CN105389883A - Banknote prefix number identification method for currency detector - Google Patents
Banknote prefix number identification method for currency detector Download PDFInfo
- Publication number
- CN105389883A CN105389883A CN201510744123.0A CN201510744123A CN105389883A CN 105389883 A CN105389883 A CN 105389883A CN 201510744123 A CN201510744123 A CN 201510744123A CN 105389883 A CN105389883 A CN 105389883A
- Authority
- CN
- China
- Prior art keywords
- processor
- pixel
- coordinate
- crown word
- word number
- 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
Landscapes
- Image Analysis (AREA)
Abstract
Disclosed in the invention is a banknote prefix number identification method for a currency detector. The currency detector consists of a processor, a memory, an upper image sensor, and a lower image sensor; the processor is connected with the memory, the upper image sensor, and the lower image sensor electrically; and the currency detector is connected with an upper computer electrically. The method comprise the following steps: coarse positioning is carried out on a prefix number; a coordinate (fx,fy) of each pixel point at an area where the prefix number is located is extracted and bilinear interpolation fitting is carried out by using four pixel points around the (fx,fy) to obtain a prefix number sub graph A' of the banknote; binaryzation is carried out on the prefix number sub graph A' and a feature vector F of each target object in the A' is extracted by using an eight neighborhood contour extraction method; and prefix number identification and outputting are carried out. Therefore, the provided method has characteristics of high prefix number extraction speed and high identification accuracy.
Description
Technical field
The present invention relates to cash inspecting machine technical field, especially relate to the paper money number recognition methods of the cash inspecting machine that a kind of crown word number extraction rate is fast, recognition accuracy is high.
Background technology
For paper money number identification, no matter be which kind of classifying identification method, extracting stable feature is the prerequisite identified.
Usual use neural net method identifies crest number, and neural net method exists following pros and cons:
Need a large amount of sample training, sentence stagger the time when there being character, can only be attributed to network and not train, then Resurvey batch sample is trained again; Or network structure is unreasonable, the adjustment network number of plies and node number, and excitation function etc.If there is individual characters to sentence mistake, want to change this state, need to drop into a large amount of work and again train, but, can not ensure the performance that the network of new training has had will cause vicious cycle.
If by whole character picture input neural network, global feature will be caused to be similar to, and the obvious character identification rate of local feature is low.As character 0 and D, 8 and B, S and 5,2 and Z, E and F etc.And profile chain code can reflect the local feature of character, there is high discrimination to the obvious character of local feature.
Summary of the invention
Goal of the invention of the present invention is to overcome the deficiency that crown word number identification method crown word number extraction rate of the prior art is slow, recognition accuracy is low, providing the paper money number recognition methods of the cash inspecting machine that a kind of crown word number extraction rate is fast, recognition accuracy is high.
To achieve these goals, the present invention is by the following technical solutions:
A paper money number recognition methods for cash inspecting machine, described cash inspecting machine comprises processor, storer, epigraph sensor, hypograph sensor, display and alarm; Processor is electrically connected with storer, epigraph sensor, hypograph sensor, display and alarm respectively; Comprise the steps:
(1-1) when bank note is through cash inspecting machine, epigraph sensor contacts with bank note upper and lower surface respectively with hypograph sensor, and epigraph sensor and hypograph sensor gather image and image is passed to processor;
Processor identifies the angle anghori of the X-axis of bank note coboundary fitting a straight line and setting, the angle angvert of the Y-axis of bank note left margin fitting a straight line and setting, the coordinate (O of bank note top left corner pixel point
x, O
y); Column direction zoom factor coefcol and line direction zoom factor coefrow is provided with in processor;
(1-2) region at processor coarse localization crown word number place, setting (n
x, n
y) be the coordinate of any one pixel in crown word number region A after rotation correction;
Utilize formula
Calculate each pixel of crown word number region A, the pixel coordinate (f in the original image before rotation correction
x, f
y),
CH＝cos(anghori)coefcol
SV＝sin(angvert)coefrow
Wherein, SH=sin (anghori) coefcol;
CV＝cos(angvert)coefrow
Processor extracts each pixel coordinate (f of crown word number region in image gathering
x, f
y), utilize (f
x, f
y) 4 pixels around carry out bilinear interpolation matching, thus obtain the crown word number subgraph A ' of bank note;
(1-3) processor adopts Otsu dynamic threshold to carry out binaryzation to crown word number subgraph A ', adopts the method for 8 neighborhood contours extract, extracts the proper vector F of each object in A ';
(1-4) be provided with Sample Storehouse in storer, recognition threshold Q, be provided with proper vector template corresponding with each crown word number character respectively in Sample Storehouse, processor is searched for and is obtained with the proper vector F of each object apart from minimum proper vector template T
min;
(1-5) processor utilizes formula
calculate similarity ρ,
As ρ>=Q, then processor will with T
mincorresponding character is selected and preserves in memory,
As ρ < Q, then processor by "? " preserve in memory;
When all character recognition complete, processor is according to the centre coordinate order from small to large of the boundary rectangle of the outline of the object corresponding with each character, and array from left to right each character the paper money number A1 obtaining identifying;
Processor control display display A1 and A ', when any one character in A1 be "? " time, processor controls alarm equipment alarm.
The present invention includes crown word number coarse localization; Extract each pixel coordinate (f of crown word number region
x, f
y), utilize (f
x, f
y) 4 pixels around carry out bilinear interpolation matching, obtain the crown word number subgraph A ' of bank note; Binaryzation is carried out to crown word number subgraph A ', adopts the method for 8 neighborhood contours extract, extract the proper vector F of each object in A '; Crown word number identification and output.
The present invention searches for the clockwise direction of the outline of object according to 8 neighborhoods, Internal periphery is searched for according to the counter clockwise direction of 8 neighborhoods, owing to having 8 possible directions for each point of profile to next point, the tendency of chain code well simulates the tendency of character stroke, each character can be represented more accurately, while extracting chain code, have also been obtained 2 dimension profile coordinate points of object, be beneficial to and identify distinguishing of later stage similar character details, profile chain code is linear feature, relative to more existing region feature, feature extraction speed of the present invention is faster, the present invention only needs run-down image, namely the stain at most of crown word number place can be rejected, improve the accuracy rate of paper money number identification largely.
Therefore, the present invention has the advantages that crown word number extraction rate is fast, recognition accuracy is high.
As preferably, the method for described employing 8 neighborhood contours extract, the proper vector F extracting each object in A ' comprises following concrete steps:
Be provided with object number n in storer, the initial value of setting search sequence number i is 1;
(2-1) processor is to the crown word number subgraph of binaryzation according to from top to bottom, and order is from left to right lined by line scan, and when running into pixel value and being first pixel B of 255, is set to current pixel point;
Processor is crossed current pixel point and is done 8 rays, point to right from current pixel point, lower right, below, lower left, left, upper left side, top, top-right 8 rays are corresponding with mutually different integer chain code L1, L2, L3, L4, L5, L6, L7, L8 respectively
(2-2) processor according to from current pixel point to the right, upper right side, top, upper left side, left, lower left, below, a bottom-right order mobile pixel successively, when searching pixel value and being the pixel C of 255, using this pixel C as current pixel point, when the coordinate of C is identical with the coordinate of B, proceed to step (2-3);
When the coordinate time of the coordinate ≠ B of C, return step (2-2);
(2-3) in the process of search, the coordinate of each pixel searched and chain code are all stored in storer by processor, after search terminates, processor obtains the outline be made up of each pixel, processor calculates the boundary rectangle of outline, calculates the center point coordinate of object, length and width according to boundary rectangle;
(2-4) processor calculates the Internal periphery number in boundary rectangle, for each Internal periphery, processor any one pixel selected in Internal periphery is current pixel point, repeat step (2-2) to (2-3), according to from current pixel point to the right, lower right, below, lower left, left, upper left side, top, a top-right order mobile pixel successively, after search terminates, processor calculates in each Internal periphery and connects rectangle, and obtain each in connect the centre coordinate of rectangle;
The chain code combination of outline and each Internal periphery is formed object chain code length, chain code tendency;
(2-5) make i value increase by 1, as i < n, searched to object outside A ' in, repeat step (2-1) to (2-4), continue search other object;
Work as i=n, then all object search are complete; The centre coordinate combination constitutive characteristic vector F of the centre coordinate of rectangle and the boundary rectangle of outline will be met in the chain code length of each object, chain code tendency, Internal periphery number, length, width, Internal periphery.
As preferably, the angle of described adjacent ray is 45 °.
As preferably, the chain code length of each object for search complete time this object pixel total quantity of searching for, the chain code that described chain code tendency is arranged by the sequencing according to the pixel searched is formed.
As preferably, the region at described processor coarse localization crown word number place comprises the steps:
When processor makes the heads judgement of bank note according to the image gathered, processor chooses crown word number region in the lower left region of the image of epigraph sensor collection;
When processor makes the judgement of bank note face down according to the image gathered, processor chooses crown word number region in the right regions of the image in the collection of epigraph sensor;
When processor makes the anti-supine judgement of bank note according to the image gathered, processor chooses crown word number region in the lower right area of the image of hypograph sensor collection;
When processor makes the anti-ventricumbent judgement of bank note according to the image gathered, processor chooses crown word number region in the top left region of the image of hypograph sensor collection.
As preferably, described L1, L2, L3, L4, L5, L6, L7, L8 are respectively 1,2,3,4,5,6,7,0.
As preferably, Q is 88% to 93%.
Therefore, the present invention has following beneficial effect: crown word number extraction rate is fast and recognition accuracy is high.
Accompanying drawing explanation
Fig. 1 is a kind of structural representation of 8 neighborhood contour extraction methods of the present invention;
Fig. 2 is a kind of crown word number subgraph to be identified of the present invention;
Fig. 3 is a kind of chain code schematic diagram of character 0 of the present invention;
Fig. 4 is a kind of chain code schematic diagram of character 1 of the present invention;
Fig. 5 is a kind of process flow diagram of embodiments of the invention.
Embodiment
Below in conjunction with the drawings and specific embodiments, the present invention will be further described.
Embodiment is as shown in Figure 1 a kind of paper money number recognition methods of cash inspecting machine, and cash inspecting machine comprises processor, storer, epigraph sensor, hypograph sensor, display and alarm; Processor is electrically connected with storer, epigraph sensor, hypograph sensor, display and alarm respectively; Comprise the steps:
As shown in Figure 5,
Step 100, gathers image
When bank note is through cash inspecting machine, epigraph sensor contacts with bank note upper and lower surface respectively with hypograph sensor, and epigraph sensor and hypograph sensor gather image and image is passed to processor;
Processor identifies the angle anghori of the X-axis of bank note coboundary fitting a straight line and setting, the angle angvert of the Y-axis of bank note left margin fitting a straight line and setting, the coordinate (O of bank note top left corner pixel point
x, O
y); Column direction zoom factor coefcol and line direction zoom factor coefrow is provided with in processor;
Step 200, carries out bilinear interpolation matching to each pixel, obtains the crown word number subgraph of bank note
The region at processor coarse localization crown word number place, when processor makes the heads judgement of bank note according to the image gathered, processor chooses crown word number region in the lower left region of the image of epigraph sensor collection;
When processor makes the judgement of bank note face down according to the image gathered, processor chooses crown word number region in the right regions of the image in the collection of epigraph sensor;
When processor makes the anti-supine judgement of bank note according to the image gathered, processor chooses crown word number region in the lower right area of the image of hypograph sensor collection;
When processor makes the anti-ventricumbent judgement of bank note according to the image gathered, processor chooses crown word number region in the top left region of the image of hypograph sensor collection;
Setting (n
x, n
y) be the coordinate of any one pixel in crown word number region A after rotation correction;
Utilize formula
Calculate each pixel of crown word number region A, the pixel coordinate (f in the original image before rotation correction
x, f
y),
CH＝cos(anghori)coefcol
SV＝sin(angvert)coefrow
Wherein, SH=sin (anghori) coefcol;
CV＝cos(angvert)coefrow
Processor extracts each pixel coordinate (f of crown word number region in image gathering
x, f
y), utilize (f
x, f
y) 4 pixels around carry out bilinear interpolation matching, thus obtain the crown word number subgraph A ' of bank note;
Step 300, processor adopts Otsu dynamic threshold to carry out binaryzation to crown word number subgraph A ', adopts the method for 8 neighborhood contours extract, extracts the proper vector F of each object in A ';
Each character in crown word number subgraph A ' is as shown in Figure 2 object, and black is substrate, and white is character.
Be provided with object number n=10 in storer, the initial value of setting search sequence number i is 1;
Step 310, processor is to the crown word number subgraph of binaryzation as shown in Figure 2 according to from top to bottom, and order is from left to right lined by line scan, and when running into pixel value and being first pixel B of 255, is set to current pixel point; In the present embodiment, because character R is highly the highest, can first be scanned.
As shown in Figure 1, processor is crossed current pixel point and is done 8 rays, from current pixel point point to right, lower right, below, lower left, left, upper left side, top, top-right 8 rays respectively with chain code 1,2,3,4,5,6,7,0 corresponding,
Step 320, processor according to from current pixel point to the right, upper right side, top, upper left side, left, lower left, below, a bottom-right order mobile pixel successively, when searching pixel value and being the pixel C of 255, using this pixel C as current pixel point, when the coordinate of C is identical with the coordinate of B, proceed to step 330;
When the coordinate time of the coordinate ≠ B of C, return step 320;
Step 330, in the process of search, the coordinate of each pixel searched and chain code are all stored in storer by processor, after search terminates, processor obtains the outline be made up of each pixel, processor calculates the boundary rectangle of outline, calculates the center point coordinate of object, length and width according to boundary rectangle;
Step 340, processor calculates the Internal periphery number in boundary rectangle, for each Internal periphery, processor any one pixel selected in Internal periphery is current pixel point, repeat step 320 to 330, according to from current pixel point to the right, lower right, below, lower left, left, upper left side, top, a top-right order mobile pixel successively, after search terminates, processor calculates in each Internal periphery and connects rectangle, and obtain each in connect the centre coordinate of rectangle;
The chain code combination of outline and each Internal periphery is formed object chain code length, chain code tendency;
Step 350, makes i value increase by 1, as i < 10, searched to object outside A ' in, repeat step 310 to 340, continue search other object;
Work as i=10, then all object search are complete; The centre coordinate combination constitutive characteristic vector F of the centre coordinate of rectangle and the boundary rectangle of outline will be met in the chain code length of each object, chain code tendency, Internal periphery number, length, width, Internal periphery.
Step 400, identifies each object, display screen display A1 and A '
Be provided with Sample Storehouse in storer, recognition threshold Q, be provided with proper vector template corresponding with each crown word number character respectively in Sample Storehouse, processor is searched for and is obtained with the proper vector F of each object apart from minimum proper vector template T
min;
The above-mentioned feature of each character that Sample Storehouse adopts method of the present invention to extract in batch bank note sample obtains.
Step 500, processor utilizes formula
calculate similarity ρ,
As ρ>=Q, then processor will with T
mincorresponding character is selected and preserves in memory,
As ρ < Q, then processor by "? " preserve in memory;
When all character recognition complete, processor is according to the centre coordinate order from small to large of the boundary rectangle of the outline of the object corresponding with each character, and array from left to right each character the paper money number A1 obtaining identifying;
Processor control display display A1 and A ', when any one character in A1 be "? " time, processor controls alarm equipment alarm.The angle of adjacent ray is 45 °.
As shown in Figure 3, Figure 4, horizontal ordinate is the chain code length of object, and ordinate is the chain code tendency of object; Chain code length for search complete time this object pixel total quantity of searching for, the chain code that chain code tendency is arranged by the sequencing according to the pixel searched is formed.
As can be seen from Fig. 3, Fig. 4, the chain code schematic diagram difference of different objects is very large, therefore can be distinguished by kinds of characters easily according to the difference of chain code.
Wherein, the region at processor coarse localization crown word number place comprises the steps:
Q is 90%, and the present invention is for a few money detect counterfeit money machine, and crown word number recognition accuracy reaches 99.8%.
Should be understood that the present embodiment is only not used in for illustration of the present invention to limit the scope of the invention.In addition should be understood that those skilled in the art can make various changes or modifications the present invention, and these equivalent form of values fall within the application's appended claims limited range equally after the content of having read the present invention's instruction.
Claims (7)
1. a paper money number recognition methods for cash inspecting machine, described cash inspecting machine comprises processor, storer, epigraph sensor, hypograph sensor, display and alarm; Processor is electrically connected with storer, epigraph sensor, hypograph sensor, display and alarm respectively; It is characterized in that, comprise the steps:
(1-1) when bank note is through cash inspecting machine, epigraph sensor contacts with bank note upper and lower surface respectively with hypograph sensor, and epigraph sensor and hypograph sensor gather image and image is passed to processor;
Processor identifies the angle anghori of the X-axis of bank note coboundary fitting a straight line and setting, the angle angvert of the Y-axis of bank note left margin fitting a straight line and setting, the coordinate (O of bank note top left corner pixel point
x, O
y); Column direction zoom factor coefcol and line direction zoom factor coefrow is provided with in processor;
(1-2) region at processor coarse localization crown word number place, setting (n
x, n
y) be the coordinate of any one pixel in crown word number region A after rotation correction;
Utilize formula
Calculate each pixel of crown word number region A, the pixel coordinate (f in the original image before rotation correction
x, f
y),
Wherein,
Processor extracts each pixel coordinate (f of crown word number region in image gathering
x, f
y), utilize (f
x, f
y) 4 pixels around carry out bilinear interpolation matching, thus obtain the crown word number subgraph A ' of bank note;
(1-3) processor adopts Otsu dynamic threshold to carry out binaryzation to crown word number subgraph A ', adopts the method for 8 neighborhood contours extract, extracts the proper vector F of each object in A ';
(1-4) be provided with Sample Storehouse in storer, recognition threshold Q, be provided with proper vector template corresponding with each crown word number character respectively in Sample Storehouse, processor is searched for and is obtained with the proper vector F of each object apart from minimum proper vector template T
min;
(1-5) processor utilizes formula
calculate similarity ρ,
As ρ>=Q, then processor will with T
mincorresponding character is selected and preserves in memory,
As ρ < Q, then processor by "? " preserve in memory;
When all character recognition complete, processor is according to the centre coordinate order from small to large of the boundary rectangle of the outline of the object corresponding with each character, and array from left to right each character the paper money number A1 obtaining identifying;
Processor control display display A1 and A ', when any one character in A1 be "? " time, processor controls alarm equipment alarm.
2. the paper money number recognition methods of cash inspecting machine according to claim 1, is characterized in that, the method for described employing 8 neighborhood contours extract, and the proper vector F extracting each object in A ' comprises following concrete steps:
Be provided with object number n in storer, the initial value of setting search sequence number i is 1;
(2-1) processor is to the crown word number subgraph of binaryzation according to from top to bottom, and order is from left to right lined by line scan, and when running into pixel value and being first pixel B of 255, is set to current pixel point;
Processor is crossed current pixel point and is done 8 rays, point to right from current pixel point, lower right, below, lower left, left, upper left side, top, top-right 8 rays are corresponding with mutually different integer chain code L1, L2, L3, L4, L5, L6, L7, L8 respectively
(2-2) processor according to from current pixel point to the right, upper right side, top, upper left side, left, lower left, below, a bottom-right order mobile pixel successively, when searching pixel value and being the pixel C of 255, using this pixel C as current pixel point, when the coordinate of C is identical with the coordinate of B, proceed to step (2-3);
When the coordinate time of the coordinate ≠ B of C, return step (2-2);
(2-3) in the process of search, the coordinate of each pixel searched and chain code are all stored in storer by processor, after search terminates, processor obtains the outline be made up of each pixel, processor calculates the boundary rectangle of outline, calculates the center point coordinate of object, length and width according to boundary rectangle;
(2-4) processor calculates the Internal periphery number in boundary rectangle, for each Internal periphery, processor any one pixel selected in Internal periphery is current pixel point, repeat step (2-2) to (2-3), according to from current pixel point to the right, lower right, below, lower left, left, upper left side, top, a top-right order mobile pixel successively, after search terminates, processor calculates in each Internal periphery and connects rectangle, and obtain each in connect the centre coordinate of rectangle;
The chain code combination of outline and each Internal periphery is formed object chain code length, chain code tendency;
(2-5) make i value increase by 1, as i < n, searched to object outside A ' in, repeat step (2-1) to (2-4), continue search other object;
Work as i=n, then all object search are complete; The centre coordinate combination constitutive characteristic vector F of the centre coordinate of rectangle and the boundary rectangle of outline will be met in the chain code length of each object, chain code tendency, Internal periphery number, length, width, Internal periphery.
3. the paper money number recognition methods of cash inspecting machine according to claim 1, is characterized in that, the angle of described adjacent ray is 45 °.
4. the paper money number recognition methods of cash inspecting machine according to claim 2, it is characterized in that, the chain code length of each object for search complete time this object pixel total quantity of searching for, the chain code that described chain code tendency is arranged by the sequencing according to the pixel searched is formed.
5. the paper money number recognition methods of cash inspecting machine according to claim 1, is characterized in that, the region at described processor coarse localization crown word number place comprises the steps:
When processor makes the heads judgement of bank note according to the image gathered, processor chooses crown word number region in the lower left region of the image of epigraph sensor collection;
When processor makes the judgement of bank note face down according to the image gathered, processor chooses crown word number region in the right regions of the image in the collection of epigraph sensor;
When processor makes the anti-supine judgement of bank note according to the image gathered, processor chooses crown word number region in the lower right area of the image of hypograph sensor collection;
When processor makes the anti-ventricumbent judgement of bank note according to the image gathered, processor chooses crown word number region in the top left region of the image of hypograph sensor collection.
6. the paper money number recognition methods of cash inspecting machine according to claim 2, is characterized in that, described L1, L2, L3, L4, L5, L6, L7, L8 are respectively 1,2,3,4,5,6,7,0.
7. the paper money number recognition methods of the cash inspecting machine according to claim 1 or 2 or 3 or 4 or 5 or 6, it is characterized in that, Q is 88% to 93%.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510744123.0A CN105389883B (en) | 2015-11-04 | 2015-11-04 | A kind of paper money number recognition methods of cash inspecting machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510744123.0A CN105389883B (en) | 2015-11-04 | 2015-11-04 | A kind of paper money number recognition methods of cash inspecting machine |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105389883A true CN105389883A (en) | 2016-03-09 |
| CN105389883B CN105389883B (en) | 2018-01-12 |
Family
ID=55422127
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510744123.0A Active CN105389883B (en) | 2015-11-04 | 2015-11-04 | A kind of paper money number recognition methods of cash inspecting machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105389883B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106503694A (en) * | 2016-12-01 | 2017-03-15 | 重庆大学 | Digit recognition method based on eight neighborhood feature |
| CN107240185A (en) * | 2017-06-23 | 2017-10-10 | 深圳怡化电脑股份有限公司 | A kind of crown word number identification method, device, equipment and storage medium |
| CN107657251A (en) * | 2016-07-26 | 2018-02-02 | 阿里巴巴集团控股有限公司 | Determine the device and method of identity document display surface, image-recognizing method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2007108826A (en) * | 2005-10-11 | 2007-04-26 | National Printing Bureau | Reading method for organizer having punched hole group imparted with code information, and reader therefor |
| EP2372310A2 (en) * | 2010-03-31 | 2011-10-05 | Aisin Aw Co., Ltd. | Image processing system and position measurement system |
| JP2011238056A (en) * | 2010-05-11 | 2011-11-24 | National Printing Bureau | Authenticity discrimination method of image forming material |
| CN102800148A (en) * | 2012-07-10 | 2012-11-28 | 中山大学 | RMB sequence number identification method |
| CN103914675A (en) * | 2014-03-17 | 2014-07-09 | 东华大学 | Garment QD code recognition method |
| CN104408814A (en) * | 2014-12-13 | 2015-03-11 | 天津远目科技有限公司 | Method for identifying RMB code |
-
2015
- 2015-11-04 CN CN201510744123.0A patent/CN105389883B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2007108826A (en) * | 2005-10-11 | 2007-04-26 | National Printing Bureau | Reading method for organizer having punched hole group imparted with code information, and reader therefor |
| EP2372310A2 (en) * | 2010-03-31 | 2011-10-05 | Aisin Aw Co., Ltd. | Image processing system and position measurement system |
| JP2011238056A (en) * | 2010-05-11 | 2011-11-24 | National Printing Bureau | Authenticity discrimination method of image forming material |
| CN102800148A (en) * | 2012-07-10 | 2012-11-28 | 中山大学 | RMB sequence number identification method |
| CN103914675A (en) * | 2014-03-17 | 2014-07-09 | 东华大学 | Garment QD code recognition method |
| CN104408814A (en) * | 2014-12-13 | 2015-03-11 | 天津远目科技有限公司 | Method for identifying RMB code |
Non-Patent Citations (2)
| Title |
|---|
| 梁杨 等: ""纸币冠字号预处理及组合特征识别方法"", 《计算机工程与设计》 * |
| 王静娇 等: ""基于TMS320DM642的人民币图像特征识别系统"", 《数据采集与处理》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107657251A (en) * | 2016-07-26 | 2018-02-02 | 阿里巴巴集团控股有限公司 | Determine the device and method of identity document display surface, image-recognizing method |
| CN106503694A (en) * | 2016-12-01 | 2017-03-15 | 重庆大学 | Digit recognition method based on eight neighborhood feature |
| CN107240185A (en) * | 2017-06-23 | 2017-10-10 | 深圳怡化电脑股份有限公司 | A kind of crown word number identification method, device, equipment and storage medium |
| CN107240185B (en) * | 2017-06-23 | 2019-09-20 | 深圳怡化电脑股份有限公司 | A kind of crown word number identification method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105389883B (en) | 2018-01-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Qiao et al. | Lgpma: Complicated table structure recognition with local and global pyramid mask alignment | |
| CN110084292B (en) | Target detection method based on DenseNet and multi-scale feature fusion | |
| Ma et al. | Arbitrary-oriented scene text detection via rotation proposals | |
| Sirmacek et al. | Urban-area and building detection using SIFT keypoints and graph theory | |
| JP5837205B2 (en) | Text detection using image area | |
| CN106600600A (en) | Wafer defect detection method based on characteristic matching | |
| CN104464079B (en) | Multiple Currencies face amount recognition methods based on template characteristic point and topological structure thereof | |
| CN109685013B (en) | Method and device for detecting head key points in human body posture recognition | |
| CN107145829B (en) | Palm vein identification method integrating textural features and scale invariant features | |
| CN101520783B (en) | Method and device for searching keywords based on image content | |
| CN105069811A (en) | Multi-temporal remote sensing image change detection method | |
| CN109426814A (en) | A kind of positioning of the specific plate of invoice picture, recognition methods, system, equipment | |
| Daixian | SIFT algorithm analysis and optimization | |
| Zhu et al. | Deep residual text detection network for scene text | |
| CN109697441A (en) | A kind of object detection method, device and computer equipment | |
| Neuhausen et al. | Automatic window detection in facade images | |
| CN105389883A (en) | Banknote prefix number identification method for currency detector | |
| CN105405204A (en) | Banknote crown word number recognition method of currency detector | |
| CN104282001A (en) | Method for enhancing image feature two-value descriptor performance | |
| Liu et al. | Extraction of earthquake-induced collapsed buildings from bi-temporal VHR images using object-level homogeneity index and histogram | |
| CN104036494A (en) | Fast matching computation method used for fruit picture | |
| Wang et al. | Detection of circular oil tanks based on the fusion of SAR and optical images | |
| CN101980296A (en) | Spectral de-aliasing-based Yellow River mainstream line detection method | |
| Yu et al. | A novel algorithm in buildings/shadow detection based on Harris detector | |
| CN109636838A (en) | A kind of combustion gas Analysis of Potential method and device based on remote sensing image variation detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210827 Address after: 310000 No. 66 Dongxin Avenue, Binjiang District, Hangzhou City, Zhejiang Province Patentee after: EASTCOM Inc. Address before: 310000 A318, R & D building, Dongxin City, No. 66, Dongxin Avenue, Binjiang District, Hangzhou City, Zhejiang Province Patentee before: EASTCOM Inc. Patentee before: Hangzhou Eastcom Financial Technology Service Co.,Ltd. |
|
| TR01 | Transfer of patent right |