CN106652167A - FPGA-based RMB crown word identification system - Google Patents
FPGA-based RMB crown word identification system Download PDFInfo
- Publication number
- CN106652167A CN106652167A CN201611165429.1A CN201611165429A CN106652167A CN 106652167 A CN106652167 A CN 106652167A CN 201611165429 A CN201611165429 A CN 201611165429A CN 106652167 A CN106652167 A CN 106652167A
- Authority
- CN
- China
- Prior art keywords
- image
- prefix
- rmb
- module
- submodule
- 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
- G07—CHECKING-DEVICES
- G07D—HANDLING OF COINS OR VALUABLE PAPERS, e.g. TESTING, SORTING BY DENOMINATIONS, COUNTING, DISPENSING, CHANGING OR DEPOSITING
- G07D7/00—Testing specially adapted to determine the identity or genuineness of valuable papers or for segregating those which are unacceptable, e.g. banknotes that are alien to a currency
- G07D7/20—Testing patterns thereon
- G07D7/2016—Testing patterns thereon using feature extraction, e.g. segmentation, edge detection or Hough-transformation
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an FPGA-based RMB crown word identification system, aiming to solve the problem of low integration and low speed in the prior art. The FPGA-based RMB crown word identification system includes an upper image sensor, a lower image sensor (1), two A/D converters (2), an FPGA unit (3), two drivers (4), an SDRAM memory (5), an FPGA external expansion USB interface (6), and an FPGA external expansion network interface (7); when the FPGA detects the rotation of the motor, the light sources of the image sensors are turned on through the drivers, the analog image signals output by the image sensors are transmitted to the FPGA after the A/D conversion, and the FPGA completes the RMB image data acquisition, crown word image search, crown word image identification, and outputs the identification results to an external device. The FPGA unit is taken as a main processor, the hardware structure is simple, the processing speed is fast, and the identification system can be applied to a class-A currency count machine and an RMB sorting machine.
Description
Technical field
The invention belongs to digital processing technology field, specifically a kind of RMB serial number identification system based on FPGA
System, can be used for A classes paper money counter and RMB cleaning-sorting machine.
Technical background
The hat number and numeral numbering of RMB is for recording note-issue's sequence, because serial number is with unique
Property, therefore the number of every RMB is recorded when RMB goes out and puts in storage, and coin code archives are set up, financial department etc. can be made effectively
Tracking RMB is effectively supervised in social current intelligence, so as to advantageously account for similar robbery armoured van, money laundering etc.
Social Events.Require according to national standard, the financial industry such as bank must be equipped with serial number extraction, identification function
A class paper money counters.
China's serial number is extracted and technology of identification is started late.But recent years quickly grows.Wherein great majority are bases
In the image acquisition and prefix code identifying system of contact-type image sensor.Identifying system is substantially based on DSP or special place
The prefix identifying schemes of reason chip, these methods can meet current identification requirement substantially, but these systems or recognition speed
It is relatively low, or level of integrated system is poor, is unfavorable for the miniaturization of equipment and the requirement of high integration.
Chengdu Shu You Science and Technology Ltd.s apply for a patent the " extraction of high speed RMB serial number and knowledge based on DSP at it
Disclose a kind of based on FPGA+DSP's in other system " (publication number CN103729934A, application number CN201410004812.3)
System architecture, this system architecture method completes image acquisition using FPGA, and using DSP as primary processor image procossing is completed
With the related work of image recognition.Due to DSP can only serial complete data processing, therefore have in processing speed natural bad
Gesture.
Ningbo Shu You Science and Technology Ltd.s apply for a patent a kind of " high speed RMB serial number identifying system and in fact at it
Existing method " (publication number CN104103119A, application number CN201410313854.5) disclose using USB3.0 control chips+
The system architecture of FPGA, this framework method completes image acquisition and serial number positioning using FPGA, is controlled using USB3.0
Chip solidification code completes image recognition work.This system architecture method using USB3.0 chips complete high rate burst communication and
Evaluation work, is an innovation relative to the processing scheme based on DSP.Shortcoming is to use USB3.0 chips to equipment interface
A considered critical is done, the application scenario of USB interface is may be only available for, and the integrated level of system is not high, poor expandability.
Technical scheme
It is an object of the invention to provide a kind of RMB prefix identifying system based on FPGA, mainly solves existing people
People's coin serial number is extracted and identifying system has that low level of integrated system, high cost, processing speed are slow.
For achieving the above object, RMB prefix identifying system of the present invention based on FPGA, including upper and lower two images is passed
USB interface and FPGA that sensor, two-way A/D converter, FPGA unit, two-way driver, SDRAM memory, FPGA are extended out
The network interface for extending out, imageing sensor is contact-type image sensor, and A/D converter turns the analogue signal of imageing sensor
For digital signal, the light source of driver control imageing sensor, it is characterised in that:
FPGA unit includes:
Control module:For controlling opening or the closing of driver;
Image capture module:It is connected with A/D converter, for completing the collection of digital picture;
Sdram interface module:For realizing sdram interface control logic, the image that image capture module is gathered is stored in
In SDRAM memory, respond the reading order that provides of prefix image zooming-out module, by the prefix image in RMB image from
Read in SDRAM memory, giving prefix picture recognition module carries out prefix identification;
Prefix image search module:For completing the judgement to prefix image-region position, and by RMB image side
Edge information calculates the anglec of rotation of RMB image;
Prefix image zooming-out module:Picture position and angle information for prefix image search module to be drawn is converted to
Sdram interface module is given in the address of SDRAM, and sdram interface module reads the image of appropriate address and gives prefix image recognition
Module;
Prefix picture recognition module:For completing the identification work of prefix image, i.e., first binaryzation is carried out to prefix image
Operation, further according to the boundary information of character Character segmentation is carried out, and the image after finishing to segmentation carries out template matching, completes right
The identification of prefix image;
On-chip processor module:Serial number for prefix picture recognition module to be identified passes through USB interface and net
Network interface is sent to external equipment.
Further, the prefix image search module includes:
RMB image binaryzation submodule, for input RMB image to be switched to into binary image output;
Enter paper money walking direction submodule, for judge RMB paper money direction is entered, that is, judge that RMB is that forward direction enters paper money also
It is reversely to enter paper money, and will determine that result gives prefix picture position calculating sub module;
Positive inverse video judging submodule, for judge RMB direct picture be by upper images sensor obtain or under
Square imageing sensor is obtained, and will determine that result gives prefix picture position calculating sub module;
Enter paper money angle calculation submodule, for calculating the anglec of rotation of RMB image, and prefix is given by result of calculation
Picture position calculating sub module;
Prefix picture position calculating sub module, enters paper money directional information, positive and negative image information and enters paper money angle for basis,
Calculate the position that prefix image is located in whole RMB image.
Further, the prefix picture recognition module includes:
Prefix image binaryzation submodule, for binaryzation to be carried out to prefix image using maximum between-cluster variance algorithm, and
Prefix image segmentation submodule is given by the image after binaryzation;
Prefix image segmentation submodule, for splitting to binary image, and gives prefix by the image after segmentation
Image recognition submodule;
Prefix image recognition submodule, for being identified to the prefix image after segmentation using template matching algorithm.
The present invention has compared with prior art following income and effect:
(1) simple structure
Invention is on hardware circuit using the design of FPGA+DSP or FPGA+ ppus, the present invention
Hardware circuit only uses FPGA complete design targets, and hardware configuration is simple, effectively saves veneer area;
(2) processing speed is fast
The present invention completes image recognition work using FPGA, and image is completed using DSP or other processors with other inventions
Identification work is compared, and because FPGA is using pipelining and parallel processing technique, therefore improves the speed of image recognition.
Description of the drawings
Fig. 1 is the structured flowchart of present system.
Fig. 2 is the FPGA unit block diagram in the present invention.
Fig. 3 is prefix image search module block diagram in FPGA unit of the invention.
Fig. 4 is prefix picture recognition module block diagram in FPGA unit of the invention.
Fig. 5 is the sequence chart sued for peace by row to prefix image after binaryzation in the present invention.
Specific embodiment
The present invention is described in further details with reference to embodiment, but embodiments of the present invention not limited to this.
With reference to Fig. 1, system of the present invention includes:Imageing sensor 1, A/D converter 2, FPGA unit 3, driver 4,
The USB interface 6 that SDRAM memory 5, FPGA is extended out and the network interface 7 that FPGA is extended out.A/D converter 2 is used to pass image
The analogue signal of sensor 1 switchs to digital signal, and driver 4 is used to control the light source of imageing sensor 1.FPGA3 is used to detect electricity
Machine turn signal, when detecting motor and rotating controller 4 is opened, and then completes RMB image data acquiring, RMB figure
Prefix picture search as in, and RMB view data is stored in SDRAM memory 5, then take from SDRAM memory 5
Go out prefix image, complete prefix image recognition work, recognition result is sent out to external equipment by USB port and network interface.
Wherein imageing sensor 1 is contact-type image sensor, is divided into the hypograph sensor 12 of epigraph sensor 11;A/D converter 2
It is divided into the first A/D converter 21 and the second A/D converter 22,4 points of driver is the first driver 41 and the second driver 42.
When motor is rotated, FPGA3 control drivers 4 open the light source of imageing sensor 1, wherein the first controller 41 is negative
The light source of duty control epigraph sensor 11, second controller 42 is responsible for the light source of control hypograph sensor 12.Epigraph is passed
Sensor 11 and hypograph sensor 12 gather respectively the image information on two surfaces of RMB, carry out being converted to mould after opto-electronic conversion
Intend electric signal output to A/D converter 2, wherein epigraph sensor 11 outputs analog signal to the first A/D converter 21, under
Imageing sensor 12 exports image to the second A/D converter 22.A/D converter carries out obtaining upper and lower surface people after analog digital conversion
People's coin data image signal is exported to FPGA3, and FPGA3 completes upper and lower surface RMB digital image acquisition, by upper and lower surface people
People's coin digital picture is stored in SDRAM memory 5 simultaneously, while carrying out positive and negative information, entering paper money direction to the two sides image of RMB
Information judges and enters paper money angle calculation, and after synthesis position of the prefix image in RMB image is obtained.Then deposit from SDRAM
Prefix image is taken out in reservoir 5 carries out prefix image recognition operations, finally connects the result of identification by USB interface 6 and network
Mouth 7 is sent to external equipment.
Reference Fig. 2, the FPGA unit 3 of the present invention, including:Control module 31, image capture module 32, sdram interface mould
Block 33, prefix image search module 34, prefix image zooming-out module 35, prefix picture recognition module 36 and on-chip processor
37.Wherein:
Control module 31, for controlling the Push And Release of driver 4;
Image capture module 32, for collection from the upper and lower surface RMB data image signal of A/D converter, will adopt
The two sides RMB image that collection is obtained is while give sdram interface module 33 and prefix image search module 34;
Sdram interface module 33, for two sides RMB image to be stored in SDRAM memory 5, then according to hat simultaneously
The address information that word image zooming-out module 35 is given is taken out RMB prefix image and gives prefix picture recognition module 36;
Prefix image search module 34, for analyzing upper and lower surface RMB digital image information, judges that RMB is positive and negative
Information, RMB enters paper money directional information and RMB rotation angle information, obtains RMB prefix picture position with this and gives
Prefix image zooming-out module 35;
RMB prefix image zooming-out module 35, for being calculated according to RMB prefix position and rotation angle information
Positional information of the RMB prefix in SDRAM memory, gives sdram interface module 33 and takes out RMB prefix image;
Picture recognition module 36, for being identified to prefix image, identifies that the corresponding crown word number of prefix image is given
On-chip processor 37;
On-chip processor 37, the serial number for will identify that is sent to outside and sets by USB interface and network interface
It is standby.
Reference Fig. 3, the prefix image search module 34, including:RMB image binaryzation submodule 341, enters paper money side
To judging submodule 342, positive inverse video judging submodule 343, enter paper money angle calculation submodule 344 and prefix picture position meter
Operator module 345.Wherein:
RMB image binaryzation submodule 341, for carrying out binaryzation to the RMB image being input into, after binaryzation
Image is given into paper money walking direction submodule 342, positive inverse video judging submodule 343 and enters paper money angle calculation submodule simultaneously
344;
Enter paper money walking direction submodule 342, for judge RMB paper money direction is entered, that is, judge that RMB is that forward direction enters paper money
Still reversely enter paper money, and the judged result in paper money direction will be entered while giving positive inverse video judging submodule 343 and entering paper money goniometer
Operator module 344;
Positive inverse video judging submodule 343, for judging that RMB direct picture is by epigraph sensor or figure below
As sensor is obtained, and prefix picture position calculating sub module 345 is given by the result that positive inverse video judges;
Enter paper money angle calculation submodule 344, for calculating the anglec of rotation of RMB image, what is calculated enters paper money angle letter
Breath gives prefix picture position calculating sub module 345;
Prefix picture position calculating sub module 345, for calculating the position that prefix image is located in whole RMB image
Put, and RMB prefix image zooming-out module 35 is given by the prefix picture position result for calculating.
Reference Fig. 4, the prefix picture recognition module 36, including:Prefix image binaryzation submodule 361, prefix image
Segmentation submodule 362 and prefix image recognition submodule 363.Wherein:
Prefix image binaryzation submodule 361, for prefix image to be carried out into binaryzation, the image after binaryzation gives figure
As segmentation submodule 362;
Prefix image segmentation submodule 362, for splitting to the binary image that image binaryzation submodule is exported,
Image after segmentation gives image recognition submodule 363;
Prefix image recognition submodule 363, for the prefix image after the segmentation that exports to prefix image segmentation submodule
It is identified, the result of identification gives on-chip processor module 37.
The operation principle of whole FPGA unit 3 is as follows:
FPGA unit 3 enters idle condition after upper electricity, in an idle state the detection of FPGA unit 3 motor rotation situation:When
Detect control module 31 when motor is not rotated and close driver, control module 31 is opened when detecting motor and starting to rotate
Driver;
Control module 31 starts image capture module 32 and starts working after driver is opened, and gathers two-way A/D converter
The RMB image of two-way up and down of input, gives two-way image sdram interface module 33 and is stored in SDRAM memory, and by its
Simultaneously being sent to prefix image search module 34 proceeds by prefix picture search;
Upper and lower two-way RMB image is stored in SDRAM memory by sdram interface module 33 simultaneously;
In prefix image search module 34 in search procedure, RMB image binaryzation submodule 341 therein is first
The RMB of two-way up and down image to being input into carries out binaryzation, arranges binary-state threshold TR=65, and input pixel value is entered with TR
Row generates more afterwards the output valve of binary image, if input pixel value is output as 1 more than TR, if input pixel value is less than or equal to
TR is then output as 0;And the image after binaryzation is judged into submodule while giving into paper money walking direction submodule 342, both forward and reverse directions
Block 343 and enter paper money angle calculation submodule 344;
Enter paper money walking direction submodule 342, assume initially that the two RMB images up and down after binaryzation are front elevation
Picture, judges that judge process is simultaneously to upper and lower two RMB images:The part of a left side 1/4 to binary image and right 1/4
Part is sued for peace respectively, obtain left 1/4 part and value X1 and right 1/4 part with value X2, then X1 is compared with X2, if X1
>X2, then the brighter region of image is in left side, conversely, the brighter region of image is on right side, according to the feature of image of RMB, such as
Fruit in left side, then enters paper money compared with bright part for forward direction, otherwise is then reversely to enter paper money;
Two binary images up and down for receiving are designated as epigraph and hypograph by both forward and reverse directions judging submodule 343, first
First assume that epigraph and hypograph are direct picture, do to epigraph according to the result for entering paper money walking direction submodule 342 and sentence
Disconnected, if forward direction enters paper money, then the upper left corner is national emblem image, if reversely entering paper money, the lower right corner is national emblem image, to this part
Pixel value summation is obtained and value Y1;Likewise, hypograph is judged according to the result for entering paper money walking direction submodule 342,
If forward direction enters paper money, then the upper left corner is national emblem image, if reversely entering paper money, the lower right corner is national emblem image, to this part picture
Plain value summation is obtained and value Y2;Compare the size of Y1 and Y2, if Y1<Y2, then epigraph is direct picture, conversely, hypograph
For direct picture;
Enter paper money angle calculation submodule 344, first to the marginal information of upper and lower two binary images of lookup, binary picture
As first position from 0 to 1 saltus step, as the left and right edges position of image in left-right parts pixel value, it is assumed that search the 50th
Capable left hand edge position is h50, and the left hand edge position of the 100th row is h100, then the slope of image border isSeek slope
Arc tangent obtain the rotation angle θ of signal, i.e.,It is defeated due to calculating upper and lower two images simultaneously
Go out two anglecs of rotation, note epigraph corresponding rotation angle is θ 1, and hypograph corresponding rotation angle is θ 2;
Enter paper money walking direction submodule 342, positive inverse video judging submodule 343 and enter paper money angle calculation submodule 344
Result give prefix picture position calculating sub module 345, prefix picture position calculating sub module 345 is first according to entering paper money side
To judging submodule 342 and the result of positive inverse video judging submodule 343, obtain which imageing sensor prefix image comes from
And in the upper left corner or the lower right corner of RMB image, and the anglec of rotation of direct picture is selected as the rotation for finally using
Angle, it is assumed that the final anglec of rotation is θ 3, if direct picture is epigraph, θ 3=θ 1, if direct picture is hypograph, θ 3=θ
2;Then the position coordinateses in the correspondence upper left corner or the lower right corner are obtained according to the marginal information of RMB image, this position is designated as and is sat
It is designated as (h0, v0).Because the relative position of RMB prefix image-region and corner location is fixed value, it is assumed that its lateral separation
For H, fore-and-aft distance is V, and prefix picture position is obtained according to corner location coordinate (h0, v0) and rotation angle θ 3, it is assumed that
Prefix picture position is (h1, v1), and the computing formula of coordinate (h1, v1) is as follows:
H1=h0+Hcos (θ 3)-Vsin (θ 3)
V1=v0+Vcos (θ 3)+Hsin (θ 3);
Whole prefix image search module 34 completes to search after element, and prefix picture position and image rotation angle information are given
Prefix image zooming-out module 35, prefix image zooming-out module 35 is according to above- mentioned information calculating prefix image in SDRAM memory
In address, give sdram interface module 33 will prefix image read after give prefix picture recognition module 36 again;
Prefix picture recognition module 36 in identification process, first by prefix image binaryzation submodule 361 pairs therein
Prefix image carries out binaryzation, and binary-state threshold is calculated with maximum between-cluster variance algorithm before binaryzation, and its computational methods is as follows:
For total pixel number is the image of N, the mean μ of pixel value is calculated, it is assumed that currently binary-state threshold is
T0, pixel value of the statistics more than T0 has N0, and it is μ 0 to be calculated its average pixel value, and pixel value of the statistics less than or equal to T0 is
N1, it is μ 1 to be calculated its average pixel value, calculates inter-class variance and is
In the range of (0~255) travel through T0 cause g for maximum, take g it is maximum when corresponding T0 values be optimal threshold T;
The output valve of binary image is generated after input pixel value and T are compared, if input pixel value is more than T,
1 is output as, if input pixel value is less than or equal to T, 0 is output as.Prefix image after binaryzation gives prefix image segmentation
Module 362;
Prefix image segmentation submodule 362 is split according to the following procedure to the prefix image after binaryzation:
First, the left and right edges position of ten prefix subgraphs in a width prefix image is found, i.e., prefix image is asked by row
With the sequence after being sued for peace is as shown in figure 5, the characteristic of the sequence is:Numerical value is less at prefix subgraph, shows as valley,
Maximum is reached in prefix subgraph gap location numerical value, peak value is shown as;
Secondly, the marginal position of prefix subgraph is determined according to peak value and valley:
The position that peak value is changed to valley into prefix subgraph left hand edge position, the position that valley is changed to peak value is into prefix
The right edge position of subgraph;
Into the 1st prefix subgraph left hand edge position, the 1st valley is to peak value for the position that 1st peak value is changed to valley
Translation bit is set to the right edge position of the 1st prefix subgraph;
Into the 2nd prefix subgraph left hand edge position, the 2nd valley is to peak value for the position that 2nd peak value is changed to valley
Translation bit is set to the right edge position of the 2nd prefix subgraph, by that analogy;
Then, 10 prefix subgraphs are taken out respectively according to the position at respective edge from prefix image, that is, realizes hat
The segmentation of word image, the prefix subgraph after segmentation gives image recognition submodule 363;
Ten prefix subgraphs of image recognition submodule 363 pair are identified, and image recognition uses template matching algorithm, its knowledge
Other process is as follows:
First, in 363 storage inside of image recognition submodule, 26 letters and the standard form image of 10 numerals;
Secondly, operation is zoomed in and out to images to be recognized, images to be recognized is scaled into the size same with template image;
Then, image after scaling is carried out into xor operation with each template image by pixel successively, and to the knot after XOR
Fruit is sued for peace, and obtains 36 XOR summed results;
Finally, this 36 results are compared, find the corresponding template of maximum of which result, this template is corresponding
Numeral or letter are the result for identifying;
Ten prefix subgraphs are identified successively in this manner, ten obtained after identification crown word number is given on piece
Processor 37;
The crown word number that on-chip processor module 37 will identify that is sent to external equipment by USB interface and network interface,
So complete the identification work of image.
Above description is only example of the present invention, does not constitute any limitation of the invention, it is clear that for this
For the professional in field, after present invention and principle is understood, all may be without departing substantially from the principle of the invention, the feelings of structure
Under condition, various amendments and the change in form and details is carried out, but these amendments and change based on inventive concept still exist
Within the claims of the present invention.
Claims (3)
1. a kind of RMB prefix identifying system based on FPGA, including:
Upper and lower two imageing sensors (1), two-way A/D converter (2), FPGA unit (3), two-way driver (4), SDRAM are deposited
The USB interface (6) that reservoir (5), FPGA are extended out and the network interface (7) that FPGA is extended out, imageing sensor is contact type image
The analogue signal of imageing sensor (1) is switched to digital signal by sensor, A/D converter (2), and driver (4) control image is passed
The light source of sensor (1), it is characterised in that:
FPGA unit (3) includes:
Control module (31):For controlling opening or the closing of driver (4);
Image capture module (32):It is connected with A/D converter, for completing the collection of digital picture;
Sdram interface module (33):For realizing sdram interface control logic, the image that image capture module (32) is gathered
In being stored in SDRAM memory (5), the reading order that prefix image zooming-out module (35) is provided is responded, by RMB image
Prefix image reads from SDRAM, and giving prefix identification module (36) carries out prefix identification;
Prefix image search module (34):For completing the judgement to prefix image-region position, and by RMB image side
Edge information calculates the anglec of rotation of RMB image;
Prefix image zooming-out module (35):Picture position and angle information for prefix image search module to be drawn is converted to
Sdram interface module is given in the address of SDRAM, and sdram interface module reads the prefix image of appropriate address and gives prefix image
Identification module (36);
Prefix picture recognition module (36):For completing the identification work of prefix image, i.e., first binaryzation is carried out to prefix image
Operation, further according to the marginal information of prefix subgraph prefix image segmentation is carried out, and finally the prefix subgraph to being partitioned into uses template
The mode of matching is identified;
On-chip processor module (37):Serial number for prefix picture recognition module (36) to be identified passes through USB interface
And network interface (7) is sent to external equipment (6).
2., according to system according to claim 1, wherein prefix image search module (34) includes:
RMB image binaryzation submodule (341), enters paper money walking direction submodule (342), positive inverse video judging submodule
(343) paper money angle calculation submodule (344) and prefix picture position calculating sub module (345), are entered;
RMB image binaryzation submodule (341), for carrying out binaryzation to the RMB image being input into, the figure after binaryzation
As giving into paper money walking direction submodule (342), positive inverse video judging submodule (343) and entering paper money angle calculation submodule simultaneously
Block (344);
Enter paper money walking direction submodule (342), for judge RMB paper money direction is entered, that is, judge that RMB is that forward direction enters paper money also
It is reversely to enter paper money, and will determine that result gives prefix picture position calculating sub module (345);
Positive inverse video judging submodule (343), for judge RMB direct picture be obtained by upper images sensor or
Lower images sensor is obtained, and will determine that result gives prefix picture position calculating sub module (345);
Enter paper money angle calculation submodule (344), for calculating the anglec of rotation of RMB image, and prefix is given by result of calculation
Picture position calculating sub module (345);
Prefix picture position calculating sub module (345), enters paper money directional information, positive and negative image information and enters paper money angle for basis
Degree, calculates the position that prefix image is located in whole RMB image.
3. system according to claim 1, wherein prefix picture recognition module (36) includes:
Prefix image binaryzation submodule (361), for binaryzation to be carried out to prefix image using maximum between-cluster variance algorithm, and
Prefix image segmentation submodule (362) is given by the image after binaryzation;
Prefix image segmentation submodule (362), for splitting to binary image, and gives prefix by the image after segmentation
Image recognition submodule (363);
Prefix image recognition submodule (363), for being identified to the prefix image after segmentation using template matching algorithm.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611165429.1A CN106652167A (en) | 2016-12-16 | 2016-12-16 | FPGA-based RMB crown word identification system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611165429.1A CN106652167A (en) | 2016-12-16 | 2016-12-16 | FPGA-based RMB crown word identification system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106652167A true CN106652167A (en) | 2017-05-10 |
Family
ID=58823634
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611165429.1A Pending CN106652167A (en) | 2016-12-16 | 2016-12-16 | FPGA-based RMB crown word identification system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106652167A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110010019A (en) * | 2019-04-15 | 2019-07-12 | 珠海格力智能装备有限公司 | Control method and device for assembling LED screen |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102262726A (en) * | 2011-06-17 | 2011-11-30 | 西安电子科技大学 | Multi-core license plate recognition system based on field programmable gate array (FPGA) |
JP2012137814A (en) * | 2010-12-24 | 2012-07-19 | Hitachi Omron Terminal Solutions Corp | Device and method for identifying paper sheet |
CN103679917A (en) * | 2014-01-06 | 2014-03-26 | 成都术有科技有限公司 | High-speed banknote serial number extraction and identification system based on FPGA (field programmable gate array) and implementation method thereof |
CN103729934A (en) * | 2014-01-06 | 2014-04-16 | 成都术有科技有限公司 | High-speed paper money crown word number extraction and recognition system based on DSP |
US20150221155A1 (en) * | 2014-01-31 | 2015-08-06 | Ncr Corporation | Media item re-orientation |
CN105261110A (en) * | 2015-10-26 | 2016-01-20 | 江苏国光信息产业股份有限公司 | Efficient DSP banknote serial number recognizing method |
CN105989659A (en) * | 2016-04-15 | 2016-10-05 | 新达通科技股份有限公司 | Similar character recognition method and paper currency crown code recognition method |
-
2016
- 2016-12-16 CN CN201611165429.1A patent/CN106652167A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012137814A (en) * | 2010-12-24 | 2012-07-19 | Hitachi Omron Terminal Solutions Corp | Device and method for identifying paper sheet |
CN102262726A (en) * | 2011-06-17 | 2011-11-30 | 西安电子科技大学 | Multi-core license plate recognition system based on field programmable gate array (FPGA) |
CN103679917A (en) * | 2014-01-06 | 2014-03-26 | 成都术有科技有限公司 | High-speed banknote serial number extraction and identification system based on FPGA (field programmable gate array) and implementation method thereof |
CN103729934A (en) * | 2014-01-06 | 2014-04-16 | 成都术有科技有限公司 | High-speed paper money crown word number extraction and recognition system based on DSP |
US20150221155A1 (en) * | 2014-01-31 | 2015-08-06 | Ncr Corporation | Media item re-orientation |
CN105261110A (en) * | 2015-10-26 | 2016-01-20 | 江苏国光信息产业股份有限公司 | Efficient DSP banknote serial number recognizing method |
CN105989659A (en) * | 2016-04-15 | 2016-10-05 | 新达通科技股份有限公司 | Similar character recognition method and paper currency crown code recognition method |
Non-Patent Citations (1)
Title |
---|
胡玉琛: ""基于FPGA的车牌识别***研究与实现"", 《电子测量技术》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110010019A (en) * | 2019-04-15 | 2019-07-12 | 珠海格力智能装备有限公司 | Control method and device for assembling LED screen |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10096091B2 (en) | Image generating method and apparatus | |
CN102103698B (en) | Image processing apparatus and image processing method | |
Ali et al. | A real-time deformable detector | |
CN103345631A (en) | Image characteristic extraction, training, detection method, module, device and system | |
Li et al. | Action recognition from depth sequence using depth motion maps-based local ternary patterns and CNN | |
Weiyao et al. | Human action recognition using multilevel depth motion maps | |
Zhou et al. | TMFNet: Three-input multilevel fusion network for detecting salient objects in RGB-D images | |
Guo et al. | Real-time hand detection based on multi-stage HOG-SVM classifier | |
Ding et al. | Recognition of hand-gestures using improved local binary pattern | |
Cao et al. | Local binary pattern features for pedestrian detection at night/dark environment | |
Sangeetha et al. | A low-cost and high-performance architecture for robust human detection using histogram of edge oriented gradients | |
Keceli | Viewpoint projection based deep feature learning for single and dyadic action recognition | |
WO2018113206A1 (en) | Image processing method and terminal | |
Wang et al. | Real-time visual static hand gesture recognition system and its FPGA-based hardware implementation | |
CN106652167A (en) | FPGA-based RMB crown word identification system | |
Zhao et al. | Hand detection using multi-resolution HOG features | |
JPH08272973A (en) | Image process for face feature extraction | |
Gao et al. | Adaptive HOG-LBP based learning for palm tracking | |
Yu et al. | Embedding pose information for multiview vehicle model recognition | |
Elsayed et al. | Hybrid method based on multi-feature descriptor for static sign language recognition | |
KR20150136971A (en) | Method and apparatus for detecting face with low energy or low resolution | |
Xu et al. | A static gesture recognition method based on improved SURF algorithm and Bayesian regularization BP neural network | |
Bhame et al. | Vision based calculator for speech and hearing impaired using hand gesture recognition | |
Zhao et al. | Hierarchical codebook background model using haar-like features | |
Wonneberger et al. | Parallel feature extraction and heterogeneous object-detection for multi-camera driver assistance systems |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170510 |