CN106778754A - A kind of industrial ammeter digit recognition method of robust - Google Patents
A kind of industrial ammeter digit recognition method of robust Download PDFInfo
- Publication number
- CN106778754A CN106778754A CN201611031147.2A CN201611031147A CN106778754A CN 106778754 A CN106778754 A CN 106778754A CN 201611031147 A CN201611031147 A CN 201611031147A CN 106778754 A CN106778754 A CN 106778754A
- Authority
- CN
- China
- Prior art keywords
- image
- ammeter
- numeral
- robust
- digit
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/14—Image acquisition
- G06V30/148—Segmentation of character regions
- G06V30/153—Segmentation of character regions using recognition of characters or words
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of industrial ammeter digit recognition method of robust, belong to image procossing, mode identification technology.The invention mainly comprises two parts content：Digital segmentation and numeral are recognized.First, the digital edge in industrial electro surface plate is detected by Canny operators；Can effectively be positioned by sciagraphy again and cut out numeric area；Finally, based on the numeral in training set, by KNN algorithm construction graders, numeral to be identified is classified.Industrial ammeter digit recognition method of the invention, noise region and numeric area can be efficiently differentiated by Canny rim detections so that algorithm more robust.Meanwhile, efficiency of algorithm can not only be effectively improved using KNN algorithms when numeral is recognized, can also cause that numeral identification is more accurate.
Description
Technical field
The invention belongs to Digital Image Processing, mode identification technology, the industrial ammeter numeral of particularly a kind of robust
Recognition methods.
Background technology
From 19th century electricity generation, and 20th century electricity commonly use beginning.In industrial quarters, ammeter is just as a kind of
Important instrument, for showing the various data of generation department in enterprise.Especially in power industry, power department is in order to unite
The electricity consumption data of each instrument is counted, often arranges special messenger to carry out work of checking meter.In existing big electrometer, having much be located at remote
, inconvenience, the position of high-risk.This is accomplished by expending substantial amounts of manpower and materials arrangement workman checks meter, and due to workman's
Easily there is meter reading mistake in long-term work.Certainly, as power industry, information industry are continued to develop, large quantities of intelligence electricity
Table is developed, and this ammeter is capable of the numeral of automatic identification itself and and can be transferred to control system.However, it is desirable to
All change original old-fashioned ammeter into limitation that new ammeter is frequently subjected to economy, therefore, such ammeter is not widely used.
The low discrimination that digital recognizer in existing some industrial ammeters is present, not the shortcomings of robustness.
The content of the invention
The technical problems to be solved by the invention are：A kind of industrial ammeter digit recognition method of robust is provided, is passed through
Canny detections edge algorithms can effectively distinguish noise region and numeric area so that the numeral cutting in industrial electro surface plate
It is more accurate；The efficiency of digital recognizer can be not only effectively improved by using KNN algorithms, digital identification is also can guarantee that
Precision.
The present invention uses following technical scheme to solve above-mentioned technical problem：A kind of industrial ammeter numeral identification side of robust
Method, comprises the following steps：
Step 1, the industrial ammeter image I_rgb that comes will be gathered first be converted to gray level image I_gray, in gray-scale map
As upper use mean filter, preliminary denoising is carried out.
Step 2, the gray level image I_noise to having removed noise use the edge of Canny operator extractions industry ammeter image.
Canny operators carry out rim detection can effectively suppress noise, while the position at edge can accurately be determined.
Step 3, based on the edge image I_dilate after expansion, it is first unexpectedly possible logical to the region containing numeral in image
The experience of mistake carries out coarse positioning.Then precise positioning cutting is carried out to numeric area by horizontal, longitudinal projection's method respectively.
Step 4, precise positioning is cut after the gray level image I_gray containing numeral binaryzation carried out by threshold method obtain
To bianry image I_binary；
Step 5, further by sciagraphy, only carry out longitudinal projection, detailed process such as step 3 cuts numeric area
It is the zonule of individual digit,
Step 6, based on KNN (k nearest neighbor) algorithm, individual digit is recognized, finally consider scaling position by step 5, obtain
To final ammeter reading.
Preferably, Canny edge detection algorithms flow is described in step 2：
Step1：Initial pictures are smoothed with Gaussian filter, similar to the mean filter in step one；
Step2：Amplitude and the direction of gradient are calculated with the finite difference of single order local derviation；
Step3：Non-maxima suppression is carried out to gradient magnitude；
Step4：Detected with dual threashold value-based algorithm and connection edge；
Preferably, it is to the method that test image carries out Classification and Identification using KNN algorithms described in step 6：For each
Test pictures I_test, we find K maximum with test pictures IOU in 10N training set I_train, then at K
In training set, there are most numerals for being test pictures in certain digital classification.
Preferably, the threshold value of the Canny edge detection algorithms in step 2 is 0.6；What the projection vector in step 3 cut
Threshold value thresh=15；Training set number in step 6：The K values of N=10, KNN algorithm are K=10.
The present invention compared with prior art, has the advantages that：1) present invention proposes a kind of efficient work of robust
Industry ammeter digit recognition method, the digital edge in industrial electro surface plate is detected by canny operators, can effectively be positioned and be cut
Cut out numeric area；2) based on the numeral in training set, by KNN algorithm construction graders, numeral to be identified is classified；
3) industrial ammeter digit recognition method of the invention can efficiently differentiate noise region and digital block by canny rim detections
Domain so that algorithm more robust；4) efficiency of algorithm can not only be effectively improved using KNN algorithms when numeral is recognized, can be also made
Obtain digital identification more accurate.
Brief description of the drawings
Fig. 1 is the artwork of industry ammeter of the invention.
Fig. 2 is the gray-scale map after the denoising that step one of the present invention is calculated.
Fig. 3 is the image border that step 2 Canny edge detection algorithms of the present invention are detected.
Fig. 4 is the image after step 2 of the present invention expands to the image border that Canny is detected.
Fig. 5 is the area image containing numeral that step 3 of the present invention cuts out.
Fig. 6 is that gray level image of the step 4 of the present invention to cutting out carries out the binary map that binaryzation is obtained.
Fig. 7 is that the present invention carries out the result figure that individual digit cuts and recognizes to binary map.
Specific embodiment
With reference to accompanying drawing, a kind of industrial ammeter digit recognition method of robust of the invention, the numeral identification in industrial ammeter
Algorithm mainly applies the Image Processing and Pattern Recognition related algorithm such as image preprocessing, image segmentation, character recognition.Including with
Lower step：
Step 1, the industrial ammeter image I_rgb that will be collected are converted to gray level image I_gray, and place is filtered afterwards
Reason, removes noise；Mean filter core during filtering process is：The process of mean filter is convolution
Calculating process：In formula,
I, j have been the I_noise transverse and longitudinal subscripts of noise image, and m, n are the transverse and longitudinal subscript of mean filter core K.
Step 2, the gray level image I_noise to having removed noise use the edge of Canny operator extractions industry ammeter image；
And morphological dilation treatment is carried out to edge image；
Use Canny operator extractions industry ammeter image edge flow for：
Step1：Initial pictures are smoothed with Gaussian filter；
Step2：Amplitude and direction with the finite difference formulations gradient of single order local derviation；
Step3：Non-maxima suppression is carried out to gradient magnitude；
Step4：Detected with dual threashold value-based algorithm and connection edge；
The image I_dilate computing formula after morphological dilation is expanded are carried out to edge image I_canny
For：Wherein strel (5) is the structural element of expansion.
The threshold value of the Canny edge detection algorithms is 0.6, and the value of the structural element strel (5) of expansion is
Step 3, based on edge image, coarse positioning first is carried out to the region containing numeral in image, then by sciagraphy pair
Numeric area carries out precise positioning cutting；Specially：Based on the edge image I_dilate after expansion, first to containing numeral in image
Region carry out coarse positioning, then respectively by laterally, longitudinal projection's method precise positioning cutting, transverse projection are carried out to numeric area
Vector isLongitudinal projection's vector is
According to two projection vector v, given threshold thresh, calculate up and down four cut points (up, down, left,
right)：=index (v ＜ thresh), then carries out further fine cut and obtains I_fine to numeric area.Projection vector
The threshold value thresh=15 of cutting.
Step 4, precise positioning is cut after containing numeral gray level image binary conversion treatment is carried out by threshold method；
Step 5, by sciagraphy, numeric area is cut into the zonule of individual digit, the single number of identification after having cut
Word lower right corner decimal point；It is specific as follows：
Step 5-1, by sciagraphy, only carry out longitudinal projection, numeric area is cut into the zonule of individual digit, cut
Cutpoint computing formula point:=index (v=0), obtains the image I_single of individual digit after having cut；
Step 5-2, for each individual digit picture for cutting out, continue to use longitudinal projection, obtain single number
The projection vector v_single of word image；
Step 5-3, given threshold th=row (I_single)/5, the cutting position of decimal point is point:=index (v_
Single ＜ th), wherein index () is the subscript computing formula of vector；Calculate the cutting of each digital lower right corner decimal point away from
From the numeral of ultimate range is the numeral containing decimal point, and computing formula is：
Step 6, based on KNN algorithms, individual digit is recognized, and recognize decimal point, obtain final ammeter numeral.Specifically
For：
Step 6-1, by way of above-mentioned steps cut individual digit image training set is made, each numeral makes N
Picture, thus 10 total 10N training set I_train of numeral；
Step 6-2, for the distance function in every test pictures I_test, KNN algorithm it is：
Step 6-3, for each test pictures I_test, we find and test in 10N training set I_train
Picture IOU maximum K, then in K training set, there are most numerals for being test pictures in certain digital classification；
Step 6-4, determine scaling position by the method for step 5, obtain final ammeter reading.
The training set number：The K values of N=10, KNN algorithm are K=10.
The present invention proposes a kind of efficient industry ammeter digit recognition method of robust, and industry is detected by canny operators
Digital edge in ammeter panel, can effectively position and cut out numeric area.
Embodiments of the present invention are described below in detail, the example of the implementation method is shown in the drawings, wherein ad initio
Same or similar element or element with same or like function are represented to same or similar label eventually.Below by ginseng
The implementation method for examining Description of Drawings is exemplary, is only used for explaining the present invention, and is not construed as limiting the claims.
Embodiment
The present invention detects the digital edge in industrial electro surface plate by canny operators first, can effectively position and cut
Cut out numeric area；Then, based on the numeral in training set, by KNN algorithm construction graders, numeral to be identified is divided
Class.Comprise the following steps that：
Step 1, collection comes industrial ammeter image I_rgb first, as shown in figure 1, being converted to gray level image I_
Gray, uses mean filter on gray level image, carries out preliminary denoising, as shown in Figure 2.Mean filter core is：The process of mean filter is convolutional calculation process：
Step 2, the gray level image I_noise to having removed noise use the edge of Canny operator extractions industry ammeter image,
As shown in Figure 3.Canny operators carry out rim detection can effectively suppress noise, while the position at edge can accurately be determined.To edge
Image I_canny carries out the image I_dilate after morphological dilation is expanded, as shown in figure 4, computing formula is：Wherein strel (5) is the structural element of expansion, and its value isCanny edges
The threshold value of detection algorithm is 0.6.
Step 3, based on the edge image I_dilate after expansion, it is first unexpectedly possible to the region containing numeral in image to pass through
Experience carries out coarse positioning.Then respectively by laterally, longitudinal projection's method carry out precise positioning cutting to numeric area, transverse projection to
Measure and beLongitudinal projection's vector is
According to two projection vector v, given threshold thresh, calculate up and down four cut points (up, down, left,
right)：=index (v ＜ thresh), then carries out further fine cut and obtains I_fine, projection vector to numeric area
The threshold value thresh=15 of cutting, as shown in Figure 5.
Step 4, precise positioning is cut after the gray level image I_gray containing numeral binaryzation carried out by threshold method obtain
To bianry image I_binary, as shown in Figure 6；
Step 5, further by sciagraphy, only carry out longitudinal projection, detailed process such as step 3 cuts numeric area
It is the zonule of individual digit, cut point computing formula point:=index (v=0), obtains the figure of individual digit after having cut
As I_single.Consider each individual digit picture for cutting out, continue to use longitudinal projection, obtain individual digit image
Projection vector v_single.Given threshold th=row (I_single)/5, the cutting position of decimal point is point:=
Index (v_single ＜ th), calculates the cutting distance of each digital lower right corner decimal point, the numeral of ultimate range be containing
The numeral of decimal point, computing formula is：
Step 6, based on KNN (k nearest neighbor) algorithm, we first pass through step one to five cutting first is recognized to individual digit
The mode of individual digit image makes training set, and we make N pictures to each numeral, thus 10 total 10N of numerals open instruction
Practice collection I_train.For every test pictures I_test, the distance function in KNN is defined as：Finally consider scaling position by step 5, obtain final ammeter and read
Number, wherein training set number：The K values of N=10, KNN algorithm are K=10, as shown in Figure 7.
From the foregoing, it will be observed that the present invention proposes a kind of efficient industry ammeter digit recognition method of robust, by canny operators
Digital edge in detection industrial electro surface plate, can effectively position and cut out numeric area.
Claims (9)
1. the industrial ammeter digit recognition method of a kind of robust, it is characterised in that comprise the following steps：
Step 1, the industrial ammeter image I_rgb that will be collected are converted to gray level image I_gray, and treatment is filtered afterwards, go
Except noise；
Step 2, the gray level image I_noise to having removed noise use the edge of Canny operator extractions industry ammeter image；And it is right
Edge image carries out morphological dilation treatment；
Step 3, based on edge image, coarse positioning first is carried out to the region containing numeral in image, then by sciagraphy logarithm
Block domain carries out precise positioning cutting；
Step 4, precise positioning is cut after containing numeral gray level image binary conversion treatment is carried out by threshold method；
Step 5, by sciagraphy, numeric area is cut into the zonule of individual digit, identification individual digit is right after having cut
Inferior horn decimal point；
Step 6, based on KNN algorithms, individual digit is recognized, and recognize decimal point, obtain final ammeter numeral.
2. the industrial ammeter digit recognition method of a kind of robust as claimed in claim 1, it is characterised in that：At step 1 filtering
Mean filter core during reason is：The process of mean filter is convolutional calculation process：In formula, i, j are to have gone
The I_noise transverse and longitudinal subscripts of noise image, m, n are the transverse and longitudinal subscript of mean filter core K.
3. the industrial ammeter digit recognition method of a kind of robust as claimed in claim 1, it is characterised in that：Step 2 is used
The flow at the edge of Canny operator extractions industry ammeter image is：
Step1：Initial pictures are smoothed with Gaussian filter；
Step2：Amplitude and direction with the finite difference formulations gradient of single order local derviation；
Step3：Non-maxima suppression is carried out to gradient magnitude；
Step4：Detected with dual threashold value-based algorithm and connection edge；
The image I_dilate computing formula after morphological dilation is expanded are carried out to edge image I_canny is：Wherein strel (5) is the structural element of expansion.
4. the industrial ammeter digit recognition method of a kind of robust as claimed in claim 1, it is characterised in that：
Step 3 is specially：Based on the edge image I_dilate after expansion, it is slightly fixed that first the region containing numeral in image is carried out
Position, then carries out precise positioning cutting to numeric area by horizontal, longitudinal projection's method respectively, and transverse projection vector isLongitudinal projection's vector is
According to two projection vector v, given threshold thresh, calculate up and down four cut points (up, down, left,
right)：=index (v ＜ thresh), then carries out further fine cut and obtains I_fine to numeric area.
5. the industrial ammeter digit recognition method of a kind of robust as claimed in claim 1, it is characterised in that：
Numeric area is cut into the zonule of individual digit, identification individual digit lower right corner decimal point after having cut in step 5
It is specific as follows：
Step 5-1, by sciagraphy, only carry out longitudinal projection, numeric area is cut into the zonule of individual digit, cut point
Computing formula point:=index (v=0), obtains the image I_single of individual digit after having cut；
Step 5-2, for each individual digit picture for cutting out, continue to use longitudinal projection, obtain individual digit figure
The projection vector v_single of picture；
Step 5-3, given threshold th=row (I_single)/5, the cutting position of decimal point is point:=index (v_
Single ＜ th), wherein index () is the subscript computing formula of vector；Calculate the cutting of each digital lower right corner decimal point away from
From the numeral of ultimate range is the numeral containing decimal point, and computing formula is：
6. the industrial ammeter digit recognition method of a kind of robust as claimed in claim 1, it is characterised in that：Step 6 is based on KNN
Algorithm, recognizes to individual digit, and recognizes decimal point, specially：
Step 6-1, above-mentioned steps cut individual digit image by way of make training set, each numeral make N pictures,
Thus 10 total 10N of numeral open training set I_train；
Step 6-2, for the distance function in every test pictures I_test, KNN algorithm it is：
Step 6-3, for each test pictures I_test, we find and test pictures in 10N training set I_train
IOU maximum K, then in K training set, there are most numerals for being test pictures in certain digital classification；
Step 6-4, determine scaling position by the method for step 5, obtain final ammeter reading.
7. the industrial ammeter digit recognition method of a kind of robust as claimed in claim 3, it is characterised in that：In step 2
The threshold value of Canny edge detection algorithms is 0.6, and the value of the structural element strel (5) of expansion is
8. the industrial ammeter digit recognition method of a kind of robust as claimed in claim 4, it is characterised in that：Throwing in step 3
The threshold value thresh=15 of shadow vector cutting.
9. the industrial ammeter digit recognition method of a kind of robust as claimed in claim 6, it is characterised in that：Instruction in step 6
Practice collection number：The K values of N=10, KNN algorithm are K=10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611031147.2A CN106778754A (en) | 2016-11-22 | 2016-11-22 | A kind of industrial ammeter digit recognition method of robust |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611031147.2A CN106778754A (en) | 2016-11-22 | 2016-11-22 | A kind of industrial ammeter digit recognition method of robust |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106778754A true CN106778754A (en) | 2017-05-31 |
Family
ID=58970534
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611031147.2A Pending CN106778754A (en) | 2016-11-22 | 2016-11-22 | A kind of industrial ammeter digit recognition method of robust |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106778754A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109063614A (en) * | 2018-07-19 | 2018-12-21 | 江苏省计量科学研究院 | A kind of liquid crystal display digit recognition method |
| CN109344820A (en) * | 2018-08-06 | 2019-02-15 | 北京邮电大学 | Digital electric meter Recognition of Reading method based on computer vision and deep learning |
| CN109492448A (en) * | 2018-11-13 | 2019-03-19 | 国网河北省电力有限公司电力科学研究院 | Label coding consistency desired result method and device |
| CN110084241A (en) * | 2019-05-05 | 2019-08-02 | 山东大学 | A kind of ammeter automatic reading method based on image recognition |
| CN110232376A (en) * | 2019-06-11 | 2019-09-13 | 重庆邮电大学 | A kind of gear type digital instrument recognition methods returned using projection |
| CN110287967A (en) * | 2019-06-28 | 2019-09-27 | 哈尔滨工业大学 | A kind of number and stem-winder digit recognition method based on image |
| CN110807416A (en) * | 2019-10-31 | 2020-02-18 | 国网湖北省电力有限公司电力科学研究院 | Digital instrument intelligent recognition device and method suitable for mobile detection device |
| CN112770080A (en) * | 2019-11-01 | 2021-05-07 | 中移物联网有限公司 | Meter reading method, meter reading device and electronic equipment |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103955694A (en) * | 2014-04-09 | 2014-07-30 | 广州邦讯信息系统有限公司 | Image recognition meter reading system and method |
| CN106023173A (en) * | 2016-05-13 | 2016-10-12 | 浙江工业大学 | Number identification method based on SVM |
-
2016
- 2016-11-22 CN CN201611031147.2A patent/CN106778754A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103955694A (en) * | 2014-04-09 | 2014-07-30 | 广州邦讯信息系统有限公司 | Image recognition meter reading system and method |
| CN106023173A (en) * | 2016-05-13 | 2016-10-12 | 浙江工业大学 | Number identification method based on SVM |
Non-Patent Citations (2)
| Title |
|---|
| 徐剑峰: "电能表表号点状数字识别算法研究", 《中国优秀硕士学位论文全文数据库 工程科技II辑》 * |
| 陈晨: "智能交通系统中车牌识别的关键技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109063614A (en) * | 2018-07-19 | 2018-12-21 | 江苏省计量科学研究院 | A kind of liquid crystal display digit recognition method |
| CN109344820A (en) * | 2018-08-06 | 2019-02-15 | 北京邮电大学 | Digital electric meter Recognition of Reading method based on computer vision and deep learning |
| CN109344820B (en) * | 2018-08-06 | 2021-09-17 | 北京邮电大学 | Digital ammeter reading identification method based on computer vision and deep learning |
| CN109492448A (en) * | 2018-11-13 | 2019-03-19 | 国网河北省电力有限公司电力科学研究院 | Label coding consistency desired result method and device |
| CN110084241A (en) * | 2019-05-05 | 2019-08-02 | 山东大学 | A kind of ammeter automatic reading method based on image recognition |
| CN110084241B (en) * | 2019-05-05 | 2023-05-30 | 山东大学 | Automatic ammeter reading method based on image recognition |
| CN110232376A (en) * | 2019-06-11 | 2019-09-13 | 重庆邮电大学 | A kind of gear type digital instrument recognition methods returned using projection |
| CN110232376B (en) * | 2019-06-11 | 2021-04-20 | 重庆邮电大学 | Gear type digital instrument identification method by utilizing projection regression |
| CN110287967A (en) * | 2019-06-28 | 2019-09-27 | 哈尔滨工业大学 | A kind of number and stem-winder digit recognition method based on image |
| CN110807416A (en) * | 2019-10-31 | 2020-02-18 | 国网湖北省电力有限公司电力科学研究院 | Digital instrument intelligent recognition device and method suitable for mobile detection device |
| CN112770080A (en) * | 2019-11-01 | 2021-05-07 | 中移物联网有限公司 | Meter reading method, meter reading device and electronic equipment |
| CN112770080B (en) * | 2019-11-01 | 2023-01-03 | 中移物联网有限公司 | Meter reading method, meter reading device and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106778754A (en) | A kind of industrial ammeter digit recognition method of robust | |
| CN107808378B (en) | Method for detecting potential defects of complex-structure casting based on vertical longitudinal and transverse line profile features | |
| CN106557764B (en) | A kind of water level recognition methods based on binary-coded character water gauge and image procossing | |
| CN104751142B (en) | A kind of natural scene Method for text detection based on stroke feature | |
| CN102999886B (en) | Image Edge Detector and scale grating grid precision detection system | |
| CN106446952A (en) | Method and apparatus for recognizing score image | |
| CN110443128B (en) | Finger vein identification method based on SURF feature point accurate matching | |
| CN102779267B (en) | The method and apparatus of specific object region in detected image | |
| CN104680130A (en) | Chinese character recognition method for identification cards | |
| CN103942797B (en) | Scene image text detection method and system based on histogram and super-pixels | |
| CN108182383B (en) | Vehicle window detection method and device | |
| CN104156704A (en) | Novel license plate identification method and system | |
| CN106529532A (en) | License plate identification system based on integral feature channels and gray projection | |
| CN102693423A (en) | Method for precise positioning of license plate in strong light conditions | |
| CN110070090A (en) | A kind of logistic label information detecting method and system based on handwriting identification | |
| CN107016394B (en) | Cross fiber feature point matching method | |
| CN104680550A (en) | Method for detecting defect on surface of bearing by image feature points | |
| CN107480585A (en) | Object detection method based on DPM algorithms | |
| CN105447489A (en) | Character and background adhesion noise elimination method for image OCR system | |
| CN105184225A (en) | Multinational paper money image identification method and apparatus | |
| CN103500323B (en) | Based on the template matching method of self-adaptation gray level image filtering | |
| CN105404868A (en) | Interaction platform based method for rapidly detecting text in complex background | |
| CN104966047A (en) | Method and device for identifying vehicle license | |
| Pandya et al. | Morphology based approach to recognize number plates in India | |
| CN112734729A (en) | Water gauge water level line image detection method and device suitable for night light supplement condition and storage medium |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170531 |
|
| RJ01 | Rejection of invention patent application after publication |