CN110533003A - A kind of threading method license plate number recognizer and equipment - Google Patents
A kind of threading method license plate number recognizer and equipment Download PDFInfo
- Publication number
- CN110533003A CN110533003A CN201910842942.7A CN201910842942A CN110533003A CN 110533003 A CN110533003 A CN 110533003A CN 201910842942 A CN201910842942 A CN 201910842942A CN 110533003 A CN110533003 A CN 110533003A
- Authority
- CN
- China
- Prior art keywords
- region
- value
- cnt
- license plate
- layer
- 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
Classifications
-
- 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
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/625—License plates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
Abstract
The invention belongs to a kind of digital recognizer and equipment, are related to image real time transfer, which identifies suitable for license plate number.A kind of threading method license plate number recognizer and equipment include the following steps: to acquire license plate image;Collected license plate image is subjected to gray scale and binarization operation operation;It realizes License Plate, obtains location data;Realize that each number is divided into region 1, region 2, region 3 and region 4 from top to bottom in the second frame image by location data, and reduced value is set 4 effective coverages respectively, and successively " 1 " in every a line image data in region 1, region 2, region 3 and region 4 is counted respectively, the maximum value or minimum value for taking count value are as characteristic value, it is compared with the reduced value in the region, and then identifies number.The present invention counts four row image datas, and fast speed, precision are higher, has bigger range compared to single data line of only choosing, reliability is higher.
Description
Technical field
The invention belongs to a kind of digital recognizer and equipment, are related to image real time transfer, which is suitable for license plate number
Word identification.
Background technique
It is proposed first FPGA of the world from Xilinx company in 1984, so far the time less than 40 years.But it is set because it has
Count that the period is short, programmable, calculating speed is fast, low-power consumption and low cost advantage, certain fields be substituted asic chip at
For mainstream.It in recent years, is core as the development of the technologies such as image processing techniques, artificial intelligence and big data results in FPGA
Data processing system be developed rapidly.Currently, Xinlinx and Intel occupies absolute advantage in the market FPGA.This hair
Quartus II software and the DE1_SOC exploitation building the realization with algorithm and be all based on Intel Company of bright middle identifying system
Plate.
Germany scientist Tansheck proposes the concept of optical character identification (OCR) in nineteen twenty-nine, with image procossing skill
The development of art and society, the recognizer of printing digital such as license plate number tend to be mature and have very big application space.Mesh
Preceding number recognizer mainly has three classes: one is template matching method;The second is various neural network algorithms, including convolutional Neural
Network, impulsive neural networks etc.;The third is character feature extraction method.Template matching method is by being pre-stored corresponding template, so
The data of acquisition and template are compared afterwards, to obtain a result, the anti-interference ability of template matching method is low at present, poor accuracy,
Application range is narrow, is no longer satisfied most of identifications and requires;Although and neural network algorithm resolution is very high, needs to acquire
A large amount of training data obtains accurate parameter by these training datas, considerably increases hardware cost;Printing digital is known
Feature extraction is not used not generally, since printing digital has notable feature, so algorithm accuracy of identification with higher.Most
Common character feature extraction method is threading method, and threading method needs to extract characteristic image specifically a few row data and a few column datas,
By calculating the situation of change (rising edge of 0-1 and the failing edge of 1-0) of data on these row and columns, then compare row data variation
Position and column data variation position obtain judgement as a result, this feature extraction be undoubtedly hardware system identification block letter number
The method of word most convenient and efficient.
Traditional method of scoring generally draws horizontal line and vertical line, and horizontal line represents the row data for choosing image, and vertical line, which represents, chooses figure
The column data of picture.The two row data and a column data at least needing to extract image, by judging these data variations and comparing hair
The position for changing obtains judging result, and image data first is to transmit by row, while extracting row data and column data more
Trouble, needs to cache image data, influences the speed of identification, meanwhile, it is affected by noise larger, it needs with high accuracy
Image-region where each number is cut and positioned.
Summary of the invention
The present invention is directed in view of the above-mentioned problems, proposing a kind of threading method license plate number recognizer.
Technical program of the present invention lies in:
A kind of threading method license plate number recognizer, includes the following steps:
Acquire license plate image;
Collected license plate image is subjected to image preprocessing;
It is characterized by: the detailed process of described image processing are as follows:
(1) image preprocessing: collected license plate image is subjected to gray scale and binarization operation, obtains binary image data;
Image is set to be converted into the black white image of only data " 0 " and " 1 ";Wherein, digital region is white, license plate background colour blue
For black;
(2) image recognition:
1) it realizes License Plate, obtains location data;
The width of position and license plate of the license plate in entire binary image data is obtained with the first frame image data;
2) realize that the region of each number divides in the second frame image by location data;
3) each number is divided into 4 effective coverages from top to bottom and carries out data identification;Respectively region 1, region 2, region 3
And region 4, and reduced value is set 4 effective coverages respectively, and successively to each in region 1, region 2, region 3 and region 4
" 1 " in row image data is counted respectively, and the maximum value or minimum value for taking count value are as in characteristic value, with the region
Reduced value is compared, and then identifies number.
Preferably, in region 1,3,4, choose the maximum value of count value as characteristic value, be denoted as respectively cnt_1, cnt_3 and
cnt_4;
In region 2, the minimum value of count value is chosen as characteristic value, is denoted as cnt_2.
It is highly preferred that each number is divided into the detailed process that 4 effective coverages carry out data identification from top to bottom are as follows:
Region 1: " 1 " in every a line image data is counted respectively;It takes maximum value therein as characteristic value, is denoted as
cnt_1;Set reduced value layer_1;Compare the size of cnt_1 and layer_1;If cnt_1 > layer_1, remembers instruction value
It is 00;If cnt_1 < layer_1, remember that instruction value is 11;
Region 2: " 1 " in every a line image data is counted respectively;It takes minimum value therein as characteristic value, is denoted as
cnt_2;Set reduced value layer_2;Compare the size of cnt_2 and layer_2;If cnt_2 < layer_2, remembers instruction value
It is 1;If cnt_2 > layer_2, remember that instruction value is 0;Instruction value is updated at this time;If instruction value includes unique value, this is right
Answer the number of instruction value identified, i.e., number 4 is identified;
Region 3: unidentified number out is grouped according to same instructions value;
First group: number 0,8,9;Instruction value is 000;
Second group: number 1,6;Instruction value is 111;
Third group: number 2,3,5,7;Instruction value is 001;
And 3 reduced values are set for grouping situation, it is denoted as layer_3_0, layer_3_1 and layer_3_2 respectively;
" 1 " in every a line image data is counted respectively;It takes maximum value therein as characteristic value, is denoted as cnt_3;It is logical
It crosses layer_3_0 differentiation to distinguish 1,2,7 and other numbers, while may recognize that number 6;Remember cnt_3 < layer_3_0
It is 10, layer_3_1 < cnt_3 < layer_3_2 for 11, layer_3_0 < cnt_3 < layer_3_1 is 00, cnt_3 >
Layer_3_2 is 01;Identify 0,1,3,4,5 and 6;Remaining 2,7 and 8,9;
Region 4: " 1 " in every a line image data is counted respectively;It takes maximum value therein as characteristic value, is denoted as
cnt_4;Set reduced value layer_4;It is 1 that note cnt_4 > layer_4, which is 0, cnt_4 < layer_4,;To distinguish 2,7,8 and 9.
Or preferably, it realizes License Plate, obtains the specific implementation step of location data are as follows: the first frame image data arrives
When coming, " 0 " in every data line is counted, the black portions in " 0 " representative image, when the number of wherein " 0 " of a line
Amount can determine that this row is the position that license plate starts when being greater than the set value, the quantity of " 0 " in this line image data is also
It is the width of license plate.
A kind of threading method license plate digital identifier device uses threading method license plate digit recognition method as described above, knot
Structure includes the video acquisition module being sequentially connected electrically, FPGA image processing module and display module;Video acquisition module is acquisition
Size is the OV5640 camera of the image of 1024*720;Display module is display;FPGA image processing module include with
Two branches of video acquisition module connection, one article of branch includes sequentially connected first fifo, sdram and the 3rd fifo, separately
One article of branch includes sequentially connected preprocessing module, the 2nd fifo, algorithm realization module and LED;Wherein, the first fifo also with
Preprocessing module connection, the 3rd fifo are connect with display.
The technical effects of the invention are that:
1, this algorithm counts four row image datas, can complete in data transmission procedure using four horizontal lines are drawn, speed
Degree is very fast;
2, this algorithm is in accuracy, due to being to count the number of " 1 " in four row image datas rather than judge rising edge under
Drop can reduce influence of the noise to algorithm accuracy along position;
3, the selection of four characteristic values in this algorithm is then selected by counting to every data line in four regions
Maximum or minimum value has bigger range compared to single data line of only choosing, and reliability is higher.
Detailed description of the invention
Fig. 1 is present system structure chart.
Fig. 2 is that inventive algorithm realizes schematic diagram.
Specific embodiment
Embodiment 1
A kind of threading method license plate number recognizer, includes the following steps:
(1) license plate image is acquired;
(2) collected license plate image is subjected to image preprocessing;Detailed process are as follows:
Image preprocessing: collected license plate image is subjected to gray scale and binarization operation, obtains binary image data;Make figure
Black white image as being converted into only data " 0 " and " 1 ";Wherein, digital region is white, and license plate background colour blue is black
Color;
Image recognition:
1) it realizes License Plate, obtains location data;
The width of position and license plate of the license plate in entire binary image data is obtained with the first frame image data;
2) realize that the region of each number divides in the second frame image by location data;
3) each number is divided into 4 effective coverages from top to bottom and carries out data identification;Respectively region 1, region 2, region 3
And region 4, and reduced value is set 4 effective coverages respectively, and successively to each in region 1, region 2, region 3 and region 4
" 1 " in row image data is counted respectively, and the maximum value or minimum value for taking count value are as in characteristic value, with the region
Reduced value is compared, and then identifies number.
Embodiment 2
A kind of threading method license plate number recognizer, includes the following steps:
(1) license plate image is acquired by the OV5640 camera of video acquisition module, image size is 1024*720, data format
For RGB565;
(2) collected license plate image is subjected to image preprocessing;Detailed process are as follows:
Image preprocessing: collected license plate image is subjected to gray scale and binarization operation, obtains binary image data;To scheme
It whether is more than this threshold value as the R component of data is as gradation data, then by one threshold decision R data of setting, in image
The R component of data be greater than threshold value when obtain new image data " 1 ", less than when obtain " 0 ", entire image will be converted into only
The black white image of data " 0 " and " 1 ";For license plate, image-region where number is displayed in white, and blue bottom region shows black;
(3) width of position and license plate of the license plate in whole image is counted by the first frame image data;Detailed process are as follows:
When first frame image data arrives, " 0 " in every data line is counted, the black portions in " 0 " representative image,
When the quantity of wherein " 0 " of a line is greater than the set value, that is, it can determine that this row is the position that license plate starts, this line picture number
The quantity of " 0 " in i.e. the width of license plate;
(4) realize that the region of each number divides in the second frame image;Detailed process are as follows:
Each number is divided into 4 effective coverages from top to bottom and carries out data identification;Respectively region 1, region 2, region 3 and
Region 4;
(5) data identification successively is carried out to each region;Detailed process are as follows:
Region 1: it to " 1 " in every a line image data, i.e. image white part, is counted respectively;All row meters in region 1
It counts up into later, takes maximum value therein as characteristic value, be denoted as cnt_1;Set reduced value layer_1;Compare cnt_1 and
The size of layer_1;If cnt_1 > layer_1, remember that instruction value is 00;If cnt_1 < layer_1 remembers that instruction value is
11;
Since the cnt_1 of number 1,4,6 is significantly less than other numbers, therefore each digital decision instruction front two is as follows:
;
Region 2: " 1 " in every a line image data is counted respectively;It takes minimum value therein as characteristic value, is denoted as
cnt_2;Set reduced value layer_2;Compare the size of cnt_2 and layer_2;If cnt_2 < layer_2, remembers instruction value
It is 1;If cnt_2 > layer_2, remember that instruction value is 0;Instruction value is updated at this time;It is as follows:
;
If instruction value includes unique value, the number of the corresponding instruction value is identified;Wherein, the decision instruction of number 4 is
110, all digital 4 are identified;
Region 3: unidentified number out is grouped according to same instructions value;
First group: number 0,8,9;Instruction value is 000;
Second group: number 1,6;Instruction value is 111;
Third group: number 2,3,5,7;Instruction value is 001;
And 3 reduced values are set for grouping situation, it is denoted as layer_3_0, layer_3_1 and layer_3_2 respectively;To each
" 1 " in row image data is counted respectively;It takes maximum value therein as characteristic value, is denoted as cnt_3;
The cnt_3 gap of second group of number 1,6 is obvious, and has the cnt_3 gap of number 2,7 and number 3,5 obvious and sum number
Word 1 is close, i.e., selection layer_3_0 distinguishes number 1,2,7 and other numbers, while may recognize that number 6;Note
Cnt_3 < layer_3_0 is 11;
The cnt_3 of number 0 is less than the cnt_3 of number 8 and 9 in first group, and note layer_3_0 < cnt_3 < layer_3_1 is
10;
The cnt_3 of number 3 is the cnt_3 for being significantly less than number 5 simultaneously, then distinguished number 3 and 5 with layer_3_2,
It is 01 that note layer_3_1 < cnt_3 < layer_3_2, which is 00, cnt_3 > layer_3_2,;Here due to being that grouping determines, have
A little numbers may meet multiple conditions simultaneously, thus will likely instruction write, to guarantee accuracy, instruction is as follows:
;
In upper table as can be seen that having identified 0,1,3,4,5 and 6 herein;Residue 2,7 and 8,9 is unidentified out;
Region 4: " 1 " in every a line image data is counted respectively;It takes maximum value therein as characteristic value, is denoted as
Cnt_4 sets reduced value layer_4;Wherein the cnt_4 of number 2,8 is significantly greater than the cnt_4 of number 7,9, remembers cnt_4 >
Layer_4 is that 0, cnt_4 < layer_4 is 1;Following decision instruction can then be obtained:
;
As above, each number corresponds to different identification instructions, i.e., each number can be identified.
Embodiment 3
A kind of threading method license plate digital identifier device uses threading method license plate digit recognition method as described above, structure packet
Include the video acquisition module being sequentially connected electrically, FPGA image processing module and display module;Video acquisition module is acquisition size
For the OV5640 camera of the image of 1024*720;Display module is display;FPGA image processing module includes and video
Two branches of acquisition module connection, one article of branch includes sequentially connected first fifo, sdram and the 3rd fifo, and another
Branch includes sequentially connected preprocessing module, the 2nd fifo, algorithm realization module and LED;Wherein, the first fifo also locates with pre-
Module connection is managed, the 3rd fifo is connect with display.
Claims (5)
1. a kind of threading method license plate number recognizer, includes the following steps:
Acquire license plate image;
Collected license plate image is subjected to image preprocessing;
It is characterized by: the detailed process of described image processing are as follows:
(1) image preprocessing: collected license plate image is subjected to gray scale and binarization operation, obtains binary image data;
Image is set to be converted into the black white image of only data " 0 " and " 1 ";Wherein, digital region is white, license plate background colour blue
For black;
(2) image recognition:
1) it realizes License Plate, obtains location data;
The width of position and license plate of the license plate in entire binary image data is obtained with the first frame image data;
2) realize that the region of each number divides in the second frame image by location data;
3) each number is divided into 4 effective coverages from top to bottom and carries out data identification;Respectively region 1, region 2, region 3
And region 4, and reduced value is set 4 effective coverages respectively, and successively to each in region 1, region 2, region 3 and region 4
" 1 " in row image data is counted respectively, and the maximum value or minimum value for taking count value are as in characteristic value, with the region
Reduced value is compared, and then identifies number.
2. threading method license plate number recognizer according to claim 1, includes the following steps:
In region 1,3,4, the maximum value of count value is chosen as characteristic value, is denoted as cnt_1, cnt_3 and cnt_4 respectively;
In region 2, the minimum value of count value is chosen as characteristic value, is denoted as cnt_2.
3. threading method license plate number recognizer according to claim 2 includes the following steps: each number from top to bottom
It is divided into the detailed process that 4 effective coverages carry out data identification are as follows:
Region 1: " 1 " in every a line image data is counted respectively;It takes maximum value therein as characteristic value, is denoted as
cnt_1;Set reduced value layer_1;Compare the size of cnt_1 and layer_1;If cnt_1 > layer_1, remembers instruction value
It is 00;If cnt_1 < layer_1, remember that instruction value is 11;
Region 2: " 1 " in every a line image data is counted respectively;It takes minimum value therein as characteristic value, is denoted as
cnt_2;Set reduced value layer_2;Compare the size of cnt_2 and layer_2;If cnt_2 < layer_2, remembers instruction value
It is 1;If cnt_2 > layer_2, remember that instruction value is 0;Instruction value is updated at this time;If instruction value includes unique value, this is right
Answer the number of instruction value identified, i.e., number 4 is identified;
Region 3: unidentified number out is grouped according to same instructions value;
First group: number 0,8,9;Instruction value is 000;
Second group: number 1,6;Instruction value is 111;
Third group: number 2,3,5,7;Instruction value is 001;
And 3 reduced values are set for grouping situation, it is denoted as layer_3_0, layer_3_1 and layer_3_2 respectively;
" 1 " in every a line image data is counted respectively;It takes maximum value therein as characteristic value, is denoted as cnt_3;It is logical
It crosses layer_3_0 differentiation to distinguish 1,2,7 and other numbers, while may recognize that number 6;Remember cnt_3 < layer_3_0
It is 10, layer_3_1 < cnt_3 < layer_3_2 for 11, layer_3_0 < cnt_3 < layer_3_1 is 00, cnt_3 >
Layer_3_2 is 01;Identify 0,1,3,4,5 and 6;Remaining 2,7 and 8,9;
Region 4: " 1 " in every a line image data is counted respectively;It takes maximum value therein as characteristic value, is denoted as
cnt_4;Set reduced value layer_4;It is 1 that note cnt_4 > layer_4, which is 0, cnt_4 < layer_4,;To distinguish 2,7,8 and 9.
4. threading method license plate number recognizer according to claim 1, it is characterised in that: realize License Plate, determined
The specific implementation step of position data are as follows: when the first frame image data arrives, " 0 " in every data line is counted, " 0 " generation
Black portions in table image can determine that this row is that license plate starts when the quantity of wherein " 0 " of a line is greater than the set value
Position, the quantity of " 0 " in this line image data i.e. the width of license plate.
5. a kind of threading method license plate digital identifier device, it is characterised in that: using described in claims 1 or 2 as above or 3 or 4
Threading method license plate digit recognition method, structure include the video acquisition module being sequentially connected electrically, FPGA image processing module and
Display module;It is characterized by: video acquisition module is the OV5640 camera for acquiring the image that size is 1024*720;Display
Module is display;FPGA image processing module includes two branches connecting with video acquisition module, and a branch includes
Sequentially connected first fifo, sdram and the 3rd fifo, another branch include sequentially connected preprocessing module, second
Fifo, algorithm realize module and LED;Wherein, the first fifo is also connect with preprocessing module, and the 3rd fifo is connect with display.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910842942.7A CN110533003B (en) | 2019-09-06 | 2019-09-06 | Threading method license plate number recognition method and equipment |
AU2020102152A AU2020102152A4 (en) | 2019-09-06 | 2020-09-06 | Algorithm and Device for License Plate Number Recognition Based on Line Drawing |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910842942.7A CN110533003B (en) | 2019-09-06 | 2019-09-06 | Threading method license plate number recognition method and equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110533003A true CN110533003A (en) | 2019-12-03 |
CN110533003B CN110533003B (en) | 2022-09-20 |
Family
ID=68667645
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910842942.7A Active CN110533003B (en) | 2019-09-06 | 2019-09-06 | Threading method license plate number recognition method and equipment |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN110533003B (en) |
AU (1) | AU2020102152A4 (en) |
Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101000653A (en) * | 2006-12-31 | 2007-07-18 | 沈阳工业大学 | Automatic recognising method and automatic recognising recording system for number of paper money |
US20090185236A1 (en) * | 2008-01-17 | 2009-07-23 | Zeng Yijun | Image binarization using dynamic sub-image division |
US20100108754A1 (en) * | 2008-11-04 | 2010-05-06 | Jerome Kahn | Thread identification system |
CN102306283A (en) * | 2011-07-23 | 2012-01-04 | 浙江大学 | Image identification method of indicating values of LCD (liquid crystal display) digital display instrument |
US20130086328A1 (en) * | 2011-06-13 | 2013-04-04 | Paneve, Llc | General Purpose Digital Data Processor, Systems and Methods |
CN103593653A (en) * | 2013-11-01 | 2014-02-19 | 浙江工业大学 | Character two-dimensional bar code recognition method based on scanning gun |
CN105335745A (en) * | 2015-11-27 | 2016-02-17 | 小米科技有限责任公司 | Recognition method, device and equipment for numbers in images |
CN105447508A (en) * | 2015-11-10 | 2016-03-30 | 上海珍岛信息技术有限公司 | Identification method and system for character image verification codes |
US20170124719A1 (en) * | 2015-10-30 | 2017-05-04 | Xiaomi Inc. | Method, device and computer-readable medium for region recognition |
CN106709484A (en) * | 2015-11-13 | 2017-05-24 | 国网吉林省电力有限公司检修公司 | Number identification method of digital instrument |
CN106875546A (en) * | 2017-02-10 | 2017-06-20 | 大连海事大学 | A kind of recognition methods of VAT invoice |
CN107704855A (en) * | 2017-09-28 | 2018-02-16 | 中北大学 | A kind of liquid crystal digital display instrument digital recognition apparatus |
CN110084241A (en) * | 2019-05-05 | 2019-08-02 | 山东大学 | A kind of ammeter automatic reading method based on image recognition |
-
2019
- 2019-09-06 CN CN201910842942.7A patent/CN110533003B/en active Active
-
2020
- 2020-09-06 AU AU2020102152A patent/AU2020102152A4/en active Active
Patent Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101000653A (en) * | 2006-12-31 | 2007-07-18 | 沈阳工业大学 | Automatic recognising method and automatic recognising recording system for number of paper money |
US20090185236A1 (en) * | 2008-01-17 | 2009-07-23 | Zeng Yijun | Image binarization using dynamic sub-image division |
US20100108754A1 (en) * | 2008-11-04 | 2010-05-06 | Jerome Kahn | Thread identification system |
US20130086328A1 (en) * | 2011-06-13 | 2013-04-04 | Paneve, Llc | General Purpose Digital Data Processor, Systems and Methods |
CN102306283A (en) * | 2011-07-23 | 2012-01-04 | 浙江大学 | Image identification method of indicating values of LCD (liquid crystal display) digital display instrument |
CN103593653A (en) * | 2013-11-01 | 2014-02-19 | 浙江工业大学 | Character two-dimensional bar code recognition method based on scanning gun |
US20170124719A1 (en) * | 2015-10-30 | 2017-05-04 | Xiaomi Inc. | Method, device and computer-readable medium for region recognition |
CN105447508A (en) * | 2015-11-10 | 2016-03-30 | 上海珍岛信息技术有限公司 | Identification method and system for character image verification codes |
CN106709484A (en) * | 2015-11-13 | 2017-05-24 | 国网吉林省电力有限公司检修公司 | Number identification method of digital instrument |
CN105335745A (en) * | 2015-11-27 | 2016-02-17 | 小米科技有限责任公司 | Recognition method, device and equipment for numbers in images |
CN106875546A (en) * | 2017-02-10 | 2017-06-20 | 大连海事大学 | A kind of recognition methods of VAT invoice |
CN107704855A (en) * | 2017-09-28 | 2018-02-16 | 中北大学 | A kind of liquid crystal digital display instrument digital recognition apparatus |
CN110084241A (en) * | 2019-05-05 | 2019-08-02 | 山东大学 | A kind of ammeter automatic reading method based on image recognition |
Non-Patent Citations (5)
Title |
---|
NAN DONG: ""Digital recognition based on sewing method"", 《 2018 INTERNATIONAL CONFERENCE ON ADVANCED MECHATRONIC SYSTEMS (ICAMECHS)》 * |
于海跃: ""工厂仪表数码管数字的识别"", 《电子世界》 * |
刘文亮: ""七段式数显仪表中数字识别的研究与实现"", 《信息科技》 * |
宗亚辉: ""印刷体数学表达式识别技术的研究"", 《信息科技》 * |
裴孟菲: ""基于数字识别和压缩传输的燃气表自动抄送系统"", 《信息科技-工程科技Ⅱ辑》 * |
Also Published As
Publication number | Publication date |
---|---|
CN110533003B (en) | 2022-09-20 |
AU2020102152A4 (en) | 2020-10-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107016405B (en) | A kind of pest image classification method based on classification prediction convolutional neural networks | |
CN108562589A (en) | A method of magnetic circuit material surface defect is detected | |
CN110276386A (en) | A kind of apple grading method and system based on machine vision | |
CN101916364B (en) | Adaptive dial digital identification method | |
CN112132843A (en) | Hematoxylin-eosin staining pathological image segmentation method based on unsupervised deep learning | |
CN110663971B (en) | Red date quality classification method based on double-branch deep fusion convolutional neural network | |
CN108596891A (en) | A kind of method of counting towards multiple types mixing silk cocoon | |
CN109271991A (en) | A kind of detection method of license plate based on deep learning | |
CN108388905B (en) | A kind of Illuminant estimation method based on convolutional neural networks and neighbourhood context | |
CN108154195A (en) | Tealeaves recognition methods and the tealeaves sorting equipment using this method | |
CN105513066B (en) | It is a kind of that the generic object detection method merged with super-pixel is chosen based on seed point | |
CN106934386A (en) | A kind of natural scene character detecting method and system based on from heuristic strategies | |
CN109740572A (en) | A kind of human face in-vivo detection method based on partial color textural characteristics | |
CN106157323A (en) | The insulator division and extracting method that a kind of dynamic division threshold value and block search combine | |
CN105095907B (en) | A kind of heterogeneous fibre identification method of cotton based on RBF neural | |
CN107480682A (en) | A kind of commodity packaging date of manufacture detection method | |
CN107194348A (en) | The domain color recognition methods of target object in a kind of image | |
CN109145848A (en) | A kind of wheat head method of counting | |
CN106932408A (en) | A kind of paper cocooning frame pollution silk cocoon detection method based on image procossing | |
CN106872473A (en) | A kind of potato defects detection identifying system design based on machine vision | |
CN106709474A (en) | Handwritten telephone number identification, verification and information sending system | |
CN110348505A (en) | Vehicle color disaggregated model training method, device and vehicle color identification method | |
CN116052222A (en) | Cattle face recognition method for naturally collecting cattle face image | |
CN116416613A (en) | Citrus fruit identification method and system based on improved YOLO v7 | |
CN116310548A (en) | Method for detecting invasive plant seeds in imported seed products |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |