Concrete construction defect detection and repair system based on image recognition

Abstract

The embodiment of the invention discloses a concrete construction defect detection and repair system based on image recognition. According to a specific embodiment of the system, concrete ultrasonic equipment is configured to acquire a concrete defect image set and send the concrete defect image set to image processing equipment in response to determining that defects exist on the surface of concrete; the image processing equipment is configured to generate a marked concrete defect image set according to the received concrete defect image set, and send the marked concrete defect image set to the defect feature detection equipment; and the concrete defect repairing equipment is configured to perform concrete defect repairing on each concrete defect position coordinate contained in the early warning information in response to the received early warning information sent by the early warning terminal. According to the embodiment, the quality of defect repairing is improved, the timeliness of defect repairing is improved, and the period of defect repairing is shortened.

Description

Technical Field

[0001] The embodiments disclosed herein relate to the field of concrete construction defect repair, and more specifically to a concrete construction defect detection and repair system based on image recognition. Background Technology

[0002] Image recognition-based concrete construction defect repair is a technology for repairing defects in concrete after construction. Currently, the common method for repairing concrete defects is to visually inspect the concrete surface for defects. For example, when repairing concrete defects, it is necessary to first visually inspect the concrete for defects before proceeding with the repair.

[0003] However, the inventors discovered that when using the above method to repair defects in concrete, the following technical problems often arise: Tiny defects on concrete surfaces are difficult to detect with the naked eye, resulting in poor quality repairs. Furthermore, the slow speed of visual inspection makes it difficult to cover large areas of concrete in a short time, leading to prolonged repair cycles. The inability to provide real-time feedback on inspection results further delays repairs, increasing the risk of wall or column collapse.

[0004] The information disclosed in this background section is only intended to enhance the understanding of the background of the inventive concept, and therefore may contain information that does not form prior art known to those skilled in the art. Summary of the Invention

[0005] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0006] Some embodiments of this disclosure propose an image recognition-based concrete construction defect detection and repair system to address the technical problems mentioned in the background section above.

[0007] In a first aspect, some embodiments of this disclosure provide a concrete construction defect detection and repair system based on image recognition. This system includes: a concrete ultrasonic device, an image processing device, a defect feature detection device, an early warning terminal, a concrete defect repair device, and a concrete layer segmentation device. Specifically: the concrete ultrasonic device is configured to: in response to determining that a defect exists on the concrete surface, acquire a set of concrete defect images and send the set of concrete defect images to the image processing device, wherein the set of concrete defect images is a collection of defect images taken after the concrete used in construction has solidified; the image processing device is configured to: generate an annotated set of concrete defect images based on the received set of concrete defect images and send the annotated set of concrete defect images to the defect feature detection device; the defect feature detection device is configured to: input the received annotated set of concrete defect images into a pre-trained defect feature detection model to obtain concrete defect image features. The system receives and distributes the received concrete defect image feature set to a concrete layer segmentation device, wherein the concrete defect image features in the received concrete defect image feature set characterize the concrete defect features of the image; the concrete layer segmentation device is configured to: perform layer segmentation on the received concrete defect image feature set to obtain a segmented concrete defect image feature set, and send the segmented concrete defect image feature set to an early warning terminal, wherein the image defect degree corresponding to each segmented concrete defect image feature set in the segmented concrete defect image feature set is different; the early warning terminal is configured to: generate early warning information corresponding to the segmented concrete defect image feature set, and send the early warning information to a concrete defect repair device, wherein the early warning information includes the coordinates of each concrete defect location; the concrete defect repair device is configured to: in response to receiving the early warning information sent by the early warning terminal, perform concrete defect repair on the concrete defect location coordinates included in the early warning information.

[0008] Secondly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the concrete construction defect detection and repair system described in any implementation of the first aspect above.

[0009] Thirdly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the concrete construction defect detection and repair system described in any of the implementations of the first aspect above.

[0010] The above-described embodiments of this disclosure have the following beneficial effects: the image recognition-based concrete construction defect detection and repair system of some embodiments of this disclosure improves the quality of defect repair, enhances the timeliness of defect repair, and shortens the defect repair cycle. Specifically, the reasons for poor defect repair quality, untimely defect repair, and long defect repair cycle are: it is difficult to detect tiny defects on the concrete surface with the naked eye, resulting in poor defect repair quality; the slow speed of naked eye detection makes it difficult to cover a large area of ​​the concrete surface in a short time, leading to a long defect repair cycle; and the inability to provide real-time feedback on the detected defect results leads to untimely defect repair, resulting in a long defect repair cycle, thereby increasing the risk of wall or column collapse. Based on this, the image recognition-based concrete construction defect detection and repair system of some embodiments of this disclosure firstly introduces the detection of concrete defect images using ultrasonic concrete equipment to accurately acquire concrete defect images. Specifically, the ultrasonic concrete equipment is configured to: in response to determining that a defect exists on the concrete surface, acquire a set of concrete defect images and send the set of concrete defect images to an image processing device, wherein the set of concrete defect images is a collection of defect images taken after the concrete used in construction has solidified. Therefore, by using ultrasonic testing equipment to inspect concrete surfaces without visual inspection, the quality of defect repair is improved. The image processing equipment is then configured to generate an annotated set of concrete defect images from the received set of concrete defect images, and to send this annotated set to a defect feature detection device. This allows for data preprocessing of the concrete defect image set, enhancing the features of the concrete defect images. Subsequently, the defect feature detection device is configured to input the received annotated set of concrete defect images into a pre-trained defect feature detection model to obtain a concrete defect image feature set, and to send this feature set to a concrete layer segmentation device. The concrete defect image features in this feature set characterize the concrete defect features of the image. This increases the detection speed, enabling the coverage of large areas of concrete surfaces in a short time and shortening the defect repair cycle. Subsequently, the aforementioned concrete layering device is configured to: perform layered segmentation on the received concrete defect image feature set to obtain segmented concrete defect image feature sets; and send the segmented concrete defect image feature sets to an early warning terminal. The different segmented concrete defect image feature sets correspond to different degrees of image defects. This allows for the segmentation of concrete defect images of the same type, facilitating subsequent processing.Next, the aforementioned early warning terminal is configured to generate early warning information corresponding to the segmented concrete defect image feature set, and to send the early warning information to the concrete defect repair equipment. The early warning information includes the coordinates of each concrete defect location. This allows for real-time feedback of the detected defect results, improving the timeliness of defect repair. Finally, the aforementioned concrete defect repair equipment is configured to, in response to receiving the early warning information from the aforementioned early warning terminal, perform concrete defect repair based on the coordinates of each concrete defect location included in the early warning information. This allows for real-time feedback of the detected defect results, timely defect repair, shortened defect repair cycle, and thus reduced risk of wall or column collapse. Attached Figure Description

[0011] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0012] Figure 1 This is a schematic diagram of an application scenario of an image recognition-based concrete construction defect detection and repair system according to some embodiments of this disclosure; Figure 2 This is a flowchart of some embodiments of the image recognition-based concrete construction defect detection and repair system according to the present disclosure; Figure 3 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0013] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0014] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0015] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0016] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0017] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0018] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0019] Figure 1 This is a schematic diagram of an application scenario of an image recognition-based concrete construction defect detection and repair system according to some embodiments of this disclosure.

[0020] exist Figure 1In application scenarios, firstly, the concrete construction defect detection and repair system based on image recognition can include: a concrete ultrasonic device 101, an image processing device 102, a defect feature detection device 103, an early warning terminal 104, a concrete defect repair device 105, and a concrete layer division device 106. First, the aforementioned ultrasonic concrete device 101 is configured to: in response to determining that a defect exists on the concrete surface, acquire a set of concrete defect images and send the set of concrete defect images to an image processing device 102, wherein the set of concrete defect images is a collection of defect images taken after the concrete used in construction has solidified; then, the image processing device 102 is configured to: generate an annotated set of concrete defect images based on the received set of concrete defect images and send the annotated set of concrete defect images to a defect feature detection device 103; next, the defect feature detection device 103 is configured to: input the received annotated set of concrete defect images into a pre-trained defect feature detection model to obtain a set of concrete defect image features, and send the set of concrete defect image features to a concrete layer segmentation device 106, wherein the concrete defect images in the set of concrete defect image features are... The first step involves identifying the concrete defect features in the image. The second step involves the concrete layering device 106 being configured to: perform layered segmentation on the received concrete defect image feature set to obtain a segmented set of concrete defect image features, and send this segmented set of features to the early warning terminal 104. The image defect severity varies among the segmented concrete defect image feature sets. The third step involves the early warning terminal 104 being configured to: generate early warning information corresponding to the segmented set of concrete defect image features, and send this early warning information to the concrete defect repair device 105. The early warning information includes the coordinates of each concrete defect location. Finally, the concrete defect repair device 105 is configured to: in response to receiving the early warning information from the early warning terminal, repair the concrete defects based on the coordinates of each concrete defect location included in the early warning information.

[0021] It should be noted that the aforementioned image recognition-based concrete construction defect detection and repair system can be either hardware or software. When the computing device is hardware, it can be implemented as a distributed cluster of multiple servers or terminal devices, or as a single server or terminal device. When the computing device is software, it can be installed on the hardware devices listed above. It can be implemented as, for example, multiple software programs or software modules to provide distributed services, or as a single software program or software module. No specific limitations are made here. It should be understood that... Figure 1In the image recognition-based concrete construction defect detection and repair system, the number of concrete ultrasonic equipment, image processing equipment, defect feature detection equipment, and concrete defect repair equipment can be arbitrary, depending on the implementation needs.

[0022] Continue to refer to Figure 2 The diagram illustrates a flow 200 of some embodiments of an image recognition-based concrete construction defect detection and repair system according to the present disclosure. This image recognition-based concrete construction defect detection and repair system includes: a concrete ultrasonic device, an image processing device, a defect feature detection device, an early warning terminal, a concrete defect repair device, and a concrete layer segmentation device, and includes the following steps: Step 201, the concrete ultrasonic device is configured to: in response to determining that a defect is detected on the concrete surface, acquire a set of concrete defect images, and send the set of concrete defect images to an image processing device.

[0023] In some embodiments, the concrete ultrasonic device is configured to: in response to determining that a defect is detected on the concrete surface, acquire a set of concrete defect images and send the set of concrete defect images to an image processing device, wherein the set of concrete defect images is a collection of defect images taken after the concrete used in building construction has solidified.

[0024] Here, the defect images in the aforementioned defect image set can refer to images of the surface defects in the concrete of walls or columns. For example, the defect images in the aforementioned defect image set can refer to honeycomb images formed by the concrete of walls or columns. Alternatively, the defect images in the aforementioned defect image set can refer to pitted images formed by the concrete of walls or columns. The aforementioned ultrasonic concrete device can be used to detect defects and crack depths in concrete. For example, the aforementioned ultrasonic concrete device can refer to the HC-U81 ultrasonic concrete detector. The aforementioned image processing device can be used to process the image data. For example, the aforementioned image processing device can refer to a computing terminal containing Adobe Photoshop (image processing software). Here, the ultrasonic concrete device can establish a communication connection with the aforementioned image processing device within a certain distance via wireless communication, allowing the image processing device to receive the set of concrete defect images from the ultrasonic concrete device.

[0025] Step 202: The image processing device is configured to generate an annotated concrete defect image set based on the received concrete defect image set, and to send the annotated concrete defect image set to the defect feature detection device.

[0026] In some embodiments, the image processing device is configured to: generate an annotated set of concrete defect images based on the received set of concrete defect images, and send the annotated set of concrete defect images to the defect feature detection device.

[0027] Here, the labeled concrete defect images in the aforementioned set of labeled concrete defect images can refer to images of concrete defects after solidification, with the coordinates of the labeled area indicated. The aforementioned area coordinates can refer to (x, y). The aforementioned defect feature detection equipment can be used to detect defect features in the images. For example, the aforementioned defect feature detection equipment can refer to a FLIR E8 thermal imaging camera. Here, the defect feature detection equipment can establish a communication connection with the aforementioned image processing equipment within a certain distance via wireless communication.

[0028] In some optional implementations of certain embodiments, the image processing device generates an annotated set of concrete defect images based on the received set of concrete defect images, including: The first step is to crop the above set of concrete defect images to obtain a cropped set of concrete defect images.

[0029] Here, the cropped concrete defect images in the above-mentioned set of cropped concrete defect images can refer to images containing areas of concrete defects.

[0030] As an example, an image processing device can use an edge detection algorithm to crop the aforementioned set of concrete defect images, resulting in a cropped set of concrete defect images. The aforementioned edge detection algorithm may refer to Canny edge detection.

[0031] The second step is to perform denoising on the above cropped concrete defect image set to obtain a denoised concrete defect image set.

[0032] As an example, the image processing device can use Gaussian filtering to denoise the above-mentioned cropped concrete defect image set to obtain a denoised concrete defect image set.

[0033] The third step is to sharpen the concrete defect image set after the above denoising process to obtain a sharpened concrete defect image set.

