CN105787894A - Barrel distortion container number correction method - Google Patents
Barrel distortion container number correction method Download PDFInfo
- Publication number
- CN105787894A CN105787894A CN201610105083.XA CN201610105083A CN105787894A CN 105787894 A CN105787894 A CN 105787894A CN 201610105083 A CN201610105083 A CN 201610105083A CN 105787894 A CN105787894 A CN 105787894A
- Authority
- CN
- China
- Prior art keywords
- image
- container number
- distortion
- coordinate
- barrel distortion
- 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
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000012937 correction Methods 0.000 title claims abstract description 22
- 238000013507 mapping Methods 0.000 claims abstract description 24
- 230000009466 transformation Effects 0.000 claims abstract description 9
- HUTDUHSNJYTCAR-UHFFFAOYSA-N ancymidol Chemical compound C1=CC(OC)=CC=C1C(O)(C=1C=NC=NC=1)C1CC1 HUTDUHSNJYTCAR-UHFFFAOYSA-N 0.000 claims description 8
- 238000002372 labelling Methods 0.000 claims description 8
- 238000001914 filtration Methods 0.000 claims description 7
- 238000001514 detection method Methods 0.000 claims description 5
- 230000015556 catabolic process Effects 0.000 claims description 4
- 230000008030 elimination Effects 0.000 claims description 4
- 238000003379 elimination reaction Methods 0.000 claims description 4
- 238000013519 translation Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 abstract description 3
- 230000008569 process Effects 0.000 description 5
- 230000011218 segmentation Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/80—Geometric correction
-
- 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
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The present invention relates to a barrel distortion container number correction method. The method comprises: the barrel distortion container number image is converted to a distortion-free container number image; the original image of the barrel distortion container number is decomposed to a R graph, a G graph and a B graph; the distortion coefficient of a camera is obtained through a calibrated camera, the space coordinate mapping from the original image to the corrected image is completed according to the distortion mapping relation, and obtaining the R graph, the G graph and the B graph of the distortion-free container number through adoption of a nearest neighbor algorithm; combining the R graph, the G graph and the B into a distortion-free RGB container number image; performing preprocessing of the distortion-free RGB container number image to obtain a container number binary image; and finally, performing inclination correction of the distortion-free container number binary image through adoption of an improved container number inclination correction algorithm based on a Hough transformation. The barrel distortion container number correction method is simple, easy to operate and convenient with no need for making various high-precision templates and complex spherical surface models.
Description
Technical field
The present invention relates to a kind of image processing techniques, be specifically related to the container number bearing calibration of a kind of barrel distortion, it is adaptable to the container number correction of barrel distortion in Intelligence Recognition of Container Code by Means.
Background technology
Container number is the unique mark identified in the container whole world, is equivalent to the ID (identity number) card No. of container.By the identification to container number, understanding all information of container, greatly facilitate the management of Container Transport, it is very necessary for therefore realizing container number Intelligent Recognition.Intelligence Recognition of Container Code by Means generally comprises several major parts such as image acquisition, case number (CN) Image semantic classification, case number (CN) zone location, case number (CN) Character segmentation, case number (CN) character recognition, and various piece is closely coupled.The barrel aberrance emendation of container number facilitates container number zone location, Character segmentation and raising character identification rate.Container number image owing to being obtained by wide-angle lens contains serious geometric distortion, severely impact container number Intelligent Recognition effect, so necessary geometric distortion correction must be carried out to eliminate the barrel distortion that wide-angle lens brings, thus obtaining a distortionless container number image.
The bearing calibration of radial distortion model is had than more typical geometric distortion correction method, bearing calibration and the bearing calibration (Liu Jiwei adopting lattice-shaped calibration template based on hexagonal lattice pattern, monitor video image figure adjustment technical research, the National University of Defense Technology, 2008, 21 27), but these methods are required for making high-precision template (Ren Xiaokui, Jin Lin, pay Wen Bin, camera lens distortion correction based on IACPSO algorithm, computer utility is studied, 2015, 32 (6): 1,865 1868), the population of this algorithm needs a fixing inertia weight, and this inertia weight value can only have good effect under certain condition.
Therefore, container number bearing calibration that is a kind of simple and that can effectively eliminate barrel distortion is wished in this area, to correct the container number image of barrel distortion so that facilitating container number zone location, Character segmentation and raising character identification rate.
Summary of the invention
The present invention is directed to the container number image rectification problem of barrel distortion, it is provided that the container number bearing calibration of a kind of barrel distortion.The barrel distortion container number image rectification of shooting is had good result by the method, and has good practical value, it is simple to the case number (CN) zone location of Intelligence Recognition of Container Code by Means, Character segmentation and raising character identification rate.
In order to achieve the above object, the present invention adopts the following technical scheme that:
A kind of container number bearing calibration of barrel distortion, the step of the method is as follows:
Step one, the container number R that RGB container number picture breakdown is barrel distortion of barrel distortion is schemed, G figure and B figure；
Step 2, respectively the container number R of barrel distortion is schemed, G figure and B figure be adjusted in the horizontal direction and the vertical direction distortionless container number R figure, G figure and B figure；Specifically comprise the following steps that
Step 21, barrel distortion coefficient acquisition: barrel distortion is expressed as δ (r)=k1r2+k2r4+ ..., wherein r is the radius from center of distortion to distortion point, k1、k2For unknown coefficient, obtain k by MATLAB image calibration workbox1=-0.00000085, k2=-0.00000225, and ignore k2After other coefficients；Quadratic term coefficient k during due to generation barrel distortion1、k2For negative, its timing should be carried out contrary operation；
Step 22, forward direction coordinate map: first output image range is carried out suitable expanded range, by adjusting so that be enough to show the image that view picture is corrected；Then by the coordinate of input picture being mapped in output image one by one, input picture rounded coordinate is likely to be mapped in the middle of four coordinates of output image, and the gray value of described point coordinates is allocated between four coordinate points at it described；Taking 2.5 times that initial range is former fault image size of correction chart picture, then take the geometric center of mapped image, namely the half of length and width is as the picture centre point coordinates after mapping；If the image after mapping is beyond scope set in advance, just whole image is made to be positioned at the centre of whole picture by adjustment coordinate range and the coordinate translation by first pixel；Then the coordinate figure after the mapping updating formula by distorting obtained is mapped to the corresponding coordinate place of correction chart picture one by one, completes the mapping one by one of the two；
Step 23, arest neighbors interpolation: take around interpolation point the gray value of 4 nearest points of the neighbor pixel mid-range objectives point gray value as this point to make up the pinhole produced after coordinate maps；
Step 3, by distortionless container number R scheme, G figure and B figure merge into distortionless RGB container number figure；
Step 4, distortionless RGB container number Image semantic classification is obtained distortionless container number bianry image；Specifically comprise the following steps that
Step 41, image is carried out greyscale transformation；
Step 42, image is carried out Wiener filtering；
Step 43, filtered image is carried out background elimination: select the square structure element being sized to 6*6 that image is carried out opening operation and obtain Background；Then the image after background eliminates is obtained with original image subtracting background figure；
Step 44, Otsu method is used to carry out binaryzation to eliminating the image after background；
Step 5, distortionless container number bianry image is corrected to nonangular container number bianry image；Specifically include following step by step:
Step 51, first container number bianry image is carried out connected component labeling, mark each character connected region of case number (CN)；
Step 52, all character connected regions of case number (CN) according to step 51 labelling, add up and record the right margin pixel of these case number (CN) character zones；Then adopting method of least square that described right margin pixel is carried out fitting a straight line, namely the straight line finally simulated is container number right margin straight line.
Step 53, case number (CN) image is carried out Hough detection, in the two-dimensional array accumulator in parameter space, find out maximum and secondary maximum peak value, simulate boundary straight line；Then the angle that the boundary straight line simulated described in basis and the angle theta of vertical direction tilt as case number (CN), and carry out described case number (CN) image reversely rotating θ.
Accompanying drawing explanation
The present invention is further illustrated below in conjunction with the drawings and specific embodiments.
Fig. 1 is the container number correcting process figure of barrel distortion.
Detailed description of the invention
For the technological means making the present invention realize, creation characteristic, reach purpose and effect and be easy to understand, below in conjunction with being specifically illustrating, the present invention is expanded on further.
The present invention is directed to the container number image rectification problem of barrel distortion, the container number bearing calibration of a kind of barrel distortion is provided, in the process of container number image acquisition, due to reasons such as camera lens, sometimes the container number image gathered can occur to distort in various degree, the case number (CN) zone location of Intelligence Recognition of Container Code by Means, Character segmentation and raising character identification rate are caused and severely impact, it is therefore desirable to the container number of barrel distortion is corrected, the container number correcting process of barrel distortion is as shown in Figure 1.The step of the method is as follows:
Step one, the container number R that RGB container number picture breakdown is barrel distortion of barrel distortion is schemed, G figure and B figure；
Step 2, respectively the container number R of barrel distortion is schemed, G figure and B figure be adjusted in the horizontal direction and the vertical direction distortionless container number R figure, G figure and B figure；Specifically comprise the following steps that
Step 21, barrel distortion coefficient acquisition: barrel distortion is expressed as δ (r)=k1r2+k2r4+ ..., wherein r is the radius from center of distortion to distortion point, k1、k2For unknown coefficient, obtain k by MATLAB image calibration workbox1=-0.00000085, k2=-0.00000225, and ignore k2After other coefficients；Quadratic term coefficient k during due to generation barrel distortion1、k2For negative, its timing should be carried out contrary operation；
Step 22, forward direction coordinate map: first output image range is carried out suitable expanded range, by adjusting so that be enough to show the image that view picture is corrected；Then by the coordinate of input picture being mapped in output image one by one, input picture rounded coordinate is likely to be mapped in the middle of four coordinates of output image, and the gray value of described point coordinates is allocated between four coordinate points at it described；Taking 2.5 times that initial range is former fault image size of correction chart picture, then take the geometric center of mapped image, namely the half of length and width is as the picture centre point coordinates after mapping；If the image after mapping is beyond scope set in advance, just whole image is made to be positioned at the centre of whole picture by adjustment coordinate range and the coordinate translation by first pixel；Then the coordinate figure after the mapping updating formula by distorting obtained is mapped to the corresponding coordinate place of correction chart picture one by one, completes the mapping one by one of the two；
Step 23, arest neighbors interpolation: take around interpolation point the gray value of 4 nearest points of the neighbor pixel mid-range objectives point gray value as this point to make up the pinhole produced after coordinate maps；
Step 3, by distortionless container number R scheme, G figure and B figure merge into distortionless RGB container number figure；
Step 4, distortionless RGB container number Image semantic classification is obtained distortionless container number bianry image；Specifically comprise the following steps that
Step 41, image is carried out greyscale transformation；
Step 42, image is carried out Wiener filtering；
Step 43, filtered image is carried out background elimination: select the square structure element being sized to 6*6 that image is carried out opening operation and obtain Background；Then the image after background eliminates is obtained with original image subtracting background figure；
Step 44, Otsu method is used to carry out binaryzation to eliminating the image after background；
Step 5, distortionless container number bianry image is corrected to nonangular container number bianry image；Specifically include following step by step:
Step 51, first container number bianry image is carried out connected component labeling, mark each character connected region of case number (CN)；
Step 52, all character connected regions of case number (CN) according to step 51 labelling, add up and record the right margin pixel of these case number (CN) character zones；Then adopting method of least square that described right margin pixel is carried out fitting a straight line, namely the straight line finally simulated is container number right margin straight line.
Step 53, case number (CN) image is carried out Hough detection, in the two-dimensional array accumulator in parameter space, find out maximum and secondary maximum peak value, simulate boundary straight line；Then the angle that the boundary straight line simulated described in basis and the angle theta of vertical direction tilt as case number (CN), and carry out described case number (CN) image reversely rotating θ.
Based on above-mentioned principle, specific embodiment of the invention process is as follows:
Step one, the container number R that RGB container number picture breakdown is barrel distortion of barrel distortion is schemed, G figure and B figure, RGB image contains much information, not easily operates.
Step 2, respectively the container number R of barrel distortion is schemed, G figure and B figure be adjusted in the horizontal direction and the vertical direction.Schemed by the container number R of barrel distortion, G figure and the coordinate of B figure are mapped to the correspondence position of correction chart picture according to the forward mapping method of coordinate, then passing through arest neighbors interpolation algorithm and complete the filling of cavity point, its algorithm is mainly made up of three parts: Part I is the barrel distortion coefficient obtaining container number；Part II is the forward mapping of coordinate；Part III has been grey scale interpolation.
Step 21, barrel distortion coefficient acquisition.According to the deformation that wide-angle lens produces, the image height of field of view edge, less than desirable image height, is called barrel distortion, also referred to as undercorrection；Distortion Law may generally be expressed as δ (r)=k1r2+k2r4+ ..., wherein r is the radius from center of distortion to distortion point, k1、k2For unknown coefficient, namely calculative barrel distortion coefficient.For different camera lenses, k1、k2Value be different, in order to calculate in simple and convenient and sufficiently exact situation, only take and obtained k by MATLAB image calibration workbox1=-0.00000085, k2=-0.00000225, and directly have ignored coefficient k3,k4,….Quadratic term coefficient k during due to generation barrel distortion1、k2For negative, its timing should be carried out contrary operation, therefore take k1、k2For on the occasion of.Experiments show that, only take quadratic term coefficient k1、k2Enough having represented barrel distortion rule, too much high-order term can cause that system is unstable on the contrary and the complexity calculated is greatly increased.
Step 22, forward direction coordinate map.The forward mapping of coordinate refers to that input picture rounded coordinate is likely to be mapped in the middle of four coordinates of output image, and therefore, the gray value of this point coordinates is just allocated between these four points around it by the coordinate of input picture being mapped in output image one by one.Outside being likely to be mapped to the coordinate range of output image due to coordinate, therefore, first output image range is carried out suitable expanded range, by adjusting so that be enough to show the image that view picture is corrected.Owing to may be exceeded the coordinate range (Wang Jian of original image by the coordinate after forward mapping, the design of geometric correction card and realization based on FPGA, Xian Institute of Posts and Telecoms, 2011,18-20.), therefore before carrying out coordinate mapping, first a general image mapping range is determined, experiment takes 2.5 times that initial range is former fault image size of correction chart picture, then it needs to be determined that the center point coordinate of mapped image, generally taking the geometric center of mapped image, namely the half of length and width is as the picture centre point coordinates after mapping.If the image after mapping is beyond scope set in advance, whole image can be made to be positioned at the centre of whole picture by adjustment coordinate range and the coordinate translation by first pixel.Then the coordinate figure after the mapping updating formula by distorting obtained is mapped to the corresponding coordinate place of correction chart picture one by one, completes the mapping one by one of the two, figuratively, it is simply that complete the resettlement of coordinate.
Step 23, arest neighbors interpolation.After forward direction coordinate maps, owing to the scope of correction chart picture is certain to bigger than original image scope, so can make generation pinhole between image pixel, in order to well carry out holes filling, obtain the image of a good visual effect, be generally adopted gray-level interpolation to make up this defect after coordinate maps.The method of gray-level interpolation has a lot, and conventional interpolation method has arest neighbors interpolation, bilinear interpolation, bicubic interpolation etc..Nearest neighbor point interpolation for two dimensional image, is the gray value taking around interpolation point 4 nearest points of the neighbor pixel mid-range objectives point gray value as this point, and the advantage of nearest neighbor point interpolation is in that algorithm is simple, fast operation.Bilinear interpolation is first to carrying out first-order linear interpolation in horizontal direction, then again to carrying out first-order linear interpolation in vertical direction, or in turn, finally both is combined.Four pixel values contiguous around this kind of method impact point, as interpolation object, carry out linear interpolation in the two directions respectively to obtain the gray value of impact point.Bicubic interpolation, it is the interpolation method more more complicated than bilinear interpolation, it considers not only the dependency of four direct adjoint point gray values, it is also contemplated that the impact of gray-value variation rate between each adjoint point, make use of around impact point in bigger contiguous range the gray value of pixel as the object of interpolation.The simplest efficient arest neighbors interpolation algorithm might as well be adopted, arest neighbors interpolation method is exactly as the gray scale of this point using the gray scale of the closest pixel of coordinate position after conversion, this interpolation computing method is very simple and practical, also greatly reducing operand when reaching experiment purpose.
Step 3, by distortionless container number R scheme, G figure and B figure merge into distortionless RGB container number figure.
Step 4, include 4 steps at image pre-processing module: greyscale transformation, Wiener filtering, background eliminate and binaryzation (Wang Yan, He Junji, container number fast locating algorithm based on mathematical morphology, computer engineering and design, 2015,36 (8): 2162-2166.).
Step 41, image is carried out greyscale transformation.The reason of greyscale transformation is owing to the collection of container representation is obtained by digital camera.Image before pre-processing is coloured image entirely, and the quantity of information that coloured image comprises is very big, so needs bigger memory space, and therefore the speed of image procossing also can be slack-off therewith.Therefore image carries out greyscale transformation can make image only comprise monochrome information, does not comprise colour information.Greyscale transformation in order that improve image processing speed, make image convenient calculating in processes.
Step 42, image is carried out Wiener filtering.Reason to image filtering is when container representation is processed, and can run into the impact that noise brings.The information of noise on image is a kind of infringement.So the container representation collected being filtered very necessary.Wiener filter is assuming that noise is Gauss and additivity.And the optimal filter (GuoJM obtaining under Minimum Mean Square Error meaning on the basis mutually brought of signal and noise, LiuYF, Licenseplatelocalizationandcharactersegmentationwithfeed backself learningandhybridbinarizationtechniques, VehicularTechnology, IEEETransactionson, 2008,57 (3): 1,417 1424.), Wiener filtering in order that by tiny noise remove.
Step 43, filtered image is carried out background elimination.The reason so processed is that container body is subject to dust pollution in addition, brushes superincumbent character and can be influenced by impact, thus reducing contrast owing to the image of shooting often has the situation of uneven illumination.Owing in the container representation of present invention shooting, the distance of camera and container is 3 to 4m, the method that therefore in the present invention, container background eliminates is: 1. selects the square structure element being sized to 6*6 that image is carried out opening operation and obtains Background；2. image after background eliminates namely is obtained with original image subtracting background figure.
Step 44, carry out binaryzation to eliminating the image after background.The present invention uses Otsu method that image is carried out binaryzation.The container representation gathered can be impacted by the angle when reason so processed is because weather, shooting and the factor such as greasy dirt on casing, causes that the mass ratio of a lot of container representation collected is relatively low.Some images can exist serious uneven illumination, the relatively low and unconspicuous character edge of contrast.The relatively conventional method of Otsu method more can meet the present invention requirement to bianry image.
Step 5, case number (CN) slant correction.Owing to container number does not have left and right (or upper and lower) frame, if directly case number (CN) being carried out Hough detection, often will detect that the straight line on non-case number (CN) border.Therefore the present invention propose a kind of improvement based on Hough transform container number slant correction algorithm.Concrete as follows step by step:
Step 51, first container number bianry image is carried out connected component labeling, mark each character connected region of case number (CN).
Step 52, all character connected regions of case number (CN) according to step 1 labelling, add up and record the right margin pixel of these case number (CN) character zones.Then (Dinke is good to adopt method of least square, Shen Yunzhong, Ou Jikun, Least Square method fitting a straight line, Liaoning Project Technology University's journal: natural science edition, 2010 (1): 44-47.) these right margin pixels being carried out fitting a straight line, namely the straight line finally simulated is container number right margin straight line.
Step 53, case number (CN) image is carried out Hough detection, in two-dimensional array accumulator in parameter space, find out maximum and secondary maximum peak value (namely simulating boundary straight line), then the angle tilted as case number (CN) according to the angle theta of this straight line Yu vertical direction, and carry out this case number (CN) image reversely rotating θ.
The ultimate principle of the present invention, principal character and advantages of the present invention have more than been shown and described.Skilled person will appreciate that of the industry; the present invention is not restricted to the described embodiments; described in above-described embodiment and description is that principles of the invention is described; without departing from the spirit and scope of the present invention; the present invention also has various changes and modifications, and these changes and improvements both fall within the claimed scope of the invention.Claimed scope is defined by appending claims and equivalent thereof.
Claims (1)
1. the container number bearing calibration of a barrel distortion, it is characterised in that described bearing calibration comprises the steps:
Step one, the container number R that RGB container number picture breakdown is barrel distortion of barrel distortion is schemed, G figure and B figure；
Step 2, respectively the container number R of barrel distortion is schemed, G figure and B figure be adjusted in the horizontal direction and the vertical direction distortionless container number R figure, G figure and B figure；Specifically comprise the following steps that
Step 21, barrel distortion coefficient acquisition: barrel distortion is expressed as δ (r)=k1r2+k2r4+ ..., wherein r is the radius from center of distortion to distortion point, k1、k2For unknown coefficient, obtain k by MATLAB image calibration workbox1=-0.00000085, k2=-0.00000225, and ignore k2After other coefficients；Quadratic term coefficient k during due to generation barrel distortion1、k2For negative, its timing should be carried out contrary operation；
Step 22, forward direction coordinate map: first output image range is carried out suitable expanded range, by adjusting so that be enough to show the image that view picture is corrected；Then by the coordinate of input picture being mapped in output image one by one, input picture rounded coordinate is likely to be mapped in the middle of four coordinates of output image, and the gray value of described point coordinates is allocated between four coordinate points at it described；Taking 2.5 times that initial range is former fault image size of correction chart picture, then take the geometric center of mapped image, namely the half of length and width is as the picture centre point coordinates after mapping；If the image after mapping is beyond scope set in advance, just whole image is made to be positioned at the centre of whole picture by adjustment coordinate range and the coordinate translation by first pixel；Then the coordinate figure after the mapping updating formula by distorting obtained is mapped to the corresponding coordinate place of correction chart picture one by one, completes the mapping one by one of the two；
Step 23, arest neighbors interpolation: take around interpolation point the gray value of 4 nearest points of the neighbor pixel mid-range objectives point gray value as this point to make up the pinhole produced after coordinate maps；
Step 3, by distortionless container number R scheme, G figure and B figure merge into distortionless RGB container number figure；
Step 4, distortionless RGB container number Image semantic classification is obtained distortionless container number bianry image；Specifically comprise the following steps that
Step 41, image is carried out greyscale transformation；
Step 42, image is carried out Wiener filtering；
Step 43, filtered image is carried out background elimination: select the square structure element being sized to 6*6 that image is carried out opening operation and obtain Background；Then the image after background eliminates is obtained with original image subtracting background figure；
Step 44, Otsu method is used to carry out binaryzation to eliminating the image after background；
Step 5, distortionless container number bianry image is corrected to nonangular container number bianry image；Specifically include following step by step:
Step 51, first container number bianry image is carried out connected component labeling, mark each character connected region of case number (CN)；
Step 52, all character connected regions of case number (CN) according to step 51 labelling, add up and record the right margin pixel of these case number (CN) character zones；Then adopting method of least square that described right margin pixel is carried out fitting a straight line, namely the straight line finally simulated is container number right margin straight line.
Step 53, case number (CN) image is carried out Hough detection, in the two-dimensional array accumulator in parameter space, find out maximum and secondary maximum peak value, simulate boundary straight line；Then the angle that the boundary straight line simulated described in basis and the angle theta of vertical direction tilt as case number (CN), and carry out described case number (CN) image reversely rotating θ.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610105083.XA CN105787894A (en) | 2016-02-25 | 2016-02-25 | Barrel distortion container number correction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610105083.XA CN105787894A (en) | 2016-02-25 | 2016-02-25 | Barrel distortion container number correction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105787894A true CN105787894A (en) | 2016-07-20 |
Family
ID=56403755
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610105083.XA Pending CN105787894A (en) | 2016-02-25 | 2016-02-25 | Barrel distortion container number correction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105787894A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106770115A (en) * | 2016-12-26 | 2017-05-31 | 天津大学 | Pipeline inner circumferential liquid film characteristic optical distortion correction and measuring method |
| CN107067439A (en) * | 2017-04-26 | 2017-08-18 | 北京航天自动控制研究所 | A kind of container truck positioning detected based on headstock and bootstrap technique |
| CN109712087A (en) * | 2018-12-13 | 2019-05-03 | 上海畅联智融通讯科技有限公司 | A kind of barrel aberrance emendation algorithm |
| CN110940320A (en) * | 2019-07-19 | 2020-03-31 | 华北电力大学（保定） | Open stock ground monitored control system based on unmanned aerial vehicle cruises |
| CN111353952A (en) * | 2020-01-21 | 2020-06-30 | 佛山科学技术学院 | Method for eliminating black boundary after image distortion correction |
| CN112053369A (en) * | 2020-08-27 | 2020-12-08 | 西安迪威码半导体有限公司 | Low-delay image distortion removal and barrel mapping fusion algorithm |
| CN115100661A (en) * | 2022-08-24 | 2022-09-23 | 上海安维尔信息科技股份有限公司 | Container number extraction method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102024144A (en) * | 2010-11-23 | 2011-04-20 | 上海海事大学 | Container number identification method |
| CN102184398A (en) * | 2011-06-17 | 2011-09-14 | 电子科技大学 | Edge-detection-based container number positioning method |
-
2016
- 2016-02-25 CN CN201610105083.XA patent/CN105787894A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102024144A (en) * | 2010-11-23 | 2011-04-20 | 上海海事大学 | Container number identification method |
| CN102184398A (en) * | 2011-06-17 | 2011-09-14 | 电子科技大学 | Edge-detection-based container number positioning method |
Non-Patent Citations (8)
| Title |
|---|
| YAJIE ZHU ET AL.: "A NEW MODEL OF ADAPTIVE CONTAINER IMAGE PREPROCESSING", 《4TH INTERNATIONAL CONFERENCE ON MECHATRONICS, MATERIALS, CHEMISTRY AND COMPUTER ENGINEERING 》 * |
| 丁克良 等: "整体最小二乘法直线拟合", 《辽宁工程技术大学学报（自然科学版）》 * |
| 刘亚强 等: "桶形畸变图像的一种校正方法", 《西安邮电学院学报》 * |
| 孙凌红: "集装箱箱号智能识别算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 张涛: "集装箱扫描图像分析系统研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 王永亮: "基于视频的集装箱号码识别方法的研究与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 王炎 等: "基于数学形态学的集装箱箱号快速定位算法", 《计算机工程与设计》 * |
| 王雨: "车牌定位与字符分割技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106770115A (en) * | 2016-12-26 | 2017-05-31 | 天津大学 | Pipeline inner circumferential liquid film characteristic optical distortion correction and measuring method |
| CN107067439A (en) * | 2017-04-26 | 2017-08-18 | 北京航天自动控制研究所 | A kind of container truck positioning detected based on headstock and bootstrap technique |
| CN107067439B (en) * | 2017-04-26 | 2020-04-10 | 北京航天自动控制研究所 | Container truck positioning and guiding method based on vehicle head detection |
| CN109712087A (en) * | 2018-12-13 | 2019-05-03 | 上海畅联智融通讯科技有限公司 | A kind of barrel aberrance emendation algorithm |
| CN110940320A (en) * | 2019-07-19 | 2020-03-31 | 华北电力大学（保定） | Open stock ground monitored control system based on unmanned aerial vehicle cruises |
| CN111353952A (en) * | 2020-01-21 | 2020-06-30 | 佛山科学技术学院 | Method for eliminating black boundary after image distortion correction |
| CN111353952B (en) * | 2020-01-21 | 2023-05-05 | 佛山科学技术学院 | Method for eliminating black boundary after image distortion correction |
| CN112053369A (en) * | 2020-08-27 | 2020-12-08 | 西安迪威码半导体有限公司 | Low-delay image distortion removal and barrel mapping fusion algorithm |
| CN112053369B (en) * | 2020-08-27 | 2024-07-19 | 西安迪威码半导体有限公司 | Fusion algorithm for low-delay image de-distortion and barrel mapping |
| CN115100661A (en) * | 2022-08-24 | 2022-09-23 | 上海安维尔信息科技股份有限公司 | Container number extraction method and system |
| CN115100661B (en) * | 2022-08-24 | 2023-01-10 | 上海安维尔信息科技股份有限公司 | Container number extraction method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105787894A (en) | Barrel distortion container number correction method | |
| CN109086714B (en) | Form recognition method, recognition system and computer device | |
| CN107045629B (en) | Multi-lane line detection method | |
| CN105488501B (en) | The method of license plate sloped correction based on rotation projection | |
| CN110232389B (en) | Stereoscopic vision navigation method based on invariance of green crop feature extraction | |
| WO2022121039A1 (en) | Bankcard tilt correction-based detection method and apparatus, readable storage medium, and terminal | |
| CN103942797B (en) | Scene image text detection method and system based on histogram and super-pixels | |
| CN108537099A (en) | A kind of licence plate recognition method of complex background | |
| CN111353961B (en) | Document curved surface correction method and device | |
| KR20020018936A (en) | Apparatus and method for correcting distortion of input image | |
| CN110472521B (en) | Pupil positioning calibration method and system | |
| CN110245600B (en) | Unmanned aerial vehicle road detection method for self-adaptive initial quick stroke width | |
| CN104240259B (en) | High photographing instrument voucher intelligence cutting edge correction system and method based on contours segmentation | |
| CN111354047A (en) | Camera module positioning method and system based on computer vision | |
| CN111145124A (en) | Image tilt correction method and device | |
| CN115861409B (en) | Soybean leaf area measuring and calculating method, system, computer equipment and storage medium | |
| CN110610455B (en) | Effective region extraction method for fisheye image | |
| CN115273114B (en) | Document image optimization method and medium | |
| CN110969164A (en) | Low-illumination imaging license plate recognition method and device based on deep learning end-to-end | |
| CN103400343A (en) | Method for compensating uneven brightness of bottom view image under nighttime infrared | |
| CN106778766B (en) | Positioning point-based rotating number identification method and system | |
| CN110263784A (en) | The English paper achievement of intelligence identifies input method | |
| CN116469085B (en) | Monitoring method and system for risk driving behavior | |
| Tong et al. | Correction of perspective text image based on gradient method | |
| CN109410227B (en) | GVF model-based land utilization pattern spot contour extraction algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160720 |
|
| RJ01 | Rejection of invention patent application after publication |