Underground pipeline construction quality inspection method and device, electronic equipment and storage medium

Abstract

The invention discloses an underground pipeline construction quality inspection method and device, electronic equipment and a storage medium. The method comprises the following steps: determining a plurality of target construction tube wells corresponding to a target construction pipeline to be subjected to quality inspection; for each target construction tube well, determining a plurality of construction related images corresponding to the target construction tube well and actual position data corresponding to the target construction tube well; performing feature extraction on the input construction-related images through a feature extraction model, determining construction-related features corresponding to each construction-related image, and performing feature fusion on the plurality of construction-related features to obtain target fusion features corresponding to the target construction tube well; determining a position evaluation score corresponding to the target construction tube well according to the actual position data; and determining a tube well quality inspection result of the target construction tube well according to the target fusion feature and the position evaluation score, and determining a pipeline quality inspection result of the target construction pipeline according to the tube well quality inspection result. Based on the technical scheme, the accuracy of underground pipeline construction quality inspection can be improved.

Description

Technical Field

[0001] The present invention relates to the field of computer application technologies, and in particular, to a method, device, electronic device, and storage medium for quality inspection of underground pipeline construction. Background Art

[0002] After underground pipeline construction, the construction quality is usually inspected to ensure the construction quality of underground pipelines and the normal operation of pipelines after construction.

[0003] Currently, usually a field image of completed underground pipeline construction is obtained first, and then the quality inspection result is determined by multiple manual re-inspections. However, the data that needs to be quality inspected in the image is relatively redundant, and there is usually a situation where the pipeline operation is abnormal after passing the quality inspection. That is to say, the accuracy of the current quality inspection for underground pipeline construction is relatively low. Summary of the Invention

[0004] The present invention provides a method, device, electronic device, and storage medium for quality inspection of underground pipeline construction to solve the technical problem of relatively low accuracy of the current quality inspection for underground pipeline construction.

[0005] According to one aspect of the present invention, a method for quality inspection of underground pipeline construction is provided. The method includes:

[0006] Determine a plurality of target construction manholes corresponding to the target construction pipeline to be quality inspected;

[0007] For each of the target construction manholes, determine multiple construction-related images corresponding to the target construction manhole and the actual position data corresponding to the target construction manhole;

[0008] Extract features from the input construction-related images through a feature extraction model, determine the construction-related features corresponding to each construction-related image, and perform feature fusion on the multiple construction-related features to obtain the target fusion feature corresponding to the target construction manhole;

[0009] Determine the position evaluation score corresponding to the target construction manhole according to the actual position data;

[0010] Determine the manhole quality inspection result of the target construction manhole according to the target fusion feature and the position evaluation score, and determine the pipeline quality inspection result of the target construction pipeline according to the manhole quality inspection result.

[0011] According to another aspect of the present invention, a device for quality inspection of underground pipeline construction is provided. The device includes:

[0012] A pipeline determination module, configured to determine a plurality of target construction manholes corresponding to the target construction pipeline to be quality inspected;

[0013] A data acquisition module, configured to determine, for each of the target construction wells, a plurality of construction-related images corresponding to the target construction well and actual position data corresponding to the target construction well;

[0014] A feature processing module, configured to perform feature extraction on the input construction-related images through a feature extraction model, determine construction-related features corresponding to each of the construction-related images, and perform feature fusion on the multiple construction-related features to obtain target fusion features corresponding to the target construction well;

[0015] A position evaluation module, configured to determine a position evaluation score corresponding to the target construction well according to the actual position data;

[0016] A result determination module, configured to determine a well quality inspection result of the target construction well according to the target fusion features and the position evaluation score, and determine a pipeline quality inspection result of the target construction pipeline according to the well quality inspection result.

[0017] According to another aspect of the present invention, there is provided an electronic device, the electronic device including:

[0018] At least one processor; and

[0019] A memory communicatively connected to the at least one processor; wherein,

[0020] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the underground pipeline construction quality inspection method according to any embodiment of the present invention.

[0021] According to another aspect of the present invention, there is provided a computer-readable storage medium, the computer-readable storage medium storing computer instructions, and the computer instructions are used to implement the underground pipeline construction quality inspection method according to any embodiment of the present invention when executed by a processor.

[0022] In the technical solution of the embodiment of the present invention, multiple target construction manholes corresponding to the target construction pipelines to be quality inspected are determined; for each of the target construction manholes, multiple construction-related images corresponding to the target construction manhole and the actual position data corresponding to the target construction manhole are determined; the construction-related features corresponding to each of the construction-related images are determined by performing feature extraction on the input construction-related images through a feature extraction model, and the construction-related features are feature-fused to obtain the target fusion features corresponding to the target construction manhole; a position evaluation score corresponding to the target construction manhole is determined according to the actual position data; the manhole quality inspection result of the target construction manhole is determined according to the target fusion features and the position evaluation score, and the pipeline quality inspection result of the target construction pipeline is determined according to the manhole quality inspection result. The present invention performs construction quality inspection on underground pipelines from multiple dimensions (image feature extraction and fusion and position evaluation), achieving the effect of improving the accurate determination of underground pipeline construction quality inspection.

[0023] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0025] Figure 1 is a flowchart of a method for quality inspecting underground pipeline construction according to Embodiment 1 of the present invention;

[0026] Figure 2 is an example diagram of a manhole and a protruding pipeline in the manhole according to an embodiment of the present invention;

[0027] Figure 3 is a flowchart of performing pipeline quality inspection through a deep learning model according to an embodiment of the present invention;

[0028] Figure 4 is a flowchart of a method for quality inspecting underground pipeline construction according to Embodiment 2 of the present invention;

[0029] Figure 5 is a flowchart of determining a positive and negative sample combination according to an embodiment of the present invention;

[0030] Figure 6It is a flowchart for training a contrastive learning model provided by an embodiment of the present invention;

[0031] Figure 7 It is a flowchart for extracting construction-related features provided by an embodiment of the present invention;

[0032] Figure 8 It is a schematic structural diagram of an underground pipeline construction quality inspection device provided by Embodiment 3 of the present invention;

[0033] Figure 9 It is a schematic structural diagram of an electronic device for implementing the underground pipeline construction quality inspection method of the embodiment of the present invention. Detailed implementation manners

[0034] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0035] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that such data may be interchanged under appropriate circumstances, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0036] Embodiment 1

[0037] Figure 1 This is a flowchart of an underground pipeline construction quality inspection method provided by Embodiment 1 of the present invention. This embodiment is applicable to the situation of quality inspection of construction projects through image processing technology. This method can be executed by an underground pipeline construction quality inspection device, which can be implemented in the form of hardware and / or software, and the underground pipeline construction quality inspection device can be configured in a computer. As Figure 1 shown, this method includes:

[0038] S110. Determine a plurality of target construction manholes corresponding to the target construction pipeline to be inspected.

[0039] Among them, the target construction pipeline can be understood as the underground construction pipeline to be inspected for quality. Usually, each underground construction pipeline corresponds to multiple construction manholes. That is, to inspect the quality of the underground construction pipeline, it is necessary to first inspect all the construction manholes corresponding to the underground construction pipeline. When all the manholes pass the quality inspection, it is considered that the underground construction pipeline passes the quality inspection.

[0040] The target construction manhole can be understood as the construction manhole corresponding to the target construction pipeline.

[0041] S120. For each of the target construction manholes, determine multiple construction-related images corresponding to the target construction manhole and the actual position data corresponding to the target construction manhole.

[0042] Among them, the construction-related images can be understood as images related to manhole construction. In the embodiments of the present invention, the construction-related images can be preset according to scene requirements and are not specifically limited here. Optionally, the construction-related images may include manhole construction images, manhole pipeline images, or protruding pipeline images. Among them, the manhole construction image can be understood as the on-site image of manhole construction. The manhole construction image may include information such as well wall plastering, pipe hole bell mouths, manholes, and pipelines. The manhole pipeline image can be understood as an image including the manhole shape and the pipeline shape. The manhole pipeline image may include information such as the manhole shape and the pipeline shape. The protruding pipeline image can be understood as an image including the shape of the protruding pipeline in the manhole. The protruding pipeline image may include information such as the shape of the protruding pipeline. Refer to Figure 2 , Figure 2 is an example diagram of a manhole and a protruding pipeline in the manhole provided according to the embodiments of the present invention. Among them, 11 represents the manhole, and 22 represents the protruding pipeline in the manhole. Obviously, Figure 2 can also be used as one of the manhole pipeline images. The image corresponding to the 22 area can be used as a protruding pipeline image.

[0043] The actual position data can be understood as the data characterizing the actual construction position of the construction manhole. Optionally, the actual position data may be the actual longitude and latitude data of the construction manhole.

[0044] S130. Through the feature extraction model, perform feature extraction on the input construction-related images, determine the construction-related features corresponding to each construction-related image, and perform feature fusion on the multiple construction-related features to obtain the target fusion feature corresponding to the target construction manhole.

[0045] Among them, the feature extraction model can be understood as a model with feature extraction capabilities. Optionally, the feature extraction model can be obtained by training a contrast learning model based on construction sample images, pipe well sample images, and pipeline sample images. Among them, the construction sample images can be sample images of the same category as the pipe well construction images, the pipe well sample images can be sample images of the same category as the pipe well pipeline images, and the pipeline sample images can be sample images of the same category as the protruding pipeline images. In summary, for sample images, no further elaboration will be provided here.

[0046] The construction-related features can be understood as features related to pipe orifice construction. The construction-related features can be digital features extracted from construction-related images based on the feature extraction model.

[0047] The target fusion feature can be understood as a fusion feature of multiple construction-related features. In the embodiments of the present invention, the specific method of feature fusion for multiple construction-related features can be preset according to scene requirements and will not be specifically limited here.

[0048] S140. Determine the position evaluation score corresponding to the target construction pipe well according to the actual position data.

[0049] Among them, the position evaluation score can be understood as an evaluation score related to the construction position of the pipe well. Optionally, the position evaluation score includes a first evaluation score for the actual construction position of each target construction pipe well and a second evaluation score for the actual distance between every two adjacent target construction pipe wells.

[0050] Under normal circumstances, for a target construction pipeline, the actual construction position of each target construction pipe well needs to be consistent with the preset target construction position. However, there may be a possibility of deviation between the actual construction position and the target construction position. Therefore, the present invention combines the evaluation of the actual construction position of the construction pipe well to conduct a comprehensive quality inspection of the pipeline construction.

[0051] Optionally, the determining the position evaluation score corresponding to the target construction pipe well according to the actual position data includes:

[0052] For each target construction pipe well, determine the preset target position data corresponding to the target construction pipe well;

[0053] Determine a preset position evaluation threshold, and determine the position evaluation score corresponding to the target construction pipe well according to the actual position data, the target position data, and the position evaluation threshold.

[0054] Among them, the position evaluation threshold can be understood as a preset threshold related to position evaluation. In the embodiments of the present invention, the position evaluation threshold can be preset according to the scene requirements, and no specific limitation is made here. The position evaluation threshold can include an evaluation threshold related to the actual construction position and an evaluation threshold related to the actual distance.

[0055] The target position data can be understood as data representing the standard construction position of the construction pipe well. Optionally, the target position data can be the target longitude and latitude data of the construction pipe well.

[0056] Specifically, the determination of the first evaluation score corresponding to the target construction pipe well according to the position evaluation threshold and the actual position data can be implemented based on the following formula:

[0057]

[0058] Among them, A pos represents the actual longitude and latitude data; D pos represents the target longitude and latitude data;

[0059] ρ(A pos -D pos ) represents the position difference between A pos and D pos ; L3 and L4 respectively represent two position evaluation thresholds, and L3 is less than L4; S pos represents the first evaluation score.

[0060] Specifically, the determination of the second evaluation score corresponding to the target construction pipe well according to the actual position data, the target position data, and the position evaluation threshold can be implemented based on the following formula:

[0061]

[0062] Among them, A dis represents the actual distance between two adjacent target construction pipe wells, and A dis can be determined based on the actual position data; D dis represents the preset target (standard) distance between two adjacent target construction pipe wells, and the target distance can be determined based on the target position data; S dis represents the second evaluation score.

[0063] Based on the above embodiment solutions, the accuracy of the pipe well construction position evaluation can be improved.

[0064] S150. Determine the inspection result of the target construction pipe well based on the target fusion feature and the position evaluation score, and determine the inspection result of the target construction pipeline based on the inspection result of the pipe well.

[0065] Among them, the inspection result of the pipe well can be understood as the inspection result indicating whether the target construction pipe well is constructed qualified. Optionally, the inspection result of the pipe well can include qualified pipe well inspection or unqualified pipe well inspection.

[0066] The inspection result of the pipeline can be understood as the inspection result indicating whether the target construction pipeline is constructed qualified. Optionally, the inspection result of the pipeline can include qualified pipeline inspection or unqualified pipeline inspection.

[0067] Optionally, the determining the inspection result of the target construction pipeline based on the inspection result of the pipe well includes:

[0068] Determine the inspection result of the target construction pipeline based on the inspection result of the pipe well and the position evaluation score.

[0069] Optionally, the position evaluation score includes a first evaluation score for the actual construction position of each target construction pipe well and a second evaluation score for the actual distance between every two adjacent target construction pipe wells;

[0070] The determining the inspection result of the target construction pipe well based on the target fusion feature and the position evaluation score includes:

[0071] Determine the inspection result of the target construction pipe well based on the target fusion feature and the first evaluation score;

[0072] The determining the inspection result of the target construction pipeline based on the inspection result of the pipe well and the position evaluation score includes:

[0073] Determine the inspection result of the target construction pipeline based on the inspection result of the pipe well and the second evaluation score.

[0074] Among them, the actual construction position can be understood as the actual construction location. In the embodiments of the present invention, the actual construction positions corresponding to different target construction pipe wells can be different.

[0075] The actual distance can be understood as the actual distance between two adjacent target construction pipe wells. Optionally, the actual distance can be determined according to the actual construction position.

