CN113371035A - Train information identification method and system - Google Patents
Train information identification method and system Download PDFInfo
- Publication number
- CN113371035A CN113371035A CN202110934199.5A CN202110934199A CN113371035A CN 113371035 A CN113371035 A CN 113371035A CN 202110934199 A CN202110934199 A CN 202110934199A CN 113371035 A CN113371035 A CN 113371035A
- Authority
- CN
- China
- Prior art keywords
- information
- image data
- identification
- area
- train
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000002372 labelling Methods 0.000 claims abstract description 47
- 238000004364 calculation method Methods 0.000 claims abstract description 12
- 238000001914 filtration Methods 0.000 claims abstract description 9
- 238000012545 processing Methods 0.000 claims description 11
- 238000007689 inspection Methods 0.000 claims description 10
- 238000012549 training Methods 0.000 claims description 9
- 238000013136 deep learning model Methods 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 description 10
- 230000009471 action Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 239000012535 impurity Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000010422 painting Methods 0.000 description 2
- 206010047571 Visual impairment Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61L—GUIDING RAILWAY TRAFFIC; ENSURING THE SAFETY OF RAILWAY TRAFFIC
- B61L25/00—Recording or indicating positions or identities of vehicles or trains or setting of track apparatus
- B61L25/02—Indicating or recording positions or identities of vehicles or trains
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61L—GUIDING RAILWAY TRAFFIC; ENSURING THE SAFETY OF RAILWAY TRAFFIC
- B61L25/00—Recording or indicating positions or identities of vehicles or trains or setting of track apparatus
- B61L25/02—Indicating or recording positions or identities of vehicles or trains
- B61L25/021—Measuring and recording of train speed
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61L—GUIDING RAILWAY TRAFFIC; ENSURING THE SAFETY OF RAILWAY TRAFFIC
- B61L25/00—Recording or indicating positions or identities of vehicles or trains or setting of track apparatus
- B61L25/02—Indicating or recording positions or identities of vehicles or trains
- B61L25/023—Determination of driving direction of vehicle or train
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61L—GUIDING RAILWAY TRAFFIC; ENSURING THE SAFETY OF RAILWAY TRAFFIC
- B61L25/00—Recording or indicating positions or identities of vehicles or trains or setting of track apparatus
- B61L25/02—Indicating or recording positions or identities of vehicles or trains
- B61L25/04—Indicating or recording train identities
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- Image Analysis (AREA)
- Electric Propulsion And Braking For Vehicles (AREA)
- Train Traffic Observation, Control, And Security (AREA)
Abstract
The application provides a train information identification method and system, which relate to the field of information identification, and the identification method comprises the following steps: acquiring image data corresponding to a target train; taking an area containing data content as a labeling area, and filtering the image data to obtain an information frame containing the labeling area; inputting the information frame into an information identification model, taking a residual error network as a backbone network, taking the marked area as a remarkable characteristic, and extracting the picture characteristics in the marked area in a layering way; constructing a feature pyramid of the picture features; and carrying out full convolution network calculation on the characteristic pyramid to obtain the vehicle information of the target train. According to the method and the device, manual recording operation is not needed, the recognition accuracy is improved by using the information recognition model, and the recognition accuracy of carriage information is guaranteed.
Description
Technical Field
The present application relates to the field of information identification, and in particular, to a train information identification method and system.
Background
In the current railway system, freight trains are various in types, complex in models, large in vehicle flow quantity and wide in flow regions, so that the management, monitoring, statistics and tracking of the operation conditions of the vehicles are very complex work, errors are not allowed in vehicle inspection, vehicle transfer, loading and unloading and the like in the whole railway transportation process, the carriage number is a unique identification unit of all works, and the identification work of the carriage number runs through all links. At present, carriage information is mainly registered and verified manually, the workload is large, the efficiency is low, the accuracy cannot be guaranteed, and in order to improve the running efficiency of a railway, firstly, the problem that the composition of a train and the flowing condition of the train must be accurately mastered in time, the information of the train such as the type, carriage number, load, dead weight, volume, length change, scheduled inspection period and the like during running of the train, the running direction of the train and the running speed information of the train must be automatically identified is solved. In the prior art, the national railway vehicle provided with the electronic tag of the railway department truck automatically reads the vehicle information in the electronic tag by adopting a radio frequency technology, and the problems of incomplete vehicle information, skip number, easy damage and loss of a chip and the like exist firstly; secondly, the carriage information is updated to the electronic tag only when the carriage information is periodically overhauled, and the carriage information cannot be updated in time; and electronic tags are not installed on part of domestic railway vehicles and self-contained vehicles, so that a large amount of manual intervention is still needed at present, the registration and verification work of the carriage information is carried out, and the human resources are extremely consumed.
Disclosure of Invention
The application aims to provide a train information identification method and an identification system, which can automatically identify train information of a train.
In order to solve the technical problem, the application provides a train information identification method, which has the following specific technical scheme:
acquiring image data corresponding to a target train;
taking an area containing data content as a labeling area, and filtering the image data to obtain an information frame containing the labeling area;
inputting the information frame into an information identification model, taking a residual error network as a backbone network, taking the marked area as a remarkable characteristic, and extracting the picture characteristics in the marked area in a layering way;
constructing a feature pyramid of the picture features;
and carrying out full convolution network calculation on the characteristic pyramid to obtain the vehicle information of the target train.
Optionally, the method further includes:
acquiring a historical carriage image data set;
labeling each image data in the historical carriage image data set, and determining the corresponding identification degree of the image data according to a preset identification degree classification table;
and training all image data containing the recognition degree by utilizing a preset deep learning model to obtain the information recognition model.
Optionally, labeling each image data in the historical car image data set:
labeling each image data in the historical carriage image data set according to the content of a labeling area; the first target comprises a carriage number area, an attribute area and scheduled inspection period information, the second target comprises character information, and the third target comprises a carriage shaft and carriage type information.
Optionally, determining the degree of identification corresponding to the image data according to the preset degree of identification classification table includes:
and determining the corresponding recognition degree of the image data according to the resolution of the image data and the number of the unidentifiable labeling areas.
Optionally, after the information frame is input to an information recognition model and vehicle information of the target train is obtained, the method further includes:
and uploading the vehicle information to an upper computer so that the upper computer can compile a train information table according to the vehicle information.
Optionally, the vehicle information includes any one or a combination of any several of a vehicle type, a carriage number, a load, a self weight, a volume, a length change, a scheduled inspection period, a driving speed and a driving direction.
The present application further provides a train information recognition system, including:
the data acquisition module is used for acquiring image data corresponding to the target train;
the data processing module is used for taking an area containing data content as a labeling area, filtering the image data and obtaining an information frame containing the labeling area;
the characteristic extraction module is used for inputting the information frame into an information identification model, taking a residual error network as a backbone network, taking the marked area as a remarkable characteristic, and extracting the picture characteristics in the marked area in a layering manner;
the characteristic structure building module is used for building a characteristic pyramid of the picture characteristics;
and the identification module is used for carrying out full convolution network calculation on the characteristic pyramid to obtain an identification result.
The application provides a train information identification method, which comprises the following steps: acquiring image data corresponding to a target train; taking an area containing data content as a labeling area, and filtering the image data to obtain an information frame containing the labeling area; inputting the information frame into an information identification model, taking a residual error network as a backbone network, taking the marked area as a remarkable characteristic, and extracting the picture characteristics in the marked area in a layering way; constructing a feature pyramid of the picture features; and carrying out full convolution network calculation on the characteristic pyramid to obtain the vehicle information of the target train.
According to the method and the device, after the image data corresponding to the target train is obtained, the labeling area with the effective information in the image data is labeled, so that the non-labeling area without the information identification value in the image data is filtered. And then, the information identification model is used for carrying out model identification on the information frame, the vehicle information of the target train is directly output, manual recording operation is not needed, the human resources required by train information identification are saved, the identification accuracy is improved by using the information identification model, the identification accuracy of carriage information is ensured, and information support is provided for scientific management of railway vehicles.
The application also provides a train information identification system, which has the beneficial effects that the identification system is not repeated herein.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, it is obvious that the drawings in the following description are only 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 train information identification method according to an embodiment of the present disclosure;
fig. 2 is a schematic diagram of an identification process of train information according to an embodiment of the present application;
fig. 3 is a structural diagram of an identification system for train information according to an embodiment of the present disclosure;
in fig. 2, 21 denotes a train, 22 denotes a camera, and 23 denotes a fill light.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
Referring to fig. 1, fig. 1 is a flowchart of a method for identifying train information according to an embodiment of the present application, where the method includes:
s101: acquiring image data corresponding to a target train;
this step aims to acquire image data corresponding to the target train. How to acquire the image data is not particularly limited, and image data captured by cameras on both sides of the target train track may be used, or other high-speed cameras may be used. In addition, since the target train is usually in a driving state when being photographed, the acquired image data may be video data, picture data, or both.
When the cameras on the two sides of the target train track are used for acquiring image data, corresponding image data can be acquired when the target train is detected to pass through. The identification process of the train information is not performed when the track is in the empty track state.
It should be understood that, if the image data is video data, the step includes a processing procedure of the video data by default, for example, parsing the video data, and obtaining corresponding picture data by using a data frame as a unit.
S102: taking an area containing data content as a labeling area, and filtering the image data to obtain an information frame containing the labeling area;
and after the image data is obtained, labeling the image data to obtain a corresponding labeling area. The callout region refers to a region containing the corresponding data content. Since most of the areas on the surface of the target train do not include the vehicle information, only the areas including the data content are treated as the label areas in order to reduce the data processing amount.
During labeling, the contents of the labeling area divided in advance can be classified. The classification process of the label area is not limited herein. For example, image data is mainly classified into three categories: the car type, the connecting shaft and the painting information, and the painting information may include at least one of car number information, a car attribute information image and a car scheduled inspection period, the marking region in the image data may be determined one by one as above.
In labeling, the labeling may also be performed in a preset labeling order, for example, from left to right, from top to bottom.
After the image data are labeled, filtering the image data, thereby removing a non-labeling area in the image data, wherein the labeling area contains the vehicle information of the target train, and obtaining an information frame containing the vehicle information. The specific content of the vehicle information in the present embodiment is not limited, and the vehicle information may include information such as a vehicle type, a car number, a load, a self weight, a volume, a length change, a scheduled inspection period, a traveling speed, and a traveling direction, may be any one or a combination of any two of them, and may be set by a person skilled in the art.
S103: inputting the information frame into an information identification model, taking a residual error network as a backbone network, taking the marked area as a remarkable characteristic, and extracting the picture characteristics in the marked area in a layering way;
s104: constructing a feature pyramid of the picture features;
s105: and carrying out full convolution network calculation on the characteristic pyramid to obtain the vehicle information of the target train. And then inputting the information frame into an information identification model, and identifying the information frame by using the information identification model to obtain the vehicle information of the target train.
Specifically, in this step, the information frame may be input to the information identification model, the residual network is used as the backbone network, the labeled area is used as the salient feature, the picture features in the labeled area are extracted in a layered manner, then a feature pyramid of the picture features is constructed, and the feature pyramid is subjected to full convolution network calculation, so that the identification result can be obtained. Features of C2-C5 layers in a comprehensive residual error network in the information identification model form a feature pyramid, small target features are mainly concerned by C2 and C3 layers, large target features are mainly concerned by C4 and C5 layers, and therefore the feature pyramid after the integration can consider both the large target and the small target identification features.
According to the method and the device, after the image data corresponding to the target train is obtained, the labeling area with the effective information in the image data is labeled, so that the non-labeling area without the information identification value in the image data is filtered. And then, the information identification model is used for carrying out model identification on the information frame, the vehicle information of the target train is directly output, manual recording operation is not needed, the human resources required by train information identification are saved, the identification accuracy is improved by using the information identification model, the identification accuracy of carriage information is ensured, and information support is provided for scientific management of railway vehicles.
Further, on the basis of this embodiment, the information frame is input to the information recognition model, and after the vehicle information of the target train is obtained, the vehicle information can also be uploaded to the upper computer, so that the upper computer can compile a train information table according to the vehicle information. Because the embodiment can identify the vehicle information of the target train, the information of the target train can be summarized to obtain the train information table, the train information table can contain the passing time, the carriage information and the like of the train, and the operation information of the train can be conveniently summarized and controlled.
On the basis of the foregoing embodiments, the present embodiment further describes an information recognition model, and herein provides a generation process of the information recognition model, which is specifically as follows:
s201: acquiring a historical carriage image data set;
the historical car image data set is acquired first, and the specific number of image data in the historical car image data set is not limited, and it is easily understood that the more the historical car image data set is, the more accurate the information recognition model is finally generated. For example, not less than 10 or 20 ten thousand historical car image data sets may be used as learning samples for training.
S202: labeling each image data in the historical carriage image data set, and determining the corresponding identification degree of the image data according to a preset identification degree classification table;
in this step, each image data of the historical car image data set needs to be labeled, and each image data of the car image data set can be labeled according to the content of the labeling area. The first object includes a car number area, an attribute area, and a scheduled inspection period information, the second object includes character information, and the third object includes a car axle and car type information. By classifying the label region, it is possible to improve the corresponding image recognition efficiency when the specified target data is recognized by the information recognition model thereafter. For example, if the attribute region is identified, which belongs to the first target and is a type trained in the information recognition model, the specific information recognition can be performed on the first target.
After labeling, the degree of identification corresponding to each image data needs to be determined. The classification standard of the preset identification degree classification table is not limited, and the identification degree corresponding to the image data can be determined according to the resolution of the image data and the number of the unrecognized marked areas. If the resolution of the image data is the definition of the target train in the image data, since the image data captured by the target train in the high-speed running process may include factors affecting information identification, such as light spots or afterimages, the corresponding identification degree can be determined according to the resolution of the image data. The definition of the degree of identification may also be performed according to the unrecognizable marked area, and there may be image data with high image data resolution but difficult marked area content identification, and at this time, the degree of identification of the image data may be defined according to the unrecognizable marked area, and of course, the greater the number of the unrecognizable marked areas, the lower the degree of identification. The resolution of the image data and the number of unrecognizable marked areas can also be used as the criterion for determining the degree of identification, and a preferred predetermined classification table for the degree of identification is provided below:
and (3) carrying out identification degree scoring on each image data according to a preset identification degree classification table, wherein the scoring interval is 0-10 points, and the scoring standard is as follows:
10 minutes, the marking area is complete, the position is in the middle of the picture, the whole is clear, no light spots or other impurities exist around the marking area, and all the marking areas are clear;
9 min, the marked area is complete, each marked area can be clearly seen, and light spots or other impurities are arranged around the marked area;
8, marking areas are complete, 1-2 marking areas are not clear enough and can be recognized, and individual characters are incomplete or have shadows and can be recognized;
7, marking areas are complete, and 3-4 marking areas are not clear enough but can be identified;
6, marking the area completely, and making the whole fuzzy but recognizable;
5, 1 marked area cannot be identified;
4, 2 marked areas cannot be identified;
3, 3 marked areas cannot be identified;
2, 4 marked areas cannot be identified;
1, more than half of the marked areas cannot be identified;
score 0 tends to be blank and can only recognize 1-2 characters.
Of course, the above provides a more specific identification degree determination method for the present embodiment. Other variations of the preset recognition degree classification table, such as an image recognition degree determination model including the preset recognition degree classification table, may be used by those skilled in the art to perform recognition degree determination on the image data, which is not limited herein.
When the information recognition model is used to obtain the vehicle information of the target train, the recognition degree of each information frame can be determined first, and the information frame with higher recognition degree can be selected for recognition.
S203: and training all image data containing the recognition degree by utilizing a preset deep learning model to obtain the information recognition model.
A Matrix-LirsNet deep learning model can be utilized, wherein the model takes a residual error network (ResNet) as a backbone network, and picture features are extracted in a layered mode (C2-C5); according to the remarkable characteristics of the identified information (large targets such as carriage intervals and marked information areas, and small targets such as information characters), C3-C5 layers of picture features are selected as the basis, and a Feature Pyramid (FPN) is constructed, so that the visual field range of the picture features participating in model training can cover the first target to the third target, and the full coverage of the large targets and the small targets is realized. And respectively carrying out full convolution network calculation on the target category and the target position by the model on the basis of the picture characteristic pyramid to obtain a final recognition result.
In order to achieve the above object, a specific application process of the present application is described below, referring to fig. 2, including: the system comprises a camera 22, a switch, an image processing unit, an upper computer and an output unit.
The camera 22 is used for acquiring high-definition video signals of the running freight train, and can further comprise a light supplement lamp 23, so that the camera is suitable for shooting clear videos in a dark environment. The video signal is transmitted as image data to the image processing unit through the switch. The plurality of cameras 22 may be installed at both sides of the freight train track, and as shown in fig. 2, the cameras capture the traveling freight train 21 to obtain high-definition video.
The switch is used to transmit the high definition video signal captured by the camera 22 to the image processing unit.
The image processing unit is used for analyzing and identifying the received image data to obtain the vehicle information of the freight train. Specifically, a deep learning model calculation engine is used for selecting a specific target characteristic layer by taking a residual error network as a backbone network according to the remarkable characteristics of the identified information, and performing full convolution network calculation on the target category and the target position respectively to obtain a final identification result which is sent to an upper computer.
The upper computer is used for integrating the vehicle information of the freight trains to form a whole train of freight train information configuration table, and the output unit provides the whole train of freight train information configuration table to the service application system. It is easy to understand that the switch, the image processing unit, the upper computer and the output unit are not necessarily arranged at the rail accessory, and since the high-definition video signal acquired by the camera 22 can be remotely transmitted through the switch, the image processing unit, the upper computer and the output unit can be located locally, so that a user can obtain finally identified vehicle information locally.
In the following, the train information identification system provided by the embodiment of the present application is introduced, and the identification system described below and the train information identification method described above may be referred to correspondingly.
Fig. 3 is a structural diagram of an identification system of train information according to an embodiment of the present application, and the present application further provides an identification system of train information, including:
the data acquisition module is used for acquiring image data corresponding to the target train;
the data processing module is used for taking an area containing data content as a labeling area, filtering the image data and obtaining an information frame containing the labeling area;
the characteristic extraction module is used for inputting the information frame into an information identification model, taking a residual error network as a backbone network, taking the marked area as a remarkable characteristic, and extracting the picture characteristics in the marked area in a layering manner;
the characteristic structure building module is used for building a characteristic pyramid of the picture characteristics;
and the identification module is used for carrying out full convolution network calculation on the characteristic pyramid to obtain an identification result.
Based on the above embodiment, as a preferred embodiment, the method further includes:
the model training module is used for acquiring a historical carriage image data set; labeling each image data in the historical carriage image data set, and determining the corresponding identification degree of the image data according to a preset identification degree classification table; and training all image data containing the recognition degree by utilizing a preset deep learning model to obtain the information recognition model.
Based on the above embodiment, as a preferred embodiment, the model training module includes:
the labeling unit is used for labeling each image data of the compartment image data set according to the content of a labeling area; the first target comprises a carriage number area, an attribute area and scheduled inspection period information, the second target comprises character information, and the third target comprises a carriage shaft and carriage type information.
Based on the above embodiment, as a preferred embodiment, the model training module includes:
and the identification degree determining unit is used for determining the identification degree corresponding to the image data according to the resolution of the image data and the number of the unidentifiable labeling areas.
Based on the above embodiment, as a preferred embodiment, the method may further include:
and the information counting module is used for inputting the information frame into an information identification model, and uploading the vehicle information to the upper computer after the vehicle information of the target train is obtained, so that the upper computer can compile a train information table according to the vehicle information.
The embodiments are described in a progressive manner in the specification, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system provided by the embodiment, the description is relatively simple because the system corresponds to the method provided by the embodiment, and the relevant points can be referred to the method part for description.
The principles and embodiments of the present application are explained herein using specific examples, which are provided only to help understand the method and the core idea of the present application. It should be noted that, for those skilled in the art, it is possible to make several improvements and modifications to the present application without departing from the principle of the present application, and such improvements and modifications also fall within the scope of the claims of the present application.
It is further noted that, in the present specification, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, 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 an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
Claims (7)
1. A method for identifying train information, comprising:
acquiring image data corresponding to a target train;
taking an area containing data content as a labeling area, and filtering the image data to obtain an information frame containing the labeling area;
inputting the information frame into an information identification model, taking a residual error network as a backbone network, taking the marked area as a remarkable characteristic, and extracting the picture characteristics in the marked area in a layering way;
constructing a feature pyramid of the picture features;
and carrying out full convolution network calculation on the characteristic pyramid to obtain the vehicle information of the target train.
2. The identification method according to claim 1, further comprising:
acquiring a historical carriage image data set;
labeling each image data in the historical carriage image data set, and determining the corresponding identification degree of the image data according to a preset identification degree classification table;
and training all image data containing the recognition degree by utilizing a preset deep learning model to obtain the information recognition model.
3. The identification method according to claim 2, wherein labeling each image data in the historical car image data set comprises:
labeling each image data in the historical carriage image data set according to the content of a labeling area; the first target comprises a carriage number area, an attribute area and scheduled inspection period information, the second target comprises character information, and the third target comprises a carriage shaft and carriage type information.
4. The identification method of claim 3, wherein determining the degree of identification corresponding to the image data according to the predetermined degree of identification classification table comprises:
and determining the corresponding recognition degree of the image data according to the resolution of the image data and the number of the unidentifiable labeling areas.
5. The identification method according to claim 1, wherein after inputting the information frame to an information identification model and obtaining the vehicle information of the target train, the method further comprises:
and uploading the vehicle information to an upper computer so that the upper computer can compile a train information table according to the vehicle information.
6. The identification method according to claim 1, wherein the vehicle information includes any one of a vehicle type, a compartment number, a load, a self weight, a volume, a length change, a scheduled inspection period, a running speed, and a running direction or a combination of any several items.
7. A train information recognition system, comprising:
the data acquisition module is used for acquiring image data corresponding to the target train;
the data processing module is used for taking an area containing data content as a labeling area, filtering the image data and obtaining an information frame containing the labeling area;
the characteristic extraction module is used for inputting the information frame into an information identification model, taking a residual error network as a backbone network, taking the marked area as a remarkable characteristic, and extracting the picture characteristics in the marked area in a layering manner;
the characteristic structure building module is used for building a characteristic pyramid of the picture characteristics;
and the identification module is used for carrying out full convolution network calculation on the characteristic pyramid to obtain an identification result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110934199.5A CN113371035B (en) | 2021-08-16 | 2021-08-16 | Train information identification method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110934199.5A CN113371035B (en) | 2021-08-16 | 2021-08-16 | Train information identification method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113371035A true CN113371035A (en) | 2021-09-10 |
| CN113371035B CN113371035B (en) | 2021-11-23 |
Family
ID=77577191
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110934199.5A Active CN113371035B (en) | 2021-08-16 | 2021-08-16 | Train information identification method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113371035B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113895481A (en) * | 2021-10-26 | 2022-01-07 | 卡斯柯信号有限公司 | Train positioning and tracking management method, equipment and medium based on pattern recognition |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005254930A (en) * | 2004-03-10 | 2005-09-22 | Mitsubishi Electric Corp | Device, system and method for monitoring unauthorized vehicle entering crossing |
| CN109190444A (en) * | 2018-07-02 | 2019-01-11 | 南京大学 | A kind of implementation method of the lane in which the drivers should pay fees vehicle feature recognition system based on video |
| CN110909745A (en) * | 2019-12-24 | 2020-03-24 | 新疆九洲熏蒸消毒有限责任公司 | Train disinfection channel identification system |
| CN110991447A (en) * | 2019-11-25 | 2020-04-10 | 北京交通大学 | Train number accurate positioning and identification method based on deep learning |
-
2021
- 2021-08-16 CN CN202110934199.5A patent/CN113371035B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005254930A (en) * | 2004-03-10 | 2005-09-22 | Mitsubishi Electric Corp | Device, system and method for monitoring unauthorized vehicle entering crossing |
| CN109190444A (en) * | 2018-07-02 | 2019-01-11 | 南京大学 | A kind of implementation method of the lane in which the drivers should pay fees vehicle feature recognition system based on video |
| CN110991447A (en) * | 2019-11-25 | 2020-04-10 | 北京交通大学 | Train number accurate positioning and identification method based on deep learning |
| CN110909745A (en) * | 2019-12-24 | 2020-03-24 | 新疆九洲熏蒸消毒有限责任公司 | Train disinfection channel identification system |
Non-Patent Citations (1)
| Title |
|---|
| 晏世武等: "基于卷积神经网络车型分类的研究", 《智能计算机与应用》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113895481A (en) * | 2021-10-26 | 2022-01-07 | 卡斯柯信号有限公司 | Train positioning and tracking management method, equipment and medium based on pattern recognition |
| CN113895481B (en) * | 2021-10-26 | 2024-01-23 | 卡斯柯信号有限公司 | Train positioning and tracking management method, device and medium based on pattern recognition |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113371035B (en) | 2021-11-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110660222B (en) | Intelligent environment-friendly electronic snapshot system for black-smoke road vehicle | |
| CN102759347B (en) | Online in-process quality control device and method for high-speed rail contact networks and composed high-speed rail contact network detection system thereof | |
| CN109784326A (en) | A kind of vehicle chassis detection method based on deep learning | |
| CN106469518A (en) | The method of the intelligent parking guiding based on video tracking identification, server and system | |
| US11010640B1 (en) | Automated training data quality process | |
| CN105259304B (en) | On-line monitoring system and method for pollutants in vehicle tail gas | |
| CN109815912B (en) | Highway safety inspection system based on artificial intelligence | |
| CN109993138A (en) | A kind of car plate detection and recognition methods and device | |
| CN204945122U (en) | A kind of Vehicular exhaust pollutant monitoring system | |
| CN111222375A (en) | Vehicle driving track generation method, system and server | |
| CN113371035B (en) | Train information identification method and system | |
| CN113788051A (en) | Train on-station running state monitoring and analyzing system | |
| CN113792578A (en) | Method, device and system for detecting abnormity of transformer substation | |
| CN109784161A (en) | Video monitoring intelligent analysis system and method in a kind of shop | |
| CN113378646A (en) | Freight train information identification system and identification method | |
| RU133082U1 (en) | IDENTIFICATION OF MOBILE RAILWAY UNITS | |
| KR102562757B1 (en) | Prediction and recognition method of road marking information and road maintenance method | |
| CN110263622A (en) | Train fire monitoring method, apparatus, terminal and storage medium | |
| CN115937765A (en) | Image identification method, AGV material sorting method and system | |
| CN115937591A (en) | Method and system for improving traffic scene target detection classification precision | |
| CN112949341B (en) | Information obtaining method and device, electronic equipment and storage medium | |
| CN212302511U (en) | Car number and case number identification association system monitoring facilities | |
| CN115439933A (en) | Garbage classification release site detection method based on multiple model recognition strategies | |
| CN114399671A (en) | Target identification method and device | |
| CN114241390A (en) | Railway foreign matter detection and identification method based on deep learning theory |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |