into the target defect detection model to obtain the corresponding defect detection results.

[0054] In this embodiment, the processing device, in response to the received detection command, acquires the current concrete image and inputs it into the target defect detection model. The target defect detection model receives and responds to the current concrete image, performs defect detection on the concrete based on the current concrete image, obtains the corresponding defect detection result, and outputs the defect detection result to the processing device. The processing device receives the defect detection result output by the target defect detection model. The target defect detection model can be obtained based on a Single Shot Multi-Box Detector (SSD) model and a texture enhancement model. The texture enhancement model can be used to enhance and extract the texture features of the current concrete image. This realizes the target defect detection model based on the Single Shot Multi-Box Detector (SSD) model and the texture enhancement model to perform defect detection on concrete. The detection time is short, which improves the detection efficiency and accuracy of concrete defect detection.

[0055] Furthermore, constructing a target defect detection model based on a single-step multi-frame target detection model and a texture enhancement model is beneficial to improving the target defect detection model's ability to learn the texture features of the current concrete image, making concrete defects easier to detect and further improving the accuracy of concrete defect detection.

[0056] The defect detection results may include a first detection result indicating that the current concrete image contains current defect information, and a second detection result indicating that the current concrete image does not contain current defect information.

[0057] The current defect information can include defect type (e.g., crack, damage, exposed rebar), defect location, and defect size. The type of current defect information is not limited here and can be set according to actual needs.

[0058] The solution provided in this application, in response to a received detection command, acquires the current concrete image. The detection command instructs the concrete to be inspected for defects, and inputs the current concrete image into the target defect detection model to obtain the corresponding defect detection result. The target defect detection model is based on a single-step multi-frame target detection model and a texture enhancement model. The texture enhancement model is used to enhance and extract the texture features of the current concrete image. This solution enables the target defect detection model based on the single-step multi-frame target detection model and the texture enhancement model to perform defect detection on concrete. The detection time is short, which improves the detection efficiency and accuracy of concrete defect detection.

[0059] Furthermore, constructing a target defect detection model based on a single-step multi-frame target detection model and a texture enhancement model is beneficial to improving the target defect detection model's ability to learn the texture features of the current concrete image, making concrete defects easier to detect and further improving the accuracy of concrete defect detection.

[0060] Please see Figure 3 This illustrates a flowchart of a concrete defect detection method according to another embodiment of this application. In a specific embodiment, the concrete defect detection method can be applied to, for example... Figure 1 The concrete defect detection system shown includes a processing device 300. The following section uses processing device 300 as an example to discuss... Figure 3 The process shown is described in detail. The concrete defect detection method may include the following steps S210 to S250.

[0061] Step S210: Construct a texture enhancement model.

[0062] In this embodiment, the processing device can fuse the Central Difference Convolution (CDC) module, the Horizontal Differential Convolution (HDC) module, the Vertical Differential Convolution (VDC) module, the Angular Differential Convolution (ADC) module, and the Channel Shuffle (CS) module to obtain a texture enhancement model.

[0063] The CDC module is used to calculate the difference in texture features between the center pixel and the surrounding neighborhood of the current concrete image to obtain a first feature map, where the center pixel represents the pixel at the center of the current concrete image. The HDC module is used to extract the horizontal texture features of the first feature map to obtain a second feature map. The VDC module is used to extract the vertical texture features of the second feature map to obtain a third feature map. The ADC module is used to extract the texture features of the third feature map from multiple angles to obtain a fourth feature map. The CS module is used to enhance the cross-channel communication of the texture features of the first, second, third, and fourth feature maps to obtain a fifth feature map.

[0064] Texture features can be used to characterize the grayscale of the current concrete image, and texture feature differences can be used to characterize the grayscale differences of the current concrete image.

[0065] like Figure 4 As shown, the texture enhancement model can include a CDC module, an HDC module, a VDC module, an ADC module, and a CS module connected in sequence. The CDC module, HDC module, and VDC module can all be connected to the CS module via residuals (shortcuts). This shortcut connection allows the texture enhancement model to learn deeper texture features, which is beneficial for improving the model's ability to learn the texture features of the current concrete image.

[0066] Step S220: Fuse the single-step multi-frame target detection model and the texture enhancement model to obtain the initial defect detection model.

[0067] In this embodiment, after the processing device constructs the texture enhancement model, it can fuse the Visual Geometry Group (VGG) module of the SSD model, the top-down module in the Feature Pyramid Networks (FPN), and the texture enhancement model to obtain an initial defect detection model. The texture enhancement model can enhance and extract the texture features of the current concrete image, which is beneficial to improving the detection accuracy of the initial defect detection model in detecting defects in concrete.

[0068] The SSD model can include a VGG module and an FPN, and the FPN can include a top-down module and a bottom-up module.

[0069] The processing device can stitch the texture enhancement model between the VGG module and the top-down module in FPN to obtain an initial defect detection model. For example... Figure 5 As shown, the initial defect detection model may include a VGG module, a texture enhancement model, a top-down module, and a bottom-up module connected in sequence.

[0070] Step S230: Train the initial defect detection model to obtain the target defect detection model.

[0071] In this embodiment, after the processing device fuses the SSD model and the texture enhancement model to obtain the initial defect detection model, it can acquire historical concrete defect images and obtain corresponding sample sets based on the historical concrete defect images. The sample sets may include training sets and test sets. The training set is input into the initial defect detection model, which receives and responds to the training set and trains according to the training set to obtain the target defect detection model. The target defect detection model is obtained by training the initial defect detection model constructed based on the single-step multi-frame target detection model and the texture enhancement model. The texture enhancement model can enhance and extract the texture features of the current concrete image, which is beneficial to improving the detection accuracy of the target defect detection model in detecting concrete defects.

[0072] The processing equipment can annotate historical concrete defect images to obtain annotated images, and divide the annotated images according to preset division rules to obtain training and test sets.

[0073] When annotating historical concrete defect images, the processing equipment mainly annotates the defect type, defect location, and defect size in the historical concrete defect images.

[0074] For example, the processing equipment can use annotation boxes to annotate the defect types in historical concrete defect images, as well as the coordinates of the corner points of the annotation boxes and the size of the annotation boxes.

[0075] The preset partitioning rules can be manual. For example, a training set to test set ratio of 7:3 would result in 14,000 training images and 6,000 test images when there are 20,000 historical concrete defect images. Alternatively, a training set to test set ratio of 9:1 would result in 27,000 training images and 3,000 test images when there are 30,000 historical concrete defect images. The specific partitioning method is not limited here and can be set according to actual needs.

[0076] Regarding the process of acquiring historical concrete defect images by the aforementioned processing equipment, in some embodiments, the processing equipment pre-stores historical concrete defect images, and the processing equipment can read the pre-stored historical concrete defect images.

[0077] Regarding the process of the aforementioned processing equipment acquiring historical concrete defect images, in some embodiments, the processing equipment can generate prompt information and receive historical concrete defect images uploaded by inspection personnel based on the prompt information.

[0078] The notification information can be used to prompt inspectors to upload historical concrete defect images to the processing equipment. The notification information can include at least one of the following: text notifications, sound notifications, and light notifications. The type of notification information is not limited here and can be set according to actual needs.

[0079] Regarding the process of acquiring historical concrete defect images by the aforementioned processing equipment, in some embodiments, the concrete defect detection system may further include a detection service platform. The detection service platform pre-stores historical concrete defect images and can be connected to the processing equipment via a network to interact with the detection service platform through the network.

[0080] The processing device can send a second acquisition command to the detection service platform via the network. The detection service platform receives and responds to the second acquisition command, and sends historical concrete defect images to the processing device via the network. The processing device receives the historical concrete defect images returned by the detection service platform.

[0081] The detection service platform can be an independent physical cloud platform, a cloud platform cluster or distributed system composed of multiple physical cloud platforms, or any of the cloud platforms that provide basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), big data or artificial intelligence platforms, etc. There are no restrictions here.

[0082] In some implementations, after the processing device fuses the SSD model and the texture enhancement model to obtain an initial defect detection model, it can acquire historical concrete defect images and obtain corresponding sample sets based on these images. These sample sets may include training and testing sets. The training set is then augmented to obtain an enhanced training set, which is input into the initial defect detection model. The initial defect detection model receives and responds to the enhanced training set, and is trained based on it to obtain the target defect detection model. This achieves training the initial defect detection model using the enhanced training set, avoiding the lower robustness and generalization ability of the target defect detection model obtained from training the initial model with a limited training set due to a small number of historical concrete defect images. This improves the robustness and generalization ability of the target defect detection model.

[0083] The data enhancement processing can include at least one of the following: brightness enhancement processing, grayscale enhancement processing, contrast enhancement processing, affine transformation enhancement processing, and transparency enhancement processing. The type of data enhancement processing is not limited here, and can be set according to actual needs.

[0084] Step S240: In response to the received detection command, acquire the current concrete image.

[0085] Step S250: Input the current concrete image into the target defect detection model to obtain the corresponding defect detection results.

[0086] In this embodiment, steps S240 and S250 can be found in the corresponding steps in the foregoing embodiments, and will not be repeated here.

[0087] The solution provided in this embodiment constructs a texture enhancement model and integrates a single-step multi-frame target detection model with the texture enhancement model to obtain an initial defect detection model. The initial defect detection model is then trained to obtain a target defect detection model. In response to the received detection command, the current concrete image is acquired, and the current concrete image is input into the target defect detection model to obtain the corresponding defect detection result. This achieves defect detection of concrete based on the target defect detection model constructed by the single-step multi-frame target detection model and the texture enhancement model. The detection time is short, which improves the detection efficiency and accuracy of concrete defect detection.

[0088] Furthermore, by training the initial defect detection model based on the single-step multi-frame target detection model and the texture enhancement model, a target defect detection model is obtained. The texture enhancement model can enhance and extract the texture features of the current concrete image, which is beneficial to improving the detection accuracy of the target defect detection model in detecting defects in concrete.

[0089] Please see Figure 6 This document illustrates a flowchart of a concrete defect detection method provided in another embodiment of this application. In a specific embodiment, the concrete defect detection method can be applied to, for example... Figure 1 The concrete defect detection system shown includes a processing device 300. The following section uses processing device 300 as an example to discuss... Figure 6 The process shown is described in detail. The concrete defect detection method may include the following steps S310 to S340.

[0090] Step S310: In response to the received detection command, acquire the current concrete image.

[0091] Step S320: Input the current concrete image into the target defect detection model to obtain the corresponding defect detection results.

[0092] In this embodiment, steps S310 and S320 can be referred to the corresponding steps in the previous embodiments, and will not be repeated here.

[0093] Step S330: When the defect detection result is the first detection result used to characterize the current concrete image containing current defect information, input the first detection result into the remaining life prediction model to obtain the remaining life of the concrete.

[0094] In this embodiment, the processing device inputs the current concrete image to the target defect detection model and obtains the corresponding defect detection result. When the defect detection result is a first detection result used to characterize that the current concrete image contains current defect information, the first detection result is input to the remaining life prediction model. The remaining life prediction model receives and responds to the first detection result, obtains the remaining life of the concrete, and outputs the remaining life of the concrete to the processing device. The processing device receives the remaining life of the concrete output by the remaining life prediction model.

[0095] Among them, the remaining life prediction model can be obtained by training a deep learning neural network model based on historical defect information.

[0096] Historical defect information can include defect type, defect location, and defect size. The type of historical defect information is not limited here and can be set according to actual needs.

[0097] Deep learning neural network models can be Convolutional Neural Networks (CNN), Deep Belief Networks (DBN), Stacked Auto Encoder Networks (SAE), Recurrent Neural Networks (RNN), Deep Neural Networks (DNN), Long Short-Term Memory (LSTM), or Gated Recurring Units (GRU), etc. The type of deep learning neural network model is not limited here; it can be set according to actual needs.

[0098] Step S340: When the remaining life of the concrete is less than or equal to the remaining life threshold, generate the corresponding warning message.

[0099] In this embodiment, when the remaining life of the concrete is less than or equal to the remaining life threshold, a corresponding warning message can be generated to alert the inspection personnel so that they can deal with the concrete in a timely manner. This can suppress the problem of increased safety risk of concrete when the remaining life of the concrete is low and the minimum remaining life is reached, thereby reducing the safety risk of the concrete.

[0100] The remaining life threshold can be used to characterize the minimum remaining life of concrete at which a safety risk occurs. The remaining life threshold can be a life value preset by the inspector, or it can be a life value automatically generated by the processing equipment based on the process of multiple defect detections of concrete. There is no limitation here, and it can be set according to actual needs.

[0101] Warning messages can be at least one of the following: sound warning messages, text warning messages, or light warning messages. There is no limitation on the type of warning message here, and the specific settings can be made according to actual needs.

[0102] The solution provided in this embodiment, in response to a received detection command, acquires the current concrete image and inputs it into the target defect detection model to obtain the corresponding defect detection result. When the defect detection result is a first detection result used to characterize that the current concrete image contains current defect information, the first detection result is input into the remaining life prediction model to obtain the remaining life of the concrete. When the remaining life of the concrete is less than or equal to the remaining life threshold, a corresponding warning message is generated. This solution enables the target defect detection model based on a single-step multi-frame target detection model and a texture enhancement model to perform defect detection on concrete. The detection time is short, improving the detection efficiency and accuracy of concrete defect detection.

[0103] Furthermore, when the remaining life of the concrete is less than or equal to the remaining life threshold, a corresponding warning message can be generated to alert the testing personnel so that they can take timely action to address the concrete issue. This can mitigate the problem of increased safety risks associated with concrete when its remaining life is low and the minimum remaining life threshold is reached, thereby reducing the overall safety risk of the concrete.

[0104] Please see Figure 7 This document illustrates a flowchart of a concrete defect detection method according to another embodiment of this application. In a specific embodiment, the concrete defect detection method can be applied to, for example... Figure 1 The concrete defect detection system shown includes a processing device 300. The following section uses processing device 300 as an example to discuss... Figure 7 The process shown is described in detail. The concrete defect detection method may include the following steps S410 to S440.

[0105] Step S410: In response to the received detection command, acquire the current concrete image.

[0106] Step S420: Input the current concrete image into the target defect detection model to obtain the corresponding defect detection results.

[0107] In this embodiment, steps S410 and S420 can be referred to the corresponding steps in the previous embodiments, and will not be repeated here.

[0108] Step S430: When the defect detection result is a first detection result used to characterize the current concrete image containing current defect information, obtain the defect processing scheme corresponding to the first detection result.

[0109] In this embodiment, after the processing device inputs the current concrete image to the target defect detection model and obtains the corresponding defect detection result, when the defect detection result is a first detection result used to characterize that the current concrete image contains current defect information, the defect processing scheme corresponding to the first detection result can be obtained.

[0110] In some implementations, the processing device may pre-store a scheme table, which can be used to characterize the correspondence between defect information and processing schemes.

[0111] After the processing device inputs the current concrete image into the target defect detection model and obtains the corresponding defect detection result, when the defect detection result is the first detection result used to characterize the current concrete image containing the current defect information, the solution table can be searched based on the first detection result to obtain the defect processing solution.

[0112] For example, defect information may include first sub-defect information (defect type A1, defect location B1, and defect size C1), second sub-defect information (defect type A1, defect location B1, and defect size C2), third sub-defect information (defect type A1, defect location B2, and defect size C1), fourth sub-defect information (defect type A1, defect location B2, and defect size C2), fifth sub-defect information (defect type A2, defect location B1, and defect size C1), sixth sub-defect information (defect type A2, defect location B1, and defect size C2), seventh sub-defect information (defect type A2, defect location B2, and defect size C1), and eighth sub-defect information (defect type A2, defect location B2, and defect size C2).

[0113] The processing schemes may include processing scheme D1, processing scheme D2, processing scheme D3, processing scheme D4, processing scheme D5, processing scheme D6, processing scheme D7, and processing scheme D8.

[0114] The correspondence between defect information and handling solutions can be shown in Table 1, i.e., the solution table. Based on this correspondence, defect handling solutions can be obtained.

[0115] Table 1

[0116] Defect Information Solution First sub-defect information D2 Second sub-defect information D5 Third sub-defect information D1 Fourth Sub-Defect Information D3 Fifth Sub-Defect Information D4 Sixth Sub-Defect Information D8 Seventh Sub-Defect Information D6 Eighth Sub-Defect Information D7

[0117] It should be noted that the correspondence between defect information and handling solutions is not limited to that shown in Table 1, and can be set according to actual needs.

[0118] In some implementations, the concrete defect detection system may also include a detection service platform, which may pre-store a solution table, which can be used to establish the correspondence between defect information and treatment solutions.

[0119] After the processing device inputs the current concrete image into the target defect detection model and obtains the corresponding defect detection result, when the defect detection result is the first detection result used to characterize that the current concrete image contains the current defect information, the first detection result can be sent to the detection service platform via the network. The detection service platform receives and responds to the first detection result, looks up the solution table, obtains the defect handling solution, and sends the defect handling solution to the processing device via the network. The processing device receives the defect handling solution returned by the detection service platform.

[0120] Step S440: Send the defect handling plan to the client associated with the inspection personnel.

[0121] In this embodiment, the concrete defect detection system may also include a client associated with the inspector. When the defect detection result is a first detection result that characterizes the current concrete image as containing current defect information, the processing device obtains the defect handling plan corresponding to the first detection result and then sends the defect handling plan to the client via the network so that the inspector can process the concrete in a timely manner according to the defect handling plan, thereby improving the user experience of the inspector during the process of detecting defects in concrete.

[0122] The solution provided in this embodiment, in response to a received detection command, acquires the current concrete image and inputs it into the target defect detection model to obtain the corresponding defect detection result. When the defect detection result is a first detection result used to characterize that the current concrete image contains current defect information, the solution obtains the defect handling scheme corresponding to the first detection result and sends the defect handling scheme to the client associated with the inspection personnel. This realizes the target defect detection model based on a single-step multi-frame target detection model and a texture enhancement model to perform defect detection on concrete. The detection time is short, which improves the detection efficiency and accuracy of concrete defect detection.

[0123] Furthermore, the defect handling plan corresponding to the first test result is sent to the client associated with the tester, so that the tester can handle the concrete in a timely manner according to the defect handling plan, thereby improving the user experience of the tester in the process of detecting defects in concrete.

[0124] Please see Figure 8This illustrates a concrete defect detection device 500 provided in one embodiment of this application. The concrete defect detection device 500 can be applied to, for example... Figure 1 The concrete defect detection system shown includes a processing device 300. The following section uses processing device 300 as an example to discuss... Figure 8 The concrete defect detection device 500 shown will be described in detail. The concrete defect detection device 500 may include a first acquisition module 510 and a first input module 520.

[0125] The first acquisition module 510 can be used to acquire the current concrete image in response to the received detection command. The detection command can be used to instruct the concrete to perform defect detection. The first input module 520 can be used to input the current concrete image into the target defect detection model to obtain the corresponding defect detection result. The target defect detection model can be obtained based on the single-step multi-frame target detection model and the texture enhancement model. The texture enhancement model can be used to enhance and extract the texture features of the current concrete image.

[0126] In some embodiments, the concrete defect detection device 500 may also include a construction module, a fusion module, and a training module.

[0127] The construction module can be used to construct a texture enhancement model before the first acquisition module 510 acquires the current concrete image of the concrete in response to the received detection command; the fusion module can be used to fuse the single-step multi-box target detection model and the texture enhancement model to obtain an initial defect detection model; the training module can be used to train the initial defect detection model to obtain a target defect detection model.

[0128] In some implementations, the building module may include a fusion unit.

[0129] The fusion unit can be used to fuse the central difference convolution module, the horizontal difference convolution module, the vertical difference convolution module, the angular difference convolution module, and the channel shuffling module to obtain a texture enhancement model.

[0130] The center difference convolution module can be used to calculate the difference in texture features between the center pixel of the current concrete image and its surrounding neighborhood to obtain the first feature map; the horizontal difference convolution module can be used to extract the texture features in the horizontal direction of the first feature map to obtain the second feature map; the vertical difference convolution module can be used to extract the texture features in the vertical direction of the second feature map to obtain the third feature map; the angular difference convolution module can be used to extract the texture features in multiple angles of the third feature map to obtain the fourth feature map; the channel shuffling module can be used to enhance the cross-channel communication of the texture features of the first feature map, the second feature map, the third feature map, and the fourth feature map to obtain the fifth feature map.

[0131] In some implementations, the training module may include an acquisition unit and an input unit.

[0132] The acquisition unit can be used to acquire the corresponding sample set based on historical concrete defect images. The sample set can include at least the training set. The input unit can be used to input the training set into the initial defect detection model for training to obtain the target defect detection model.

[0133] In some embodiments, the concrete defect detection device 500 may also include a reinforcement module.

[0134] The augmentation module can be used to augment the training set before the target defect detection model is obtained by inputting the training set into the input unit for training.

[0135] In some implementations, the input unit may include an input subunit.

[0136] The input sub-unit can be used to input the enhanced training set into the initial defect detection model for training, thereby obtaining the target defect detection model.

[0137] In some embodiments, the concrete defect detection device 500 may also include a second input module and a generation module.

[0138] The second input module can be used to input the first detection result into the remaining life prediction model when the defect detection result is a first detection result used to characterize the current concrete image containing current defect information, so as to obtain the remaining life of the concrete. The remaining life prediction model can be obtained by training a deep learning neural network model based on historical defect information. The generation module can be used to generate corresponding warning information when the remaining life of the concrete is less than or equal to the remaining life threshold.

[0139] In some embodiments, the concrete defect detection device 500 may also include a second acquisition module and a transmission module.

[0140] The second acquisition module can be used to acquire the defect handling scheme corresponding to the first detection result when the defect detection result is a first detection result used to characterize the current concrete image containing the current defect information; the sending module can be used to send the defect handling scheme to the client associated with the inspection personnel.

[0141] The solution provided in this embodiment, in response to a received detection command, acquires the current concrete image. The detection command instructs the concrete to be inspected for defects, and inputs the current concrete image into the target defect detection model to obtain the corresponding defect detection result. The target defect detection model is based on a single-step multi-frame target detection model and a texture enhancement model. The texture enhancement model is used to enhance and extract the texture features of the current concrete image. This realizes the target defect detection model based on the single-step multi-frame target detection model and the texture enhancement model to perform defect detection on concrete. The detection time is short, which improves the detection efficiency and accuracy of concrete defect detection.

[0142] Furthermore, constructing a target defect detection model based on a single-step multi-frame target detection model and a texture enhancement model is beneficial to improving the target defect detection model's ability to learn the texture features of the current concrete image, making concrete defects easier to detect and further improving the accuracy of concrete defect detection.

[0143] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For device embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to in the descriptions of the method embodiments. Any processing method described in the method embodiments can be implemented in the device embodiments through corresponding processing modules, and will not be elaborated upon further in the device embodiments.

[0144] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0145] Please see Figure 9 The diagram illustrates a functional block diagram of an electronic device 600 provided in one embodiment of the present application. The electronic device 600 may include one or more of the following components: a memory 610, a processor 620, and one or more application programs. One or more application programs may be stored in the memory 610 and configured to be executed by one or more processors 620. One or more application programs are configured to perform the methods as described in the foregoing method embodiments.

[0146] The memory 610 may include random access memory (RAM) or read-only memory. The memory 610 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 610 may include a program storage area and a data storage area. The program storage area may store instructions for implementing an operating system, instructions for implementing at least one function (such as receiving detection instructions, responding to detection instructions, acquiring the current concrete image, defect detection, inputting the current concrete image, acquiring defect detection results, enhancing texture features, extracting texture features, constructing a texture enhancement model, fusing a single-step multi-box target detection model and a texture enhancement model, obtaining an initial defect detection model, training the initial defect detection model, obtaining a target defect detection model, a fusion module, obtaining a texture enhancement model, calculating texture feature differences, obtaining a first feature map, extracting texture features, obtaining a second feature map, obtaining a third feature map, obtaining a fourth feature map, mixing texture features, obtaining a fifth feature map, acquiring a sample set, inputting a training set, data augmentation processing, obtaining an enhanced training set, inputting the enhanced training set, inputting the first detection result, obtaining the remaining life of concrete, training a deep learning neural network model, obtaining a remaining life prediction model, generating warning information, acquiring a defect handling plan, and sending the defect handling plan, etc.), and instructions for implementing the various method embodiments described below. The data storage area can also store data created by the electronic device 600 during use (such as detection instructions, concrete, current concrete image, target defect detection model, defect detection result, single-step multi-box target detection model, texture enhancement model, texture features, initial defect detection model, center difference convolution module, horizontal difference convolution module, vertical difference convolution module, angle difference convolution module, channel shuffling module, center pixel, surrounding neighborhood, texture feature difference, first feature map, horizontal direction, second feature map, vertical direction, third feature map, multiple angles, fourth feature map, fifth feature map, historical concrete defect images, sample set, training set, enhanced training set, current defect information, first detection result, remaining life prediction model, remaining concrete life, historical defect information, deep learning neural network model, remaining life threshold, warning information, defect handling plan, detection personnel, and client).

[0147] Processor 620 may include one or more processing cores. Processor 620 connects to various parts within the electronic device 600 using various interfaces and lines, and performs various functions and processes data of the electronic device 600 by running or executing instructions, programs, code sets, or instruction sets stored in memory 610, and by calling data stored in memory 610. Optionally, processor 620 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). Processor 620 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and Modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the displayed content; and the modem handles wireless communication. It is understood that the modem may also not be integrated into processor 620 and may be implemented separately using a communication chip.

[0148] Please refer to Figure 10 This diagram illustrates a structural block diagram of a computer-readable storage medium provided in an embodiment of this application. The computer-readable storage medium 700 stores program code 710, which can be called by a processor to execute the methods described in the above method embodiments.

[0149] The computer-readable storage medium 700 may be an electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read-Only Memory), EPROM, hard disk, or ROM. Optionally, the computer-readable storage medium 700 includes a non-transitory computer-readable storage medium. The computer-readable storage medium 700 has storage space for program code 710 that performs any of the method steps described above. This program code can be read from or written to one or more computer program products. The program code 710 may, for example, be compressed in a suitable form.

[0150] Please refer to Figure 11This diagram illustrates a structural block diagram of a computer program product 800 provided in an embodiment of this application. The computer program product 800 includes a computer program / instructions 810, which is stored in a computer-readable storage medium of a computer device. When the computer program product 800 is executed on the computer device, the processor of the computer device reads the computer program / instructions 810 from the computer-readable storage medium, and executes the computer program / instructions 810, causing the computer device to perform the methods described in the above method embodiments.

[0151] The solution provided in this embodiment, in response to a received detection command, acquires the current concrete image. The detection command instructs the concrete to be inspected for defects, and inputs the current concrete image into the target defect detection model to obtain the corresponding defect detection result. The target defect detection model is based on a single-step multi-frame target detection model and a texture enhancement model. The texture enhancement model is used to enhance and extract the texture features of the current concrete image. This realizes the target defect detection model based on the single-step multi-frame target detection model and the texture enhancement model to perform defect detection on concrete. The detection time is short, which improves the detection efficiency and accuracy of concrete defect detection.

[0152] Furthermore, constructing a target defect detection model based on a single-step multi-frame target detection model and a texture enhancement model is beneficial to improving the target defect detection model's ability to learn the texture features of the current concrete image, making concrete defects easier to detect and further improving the accuracy of concrete defect detection.

[0153] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for detecting defects in concrete, characterized in that, This is applied to a concrete defect detection system, which includes a camera and processing equipment. The processing device acquires an initial defect detection model, which is obtained by concatenating a texture enhancement model between the visual geometry module of the single-step multi-box target detection model and the top-down module in the feature pyramid network. The texture enhancement model is used to enhance and extract texture features from the concrete image. The texture enhancement model is obtained by fusing a central difference convolution module, a horizontal difference convolution module, a vertical difference convolution module, an angular difference convolution module, and a channel shuffling module. These modules are connected sequentially, and all three modules are connected to the channel shuffling module via residuals. The central difference convolution module is used to calculate the difference in texture features between the center pixel of the current concrete image and its surrounding neighborhood to obtain the first feature map, where the center pixel represents the pixel at the center position of the current concrete image; the horizontal difference convolution module is used to extract the horizontal texture features of the first feature map to obtain the second feature map; the vertical difference convolution module is used to extract the vertical texture features of the second feature map to obtain the third feature map; the angular difference convolution module is used to extract the texture features of the third feature map from multiple angles to obtain the fourth feature map; the channel shuffling module is used to enhance the cross-channel communication of the texture features of the first, second, third, and fourth feature maps to obtain the fifth feature map; The processing device uses a training set to train the initial defect detection model to obtain a target defect detection model; the training set includes at least one historical concrete defect image, the historical concrete defect image is labeled, the label mainly refers to the defect type, defect location and defect size in the historical concrete defect image. In response to the received detection command, the processing device sends a first acquisition command to the camera. The first acquisition command is used to trigger the camera to acquire the current concrete image. In response to the first acquisition command, the camera acquires and sends the current concrete image to the processing device. The detection command is used to instruct the concrete to perform defect detection. The processing device inputs the current concrete image to the target defect detection model to obtain the corresponding defect detection result. The defect detection result includes a first detection result indicating that the current concrete image contains current defect information, and a second detection result indicating that the current concrete image does not contain current defect information. The current defect information includes defect type, defect location, and defect size. When the defect detection result is a first detection result used to characterize that the current concrete image contains current defect information, the defect processing scheme corresponding to the first detection result is obtained.

2. The concrete defect detection method according to claim 1, characterized in that, The central difference convolution module is used to calculate the difference in texture features between the center pixel of the current concrete image and its surrounding neighborhood to obtain a first feature map; the horizontal difference convolution module is used to extract the texture features in the horizontal direction of the first feature map to obtain a second feature map; the vertical difference convolution module is used to extract the texture features in the vertical direction of the second feature map to obtain a third feature map; the angular difference convolution module is used to extract the texture features in multiple angles of the third feature map to obtain a fourth feature map; the channel shuffling module is used to enhance the cross-channel communication of the texture features of the first feature map, the second feature map, the third feature map, and the fourth feature map to obtain a fifth feature map.

3. The concrete defect detection method according to claim 1, characterized in that, Before inputting the training set into the initial defect detection model for training to obtain the target defect detection model, the concrete defect detection method further includes: The training set is augmented to obtain an augmented training set; The step of inputting the training set into the initial defect detection model for training to obtain the target defect detection model includes: The enhanced training set is input into the initial defect detection model for training to obtain the target defect detection model.

4. The concrete defect detection method according to claim 1, characterized in that, Also includes: When the defect detection result is a first detection result used to characterize that the current concrete image contains current defect information, the first detection result is input into the remaining life prediction model to obtain the remaining life of the concrete. The remaining life prediction model is obtained by training a deep learning neural network model based on historical defect information. When the remaining life of the concrete is less than or equal to the remaining life threshold, a corresponding warning message is generated.

5. The method for detecting concrete defects according to any one of claims 1 to 4, characterized in that, After obtaining the defect handling plan corresponding to the first detection result, the method further includes: Send the defect handling plan to the client associated with the testing personnel.

6. A concrete defect detection device, characterized in that, This is applied to a concrete defect detection system, which includes a camera and processing equipment: the processing equipment includes: The fusion unit is used to fuse the central difference convolution module, the horizontal difference convolution module, the vertical difference convolution module, the angular difference convolution module, and the channel shuffling module to obtain a texture enhancement model. The central difference convolution module, the horizontal difference convolution module, the vertical difference convolution module, the angular difference convolution module, and the channel shuffling module are sequentially connected. The central difference convolution module, the horizontal difference convolution module, and the vertical difference convolution module are all connected to the channel shuffling module via residuals. The texture enhancement model is used to enhance and extract the texture features of the concrete image. The central difference convolution module is used to calculate the center pixel of the current concrete image and... The texture feature difference of the surrounding neighborhood is used to obtain the first feature map, and the center pixel is used to represent the pixel point at the center position of the current concrete image; the horizontal difference convolution module is used to extract the texture features in the horizontal direction of the first feature map to obtain the second feature map; the vertical difference convolution module is used to extract the texture features in the vertical direction of the second feature map to obtain the third feature map; the angular difference convolution module is used to extract the texture features in multiple angles of the third feature map to obtain the fourth feature map; the channel shuffling module is used to enhance the cross-channel communication of the texture features of the first feature map, the second feature map, the third feature map, and the fourth feature map to obtain the fifth feature map; The fusion module is used to splice the texture enhancement model between the visual geometry group module of the single-step multi-box target detection model and the top-down module in the feature pyramid network to obtain the initial defect detection model. The training module is used to train the initial defect detection model using the training set to obtain the target defect detection model; the training set includes at least one historical concrete defect image, the historical concrete defect image is labeled, the label mainly refers to the defect type, defect location and defect size in the historical concrete defect image. The first acquisition module is used to send a first acquisition instruction to the camera in response to a received detection instruction. The first acquisition instruction is used to trigger the camera to acquire the current concrete image of the concrete. In response to the first acquisition instruction, the camera acquires and sends the current concrete image of the concrete to the processing device. The detection instruction is used to instruct the concrete to perform defect detection. The first input module is used to input the current concrete image into the target defect detection model to obtain the corresponding defect detection result. The defect detection result includes a first detection result indicating that the current concrete image contains current defect information, and a second detection result indicating that the current concrete image does not contain current defect information. The current defect information includes defect type, defect location, and defect size. When the defect detection result is the first detection result indicating that the current concrete image contains current defect information, the defect processing scheme corresponding to the first detection result is obtained.

7. An electronic device, characterized in that, include: Memory; One or more processors are coupled to the memory; One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more applications being configured to perform the concrete defect detection method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium contains program code, which can be called by a processor to execute the concrete defect detection method as described in any one of claims 1 to 5.

Images