[0076] In the embodiment of the present invention, the feature fusion of multiple construction-related features to obtain the target fusion feature corresponding to the target construction manhole, and, determining the manhole quality inspection result of the target construction manhole according to the target fusion feature and the position evaluation score, and determining the pipeline quality inspection result of the target construction pipeline according to the manhole quality inspection result may be implemented based on a trained deep learning model.

[0077] Specifically, input the first construction feature, the second construction feature, the third construction feature, the first evaluation score of each target construction manhole, and the second evaluation score of every two adjacent target construction manholes into the trained deep learning model to obtain the pipeline quality inspection result output by the deep learning model. Figure 3 It is a flowchart of pipeline quality inspection through a deep learning model provided by an embodiment of the present invention. The following is a further elaboration in combination with Figure 3 the above process, where Figure 3 feature a, feature b, and the feature in it represent the first construction feature, the second construction feature, and the third construction feature in sequence; S dis 、 and represent the second evaluation score for every two adjacent target construction manholes (the first pipe orifice and the second pipe orifice), the first evaluation score of the first pipe orifice, and the second evaluation score of the second pipe orifice in sequence.

[0078] Specifically, first, after the features of different images obtained by extraction are subjected to feature mapping through a linear projection layer, they are input into a multi-layer deep learning model (for example, a Transformer encoder) for multi-modal fusion; so that the model uses the multi-head self-attention mechanism to perform self-attention calculation on the input features, capture the internal connections and correlations between different modalities, and generate a feature representation that fuses multi-modal information; after completing the multi-modal feature fusion, the fused feature and S dis 、 and are concatenated together and input into a classifier to obtain the quality inspection result. The classifier includes a linear layer, an activation function, and a Softmax function.

[0079] The loss function can adopt Focal Loss, and the mathematical expression is as follows:

[0080] FocalLoss=-α t (1 - s) γ log(s)

[0081] where α trepresents the scoring balance factor, which is used to balance the loss weights of different input evaluation scores; s represents the quality inspection result predicted by the model for the current input; γ represents the feature adjustment factor, which is used to adjust the loss weights of the construction features of different images. For the above formula, when γ = 0, it is considered that the Focal Loss degenerates into the standard cross-entropy loss function.

[0082] In the present invention, a dropout layer can be introduced during the training process. The working principle of the dropout layer is to randomly discard a part of neurons, which can reduce the number of model parameters, lower the complexity of the model, improve the generalization performance of the model on unknown data, further enhance the generalization ability of the model, and prevent the occurrence of overfitting.

[0083] Based on the above embodiment solutions, by combining the evaluation of the construction positions of the pipe orifices in multiple dimensions (single pipe orifice position and the distance between adjacent pipe orifices), the quality inspection of the construction pipeline can be improved in terms of comprehensiveness.

[0084] The technical solution of the embodiment of the present invention includes: determining a plurality of target construction manholes corresponding to the target construction pipeline to be quality inspected; for each of the target construction manholes, determining multiple construction-related images corresponding to the target construction manhole and the actual position data corresponding to the target construction manhole; extracting features from the input construction-related images through a feature extraction model to determine the construction-related features corresponding to each construction-related image, and performing feature fusion on the multiple construction-related features to obtain the target fusion feature corresponding to the target construction manhole; determining the position evaluation score corresponding to the target construction manhole according to the actual position data; determining the manhole quality inspection result of the target construction manhole according to the target fusion feature and the position evaluation score, and determining the pipeline quality inspection result of the target construction pipeline according to the manhole quality inspection result. The present invention performs construction quality inspection on underground pipelines from multiple dimensions (image feature extraction and fusion and position evaluation), achieving the effect of improving the accurate determination of underground pipeline construction quality inspection.

[0085] Embodiment Two

[0086] Figure 4 It is a flowchart of a method for quality inspection of underground pipeline construction provided by Embodiment Two of the present invention. This embodiment details the process of extracting features from the input construction-related images through the feature extraction model to determine the construction-related features corresponding to each construction-related image. As Figure 4 shown, the method includes:

[0087] S210. Determine a plurality of target construction manholes corresponding to the target construction pipeline to be quality inspected.

[0088] S220. For each of the target construction manholes, determine multiple construction-related images corresponding to the target construction manhole and actual position data corresponding to the target construction manhole.

[0089] In an embodiment of the present invention, the construction-related images include manhole construction images, manhole pipeline images or protruding pipeline images, and the construction-related features include first construction features, second construction features or third construction features.

[0090] S230. Extract features from the manhole construction image through a first extraction model to determine the first construction feature corresponding to the manhole construction image.

[0091] S240. Extract features from the manhole pipeline image through a second extraction model to determine the second construction feature corresponding to the manhole construction image.

[0092] S250. Extract features from the protruding pipeline image through a third extraction model to determine the third construction feature corresponding to the manhole construction image.

[0093] Among them, the first extraction model, the second extraction model and the third extraction model can be understood as feature extraction models with three different model parameters. The first extraction model, the second extraction model and the third extraction model are all obtained by training the contrast learning model based on construction sample images, manhole sample images and pipeline sample images. However, the training methods are different.

[0094] Optionally, the first extraction model can be obtained by training the contrast learning model based on the manhole sample image, the pipeline sample image and multiple enhanced images corresponding to the manhole construction image. Correspondingly, the second extraction model can be obtained by training the contrast learning model based on the manhole construction image, the protruding pipeline image and multiple enhanced images corresponding to the manhole pipeline image. Correspondingly, the second extraction model can be obtained by training the contrast learning model based on the manhole construction image, the manhole pipeline image and multiple enhanced images corresponding to the protruding pipeline image.

[0095] The first construction feature, the second construction feature and the third construction feature can be sequentially understood as the extracted features corresponding to the manhole construction image, the manhole pipeline image and the protruding pipeline image.

[0096] Optionally, before extracting features from the manhole construction image through the first extraction model to determine the first construction feature corresponding to the manhole construction image, it further includes:

[0097] Determine the construction sample image, the pipe well sample image, and the pipeline sample image;

[0098] Perform data augmentation on the construction sample image respectively through the first data augmentation algorithm and the second data augmentation algorithm to determine the first augmented image and the second augmented image;

[0099] Train the contrastive learning model based on the first augmented image, the second augmented image, the pipe well sample image, and the pipeline sample image to obtain the first extraction model.

[0100] Among them, the first data augmentation algorithm and the second data augmentation can be understood as two different data augmentation algorithms. In the embodiments of the present invention, the data augmentation algorithm can be preset according to the scenario requirements and will not be specifically limited here. Optionally, the data augmentation algorithm can include algorithms such as geometric transformation, pixel transformation, and sample mixing.

[0101] The first augmented image and the second augmented image can be understood as two different data-augmented images.

[0102] The contrastive learning model can be understood as a machine model that learns the general features of an unlabeled dataset by guiding the model which data points are similar or dissimilar.

[0103] Based on the above embodiment solution, training the contrastive learning model in combination with image enhancement technology can improve the sample richness and the accuracy of the trained model.

[0104] Optionally, the training of the contrastive learning model based on the first augmented image, the second augmented image, the pipe well sample image, and the pipeline sample image to obtain the first extraction model includes:

[0105] Take the first augmented image and the second augmented image as the positive sample combination, and take the first augmented image, the pipe well sample image, and the pipeline sample image as the negative sample combination;

[0106] Perform feature extraction on the input first augmented image, second augmented image, pipe well pipeline image, and prominent pipeline image respectively through the contrastive learning model to obtain multiple sample extraction features;

[0107] Based on the positive sample combination and the negative sample combination, perform feature comparison on the multiple sample extraction features to obtain a comparison result, and adjust the model parameters of the contrastive learning model according to the comparison result to obtain the first extraction model.

[0108] Among them, the positive sample combination can be understood as a sample pair of positive samples. The negative sample combination can be understood as a sample pair of negative samples.

[0109] Figure 5 It is a flowchart for determining positive and negative sample combinations provided according to an embodiment of the present invention. Combining Figure 5 , the following specifically elaborates on taking the first enhanced image and the second enhanced image as a positive sample combination, and taking the first enhanced image, the pipe well sample image, and the pipeline sample image as a negative sample combination. Figure 5 In it, image classifications p1, p2, and p3 respectively represent a construction sample image, a pipe well sample image, and a pipeline sample image, and images and image respectively represent the first enhanced image and the second enhanced image in sequence.

[0110] Specifically, first, randomly shuffle sample pictures of different types in the model training task, and label them as p1, p2, and p3; apply two different data augmentation algorithms to p1 to generate two enhanced pictures and Take and as a positive sample combination, and take and p2, p3 as a negative sample combination. Perform encoding processing on the samples, and encode to obtain a vector s q , and encode to obtain vectors

[0111] Furthermore, extracting features from multiple samples based on the positive sample combination and the negative sample combination for feature comparison to obtain a comparison result, and adjusting the model parameters of the contrast learning model according to the comparison result can be understood as a process of guiding the model to learn the general features between similar features extracted from multiple samples based on the positive sample combination, and guiding the model to learn the general features between non-similar features extracted from multiple samples based on the negative sample combination.

[0112] The feature extraction of the sample can be understood as the feature determined by extracting features from the sample image through a contrast learning model. In the embodiment of the present invention, the first enhanced image, the second enhanced image, the pipe well pipeline image, and the prominent pipeline image can respectively correspond to different sample extraction features.

[0113] Figure 6 It is a flowchart for training a contrast learning model provided according to an embodiment of the present invention. The following specifically elaborates on training the contrast learning model to obtain the first extraction model in combination with Figure 6 .

[0114] Specifically, the loss calculation for model selection and training can be implemented based on the following formula:

[0115]

[0116] Among them, represents the comparison result; s q represents the encoded vector of the first enhanced image; represents the encoded vector of the second enhanced image; represents the encoded vector of the well sample image; represents the encoded vector of the pipeline sample image.

[0117] In the embodiments of the present invention, the second extraction model and the third extraction model can be trained in the same manner as the first extraction model, which will not be elaborated here.

[0118] After training the first extraction model, the second extraction model, and the third extraction model, the processes of feature extraction for the associated images based on each extraction model can be executed in parallel. Refer to Figure 7 , Figure 7 is a flowchart of a construction-related feature extraction provided according to an embodiment of the present invention. Among them, 33 represents the feature extraction model, and the contrast learning encodings A, B, and C represent the first extraction model, the second extraction model, and the third extraction model in sequence, and the features a, b, and c represent the first construction feature, the second construction feature, and the third construction feature in sequence.

[0119] Based on the above embodiment solutions, by using the method of combining positive and negative samples to guide the contrast learning model to learn the general features between similar features and dissimilar features in the extraction of multiple samples, the feature extraction model is trained, which can improve the accuracy of the trained model.

[0120] S260. Perform feature fusion on the multiple construction-related features to obtain the target fusion feature corresponding to the target construction well.

[0121] S270. Determine the position evaluation score corresponding to the target construction well according to the actual position data.

[0122] S280. Determine the well quality inspection result of the target construction well according to the target fusion feature and the position evaluation score, and determine the pipeline quality inspection result of the target construction pipeline according to the well quality inspection result.

[0123] In the technical solution of the embodiment of the present invention, a first extraction model is used to extract features from the pipe well construction image to determine the first construction feature corresponding to the pipe well construction image; a second extraction model is used to extract features from the pipe well pipeline image to determine the second construction feature corresponding to the pipe well construction image; a third extraction model is used to extract features from the protruding pipeline image to determine the third construction feature corresponding to the pipe well construction image. The present invention adopts a method of parallel feature extraction using different extraction models for different images to determine construction features, improving the efficiency and accuracy of feature extraction.

[0124] Embodiment III

[0125] Figure 8 FIG. is a schematic structural diagram of an underground pipeline construction quality inspection device provided in Embodiment III of the present invention. As Figure 8 shown, the device includes: a pipeline determination module 310, a data acquisition module 320, a feature processing module 330, a position evaluation module 340, and a result determination module 350.

[0126] Among them, the pipeline determination module 310 is used to determine a plurality of target construction pipe wells corresponding to the target construction pipeline to be quality inspected; the data acquisition module 320 is used to determine, for each of the target construction pipe wells, a plurality of construction-related images corresponding to the target construction pipe well and the actual position data corresponding to the target construction pipe well; the feature processing module 330 is used to extract features from the input construction-related images through a feature extraction model, determine the construction-related features corresponding to each of the construction-related images, and perform feature fusion on the plurality of construction-related features to obtain the target fusion feature corresponding to the target construction pipe well; the position evaluation module 340 is used to determine the position evaluation score corresponding to the target construction pipe well according to the actual position data; the result determination module 350 is used to determine the pipe well quality inspection result of the target construction pipe well according to the target fusion feature and the position evaluation score, and determine the pipeline quality inspection result of the target construction pipeline according to the pipe well quality inspection result.

[0127] The technical solution of the embodiment of the present invention determines a plurality of target construction manholes corresponding to the target construction pipeline to be quality inspected; for each of the target construction manholes, determines multiple construction-related images corresponding to the target construction manhole and the actual position data corresponding to the target construction manhole; extracts features from the input construction-related images through a feature extraction model, determines the construction-related features corresponding to each construction-related image, and performs feature fusion on the multiple construction-related features to obtain the target fusion feature corresponding to the target construction manhole; determines the position evaluation score corresponding to the target construction manhole according to the actual position data; determines the manhole quality inspection result of the target construction manhole according to the target fusion feature and the position evaluation score, and determines the pipeline quality inspection result of the target construction pipeline according to the manhole quality inspection result. The present invention performs construction quality inspection on underground pipelines from multiple dimensions (image feature extraction and fusion and position evaluation), achieving the effect of improving the accurate determination of underground pipeline construction quality inspection.

[0128] Optionally, the construction-related images include manhole construction images, manhole pipeline images or protruding pipeline images, and the construction-related features include first construction features, second construction features or third construction features;

[0129] The feature processing module 330 is specifically configured to:

[0130] Extract features from the manhole construction image through a first extraction model to determine the first construction feature corresponding to the manhole construction image;

[0131] Extract features from the manhole pipeline image through a second extraction model to determine the second construction feature corresponding to the manhole construction image;

[0132] Extract features from the protruding pipeline image through a third extraction model to determine the third construction feature corresponding to the manhole construction image.

[0133] Optionally, the underground pipeline construction quality inspection device further includes: a sample determination module, an image enhancement module, and a model training module;

[0134] Among them, the sample determination module is used to determine construction sample images, manhole sample images, and pipeline sample images before extracting features from the manhole construction image through the first extraction model to determine the first construction feature corresponding to the manhole construction image;

[0135] The image enhancement module is used to perform data enhancement on the construction sample images through a first data enhancement algorithm and a second data enhancement algorithm respectively to determine a first enhanced image and a second enhanced image;

[0136] The model training module is used to train a contrastive learning model based on the first enhanced image, the second enhanced image, the manhole sample image, and the pipeline sample image to obtain the first extraction model.

[0137] Optionally, the model training module is specifically configured to:

[0138] Take the first enhanced image and the second enhanced image as a positive sample combination, and take the first enhanced image, the manhole sample image, and the pipeline sample image as a negative sample combination;

[0139] Use the contrastive learning model to respectively perform feature extraction on the input first enhanced image, second enhanced image, manhole pipeline image, and protruding pipeline image to obtain multiple sample extraction features;

[0140] Based on the positive sample combination and the negative sample combination, perform feature comparison on the multiple sample extraction features to obtain a comparison result, and adjust the model parameters of the contrastive learning model according to the comparison result to obtain the first extraction model.

[0141] Optionally, the result determination module 350 includes: a pipeline quality inspection unit for determining the pipeline quality inspection result of the target construction pipeline according to the manhole quality inspection result and the position evaluation score.

[0142] Optionally, the position evaluation score includes a first evaluation score for the actual construction position of each target construction manhole and a second evaluation score for the actual distance between every two adjacent target construction manholes;

[0143] The result determination module 350 includes: a manhole quality inspection unit for determining the manhole quality inspection result of the target construction manhole according to the target fusion feature and the first evaluation score;

[0144] A pipeline quality inspection unit for determining the pipeline quality inspection result of the target construction pipeline according to the manhole quality inspection result and the second evaluation score.

[0145] Optionally, the position evaluation module 340 is specifically configured to:

[0146] For each target construction manhole, determine the preset target position data corresponding to the target construction manhole;

[0147] Determine a preset position evaluation threshold, and determine the position evaluation score corresponding to the target construction manhole according to the actual position data, the target position data, and the position evaluation threshold.

[0148] The underground pipeline construction quality inspection device provided by the embodiments of the present invention can execute the underground pipeline construction quality inspection method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects for executing the method.

[0149] Embodiment 4

[0150] Figure 9 FIG. shows a schematic structural diagram of an electronic device 10 that can be used to implement the embodiments of the present invention. The electronic device is intended to represent various forms of digital computers, such as, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, personal digital processors, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0151] As Figure 9 shown, the electronic device 10 includes at least one processor 11, and a memory communicatively connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc. Among them, the memory stores a computer program executable by the at least one processor. The processor 11 can execute various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. The input / output (I / O) interface 15 is also connected to the bus 14.

[0152] Multiple components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0153] The processor 11 may be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the underground pipeline construction quality inspection method.

[0154] In some embodiments, the underground pipeline construction quality inspection method can be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the underground pipeline construction quality inspection method described above can be executed. Alternatively, in other embodiments, the processor 11 can be configured to execute the underground pipeline construction quality inspection method by any other suitable means (e.g., by means of firmware).

[0155] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuitry, integrated circuit systems, field-programmable gate arrays (FPGA), application-specific integrated circuits (ASIC), application-specific standard products (ASSP), systems-on-a-chip (SOC), complex programmable logic devices (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special or general-purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0156] The computer program for implementing the method of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowchart and / or block diagram are implemented. The computer program can be executed entirely on the machine, partially on the machine, as a stand-alone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0157] In the context of the present invention, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, the computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0158] To provide for interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can also be used to provide for interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0159] The systems and techniques described herein can be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0160] A computing system may include a client and a server. The client and the server are generally far from each other and usually interact via a communication network. The client-server relationship is created by computer programs running on respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system, and solves the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.

[0161] It should be understood that various forms of processes shown above can be used, steps can be reordered, added or deleted. For example, the steps described in the present invention can be executed in parallel, sequentially or in a different order, as long as the desired results of the technical solution of the present invention can be achieved, and no limitation is made herein.

[0162] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A quality inspection method for underground pipeline construction, characterized in that, Including: Determine a plurality of target construction manholes corresponding to the target construction pipeline to be inspected for quality; For each of the target construction manholes, determine multiple construction-related images corresponding to the target construction manhole and the actual position data corresponding to the target construction manhole; Extract features from the input construction-related images through a feature extraction model, determine the construction-related features corresponding to each construction-related image, and perform feature fusion on multiple construction-related features to obtain the target fusion features corresponding to the target construction manhole; Determine the position evaluation score corresponding to the target construction manhole according to the actual position data; Determine the manhole quality inspection result of the target construction manhole according to the target fusion features and the position evaluation score, and determine the pipeline quality inspection result of the target construction pipeline according to the manhole quality inspection result.

2. The method according to claim 1, wherein The construction-related images include manhole construction images, manhole pipeline images or protruding pipeline images, and the construction-related features include first construction features, second construction features or third construction features; The extracting features from the input construction-related images through the feature extraction model to determine the construction-related features corresponding to each construction-related image includes: Extract features from the manhole construction image through a first extraction model to determine the first construction feature corresponding to the manhole construction image; Extract features from the manhole pipeline image through a second extraction model to determine the second construction feature corresponding to the manhole construction image; Extract features from the protruding pipeline image through a third extraction model to determine the third construction feature corresponding to the manhole construction image.

3. The method according to claim 2, characterized in that, Before extracting features from the manhole construction image through the first extraction model to determine the first construction feature corresponding to the manhole construction image, it further includes: Determine construction sample images, manhole sample images and pipeline sample images; Perform data augmentation on the construction sample images through a first data augmentation algorithm and a second data augmentation algorithm respectively to determine a first augmented image and a second augmented image; Train a contrast learning model based on the first augmented image, the second augmented image, the manhole sample image and the pipeline sample image to obtain the first extraction model.

4. The method according to claim 3, characterized in that, The training the contrast learning model based on the first augmented image, the second augmented image, the manhole sample image and the pipeline sample image to obtain the first extraction model includes: Use the first augmented image and the second augmented image as a positive sample combination, and use the first augmented image, the manhole sample image and the pipeline sample image as a negative sample combination; Extract features from the input first augmented image, second augmented image, manhole pipeline image and protruding pipeline image through the contrast learning model to obtain multiple sample extraction features; Perform feature comparison on multiple sample extraction features based on the positive sample combination and the negative sample combination to obtain a comparison result, and adjust the model parameters of the contrast learning model according to the comparison result to obtain the first extraction model.

5. The method according to claim 1, characterized in that Determining the pipeline quality inspection result of the target construction pipeline according to the quality inspection result of the pipe well includes: Determining the pipeline quality inspection result of the target construction pipeline according to the quality inspection result of the pipe well and the position evaluation score.

6. The method according to claim 5, characterized in that The position evaluation score includes a first evaluation score for the actual construction position of each target construction pipe well and a second evaluation score for the actual distance between every two adjacent target construction pipe wells; Determining the quality inspection result of the target construction pipe well according to the target fusion feature and the position evaluation score includes: Determining the quality inspection result of the target construction pipe well according to the target fusion feature and the first evaluation score; Determining the pipeline quality inspection result of the target construction pipeline according to the quality inspection result of the pipe well and the position evaluation score includes: Determining the pipeline quality inspection result of the target construction pipeline according to the quality inspection result of the pipe well and the second evaluation score.

7. The method according to claim 1, wherein Determining the position evaluation score corresponding to the target construction pipe well according to the actual position data includes: For each target construction pipe well, determining the corresponding preset target position data of the target construction pipe well; Determining a preset position evaluation threshold, and determining the position evaluation score corresponding to the target construction pipe well according to the actual position data, the target position data, and the position evaluation threshold.

8. An underground pipeline construction quality inspection device, characterized in that, Including: A pipeline determination module, configured to determine a plurality of target construction pipe wells corresponding to a target construction pipeline to be quality inspected; A data acquisition module, configured to, for each target construction pipe well, determine a plurality of construction-related images corresponding to the target construction pipe well and the actual position data corresponding to the target construction pipe well; A feature processing module, configured to perform feature extraction on the input construction-related images through a feature extraction model, determine the construction-related features corresponding to each construction-related image, and perform feature fusion on the plurality of construction-related features to obtain the target fusion feature corresponding to the target construction pipe well; A position evaluation module, configured to determine the position evaluation score corresponding to the target construction pipe well according to the actual position data; A result determination module, configured to determine the quality inspection result of the target construction pipe well according to the target fusion feature and the position evaluation score, and determine the pipeline quality inspection result of the target construction pipeline according to the quality inspection result of the pipe well.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the underground pipeline construction quality inspection method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the processor to implement the underground pipeline construction quality inspection method according to any one of claims 1-7 when executed.

Images