[0034] As an example, the image processing device can use Laplacian sharpening to sharpen the above-mentioned denoised concrete defect image set to obtain a sharpened concrete defect image set.

[0035] The fourth step is to enhance the contrast of the above-sharpened concrete defect image set to obtain the enhanced concrete defect image set.

[0036] As an example, the image processing device can use Adaptive histogram equalization (AHE, CLAHE) to enhance the contrast of the above-sharpened concrete defect image set, resulting in an enhanced concrete defect image set.

[0037] The fifth step is to adjust the lighting on the above-mentioned enhanced concrete defect image set to obtain the adjusted concrete defect image set.

[0038] As an example, the image processing device can adjust the illumination of the above-mentioned enhanced concrete defect image set through histogram normalization to obtain the adjusted concrete defect image set.

[0039] The sixth step is to annotate the adjusted concrete defect image set to obtain an annotated concrete defect image set.

[0040] As an example, the image processing device can locate each adjusted concrete defect image in the adjusted concrete defect image set to generate adjusted concrete defect image coordinates, thus obtaining an adjusted concrete defect image coordinate set. Then, the adjusted concrete defect image coordinate set is labeled to each adjusted concrete defect image in the adjusted concrete defect image set to generate labeled concrete defect images, thus obtaining a labeled concrete defect image set.

[0041] In some optional implementations of certain embodiments, the image processing device may perform data annotation on the adjusted concrete defect image set described above by the following steps: The first step is to determine the bounding box position information for each adjusted concrete defect image in the above-mentioned adjusted concrete defect image set, so as to generate the concrete defect image bounding box position information set.

[0042] Here, the bounding box position information mentioned above refers to the position information of the defect area locked by the bounding box in the adjusted concrete defect image. For example, the bounding box position information could be (x1, y1; x2, y2). Here, x1 is the x-coordinate of the lower left vertex of the bounding box, y1 is the y-coordinate of the lower left vertex of the bounding box, x2 is the x-coordinate of the upper right vertex of the bounding box, and y2 is the y-coordinate of the upper right vertex of the bounding box.

[0043] The second step is to perform feature recognition on each adjusted concrete defect image in the above-mentioned adjusted concrete defect image set to generate recognized concrete defect image feature information, thus obtaining a set of recognized concrete defect image feature information.

[0044] As an example, the image processing device can use corner detection to extract features from each adjusted concrete defect image in the aforementioned adjusted concrete defect image set, generating extracted concrete defect image feature information, thus obtaining an extracted concrete defect image feature information set. Then, principal component analysis (PCA) is used to perform feature dimensionality reduction on each extracted concrete defect image feature information in the aforementioned extracted concrete defect image feature information set, generating dimensionality-reduced concrete defect image feature information, which serves as the identified concrete defect image feature information, thus obtaining an identified concrete defect image feature information set.

[0045] The third step is to label the adjusted concrete defect image set with categories based on the feature information set of the identified concrete defect images, thereby obtaining the category-labeled concrete defect image set.

[0046] As an example, the image processing device can classify the feature information set of the identified concrete defect images to obtain a set of concrete defect image feature categories. Then, each concrete defect image feature category in the set of concrete defect image feature categories is labeled to the corresponding adjusted concrete defect image in the set of adjusted concrete defect images to generate a set of category-labeled concrete defect images. Here, the corresponding adjusted concrete defect image can refer to an adjusted concrete defect image with the same category as the feature category of the concrete defect image. For example, if the feature category of the concrete defect image is honeycomb, then the corresponding adjusted concrete defect image also belongs to the honeycomb category, and the two categories are consistent. Here, the above categories can include, but are not limited to, at least one of the following: honeycomb category, pitted surface category, and crack category.

[0047] The fourth step is to annotate the coordinate position information of the concrete defect image bounding box based on the above-mentioned concrete defect image bounding box location information, and obtain the annotated concrete defect image set.

[0048] As an example, the image processing device can annotate the bounding box location information of each concrete defect image in the aforementioned set of concrete defect image bounding box location information to the corresponding category-annotated concrete defect image in the aforementioned set of category-annotated concrete defect images, thereby generating an annotated concrete defect image and obtaining an annotated concrete defect image set. Here, the corresponding category-annotated concrete defect image can refer to a category-annotated concrete defect image with the same location as the bounding box location information of the concrete defect image. For example, if the location of the bounding box location information of the concrete defect image is (x1, y1; x2, y2), then the location of the corresponding category-annotated concrete defect image is also (x1, y1; x2, y2). The positions of the two are consistent.

[0049] Step 203: The defect feature detection device is configured to: input the received annotated concrete defect image set into the pre-trained defect feature detection model to obtain the concrete defect image feature set, and send the concrete defect image feature set to the concrete layer division device.

[0050] In some embodiments, the defect feature detection device is configured to: input the received set of annotated concrete defect images into a pre-trained defect feature detection model to obtain a set of concrete defect image features, and send the set of concrete defect image features to a concrete hierarchical segmentation device. The concrete defect image features in the set of concrete defect image features characterize the concrete defect features of the concrete defect images. The defect feature detection device can be used to detect defect features in images. For example, the defect feature detection device can refer to a FLIR E8 thermal imaging camera. The defect feature detection model can be used to characterize the correspondence between the annotated concrete defect images in the set of annotated concrete defect images and the concrete defect image features in the set of concrete defect image features. The defect feature detection model can be a classification model that takes the set of annotated concrete defect images as input and the set of concrete defect image features as output. The defect feature detection model can sequentially compare the set of annotated concrete defect images with multiple preset sets of annotated concrete defect images in a preset concrete defect image relationship table. The aforementioned preset concrete defect image relationship table can be a correspondence table created based on the analysis of a large number of preset labeled concrete defect image sets, storing the correspondence between preset labeled concrete defect image sets and concrete defect image feature sets. Each set of preset labeled concrete defect images corresponds to a set of concrete defect image features. The aforementioned preset concrete defect image feature sets can be pre-defined concrete defect image feature sets.

[0051] Here, the aforementioned concrete layer classification device can be used to classify concrete and detect the degree of defects. The concrete layer classification device may refer to a computing terminal that includes the Pundit Live Array Pro array-type ultrasonic imaging system. The concrete layer classification device is wiredly connected to the defect feature detection device.

[0052] In some optional implementations of certain embodiments, the above-described concrete construction defect detection and repair system further includes: a defect feature detection model construction terminal; and Before inputting the received annotated concrete defect image set into the pre-trained defect feature detection model to obtain the concrete defect image feature set, the following steps are also included: The aforementioned defect feature detection model construction terminal is configured to: train the initial defect feature detection model to obtain a pre-trained defect feature detection model.

[0053] Here, the aforementioned defect feature detection model construction terminal can establish a communication connection with the defect feature detection device within a certain distance via wireless communication.

[0054] In some optional implementations of certain embodiments, the above-described training of the initial defect feature detection model to obtain a pre-trained defect feature detection model includes: The first step is to obtain a sample set of labeled concrete defect images.

[0055] Here, obtaining the labeled concrete defect image sample set refers to randomly obtaining the labeled concrete defect image sample set. Here, the labeled concrete defect image sample in the labeled concrete defect image sample set refers to the defect image of the concrete surface after solidification, which has labeled information. Here, the aforementioned labeling information may include, but is not limited to, at least one of the following: concrete defect image coordinate information and concrete defect image type information.

[0056] The second step is to select labeled concrete defect image samples from the above set of labeled concrete defect image samples as labeled concrete defect image test samples. The labeled concrete defect image test samples include: the labeled concrete defect image and the corresponding concrete defect image prediction feature label.

[0057] Here, the above selection can refer to random selection.

[0058] The third step is to input the above-annotated concrete defect image test samples into the initial defect feature detection model to obtain the predicted features of the concrete defect images.

[0059] Here, the initial defect feature detection model can be a defect feature detection model that has not been trained. The aforementioned initial defect feature detection model can refer to an initial defect feature detection model that takes labeled concrete defect image test samples as input and outputs predicted features of the concrete defect images. For example, the aforementioned initial defect feature detection model can refer to a recurrent neural network (RNN) model.

[0060] The fourth step involves determining the label difference between the predicted features of the concrete defect images and the predicted feature labels of the concrete defect images included in the labeled concrete defect image test samples, based on a preset image feature loss function.

[0061] Here, the image feature loss function mentioned above may include, but is not limited to: mean squared error loss function (MSE), cross-entropy loss function, etc.

[0062] Fifth step: In response to the above label difference value being greater than or equal to the preset label threshold, adjust the network parameters of the above initial defect feature detection model.

[0063] In some embodiments, the defect feature detection model building terminal can adjust the network parameters of the initial defect feature detection model in response to the label difference value being greater than or equal to a preset label threshold. Here, the setting of the preset label threshold is not limited. For example, the difference between the difference value and the preset label threshold can be calculated to obtain the loss difference. Based on this, methods such as backpropagation and stochastic gradient descent are used to propagate the difference value from the last layer of the model forward to adjust the parameters of each layer. Of course, as needed, the network dropout method can also be used to keep the network parameters of some layers unchanged without adjustment; no limitations are imposed on this.

[0064] In some optional implementations of certain embodiments, the defect feature detection model construction terminal described above is further configured to: In response to the fact that the label difference value is less than the preset label threshold, the initial defect feature detection model is determined as the defect feature detection model and used as the pre-trained defect feature detection model.

[0065] In some embodiments, the defect feature detection model building terminal may, in response to the label difference value being less than the preset label threshold, determine the initial defect feature detection model as a defect feature detection model, and use it as a pre-trained defect feature detection model. Here, there is no limitation on the setting of the preset label threshold.

[0066] Step 204: The concrete layer segmentation device is configured to: perform layer segmentation on the received concrete defect image feature set to obtain a segmented concrete defect image feature set, and send the segmented concrete defect image feature set to the early warning terminal.

[0067] In some embodiments, the concrete layer segmentation device is configured to: perform layer segmentation on the received concrete defect image feature set to obtain a segmented concrete defect image feature set, and send the segmented concrete defect image feature set to an early warning terminal. The degree of image defect corresponds to different segments of the segmented concrete defect image feature set.

[0068] Here, the aforementioned early warning terminal is used to generate image information from the feature set of the segmented concrete defect image. The early warning terminal communicates with the concrete layer segmentation device wirelessly.

[0069] In some optional implementations of certain embodiments, the concrete layer segmentation device described above includes: a confidence information detection processor; and The above-mentioned hierarchical division of the received concrete defect image feature set yields a segmented set of concrete defect image features, including: The first step is to perform confidence information detection on each concrete defect image feature in the received concrete defect image feature set to generate concrete defect image confidence information, thus obtaining a concrete defect image confidence information set.

[0070] Here, the aforementioned confidence information characterizes the degree of confidence in the features of concrete defect images. The aforementioned confidence information detection processor is used to process the confidence level of the model detection results.

[0071] As an example, the confidence information detection processor can determine the confidence information of each concrete defect image feature in the concrete defect image feature set using the YOLO (You Only Look Once) algorithm to generate concrete defect image confidence information and obtain a concrete defect image confidence information set.

[0072] The second step involves comparing the aforementioned confidence information set of concrete defect images with a preset confidence level to obtain a comparison result set. The confidence information detection processor can compare the aforementioned confidence information set of concrete defect images with a preset confidence level to obtain a comparison result set.

[0073] Here, the aforementioned preset confidence level can refer to a pre-set confidence level. For example, the aforementioned preset confidence level can be 0.5. Here, the comparison result can indicate that the confidence level of concrete defect images in the concrete defect image confidence information set is less than the preset confidence level; or it can indicate that the confidence level of concrete defect images in the concrete defect image confidence information set is greater than or equal to the preset confidence level.

[0074] The third step involves removing concrete defect image features corresponding to a confidence level lower than the preset confidence level from the concrete defect image feature set and the concrete defect image confidence information set, resulting in a removed concrete defect image feature set and a removed concrete defect image confidence information set. The confidence information detection processor can remove concrete defect image features corresponding to a confidence level lower than the preset confidence level from the concrete defect image feature set and the concrete defect image confidence information set, resulting in a removed concrete defect image feature set and a removed concrete defect image confidence information set.

[0075] The fourth step is to divide the feature set of the removed concrete defect images according to the confidence information set and the preset confidence level, so as to obtain the feature set of the removed concrete defect images.

[0076] Here, the aforementioned preset reliability level can refer to a pre-defined level of confidence. For example, the aforementioned preset reliability level can refer to a minor defect level, a moderate defect level, and a major defect level. The range for the minor defect level can be (0, 0.7). The range for the moderate defect level can be (0.7, 0.9). The range for the major defect level can be (0.9, 1).

[0077] In some optional implementations of some embodiments, the above-described concrete construction defect detection and repair system further includes an image segmentation processor and an image tagging processor; and: The first step, the image segmentation processor, is configured as follows: The first sub-step involves classifying the defect type for each of the segmented concrete defect image feature groups in the above-mentioned segmented concrete defect image feature group set, thereby obtaining a defect type classification result set.

[0078] Here, the image segmentation processor described above is used for classifying and segmenting images. The image labeling processor described above is used for labeling images. The image segmentation processor and the image labeling processor are connected to the concrete construction defect detection and repair system via a wired connection.

[0079] Here, the aforementioned defect types can refer to the types of defective surfaces formed after concrete has hardened. These defect types may include, but are not limited to, at least one of the following: honeycomb type, pitted surface type.

[0080] As an example, the image segmentation processor can use TensorFlow to classify the defect type of each segmented concrete defect image feature group in the above segmented concrete defect image feature group set, and obtain a defect type classification result set.

[0081] The second sub-step involves determining the coordinate positions of each concrete defect image corresponding to each segmented concrete defect image feature group in the segmented concrete defect image feature group set, based on the aforementioned defect type classification result set, to generate a set of coordinate positions of concrete defects to be repaired, thus obtaining a set of coordinate positions of concrete defects to be repaired.

[0082] As an example, the image segmentation processor determines the position coordinates of each concrete defect image corresponding to each segmented concrete defect image feature group in the above-mentioned segmented concrete defect image feature group set, so as to generate a set of coordinate positions of concrete defects to be repaired, thus obtaining a set of coordinate positions of concrete defects to be repaired.

[0083] The third sub-step involves performing repair operations on each concrete defect region corresponding to the aforementioned set of coordinate positions to be repaired for concrete defects, resulting in a set of repaired concrete defect regions for the corresponding set of coordinate positions to be repaired. This set of repaired concrete defect regions comprises a collection of different types of repaired concrete defect regions, and the sub-groups of repaired concrete defect regions within this set are combinations of repaired concrete defect regions with different degrees of defect.

[0084] As an example, the image segmentation processor sends a set of coordinate locations of concrete defects to be repaired to a concrete defect repair device. This instructs the device to control an associated filling tool to perform repair operations on each concrete defect area corresponding to the set of coordinate locations, resulting in a repaired set of concrete defect areas. Here, the filling tool can refer to a high-pressure grouting machine.

[0085] The fourth sub-step involves determining the quality level of each repaired concrete defect image corresponding to a repaired concrete defect area in the image set corresponding to the repaired concrete defect area, thereby obtaining a set of quality levels for the repaired concrete defect images.

[0086] Here, the aforementioned quality levels can refer to general, good, and excellent quality. These quality levels characterize the defect repair quality of the corresponding image of the repaired concrete defect area. The aforementioned image of the repaired concrete defect is an image of the concrete defect acquired using an image acquisition device after the concrete defect area has been repaired.

[0087] As an example, the image segmentation processor can determine the defect area of ​​each repaired concrete defect region in the image set corresponding to the repaired concrete defect region, thus obtaining a set of concrete defect areas. Then, the images corresponding to each repaired concrete defect region in the above set of concrete defect areas are sorted by quality level from largest to smallest, resulting in a sorted set of concrete defect image quality levels, which serves as the set of quality levels for repaired concrete defect images.

[0088] The fifth sub-step involves performing secondary repair on each concrete defect region corresponding to at least one of the repaired concrete defect image quality levels that meets the preset first quality level condition in the above-mentioned repaired concrete defect image quality level set, to obtain a secondary repaired concrete defect region set.

[0089] Here, the aforementioned secondary repair can be performed using a filling tool. The aforementioned preset first quality level condition can refer to conditions where the quality level is either average or good.

[0090] As an example, the image segmentation processor sends at least one quality level of the repaired concrete defect image that meets the preset first quality level condition from the aforementioned set of repaired concrete defect image quality levels to the concrete defect repair equipment. This instructs the concrete defect repair equipment to control an associated filling tool to perform repair operations on each concrete defect region corresponding to the at least one quality level of the repaired concrete defect image that meets the preset first quality level condition, resulting in a set of concrete defect regions after secondary repair. Here, the filling tool can refer to a high-pressure grouting machine.

[0091] The second step involves configuring the image tagging processor as follows: The first sub-step involves marking the defective areas of the concrete after the secondary repair to obtain the first marked defective area set.

[0092] Here, the aforementioned repair marker can refer to the "secondary repair" marker. The aforementioned defect area set after the first marking can refer to the defect area set marked with the "secondary repair" marker.

[0093] The second sub-step involves marking each concrete defect region corresponding to at least one quality level of the repaired concrete defect image that meets the preset second quality level condition in the above-mentioned repaired concrete defect image quality level group, thereby obtaining the second marked defect region group. Here, the preset second quality level condition can refer to a quality level of excellent. The defect area set after the second mark mentioned above is the defect area set marked with the "first repair" identifier.

[0094] The third sub-step involves determining the concrete defect images corresponding to the first marked defect region set and the second marked defect region set as the image set corresponding to the repaired concrete defect region.

[0095] The aforementioned first and second steps constitute an inventive point of this disclosure, solving the technical problem mentioned in the background art of "poor quality of defect repair." Factors leading to poor real-time interactivity and performance in 3D exhibition hall page rendering often include: When repairing concrete defects, the variety of concrete defect types can cause errors in the repair operation, resulting in poor quality repair. Solving these factors can improve the quality of defect repair. To achieve this, firstly, each segmented concrete defect image feature group in the aforementioned segmented concrete defect image feature group set is classified by defect type, resulting in a defect type classification result set. This allows for the determination of the concrete defect type during repair. Consequently, different repair methods are used for each different type, reducing the possibility of errors in the repair operation. Next, the quality level of the concrete defect images corresponding to each repaired concrete defect region in the aforementioned repaired concrete defect region group set is determined, resulting in a repaired concrete defect image quality level group set. This allows for the determination of the quality after defect repair, facilitating subsequent secondary repair based on the quality level. Then, the aforementioned secondary repaired concrete defect region group set is marked with repair identification tags, resulting in a first marked defect region group set. The concrete defect areas corresponding to at least one quality level of the repaired concrete defect images that meet the preset second quality level conditions are marked with repair labels, resulting in a second-marked defect area set. This allows for the re-repair of lower-quality concrete defect images to ensure repair quality, thereby improving the overall quality of defect repair.

[0096] Step 205: The early warning terminal is configured to generate early warning information corresponding to the above-described segmented concrete defect image feature set, and to send the early warning information to the concrete defect repair equipment.

[0097] In some embodiments, the upper warning terminal is configured to: generate warning information corresponding to the aforementioned segmented set of concrete defect image features, and send the warning information to the concrete defect repair device. The warning information includes the coordinates of each concrete defect location.

[0098] Here, the concrete defect location coordinates in the above-mentioned concrete defect location coordinates can refer to the location coordinates of the concrete defect. For example, the concrete defect location coordinates in the above-mentioned concrete defect location coordinates can refer to (10, 15).

[0099] As an example, the early warning terminal responds to determining that it has received a set of segmented concrete defect image features, issues an early warning based on the segmented concrete defect image feature set, and determines the information generated after the early warning as early warning information. The information generated after the early warning may refer to the location coordinates of the location where the warning is issued.

[0100] Step 206, the concrete defect repair equipment is configured to: in response to receiving the warning information sent by the aforementioned warning terminal, repair the concrete defects according to the coordinates of each concrete defect location contained in the aforementioned warning information.

[0101] In some embodiments, the concrete defect repair device in the concrete defect repair equipment is configured to: in response to receiving the warning information sent by the warning terminal, repair the concrete defects by performing concrete defect repair on the coordinates of each concrete defect location contained in the warning information.

[0102] Here, the concrete defect repair equipment mentioned above is a device used to repair defects on the surface of concrete.

[0103] As an example, concrete defect repair equipment can use a filling tool to repair concrete defects based on the coordinates of the locations of each concrete defect included in the aforementioned warning information. Here, the filling tool can refer to a high-pressure grouting machine.

[0104] In some optional implementations of certain embodiments, the aforementioned concrete defect repair equipment includes: a handheld terminal set for construction personnel and a defect repair control terminal; and In response to receiving the warning information from the aforementioned warning terminal, the above-mentioned concrete defect repair is performed on the coordinates of each concrete defect location corresponding to the aforementioned segmented concrete defect image feature set, including: The first step involves configuring each worker's handheld terminal within the aforementioned concrete defect repair equipment as follows: The first sub-step involves, in response to receiving the warning information from the aforementioned warning terminal, determining the warning level of each concrete defect coordinate corresponding to the aforementioned divided set of concrete defect image features, and thus obtaining a set of warning levels.

[0105] Here, the aforementioned handheld terminal for construction workers can be a terminal used by construction workers to receive early warning information sent by the early warning terminal. The handheld terminal can move with the construction workers, meaning the position of the construction worker and the position of the handheld terminal are the same. The handheld terminal can be a mobile terminal (MT). The aforementioned defect repair control terminal can be a terminal used to control defect repair. The handheld terminal and the concrete defect repair equipment communicate wirelessly, and the defect repair control terminal and the concrete defect repair equipment also communicate wirelessly.

[0106] Here, the aforementioned warning level groups may include: a cellular warning level group and a surface roughness warning level group. The cellular warning level group includes low-level cellular warning, medium-level cellular warning, and high-level cellular warning. The surface roughness warning level group includes low-level surface roughness warning, medium-level surface roughness warning, and high-level surface roughness warning. The cellular warning level group may refer to a warning level group for concrete defects that are cellular defects. The surface roughness warning level group may refer to a warning level group for concrete defects that are surface roughness defects.

[0107] The second sub-step involves determining the location information of the handheld terminals used by the construction workers to obtain the location information of concrete defects.

[0108] Here, the aforementioned location information may refer to information determined via GPS (Global Positioning System).

[0109] The third sub-step involves determining the distance between the concrete defect location information and the coordinate positions corresponding to each warning level in the warning level sequence, thereby obtaining a distance information sequence. This distance information sequence is a sequence of distances between the concrete defect location information and the coordinate positions of the construction area corresponding to each warning level in the warning level sequence.

[0110] As an example, construction workers can use a handheld terminal to determine the distance between the above concrete defect location information and the coordinate positions corresponding to each warning level in the warning level sequence using Euclidean distance, thus obtaining a distance information sequence.

[0111] The fourth sub-step involves determining the repair task information for the coordinate position of the area to be repaired corresponding to the minimum distance information in the above distance information sequence, obtaining concrete defect repair task information, and sending the concrete defect repair task information to the defect repair control terminal.

[0112] Here, the aforementioned repair task information may refer to the task information for repairing concrete defects corresponding to the coordinates of the area to be repaired.

[0113] The second step involves configuring the defect repair control terminal as follows: The first sub-step involves deduplicating the received concrete defect repair tasks to obtain a deduplicated concrete defect repair task group, and controlling the associated repair equipment to perform construction repair operations based on the deduplicated concrete defect repair task group.

[0114] Here, each deduplicated concrete defect repair task in the aforementioned deduplicated concrete defect repair task group represents a unique concrete defect repair task corresponding to each worker's handheld terminal. For example, the concrete defect repair task corresponding to worker handheld terminal A is to repair the concrete pitting defect at coordinates (10, 20). The concrete defect repair task corresponding to worker handheld terminal B is to repair the concrete honeycomb defect at coordinates (15, 10). The concrete defect repair task corresponding to worker handheld terminal C is to repair the concrete honeycomb defect at coordinates (30, 5). It is important to note that the defect repair tasks of worker handheld terminals A, B, and C are different.

[0115] As an example, the defect repair control terminal can deduplicate duplicate concrete defect repair tasks among the received concrete defect repair tasks according to the distance information of the construction worker's handheld terminal, resulting in a deduplicated concrete defect repair task group. The terminal then sends the received deduplicated concrete defect repair task group to the associated repair equipment, allowing each equipment to perform repair operations using filling tools. Here, the filling tool can refer to a high-pressure grouting machine. The associated repair equipment can refer to the equipment receiving the deduplicated concrete defect repair tasks from the deduplicated concrete defect repair task group. The repair operation can refer to the operation of injecting filling material into the concrete defect area using a high-pressure grouting machine. The filling material can refer to mortar. For example, if the coordinates of the construction worker's handheld terminal A are (9, 19), the concrete defect repair task is to repair the concrete surface defect at coordinates (10, 20). Similarly, if the coordinates of the construction worker's handheld terminal D are (6, 16), the concrete defect repair task is also to repair the concrete surface defect at coordinates (10, 20). The recurring concrete defect repair task is the task to repair the concrete surface defects at coordinates (10, 20). It can be seen that the coordinates of the construction worker's handheld terminal A are closer to the concrete surface defects repair task at coordinates (10, 20), so the concrete defect repair task corresponding to coordinates (6, 16) of the construction worker's handheld terminal D is removed.

[0116] The second sub-step involves assigning defect repair tasks to the handheld terminals of construction personnel corresponding to each of the aforementioned removed concrete defect repair task groups, thereby generating the assigned concrete defect repair tasks and obtaining the assigned concrete defect repair task groups. Based on the assigned concrete defect repair task groups, the associated repair equipment is controlled to perform construction repair operations.

[0117] As an example, the defect repair control terminal can first sort the warning level groups corresponding to the removed concrete defect repair tasks in the removed concrete defect repair task group from largest to smallest to obtain a set of warning information sequences to be repaired. Then, the warning information sequences to be repaired in the set of sequences to be repaired are sorted from nearest to farthest distance from each worker's handheld terminal corresponding to the removed concrete defect repair task group to generate a set of sequences to be assigned. Next, each sequence to be assigned in the set of sequences to be assigned is assigned a task based on minimum distance to generate assigned concrete defect repair tasks, resulting in a set of assigned concrete defect repair tasks. The removed concrete defect repair task group can be: {Concrete pitting defect repair task with worker handheld terminal D at coordinates (10, 20); Concrete honeycomb defect repair task with worker handheld terminal E at coordinates (15, 10)}. The aforementioned sequence of early warning information to be repaired can be: {Construction worker handheld terminal D: Repair task for concrete surface defects at coordinates (10, 20) with early warning level 3; Repair task for concrete honeycomb defects at coordinates (15, 10) with early warning level 2; Repair task for concrete honeycomb defects at coordinates (9, 19) with early warning level 1; Construction worker handheld terminal E: Repair task for concrete honeycomb defects at coordinates (15, 10) with early warning level 3; Repair task for concrete surface defects at coordinates (10, 20) with early warning level 2; Repair task for concrete honeycomb defects at coordinates (9, 19) with early warning level 1}. The coordinates of the aforementioned construction worker handheld terminal D can be (6, 16). The coordinates of the aforementioned construction worker handheld terminal E can be (16, 11). The aforementioned sequence group to be assigned can refer to {Construction worker handheld terminal D: Concrete honeycomb defect repair task at coordinates (9, 19), concrete pitting defect repair task at coordinates (10, 20), and concrete honeycomb defect repair task at coordinates (15, 10); Construction worker handheld terminal E: Concrete honeycomb defect repair task at coordinates (15, 10), concrete pitting defect repair task at coordinates (10, 20), and concrete honeycomb defect repair task at coordinates (9, 19)}. The assigned concrete defect repair task group can be {Construction worker handheld terminal D: Concrete pitting defect repair task at coordinates (9, 19); Construction worker handheld terminal E: Concrete honeycomb defect repair task at coordinates (15, 10)}. The assigned concrete defect repair task group is then sent to the associated repair equipment, allowing each equipment to perform repair operations using filling tools. Here, the filling tool can refer to a high-pressure grouting machine. The associated repair equipment can refer to the repair equipment for the assigned concrete defect repair task in the assigned concrete defect repair task group.The aforementioned construction and repair operation refers to the process of injecting filling material into the defective area of ​​concrete using a high-pressure grouting machine. The filling material can be mortar.

[0118] The content in steps one and two above constitutes an inventive point of this disclosure, solving the technical problem mentioned in the background art: "long time for allocating repair tasks and long repair task cycles." Factors leading to long repair task allocation times and cycles are often as follows: The inability to respond promptly to repair tasks after detecting concrete defects results in a longer allocation time. Furthermore, the location of concrete defects is often uncertain, making it impossible to promptly allocate repair tasks to the nearest concrete defect receiving device, thus lengthening the repair task cycle. Solving these factors can reduce the time for allocating repair tasks and shorten the repair task cycle. To achieve this effect, firstly, each worker's handheld terminal in the aforementioned concrete defect repair equipment is configured to: in response to receiving warning information from the aforementioned warning terminal, determine the warning level of each concrete defect coordinate corresponding to the aforementioned divided concrete defect image feature set, obtaining a warning level set. Therefore, prioritizing the processing of locations with lower quality according to the warning level allows for timely response to repair tasks and facilitates the allocation of subsequent repair tasks. Next, the distance between the concrete defect location information and the coordinates corresponding to each warning level in the warning level sequence is determined, resulting in a distance information sequence. This sequence represents the distances between the concrete defect location information and the coordinates of the construction area corresponding to each warning level in the warning level sequence. This allows the location of the concrete defect to be determined. Then, the repair task information is determined for the coordinates of the area to be repaired corresponding to the minimum distance in the distance information sequence, resulting in concrete defect repair task information. This concrete defect repair task information is then sent to the defect repair control terminal. Defect repair tasks are assigned to the handheld terminals of the construction personnel corresponding to each rejected concrete defect repair task in the rejected concrete defect repair task group, generating a rejected concrete defect repair task group. Based on this group, the associated repair equipment is controlled to perform construction repair operations. This allows for timely assignment of repair tasks to the nearest concrete defect receiving device, thereby shortening the repair task cycle. Therefore, the time for assigning repair tasks is reduced, and the repair task cycle is shortened.

[0119] The following is for reference. Figure 3 It illustrates a schematic diagram of the structure of an electronic device (such as a computing device) suitable for implementing some embodiments of the present disclosure. Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality or scope of the embodiments of this disclosure. Figure 3 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The memory may include non-volatile storage media and internal memory. The non-volatile storage media may store an operating system and computer programs. The computer program includes program instructions that, when executed, cause the processor to execute any of the aforementioned systems. The processor provides computing and control capabilities to support the operation of the entire computer device. The internal memory provides an environment for the execution of the computer program in the non-volatile storage media; when executed by the processor, this program causes the processor to execute any of the aforementioned systems. The network interface is used for network communication, such as sending assigned tasks. Those skilled in the art will understand that... Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present disclosure and does not constitute a limitation on the computer device to which the present disclosure is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0120] It should be understood that the processor can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among these, a general-purpose processor can be a microprocessor or any conventional processor.

[0121] In one embodiment, the processor is configured to run a computer program stored in a memory to perform the following steps: the ultrasonic concrete device is configured to: in response to determining that a defect is detected on the concrete surface, acquire a set of concrete defect images and send the set of concrete defect images to an image processing device, wherein the set of concrete defect images is a collection of defect images taken after the concrete used in construction has solidified; the image processing device is configured to: generate an annotated set of concrete defect images based on the received set of concrete defect images and send the annotated set of concrete defect images to a defect feature detection device; the defect feature detection device is configured to: input the received annotated set of concrete defect images into a pre-trained defect feature detection model to obtain a set of concrete defect image features and send the set of concrete defect image features to a concrete layer segmentation device, wherein the above... The concrete defect image features in the concrete defect image feature set characterize the concrete defect features of the image; the concrete layer segmentation device is configured to: perform layer segmentation on the received concrete defect image feature set to obtain a segmented concrete defect image feature set, and send the segmented concrete defect image feature set to an early warning terminal, wherein the image defect degree corresponding to each segmented concrete defect image feature set in the segmented concrete defect image feature set is different; the early warning terminal is configured to: generate early warning information corresponding to the segmented concrete defect image feature set, and send the early warning information to a concrete defect repair device, wherein the early warning information includes the coordinates of each concrete defect location; the concrete defect repair device is configured to: in response to receiving the early warning information sent by the early warning terminal, perform concrete defect repair on the concrete defect location coordinates included in the early warning information.

[0122] This disclosure also provides a computer-readable storage medium storing a computer program, the computer program including program instructions, and the system implemented when the program instructions are executed can be referred to the various embodiments of the system described above.

[0123] The aforementioned computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiments, such as the hard disk or memory of the computer device. Alternatively, the aforementioned computer-readable storage medium may be an external storage device of the computer device, such as a plug-in hard disk, SmartMedia Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the computer device.

[0124] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, system, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, system, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, system, article, or system that includes that element.

[0125] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. An image recognition based concrete construction defect detection and repair system, characterized in that, The concrete construction defect detection and repair system comprises a concrete ultrasonic device, an image processing device, a defect feature detection device, a warning terminal, a concrete defect repair device and a concrete level division device, wherein: The concrete ultrasonic device is configured to: in response to determining that a defect exists on the surface of the concrete, collect a set of concrete defect images, and send the set of concrete defect images to the image processing device, wherein the set of concrete defect images is a collection of defect images taken after the concrete used for construction is solidified; The image processing device is configured to: generate a set of labeled concrete defect images based on the received set of concrete defect images, and send the set of labeled concrete defect images to the defect feature detection device; The defect feature detection device is configured to: input the received set of labeled concrete defect images into a pre-trained defect feature detection model to obtain a set of concrete defect image features, and send the set of concrete defect image features to the concrete level division device, wherein the concrete defect image features in the set of concrete defect image features represent the concrete defect features of the images; The concrete level division device is configured to: perform hierarchical division on the received set of concrete defect image features to obtain a set of divided concrete defect image feature groups, and send the set of divided concrete defect image feature groups to the warning terminal, wherein each divided concrete defect image feature group in the set of divided concrete defect image feature groups corresponds to images with different defect levels; The warning terminal is configured to: generate warning information corresponding to the set of divided concrete defect image feature groups, and send the warning information to the concrete defect repair device, wherein the warning information includes coordinates of each concrete defect location; The concrete defect repair device is configured to: in response to receiving the warning information sent by the warning terminal, perform concrete defect repair on the coordinates of each concrete defect location included in the warning information.

2. The concrete construction defect detection and repair system of claim 1, wherein, The image processing device is further configured to: perform image cropping on the set of concrete defect images to obtain a set of cropped concrete defect images; perform denoising processing on the set of cropped concrete defect images to obtain a set of denoising processed concrete defect images; perform sharpening processing on the set of denoising processed concrete defect images to obtain a set of sharpening processed concrete defect images; perform contrast enhancement on the set of sharpening processed concrete defect images to obtain a set of enhanced concrete defect images; perform illumination adjustment on the set of enhanced concrete defect images to obtain a set of adjusted concrete defect images; perform data labeling on the set of adjusted concrete defect images to obtain a set of labeled concrete defect images.

3. The concrete construction defect detection and repair system of claim 2, wherein, The image processing device is further configured to: perform boundary box position information determination on each adjusted concrete defect image in the set of adjusted concrete defect images to generate concrete defect image boundary box position information, thereby obtaining a set of concrete defect image boundary box position information; feature recognition is performed on each of the adjusted concrete defect images in the adjusted concrete defect image set to generate identified concrete defect image feature information, to obtain an identified concrete defect image feature information set; According to the identified concrete defect image feature information set, the adjusted concrete defect image set is labeled by category to obtain a category-labeled concrete defect image set; According to the concrete defect image bounding box position information set, coordinate position information is labeled on the category-labeled concrete defect image set to obtain a labeled concrete defect image set.

4. The concrete construction defect detection and repair system of claim 1, wherein, The concrete hierarchical division device comprises a confidence information detection processor; and The confidence information detection processor is configured to: confidence information detection is performed on each of the received concrete defect image features in the concrete defect image feature set to generate concrete defect image confidence information, to obtain a concrete defect image confidence information set; The concrete defect image confidence information set is compared with a preset confidence to obtain a comparison result set; The concrete defect image features corresponding to the comparison result set smaller than the preset confidence are removed from the concrete defect image feature set and the concrete defect image confidence information set, to obtain a removed concrete defect image feature set and a removed concrete defect image confidence information set; According to the removed concrete defect image confidence information set and a preset confidence level, the removed concrete defect image feature set is divided to obtain a divided concrete defect image feature group set.

5. The concrete construction defect detection and repair system of claim 1, wherein, The concrete construction defect detection and repair system further comprises a defect feature detection model construction terminal; and The defect feature detection model construction terminal is configured to: perform model training on an initial defect feature detection model to obtain a pre-trained defect feature detection model.

6. The concrete construction defect detection and repair system of claim 5, wherein, The defect feature detection model construction terminal is further configured to: obtain a labeled concrete defect image sample set; select a labeled concrete defect image sample from the labeled concrete defect image sample set as a labeled concrete defect image test sample, wherein the labeled concrete defect image test sample comprises a sample labeled concrete defect image and a corresponding concrete defect image predicted feature label; input the labeled concrete defect image test sample into the initial defect feature detection model to obtain a concrete defect image predicted feature; based on a preset image feature loss function, determine a label difference value between the concrete defect image predicted feature and a concrete defect image predicted feature label included in the labeled concrete defect image test sample; in response to the label difference value being greater than or equal to a preset label threshold, adjust network parameters of the initial defect feature detection model.

7. The concrete construction defect detection and repair system of claim 6, wherein, The defect feature detection model construction terminal is further configured to: in response to the label difference value being less than the preset label threshold, determine the initial defect feature detection model as a defect feature detection model as a pre-trained defect feature detection model.

8. An electronic device, comprising: comprise: one or more processors; a storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the one or more processors implement the concrete construction defect detection and repair system as claimed in any one of claims 1 to 7.

9. A computer readable medium characterized by A computer program product, having stored thereon a computer program, wherein the computer program is executed by a processor to implement the concrete construction defect detection and repair system as claimed in any one of claims 1 to 7.

Images