A hydropower station dam risk assessment method, device, equipment and medium

Abstract

The present invention relates to the field of hydropower station technology and discloses a method, apparatus, device, and medium for hydropower station dam risk assessment. The present invention identifies defects in multiple sets of two consecutive real-time sonar images within a real-time sonar image set. If the individual defect types in the two consecutive real-time sonar images are inconsistent, the two different types of individual defects are fitted into a single individual defect. The defect risk of the hydropower station dam is then assessed by calculating the defect area of ​​the fitted individual defect. Therefore, by implementing the present invention, the risk of a hydropower station dam can be accurately assessed even if different individual defects are closely spaced.

Description

Technical Field

[0001] The present invention relates to the technical field of hydropower stations, and in particular to a method, device, equipment and medium for risk assessment of a hydropower station dam. Background Art

[0002] Defect detection of hydropower station dams, especially defect detection of the underwater part of the dam, generally adopts sonar detection combined with computer-assisted image recognition technology. The identified defects are divided into individual defects and overall defects. Individual defects refer to independent defect shapes such as cracks and pits. These individual cracks and pits themselves can cause local safety hazards of the dam. Studies have found that under certain conditions, the correlation of multiple cracks and pits can also cause safety accidents. Overall defects refer to safety hazards caused by the correlation of multiple cracks and pits.

[0003] At present, technicians focus more on the assessment of the expected risk of individual defects, while there is little research on the identification and assessment of the expected risk of overall defects. In particular, research has found that the closer the distance between different individual defects, the greater the risk of mutual correlation causing overall defects. However, when the distance between different individual defects is very close, image recognition technology cannot accurately assess the risk because it is difficult to distinguish the boundaries between different individual defects. Summary of the Invention

[0004] In view of this, the present invention provides a hydropower station dam risk assessment method, device, equipment and medium to solve the problem that when the distance between different individual defects is very close, image recognition technology is difficult to distinguish the boundaries between different individual defects, resulting in an inability to accurately assess the risk.

[0005] In a first aspect, the present invention provides a method for risk assessment of a hydropower station dam, the method comprising:

[0006] A real-time sonar image set of a dam to be identified is obtained; defects are identified on multiple groups of two consecutive real-time sonar images in the real-time sonar image set to obtain a first individual defect type and a second individual defect type for each group of two consecutive real-time sonar images; based on each group of two consecutive real-time sonar images, if the first individual defect type and the second individual defect type are inconsistent, the first individual defect type and the second individual defect type are fitted to obtain a target individual defect type; the defect area of ​​the fitted target individual defect type is calculated; based on the defect area, a risk assessment is performed on the dam to be identified to obtain a risk assessment result of the dam to be identified.

[0007] The hydropower station dam risk assessment method provided by this invention identifies defects in multiple sets of two consecutive real-time sonar images within a real-time sonar image set. If the individual defect types in the two consecutive real-time sonar images are inconsistent, the two different types of individual defects are fitted into a single individual defect. The defect risk of the hydropower station dam is then assessed by calculating the defect area of ​​the fitted individual defect. Therefore, by implementing this invention, the risk of a hydropower station dam can be accurately assessed even if different individual defects are very close to each other.

[0008] In an optional embodiment, defect recognition is performed on multiple groups of two consecutive real-time sonar images in a real-time sonar image set to obtain a first individual defect type and a second individual defect type for each group of two consecutive real-time sonar images, including:

[0009] Obtain a historical sonar image set of the dam to be identified; use the historical sonar image set to establish a target full defect model and multiple target single defect models; use the target full defect model and multiple target single defect models to perform defect identification on multiple groups of two consecutive real-time sonar images in the real-time sonar image set to obtain defect identification results; based on the defect identification results, use a preset defect comprehensive identification method to perform defect identification on multiple groups of two consecutive real-time sonar images in the real-time sonar image set to obtain the first individual defect type and the second individual defect type of each group of two consecutive real-time sonar images.

[0010] The hydropower station dam risk assessment method provided by the present invention can perform preliminary defect identification by establishing a target full defect model and multiple target single defect models, and then continue to use a preset defect comprehensive identification method to identify defects, thereby improving the accuracy of defect identification.

[0011] In an optional embodiment, a target full defect model and multiple target single defect models are established using a historical sonar image set, including:

[0012] Defect recognition is performed on multiple groups of two consecutive historical sonar images in the historical sonar image set to obtain the third individual defect type and the fourth individual defect type of each group of two consecutive historical sonar images; based on the pre-trained model, the third individual defect type and the fourth individual defect type of each group of two consecutive historical sonar images are used to establish a target full defect model and multiple target single defect models.

[0013] The hydropower station dam risk assessment method provided by the present invention, based on the pre-trained model, combines the third individual defect type and the fourth individual defect type of each group of two consecutive historical sonar images to establish a corresponding target full defect model and multiple target single defect models, providing support for subsequent improvement of the accuracy of defect identification.

[0014] In an optional embodiment, based on the pre-trained model, a target full defect model and multiple target single defect models are established using the third individual defect type and the fourth individual defect type of each set of two consecutive historical sonar images, including:

[0015] Based on the pre-training model, an initial full defect model and multiple initial single defect models are established using the third individual defect type and the fourth individual defect type of each set of two consecutive historical sonar images; based on the third individual defect type and the fourth individual defect type of each set of two consecutive historical sonar images, the historical sonar image set is processed to obtain the first dam historical defect image set and the second dam historical defect image set; the initial full defect model is trained using the historical defect images of the first dam to obtain a target full defect model; and the multiple initial single defect models are trained using the historical defect image set of the second dam to obtain multiple target single defect models.

[0016] In an optional embodiment, a risk assessment is performed on the dam to be identified based on the defect area, and a risk assessment result of the dam to be identified is obtained, including:

[0017] The dam area of ​​the dam to be identified is obtained; the ratio of the defect area to the dam area is calculated; and the risk assessment result of the dam to be identified is determined based on the ratio.

[0018] The hydropower station dam risk assessment method provided by the present invention can realize risk assessment of the dam to be identified through the ratio of the defect area to the dam area of ​​the dam to be identified.

[0019] In an optional embodiment, the method further includes:

[0020] Based on each set of two consecutive real-time sonar images, if the first individual defect type is consistent with the second individual defect type, the risk assessment of the dam to be identified is performed using the preset individual defect expected risk assessment method to obtain the risk assessment result of the dam to be identified.

[0021] The hydropower station dam risk assessment method provided by the present invention can continue to perform risk assessment according to a preset individual defect expected risk assessment method if the individual defect types of two consecutive real-time sonar images are consistent.

[0022] In a second aspect, the present invention provides a hydropower station dam risk assessment device, the device comprising:

[0023] An acquisition module is used to acquire a real-time sonar image set of a dam to be identified; an identification module is used to perform defect identification on multiple groups of two consecutive real-time sonar images in the real-time sonar image set to obtain a first individual defect type and a second individual defect type for each group of two consecutive real-time sonar images; a fitting module is used to fit the first individual defect type and the second individual defect type based on each group of two consecutive real-time sonar images to obtain a target individual defect type if the first individual defect type and the second individual defect type are inconsistent; a calculation module is used to calculate the defect area of ​​the fitted target individual defect type; a first assessment module is used to perform risk assessment on the dam to be identified based on the defect area to obtain a risk assessment result of the dam to be identified.

[0024] In a third aspect, the present invention provides a computer device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the hydropower station dam risk assessment method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0025] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the hydropower station dam risk assessment method of the first aspect or any corresponding embodiment thereof.

[0026] In a fifth aspect, the present invention provides a computer program product comprising computer instructions for causing a computer to execute the hydropower station dam risk assessment method according to the first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0028] Figure 1 is a flow chart of a method for risk assessment of a hydropower station dam according to an embodiment of the present invention;

[0029] Figure 2 is a flow chart of another hydropower station dam risk assessment method according to an embodiment of the present invention;

[0030] Figure 3 is a flow chart of another hydropower station dam risk assessment method according to an embodiment of the present invention;

[0031] Figure 4 is a structural block diagram of a hydropower station dam risk assessment device according to an embodiment of the present invention;

[0032] Figure 5 Schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0033] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of the present invention.

[0034] An embodiment of the present invention provides a method for assessing the risk of a hydropower station dam. By fitting two different types of individual defects in two consecutive real-time sonar images into one individual defect, the risk of the hydropower station dam can be accurately assessed even if the distance between different individual defects is very close.

[0035] According to an embodiment of the present invention, an embodiment of a method for risk assessment of a hydropower station dam is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0036] In this embodiment, a hydropower station dam risk assessment method is provided, which can be used in electronic devices such as computers, mobile phones, tablet computers, etc. Figure 1 FIG. 1 is a flow chart of a method for risk assessment of a hydropower station dam according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0037] Step S101: Acquire a real-time sonar image set of the dam to be identified.

[0038] Specifically, obtain n grayscale images, which are marked as M1, M2, M3, ..., M n And form a real-time sonar image set.

[0039] In this embodiment, the imaging area of ​​each grayscale image is limited to be at least smaller than a fixed value.

[0040] Furthermore, since the larger the ratio of the imaging area to the overall area of ​​the dam, the more likely it is that interrelated overall defects will occur, the fixed value is set as a ratio relative to the overall area of ​​the dam, rather than the absolute value of the area.

[0041] In one embodiment, the fixed value is less than 0.003-0.005.

[0042] Step S102 : performing defect recognition on a plurality of groups of two consecutive real-time sonar images in the real-time sonar image set to obtain a first individual defect type and a second individual defect type for each group of two consecutive real-time sonar images.

[0043] Specifically, traverse all two consecutive images in the real-time sonar image set, such as M2, M3, or M k , M k+1 , and identifies whether individual defects appear in both consecutive images, and obtains the first individual defect type and the second individual defect type corresponding to the two consecutive images. Individual defects can include cracks, pits, trachoma, bulging, etc.

[0044] In step S103 , based on each set of two consecutive real-time sonar images, if the first individual defect type and the second individual defect type are inconsistent, the first individual defect type and the second individual defect type are fitted to obtain a target individual defect type.

[0045] Specifically, by comparing with a preset defect database, it can be identified whether the first individual defect type and the second individual defect type corresponding to the two consecutive images are consistent.

[0046] Furthermore, if the first individual defect type is inconsistent with the second individual defect type, individual defects of different types in two consecutive real-time sonar images are fitted into one individual defect.

[0047] Furthermore, the above operations are performed on two consecutive real-time sonar images respectively until the final fitted target individual defect type is obtained.

[0048] Step S104: Calculate the defect area of ​​the target individual defect type after fitting.

[0049] Specifically, the finite element partitioning method can be used to divide the individual defect surface into n micro units, and the area of ​​each micro unit and the number of micro units can be calculated to obtain the defect area.

[0050] Furthermore, the above calculation process ignores the error in the microcell area at the defect boundary. The area of ​​each microcell is preset by the system and is related to the sonar acquisition device's "field of view." For example, if the device's "field of view" accommodates m microcells, the number or area of ​​individual defect microcells contained in each "field of view" is n / m "fields of view." The "field of view" area range is fixed and related to the device configuration.

[0051] Step S105: Based on the defect area, a risk assessment is performed on the dam to be identified to obtain a risk assessment result of the dam to be identified.

[0052] Specifically, the defect risk of the hydropower station dam can be evaluated in combination with the calculated defect area, and the corresponding risk assessment results can be obtained.

[0053] The hydropower station dam risk assessment method provided in this embodiment identifies defects in multiple sets of two consecutive real-time sonar images within a real-time sonar image set. If the individual defect types in the two consecutive real-time sonar images are inconsistent, the two different types of individual defects are fitted into a single individual defect. The defect risk of the hydropower station dam is then assessed by calculating the defect area of ​​the fitted individual defect. Therefore, by implementing this method, the risk of a hydropower station dam can be accurately assessed even if different individual defects are very close to each other.

[0054] In this embodiment, a hydropower station dam risk assessment method is provided, which can be used in electronic devices such as computers, mobile phones, tablet computers, etc. Figure 2 FIG. 1 is a flow chart of a method for risk assessment of a hydropower station dam according to an embodiment of the present invention. Figure 2 As shown, the process includes the following steps:

[0055] Step S201: Acquire a real-time sonar image set of the dam to be identified. Figure 1 Step S101 of the illustrated embodiment will not be described in detail here.

[0056] Step S202 : performing defect recognition on a plurality of groups of two consecutive real-time sonar images in the real-time sonar image set to obtain a first individual defect type and a second individual defect type for each group of two consecutive real-time sonar images.

[0057] Specifically, the above step S202 includes:

[0058] Step S2021: Obtain a historical sonar image set of the dam to be identified.

[0059] The specific process can be referred to the description of step S101 above, which will not be repeated here.

[0060] Step S2022: Using the historical sonar image set, a target full defect model and multiple target single defect models are established.

[0061] Among them, the target full defect model includes all defects in the image and can be used to identify which defects are in the image and their locations; the target single defect model only includes a certain type of defect and can be used to identify whether a certain type of defect is in the image and its location.

[0062] In some optional implementations, the above step S2022 includes:

[0063] Step a1: performing defect recognition on multiple groups of two consecutive historical sonar images in the historical sonar image set to obtain the third individual defect type and the fourth individual defect type of each group of two consecutive historical sonar images.

[0064] In step a2, based on the pre-trained model, a target full defect model and multiple target single defect models are established using the third individual defect type and the fourth individual defect type of each set of two consecutive historical sonar images.

[0065] Specifically, the existing defect recognition method can be used to perform defect recognition on multiple groups of two consecutive historical sonar images in the historical sonar image set, and the third individual defect type and the fourth individual defect type corresponding to each group of two consecutive historical sonar images can be obtained.

[0066] Furthermore, a pretrained model was selected using the TensorFlow framework. Pretrained models can be downloaded from the open-source TensorFlow project on GitHub. The TensorFlow pretrained model library provides multiple models trained on datasets such as COCO, KITTI, Open Images, and AVAv2.1. After testing and comparison, this example selected the FasterR-CNN InceptionResNet pretrained model, which yielded a highly effective defect recognition model.

[0067] Furthermore, using the Tensorflow framework, on the basis of the pre-trained model, the corresponding target full defect model and multiple target single defect models can be trained and obtained according to the third individual defect type and the fourth individual defect type of each group of two consecutive historical sonar images.

[0068] In some optional implementations, the above step a2 includes:

[0069] In step a21, based on the pre-trained model, an initial full defect model and multiple initial single defect models are established using the third individual defect type and the fourth individual defect type of each set of two consecutive historical sonar images.

[0070] Step a22: Based on the third individual defect type and the fourth individual defect type of each set of two consecutive historical sonar images, the historical sonar image set is processed to obtain a first dam historical defect image set and a second dam historical defect image set.

[0071] Step a23: Use the historical defect image of the first dam to train the initial full defect model to obtain the target full defect model.

[0072] Step a24: Use the second dam historical defect image set to train multiple initial single defect models to obtain multiple target single defect models.

[0073] Specifically, the Tensorflow framework is used, and on the basis of the pre-trained model, an initial full defect model and multiple initial single defect models are established according to the third individual defect type and the fourth individual defect type of each group of two consecutive historical sonar images.

[0074] Furthermore, the corresponding dam historical defect image set can be determined based on the third individual defect type and the fourth individual defect type of each group of two consecutive historical sonar images, and the obtained dam historical defect image set is stored in the image subfolder in the model folder. After sample processing, the corresponding first dam historical defect image set and second dam historical defect image set are obtained and used for model training and evaluation.

[0075] The sample processing process may include sample annotation, sample inspection, and sample file (including image files and annotation files) conversion.

[0076] Furthermore, at least 200 image samples are required in the first dam historical defect image set and the second dam historical defect image set.

[0077] Furthermore, the processed first dam historical defect image set and the second dam historical defect image set are used to train a full defect model and multiple single defect models respectively.

[0078] Among them, the full-defect model is trained using image samples labeled with all defects in the image, namely the first dam historical defect image set; the single-defect model is trained using image samples labeled with only the corresponding defects, namely the second dam historical defect image set.

[0079] In step S2023 , the target full defect model and multiple target single defect models are used to perform defect recognition on multiple groups of two consecutive real-time sonar images in the real-time sonar image set to obtain defect recognition results.

[0080] Specifically, the trained target full-defect model and multiple target single-defect models can be integrated into the inspection APP software and video monitoring software.

[0081] Furthermore, the trained target full-defect model and multiple target single-defect models can be used to perform defect recognition on multiple groups of two consecutive real-time sonar images in the real-time sonar image set obtained by the inspection APP software and video monitoring software, and obtain corresponding defect recognition results.

[0082] Step S2024: Based on the defect recognition result, a preset comprehensive defect recognition method is used to perform defect recognition on multiple groups of two consecutive real-time sonar images in the real-time sonar image set to obtain a first individual defect type and a second individual defect type for each group of two consecutive real-time sonar images.

[0083] The preset comprehensive defect identification method may be the predict method of the keras.Model class in the TensorFlow framework or a method in another framework.

[0084] Specifically, based on the defect identification of the target full defect model and multiple target single defect models, the preset defect comprehensive identification method is continued to be used for defect identification, and the first individual defect type and the second individual defect type of each group of two consecutive real-time sonar images can be finally determined.

[0085] Step S203: Based on each set of two consecutive real-time sonar images, if the first individual defect type and the second individual defect type are inconsistent, the first individual defect type and the second individual defect type are fitted to obtain the target individual defect type. Figure 1 Step S103 of the illustrated embodiment will not be described in detail here.

[0086] Step S204, calculate the defect area of ​​the target individual defect type after fitting. Figure 1 Step S104 of the illustrated embodiment will not be described in detail here.

[0087] Step S205: Based on the defect area, a risk assessment is performed on the dam to be identified to obtain a risk assessment result for the dam to be identified. Figure 1 Step S105 of the illustrated embodiment will not be described in detail here.

[0088] The hydropower station dam risk assessment method provided in this embodiment, based on the pre-trained model, combines the third individual defect type and the fourth individual defect type of each group of two consecutive historical sonar images identified to establish a corresponding target full defect model and multiple target single defect models. Furthermore, by establishing the target full defect model and multiple target single defect models, preliminary defect identification can be performed, and then the preset defect comprehensive identification method is further used to perform defect identification, thereby improving the accuracy of defect identification. Furthermore, defect identification is performed on multiple groups of two consecutive real-time sonar images in the real-time sonar image set. If the individual defect types of the two consecutive real-time sonar images are inconsistent, the two different types of individual defects are fitted into one individual defect, and then the defect risk of the hydropower station dam can be assessed by calculating the defect area of ​​the fitted individual defect. Therefore, by implementing the present invention, the risk of a hydropower station dam can be accurately assessed even if the distance between different individual defects is very close.

[0089] In this embodiment, a hydropower station dam risk assessment method is provided, which can be used in electronic devices such as computers, mobile phones, tablet computers, etc. Figure 3 FIG. 1 is a flow chart of a method for risk assessment of a hydropower station dam according to an embodiment of the present invention. Figure 3 As shown, the process includes the following steps:

[0090] Step S301: Acquire a real-time sonar image set of the dam to be identified. Figure 1 Step S101 of the illustrated embodiment will not be described in detail here.

[0091] Step S302: perform defect recognition on multiple sets of two consecutive real-time sonar images in the real-time sonar image set to obtain the first individual defect type and the second individual defect type of each set of two consecutive real-time sonar images. Figure 2 Step S202 of the illustrated embodiment will not be described in detail here.

[0092] Step S303: Based on each set of two consecutive real-time sonar images, if the first individual defect type and the second individual defect type are inconsistent, the first individual defect type and the second individual defect type are fitted to obtain the target individual defect type. Figure 1 Step S103 of the illustrated embodiment will not be described in detail here.

[0093] Step S304, calculate the defect area of ​​the target individual defect type after fitting. Figure 1 Step S104 of the illustrated embodiment will not be described in detail here.

[0094] Step S305: Based on the defect area, a risk assessment is performed on the dam to be identified to obtain a risk assessment result of the dam to be identified.

[0095] Specifically, the above step S305 includes:

[0096] Step S3051: Obtain the dam area of ​​the dam to be identified.

[0097] Specifically, the overall dam area of ​​the dam to be identified is obtained.

[0098] Step S3052, calculate the ratio of the defect area to the dam area.

[0099] Specifically, the ratio of the defect area to the overall dam area is calculated.

[0100] Step S3053: Determine the risk assessment result of the dam to be identified based on the ratio.

[0101] Specifically, the larger the ratio of the defect area to the overall dam area, the higher the risk of the dam to be identified.

[0102] In this embodiment, when the ratio of the defect area to the overall dam area is greater than 0.0015-0.0018, the dam to be identified is deemed to be at risk.

[0103] In some optional implementations, after the above step S302, the process further includes the following steps:

[0104] Step S306: Based on each set of two consecutive real-time sonar images, if the first individual defect type is consistent with the second individual defect type, a risk assessment is performed on the dam to be identified using a preset individual defect expected risk assessment method to obtain a risk assessment result of the dam to be identified.

[0105] Specifically, if the first individual defect type and the second individual defect type corresponding to the two consecutive images are consistent, a conventional assessment can be directly performed according to the existing expected risk assessment method for individual defects.

[0106] The hydropower station dam risk assessment method provided in this embodiment identifies defects in multiple sets of two consecutive real-time sonar images within a real-time sonar image set. If the individual defect types in the two consecutive real-time sonar images are inconsistent, the two different types of individual defects are fitted into a single individual defect. The defect risk of the hydropower station dam can then be assessed by calculating the defect area of ​​the fitted individual defect. Furthermore, if the individual defect types in the two consecutive real-time sonar images are consistent, the risk assessment can continue according to a preset individual defect expected risk assessment method. Therefore, by implementing the present invention, the risk of a hydropower station dam can be accurately assessed even if different individual defects are very close together.

[0107] In one example, a fuzzy identification method for overall defects of a hydropower station dam is provided, specifically comprising:

[0108] S1, obtain sonar image

[0109] Specifically, obtain n grayscale images, which are marked as M1, M2, M3, ..., M n ;

[0110] Among them, in order to ensure that this method can be truly used to blur the overall defects, rather than forcibly fitting two individual defects that are far apart into an overall defect, the imaging area of ​​each grayscale image must be limited to at least less than a value. This value is not the absolute value of the area, but the ratio relative to the entire area of ​​the dam. Because the larger the ratio of the imaging area to the overall area of ​​the dam, the easier it is to produce interrelated overall defects. In a preferred embodiment, the ratio can be less than 0.003-0.005.

[0111] S2, Identify different types of individual defects

[0112] Specifically, traverse all two consecutive pictures, such as M2, M3, or M k , M k+1 , identify whether individual defects, including cracks, pits, trachoma, and arching, appear in two consecutive images;

[0113] Furthermore, the system compares the two consecutive images to the database to identify whether the individual defect types appearing in each image are consistent. If they are consistent, a conventional assessment is performed according to the expected risk assessment method for individual defects. If they are inconsistent, the different types of individual defects in the two images are fitted into one individual defect.

[0114] The method for identifying different types of individual defects uses transfer learning technology, which transfers knowledge from a model that has been trained on a large-scale dataset to another model, retaining the feature extractor and retraining the classifier on a new dataset with similar features. Since the feature extractor does not need to be retrained, the number of parameters that need to be trained is greatly reduced, and only a small sample is required. Specifically, the following are some of the methods:

[0115] (1) Select a pre-trained model using the TensorFlow framework. The pre-trained model can be downloaded from the open source project TensorFlow on the GitHub hosting platform. The TersonFlow pre-trained model library provides multiple models that have been trained on datasets such as COCO, Kitti, OpenImages, and AVAv2.1. After testing and comparison, the present invention selected the Faster R-CNNInceptionResNet pre-trained model, which can be trained to obtain a better earth-rock dam defect recognition model;

[0116] (2) Establish a defect recognition model. Using the Tensorflow framework, based on the Faster R-CNN InceptionResNet pre-trained model, a defect recognition model is established according to the defect types obtained in the previous step. The defect recognition model includes a full defect model and multiple single defect models. The full defect model contains all defects and is used to identify which defects are in the image and their locations. The single defect model only contains a certain type of defect and is used to identify whether a certain type of defect is in the image and its location. The combination of the two improves the accuracy of defect recognition.

[0117] (3) Sample processing: The collected dam defect images are stored in the image subfolder of the model folder. They must be processed before they can be used for model training and evaluation. The sample processing process includes sample annotation, sample inspection, and sample file (including image files and annotation files) conversion. This method requires at least 200 image samples for each defect;

[0118] (4) Model training and evaluation: Use the processed samples to train a full-defect model and multiple single-defect models. The full-defect model training uses image samples labeled with all defects in the image, and the single-defect model training only uses image samples labeled with the corresponding defects;

[0119] (5) Defect recognition: After training, the defect recognition model can be integrated into the inspection APP software and video surveillance software. The predict method of the keras.Model class in the TensorFlow framework or other methods in other frameworks can be used to identify defects in inspection images or video surveillance images. After the inspection image is identified by the defect recognition model, the defect type is finally determined using the comprehensive defect recognition method.

[0120] S3, calculate the individual defect area after fitting;

[0121] S4. Evaluate the expected risk of the overall defect based on the fitted individual defect area. The larger the ratio of the fitted individual defect area to the overall dam area, the greater the expected risk. In a preferred embodiment, when the ratio is set to be greater than 0.0015-0.0018, it is determined to be risky.

[0122] This embodiment also provides a hydropower station dam risk assessment device, which is used to implement the above-mentioned embodiments and preferred implementations. Details already described will not be repeated here. As used below, the term "module" may refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and contemplated.

[0123] This embodiment provides a hydropower station dam risk assessment device, such as Figure 4 As shown, the device includes:

[0124] The acquisition module 401 is used to acquire a real-time sonar image set of the dam to be identified.

[0125] The identification module 402 is configured to perform defect identification on a plurality of groups of two consecutive real-time sonar images in the real-time sonar image set, and obtain a first individual defect type and a second individual defect type for each group of two consecutive real-time sonar images.

[0126] The fitting module 403 is configured to fit the first individual defect type and the second individual defect type based on each set of two consecutive real-time sonar images to obtain a target individual defect type if the first individual defect type and the second individual defect type are inconsistent.

[0127] The calculation module 404 is used to calculate the defect area of ​​the target individual defect type after fitting.

[0128] The first assessment module 405 is configured to perform risk assessment on the dam to be identified based on the defect area, and obtain a risk assessment result of the dam to be identified.

[0129] In some optional implementations, the identification module 402 includes:

[0130] The first acquisition submodule is used to acquire a historical sonar image set of the dam to be identified.

[0131] A submodule is established to use historical sonar image sets to establish a target full defect model and multiple target single defect models.

[0132] The first recognition submodule is used to perform defect recognition on multiple groups of two consecutive real-time sonar images in the real-time sonar image set using a target full-defect model and multiple target single-defect models to obtain defect recognition results.

[0133] The second identification submodule is used to perform defect identification on multiple groups of two consecutive real-time sonar images in the real-time sonar image set based on the defect identification results using a preset defect comprehensive identification method to obtain the first individual defect type and the second individual defect type of each group of two consecutive real-time sonar images.

[0134] In some optional implementations, establishing the submodule includes:

[0135] The recognition unit is used to perform defect recognition on multiple groups of two consecutive historical sonar images in the historical sonar image set, and obtain the third individual defect type and the fourth individual defect type of each group of two consecutive historical sonar images.

[0136] A unit is established for establishing a target full defect model and multiple target single defect models based on a pre-trained model and using the third individual defect type and the fourth individual defect type of each group of two consecutive historical sonar images.

[0137] In some optional embodiments, the establishing unit includes:

[0138] A subunit is established for establishing an initial full defect model and multiple initial single defect models based on a pre-trained model by using the third individual defect type and the fourth individual defect type of each group of two consecutive historical sonar images.

[0139] The processing subunit is used to process the historical sonar image set based on the third individual defect type and the fourth individual defect type of each group of two consecutive historical sonar images to obtain the first dam historical defect image set and the second dam historical defect image set.

[0140] The first training subunit is used to train the initial full defect model using the historical defect image of the first dam to obtain a target full defect model.

[0141] The second training subunit is used to train multiple initial single defect models using the second dam historical defect image set to obtain multiple target single defect models.

[0142] In some optional implementations, the first evaluation module 405 includes:

[0143] The second acquisition submodule is used to acquire the dam area of ​​the dam to be identified.

[0144] The calculation submodule is used to calculate the ratio of the defect area to the dam area.

[0145] The determination submodule is used to determine the risk assessment result of the dam to be identified based on the ratio.

[0146] In some optional embodiments, the device further comprises:

[0147] The second evaluation module is used to perform risk assessment on the dam to be identified based on each set of two consecutive real-time sonar images. If the first individual defect type is consistent with the second individual defect type, a preset individual defect expected risk assessment method is used to perform risk assessment on the dam to be identified, thereby obtaining a risk assessment result for the dam to be identified.

[0148] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0149] The hydropower station dam risk assessment device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0150] The embodiment of the present invention also provides a computer device having the above Figure 4 The hydropower station dam risk assessment device shown.

[0151] See also Figure 5 , Figure 5 is a structural diagram of a computer device provided by an optional embodiment of the present invention, such as Figure 5As shown, the computer device includes: one or more processors 10, memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components utilize different buses to communicate with each other and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in the memory or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Equally, multiple computer devices can be connected, and each device provides part of the necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 5 A processor 10 is taken as an example.

[0152] The processor 10 may be a central processing unit, a network processor, or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit, a programmable logic device, or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic, or any combination thereof.

[0153] The memory 20 stores instructions that can be executed by at least one processor 10, so as to enable at least one processor 10 to execute the method shown in the above embodiment.

[0154] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created based on the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0155] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0156] The computer device further includes a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0157] The embodiment of the present invention also provides a computer-readable storage medium. The above-mentioned method according to the embodiment of the present invention can be implemented in hardware, firmware, or implemented as a computer code that can be recorded in a storage medium, or implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor or hardware, the method shown in the above embodiment is implemented.

[0158] A portion of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the form in which the computer program instruction exists in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc. Accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium that can be accessed by the computer.

[0159] Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention. Such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for risk assessment of a hydropower station dam, characterized in that: The method comprises: Obtain a real-time sonar image set of the dam to be identified; performing defect recognition on a plurality of groups of two consecutive real-time sonar images in the real-time sonar image set to obtain a first individual defect type and a second individual defect type for each group of two consecutive real-time sonar images; Based on each set of two consecutive real-time sonar images, if the first individual defect type and the second individual defect type are inconsistent, the first individual defect type and the second individual defect type are fitted to obtain a target individual defect type; Calculate the defect area of ​​the target individual defect type after fitting; performing a risk assessment on the dam to be identified based on the defect area to obtain a risk assessment result of the dam to be identified; The defect recognition is performed on a plurality of groups of two consecutive real-time sonar images in the real-time sonar image set to obtain a first individual defect type and a second individual defect type for each group of two consecutive real-time sonar images, including: Acquire a historical sonar image set of the dam to be identified; Using the historical sonar image set, establishing a target full defect model and multiple target single defect models; Using the target full defect model and the multiple target single defect models, defect recognition is performed on multiple groups of two consecutive real-time sonar images in the real-time sonar image set to obtain defect recognition results; Based on the defect recognition result, defect recognition is performed on multiple groups of two consecutive real-time sonar images in the real-time sonar image set using a preset comprehensive defect recognition method to obtain the first individual defect type and the second individual defect type of each group of two consecutive real-time sonar images.

2. The method according to claim 1, characterized in that Using the historical sonar image set, a target full defect model and multiple target single defect models are established, including: performing defect recognition on a plurality of groups of two consecutive historical sonar images in the historical sonar image set to obtain a third individual defect type and a fourth individual defect type of each group of two consecutive historical sonar images; Based on the pre-training model, the target full defect model and the multiple target single defect models are established by utilizing the third individual defect type and the fourth individual defect type of each group of two consecutive historical sonar images.

3. The method according to claim 2, characterized in that Based on the pre-trained model, the target full defect model and the multiple target single defect models are established using the third individual defect type and the fourth individual defect type of each set of two consecutive historical sonar images, including: Based on the pre-trained model, an initial full defect model and multiple initial single defect models are established using the third individual defect type and the fourth individual defect type of each set of two consecutive historical sonar images; Based on the third individual defect type and the fourth individual defect type of each set of two consecutive historical sonar images, the historical sonar image set is processed to obtain a first dam historical defect image set and a second dam historical defect image set; Using the first dam historical defect image to train the initial full defect model to obtain the target full defect model; The plurality of initial single defect models are trained using the second dam historical defect image set to obtain the plurality of target single defect models.

4. The method according to claim 1, wherein Based on the defect area, a risk assessment is performed on the dam to be identified to obtain a risk assessment result of the dam to be identified, including: Obtaining the dam area of ​​the dam to be identified; Calculating a ratio of the defect area to the dam area; Based on the ratio, the risk assessment result of the dam to be identified is determined.

5. The method according to claim 1, wherein The method further comprises: Based on each group of two consecutive real-time sonar images, if the first individual defect type is consistent with the second individual defect type, a risk assessment of the dam to be identified is performed using a preset individual defect expected risk assessment method to obtain the risk assessment result of the dam to be identified.

6. A hydropower station dam risk assessment device, characterized in that: The device comprises: An acquisition module is used to acquire a real-time sonar image set of the dam to be identified; An identification module is configured to perform defect identification on a plurality of groups of two consecutive real-time sonar images in the real-time sonar image set, and obtain a first individual defect type and a second individual defect type for each group of two consecutive real-time sonar images; a fitting module, configured to fit the first individual defect type and the second individual defect type based on each set of two consecutive real-time sonar images to obtain a target individual defect type if the first individual defect type and the second individual defect type are inconsistent; A calculation module, used to calculate the defect area of ​​the target individual defect type after fitting; A first assessment module is configured to perform a risk assessment on the dam to be identified based on the defect area to obtain a risk assessment result of the dam to be identified; Wherein, the identification module includes: A first acquisition submodule is configured to acquire a historical sonar image set of the dam to be identified; Establishing a submodule for establishing a target full defect model and multiple target single defect models using the historical sonar image set; A first identification submodule is configured to perform defect identification on a plurality of groups of two consecutive real-time sonar images in the real-time sonar image set using the target full defect model and the plurality of target single defect models to obtain a defect identification result; The second identification submodule is used to perform defect identification on multiple groups of two consecutive real-time sonar images in the real-time sonar image set based on the defect identification result using a preset defect comprehensive identification method to obtain the first individual defect type and the second individual defect type of each group of two consecutive real-time sonar images.

7. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the hydropower station dam risk assessment method according to any one of claims 1 to 5 by executing the computer instructions.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the hydropower station dam risk assessment method according to any one of claims 1 to 6.

9. A computer program product, characterized in that The method comprises computer instructions for causing a computer to execute the hydropower station dam risk assessment method according to any one of claims 1 to 5.

Images