CN114299475A - Method for detecting corrosion of damper and related equipment - Google Patents
Method for detecting corrosion of damper and related equipment Download PDFInfo
- Publication number
- CN114299475A CN114299475A CN202111592340.4A CN202111592340A CN114299475A CN 114299475 A CN114299475 A CN 114299475A CN 202111592340 A CN202111592340 A CN 202111592340A CN 114299475 A CN114299475 A CN 114299475A
- Authority
- CN
- China
- Prior art keywords
- damper
- image
- trained
- unit
- detection
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The invention discloses a method and related equipment for detecting corrosion of a damper, which can obtain a target image to be subjected to the corrosion detection of the damper, detect whether the damper exists in the target image by utilizing a trained damper detection model, if so, obtain coordinate information of a damper area in the target image output by the trained damper detection model, intercept the damper area image from the target image based on the coordinate information of the damper area, and detect whether the damper in the damper area image is corroded by utilizing the trained corrosion classification model. According to the invention, the anti-vibration hammer corrosion detection can be carried out on the target image through the anti-vibration hammer detection model and the corrosion classification model, manual detection is not needed, human resources required by manual detection are reduced, when more images are required to be subjected to the anti-vibration hammer corrosion detection, the detection efficiency can be effectively improved, the possible missing detection and false detection of the manual detection can be avoided, and the accuracy of the anti-vibration hammer corrosion detection is effectively ensured.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a method for detecting corrosion of a damper and related equipment.
Background
In the frame control transmission line, the damper can be used for guaranteeing the safety and reliability of normal work of the transmission line.
The damper may be a protection fitting installed on the wire, and is used to suppress or reduce vibration of the wire due to wind. The damper can be corroded under the action of long-term external environment, the corrosion can affect the overall performance of the damper, the connection strength of the damper and other parts is reduced, and the reliability of normal work of the power transmission line is reduced. In daily patrolling and examining transmission line, the staff can shoot the damper through camera device, and whether corrosion appears in the damper is confirmed to the outward appearance form of the damper in the artifical observation shooting obtained image to in time handle when the corrosion appears in the damper, ensure transmission line's working property.
However, when a large number of images of the damper are captured, a manual observation is required to determine whether the damper is rusted, which results in high labor consumption and low efficiency.
Disclosure of Invention
In view of the above problems, the present invention provides a method and related apparatus for detecting corrosion of a damper, which overcomes or at least partially solves the above problems, and adopts the following technical solutions:
a method for detecting corrosion of a damper comprises the following steps:
obtaining a target image to be subjected to the corrosion detection of the damper;
detecting whether the damper exists in the target image or not by using the trained damper detection model, and if so, obtaining coordinate information of a damper area in the target image output by the trained damper detection model;
intercepting a damper area image from the target image based on the coordinate information of the damper area;
and detecting whether the damper in the damper region image is rusted or not by using the trained rust classification model.
Optionally, the trained rust classification model includes a multilayer network; the trained rust classification model comprises a first layer network, a second layer network, a third layer network, a thirty-fourth layer network, an average pooling layer and a thirty-sixth layer network, wherein the first layer network in the trained rust classification model is a 7 x 7 convolutional layer with the use step size of 2, the second layer network is a pooling layer with the use step size of 2, the third layer to the thirty-fourth layer are residual block structures, the thirty-fifth layer is an average pooling layer, and the thirty-sixth layer is an output layer formed by all-connected layers.
Optionally, the method further includes:
obtaining a plurality of damper images; wherein each damper image is a patrol inspection image containing a damper;
respectively preprocessing each damper image according to a predefined image preprocessing mode to obtain each damper image after preprocessing;
using the at least partially preprocessed damper image, creating a first image set for the damper detection model to be trained;
and training the damper detection model to be trained by using the first image set to obtain the trained damper detection model.
Optionally, the method further includes:
inputting each preprocessed damper image into the trained damper detection model; obtaining coordinate information of a damper area in each damper image respectively output by the trained damper detection model;
intercepting damper area images from each damper image respectively based on coordinate information of the damper area in each damper image;
preprocessing each damper area image cut out from each damper image according to the image preprocessing mode to obtain each damper area image after preprocessing;
creating a second image set for training the rust classification model to be trained by using the preprocessed each damper area image;
and training the rust classification model to be trained by using the second image set to obtain the trained rust classification model.
A damper rust detection device comprising: the device comprises a first obtaining unit, a first detecting unit, a second obtaining unit, a first intercepting unit and a second detecting unit; wherein:
the first obtaining unit is used for obtaining a target image to be subjected to the corrosion detection of the damper;
the first detection unit is used for detecting whether the damper exists in the target image by utilizing a trained damper detection model, and if so, the second obtaining unit is triggered;
the second obtaining unit is configured to obtain coordinate information of a damper area in the target image output by the trained damper detection model;
the first intercepting unit is used for intercepting a damper area image from the target image based on the coordinate information of the damper area;
and the second detection unit is used for detecting whether the damper in the damper area image is rusted or not by using the trained rust classification model.
Optionally, the trained rust classification model includes a multilayer network; the trained rust classification model comprises a first layer network, a second layer network, a third layer network, a thirty-fourth layer network, an average pooling layer and a thirty-sixth layer network, wherein the first layer network in the trained rust classification model is a 7 x 7 convolutional layer with the use step size of 2, the second layer network is a pooling layer with the use step size of 2, the third layer to the thirty-fourth layer are residual block structures, the thirty-fifth layer is an average pooling layer, and the thirty-sixth layer is an output layer formed by all-connected layers.
Optionally, the apparatus further comprises: the device comprises a third obtaining unit, a first preprocessing unit, a fourth obtaining unit, a first creating unit, a first training unit and a fifth obtaining unit; wherein:
the third obtaining unit is used for obtaining a plurality of damper images; wherein each damper image is a patrol inspection image containing a damper;
the first preprocessing unit is used for preprocessing each damper image according to a predefined image preprocessing mode;
the fourth obtaining unit is configured to obtain each preprocessed damper image;
the first creating unit is used for creating a first image set of the damper detection model to be trained by using at least part of the preprocessed damper images;
the first training unit is used for training the damper detection model to be trained by using the first image set;
and the fifth obtaining unit is used for obtaining the trained damper detection model.
Optionally, the apparatus further comprises: the device comprises a first input unit, a sixth obtaining unit, a second intercepting unit, a second preprocessing unit, a seventh obtaining unit, a second creating unit, a second training unit and an eighth obtaining unit; wherein:
the first input unit is used for respectively inputting the preprocessed damper images to the trained damper detection model;
the sixth obtaining unit is configured to obtain coordinate information of a damper area in each damper image that is output by the trained damper detection model;
the second intercepting unit is used for intercepting damper area images from the damper images respectively based on the coordinate information of the damper areas in the damper images;
the second preprocessing unit is configured to respectively preprocess each of the damper region images captured from each of the damper images according to the image preprocessing method;
the seventh obtaining unit is configured to obtain preprocessed each damper region image;
the second creating unit is configured to create a second image set used for training the rust classification model to be trained, using the preprocessed each of the damper region images;
the second training unit is used for training the rust classification model to be trained by using the second image set;
and the eighth obtaining unit is used for obtaining the trained corrosion classification model.
A computer readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements any of the damper rust detection methods described above.
A processor for running a program, wherein the program when running implements any of the above-described damper rust detection methods.
Optionally, an electronic device includes:
one or more processors;
a storage device having one or more programs stored thereon;
the one or more programs, when executed by the one or more processors, cause the one or more processors to implement any of the damper rust detection methods described above.
The method and the related equipment for detecting the corrosion of the damper can obtain a target image to be subjected to the corrosion detection of the damper, detect whether the damper exists in the target image by using a trained damper detection model, if so, obtain the coordinate information of a damper area in the target image output by the trained damper detection model, intercept the image of the damper area from the target image based on the coordinate information of the damper area, and detect whether the damper in the image of the damper area has the corrosion by using the trained corrosion classification model. According to the invention, the anti-vibration hammer corrosion detection can be carried out on the target image through the anti-vibration hammer detection model and the corrosion classification model, manual detection is not needed, human resources required by manual detection are reduced, when more images are required to be subjected to the anti-vibration hammer corrosion detection, the detection efficiency can be effectively improved, the possible missing detection and false detection of the manual detection can be avoided, and the accuracy of the anti-vibration hammer corrosion detection is effectively ensured.
The foregoing description is only an overview of the technical solutions of the present invention, and the following detailed description of the present invention is provided to enable the technical means of the present invention to be more clearly understood, and to enable the above and other objects, features, and advantages of the present invention to be more clearly understood.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
Fig. 1 is a flowchart illustrating a first method for detecting corrosion of a damper according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram illustrating a first corrosion detection apparatus for a damper according to an embodiment of the present invention;
fig. 3 shows a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
Exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.
As shown in fig. 1, the present embodiment proposes a first method for detecting corrosion of a damper, which may include the steps of:
s101, obtaining a target image to be subjected to corrosion detection of the damper;
it should be noted that the present invention can be applied to target electronic devices, such as a mobile phone, a tablet computer, and a server.
The target image may be an image to be detected whether the tarnishing of the damper is present. It is understood that the detection of the tarnish of the damper on the target image may include detecting the presence of the damper in the target image, and if so, further detecting the presence of tarnish in the damper in the target image.
Alternatively, the target image may be an image containing one or more dampers; alternatively, the target image may be an image not including the damper.
Optionally, the target image may be obtained by shooting the damper on the power transmission line by a certain shooting means when the worker inspects the power transmission line. Wherein, the means of shooing can include that remote control unmanned aerial vehicle shoots and handheld mobile terminal who sets up the camera shoots the means such as.
Optionally, the target image may also be obtained by shooting the damper on the power transmission line by using a camera device arranged around the power transmission line.
Alternatively, the target image may be an image acquired from the internet.
S102, detecting whether the damper exists in the target image by using the trained damper detection model, and if so, executing a step S103; otherwise, forbidding to execute the subsequent steps, and avoiding the unnecessary consumption of resources.
The damper detection model may be a machine learning model for detecting whether a damper exists in a certain image, and for labeling each damper existing in the image by using a frame (such as a rectangular frame or a circular frame).
Optionally, the trained damper detection model may be obtained by training the damper detection model to be trained;
alternatively, the damper detection model may be a target detection model obtained by machine learning a neural network model using a deep learning technique. The specific type of the damper detection model is not limited, for example, the damper detection model can be a Cascade R-CNN model and other common target detection models.
Specifically, the method detects whether the damper exists in the target image through a trained damper detection model. According to the invention, after the target image is obtained, the target image is input into the trained damper detection model, and the detection result output by the trained damper detection model is obtained.
If the detection result shows that the damper does not exist in the target image, the method can forbid the follow-up step for detecting whether the damper is rusted or not, thereby avoiding the consumption of useless resources and reducing the consumption of resources.
If the detection result is that the damper exists in the target image, the method can execute the subsequent step for detecting whether the damper is rusted or not so as to determine whether the damper can continuously realize the function of inhibiting the vibration of the power transmission line or not and guarantee the safety and the reliability of the normal work of the power transmission line.
S103, obtaining coordinate information of a damper area in a target image output by the trained damper detection model;
the damper region may be a region including a damper in the target image.
Optionally, the damper region may be a local region in the target image; the damper region may also be the full volume region of the target image;
optionally, the damper area may be a rectangular area, such as a rectangle; the damper area may also be an area of other shapes, such as circular, oval.
It should be noted that, when the target image includes one damper, the damper area in the target image may be one; when the target image includes a plurality of dampers, the target image may include a plurality of damper regions, where each damper may correspond to each damper region.
The coordinate information is information that can identify coordinates of all positions covered by a certain damper area in the target image. Specifically, when a damper area is a rectangular area, the coordinate information may be coordinates of four vertices of the damper area; when a damper area is a circular area, the coordinate information may be the center coordinate and radius of the damper area.
It should be noted that, when the target image includes a damper area, the invention can obtain the coordinate information of the damper area output by the damper detection model; when the target image includes a plurality of damper areas, the invention can obtain the coordinate information of each damper area output by the damper detection model, for example, when the target image includes a first damper area and a second damper area, the invention can obtain the coordinate information of the first damper area and the coordinate information of the second damper area output by the damper detection model.
S104, intercepting a damper area image from the target image based on the coordinate information of the damper area;
specifically, after coordinate information of a certain damper area in the target image is obtained, an image covered by the damper area in the target image, that is, a damper area image, can be captured from the target image based on the coordinate information.
Optionally, when a plurality of damper areas exist in the target image, the invention intercepts each damper area image from the target image based on each coordinate information.
It can be understood that when the damper area is a local area in the target image, the damper area image can be a local image in the target image, at this time, the damper area image intercepted by the method is a smaller image relative to the target image, and when subsequent corrosion detection is performed, the method can be beneficial to reducing image data needing to be processed, improving the data processing efficiency and improving the resource utilization rate, and can prevent the possibility of occurrence of over-fitting through data enhancement, and improve the performance of corrosion classification of a subsequent model by increasing more characteristic information; when the damper area is the full-scale area in the target image, the damper area image can be the target image.
It should be noted that the present invention may use the existing image capturing technology to capture the image of the damper area from the target image based on the coordinate information.
And S105, detecting whether the damper in the damper area image is rusted or not by using the trained rust classification model.
The rust classification model can be a machine learning model used for classifying the damper in a certain damper area image and determining whether the damper in the certain damper area image is rusted.
Optionally, the trained rust classification model can be obtained by training a rust classification model to be trained by the invention;
alternatively, the rust classification model may be a target detection model obtained by machine learning a neural network model using a deep learning technique. The invention is not limited to the specific type of rust classification model, for example, the rust classification model can be ResNet34 model and other common classification models.
Optionally, when the rust classification model is a ResNet model, the network structure thereof may include:
an activation function, the activation function of a neuron defining a set of outputs given a set of inputs in the ResNet network;
a residual block, which eliminates the problem of gradient disappearance or gradient explosion of the ResNet network;
and the network structure outputs a result through the activation function.
It should be noted that the problem of gradient disappearance or gradient explosion and the problem of learning efficiency degradation can be solved through the ResNet network residual block, and the problems of overfitting caused by an excessive IoU threshold value and mismatching between a IoU threshold value and a threshold value in a trainer in the prior art can be solved by using a Cascade R-CNN model as a vibration damper detection model, so that the detection capability of the vibration damper is improved, and the inspection work efficiency is improved.
Optionally, the trained rust classification model includes a multilayer network; the first layer network in the trained rust classification model is a 7 x 7 convolutional layer with the use step length of 2, the second layer network is a pooling layer with the use step length of 2, the third layer to the thirty-fourth layer are residual block structures, the thirty-fifth layer is an average pooling layer, and the thirty-sixth layer is an output layer formed by full connection layers. The result can then be output by the softmax activation function.
Specifically, after a certain damper area image is obtained, the damper area image can be input into a trained corrosion classification model to obtain a classification result output by the corrosion classification model, so that whether the damper in the damper area image is corroded is determined. The classification result output by the corrosion classification model can be that the damper in the damper region image is a normal damper, a corrosion damper, a slight corrosion damper and/or a severe corrosion damper, etc.
It should be noted that, through steps S101, S102, S103, S104 and S105 in fig. 1, the present invention can perform the detection of the tarp corrosion on the target image without manual detection, so that human resources required by the manual detection can be reduced, and when there are many images to be detected by the tarp corrosion, the detection efficiency can be effectively improved. In addition, the inspection image has the characteristics of complex background, variable vibration damper angle and the like, when the inspection image is manually detected to determine whether the vibration damper is corroded, conditions such as missing detection and false detection are easy to occur, and the maintenance cost is increased. And, because the distance span of the transmission line is large, the natural environment is complex and various, and is influenced by external factors such as strong wind, rain, lightning stroke, etc., the transmission line is easy to have the problem of the corrosion of the damper, the manual inspection has the defects of large workload of workers, no guarantee on safety, low inspection efficiency, untimely failure solution, etc., and the real-time inspection requirement of the transmission line cannot be met by only using the manual inspection to perform the troubleshooting, therefore, the invention can acquire the inspection image by inspecting the transmission line through the unmanned aerial vehicle, and perform the detection of the corrosion of the damper on the inspection image through the method shown in fig. 1.
The method for detecting corrosion of a damper according to this embodiment can obtain a target image to be subjected to damper corrosion detection, detect whether a damper exists in the target image by using a trained damper detection model, if so, obtain coordinate information of a damper area in the target image output by the trained damper detection model, intercept out a damper area image from the target image based on the coordinate information of the damper area, and detect whether the damper exists in the damper area image by using a trained corrosion classification model. According to the invention, the anti-vibration hammer corrosion detection can be carried out on the target image through the anti-vibration hammer detection model and the corrosion classification model, manual detection is not needed, human resources required by manual detection are reduced, when more images are required to be subjected to the anti-vibration hammer corrosion detection, the detection efficiency can be effectively improved, the possible missing detection and false detection of the manual detection can be avoided, and the accuracy of the anti-vibration hammer corrosion detection is effectively ensured.
Based on the steps shown in fig. 1, the present embodiment proposes a second method for detecting corrosion of a damper, as shown in fig. 2. The method may further comprise the steps of:
s201, obtaining a plurality of damper images; wherein each damper image is a patrol inspection image containing a damper;
specifically, the damper image may be an image containing one or more dampers.
Optionally, the multiple damper images may be inspection images selected by the worker from the inspection image set. Specifically, the inspection image set may include a plurality of inspection images, and the inspection images may be images for recording the appearance shape, the working state, and/or the surrounding environment of the power transmission line. Specifically, the inspection image may be obtained by shooting the damper or other components on the power transmission line by using a camera, a remote control unmanned aerial vehicle and/or a remote control camera device in the inspection process of the power transmission line by a worker, or obtained by shooting the damper or other components on the power transmission line by a camera device arranged around the power transmission line.
The number of images of the damper image is not limited in the present invention.
S202, preprocessing each damper image according to a predefined image preprocessing mode;
specifically, the present invention may be configured to obtain a plurality of damper images, and then perform preprocessing on each damper image using an image preprocessing method.
Optionally, the image preprocessing method at least includes: image quality evaluation, image data cleaning and/or damper deformation defect image data annotation.
Optionally, the image preprocessing mode may further include image normalization processing and amplification of the damper deformation image data by using translation, mirroring, rotation, occlusion, and random noise addition.
S203, obtaining each preprocessed damper image;
specifically, the present invention may obtain each preprocessed damper image after preprocessing each damper image.
S204, using at least part of the preprocessed damper images to create a first image set of a damper detection model to be trained;
in particular, the present invention may use at least a portion of the preprocessed damper image and other images that do not contain dampers to create the first set of images.
Optionally, the first set of images may include a training set, a validation set, and/or a test set used in training the damper detection model. Wherein, the training set can comprise a plurality of positive samples and negative samples for training the damper detection model. The positive sample may be an image containing the damper, and the negative sample may be an image not containing the damper.
The training method of the vibration damper detection model used in the present invention is not limited, and examples include supervised training, semi-supervised training, and unsupervised training.
Alternatively, the positive samples in the training set may be images with the damper marked out using a border.
The precision evaluation can be carried out on the damper detection model after each training by using the test set, whether the output detection accuracy rate meets the requirement or not is evaluated, if the output detection accuracy rate meets the requirement, the damper detection model can be determined to be well trained, the damper detection model can be the trained damper detection model, the detection work can be put into operation, and otherwise, the model retraining and parameter optimization can be carried out on the damper detection model.
S205, training the damper detection model to be trained by using the first image set to obtain the trained damper detection model.
Specifically, the method can use the training set, the verification set and/or the test set in the first image set to perform machine learning on the damper detection model until the trained damper detection model is obtained.
In the process of training the damper detection model, different Intersection-over-unity (IoU) thresholds can be adopted to divide positive and negative samples, so that the detector at each stage is focused on detecting the predicted value of the IOU within a certain range, and the output IOU is generally larger than the input IOU; furthermore, the invention can adopt a sequential multi-stage extension detection, and utilize the output of the previous stage to train the next stage, and use a higher IoU threshold value after the next stage, thereby generating a higher quality bndbobox.
It should be noted that, in the present invention, the trained damper detection model can be obtained by training the damper detection model to be trained through the above steps S201, S202, S203, S204, and S205. It can be understood that the trained damper detection model can be regarded as the damper detection model to be trained, the training is continued, the damper detection model is optimized, and the detection performance of the damper detection model on the damper is further improved.
Optionally, the trained damper detection model may use ResNet101 as a backbone network characteristic extraction network. At this time, when a trained damper detection model receives a certain preprocessed damper image, the ResNet101 can be used to extract the damper image features in the preprocessed damper image, and generate a feature map including depth convolution features.
Optionally, in the second method for detecting corrosion of a damper, the method may further include the following steps:
s301, inputting each preprocessed damper image into a trained damper detection model;
specifically, after a trained damper detection model is obtained, the acquired preprocessed damper images can be input into the trained damper detection model.
S302, obtaining coordinate information of the damper areas in the damper images respectively output by the trained damper detection model;
specifically, the invention can obtain the coordinate information of the damper area in each damper image respectively output by the trained damper detection model after inputting each preprocessed damper image into the trained damper detection model. For example, the present invention may obtain coordinate information of the hammer zone in the first damper image output by the trained damper detection model and obtain coordinate information of the hammer zone in the second damper image output by the trained damper detection model after inputting the preprocessed first damper image and second damper image into the trained damper detection model.
Optionally, after the preprocessed damper images are input into the trained damper detection model, the trained damper detection model can also output the damper images in which the damper areas are marked by using the frames. For example, after the preprocessed first damper image and the preprocessed second damper image are input into the trained damper detection model, the first damper image output by the trained damper detection model and marked with the damper area by using the frame can be obtained, and the second damper image output by the trained damper detection model and marked with the damper area by using the frame can be obtained.
S303, intercepting damper area images from the damper images respectively based on the coordinate information of the damper areas in the damper images;
specifically, the invention can respectively intercept the images of the damper areas from the images of the dampers after obtaining the coordinate information of the damper areas in the images of the dampers. For example, the present invention may intercept the image of the damper region from the first damper image and the image of the damper region from the second damper image after obtaining the coordinate information of the damper region in the first damper image and the coordinate information of the damper region in the second damper image.
S304, preprocessing each damper area image cut out from each damper image according to an image preprocessing mode;
specifically, the image preprocessing method here may be the same as the image preprocessing method in step S202.
S305, obtaining preprocessed anti-vibration hammer area images;
specifically, the present invention may obtain each preprocessed damper area image after preprocessing each damper area image.
S306, using the preprocessed vibration damper area images to create a second image set used for training the rust classification model to be trained;
specifically, the present invention may be implemented by a worker to mark whether the damper included in the preprocessed image of each damper area is rusted or not, so as to create a positive sample and a negative sample, thereby creating the second image set.
In particular, the second set of images may include a training set, a validation set, and a test set. Wherein the training set may comprise a plurality of positive and negative examples for training the rust classification model. Wherein, the positive sample can be an image containing a rust damper, and the negative sample can be an image of a normal damper;
s307, training the to-be-trained rust classification model by using the second image set to obtain the trained rust classification model.
Specifically, the invention can use the training set, the verification set and/or the test set in the second image set to perform machine learning on the rust classification model until a trained rust classification model is obtained.
It should be noted that the training mode of the rust classification model adopted by the invention is not limited, and includes supervised training, semi-supervised training, unsupervised training and the like.
The invention can utilize a test set to perform precision evaluation on the rust classification model after each training, evaluate whether the output classification accuracy and the recall rate meet the inspection requirement, and can determine that the rust classification model is trained and can be put into classification work if the output classification accuracy and the recall rate meet the inspection requirement, or else, can retrain the model and optimize parameters.
It should also be noted that the invention can utilize the trained damper detection model to train the rust classification model, thereby improving the training efficiency and the training effect of the rust classification model.
The damper corrosion detection method provided by the embodiment can be used for training the damper detection model and the corrosion classification model respectively, and in the training process, the trained damper detection model is used for training the corrosion classification model, so that the training efficiency and the training effect of the corrosion classification model are improved, and the detection effect of the damper corrosion detection is improved.
Corresponding to the steps of fig. 1, the present embodiment proposes a first kind of corrosion detection device for damper, as shown in fig. 2. The apparatus may include: a first obtaining unit 101, a first detecting unit 102, a second obtaining unit 103, a first truncating unit 104, and a second detecting unit 105; wherein:
a first obtaining unit 101 for obtaining a target image to be subjected to the detection of the corrosion of the damper;
the first detection unit 102 is configured to detect whether a damper exists in the target image by using the trained damper detection model, and if so, trigger the second obtaining unit 103;
a second obtaining unit 103, configured to obtain coordinate information of a damper area in a target image output by the trained damper detection model;
a first intercepting unit 104, configured to intercept a damper area image from the target image based on coordinate information of the damper area;
and the second detection unit 105 is used for detecting whether the damper in the damper area image is rusted or not by using the trained rust classification model.
It should be noted that specific processing procedures of the first obtaining unit 101, the first detecting unit 102, the second obtaining unit 103, the first truncating unit 104 and the second detecting unit 105 and technical effects brought by the processing procedures can refer to steps S101, S102, S103, S104 and S105 in fig. 1, respectively, and relevant descriptions are not repeated here.
Optionally, the trained rust classification model includes a multilayer network; the first layer network in the trained rust classification model is a 7 x 7 convolutional layer with the use step length of 2, the second layer network is a pooling layer with the use step length of 2, the third layer to the thirty-fourth layer are residual block structures, the thirty-fifth layer is an average pooling layer, and the thirty-sixth layer is an output layer formed by full connection layers.
Optionally, the apparatus further comprises: the device comprises a third obtaining unit, a first preprocessing unit, a fourth obtaining unit, a first creating unit, a first training unit and a fifth obtaining unit; wherein:
a third obtaining unit configured to obtain a plurality of damper images; wherein each damper image is a patrol inspection image containing a damper;
the first preprocessing unit is used for respectively preprocessing each damper image according to a predefined image preprocessing mode;
a fourth obtaining unit configured to obtain each preprocessed damper image;
a first creating unit, configured to create a first image set for a damper detection model to be trained, using at least part of the preprocessed damper images;
the first training unit is used for training the damper detection model to be trained by using the first image set;
and the fifth obtaining unit is used for obtaining the trained damper detection model.
Optionally, the apparatus further comprises: the device comprises a first input unit, a sixth obtaining unit, a second intercepting unit, a second preprocessing unit, a seventh obtaining unit, a second creating unit, a second training unit and an eighth obtaining unit; wherein:
the first input unit is used for respectively inputting the preprocessed damper images to the trained damper detection model;
a sixth obtaining unit, configured to obtain coordinate information of the damper area in each damper image output by the trained damper detection model;
the second intercepting unit is used for respectively intercepting the images of the damper areas from the images of the dampers on the basis of the coordinate information of the damper areas in the images of the dampers;
the second preprocessing unit is used for respectively preprocessing each damper area image cut out from each damper image according to an image preprocessing mode;
a seventh obtaining unit, configured to obtain preprocessed images of each damper area;
the second creating unit is used for creating a second image set used for training the rust classification model to be trained by using the preprocessed vibration damper area images;
the second training unit is used for training the rust classification model to be trained by using the second image set;
and the eighth obtaining unit is used for obtaining the trained rust classification model.
The utility model provides a damper corrosion detection device can carry out damper corrosion detection to the target image through damper detection model and corrosion classification model, and need not to carry out artifical the detection, reduces artifical required manpower resources that detects, when the image that the damper corrosion detection was carried out to needs is more, can effectively improve detection efficiency to can avoid artifical leak hunting and the false retrieval that detects the probably appearance, effectively ensure the accuracy that damper corrosion detected.
As shown in fig. 3, the electronic device 300 may include a processor 301, a memory 302, a communication interface 303, an input unit 304, an output unit 305, and a communication bus 306. Wherein the processor 301 and the memory 302 are connected to each other by a communication bus 306. A communication interface 303, an input unit 304 and an output unit 305 are also connected to the communication bus 306.
The communication interface 303 may be an interface of a communication module, such as an interface of a GSM module. The communication interface 303 may be used to obtain data or instructions sent by other devices. The communication interface 303 is also used to transmit data or instructions to other devices.
In the embodiment of the present invention, the processor 301 may be a Central Processing Unit (CPU), an application-specific integrated circuit (ASIC), a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf programmable gate array (FPGA) or other programmable logic devices.
In one possible implementation, the memory 302 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, application programs required for at least one function (such as image detection and image classification functions), and the like; the storage data area may store data created during use of the computer, such as user data, user access data, image data, and the like.
Further, the memory 302 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device or other volatile solid state storage device.
The processor 301 may call a program stored in the memory 302, and in particular, the processor 301 may perform any of the above methods for detecting the tarnish of the damper.
The memory 302 is used for storing one or more programs, the program may include program codes, the program codes include computer operation instructions, and in the embodiment of the present invention, at least the program for realizing the following functions is stored in the memory 302:
obtaining a target image to be subjected to the corrosion detection of the damper;
detecting whether the damper exists in the target image by using the trained damper detection model, and if so, obtaining coordinate information of a damper area in the target image output by the trained damper detection model;
intercepting a damper area image from a target image based on coordinate information of the damper area;
and detecting whether the damper in the damper region image is rusted or not by using the trained rust classification model.
In one possible implementation, the electronic device 300 may include: one or more processors 301;
one or more processors 301;
a storage device having one or more programs stored thereon;
the one or more programs, when executed by the one or more processors 301, cause the one or more processors 301 to implement any of the damper rust detection methods described above.
The present invention may further include an input unit 304, and the input unit 304 may include at least one of a touch sensing unit sensing a touch event on the touch display panel, a keyboard, a mouse, a camera, a microphone, and the like.
The output unit 305 may include: at least one of a display, a speaker, a vibration mechanism, a light, and the like. The display may comprise a display panel, such as a touch display panel or the like. In one possible case, the Display panel may be configured in the form of a Liquid Crystal Display (LCD), an Organic Light-Emitting Diode (OLED), or the like. The vibration mechanism is operable to displace the electronic device 300, and in one possible implementation, the vibration mechanism includes a motor and an eccentric vibrator, and the motor drives the eccentric vibrator to rotate so as to generate vibration. The brightness and/or color of the lamp can be adjusted, in a possible implementation manner, different information can be embodied through at least one of the on-off, brightness and color of the lamp, for example, the alarm information can be embodied through red light emitted by the lamp.
Of course, the structure of the electronic device 300 shown in fig. 3 does not constitute a limitation of the electronic device in the embodiment of the present invention, and in practical applications, the electronic device may include more or less components than those shown in fig. 3, or some components may be combined.
The embodiment of the invention provides a computer readable medium, wherein a computer program is stored on the computer readable medium, and when the computer program is executed by a processor, the computer readable medium implements any one of the above-mentioned methods for detecting corrosion of a damper.
The embodiment of the invention provides a processor, which is used for running a program, wherein when the program runs, any one of the vibration damper corrosion detection methods is realized.
The present invention also provides a computer program product which, when executed on data processing apparatus, causes the data processing apparatus to implement any of the above-described damper rust detection methods.
In addition, the electronic device, the processor, the computer readable medium, or the computer program product provided in the foregoing embodiments of the present invention may be all used for executing the corresponding methods provided above, and therefore, the beneficial effects achieved by the electronic device, the processor, the computer readable medium, or the computer program product may refer to the beneficial effects in the corresponding methods provided above, and are not described herein again.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). The memory is an example of a computer-readable medium.
Computer-readable media, which include both non-transitory and non-transitory, removable and non-removable media, may implement the information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and the technical principles applied, and is not intended to limit the present invention. Various modifications and alterations to this invention will become apparent to those skilled in the art. The scope of the present invention is not limited to the specific combinations of the above-described features, and may also include other features formed by arbitrary combinations of the above-described features or their equivalents without departing from the spirit of the present invention. For example, the above features and (but not limited to) features having similar functions disclosed in the present invention are mutually replaced to form the technical solution.
Claims (10)
1. A method for detecting corrosion of a damper is characterized by comprising the following steps:
obtaining a target image to be subjected to the corrosion detection of the damper;
detecting whether the damper exists in the target image or not by using the trained damper detection model, and if so, obtaining coordinate information of a damper area in the target image output by the trained damper detection model;
intercepting a damper area image from the target image based on the coordinate information of the damper area;
and detecting whether the damper in the damper region image is rusted or not by using the trained rust classification model.
2. The method for detecting the corrosion of the damper according to claim 1, wherein the trained corrosion classification model comprises a plurality of layers of networks; the trained rust classification model comprises a first layer network, a second layer network, a third layer network, a thirty-fourth layer network, an average pooling layer and a thirty-sixth layer network, wherein the first layer network in the trained rust classification model is a 7 x 7 convolutional layer with the use step size of 2, the second layer network is a pooling layer with the use step size of 2, the third layer to the thirty-fourth layer are residual block structures, the thirty-fifth layer is an average pooling layer, and the thirty-sixth layer is an output layer formed by all-connected layers.
3. The method of claim 1, further comprising:
obtaining a plurality of damper images; wherein each damper image is a patrol inspection image containing a damper;
respectively preprocessing each damper image according to a predefined image preprocessing mode to obtain each damper image after preprocessing;
using the at least partially preprocessed damper image, creating a first image set for the damper detection model to be trained;
and training the damper detection model to be trained by using the first image set to obtain the trained damper detection model.
4. The damper rust detection method according to claim 3, further comprising:
inputting each preprocessed damper image into the trained damper detection model; obtaining coordinate information of a damper area in each damper image respectively output by the trained damper detection model;
intercepting damper area images from each damper image respectively based on coordinate information of the damper area in each damper image;
preprocessing each damper area image cut out from each damper image according to the image preprocessing mode to obtain each damper area image after preprocessing;
creating a second image set for training the rust classification model to be trained by using the preprocessed each damper area image;
and training the rust classification model to be trained by using the second image set to obtain the trained rust classification model.
5. The utility model provides a damper corrosion detection device which characterized in that includes: the device comprises a first obtaining unit, a first detecting unit, a second obtaining unit, a first intercepting unit and a second detecting unit; wherein:
the first obtaining unit is used for obtaining a target image to be subjected to the corrosion detection of the damper;
the first detection unit is used for detecting whether the damper exists in the target image by utilizing a trained damper detection model, and if so, the second obtaining unit is triggered;
the second obtaining unit is configured to obtain coordinate information of a damper area in the target image output by the trained damper detection model;
the first intercepting unit is used for intercepting a damper area image from the target image based on the coordinate information of the damper area;
and the second detection unit is used for detecting whether the damper in the damper area image is rusted or not by using the trained rust classification model.
6. The apparatus of claim 5, wherein the trained corrosion classification model comprises a multi-layer network; the trained rust classification model comprises a first layer network, a second layer network, a third layer network, a thirty-fourth layer network, an average pooling layer and a thirty-sixth layer network, wherein the first layer network in the trained rust classification model is a 7 x 7 convolutional layer with the use step size of 2, the second layer network is a pooling layer with the use step size of 2, the third layer to the thirty-fourth layer are residual block structures, the thirty-fifth layer is an average pooling layer, and the thirty-sixth layer is an output layer formed by all-connected layers.
7. The apparatus of claim 5, further comprising: the device comprises a third obtaining unit, a first preprocessing unit, a fourth obtaining unit, a first creating unit, a first training unit and a fifth obtaining unit; wherein:
the third obtaining unit is used for obtaining a plurality of damper images; wherein each damper image is a patrol inspection image containing a damper;
the first preprocessing unit is used for preprocessing each damper image according to a predefined image preprocessing mode;
the fourth obtaining unit is configured to obtain each preprocessed damper image;
the first creating unit is used for creating a first image set of the damper detection model to be trained by using at least part of the preprocessed damper images;
the first training unit is used for training the damper detection model to be trained by using the first image set;
and the fifth obtaining unit is used for obtaining the trained damper detection model.
8. The apparatus of claim 7, further comprising: the device comprises a first input unit, a sixth obtaining unit, a second intercepting unit, a second preprocessing unit, a seventh obtaining unit, a second creating unit, a second training unit and an eighth obtaining unit; wherein:
the first input unit is used for respectively inputting the preprocessed damper images to the trained damper detection model;
the sixth obtaining unit is configured to obtain coordinate information of a damper area in each damper image that is output by the trained damper detection model;
the second intercepting unit is used for intercepting damper area images from the damper images respectively based on the coordinate information of the damper areas in the damper images;
the second preprocessing unit is configured to respectively preprocess each of the damper region images captured from each of the damper images according to the image preprocessing method;
the seventh obtaining unit is configured to obtain preprocessed each damper region image;
the second creating unit is configured to create a second image set used for training the rust classification model to be trained, using the preprocessed each of the damper region images;
the second training unit is used for training the rust classification model to be trained by using the second image set;
and the eighth obtaining unit is used for obtaining the trained corrosion classification model.
9. A computer readable medium having stored thereon a computer program, wherein the program when executed by a processor implements a damper rust detection method as claimed in any one of claims 1 to 4.
10. A processor for running a program, wherein the program when run implements the damper rust detection method of any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111592340.4A CN114299475A (en) | 2021-12-23 | 2021-12-23 | Method for detecting corrosion of damper and related equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111592340.4A CN114299475A (en) | 2021-12-23 | 2021-12-23 | Method for detecting corrosion of damper and related equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114299475A true CN114299475A (en) | 2022-04-08 |
Family
ID=80970555
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111592340.4A Pending CN114299475A (en) | 2021-12-23 | 2021-12-23 | Method for detecting corrosion of damper and related equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114299475A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115318760A (en) * | 2022-07-29 | 2022-11-11 | 武汉理工大学 | Unmanned aerial vehicle laser cleaning method and system for power transmission tower |
-
2021
- 2021-12-23 CN CN202111592340.4A patent/CN114299475A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115318760A (en) * | 2022-07-29 | 2022-11-11 | 武汉理工大学 | Unmanned aerial vehicle laser cleaning method and system for power transmission tower |
| CN115318760B (en) * | 2022-07-29 | 2024-04-16 | 武汉理工大学 | Unmanned aerial vehicle laser cleaning method and system for power transmission tower |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11488294B2 (en) | Method for detecting display screen quality, apparatus, electronic device and storage medium | |
| Lestari et al. | Fire hotspots detection system on CCTV videos using you only look once (YOLO) method and tiny YOLO model for high buildings evacuation | |
| CN113283344A (en) | Mining conveying belt deviation detection method based on semantic segmentation network | |
| CN113947188A (en) | Training method of target detection network and vehicle detection method | |
| CN115240117A (en) | Helmet wearing detection method in construction site construction scene | |
| CN114299475A (en) | Method for detecting corrosion of damper and related equipment | |
| Li et al. | An improved YOLOv3 for foreign objects detection of transmission lines | |
| CN113487610B (en) | Herpes image recognition method and device, computer equipment and storage medium | |
| CN112784675B (en) | Target detection method and device, storage medium and terminal | |
| CN112001300A (en) | Building monitoring method and device based on cross entropy according to position and electronic equipment | |
| CN116563762A (en) | Fire detection method, system, medium, equipment and terminal for oil and gas station | |
| Choi et al. | Deep‐learning‐based nuclear power plant fault detection using remote light‐emitting diode array data transmission | |
| CN115187831A (en) | Model training and smoke detection method and device, electronic equipment and storage medium | |
| CN115147716A (en) | Forest area felling detection method and device, electronic equipment and storage medium | |
| CN111340149B (en) | Excavator real-time detection method and system based on background difference method and deep learning | |
| CN114241189A (en) | Ship black smoke identification method based on deep learning | |
| CN113780136A (en) | VOCs gas leakage detection method, system and equipment based on space-time texture recognition | |
| Mysiuk et al. | Detection of Surface Defects Inside Concrete Pipelines Using Trained Model on JetRacer Kit | |
| CN112101134A (en) | Object detection method and device, electronic device and storage medium | |
| CN110852174A (en) | Early smoke detection method based on video monitoring | |
| CN114639037B (en) | Method for determining vehicle saturation of high-speed service area and electronic equipment | |
| CN114581890B (en) | Method and device for determining lane line, electronic equipment and storage medium | |
| Ergasheva et al. | Advancing Maritime Safety: Early Detection of Ship Fires through Computer Vision, Deep Learning Approaches, and Histogram Equalization Techniques | |
| CN112949526A (en) | Face detection method and device | |
| Friederich et al. | Security Fence Inspection at Airports Using Object Detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |