CN104463130B - A kind of license plate image photo-irradiation treatment method based on assessment reponse system - Google Patents
A kind of license plate image photo-irradiation treatment method based on assessment reponse system Download PDFInfo
- Publication number
- CN104463130B CN104463130B CN201410788228.1A CN201410788228A CN104463130B CN 104463130 B CN104463130 B CN 104463130B CN 201410788228 A CN201410788228 A CN 201410788228A CN 104463130 B CN104463130 B CN 104463130B
- Authority
- CN
- China
- Prior art keywords
- image
- license plate
- photo
- irradiation treatment
- assessment
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
-
- 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
- G06V10/30—Noise filtering
-
- 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
Abstract
The invention discloses a kind of license plate image photo-irradiation treatment methods based on assessment reponse system, include the following steps：(1) it is HSV models by RGB model conversions by the license plate image of input；(2) a series of photo-irradiation treatments are carried out to V component；(3) local auto-adaptive threshold process is carried out to enhanced license plate image and obtains car plate bianry image；(4) characters on license plate situation in car plate bianry image is assessed；(5) judge whether to meet the condition for stopping feedback；(6) judge whether to need the selection of eye-observation, if desired then to car plate bianry image skeletonizing, then carry out region-growing method on car plate enhancing image as seed point, and to growth district into the promotion of row pixel value；Recognition of License Plate Characters is directly carried out if not needing to and exports computer recognition result.The problem of being difficult under complex illumination environment by human eye and computer identification present invention efficiently solves the number-plate number, so as to improve the discrimination of the number-plate number.
Description
Technical field
The present invention relates to the research field of Car license recognition, more particularly to a kind of license plate image light based on assessment reponse system
According to processing method.
Background technology
The number-plate number identifies the important component indispensable as intelligent transportation system, has been developing progressively as vehicle
The main means of identification are simultaneously widely used in monitoring, alarming, punishment on contravention of regulation, room entry/exit management and expressway tol lcollection management
The fields of grade.In past 10 years, although license plate recognition technology has been achieved with remarkable progress, but in some practical applications, light
Then discrimination is influenced to some extent according to, many factors such as angle, resolution ratio and car plate type, and wherein illumination is more typical one
Kind disturbing factor.Photo-irradiation treatment method is broadly divided into three classes at present：The method of method, feature based based on study and based on returning
One method changed.It is trained based on the method for study by collecting the sample under different illumination scenes and being directly used in comprising sample
The world model of middle difference illumination situation, such as linear subspaces and Flow model, the unknown light being then generalized to again in new picture
According to situation；As for feature based method by extracting illumination invariant feature in image, these features include geometric properties and line
Manage feature；It is finally based on normalized method, this method handles illumination component in removal image by " normalization ", so as to subtract
Influence of the low illumination to identification, wherein in recent years, the LTV model photo-irradiation treatment methods based on frequency domain cause larger concern.It but should
Method is susceptible to illumination invariant when illumination variation is violent and divides the problem of inaccurate and parameter optimization is excessively random, together
When algorithm complexity is higher leads to operation time.In conclusion the shortcomings that existing photo-irradiation treatment method, is mainly reflected in following two
Point：1st, the violent scene stability inferior of illumination variation is poor；2nd, the selection of parameter or feature needs experience, not smart enough.
Invention content
The shortcomings that it is a primary object of the present invention to overcome the prior art and deficiency provide a kind of based on assessment reponse system
License plate image photo-irradiation treatment method, this method efficiently solve the number-plate number be difficult under complex illumination environment by human eye and
The problem of computer identifies, so as to improve the discrimination of the number-plate number.
In order to reach above-mentioned purpose, the present invention uses following technical scheme：
A kind of license plate image photo-irradiation treatment method based on assessment reponse system, includes the following steps：
(1) it is HSV models by RGB model conversions by the license plate image of input；
(2) a series of photo-irradiation treatments are carried out to V component, corrects enhancing brightness of image with Gamma first, then carry out Gaussian difference
Divide the high fdrequency component in filtering removal illumination, the region being further processed then is needed according to the extraction of mask function, is finally used
The contrast of contrast equalization enhancing image irradiation, so as to reach uneven in removal illumination while utmostly retain image
Useful details simultaneously enhances picture contrast；
(3) local auto-adaptive threshold process is carried out to enhanced license plate image and obtains car plate bianry image；
(4) characters on license plate situation in car plate bianry image is assessed, that is, carries out the connected component labeling based on eight neighborhood and calculating
Each spacing parameter between the area of connected domain, position, length-width ratio and domain；
(5) judge whether to stop feedback, i.e., it is first determined whether meeting characters on license plate standard, field parameter and warp will be connected
It tests threshold value to compare, the bianry image is exported if meeting and exits assessment reponse system, if being unsatisfactory for comparing account of the history
And cache best result and corresponding bianry image；Then judge whether the number for the cycle upper limit for reaching setting, it is defeated if reaching
Go out in caching and bianry image and exit assessment reponse system, the new photo-irradiation treatment parameter if not up to after output adjustment and again
Carry out photo-irradiation treatment；
(6) judge whether to need the selection of eye-observation, if desired then increase car plate bianry image in step (4) and car plate
Strong image carries out image co-registration output vision enhancement as a result, i.e. to car plate bianry image skeletonizing, then as seed point in vehicle
Region-growing method is carried out on board enhancing image, and to growth district into the promotion of row pixel value；Directly into driving if not needing to
Board character recognition simultaneously exports computer recognition result.
Preferably, in step (2), the calculation formula of Gamma corrections is：
G (x, y)=c [f (x, y)]γ
F (x, y) is input picture in formula, and g (x, y) is output image, and c is used to adjust picture contrast, γ for arbitrary value
Between 0~1, for adjusting brightness of image.
Preferably, in step (2), the calculation formula of difference of Gaussian filtering is：
F (x, y) is input picture in formula, and g (x, y) is output image, and μ is mean value, σn(n=1,2) it is variance.
Preferably, in step (2), the calculation formula of mask is：
F (x, y) is input picture in formula, and g (x, y) is output image, and mask (x, y) is mask image.
Preferably, in step (2), the calculation formula of contrast equalization is：
F (x, y) is input picture in formula, gn(x, y) (n=1,2,3) is output image, and α is that compression sex index is used to drop
The influence of low big pixel value, τ is for big pixel value in amputation image after first step normalization.
Preferably, step (3) is specially：
(3-1) divides the image into the big zonules, wherein n such as m*n>m>0；
OTSU processing is carried out in (3-2) each zonule, the two-dimensional matrix that all threshold values form is denoted as Mat_T0；
(3-3) is to all Mat_T0Carry out the gaussian filtering output Mat_T of 3*31, then according to Mat_T1Carry out each area
The image binaryzation in domain.
Preferably, in step (3-2), the specific method for carrying out OTSU processing is：
(3-2-1) sets an initial threshold T, and the wherein calculation formula of T is as follows：
T=wp1+(1-w)p2
In formula, p1And p2For pixel value arbitrary in image；W is weighted value；
(3-2-2) divides image using T, generates two groups of phase pixels：All pixels composition G of the brightness value more than or equal to T1,
All pixels composition G of the brightness value less than T2；
(3-2-3) calculates G1And G2In the range of pixel average brightness value μ1And μ2；
(3-2-4) calculates new T, is shown below：
T=w μ1+(1-w)μ2
(3-2-5) repeats step (3-2-2) to (3-2-4), until the variation of T is less than a predetermined value T0Until.
Preferably, in step (5), if output effect and input parameter form unimodal letter, using following processing methods：
First five parameters of the algorithm are analyzed, omit minor parameter, leave two parameters of DOG filtering --- it is high
The meansquaredeviationσ of this filtering1/σ2, then fixed σ1/σ2, to σ1/σ2It carries out, based on the preferred of Fibonacci method, then fixing σ1/σ2Instead
Come over to be σ1/σ2It is preferred, so cycle until σ1/σ2Until variation is less than a certain range.
Preferably, in step (5), if output effect and input parameter form non-unimodal letter, using following processing methods：
First by two parameters of the algorithm --- often row grid number and each column grid number r1/r2Value range be limited in [3,
9] between, then by value range it is coarse turn to { 3,6,9 } and according to car plate the ratio of width to height provide r1/r2>r1/r2, then calculate at this time
Optimal solution r1n/r2n, then in r1n/r2nOptimal solution is calculated in the range of ± 2.
Preferably, in step (5), the number for recycling the upper limit is 10 times.
Compared with prior art, the present invention having the following advantages that and advantageous effect：
1st, photo-irradiation treatment method proposed by the present invention employs a series of simple light irradiation preprocess methods, such as Gamma schools
Just, gaussian filtering calculus of differences and contrast equalization etc., remove license plate image in reflected light ingredient, so as to retain incident light into
Get enhancing image, it is substantially suitable compared with the LTV model photo-irradiation treatment method effects based on frequency domain, but improved not in efficiency
It is few.
2nd, the present invention replaces Global thresholding using local thresholding method, and improve that character in the case of uneven illumination extracts has
Effect property.
3rd, the present invention finally innovatively adds assessment reponse system, according to assessing local threshold treated binary map
The character extraction situation of picture adjusts the parameter of photo-irradiation treatment algorithm, and photo-irradiation treatment is made to play more intelligence and stablize.
4th, the present invention provides computer character identifications and human eye vision to enhance two kinds of outputs as a result, to expand this method
Actually use range.
Description of the drawings
Fig. 1 is that the present invention is based on the general flow charts of the license plate image photo-irradiation treatment method of assessment reponse system；
Fig. 2 is the flow chart of present invention assessment reponse system.
Specific embodiment
With reference to embodiment and attached drawing, the present invention is described in further detail, but embodiments of the present invention are unlimited
In this.
Embodiment
As shown in Figure 1 and Figure 2, the license plate image photo-irradiation treatment method based on assessment reponse system, specifically includes following steps
Suddenly：
(1) input license plate image is switched into HSV models by RGB models；
(2) a series of light irradiation preprocess methods are carried out to V component to deal with complicated light environment, hereinafter referred to as TT side
Method.This method corrects enhancing brightness of image with Gamma first, then carries out difference of Gaussian filtering (DOG, Differnce of
Gaussian illumination high frequency components) are removed, i.e., then uneven part needs what is be further processed according to the extraction of mask function
Region finally equalizes the contrast of enhancing image irradiation using contrast, uneven while most in removal illumination so as to reach
Big degree retains the useful details of image and enhances picture contrast；
1. Gamma updating formulas：
G (x, y)=c [f (x, y)]γ (1)
F (x, y) is input picture in formula, and g (x, y) is output image, and c is used to adjust picture contrast, γ for arbitrary value
It is used to adjust brightness of image between 0~1；
2. Dog is filtered：
F (x, y) is input picture in formula, and g (x, y) is output image, and μ is mean value, σn(n=1,2) it is variance；
3. mask masks：
F (x, y) is input picture in formula, and g (x, y) is output image, and mask (x, y) is mask image；
4. contrast equalizes：
F (x, y) is input picture in formula, gn(x, y) (n=1,2,3) is output image, and α is that compression sex index is used to drop
The influence of low big pixel value, τ is for big pixel value in amputation image after first step normalization；
(3) local auto-adaptive threshold process is carried out to enhanced license plate image and obtains car plate bianry image, specific practice
It is as follows：
1. divide the image into m*n (n>m>0) the big zonule of a grade；
2. carrying out OTSU processing in each zonule, the two-dimensional matrix that all threshold values form is denoted as Mat_T0；
3. to all Mat_T0Carry out the gaussian filtering output Mat_T of 3*31, then according to Mat_T1Carry out each region
Image binaryzation；
Wherein, OSTU processing operand is larger, proposes a kind of simplified algorithm here：
1. an initial threshold T is set first.The wherein calculation formula of T is as follows：
T=wp1+(1-w)p2 (7)
In formula, p1And p2For pixel value (general p arbitrary in image1Take maximum brightness value, p2Take minimum luminance value)；W is power
Weight values (generally take 1/2).
2. divide image using T.This can generate two groups of phase pixels：All pixels composition G of the brightness value more than or equal to T1, it is bright
All pixels composition G of the angle value less than T2。
3. calculate the average brightness value μ of the pixel in the range of G1 and G21And μ2。
4. calculating new T, it is shown below：
T=w μ1+(1-w)μ2 (8)
5. 2. 4. repeating step arrives step, until the variation of T is less than a predetermined value T0Until.
(4) with reference to attached drawing 2, the connected component labeling based on eight neighborhood is carried out to car plate bianry image and calculates each connected domain
Area, position, the parameters such as spacing between length-width ratio and domain；
(5) judge whether to stop feedback, judge whether to meet each of characters on license plate evaluation criterion first by these parameters
Item index, will connect field parameter compared with empirical value, and the bianry image is exported if meeting and exits assessment feedback system
System compares account of the history if being unsatisfactory for and caches best result and corresponding bianry image；Then judge whether to reach in cycle
Limit 10 times exports bianry image in caching if reaching and exits assessment reponse system, if not up to new after output adjustment
Photo-irradiation treatment parameter simultaneously re-starts photo-irradiation treatment；Wherein during parameter adjustment, since parameter is more time-consuming longer, therefore
Following two accelerating algorithms have been selected herein according to whether output effect and input parameter form unimodal function：
1. optimum seeking method (unimodal function)
Optimum seeking method is according to practical application needs, and using mathematical principle, reasonable arrangement experiment is rapid with minimum test number (TN)
Find the scientific experimentation method of Best Point.This method on condition that object function be unimodal function, be applicable in TT illumination in this article
Processing Algorithm.It comprises the concrete steps that：First five parameters of the algorithm are analyzed, omit minor parameter, leave the two of DOG filtering
The meansquaredeviationσ of a parameter --- gaussian filtering1/σ2, then fixed σ1/σ2, to σ1/σ2It carries out based on the preferred of Fibonacci method, so
σ is fixed afterwards1/σ2σ is in turn1/σ2It is preferred, so cycle until σ1/σ2Until variation is less than a certain range.
2. multi-grid method (non-unimodal function)
Multi-grid method is originally differential equation fast solution method, it is extended on the problem of parameter is preferred here,
Its basic ideas is that problem is roughened, being calculated on former Grid Projection to a fairly simple new grid, until quickly receiving
After holding back former grid is returned to via interpolation arithmetic again.This method does not need to object function as unimodal function, is applicable in office in this article
Portion's OTSU Thresholding Algorithms.It comprises the concrete steps that：First by two parameters of the algorithm --- often row grid number and each column grid number
r1/r2Value range be limited between [3,9], then by value range it is coarse turn to { 3,6,9 } and according to car plate the ratio of width to height provide
r1/r2>r1/r2, then calculate optimal solution r at this time1n/r2n, then in r1n/r2nOptimal solution is calculated in the range of ± 2.
(6) judge whether to need the selection of eye-observation, if desired then increase car plate bianry image in step (4) and car plate
Strong image carries out image co-registration output vision enhancement as a result, i.e. to car plate bianry image skeletonizing, then as seed point in vehicle
Region-growing method is carried out on board enhancing image, and to growth district into the promotion of row pixel value；Directly into driving if not needing to
Board character recognition simultaneously exports computer recognition result.
Above-described embodiment is the preferable embodiment of the present invention, but embodiments of the present invention are not by above-described embodiment
Limitation, other any Spirit Essences without departing from the present invention with made under principle change, modification, replacement, combine, simplification,
Equivalent substitute mode is should be, is included within protection scope of the present invention.
Claims (8)
- A kind of 1. license plate image photo-irradiation treatment method based on assessment reponse system, which is characterized in that include the following steps：(1) it is HSV models by RGB model conversions by the license plate image of input；(2) a series of photo-irradiation treatments are carried out to V component, corrects enhancing brightness of image with Gamma first, then carry out difference of Gaussian filter High fdrequency component in wave removal illumination, then needs the region being further processed, finally using comparison according to the extraction of mask function The contrast of degree equalization enhancing image irradiation, so as to reach uneven in removal illumination while utmostly retain image useful Details simultaneously enhances picture contrast；(3) local auto-adaptive threshold process is carried out to enhanced license plate image and obtains car plate bianry image, specially：(3-1) divides the image into the big zonules, wherein n such as m*n>m>0；OTSU processing is carried out in (3-2) each zonule, the two-dimensional matrix that all threshold values form is denoted as Mat_T0；Carry out OTSU The specific method of processing is：(3-2-1) sets an initial threshold T, and the wherein calculation formula of T is as follows：T=wp1+(1-w)p2In formula, p1And p2For pixel value arbitrary in image；W is weighted value；(3-2-2) divides image using T, generates two groups of phase pixels：All pixels composition G of the brightness value more than or equal to T1, brightness value All pixels composition G less than T2；(3-2-3) calculates G1And G2In the range of pixel average brightness value μ1And μ2；(3-2-4) calculates new T, is shown below：T=w μ1+(1-w)μ2(3-2-5) repeats step (3-2-2) to (3-2-4), until the variation of T is less than a predetermined value T0Until；(3-3) is to all Mat_T0Carry out the gaussian filtering output Mat_T of 3*31, then according to Mat_T1Carry out each region Image binaryzation(4) characters on license plate situation in car plate bianry image is assessed, that is, carry out the connected component labeling based on eight neighborhood and is calculated each Spacing parameter between the area of connected domain, position, length-width ratio and domain；(5) judge whether to stop feedback, i.e., it is first determined whether meeting characters on license plate standard, field parameter and experience threshold will be connected Value compares, and exports the bianry image if meeting and exits assessment reponse system, if being unsatisfactory for comparing account of the history and delay Deposit best result and corresponding bianry image；Then judge whether the number for the cycle upper limit for reaching setting, exported if reaching slow It deposits middle bianry image and exits assessment reponse system, the new photo-irradiation treatment parameter if not up to after output adjustment simultaneously re-starts Photo-irradiation treatment；(6) judge whether to need the selection of eye-observation, if desired then scheme car plate bianry image in step (4) and car plate enhancing As carrying out image co-registration output vision enhancement as a result, i.e. to car plate bianry image skeletonizing, then increase as seed point in car plate Carry out region-growing method on strong image, and to growth district into the promotion of row pixel value；Car plate word is directly carried out if not needing to Symbol identifies and exports computer recognition result.
- 2. the license plate image photo-irradiation treatment method according to claim 1 based on assessment reponse system, which is characterized in that step Suddenly in (2), the calculation formula of Gamma corrections is：G (x, y)=c [f (x, y)] γF (x, y) is input picture in formula, and g (x, y) is output image, c for arbitrary value for adjusting picture contrast, γ 0~ Between 1, for adjusting brightness of image.
- 3. the license plate image photo-irradiation treatment method according to claim 1 based on assessment reponse system, which is characterized in that step Suddenly in (2), the calculation formula of difference of Gaussian filtering is：F (x, y) is input picture in formula, and g (x, y) is output image, and μ is mean value, σn(n=1,2) it is variance.
- 4. the license plate image photo-irradiation treatment method according to claim 1 based on assessment reponse system, which is characterized in that step Suddenly in (2), the calculation formula of mask is：F (x, y) is input picture in formula, and g (x, y) is output image, and mask (x, y) is mask image.
- 5. the license plate image photo-irradiation treatment method according to claim 1 based on assessment reponse system, which is characterized in that step Suddenly in (2), the calculation formula of contrast equalization is：F (x, y) is input picture in formula, gn(x, y) (n=1,2,3) is output image, and α is compression sex index for reducing big picture The influence of element value, τ is for big pixel value in amputation image after first step normalization.
- 6. the license plate image photo-irradiation treatment method according to claim 1 based on assessment reponse system, which is characterized in that step Suddenly in (5), if output effect and input parameter form unimodal letter, using following processing methods：First five parameters of the license plate image photo-irradiation treatment method are analyzed, omit minor parameter, leave the two of DOG filtering The meansquaredeviationσ of a parameter --- gaussian filtering1/σ2, then fixed σ1/σ2, to σ1/σ2It carries out based on the preferred of Fibonacci method, so σ is fixed afterwards1/σ2σ is in turn1/σ2It is preferred, so cycle until σ1/σ2Until variation is less than a certain range.
- 7. the license plate image photo-irradiation treatment method according to claim 1 based on assessment reponse system, which is characterized in that step Suddenly in (5), if output effect and input parameter form non-unimodal letter, using following processing methods：First by two parameters of the license plate image photo-irradiation treatment method --- often row grid number and each column grid number r1/r2Value Scope limitation between [3,9], then by value range it is coarse turn to { 3,6,9 } and according to car plate the ratio of width to height provide r1/r2>r1/ r2, then calculate optimal solution r at this time1n/r2n, then in r1n/r2nOptimal solution is calculated in the range of ± 2.
- 8. the license plate image photo-irradiation treatment method according to claim 1 based on assessment reponse system, which is characterized in that step Suddenly in (5), the number for recycling the upper limit is 10 times.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410788228.1A CN104463130B (en) | 2014-12-16 | 2014-12-16 | A kind of license plate image photo-irradiation treatment method based on assessment reponse system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410788228.1A CN104463130B (en) | 2014-12-16 | 2014-12-16 | A kind of license plate image photo-irradiation treatment method based on assessment reponse system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104463130A CN104463130A (en) | 2015-03-25 |
| CN104463130B true CN104463130B (en) | 2018-06-12 |
Family
ID=52909146
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410788228.1A Active CN104463130B (en) | 2014-12-16 | 2014-12-16 | A kind of license plate image photo-irradiation treatment method based on assessment reponse system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104463130B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106651872B (en) * | 2016-11-23 | 2020-09-15 | 北京理工大学 | Pavement crack identification method and system based on Prewitt operator |
| CN108629738B (en) * | 2017-03-16 | 2022-04-01 | 斑马智行网络(香港)有限公司 | Image processing method and device |
| CN109658706A (en) * | 2018-12-18 | 2019-04-19 | 深圳市捷顺科技实业股份有限公司 | A kind of vehicle data processing method and relevant device |
| CN111091544B (en) * | 2019-12-12 | 2020-10-16 | 哈尔滨市科佳通用机电股份有限公司 | Method for detecting breakage fault of side integrated framework of railway wagon bogie |
| CN111627025A (en) * | 2020-05-21 | 2020-09-04 | 江苏华阳智能装备股份有限公司 | Bottled liquid medicine intelligent detection method and system |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7016518B2 (en) * | 2002-03-15 | 2006-03-21 | Extreme Cctv Inc. | Vehicle license plate imaging and reading system for day and night |
| CN101373517B (en) * | 2007-08-22 | 2011-03-16 | 北京万集科技有限责任公司 | Method and system for recognizing license plate |
-
2014
- 2014-12-16 CN CN201410788228.1A patent/CN104463130B/en active Active
Non-Patent Citations (2)
| Title |
|---|
| HSV变换和同态滤波的彩色图像光照补偿;焦竹青等;《计算机工程与应用》;20101231;第46卷(第30期);第142-144页 * |
| 视觉检测系统的反馈机制研究;罗三定等;《计算机工程》;20100131;第36卷(第1期);第197-200页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104463130A (en) | 2015-03-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104463130B (en) | A kind of license plate image photo-irradiation treatment method based on assessment reponse system | |
| Raju et al. | A fast and efficient color image enhancement method based on fuzzy-logic and histogram | |
| CN103996018B (en) | Face identification method based on 4DLBP | |
| CN105354865A (en) | Automatic cloud detection method and system for multi-spectral remote sensing satellite image | |
| CN104102920A (en) | Pest image classification method and pest image classification system based on morphological multi-feature fusion | |
| CN103942577A (en) | Identity identification method based on self-established sample library and composite characters in video monitoring | |
| CN111027493A (en) | Pedestrian detection method based on deep learning multi-network soft fusion | |
| CN108960189A (en) | Image recognition methods, device and electronic equipment again | |
| Hassanzadeh et al. | Non-linear grayscale image enhancement based on firefly algorithm | |
| CN104636118A (en) | QR two-dimensional code self-adaptation binarization processing method and device based on light balance | |
| CN107315998B (en) | Vehicle class division method and system based on lane line | |
| Bak et al. | Two-stream convolutional networks for dynamic saliency prediction | |
| Nalla et al. | Image dehazing for object recognition using faster RCNN | |
| CN108108703A (en) | Deceleration strip missing detection method, device and electronic equipment | |
| CN107274425B (en) | A kind of color image segmentation method and device based on Pulse Coupled Neural Network | |
| CN103700118B (en) | Based on the moving target detection method of pulse coupled neural network | |
| CN105354547A (en) | Pedestrian detection method in combination of texture and color features | |
| Yimit et al. | Automatic image enhancement by artificial bee colony algorithm | |
| CN113902965A (en) | Multi-spectral pedestrian detection method based on multi-layer feature fusion | |
| CN110570450B (en) | Target tracking method based on cascade context-aware framework | |
| CN103093241B (en) | Based on the remote sensing image nonuniformity cloud layer method of discrimination of homogeneity process | |
| Wang et al. | Traffic thermal infrared texture generation based on siamese semantic CycleGAN | |
| CN109635764A (en) | A kind of Human bodys' response method and system based on multiple features linear temporal coding | |
| Singh et al. | Multiscale reflection component based weakly illuminated nighttime image enhancement | |
| Oh et al. | Visual adversarial attacks and defenses |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Zhang Yongdong Inventor after: Xu Yuesheng Inventor after: Tan Li Inventor before: Xu Yuesheng Inventor before: Zhang Yongdong Inventor before: Tan Li |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |