CN115880624A - Identification method for pull rod play fault on truck - Google Patents
Identification method for pull rod play fault on truck Download PDFInfo
- Publication number
- CN115880624A CN115880624A CN202211406790.4A CN202211406790A CN115880624A CN 115880624 A CN115880624 A CN 115880624A CN 202211406790 A CN202211406790 A CN 202211406790A CN 115880624 A CN115880624 A CN 115880624A
- Authority
- CN
- China
- Prior art keywords
- pull rod
- fault
- upper pull
- probability value
- bracket
- 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 application provides a method for identifying a pull rod fleeing fault on a truck, which comprises the following steps: acquiring an image to be detected; positioning the position of a component with the fault caused by the play of the upper pull rod in the image to be detected, and predicting the probability value of the fault caused by the play of the upper pull rod; the upper pull rod fault-out component comprises an upper pull rod and an upper pull rod bracket; if the fault probability value of the upper pull rod jumping-out fault is greater than a preset first fault probability value, judging that the upper pull rod jumping-out fault occurs; and if the fault probability value of the upper pull rod jumping fault is smaller than the preset first fault probability value and larger than the preset second fault probability value, judging whether the upper pull rod jumping fault occurs or not according to the position relation between the central position of the upper pull rod and the upper edge and the lower edge of the upper pull rod bracket. By the method, the detection speed and the detection accuracy rate of the fault of the pull rod shifting on the truck can be improved.
Description
Technical Field
The invention belongs to the technical field of fault detection of running states of rail wagons, and particularly relates to a method for identifying a fault of a pull rod play on a wagon.
Background
At present, in order to ensure the running safety of a railway Freight car, a railway Freight line is provided with a TFDS (railway of moving Freight car Detection System) device, the System utilizes a plurality of groups of high-speed industrial cameras arranged on a rail edge to carry out dynamic snapshot on running train vehicles, collected images are transmitted to a train inspection center through a network, a vehicle inspector uses a software platform through the TFDS System, fault judgment and analysis are carried out in a mode of manually browsing pictures, the vehicle inspection efficiency and the operation quality are greatly improved, and the manpower, material resources and financial resources are saved.
The heavy haul railway has the characteristics of large transportation capacity, long marshalling, intensive vehicle passing, short stop time, huge overhaul workload and the like, and the overhaul operation faces the challenges of short operation time, huge workload and high overhaul standard. In the practical use of the system, with the continuous increase of the transportation volume, the speed and the number of trucks can be continuously increased, and meanwhile, the number of fault vehicles can be continuously increased in the process of using new vehicles to become old vehicles. At present, the manual operation mode is completely adopted, so that the workload and the pressure for ensuring the safety of TFDS car inspectors are multiplied.
The upper pull rod is used as an important part for braking the truck, plays a vital role in the running safety of the truck, and an intelligent identification method for detecting the fault image of the truck upper pull rod jumping is urgently needed at present in order to ensure the safety of the truck in the running process.
Disclosure of Invention
In order to solve the problems, the method for identifying the pull rod shifting fault on the truck is provided, and the detection speed and the detection accuracy of the pull rod shifting fault on the truck are improved.
In a first aspect of the present application, a method for identifying a pull rod play fault on a truck is provided, which includes:
acquiring an image to be detected;
positioning the position of a part with the upper pull rod jumping out of the fault in the image to be detected, and predicting the fault probability value of the upper pull rod jumping out of the fault; the upper pull rod jumping fault component comprises an upper pull rod and an upper pull rod bracket;
if the fault probability value of the upper pull rod jumping fault is greater than a preset first fault probability value, judging that the upper pull rod jumping fault occurs according to the judgment result;
and if the fault probability value of the upper pull rod jumping fault is smaller than the preset first fault probability value and larger than the preset second fault probability value, judging whether the upper pull rod jumping fault occurs or not according to the position relation between the central position of the upper pull rod and the upper edge and the lower edge of the upper pull rod bracket.
In some embodiments, the position of the upper pull rod out-of-fault part in the image to be detected is located by adopting an upper pull rod out-of-fault location model, and the fault probability value of the upper pull rod out-of-fault is predicted.
In some embodiments, the determining whether the upper pull rod has a fleeing fault according to the position relationship between the central position of the upper pull rod and the upper and lower edges of the upper pull rod bracket includes:
if the central position of the upper pull rod is positioned between the upper edge and the lower edge of the upper pull rod bracket, judging that the upper pull rod has no play fault;
and if the central position of the upper pull rod is not positioned between the upper edge and the lower edge of the upper pull rod bracket, judging that the upper pull rod has a play fault.
In some embodiments, the determination of the center position of the upper tie bar and the upper and lower edges of the upper tie bar bracket comprises:
positioning coordinates of the upper pull rod and the upper pull rod supporting groove in the image to be detected;
and respectively calculating the central position of the upper pull rod and the upper edge and the lower edge of the upper pull rod bracket according to the position coordinates of the upper pull rod and the upper pull rod bracket in the image to be detected.
In some embodiments, the position coordinates of the upper pull rod and the upper pull rod bracket in the image to be detected are positioned by using a preset upper pull rod and upper pull rod bracket positioning model.
In some embodiments, the upper tie rod channeling fault localization model deep learning network model employs a new yolo-v3 network model body structure that employs only one output layer.
In some embodiments, the aspect ratio of the bounding box of the upper tie rod channeling fault localization model is determined from the aspect ratio of the upper tie rod samples.
In some embodiments, if the judgment result is that the upper pull rod fleeing fault occurs, uploading the upper pull rod fleeing fault information and giving an alarm.
In some embodiments, the preset first failure probability value is 0.9 and the preset second failure probability value is 0.5.
In a second aspect of the present application, there is provided a device for identifying a pull rod fleeing fault on a truck, comprising:
the image acquisition module is used for acquiring an image to be detected;
the target detection module is used for positioning the position of a part with the fault of the upper pull rod in the image to be detected and predicting the fault probability value of the fault of the upper pull rod; the upper pull rod fault-out component comprises an upper pull rod and an upper pull rod bracket;
the judging module is used for judging that the upper pull rod fleeing fault occurs if the fault probability value of the upper pull rod fleeing fault is greater than a preset first fault probability value; and
and if the fault probability value of the upper pull rod jumping fault is smaller than the preset first fault probability value and larger than the preset second fault probability value, judging whether the upper pull rod jumping fault occurs or not according to the position relation between the central position of the upper pull rod and the upper edge and the lower edge of the upper pull rod bracket.
In a third aspect of the present application, a storage medium storing a computer program executable by one or more processors is provided for implementing a method for identifying a pull-out fault in a truck upper tie rod as described above.
In a fourth aspect of the present application, an electronic device is provided, which includes a memory and a processor, where the memory stores a computer program, and the memory and the processor are communicatively connected to each other, and when the computer program is executed by the processor, the method for identifying a pull rod fleeing fault on a truck is implemented.
The technical effects of the application are as follows: 1. according to the technical scheme, the prediction speed of the upper pull rod fleeing fault positioning model is improved by utilizing the deep learning technology and optimizing the upper pull rod fleeing fault positioning model. And then detecting the upper pull rod fleeing fault part by using the trained upper pull rod fleeing fault positioning model, and predicting the fault probability value of the upper pull rod fleeing fault. Directly judging that the detected probability value of the fault component is greater than a preset first fault probability value to send an upper pull rod to break out of the fault; and if the fault probability value is less than the preset first fault probability value and greater than the preset second fault probability value, further positioning the position information of the upper pull rod bracket and the upper pull rod, and judging whether the upper pull rod is separated from the upper pull rod through the position relation between the upper pull rod bracket and the upper pull rod, so that the generation of false alarm can be effectively avoided. Thereby can be faster more accurate realization go up pull rod play fault detection to railway freight car, and then can alleviate artificial working strength greatly, shorten and overhaul the activity duration, improve the work efficiency who overhauls, reduce the probability of lou examining, reduce human cost and administrative cost to guarantee the safe operation of freight car.
2. The size ratio of a boundary frame (anchor box) in the upper pull rod play fault positioning model is fixed to be the size ratio of an upper pull rod play fault component in an actual picture, and the convergence speed of the upper pull rod play fault positioning model during training is improved. Meanwhile, the fault identification effect is increased along with the increase of the image data volume, the fault samples are gradually increased, the detection effect can be effectively improved, the calculated amount is small after the network model is improved, and the operation speed is high and accurate.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings used in the embodiments or the prior art descriptions will be briefly described below, it is obvious that the drawings in the following description are only the embodiments of the present application, 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 of a method for identifying a pull rod play fault on a truck according to an embodiment of the present disclosure;
fig. 2 is a flowchart of a method for identifying a pull rod play fault on a truck according to an embodiment of the present disclosure;
FIG. 3 is a flow chart of a training process of the upper pull rod play fault location model;
FIG. 4 is a schematic structural diagram of a main body of a fault location model of the upper pull rod jumping;
FIG. 5 is a flow chart of a training process for an upper tie rod and upper tie rod bracket positioning model;
FIG. 6 is an image to be detected;
FIG. 7 is a schematic diagram showing the position of a component with a fault caused by the play of an upper pull rod in an image to be detected;
FIG. 8 is a schematic diagram of the coordinates of the upper pull rod and the upper pull rod supporting groove in the image to be detected;
fig. 9 is a schematic structural diagram of a device for identifying a play fault of a pull rod on a truck according to a third embodiment of the present application;
fig. 10 is a connection block diagram of an electronic device.
Detailed Description
In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be described in further detail with reference to the attached drawings, the described embodiments should not be considered as limiting the present application, and all other embodiments obtained by a person of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.
In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict.
The following description will be added if a similar description of "first \ second \ third" appears in the application file, where the terms "first \ second \ third" merely distinguish similar objects and do not represent a specific ordering with respect to the objects, and it should be understood that "first \ second \ third" may be interchanged with a specific order or sequence as permitted, so that the embodiments of the application described herein can be implemented in an order other than that illustrated or described herein.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of the present application only and is not intended to be limiting of the application.
Example one
Fig. 1 and fig. 2 are a flowchart of a method for identifying a pull rod play failure in a truck according to an embodiment, and as shown in fig. 1 and fig. 2, the method according to the embodiment includes:
s1, acquiring an image to be detected.
S2, positioning a part with a fault caused by the play of the upper pull rod in the image to be detected, and predicting the probability value of the fault caused by the play of the upper pull rod; the upper pull rod fault-out component comprises an upper pull rod and an upper pull rod bracket. In the embodiment, the position of the upper pull rod jumping fault part in the image to be detected is positioned by adopting an upper pull rod jumping fault positioning model, and the fault probability value of the upper pull rod jumping fault is predicted. And the upper pull rod in the image to be detected breaks out of the position of the fault part.
The training process of the upper pull rod play fault positioning model is shown in fig. 3, and comprises the following steps: the method comprises the steps of collecting a large number of original images with faults fleeed by an upper pull rod, utilizing image enhancement to expand sample data to form an image data set A, then dividing the image data set A into a training data set and a testing data set, wherein the training data set accounts for 9/10 of the whole image data set, the testing data set accounts for 1/10 of the whole image data set, utilizing a VOC data set format to finish labeling the training data set, then training the fault location model with the upper pull rod fleeing through SGD (random gradient descent), the initial learning rate is 0.001, training 40000 times, adjusting the learning rate to 0.0001, training 5000 times, adjusting the learning rate to 0.00001, training 5000 times, storing an intermediate network model parameter file every 1000 times of training, testing the trained network model parameter file through the testing data set, and selecting an optimal network model related parameter file.
The main structure of the upper pull rod play-out fault positioning model adopts a yolo-v3 main structure, and considering that the upper pull rod is relatively large in size and fixed in size in a picture, two output layers of 26 multiplied by 255 and 52 multiplied by 255 in a yolo-v3 network structure are deleted, only the output layers of 13 multiplied by 255 are reserved, the network structure is simplified, and the model detection speed is improved. The default bounding box (anchor box) size for yolo-v3 was obtained using the k-means clustering algorithm based on the ImageNet training set. In the upper pull rod detection process, the marking frame of the upper pull rod is a rectangle with a large length-width ratio, and if a universal anchor box aspect ratio is adopted, the accuracy of the final training model is affected, so that the embodiment generates a corresponding anchor box aspect ratio according to the size of the upper pull rod sample, replaces the original default value, and enables the upper pull rod to be easier to converge when the fault positioning model is shifted out. The main structure of the upper pull rod channeling fault location model is shown in fig. 4, the size of an input image of the network model in fig. 4 is 416 × 416 × 3, and then the size of the input image is Darknet53 as a main network, wherein the main network comprises a DBL feature extraction block, and the DBL consists of a convolutional layer (cinv), a Batch Normalization (BN) and a Leaky relu activation function. The DBL feature extraction block is followed by 5 layers of resnet convolutional layers, the Darknet53 backbone network is followed by 6 DBL feature extraction layers, then one convolutional layer, and finally a 13 x 255 feature map is used for subsequent classification detection.
And S3, if the fault probability value of the upper pull rod jumping out of the fault is greater than a preset first fault probability value, judging that the upper pull rod jumping out of the fault occurs.
And S4, if the fault probability value of the upper pull rod jumping out of fault is smaller than the preset first fault probability value and larger than the preset second fault probability value, judging whether the upper pull rod jumping out of fault occurs or not according to the position relation between the central position of the upper pull rod and the upper edge and the lower edge of the upper pull rod bracket.
In some embodiments, the first failure probability value is 0.9 and the second failure probability value is 0.5.
In some embodiments, the determination of the center position of the upper tie bar and the upper and lower edges of the upper tie bar bracket comprises:
positioning coordinates of the upper pull rod and the upper pull rod supporting groove in the image to be detected; in some embodiments, the position coordinates of the upper pull rod and the upper pull rod bracket in the image to be detected are positioned by using a preset upper pull rod and upper pull rod bracket positioning model. Wherein, the training process of the upper pull rod and the upper pull rod bracket positioning model is as shown in fig. 5, and comprises the following steps: a large number of original images of an upper pull rod and an upper pull rod bracket are collected, sample data are augmented by using image enhancement to form an image data set B, then the image data set B is divided into a training data set and a testing data set, the training data set accounts for 9/10 of the whole image data set, the testing data set accounts for 1/10 of the whole image data set, marking of the upper pull rod and an upper pull rod bracket in the training data set is completed by using a VOC data set format, then training of an upper pull rod and an upper pull rod bracket positioning model is realized through SGD (random gradient descent), the initial learning rate is 0.001, training is 40000 times, the learning rate is adjusted to 0.0001 and 5000 times, the learning rate is adjusted to 0.00001 and is 5000 times, a middle network model parameter file is stored every 1000 times of training, the trained network model parameter file is tested by using the testing data set, and an optimal network model related parameter file is selected.
And respectively calculating the central position of the upper pull rod and the upper edge and the lower edge of the upper pull rod bracket according to the position coordinates of the upper pull rod and the upper pull rod bracket in the image to be detected.
In some embodiments, the determining whether a failure of the upper pull rod occurs according to a positional relationship between a central position of the upper pull rod and upper and lower edges of the upper pull rod bracket specifically includes:
if the central position of the upper pull rod is positioned between the upper edge and the lower edge of the upper pull rod bracket, judging that the upper pull rod has no play fault;
and if the central position of the upper pull rod is not positioned between the upper edge and the lower edge of the upper pull rod bracket, judging that the upper pull rod has a play fault.
And S5, if the judgment result is that the upper pull rod has a play fault, uploading fault information and giving an alarm.
According to the technical scheme, the prediction speed of the upper pull rod play-out fault positioning model is increased by utilizing the deep learning technology and optimizing the upper pull rod play-out fault positioning model. And then detecting the upper pull rod fleeing fault part by using the trained upper pull rod fleeing fault positioning model, and predicting the fault probability value of the upper pull rod fleeing fault. Directly judging that the detected probability value of the fault component is greater than a preset first fault probability value to send an upper pull rod to break out of the fault; and if the fault probability value is less than the preset first fault probability value and greater than the preset second fault probability value, further positioning the position information of the upper pull rod bracket and the upper pull rod, and judging whether the upper pull rod is separated from the upper pull rod through the position relation between the upper pull rod bracket and the upper pull rod, thereby effectively avoiding the generation of false alarm. Thereby can be faster more accurate realization go up pull rod play fault detection to railway freight car, and then can alleviate artificial working strength greatly, shorten and overhaul the activity duration, improve the work efficiency who overhauls, reduce the probability of lou examining, reduce human cost and administrative cost to guarantee the safe operation of freight car.
Example two
The embodiment provides a method for identifying a pull rod play fault on a truck, which comprises the following steps:
s1, acquiring an image to be detected. The image to be detected is shown in fig. 6.
S2, positioning the position of a component with the upper pull rod jumping out of fault in the image to be detected, and predicting the probability value of the fault with the upper pull rod jumping out of fault; the upper pull rod play fault component comprises an upper pull rod and an upper pull rod bracket. In this embodiment, the position of the component with the fault caused by the play of the upper pull rod in the image to be detected is located by using an upper pull rod play fault location model, and the fault probability value of the fault caused by the play of the upper pull rod is predicted. The position of the upper pull rod out of the fault part in the image to be detected is shown in fig. 7.
Wherein, the training process of the upper pull rod jumping fault positioning model comprises the following steps: the method comprises the steps of collecting a large number of original images with faults fleeed by an upper pull rod, utilizing image enhancement to expand sample data to form an image data set A, then dividing the image data set A into a training data set and a testing data set, wherein the training data set accounts for 9/10 of the whole image data set, the testing data set accounts for 1/10 of the whole image data set, utilizing a VOC data set format to finish labeling the training data set, then training the fault location model with the upper pull rod fleeing through SGD (random gradient descent), the initial learning rate is 0.001, training 40000 times, adjusting the learning rate to 0.0001, training 5000 times, adjusting the learning rate to 0.00001, training 5000 times, storing an intermediate network model parameter file every 1000 times of training, testing the trained network model parameter file through the testing data set, and selecting an optimal network model related parameter file.
The main structure of the upper pull rod fleeing fault positioning model adopts a main structure of yolo-v3, and considering that the upper pull rod has relatively large size in a picture and fixed size, two output layers of 26 multiplied by 255 and 52 multiplied by 255 in a yolo-v3 network structure are deleted, only the output layers of 13 multiplied by 255 are reserved, the network structure is simplified, and the model detection speed is improved. The default anchor box aspect ratio for yolo-v3 was obtained based on the ImageNet training set using the k-means clustering algorithm. In the upper pull rod detection process, the marking frame of the upper pull rod is a rectangle with a large length-width ratio, and the accuracy of the final training model is affected if a universal anchor box aspect ratio is adopted, so that the corresponding anchor box aspect ratio is generated according to the aspect ratio of the upper pull rod sample in the embodiment, the original default value is replaced, and the upper pull rod fleeing fault positioning model is easier to converge.
And S3, if the fault probability value of the upper pull rod jumping out of the fault is greater than the preset first fault probability value, judging that the upper pull rod jumping out of the fault occurs.
And S4, if the fault probability value of the upper pull rod jumping out of fault is smaller than the preset first fault probability value and larger than the preset second fault probability value, judging whether the upper pull rod jumping out of fault occurs or not according to the position relation between the central position of the upper pull rod and the upper edge and the lower edge of the upper pull rod bracket.
In some embodiments, the first failure probability value is 0.9 and the second failure probability value is 0.5.
In some embodiments, the determination of the center position of the upper tie bar and the upper and lower edges of the upper tie bar bracket comprises:
positioning the position coordinates of an upper pull rod and an upper pull rod supporting groove in the image to be detected; in some embodiments, the position coordinates of the upper pull rod and the upper pull rod bracket in the image to be detected are positioned by using a preset upper pull rod and upper pull rod bracket positioning model. Wherein, the training process of going up pull rod and upper pull rod bracket location model includes: the method comprises the steps of collecting a large number of original images of an upper pull rod and an upper pull rod bracket, utilizing image enhancement to amplify sample data to form an image data set B, then dividing the image data set B into a training data set and a testing data set, wherein the training data set accounts for 9/10 of the whole image data set, the testing data set accounts for 1/10 of the whole image data set, utilizing a VOC data set format to finish labeling the upper pull rod and the upper pull rod bracket in the training data set, then training an upper pull rod and an upper pull rod bracket positioning model through SGD (random gradient descent), adjusting the learning rate to be 0.001, training 40000 times, adjusting the learning rate to be 0.0001, training 5000 times, adjusting the learning rate to be 0.00001, training 5000 times, storing an intermediate network model parameter file once per training 1000 times, testing the trained network model parameter file through the testing data set, and selecting an optimal network model related parameter file.
And respectively calculating the central position of the upper pull rod and the upper edge and the lower edge of the upper pull rod bracket according to the position coordinates of the upper pull rod and the upper pull rod bracket in the image to be detected.
For example, as shown in fig. 8, the position coordinates of the upper pull rod and the upper pull rod receiving groove in the image to be detected are (Xs, ys, ws, hs), where (Xs, ys) represents the coordinates of the upper left point of the positioning frame for positioning the upper pull rod in the image to be detected, ws represents the width of the positioning frame of the upper pull rod, and Hs represents the height of the positioning frame of the upper pull rod.
The position coordinates of the upper pull rod bracket are (Xt, yt, wt, ht), wherein (Xt, yt) represents the coordinates of the upper left point of the positioning frame for positioning the upper pull rod bracket in the image to be detected, wt represents the width of the positioning frame of the upper pull rod bracket, and Hs represents the height of the positioning frame of the upper pull rod bracket.
The central position of the upper pull rod is the central position of the upper edge and the lower edge of the positioning frame of the upper pull rod
The upper edge and the lower edge of the upper pull rod bracket are respectively Yt and Y t +H t . If Yt<center s <(Y t +H t ) The upper pull rod is in a normal state, otherwise, the upper pull rod has a play fault.
In some embodiments, the determining whether a failure of the upper pull rod occurs according to a positional relationship between a central position of the upper pull rod and upper and lower edges of the upper pull rod bracket specifically includes:
if the central position of the upper pull rod is positioned between the upper edge and the lower edge of the upper pull rod bracket, judging that the upper pull rod has no play fault;
and if the central position of the upper pull rod is not positioned between the upper edge and the lower edge of the upper pull rod bracket, judging that the upper pull rod has a play fault.
And S5, if the judgment result shows that the upper pull rod has a play fault, uploading fault information and giving an alarm.
According to the technical scheme, the prediction speed of the upper pull rod play-out fault positioning model is increased by utilizing the deep learning technology and optimizing the upper pull rod play-out fault positioning model. And then detecting the upper pull rod jumping fault part by using the trained upper pull rod jumping fault positioning model, and predicting the fault probability value of the upper pull rod jumping fault. Directly judging that the detected probability value of the fault component is greater than a preset first fault probability value to send an upper pull rod to break out of the fault; and if the fault probability value is less than the preset first fault probability value and greater than the preset second fault probability value, further positioning the position information of the upper pull rod bracket and the upper pull rod, and judging whether the upper pull rod is separated from the upper pull rod through the position relation between the upper pull rod bracket and the upper pull rod, thereby effectively avoiding the generation of false alarm. Thereby can be faster more accurate realization go up pull rod play fault detection to railway freight car, and then can alleviate artificial working strength greatly, shorten and overhaul the activity duration, improve the work efficiency who overhauls, reduce the probability of lou examining, reduce human cost and administrative cost to guarantee the safe operation of freight car.
EXAMPLE III
The embodiment of the system can be used for executing the embodiment of the method, and for details which are not disclosed in the embodiment of the system, please refer to the embodiment of the method. Fig. 9 is a schematic structural diagram of a device for identifying a pull rod channeling failure in a truck according to an embodiment of the present application, and as shown in fig. 9, the device according to the embodiment includes:
an image obtaining module 810 for obtaining an image to be detected
The target detection module 820 is used for positioning the position of a part with the fault of the upper pull rod in the image to be detected and predicting the fault probability value of the fault of the upper pull rod; the upper pull rod play fault component comprises an upper pull rod and an upper pull rod bracket. Specifically, the target detection module 820 uses an upper pull rod crossing fault positioning model to position the position of the upper pull rod crossing fault component in the image to be detected, and predicts the fault probability value of the upper pull rod crossing fault.
The training process of the upper pull rod play fault positioning model comprises the following steps: a large number of original images of the upper pull rod which breaks out the fault are collected, sample data are augmented by using image enhancement to form an image data set A, then the image data set A is divided into a training data set and a testing data set, the training data set accounts for 9/10 of the whole image data set, the testing data set accounts for 1/10 of the whole image data set, marking of the training data set is completed by using a VOC data set format, then training of the upper pull rod which breaks out the fault positioning model is achieved through SGD (random gradient descent), the initial learning rate is 0.001, training is 40000 times, the learning rate is adjusted to be 0.0001, training is 5000 times, the learning rate is adjusted to be 0.00001, an intermediate network model parameter file is stored every 1000 times of training, the trained network model parameter file is tested through the testing data set, and an optimal network model related parameter file is selected.
The main structure of the upper pull rod play-out fault positioning model adopts a yolo-v3 main structure, and considering that the upper pull rod is relatively large in size and fixed in size in a picture, two output layers of 26 multiplied by 255 and 52 multiplied by 255 in a yolo-v3 network structure are deleted, only the output layers of 13 multiplied by 255 are reserved, the network structure is simplified, and the model detection speed is improved. The default anchor box aspect ratio for yolo-v3 was obtained using the k-means clustering algorithm based on the ImageNet training set. In the upper pull rod detection process, the marking frame of the upper pull rod is a rectangle with a large length-width ratio, and if the universal aspect box aspect ratio is adopted, the accuracy of the final training model is affected, so that the embodiment generates the corresponding aspect box aspect ratio according to the size of the upper pull rod sample, replaces the original default value, and enables the upper pull rod to be easier to converge when the fault positioning model is exposed.
And the positioning module 830 is configured to position coordinates of the upper pull rod and the upper pull rod tray in the image to be detected. In this embodiment, the positioning module 830 locates the position coordinates of the upper pull rod and the upper pull rod bracket in the image to be detected by using a preset upper pull rod and upper pull rod bracket positioning model. Wherein, the training process of the upper pull rod and the upper pull rod bracket positioning model comprises the following steps: a large number of original images of an upper pull rod and an upper pull rod bracket are collected, sample data are augmented by using image enhancement to form an image data set B, then the image data set B is divided into a training data set and a testing data set, the training data set accounts for 9/10 of the whole image data set, the testing data set accounts for 1/10 of the whole image data set, marking of the upper pull rod and an upper pull rod bracket in the training data set is completed by using a VOC data set format, then training of an upper pull rod and an upper pull rod bracket positioning model is realized through SGD (random gradient descent), the initial learning rate is 0.001, training is 40000 times, the learning rate is adjusted to 0.0001 and 5000 times, the learning rate is adjusted to 0.00001 and is 5000 times, a middle network model parameter file is stored every 1000 times of training, the trained network model parameter file is tested by using the testing data set, and an optimal network model related parameter file is selected.
And the calculating module 840 is used for respectively calculating the central position of the upper pull rod and the upper edge and the lower edge of the upper pull rod bracket according to the position coordinates of the upper pull rod and the upper pull rod bracket in the image to be detected.
Illustratively, the position coordinates of the upper pull rod in the image to be detected are (Xs, ys, ws, hs), where (Xs, ys) represents the coordinates of the upper left point of the positioning frame for positioning the upper pull rod in the image to be detected, ws represents the width of the positioning frame of the upper pull rod, and Hs represents the height of the positioning frame of the upper pull rod.
The position coordinates of the upper pull rod bracket groove are (Xt, yt, wt, ht), wherein (Xt, yt) represents the coordinates of an upper left point of a positioning frame for positioning the upper pull rod bracket in the image to be detected, wt represents the width of the positioning frame of the upper pull rod bracket, and Hs represents the height of the positioning frame of the upper pull rod bracket.
The central position of the upper pull rod is the central position of the upper edge and the lower edge of the positioning frame of the upper pull rod
The upper edge and the lower edge of the upper pull rod bracket are respectively Yt and Y t +H t . If Yt<center s <(Y t +H t ) The upper pull rod is in a normal state, otherwise, the upper pull rod has a play fault.
A determining module 850, configured to determine that the upper pull rod fleeing fault occurs if a fault probability value of the upper pull rod fleeing fault is greater than a preset first fault probability value; and if the fault probability value of the upper pull rod jumping fault is smaller than the preset first fault probability value and larger than the preset second fault probability value, judging whether the upper pull rod jumping fault occurs or not according to the position relation between the central position of the upper pull rod and the upper edge and the lower edge of the upper pull rod bracket. The method specifically comprises the following steps: if the central position of the upper pull rod is positioned between the upper edge and the lower edge of the upper pull rod bracket, judging that the upper pull rod has no play fault; and if the central position of the upper pull rod is not positioned between the upper edge and the lower edge of the upper pull rod bracket, judging that the upper pull rod has a play fault. In some embodiments, the first failure probability value is 0.9 and the second failure probability value is 0.5.
And the alarm module 860 is used for uploading fault information and giving an alarm when the upper pull rod is judged to have a fault.
According to the technical scheme, the prediction speed of the upper pull rod play-out fault positioning model is increased by utilizing the deep learning technology and optimizing the upper pull rod play-out fault positioning model. And then detecting the upper pull rod fleeing fault part by using the trained upper pull rod fleeing fault positioning model, and predicting the fault probability value of the upper pull rod fleeing fault. Directly judging that the detected probability value of the fault component is greater than a preset first fault probability value, and sending an upper pull rod to break out the fault; and if the fault probability value is less than the preset first fault probability value and greater than the preset second fault probability value, further positioning the position information of the upper pull rod bracket and the upper pull rod, and judging whether the upper pull rod is separated from the upper pull rod through the position relation between the upper pull rod bracket and the upper pull rod, thereby effectively avoiding the generation of false alarm. Thereby can be faster more accurate realization go up pull rod play fault detection to railway freight car, and then can alleviate artificial working strength greatly, shorten and overhaul the activity duration, improve the work efficiency who overhauls, reduce the probability of lou examining, reduce human cost and administrative cost to guarantee the safe operation of freight car.
Example four
The present embodiment also provides a computer-readable storage medium, where a computer program is stored, and when the computer program is executed by a processor, the method steps in the foregoing embodiments may be implemented, and details of the embodiment are not repeated herein.
The computer-readable storage medium may also include, among other things, a computer program, a data file, a data structure, etc., either alone or in combination. The computer-readable storage medium or computer program may be specifically designed and understood by those skilled in the art of computer software, or the computer-readable storage medium may be known and available to those skilled in the art of computer software. Examples of computer-readable storage media include: magnetic media such as hard disks, floppy disks, and magnetic tape; optical media such as CDROM disks and DVDs; magneto-optical media, e.g., optical disks; and hardware devices specifically configured to store and execute computer programs, e.g., read Only Memory (ROM), random Access Memory (RAM), flash memory; or a server, app application mall, etc. Examples of computer programs include both machine code, such as produced by a compiler, and files containing higher level code that may be executed by the computer using an interpreter. The described hardware devices may be configured to act as one or more software modules to perform the operations and methods described above, and vice versa. In addition, the computer readable storage medium may be distributed over network coupled computer systems so that program code or computer programs may be stored and executed in a distributed fashion.
EXAMPLE five
Fig. 10 is a connection block diagram of an electronic device according to an embodiment of the present application, and as shown in fig. 10, the electronic device 400 may include: one or more processors 410, memory 420, multimedia components 430, input/output (I/O) interfaces 440, and communication components 450.
Wherein the processor 410 is configured to perform all or a portion of the steps of the method according to an embodiment. The memory 420 is used to store various types of data, which may include, for example, instructions for any application or method in the electronic device, as well as application-related data.
The Processor 410 may be an Application Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a Programmable Logic Device (PLD), a Field Programmable Gate Array (FPGA), a controller, a microcontroller, a microprocessor, or other electronic components, and is configured to perform the method of the above embodiments.
The Memory 420 may be implemented by any type of volatile or non-volatile Memory device or combination thereof, such as Static Random Access Memory (SRAM), electrically Erasable Programmable Read-Only Memory (EEPROM), erasable Programmable Read-Only Memory (EPROM), programmable Read-Only Memory (PROM), read-Only Memory (ROM), magnetic Memory, flash Memory, magnetic disk or optical disk.
The multimedia component 430 may include a screen, which may be a touch screen, and an audio component for outputting and/or inputting audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may further be stored in a memory or transmitted through a communication component. The audio assembly also includes at least one speaker for outputting audio signals.
The I/O interface 440 provides an interface between the processor 410 and other interface modules, such as a keyboard, mouse, buttons, etc. These buttons may be virtual buttons or physical buttons.
The communication component 450 is used for wired or wireless communication between the electronic device 400 and other devices.
The wired communication includes communication through a network port, a serial port and the like; the wireless communication includes: wi-Fi, bluetooth, near Field Communication (NFC for short), 2G, 3G, 4G, 5G, or a combination of one or more of them. The corresponding communication component 415 may therefore include: wi-Fi module, bluetooth module, NFC module.
It should be further understood that the method or system disclosed in the embodiments provided in the present application may be implemented in other ways. The method or system embodiments described above are merely illustrative, for example, the flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of methods and apparatus according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, a computer program segment, or a portion of a computer program, which comprises one or more computer programs for implementing the specified logical function(s).
It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures, and in fact may be executed substantially concurrently, or in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems that perform the specified functions or acts, or combinations of special purpose hardware and computer programs.
In this application, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one of 8230a of 8230, a" does not exclude the presence of additional identical elements in the process, method, apparatus or device in which the element is comprised; if the description to "first", "second", etc. is used for descriptive purposes only, it is not to be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated or implicitly indicating the precedence of the technical features indicated; in the description of the present application, the terms "plurality" and "plurality" mean at least two unless otherwise specified; if a server is described, it should be noted that the server may be an independent physical server or terminal, or may be a server cluster formed by a plurality of physical servers, or may be a cloud server capable of providing basic cloud computing services such as a cloud server, a cloud database, a cloud storage, a CDN, and the like; if an intelligent terminal or a mobile device is described in the present application, it should be noted that the intelligent terminal or the mobile device may be a mobile phone, a tablet Computer, an intelligent watch, a netbook, a wearable electronic device, a Personal Digital Assistant (PDA), an Augmented Reality device (AR), a Virtual Reality device (VR), a smart television, a smart audio, a Personal Computer (PC), and the like, but is not limited thereto.
Finally, it should be noted that in the description of the present specification, reference to the description of "one embodiment," "some embodiments," "an example," "one example" or "some examples" or the like is intended to mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
While embodiments of the present application have been illustrated and described above, it is to be understood that the above embodiments are exemplary, and the description is only for the purpose of facilitating understanding of the present application and is not intended to limit the present application. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the disclosure as defined by the appended claims.
Claims (12)
1. A method for identifying a pull rod fleeing fault on a truck is characterized by comprising the following steps:
acquiring an image to be detected;
positioning the position of a component with the fault caused by the play of the upper pull rod in the image to be detected, and predicting the probability value of the fault caused by the play of the upper pull rod; the upper pull rod fault-out component comprises an upper pull rod and an upper pull rod bracket;
if the fault probability value of the upper pull rod jumping fault is greater than a preset first fault probability value, judging that the upper pull rod jumping fault occurs according to the judgment result;
and if the fault probability value of the upper pull rod jumping fault is smaller than the preset first fault probability value and larger than the preset second fault probability value, judging whether the upper pull rod jumping fault occurs or not according to the position relation between the central position of the upper pull rod and the upper edge and the lower edge of the upper pull rod bracket.
2. The method as claimed in claim 1, wherein a pull-up rod crossing fault location model is used to locate the position of a pull-up rod crossing fault component in the image to be detected, and the fault probability value of the pull-up rod crossing fault is predicted.
3. The method of claim 1, wherein the determining whether the upper pull rod fleeing fault occurs according to the position relationship between the central position of the upper pull rod and the upper and lower edges of the upper pull rod bracket comprises:
if the central position of the upper pull rod is positioned between the upper edge and the lower edge of the upper pull rod bracket, judging that the upper pull rod has no play fault;
and if the central position of the upper pull rod is not positioned between the upper edge and the lower edge of the upper pull rod bracket, judging that the upper pull rod has a play fault.
4. The method of claim 1, wherein the determining of the center position of the upper tie bar and the upper and lower edges of the upper tie bar bracket comprises:
positioning the position coordinates of an upper pull rod and an upper pull rod supporting groove in the image to be detected;
and respectively calculating the central position of the upper pull rod and the upper edge and the lower edge of the upper pull rod bracket according to the position coordinates of the upper pull rod and the upper pull rod bracket in the image to be detected.
5. The method of claim 4, wherein the coordinates of the upper tie rod and the upper tie rod bracket in the image to be detected are located by using a preset upper tie rod and upper tie rod bracket location model.
6. The method of claim 2, wherein the tie-up rod channeling fault location model is a deep learning network model that employs a body structure of a new yolo-v3 network model that employs only one output layer.
7. The method of claim 6, wherein the aspect ratio of the bounding box of the tie-up pole blow-out fault localization model is determined from the aspect ratio of a sample of the tie-up pole.
8. The method according to claim 1, characterized in that if the judgment result is that the upper pull rod fleeing fault occurs, the upper pull rod fleeing fault information is uploaded and an alarm is given.
9. The method of claim 1, wherein the preset first fault probability value is 0.9 and the preset second fault probability value is 0.5.
10. An identification device for a pull rod play fault on a truck, the device comprising:
the image acquisition module is used for acquiring an image to be detected;
the target detection module is used for positioning the position of a component with the fault of the upper pull rod in the image to be detected and predicting the probability value of the fault of the upper pull rod in the fault; the upper pull rod fault-out component comprises an upper pull rod and an upper pull rod bracket;
the judging module is used for judging that the upper pull rod fleeing fault occurs if the fault probability value of the upper pull rod fleeing fault is greater than a preset first fault probability value; and
and if the fault probability value of the upper pull rod jumping fault is smaller than the preset first fault probability value and larger than the preset second fault probability value, judging whether the upper pull rod jumping fault occurs or not according to the position relation between the central position of the upper pull rod and the upper edge and the lower edge of the upper pull rod bracket.
11. A computer-readable storage medium storing a computer program which, when executed by one or more processors, implements a method of identifying a pull-out malfunction on a truck draw-bar according to any one of claims 1 to 9.
12. An electronic device, comprising a memory and one or more processors, wherein the memory stores a computer program, the memory and the processors are communicatively connected, and when the computer program is executed by the processors, the method for identifying a pull-out fault on a truck upper tie is performed according to any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211406790.4A CN115880624A (en) | 2022-11-10 | 2022-11-10 | Identification method for pull rod play fault on truck |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211406790.4A CN115880624A (en) | 2022-11-10 | 2022-11-10 | Identification method for pull rod play fault on truck |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115880624A true CN115880624A (en) | 2023-03-31 |
Family
ID=85759639
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211406790.4A Pending CN115880624A (en) | 2022-11-10 | 2022-11-10 | Identification method for pull rod play fault on truck |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115880624A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116153041A (en) * | 2023-04-17 | 2023-05-23 | 江西联创光电超导应用有限公司 | Pull rod emergency early warning system for superconducting magnet |
-
2022
- 2022-11-10 CN CN202211406790.4A patent/CN115880624A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116153041A (en) * | 2023-04-17 | 2023-05-23 | 江西联创光电超导应用有限公司 | Pull rod emergency early warning system for superconducting magnet |
| CN116153041B (en) * | 2023-04-17 | 2023-08-18 | 江西联创光电超导应用有限公司 | Pull rod emergency early warning system for superconducting magnet |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111626208B (en) | Method and device for detecting small objects | |
| CN111899227A (en) | Automatic railway fastener defect acquisition and identification method based on unmanned aerial vehicle operation | |
| KR20210078530A (en) | Lane property detection method, device, electronic device and readable storage medium | |
| JP2022549541A (en) | Defect detection method, apparatus, electronic equipment and computer storage medium | |
| CN110377682B (en) | Track type determination method and device, computing equipment and storage medium | |
| CN113343461A (en) | Simulation method and device for automatic driving vehicle, electronic equipment and storage medium | |
| CN111199541A (en) | Image quality evaluation method, image quality evaluation device, electronic device, and storage medium | |
| CN115285185B (en) | Method and device for testing train speed limit processing function | |
| CN115880624A (en) | Identification method for pull rod play fault on truck | |
| CN111121797B (en) | Road screening method, device, server and storage medium | |
| CN113392793A (en) | Method, device, equipment, storage medium and unmanned vehicle for identifying lane line | |
| CN110852253A (en) | Ladder control scene detection method and device and electronic equipment | |
| CN111401229A (en) | Visual small target automatic labeling method and device and electronic equipment | |
| CN116964588A (en) | Target detection method, target detection model training method and device | |
| CN111488866B (en) | Invading object identification method and device based on deep learning and electronic equipment | |
| CN115775367A (en) | Road target detection method, detection device, electronic equipment and storage medium | |
| CN114510468A (en) | Data processing method and computer program product | |
| CN113160217A (en) | Method, device and equipment for detecting foreign matters in circuit and storage medium | |
| CN112380986A (en) | Vehicle type detection method and system for resisting complex street background interference and storage medium | |
| CN112215042A (en) | Parking space limiter identification method and system and computer equipment | |
| CN111326172A (en) | Conflict detection method and device, electronic equipment and readable storage medium | |
| CN113793330B (en) | Method and system for detecting road surface flatness area | |
| CN111753625B (en) | Pedestrian detection method, device, equipment and medium | |
| Setiawan et al. | Monitoring Road Surface Conditions with Cyclist's Smartphone Sensors | |
| CN110991447B (en) | Train number accurate positioning and identifying method based on deep learning |
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 |