CN109271980A - A kind of vehicle nameplate full information recognition methods, system, terminal and medium - Google Patents
A kind of vehicle nameplate full information recognition methods, system, terminal and medium Download PDFInfo
- Publication number
- CN109271980A CN109271980A CN201810988374.7A CN201810988374A CN109271980A CN 109271980 A CN109271980 A CN 109271980A CN 201810988374 A CN201810988374 A CN 201810988374A CN 109271980 A CN109271980 A CN 109271980A
- Authority
- CN
- China
- Prior art keywords
- image
- critical field
- vehicle
- nameplate
- local map
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
A kind of vehicle nameplate full information recognition methods disclosed by the invention, obtains the nameplate image of vehicle to be detected；Using the critical field in the nameplate image of regional nerve Network Recognition vehicle to be detected；Using convolutional neural networks to the position coordinates for carrying out feature extraction, the Local map classified, orient each critical field in the nameplate image of vehicle to be detected；The Local map of each critical field and the Local map of each critical field corresponding field content are intercepted according to the position coordinates of the critical field；By the Local map of each critical field corresponding field content and the splicing of the Local map of critical field on an image, the critical field in stitching image is identified using region convolutional neural networks, identifies the corresponding character of the critical field in stitching image using convolutional neural networks.Using the Target Recognition Algorithms of deep learning, Text region accuracy is improved, improves Text region in natural scene, the robustness of the interference such as inclination, reflective, distortion.
Description
Technical field
The present invention relates to image identification technical fields, and in particular to a kind of vehicle nameplate full information recognition methods, system, end
End and medium.
Background technique
In imported vehicle Information Statistics business, need manually to acquire imported vehicle nameplate image, then manually by nameplate
In information of vehicles (such as Vehicle Identification Number, engine model, engine power, maximum autohrozed total mass, maximum seating capacity
Deng) it is manually entered electrical form, it is counted, labor workload is big.
Existing OCR character recognition technology mainly solves the optical imagery Text region of printing document class, such as expenses for medicine document, hair
Ticket etc. after carrying out layout to document class image based on certain format or domain knowledge, identifies text therein by element
Word information, it is not good enough for the natural image treatment effect of any format.Main reason is that existing character recognition method, needs
A large amount of domain knowledge and pre-treatment step are wanted, this considerably increases the error of algorithm, furthermore existing character recognition method is all
Learn on the basis of Manual definition's feature, ability to express is directly related with the superiority and inferiority of feature, and in natural scene
The variation such as distortion, inclination, illumination of text is sensitive, and robustness is insufficient.
Summary of the invention
For the defects in the prior art, the present invention provides a kind of vehicle nameplate full information recognition methods, using depth
The Target Recognition Algorithms of habit improve Text region accuracy, improve Text region in natural scene, inclination, reflective, torsion
The robustness of the interference such as song.
In a first aspect, a kind of vehicle nameplate full information recognition methods provided in an embodiment of the present invention, comprising:
The vehicle nameplate image of different formats is selected, and correcting is orthography, the Local map of critical field is single respectively
Solely storage, and the standard form image as each critical field；
Obtain the nameplate image of vehicle to be detected；
Using the critical field in the nameplate image of regional nerve Network Recognition vehicle to be detected；
Using convolutional neural networks to feature extraction and classification is carried out in the nameplate image of vehicle to be detected, orient each
The position coordinates of the Local map of critical field；
The Local map of each critical field is intercepted according to the position coordinates of the critical field and each critical field corresponds to
The Local map of field contents；
By the Local map of each critical field corresponding field content and the splicing of the Local map of critical field on an image,
Critical field in stitching image is gone out using regional nerve Network Recognition, is identified in stitching image using convolutional neural networks
The corresponding character of critical field.
Optionally, the convolutional neural networks are to progress feature extraction and classification in the nameplate image for obtaining vehicle to be detected
Specific method include:
Convolution, down-sampling processing are successively carried out to the nameplate image of vehicle to be detected, obtain the characteristics of image of semantic level；
The characteristics of image of the semantic level suggests net by region, region of interest is generated, by interest on convolution characteristic pattern
The corresponding characteristics of image in area inputs full articulamentum processing, obtains final image feature；
Final image feature is classified using classifier and returns device and tightens bounding box, obtains classification belonging to characteristics of image
With original image coordinate.
Optionally, the Local map of each critical field and each is intercepted in the position coordinates according to the critical field
After the Local map step of critical field corresponding field content, the Local map and keyword of each critical field corresponding field content
The Local map of section splices before the step on an image further include: corrects the Local map of each critical field for orthogonal projection
Picture.
Optionally, the Local map by each critical field correct the specific method for orthography include: will be each
The Local map of critical field carries out Feature Points Matching with the standard form image of corresponding critical field, rejects invalid characteristic point
It is right, using remaining characteristic point to radiation transformation correction is carried out, obtain the orthography after each critical field is corrected.
Second aspect, a kind of vehicle nameplate full information identifying system provided in an embodiment of the present invention, including standard form figure
As unit, image acquisition unit, regional nerve network unit, convolutional neural networks unit, image interception unit and image mosaic
Unit,
The standard form elementary area is configured for storing the mark of the critical field of different format vehicle nameplate images
Quasi- template image；
Described image acquiring unit is configurable for obtaining the nameplate image of vehicle to be detected；
The regional nerve network unit is configured as identifying the critical field in the nameplate image of vehicle to be detected；
The convolutional neural networks unit is configured as carrying out feature extraction in the nameplate image to vehicle to be detected, divide
Class orients the position coordinates of the Local map of each critical field；
Described image interception unit is configured as intercepting each critical field according to the position coordinates of the critical field
The Local map of Local map and each critical field corresponding field content；
Image mosaic unit is configured as the office by the Local map of each critical field corresponding field content and critical field
Portion's figure splices on an image；
The regional nerve network unit is configured as identifying the critical field in stitching image；
The convolutional neural networks unit is configured as identifying the corresponding character of critical field in stitching image.
Optionally, the convolutional neural networks unit to carried out in the nameplate image for obtaining vehicle to be detected feature extraction and
The specific method of classification includes:
Convolution, down-sampling processing are successively carried out to the nameplate image of vehicle to be detected, obtain the characteristics of image of semantic level；
The characteristics of image of the semantic level suggests net by region, region of interest is generated, by interest on convolution characteristic pattern
The corresponding characteristics of image in area inputs full articulamentum processing, obtains final image feature；
Final image feature is classified using classifier and returns device and tightens bounding box, obtains classification belonging to characteristics of image
With original image coordinate.
Optionally, the system also includes image flame detection unit, described image correcting unit is configured as each key
The Local map correction of field is orthography.
Optionally, the specific method that described image correcting unit is corrected the Local map of each critical field as orthography
Include: that the Local map of each critical field is subjected to Feature Points Matching with the standard form image of corresponding critical field, rejects
Invalid characteristic point pair is obtained after each critical field is corrected just using remaining characteristic point to radiation transformation correction is carried out
Projection picture.
The third aspect, the embodiment of the present invention also provide a kind of intelligent terminal, including processor, input equipment, output equipment
And memory, the processor, input equipment, output equipment and memory are connected with each other, the memory is calculated for storing
Machine program, the computer program include program instruction, and the processor is configured for that described program is called to instruct, in execution
State method.
Fourth aspect, the embodiment of the present invention also provide a kind of computer readable storage medium, the computer storage medium
It is stored with computer program, the computer program includes program instruction, and described program instruction makes institute when being executed by a processor
It states processor and executes the above method.
Beneficial effects of the present invention:
A kind of vehicle nameplate full information recognition methods provided in an embodiment of the present invention, using critical field positioning, keyword
Section is known realizes that the nameplate image for different brands, different type-setting modes, different fonts identifies otherwise.It is rectified by image
Just with the Target Recognition Algorithms of deep learning, Text region is improved in natural scene, the interference such as inclination, reflective, distortion
Robustness.Deep neural network extracts image, semantic grade feature, and using area neural network RCNN framework identifies single character, quasi-
True rate is substantially improved.
A kind of vehicle nameplate full information identifying system provided in an embodiment of the present invention, using critical field positioning, keyword
Section is known realizes that the nameplate image for different brands, different type-setting modes, different fonts identifies otherwise.It is rectified by image
Just with the Target Recognition Algorithms of deep learning, Text region is improved in natural scene, the interference such as inclination, reflective, distortion
Robustness.Deep neural network extracts image, semantic grade feature, and using area neural network RCNN framework identifies single character, quasi-
True rate is substantially improved.
Detailed description of the invention
It, below will be to specific in order to illustrate more clearly of the specific embodiment of the invention or technical solution in the prior art
Embodiment or attached drawing needed to be used in the description of the prior art are briefly described.In all the appended drawings, similar element
Or part is generally identified by similar appended drawing reference.In attached drawing, each element or part might not be drawn according to actual ratio.
Fig. 1 shows a kind of flow chart of vehicle nameplate full information recognition methods first embodiment provided by the present invention；
Fig. 2 shows a kind of structural representations of vehicle nameplate full information identifying system first embodiment provided by the present invention
Figure；
Fig. 3 shows a kind of schematic structural diagram of the first embodiment of intelligent terminal provided by the present invention.
Specific embodiment
It is described in detail below in conjunction with embodiment of the attached drawing to technical solution of the present invention.Following embodiment is only used for
Clearly illustrate technical solution of the present invention, therefore be intended only as example, and cannot be used as a limitation and limit protection of the invention
Range.
It should be noted that unless otherwise indicated, technical term or scientific term used in this application should be this hair
The ordinary meaning that bright one of ordinary skill in the art are understood.
As shown in Figure 1, showing a kind of first embodiment of vehicle nameplate full information recognition methods provided by the present invention
Flow chart, this method comprises:
S11: selecting the vehicle nameplate image of different formats, and correcting is orthography, by the Local map of critical field point
It does not store not individually, and the standard form image as each critical field.
Specifically, in existing vehicle nameplate image, the image of every kind of format selects one, and vehicle nameplate image is rectified
Just it is orthography, and extracts critical field, the Local map of critical field is individually stored, the mark of each critical field is stored as
Quasi- template image.Information in vehicle nameplate image includes: vehicle manufacture producer, Vehicle Identification Number, brand, engine row
Amount, vehicle model, engine model, manufacture days, maximum autohrozed total mass, seating capacity, engine maximum net power, manufacture
State and color code and its corresponding information.By Vehicle Identification Number, brand, engine displacement, vehicle model, engine models
Number, manufacture days, maximum autohrozed total mass, seating capacity, engine maximum net power, manufacturing nation and color code be as information
The critical field of identification, need to identify is the corresponding information of these critical fielies.
S12: the nameplate image of vehicle to be detected is obtained.
Specifically, obtain the nameplate image of vehicle to be detected by way of taking pictures, or by data transfer mode or
Other modes obtain the nameplate image of existing vehicle to be detected.
S13: using the critical field in the nameplate image of regional nerve Network Recognition vehicle to be detected.Using RCNN
(Regions with CNN features) target object detecting method detects the keyword in the nameplate image of vehicle to be detected
Section.
S14: using convolutional neural networks to carrying out feature extraction in the nameplate image of vehicle to be detected, classify, orient
The position coordinates of the Local map of each critical field.
Specifically, convolutional neural networks are to progress feature extraction, classification, positioning in the nameplate image for obtaining vehicle to be detected
The specific method of the position coordinates of the Local map of each critical field includes: out
Convolution, down-sampling processing are successively carried out to the nameplate image of vehicle to be detected, obtain the characteristics of image of semantic level；
The characteristics of image of the semantic level suggests net by region, region of interest is generated, by interest on convolution characteristic pattern
The corresponding characteristics of image in area inputs full articulamentum processing, obtains final image feature；
Final image feature is classified using classifier and returns device and tightens bounding box, obtains classification belonging to characteristics of image
With original image coordinate.
S15: the Local map of each critical field is intercepted according to the position coordinates of critical field and each critical field corresponds to
The Local map of field contents.
Specifically, in order to enable the Local map of each critical field preferably to identify, by the office of each critical field
Figure correction in portion's is orthography.By the Local map of each critical field correct the specific method for orthography include: will be each
The Local map of critical field carries out Feature Points Matching with the standard form image of corresponding critical field, rejects invalid characteristic point
It is right, using remaining characteristic point to radiation transformation correction is carried out, obtain the orthography after each critical field is corrected.
S16: the Local map of each critical field corresponding field content and the splicing of the Local map of critical field are schemed at one
As upper, the critical field in stitching image is gone out using regional nerve Network Recognition, identifies spliced map using convolutional neural networks
The corresponding character of critical field as in.
A kind of vehicle nameplate full information recognition methods provided in an embodiment of the present invention, using critical field positioning, keyword
Section is known realizes that the nameplate image for different brands, different type-setting modes, different fonts identifies otherwise.It is rectified by image
Just with the Target Recognition Algorithms of deep learning, Text region is improved in natural scene, the interference such as inclination, reflective, distortion
Robustness.Deep neural network extracts image, semantic grade feature, and using area neural network RCNN framework identifies single character, quasi-
True rate is substantially improved.
As shown in Fig. 2, the structure for showing a kind of vehicle nameplate full information identifying system provided in an embodiment of the present invention is shown
It is intended to, which includes standard form elementary area 21, image acquisition unit 22, regional nerve network unit 23, convolutional Neural
Network unit 24, image interception unit 25 and image mosaic unit 26, the standard form elementary area 21 are configured for depositing
Store up the standard form image of the critical field of different format vehicle nameplate images；Described image acquiring unit 22 is configurable for
Obtain the nameplate image of vehicle to be detected；The regional nerve network unit 23 is configured as identifying the nameplate figure of vehicle to be detected
Critical field as in；The convolutional neural networks unit 24 is configured as carrying out feature in the nameplate image to vehicle to be detected
It extracts, classification, orients the position coordinates of the Local map of each critical field.Convolutional neural networks unit is to acquisition measuring car to be checked
Nameplate image in carry out feature extraction, classification and positioning specific method include: to the nameplate image of vehicle to be detected according to
Secondary progress convolution, down-sampling processing, obtain the characteristics of image of semantic level；The characteristics of image of the semantic level is built by region
Net is discussed, region of interest is generated, the corresponding characteristics of image of region of interest is inputted into full articulamentum on convolution characteristic pattern and is handled, is obtained final
Characteristics of image；Final image feature is classified using classifier and returns device and tightens bounding box, obtains class belonging to characteristics of image
Other and original image coordinate.Described image interception unit 25 is configured as intercepting each key according to the position coordinates of the critical field
The Local map of the Local map of field and each critical field corresponding field content；Image mosaic unit 26 is configured as each pass
The Local map of key field corresponding field content and the Local map of critical field splice on an image；The regional nerve network
Unit 23 is configured as identifying the critical field in stitching image；The convolutional neural networks unit 24 is configured as identifying
The corresponding character of critical field in stitching image.
Further, system further includes image flame detection unit 27, and described image correcting unit 27 is configured as each pass
The Local map correction of key field is orthography.The Local map correction of each critical field is positive and penetrates by described image correcting unit
The specific method of image includes: that the Local map of each critical field is carried out spy with the standard form image of corresponding critical field
Sign point matching, rejects invalid characteristic point pair, using remaining characteristic point to radiation transformation correction is carried out, obtains each keyword
Orthography after section correction.
It is stored with the critical field standard form image of different format vehicle nameplate images in standard form elementary area, leads to
The nameplate image that image acquisition unit obtains vehicle to be detected is crossed, nameplate image information includes man of Railway Car Plant, vehicle identification
Code name, brand, engine displacement, vehicle model, engine model, manufacture days, maximum autohrozed total mass, seating capacity, hair
Motivation maximum net power, manufacturing nation and color code and its corresponding information.By Vehicle Identification Number, brand, engine displacement,
Vehicle model, engine model, manufacture days, maximum autohrozed total mass, seating capacity, engine maximum net power, manufacturing nation
The critical field identified with color code as information, need to identify is the corresponding information of these critical fielies.Region mind
The critical field in the nameplate image of vehicle to be detected is identified through network unit, convolutional neural networks unit is to vehicle to be detected
Nameplate image successively carry out convolution, down-sampling processing, obtain the characteristics of image of semantic level；The image of the semantic level is special
Sign suggests net by region, generates region of interest, and the corresponding characteristics of image of region of interest is inputted full articulamentum on convolution characteristic pattern
Processing, obtains final image feature；Final image feature is classified using classifier and returns device and tightens bounding box, obtains image
Classification belonging to feature and critical field Local map original image coordinate.Image interception unit is intercepted according to the position coordinates of critical field
The Local map of the Local map of each critical field and each critical field corresponding field content.Image flame detection unit is by each key
The Local map of field carries out Feature Points Matching with the standard form image of corresponding critical field, rejects invalid characteristic point pair,
Using remaining characteristic point to radiation transformation correction is carried out, the orthography after each critical field is corrected is obtained.Image mosaic
Unit is by the Local map of each critical field corresponding field content and the splicing of the Local map of critical field on an image.Region
Neural network unit identifies the critical field in stitching image, and convolutional neural networks unit identifies the key in stitching image
The corresponding character value of field.
A kind of vehicle nameplate full information identifying system provided in an embodiment of the present invention, using critical field positioning, keyword
Section is known realizes that the nameplate image for different brands, different type-setting modes, different fonts identifies otherwise.It is rectified by image
Just with the Target Recognition Algorithms of deep learning, Text region is improved in natural scene, the interference such as inclination, reflective, distortion
Robustness.Deep neural network extracts image, semantic grade feature, and using area neural network RCNN framework identifies single character, quasi-
True rate is substantially improved.
The present invention also provides a kind of first embodiments of intelligent terminal, as shown in figure 3, the structure for showing intelligent terminal is shown
It is intended to, which includes processor 31, input equipment 32, output equipment 33 and memory 34, the processor 31, input equipment
32, output equipment 33 and memory 34 are connected with each other, and the memory 34 is for storing computer program, the computer program
Including program instruction, the processor 31 is configured for calling described program instruction, the method for executing above-described embodiment description.
It should be appreciated that in embodiments of the present invention, alleged processor 31 can be central processing unit (Central
Processing Unit, CPU), which can also be other general processors, digital signal processor (Digital
Signal Processor, DSP), specific integrated circuit (Application Specific Integrated Circuit,
ASIC), ready-made programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic
Device, discrete gate or transistor logic, discrete hardware components etc..General processor can be microprocessor or this at
Reason device is also possible to any conventional processor etc..
Input equipment 32 may include that Trackpad, fingerprint adopt sensor (for acquiring the finger print information and fingerprint of user
Directional information), microphone etc., output equipment 33 may include display (LCD etc.), loudspeaker etc..
The memory 34 may include read-only memory and random access memory, and provide instruction sum number to processor 31
According to.The a part of of memory 34 can also include nonvolatile RAM.It is set for example, memory 34 can also store
The information of standby type.
In the specific implementation, processor 31 described in the embodiment of the present invention, input equipment 32, output equipment 33 are executable
System described in the embodiment of the present invention also can be performed in implementation described in embodiment of the method provided in an embodiment of the present invention
The implementation of embodiment, details are not described herein.
The present invention also provides a kind of embodiment of computer readable storage medium, the computer storage medium is stored with
Computer program, the computer program include program instruction, and described program instruction makes the processing when being executed by a processor
The method that device holds above-described embodiment description.
The computer readable storage medium can be the internal storage unit of terminal described in previous embodiment, such as eventually
The hard disk or memory at end.The computer readable storage medium is also possible to the External memory equipment of the terminal, such as described
The plug-in type hard disk being equipped in terminal, intelligent memory card (Smart Media Card, SMC), secure digital (Secure
Digital, SD) card, flash card (Flash Card) etc..Further, the computer readable storage medium can also be wrapped both
The internal storage unit for including the terminal also includes External memory equipment.The computer readable storage medium is described for storing
Other programs and data needed for computer program and the terminal.The computer readable storage medium can be also used for temporarily
When store the data that has exported or will export.
Those of ordinary skill in the art may be aware that list described in conjunction with the examples disclosed in the embodiments of the present disclosure
Member and algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware
With the interchangeability of software, each exemplary composition and step are generally described according to function in the above description.This
A little functions are implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Specially
Industry technical staff can use different methods to achieve the described function each specific application, but this realization is not
It is considered as beyond the scope of this invention.
It is apparent to those skilled in the art that for convenience of description and succinctly, the end of foregoing description
The specific work process at end and unit, can refer to corresponding processes in the foregoing method embodiment, details are not described herein.
In several embodiments provided herein, it should be understood that disclosed system, terminal and method, Ke Yitong
Other modes are crossed to realize.For example, system embodiment described above is only schematical, for example, the unit is drawn
Point, only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components can
To combine or be desirably integrated into another system, or some features can be ignored or not executed.In addition, shown or discussed
Mutual coupling, direct-coupling or communication connection can be through some interfaces, the INDIRECT COUPLING of device or unit or
Communication connection is also possible to electricity, mechanical or other form connections.
Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention., rather than its limitations；To the greatest extent
Pipe present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: its according to
So be possible to modify the technical solutions described in the foregoing embodiments, or to some or all of the technical features into
Row equivalent replacement；And these are modified or replaceed, various embodiments of the present invention technology that it does not separate the essence of the corresponding technical solution
The range of scheme should all cover within the scope of the claims and the description of the invention.
Claims (10)
1. a kind of vehicle nameplate full information recognition methods characterized by comprising
The vehicle nameplate image of different formats is selected, and correcting is orthography, and the Local map of critical field is individually deposited
Storage, and the standard form image as each critical field；
Obtain the nameplate image of vehicle to be detected；
Using the critical field in the nameplate image of regional nerve Network Recognition vehicle to be detected；
Using convolutional neural networks to carrying out feature extraction in the nameplate image of vehicle to be detected, classify, orient each key
The position coordinates of the Local map of field；
According to the position coordinates of the critical field intercept each critical field Local map and each critical field corresponding field
The Local map of content；
By the Local map of each critical field corresponding field content and the splicing of the Local map of critical field on an image, use
Region convolutional neural networks identify the critical field in stitching image, are identified in stitching image using convolutional neural networks
The corresponding character of critical field.
2. vehicle nameplate full information recognition methods as described in claim 1, which is characterized in that the convolutional neural networks are to obtaining
Take in the nameplate image of vehicle to be detected carry out feature extraction, classification and positioning specific method include:
Convolution, down-sampling processing are successively carried out to the nameplate image of vehicle to be detected, obtain the characteristics of image of semantic level；
The characteristics of image of the semantic level suggests net by region, region of interest is generated, by region of interest pair on convolution characteristic pattern
The characteristics of image answered inputs full articulamentum processing, obtains final image feature；
Final image feature is classified using classifier and returns device and tightens bounding box, obtains classification belonging to characteristics of image and original
Figure coordinate.
3. vehicle nameplate full information recognition methods as described in claim 1, which is characterized in that described according to the keyword
Section position coordinates intercept each critical field Local map and each critical field corresponding field content Local map step it
Afterwards, step on an image of the Local map of each critical field corresponding field content and the splicing of the Local map of critical field it
Before further include: the Local map of each critical field is corrected as orthography.
4. vehicle nameplate full information recognition methods as claimed in claim 3, which is characterized in that described by each critical field
Local map correction is that the specific method of orthography includes: by the mark of the Local map of each critical field and corresponding critical field
Quasi- template image carries out Feature Points Matching, rejects invalid characteristic point pair, is rectified using remaining characteristic point to radiation transformation is carried out
Just, the orthography after obtaining each critical field correction.
5. a kind of vehicle nameplate full information identifying system, which is characterized in that obtained including standard form elementary area, image single
Member, regional nerve network unit, convolutional neural networks unit, image interception unit and image mosaic unit,
The standard form elementary area is configured for storing the master die of the critical field of different format vehicle nameplate images
Plate image；
Described image acquiring unit is configurable for obtaining the nameplate image of vehicle to be detected；
The regional nerve network unit is configured as identifying the critical field in the nameplate image of vehicle to be detected；
The convolutional neural networks unit is configured as carrying out feature extraction in the nameplate image to vehicle to be detected, classification, determine
Position goes out the position coordinates of the Local map of each critical field；
Described image interception unit is configured as intercepting the part of each critical field according to the position coordinates of the critical field
The Local map of figure and each critical field corresponding field content；
Image mosaic unit is configured as the Local map by the Local map of each critical field corresponding field content and critical field
Splicing is on an image；
The regional nerve network unit is configured as identifying the critical field in stitching image；
The convolutional neural networks unit is configured as identifying the corresponding character of critical field in stitching image.
6. vehicle nameplate full information identifying system as claimed in claim 5, which is characterized in that the convolutional neural networks unit
Include: to the specific method for carrying out feature extraction, classification and positioning in the nameplate image for obtaining vehicle to be detected
Convolution, down-sampling processing are successively carried out to the nameplate image of vehicle to be detected, obtain the characteristics of image of semantic level；
The characteristics of image of the semantic level suggests net by region, region of interest is generated, by region of interest pair on convolution characteristic pattern
The characteristics of image answered inputs full articulamentum processing, obtains final image feature；
Final image feature is classified using classifier and returns device and tightens bounding box, obtains classification belonging to characteristics of image and original
Figure coordinate.
7. vehicle nameplate full information identifying system as claimed in claim 5, which is characterized in that the system also includes images to rectify
Positive unit, described image correcting unit are configured as correcting the Local map of each critical field for orthography.
8. vehicle nameplate full information identifying system as claimed in claim 7, which is characterized in that described image correcting unit will be each
A critical field Local map correction be orthography specific method include: by the Local map of each critical field with it is corresponding
The standard form image of critical field carries out Feature Points Matching, rejects invalid characteristic point pair, using remaining characteristic point into
Row radiation transformation correction, the orthography after obtaining each critical field correction.
9. a kind of intelligent terminal, including processor, input equipment, output equipment and memory, the processor, input equipment,
Output equipment and memory are connected with each other, and for the memory for storing computer program, the computer program includes program
Instruction, which is characterized in that the processor is configured for calling described program instruction, executes such as any one of claim 1-4
The method.
10. a kind of computer readable storage medium, which is characterized in that the computer storage medium is stored with computer program,
The computer program includes program instruction, and described program instruction makes the processor execute such as right when being executed by a processor
It is required that the described in any item methods of 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810988374.7A CN109271980A (en) | 2018-08-28 | 2018-08-28 | A kind of vehicle nameplate full information recognition methods, system, terminal and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810988374.7A CN109271980A (en) | 2018-08-28 | 2018-08-28 | A kind of vehicle nameplate full information recognition methods, system, terminal and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109271980A true CN109271980A (en) | 2019-01-25 |
Family
ID=65154501
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810988374.7A Pending CN109271980A (en) | 2018-08-28 | 2018-08-28 | A kind of vehicle nameplate full information recognition methods, system, terminal and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109271980A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110188755A (en) * | 2019-05-30 | 2019-08-30 | 北京百度网讯科技有限公司 | A kind of method, apparatus and computer readable storage medium of image recognition |
| CN110348463A (en) * | 2019-07-16 | 2019-10-18 | 北京百度网讯科技有限公司 | The method and apparatus of vehicle for identification |
| CN110569801A (en) * | 2019-09-10 | 2019-12-13 | 四川星盾科技股份有限公司 | method for identifying key content of driving license |
| CN110705547A (en) * | 2019-09-06 | 2020-01-17 | 中国平安财产保险股份有限公司 | Method and device for recognizing characters in image and computer readable storage medium |
| CN111160342A (en) * | 2019-12-30 | 2020-05-15 | 上海眼控科技股份有限公司 | Vehicle reflective mark verification method and device, computer equipment and storage medium |
| CN111667408A (en) * | 2020-06-02 | 2020-09-15 | 中车青岛四方机车车辆股份有限公司 | Vehicle image processing method and device, storage medium and processor |
| WO2021238855A1 (en) * | 2020-05-28 | 2021-12-02 | 同方威视技术股份有限公司 | Vehicle inspection method and system |
| CN114267038A (en) * | 2022-03-03 | 2022-04-01 | 南京甄视智能科技有限公司 | Nameplate type identification method and device, storage medium and equipment |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110106823A1 (en) * | 2009-11-02 | 2011-05-05 | Harry Urbschat | System and method of using dynamic variance networks |
| US20130287284A1 (en) * | 2008-01-18 | 2013-10-31 | Mitek Systems | Systems and methods for classifying payment documents during mobile image processing |
| US20170068866A1 (en) * | 2015-09-07 | 2017-03-09 | Abbyy Development Llc | Method and system for data extraction from images of semi-structured documents |
| CN106530528A (en) * | 2016-10-11 | 2017-03-22 | 上海慧银信息科技有限公司 | Cash voucher information identification method and cash voucher information identification device |
| CN107766809A (en) * | 2017-10-09 | 2018-03-06 | 平安科技（深圳）有限公司 | Electronic installation, billing information recognition methods and computer-readable recording medium |
| CN107798299A (en) * | 2017-10-09 | 2018-03-13 | 平安科技（深圳）有限公司 | Billing information recognition methods, electronic installation and readable storage medium storing program for executing |
| CN107808154A (en) * | 2017-12-08 | 2018-03-16 | 上海慧银信息科技有限公司 | The method and apparatus for extracting cash register billing information |
| CN107862303A (en) * | 2017-11-30 | 2018-03-30 | 平安科技（深圳）有限公司 | Information identifying method, electronic installation and the readable storage medium storing program for executing of form class diagram picture |
| CN107977665A (en) * | 2017-12-15 | 2018-05-01 | 北京科摩仕捷科技有限公司 | The recognition methods of key message and computing device in a kind of invoice |
| CN108229463A (en) * | 2018-02-07 | 2018-06-29 | 众安信息技术服务有限公司 | Character recognition method based on image |
| US20180239989A1 (en) * | 2017-02-20 | 2018-08-23 | Alibaba Group Holding Limited | Type Prediction Method, Apparatus and Electronic Device for Recognizing an Object in an Image |
-
2018
- 2018-08-28 CN CN201810988374.7A patent/CN109271980A/en active Pending
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130287284A1 (en) * | 2008-01-18 | 2013-10-31 | Mitek Systems | Systems and methods for classifying payment documents during mobile image processing |
| US20110106823A1 (en) * | 2009-11-02 | 2011-05-05 | Harry Urbschat | System and method of using dynamic variance networks |
| US20170068866A1 (en) * | 2015-09-07 | 2017-03-09 | Abbyy Development Llc | Method and system for data extraction from images of semi-structured documents |
| CN106530528A (en) * | 2016-10-11 | 2017-03-22 | 上海慧银信息科技有限公司 | Cash voucher information identification method and cash voucher information identification device |
| US20180239989A1 (en) * | 2017-02-20 | 2018-08-23 | Alibaba Group Holding Limited | Type Prediction Method, Apparatus and Electronic Device for Recognizing an Object in an Image |
| CN107766809A (en) * | 2017-10-09 | 2018-03-06 | 平安科技（深圳）有限公司 | Electronic installation, billing information recognition methods and computer-readable recording medium |
| CN107798299A (en) * | 2017-10-09 | 2018-03-13 | 平安科技（深圳）有限公司 | Billing information recognition methods, electronic installation and readable storage medium storing program for executing |
| CN107862303A (en) * | 2017-11-30 | 2018-03-30 | 平安科技（深圳）有限公司 | Information identifying method, electronic installation and the readable storage medium storing program for executing of form class diagram picture |
| CN107808154A (en) * | 2017-12-08 | 2018-03-16 | 上海慧银信息科技有限公司 | The method and apparatus for extracting cash register billing information |
| CN107977665A (en) * | 2017-12-15 | 2018-05-01 | 北京科摩仕捷科技有限公司 | The recognition methods of key message and computing device in a kind of invoice |
| CN108229463A (en) * | 2018-02-07 | 2018-06-29 | 众安信息技术服务有限公司 | Character recognition method based on image |
Non-Patent Citations (2)
| Title |
|---|
| SHAOQING REN 等: "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks", 《IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE》 * |
| 谢志钢: "面向增值税发票的图像自动处理技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110188755A (en) * | 2019-05-30 | 2019-08-30 | 北京百度网讯科技有限公司 | A kind of method, apparatus and computer readable storage medium of image recognition |
| CN110348463A (en) * | 2019-07-16 | 2019-10-18 | 北京百度网讯科技有限公司 | The method and apparatus of vehicle for identification |
| CN110705547A (en) * | 2019-09-06 | 2020-01-17 | 中国平安财产保险股份有限公司 | Method and device for recognizing characters in image and computer readable storage medium |
| CN110705547B (en) * | 2019-09-06 | 2023-08-18 | 中国平安财产保险股份有限公司 | Method and device for recognizing text in image and computer readable storage medium |
| CN110569801A (en) * | 2019-09-10 | 2019-12-13 | 四川星盾科技股份有限公司 | method for identifying key content of driving license |
| CN111160342A (en) * | 2019-12-30 | 2020-05-15 | 上海眼控科技股份有限公司 | Vehicle reflective mark verification method and device, computer equipment and storage medium |
| WO2021238855A1 (en) * | 2020-05-28 | 2021-12-02 | 同方威视技术股份有限公司 | Vehicle inspection method and system |
| CN111667408A (en) * | 2020-06-02 | 2020-09-15 | 中车青岛四方机车车辆股份有限公司 | Vehicle image processing method and device, storage medium and processor |
| CN111667408B (en) * | 2020-06-02 | 2023-06-02 | 中车青岛四方机车车辆股份有限公司 | Vehicle image processing method and device, storage medium and processor |
| CN114267038A (en) * | 2022-03-03 | 2022-04-01 | 南京甄视智能科技有限公司 | Nameplate type identification method and device, storage medium and equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109271980A (en) | A kind of vehicle nameplate full information recognition methods, system, terminal and medium | |
| CN107798299B (en) | Bill information identification method, electronic device and readable storage medium | |
| CN111931664B (en) | Mixed-pasting bill image processing method and device, computer equipment and storage medium | |
| CN110298338B (en) | Document image classification method and device | |
| CN110263694A (en) | A kind of bank slip recognition method and device | |
| CN110874618B (en) | OCR template learning method and device based on small sample, electronic equipment and medium | |
| CN109934181A (en) | Text recognition method, device, equipment and computer-readable medium | |
| CN110705952A (en) | Contract auditing method and device | |
| CN112699775A (en) | Certificate identification method, device and equipment based on deep learning and storage medium | |
| CN109948549A (en) | OCR data creation method, device, computer equipment and storage medium | |
| CN113011144A (en) | Form information acquisition method and device and server | |
| CN112668575B (en) | Key information extraction method and device, electronic equipment and storage medium | |
| CN112396047B (en) | Training sample generation method and device, computer equipment and storage medium | |
| CN112581344A (en) | Image processing method and device, computer equipment and storage medium | |
| CN113012075A (en) | Image correction method and device, computer equipment and storage medium | |
| CN114332883A (en) | Invoice information identification method and device, computer equipment and storage medium | |
| CN111462388A (en) | Bill inspection method and device, terminal equipment and storage medium | |
| CN114202759A (en) | Multi-currency paper currency crown word number identification method and device based on deep learning | |
| CN110378886A (en) | Image comparison method, image comparison device, electronic equipment and medium | |
| US11030488B1 (en) | Book scanning using machine-trained model | |
| CN113486715A (en) | Image reproduction identification method, intelligent terminal and computer storage medium | |
| CN112560855A (en) | Image information extraction method and device, electronic equipment and storage medium | |
| CN110717732A (en) | Information authentication method and system | |
| CN112396057A (en) | Character recognition method and device and electronic equipment | |
| US20220207668A1 (en) | Book scanning using machine-trained model |
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 | ||
| AD01 | Patent right deemed abandoned | ||
| AD01 | Patent right deemed abandoned |
Effective date of abandoning: 20210730 |