CN104537663A - Method for rapid correction of image dithering - Google Patents
Method for rapid correction of image dithering Download PDFInfo
- Publication number
- CN104537663A CN104537663A CN201410824200.9A CN201410824200A CN104537663A CN 104537663 A CN104537663 A CN 104537663A CN 201410824200 A CN201410824200 A CN 201410824200A CN 104537663 A CN104537663 A CN 104537663A
- Authority
- CN
- China
- Prior art keywords
- image
- line
- line segment
- flating
- camera
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
-
- 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
- G06T2207/10016—Video; Image sequence
Abstract
The invention discloses a method for rapid correction of image dithering. The method is used for rapid correction of image dithering during video hydrologic observation and comprises the following steps that firstly, a characteristic line of a reference image is extracted, and a reference line of the reference image is determined; secondly, a characteristic line of a real-time image is extracted; thirdly, the characteristic line of the real-time image is matched with the reference line, and a correction parameter is worked out; fourthly, the correction parameter is used for correcting image dithering. According to the method, the angle of 360 degrees by which a camera rotates is divided into a plurality of fixed angles, and each angle corresponds to a template segment set. During practical matching, the angle of the camera is pre-judged first, a corresponding application scenario is determined, and therefore the corresponding template segment set needed for follow-up matching is determined. Thus, through one-shot pre-judging, the application scenario corresponding to each rotation change of the camera can be converted into a fixed camera scenario, characteristic line matching operating time is shortened, the correction parameter is obtained rapidly, and rapid correction of image dithering is achieved.
Description
Technical field
The present invention relates to field of video image processing, more specifically, relate to a kind of method for quickly correcting of the flating be applied in video hydrologic observation.
Background technology
Because traditional hydrologic monitoring means system is installed complicated, installation difficulty is large, and later maintenance cost is large, and need the shortcomings such as regular manual maintenance, utilizing video monitoring to carry out real-time hydrologic observation has become a kind of trend.What video hydrologic monitoring adopted is static image analysis technical limit spacing hydrographic information, namely extracts corresponding hydrological environment information by Computer Vision.Compared to the hydrologic monitoring equipment of traditional contact, easily install, not fragile, and follow-up maintenance cost is low.
And utilize Computer Vision to carry out hydrologic observation, be subject to various field environment factor (as wind, rain) impact.Before video image resolves hydrographic information, necessary image dithering correcting---Image Feature Matching must be carried out.What adopt traditionally is correct based on the matching process in region, be subject to the impact of gray scale light and shade change, in the occasion that can not conform very well change practical application, run into Changes in weather (rain, mist etc.) occasion and be matched to power reduction, and under dusk and night-time scene, feature based Region Matching method can not be used completely.Adopt characteristic curve to carry out matching and correlation compared to Character Area Matching, be not subject to half-tone information change interference, line features information has higher robustness than simple half-tone information, this is because line features is insensitive to image radiation distortion.
General image dithering correcting flow process, as figure below, is first read in image, is determined whether still image, then the feature line extraction carrying out realtime graphic with mate, calculate correction parameter, finally utilize correction parameter to carry out the correction of flating displacement.Wherein two important steps are that feature line extraction mates with characteristic curve.And in actual applications, above-mentioned image dithering correcting method life period complexity is higher, the deficiency that calculated amount is larger, reaches several seconds even tens of second its operation time.According to national Standard for observation of water level, video monitoring is higher to algorithm requirement of real time, is necessary to propose a kind of better image dithering correcting method of performance.
Known features line drawing comprises detection and the lines detection of image edge information.Edge-Detection Algorithm is existing many achievements in research both at home and abroad, and Canny algorithm is wherein comparatively ripe one.Canny operator has the following advantages: (1) signal to noise ratio (S/N ratio) is good, and namely the false retrieval Loss Rate of edge point is lower; (2) positioning performance is good, and the marginal point namely detected is as much as possible at the center of actual edge; (3) there is unique response to single edge, namely will suppress false edge as much as possible.But Canny algorithm is not one image processing algorithm fast.
In lines detection, the noiseproof feature of classical Hough transform is good, and can connect the short lines of conllinear, but shortcoming is parameter is difficult to select and the straight line resolution of calculation of complex, extraction is lower and lack local characteristics.
Based on the bearing calibration of LocalSearch algorithm characteristics lines matching, traditional shortcoming based on gradation of image information matches algorithm can be overcome, when mating by the affined transformation of image, the interference such as image quality change and image deformation are less, there is stronger robustness, can still provide good rectification when other algorithm cisco unity malfunctions, draw the correct displacement information of image.
LocalSearch is put forward in 1993 by J.Ross Beveridge, and Local Search is exactly the meaning of Local Search.The core of Local Search matching algorithm is depth-first search, degree of depth optimum search is exactly from current bit string, be in all spectra of 1, search for the initial bit string of optimum result as next iterative search at its Hamming distances, until current bit string is optimum, other field combination is not optimum, and this optimal result is exactly local optimum result.In order to ensure to obtain global optimum, must repeatedly search for, the most initial bit string is all random generation at every turn.Now time complexity drops to, but does not still reach operation time quick (below the 500ms) of project actual requirement.Because the working time of Local Search matching algorithm is long, greatly differ from each other with the requirement of real-time of actual items, therefore existing LocalSearch matching algorithm also only just rests on theoretical research stage.
Summary of the invention
In order to overcome above-mentioned the deficiencies in the prior art, the present invention proposes a kind of method for quickly correcting of flating, adopts this bearing calibration can improve the robustness of the jitter correction of video hydrologic observation and meet the requirement of its real-time monitored.
To achieve these goals, technical scheme of the present invention is:
A method for quickly correcting for flating, for the Fast Correction of flating in video hydrologic observation, comprises the following steps:
S1. extract the characteristic curve of reference picture, determine the reference line of reference picture;
S2. realtime graphic is carried out to the extraction of characteristic curve;
S3. the characteristic curve of realtime graphic is mated with reference line, calculate correction parameter;
S4. correction parameter is utilized to carry out the correction of flating;
The concrete mode of described step S1, S2 is:
S11. all line segments in reference picture are extracted: adopt edge detection algorithm setting to be applicable to adaptive threshold, extract the edge of reference picture; Adopt Line feature algorithm to obtain the vector form of straight line again, therefrom selected straight-line segment is as reference template;
S12. a mask is made to each selected straight-line segment, the mask of all selected straight-line segments is superimposed together, the mask plate of making image process;
S13. based on the scene image of mask plate to Real-time Collection, the edge detection algorithm identical with step S11, Line feature algorithm is adopted image to be carried out to the extraction of edge line segment, when real-time scene image is processed, only the region in mask plate is processed.
The mask plate that this method makes, in concrete image processing process, can reduce the calculated amount of image procossing, thus reduces the consuming time of program operation, for fast image processing provides basis.When the scene image process to Real-time Collection, only need that process is carried out to the information in mask plate region and calculate, significantly can reduce the calculated amount of image procossing, computing in mask plate region, characteristic curve can be extracted accurately equally.
Preferably, making image mask plate is the image masks version making different angles, is specially: 360 degree that are rotated by camera are divided into several fixing angles, for each angle, select out a corresponding masterplate line-segment sets, thus make corresponding image masks version.
Produce several image masks versions according to the anglec of rotation of camera, namely decrease template line segment aggregate, thus the space of coupling can be reduced, finally reach the time reducing characteristic curve coupling and run.
Preferably, described step S3 is the coupling utilizing LocalSearch characteristic curve matching process to carry out image, and calculate size---the amount of jitter size of the relative displacement of realtime graphic and reference picture, unit is pixel; In step S3 by the concrete mode that the characteristic curve of realtime graphic and reference line carry out mating be:
S31. to the anticipation of camera, according to the anglec of rotation of camera, determine the reference template picture mated, thus determine the image masks version of coupling;
S32. line segments extraction is carried out to Real-time Collection scene image, utilize Depth Priority Algorithm to search in image masks version region inside, find optimum combination, and output matching result, obtain correction parameter.
By the anticipation to camera, can the application scenarios that each of camera rotates change corresponding be converted to fixed camera scene, thus the time that reduction characteristic curve coupling is run.
Preferably, after extraction Real-time Collection scene image carries out line segment, also according to the line segment pair that the priori exclusive segment preset can not mate.
Preferably, described default priori comprise line segment between the right length ratio of angle, line segment and line segment between distance.To the data line-segment sets of input, by some prioris, as: the angle between two line segments is scope 2 ~ 7 degree, the length ratio of two line segments to be 1/7 ~ 9/2 or two distances between line segment be 6 ~ 8 pixels get rid of in advance data line segment concentrates can not with the line segment that matches in masterplate line-segment sets, can package space be reduced, thus reduce calculated amount.Line segment between the right length ratio of angle, line segment and line segment between this three parameter of distance, be the empirical value in Practical Project, do not do quantitative test and derivation, running into specific environment can adjust.In addition, the too wide matching condition that directly causes of these setting parameter restrictive conditions is loose, and result is that coupling possibility is more, and match time, complexity was larger; Restrictive condition too strictly may affect and be matched to power.
Compared with prior art, beneficial effect of the present invention is: method of the present invention is applied to the jitter correction in video hydrologic observation, overcome the shortcoming that feature based Region Matching is subject to the factor impacts such as half-tone information change, the method of mask plate is adopted to reduce computation complexity, experimental result, show that the process (comprising line drawing and lines matching) of whole computed image displacement can control within 1s, meet the requirement of scan picture.
Accompanying drawing explanation
Fig. 1 is conventional images jitter correction process flow diagram.
Fig. 2 is the process flow diagram of existing LocalSearch algorithm.
Fig. 3 is the process flow diagram of the LocalSearch algorithm improved in the present invention.
Fig. 4 is mask design drawing; Wherein d is mask buffer strip width, and AB is with reference to line segment.
Fig. 5 is mask plate areal map; White portion is mask, and characteristic curve image procossing carries out in this region.
Fig. 6 is realtime graphic feature line extraction design sketch.
Fig. 7 is emulation schematic diagram.
Embodiment
Below in conjunction with accompanying drawing, the present invention will be further described, but embodiments of the present invention are not limited to this.
Embodiment 1
The present embodiment improves traditional images match bearing calibration, is to adopt characteristic curve to mate.By improving characteristic curve matching algorithm, reducing the time (controlling within 1s) needed for calculating, making it both improve the stability of images match correction, also meeting the requirement of real-time of video surveillance applications.
Key working time of characteristic curve matching algorithm is the size of package space; Therefore the main method reducing characteristic curve coupling working time is to reduce the size of package space.Package space is determined by masterplate line-segment sets M and real-time diagram data line-segment sets N.Before not being optimized package space, the size of package space is, i.e. complete or collected works.Certainly, be the requirement that cannot meet real-time system.
Reduce characteristic matching space to consider from two aspects: one is the absolute quantity of line segment, another is the right relative populations of coupling line segment.Absolute quantity refers to only needs buildings principal outline line segment in image, as long as because have buildings principal outline line segment just can complete coupling.Reduce the absolute number value of line segment, directly can reduce the absolute size of M and N.
Another method reducing package space size is exactly utilize priori to reduce the right quantity of coupling line segment.Although in the various project of reality, when gathering picture, can relate to the various change such as translation, Rotation and Zoom, the present embodiment first only considers the situation of translation.So just can utilize as line segment between distance, line segment between angle and the priori such as the right length ratio of line segment get rid of some line segments pair that can not mate, reduce package space thus, and then reduce the time that characteristic curve coupling runs.
In hydrologic monitoring, although the position of camera and focal length are fixing, therefore put aside the convergent-divergent of picture, except translation, also can relate to the rotation of camera.Add and rotate this change, package space will be much larger, and the time of coupling is inevitable elongated.The method adopted in embodiment is 360 degree of angles being divided into several and fixing that camera is rotated, and for each angle, has the masterplate line-segment sets that corresponding.
In actual match, the first angle of anticipation camera, determines corresponding application scenarios, thus determines the masterplate line-segment sets that subsequent match will use.So, first by an anticipation, can the application scenarios that each of camera rotates change corresponding be converted to fixed camera scene, thus the time that reduction characteristic curve coupling is run.
The present embodiment is the robustness of the jitter correction improving video hydrologic observation and meets the requirement of its real-time monitored, its principal feature is: (1) feature line extraction process is that the line segment of every bar coupling makes mask, greatly reduce the calculated amount of view data, reduce consuming time, reach requirement of real-time; (2) application scenarios that characteristic curve matching rotation changes is converted into the application scenarios of fixed camera, reduces characteristic matching space, meet the requirement of real-time of Practical Project, realize the Fast Correction of flating.
Below set forth the implementing procedure that the bearing calibration of this fast jitter is total:
1, characteristic curve Detection and Extraction flow process:
(1) all line segments in known reference image are extracted in advance: adopt the good Canny edge detection operator setting of robustness to be applicable to adaptive threshold, extract the edge of image.Then, Line feature algorithm is adopted to obtain the vector form of straight line; Therefrom the selected straight-line segment be applicable to is as reference template.
The process adopting Canny edge detection operator to realize the extraction of image border line is: utilize canny algorithm to detect band phase place (angle) edge in image, the Phase Tracking edge (assuming that its phase place of marginal point is on the same line roughly the same) at recycling edge, obtain the set (referred to herein as point set) having the point of linear feature that one group of group is made up of marginal point, rejecting point set number is less than a certain threshold value (being usually set to 10-20), finally utilizes least square fitting to obtain the vector form of straight line.
At the selected straight-line segment be applicable to as during with reference to template, " standard " that its reference line segment is chosen is a partially subjective concept.Standard is in the present embodiment: 1) selected line segment as far as possible effective (fit effective with the object of reference in image), 2) comprise vertically and the straight line of horizontal both direction as far as possible, the profile of object of reference can be formed as far as possible.3) the consult straight line number generally chosen is 4-10.
(2) for each selected line segment makes a mask, as the calculation window of scan picture.The mask of all selected line segments is superimposed together, and makes the mask plate of an image procossing, and object reduces the calculated amount of image procossing, thus reduces the consuming time of program operation, for fast image processing provides possibility.
(3) according to mask plate, to the scene image of Real-time Collection, still adopt Canny algorithm image to be carried out to the extraction of edge line segment, only only the region in mask plate is calculated during image procossing.Such image procossing calculated amount obviously reduces.
Canny edge detection operator is the extraction for characteristic curve; The extraction that this image dithering correcting method comprises characteristics of image with mate, characteristics of image comprises the features such as point, edge, profile, straight-line segment.
2, characteristic curve coupling flow process
In flating fast method, LocalSearch algorithm is the coupling for characteristic curve, and images match is then find the feature of the same name of image, calculates the shake displacement of two width images (reference picture and realtime graphic).A brief description is done to the principle and characteristics of LocalSearch algorithm below:
The process flow diagram of existing LocalSearch algorithm is as shown in Figure 2:
1) first the line segment information set M(of input corresponding to template image with TXT representation of file) and line segment information set N corresponding to real time data image, two data are all the line segments of extraction in 1.
2) then determine package space according to M and N, there is the changes such as rotation, convergent-divergent and skew due to image, therefore corresponding package space is the complete or collected works of M × N;
3) Depth Priority Algorithm is then utilized to find optimum combination in the package space determined;
4) output matching result.
This LocalSearch matching algorithm only just rests on theoretical research stage, and reason is to greatly differ from each other its long working time with the requirement of real-time of actual items.
Go to be applied in Practical Project by LocalSearch matching algorithm, the present embodiment has made following optimization to reduce the working time of algorithm to it:
(1) by anticipation to camera, determine to be about to the masterplate picture for coupling, for the masterplate picture determined, line segment quantity that masterplate picture comprises can be reduced as far as possible to reduce the calculated amount of algorithm;
(2) to the data line-segment sets of input, by some prioris, as: the distance between the length ratio of the angle between two line segments, two line segments and two line segments get rid of in advance data line segment concentrate can not with the line segment that matches in masterplate line-segment sets, to reduce package space, thus reduce calculated amount.
Above-mentioned three optimum configurations have certain standard, below provide the scope that several parameter can set, and 1, angular range 2 ~ 7 degree (experiment is set as 5 degree); 2, two line segment length ratios: 1/7 ~ 4.5 times; 3, line segment distance 6 ~ 8 pixels.Illustrate: these parameters, be the empirical value in Practical Project, do not do quantitative test and derivation, running into specific environment can adjust.In addition, the too wide matching condition that directly causes of these setting parameter restrictive conditions is loose, and result is that coupling possibility is more, and match time, complexity was larger; Restrictive condition too strictly may affect and be matched to power.
By above two measures, can greatly reduce the calculated amount of LocalSearch algorithm, thus can meet the requirement of real-time of video hydrologic monitoring system, LocalSearch matching algorithm first time is applied in the middle of Practical Project.
As shown in Figure 3, key step is as follows for the mode of the characteristic curve coupling in hydrologic monitoring:
1) anticipation is carried out to camera, determine the angle that camera rotates, determined the masterplate line-segment sets that use by the angle of camera;
2) input real-time line-segment sets, set the parameter that priori is corresponding, get rid of the line segment pair that some can not mate, reduce characteristic matching space;
3) utilize Depth Priority Algorithm to search in package space, find optimum combination, and output matching result.
The method adopted in the present embodiment is applied to the jitter correction in video hydrologic observation, overcome the shortcoming that feature based Region Matching is subject to the factor impacts such as half-tone information change, the method of mask plate is adopted to reduce computation complexity, experimental result, show that the process (comprising line drawing and lines matching) of whole computed image displacement can control within 1s, meet the requirement of scan picture.
As shown in Table 1, under different image-forming conditions, when its feature line extraction success ratio is identical, what do not use the mask extraction characteristic curve time used extracts the characteristic curve time used higher than use mask far away, namely adopt method disclosed in the present embodiment significantly can reduce the time of characteristic matching, accelerate the correction of flating.
The present embodiment is mainly based on two algorithms---feature line extraction and matching algorithm.Two Algorithms T-cbmplexity itself are very high, and Canny algorithm belongs to image processing algorithm, if be not optimized improvement, its time complexity is
o (xy),
x,
ywide, high for image.Characteristic curve matching algorithm adopts LocalSearch algorithm, and time complexity is also
o (mn),
m,
nfor consult straight line and straight line number to be matched.After adopting mask technology,
x,
ycan obviously reduce.In addition, adopt mask technology, can the line segment extracted be divided into groups in advance, namely can only combine by the line segment extracted in mask corresponding to it with reference to line segment, consult straight line in coupling and between straight line to be matched the mapping relations of multi-to-multi become one-to-many, now Algorithms T-cbmplexity is
o (n), n is straight line number to be matched.After adding priori (such as rotation angle is no more than △ θ, and the length ratio of line segment can not be less than constant K etc.), get rid of some impossible combinations, Algorithms T-cbmplexity is less than
o (n).Algorithm in the present embodiment, by various improvement, makes it really reach fast jitter correcting algorithm.Put into practice in project application, algorithm controls at below 0.5s at the desk-top machine run time of office, is maximumly no more than 1.0s.And before algorithm improvement, need tens seconds even tens seconds working time.
Above-described embodiments of the present invention, do not form limiting the scope of the present invention.Any amendment done within spiritual principles of the present invention, equivalent replacement and improvement etc., all should be included within claims of the present invention.
Claims (5)
1. a method for quickly correcting for flating, for the Fast Correction of flating in video hydrologic observation, comprises the following steps:
S1. extract the characteristic curve of reference picture, determine the reference line of reference picture;
S2. realtime graphic is carried out to the extraction of characteristic curve;
S3. the characteristic curve of realtime graphic is mated with reference line, calculate correction parameter;
S4. correction parameter is utilized to carry out the correction of flating;
It is characterized in that, the concrete mode of described step S1, S2 is:
S11. all line segments in reference picture are extracted: adopt edge detection algorithm setting to be applicable to adaptive threshold, extract the edge of reference picture; Adopt Line feature algorithm to obtain the vector form of straight line again, therefrom selected straight-line segment is as reference template;
S12. a mask is made to each selected straight-line segment, the mask of all selected straight-line segments is superimposed together, the mask plate of making image process;
S13. based on the scene image of mask plate to Real-time Collection, the edge detection algorithm identical with step S11, Line feature algorithm is adopted image to be carried out to the extraction of edge line segment, when real-time scene image is processed, only the region in mask plate is processed.
2. the method for quickly correcting of flating according to claim 1, it is characterized in that, making image mask plate is the image masks version making different angles, be specially: 360 degree that are rotated by camera are divided into several fixing angles, for each angle, select out a corresponding masterplate line-segment sets, thus make corresponding image masks version.
3. the method for quickly correcting of flating according to claim 2, it is characterized in that, described step S3 is the coupling utilizing LocalSearch characteristic curve matching process to carry out image, and calculate size---the amount of jitter size of the relative displacement of realtime graphic and reference picture, unit is pixel; In step S3 by the concrete mode that the characteristic curve of realtime graphic and reference line carry out mating be:
S31. to the anticipation of camera, according to the anglec of rotation of camera, determine the reference template picture mated, thus determine the image masks version of coupling;
S32. line segments extraction is carried out to Real-time Collection scene image, utilize Depth Priority Algorithm to search in image masks version region inside, find optimum combination, and output matching result, obtain correction parameter.
4. the method for quickly correcting of flating according to claim 3, is characterized in that, after extraction Real-time Collection scene image carries out line segment, also according to the line segment pair that the priori exclusive segment preset can not mate.
5. the method for quickly correcting of flating according to claim 4, is characterized in that, described default priori comprise line segment between the right length ratio of angle, line segment and line segment between distance.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410824200.9A CN104537663B (en) | 2014-12-26 | 2014-12-26 | A kind of method for quickly correcting of flating |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410824200.9A CN104537663B (en) | 2014-12-26 | 2014-12-26 | A kind of method for quickly correcting of flating |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104537663A true CN104537663A (en) | 2015-04-22 |
| CN104537663B CN104537663B (en) | 2018-01-02 |
Family
ID=52853182
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410824200.9A Expired - Fee Related CN104537663B (en) | 2014-12-26 | 2014-12-26 | A kind of method for quickly correcting of flating |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104537663B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108366243A (en) * | 2018-01-23 | 2018-08-03 | 微幻科技(北京)有限公司 | A kind of video jitter removing method and device |
| CN109493391A (en) * | 2018-11-30 | 2019-03-19 | Oppo广东移动通信有限公司 | Camera calibration method and device, electronic equipment, computer readable storage medium |
| CN111340888A (en) * | 2019-12-23 | 2020-06-26 | 首都师范大学 | Light field camera calibration method and system without white image |
| CN113330480A (en) * | 2019-02-11 | 2021-08-31 | 康蒂-特米克微电子有限公司 | Modular image restoration method |
| CN113382177A (en) * | 2021-05-31 | 2021-09-10 | 上海东方传媒技术有限公司 | Multi-view-angle surrounding shooting method and system |
| CN115760595A (en) * | 2022-10-26 | 2023-03-07 | 中国电子科技集团公司第五十四研究所 | Ultra-wide-angle lens photo distortion correction method based on line segment characteristics |
| US11961215B2 (en) | 2019-02-11 | 2024-04-16 | Conti Temic Microelectronic Gmbh | Modular inpainting method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090028585A1 (en) * | 2007-07-27 | 2009-01-29 | Shilin Guo | Interactive visual card-selection process for mitigating light-area banding in a pagewide array |
| CN103177441A (en) * | 2013-01-24 | 2013-06-26 | 中国科学院对地观测与数字地球科学中心 | Image geometric correction method based on straight line segments |
| CN103345642A (en) * | 2013-06-28 | 2013-10-09 | 华中科技大学 | Image matching method based on dotted line dual |
-
2014
- 2014-12-26 CN CN201410824200.9A patent/CN104537663B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090028585A1 (en) * | 2007-07-27 | 2009-01-29 | Shilin Guo | Interactive visual card-selection process for mitigating light-area banding in a pagewide array |
| CN103177441A (en) * | 2013-01-24 | 2013-06-26 | 中国科学院对地观测与数字地球科学中心 | Image geometric correction method based on straight line segments |
| CN103345642A (en) * | 2013-06-28 | 2013-10-09 | 华中科技大学 | Image matching method based on dotted line dual |
Non-Patent Citations (1)
| Title |
|---|
| 李松霖等: "基于特征线匹配的城市建筑物识别方法研究", 《遥感技术与应用》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108366243A (en) * | 2018-01-23 | 2018-08-03 | 微幻科技(北京)有限公司 | A kind of video jitter removing method and device |
| CN108366243B (en) * | 2018-01-23 | 2019-10-29 | 微幻科技(北京)有限公司 | A kind of video jitter removing method and device |
| CN109493391A (en) * | 2018-11-30 | 2019-03-19 | Oppo广东移动通信有限公司 | Camera calibration method and device, electronic equipment, computer readable storage medium |
| CN113330480A (en) * | 2019-02-11 | 2021-08-31 | 康蒂-特米克微电子有限公司 | Modular image restoration method |
| US11961215B2 (en) | 2019-02-11 | 2024-04-16 | Conti Temic Microelectronic Gmbh | Modular inpainting method |
| CN111340888A (en) * | 2019-12-23 | 2020-06-26 | 首都师范大学 | Light field camera calibration method and system without white image |
| WO2021129437A1 (en) * | 2019-12-23 | 2021-07-01 | 首都师范大学 | Method and system for light calibration field camera without requiring white image |
| CN113382177A (en) * | 2021-05-31 | 2021-09-10 | 上海东方传媒技术有限公司 | Multi-view-angle surrounding shooting method and system |
| CN115760595A (en) * | 2022-10-26 | 2023-03-07 | 中国电子科技集团公司第五十四研究所 | Ultra-wide-angle lens photo distortion correction method based on line segment characteristics |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104537663B (en) | 2018-01-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104537663A (en) | Method for rapid correction of image dithering | |
| CN106650829B (en) | A kind of picture similarity calculating method | |
| CN111612760B (en) | Method and device for detecting obstacles | |
| CN102768071B (en) | Non-refrigerant thermal imager shutter-free nonuniformity correcting device based on template method | |
| CN105554447B (en) | A kind of coal working face real-time video splicing system based on image processing techniques | |
| CN103745449B (en) | Rapid and automatic mosaic technology of aerial video in search and tracking system | |
| WO2013135033A1 (en) | Tunnel deformation online monitoring system based on image analysis and application thereof | |
| CN103337077B (en) | A kind of based on multi-scale division and the visible ray of SIFT and the method for infrared image registration | |
| CN108334844B (en) | Line automatic tracking method for power transmission line inspection | |
| CN102542282B (en) | Mosaic detection method and mosaic detection device for passive images | |
| CN106650682A (en) | Method and device for face tracking | |
| CN106845410B (en) | Flame identification method based on deep learning model | |
| CN107248174A (en) | A kind of method for tracking target based on TLD algorithms | |
| CN104933403B (en) | A kind of monitoring method of the conductor galloping based on conductor spacer identification | |
| CN104253930A (en) | Real-time video defogging method | |
| CN101729911A (en) | Multi-view image color correction method based on visual perception | |
| CN110130987B (en) | Tunnel convergence deformation monitoring method based on image analysis | |
| CN104079800A (en) | Shaking preventing method for video image in video surveillance | |
| WO2021068497A1 (en) | Grate cooler working state monitoring method based on image recognition | |
| CN110569785A (en) | Face recognition method based on fusion tracking technology | |
| TWI383690B (en) | Method for image processing | |
| CN113902792A (en) | Building height detection method and system based on improved RetinaNet network and electronic equipment | |
| CN103996199A (en) | Movement detection method based on depth information | |
| CN106570834A (en) | Image correction method for pixel modulation visible light communication system | |
| CN107945196B (en) | Image edge detection method for aerial photography correction image splicing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180102 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |