CN103679623B - Solve the structure of image deformation - Google Patents
Solve the structure of image deformation Download PDFInfo
- Publication number
- CN103679623B CN103679623B CN201210345140.3A CN201210345140A CN103679623B CN 103679623 B CN103679623 B CN 103679623B CN 201210345140 A CN201210345140 A CN 201210345140A CN 103679623 B CN103679623 B CN 103679623B
- Authority
- CN
- China
- Prior art keywords
- frame
- ram
- module
- correlation
- data
- 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
Abstract
The structure solving image deformation that the present invention provides, it is characterized in that, provide module, data transmission module, a RAM to select module, the 2nd RAM to select module, correlation module, RAM storage array, statistics with histogram module, lock frame state control module, image detection module including two field picture;The structure solving image deformation that the present invention provides, it is adaptable to two or more than two adjacent have partly overlapping image correlation and calculate, and image is without spin.Achieve by the mode of ASIC that high speed, low cost, amount of calculation be little and the structure of streamlined.
Description
Technical field
The present invention relates to field of image recognition, particularly relate to solve the structure of image deformation.
Background technology
Image mosaic technology obtains the image at ultra-wide visual angle by having partly overlapping image by one group and carrying out seamless spliced.Going out of this class technology
Now make equipment more commonization of collection image.This technology is at field of video processing, field of medical image processing, fingerprint field and military neck in recent years
Territory is all widely used.Therefore, study and propose a kind of accurately and quickly merging algorithm for images and be of great practical significance.
The image geometry deformation that existing image mosaic technology has been spliced is the most serious.These a lot of links when for back-end processing all can cause sternly
The impact of weight, directly influences image recognition rate.And during the redundancy estimating neighbor map photo, existing method consumes very much hardware resource,
Or need to consume the longer process time, so that the cost performance of product reduces.
Summary of the invention
For solving above-mentioned technical problem, the structure solving image deformation that the present invention provides, it is characterised in that include that two field picture provides module, data
Transport module, the oneth RAM select module, the 2nd RAM select module, correlation module, RAM storage array, statistics with histogram module,
Lock frame state control module, image detection module;
Image sensing signal that offset data that described lock frame state control module calculates according to correlation module, image detection module provide, straight
The lower limit gray scale threshold value that side's figure statistical module provides, it is judged that this frame gives a RAM the need of lock frame or release lock frame, output lock control frame signal
Module and the 2nd RAM is selected to select module;
Described lock frame state control module judge this frame the need of lock frame or release lock frame according to being: if two two field pictures at acquisition target relative to adopting
Displacement on the main moving direction of collection window does not reaches the threshold value of regulation, then abandon present frame, uses next frame image to replace present frame and template frame
Carrying out correlation calculations, template frame is that frame currently locking frame;If the displacement between two two field pictures has reached the threshold value of regulation, then output is when advancing
The template frame of row correlation calculations, is converted into template frame lock present frame, carries out next round correlation calculations.
A described RAM selects module by the incoming correlation module of selected data in present frame RAM and template frame RAM;Described selected number
Module is selected to select to put into the frame image data of the incoming frame RAM in RAM storage array according to for a RAM.
Described RAM storage array is made up of four or more RAM;Being assigned as of described RAM storage array:
At least currently carry out the current frame image data of correlation calculations with a RAM storage, be defined as present frame RAM;At least with one
RAM storage is currently carrying out the template frame image data of correlation calculations, is defined as template frame RAM;At least current with a RAM storage
The frame image data gathered, is defined as incoming frame RAM;The view data at least spread out of with a RAM storage current release lock frame, definition
For output frame RAM;Also comprise a RAM and store current relevance matrix, be defined as correlation matrix RAM.
Described two field picture provides module to provide each two field picture collected, and selects module to select to put into RAM by a RAM two field picture
Incoming frame RAM in storage array;
Described data transmission module transmission correlation data, offset data, two field picture;
A described RAM selects module to select to put in the incoming frame RAM in RAM storage array by the frame image data of reception, and will be current
The incoming correlation module of selected data in frame RAM and template frame RAM;Described 2nd RAM select module to output frame RAM read address,
Read data and reading enable carries out processing and transfer;
Described statistics with histogram module obtains the original histogram of each frame image data in real time, obtains its corresponding lower limit gray scale respectively further according to rectangular histogram
Thresholding;
The correlation data that described image detection module is calculated by frame image data and the correlation module of Real-time Collection judges that acquisition target is relative
State in acquisition window.
The storage rule of described RAM storage array is:
Under release lock frame condition, deposit template frame RAM and become depositing output frame RAM, the template frame that output is released;Deposit present frame RAM
Become depositing template frame RAM, become the displacement template of correlation calculations comparison next time;Deposit incoming frame RAM to become depositing present frame RAM,
Correlation calculations is carried out with template frame;Deposit output frame RAM to become depositing incoming frame RAM;
Under lock frame condition, the RAM depositing present frame and the RAM depositing incoming frame exchanges, and present frame RAM is become incoming frame RAM,
And abandon present frame, incoming frame is become present frame and does computing.
This frame of described judgement also includes the need of the foundation of lock frame or release lock frame: if the image sensing signal that image detection module provides is collecting window
Mouth does not collects effective acquisition target, updates lock frame the most all the time, and resets lock frame count;
Collect effective acquisition target but acquisition target does not starts mobile if the image sensing signal that image detection module provides is acquisition window, then judge
Before and after statistics with histogram, whether the difference of two frame lower limit gray scale thresholdings reaches threshold value;Or two frame data have dependency side-play amount;Maybe reach when lock frame count
During to the threshold value specified, it is determined that present frame is forced to update lock frame, and resets lock frame count;
Effective acquisition target is collected and acquisition target starts mobile if the image sensing signal that image detection module provides is acquisition window, and now water
During the threshold component that flat or vertical offset sum reduces more than the increase with lock frame count, update lock frame, and reset lock frame count.
Being separated with frame number between described module frame and present frame, the frame number at its interval depends on that locking frame judges;If current template frame and present frame meet release
Putting lock frame bar part, the frame number between rear pattern plate frame and the present frame of release lock frame is 0;If current template frame and present frame are unsatisfactory for release lock frame requirement,
Frame number between rear pattern plate frame and the present frame of release lock frame is more than or equal to 1.
Described correlation module selects the template frame that transmits of module and present frame to calculate the dependency of two two field pictures according to the RAM received;If institute
Stating template frame is with reference to that motionless frame in two two field pictures, and the most described present frame is another frame moved up and down with reference to template frame in two two field pictures
Image.According to template frame and the coverage condition of present frame, choose the part or all of data of two two field picture overlay areas, by two two field picture overlay areas
In data one to one ask squared difference sum to obtain the correlation operation result under current coverage condition;
Two two field pictures correlation operation result composition correlation matrix under all coverage conditions;Described correlation matrix is exported by correlation module
The storage of RAM storage array and offset tracking block, and the offset data of two two field pictures is calculated by offset tracking block.
Described correlation data includes the maximum of correlation matrix, minima and coordinate thereof;Described offset data includes the two frame figures participating in computing
The side-play amount both horizontally and vertically of picture, speed both horizontally and vertically, speed weight both horizontally and vertically.
The empirical value obtained is tested according to described threshold value.
The structure solving image deformation that the present invention provides, it is adaptable to two or more than two adjacent have partly overlapping image correlation and calculate,
And image is without spin.Achieve by the mode of ASIC that high speed, low cost, amount of calculation be little and the structure of streamlined.Utilize the image that the method is spliced,
Owing to arithmetic accuracy is high, error is little, the phenomenons such as image offset, image deformation, short image, streaking can be significantly corrected, for subsequent treatment
Lay a solid foundation;It is especially suitable for the popularization of fingerprint class commercial product.
Accompanying drawing explanation
Fig. 1 is the module rack composition that the present invention locks under frame mechanism;
Fig. 2 is each frame image data sequencing schematic diagram on collecting flowchart;
Fig. 3 be in RAM storage array of the present invention each RAM redirect graph of a relation, wherein Fig. 3-1 is lock frame and each RAM when releasing lock frame
Redirecting graph of a relation, Fig. 3-2 is the transition diagram of RAM under release lock frame condition, and Fig. 3-3 is the transition diagram of RAM under lock frame condition.
Detailed description of the invention
Illustrate that the preferred embodiments of the present invention, specific embodiment described herein only in order to explain the present invention, are not used to below in conjunction with the accompanying drawings
Limit the present invention.
As it is shown in figure 1, solve image deformation structure, including two field picture provide module 1, data transmission module the 2, the oneth RAM select module 3,
2nd RAM selects module 4, correlation module 5, RAM storage array 6, statistics with histogram module 7, lock frame state control module 8, image
Detection module 9.
Two field picture provides module 1 to provide each two field picture collected, and selects module 3 to select to put into RAM by a RAM two field picture
Incoming frame RAM in storage array 6.Data transmission module 2 transmits correlation data, two field picture, threshold parameter.Oneth RAM selects module
The frame image data of reception is put in the incoming frame RAM in RAM storage array 6 by 3, and the RAM that two field picture provides module 1 provide reads
Write address, read-write data, reading write enable signal carry out processing and transfer.2nd RAM selects the RAM that data transmission module 2 is provided by module 4
Read address and reading enable carries out processing and transfer.
Correlation module 5 calculates correlation data and the offset data of present frame, and its method is two that reception the oneth RAM selects module 3 to transmit
Two field picture, and two two field pictures are set to template frame and present frame, if described template frame is with reference to that motionless frame in two two field pictures, then described
Present frame is another two field picture moved up and down with reference to template frame in two two field pictures.According to template frame and the coverage condition of present frame, choose two frames
Data one to one in two two field picture overlay areas are asked squared difference sum currently to be covered by part or all of data of image overlay area
Correlation operation result in the case of lid;Two two field pictures correlation operation result under various coverage conditions forms the correlation matrix of two two field pictures.
Described correlation data includes the maximum of correlation matrix, minima and coordinate thereof.
Correlation data is output in parallel to data transmission module 2, image detection module 9 and lock frame state control module 8 by correlation module 5, and
Correlation matrix is exported RAM storage array 6.Correlation module 5 carries out side-play amount calculating to correlation matrix, obtains the skew of present frame
Amount data, described offset data includes the two frames side-play amount both horizontally and vertically participating in computing, speed both horizontally and vertically, level and
The speed weight etc. of vertical direction.
RAM storage array 6 is used for circulating storage two field picture or correlation matrix and other data, is made up of four or more RAM.RAM
Storage array 6 is assigned as: is at least currently carrying out the current frame image data of correlation calculations with a RAM storage, is being defined as present frame
RAM;At least currently carry out the template frame image data of correlation calculations with a RAM storage, be defined as template frame RAM;At least use
The frame image data that one RAM storage is currently gathering, is defined as incoming frame RAM;At least pass with a RAM storage current release lock frame
The view data gone out, is defined as output frame RAM;Also comprise a RAM and store current relevance matrix, be defined as correlation matrix RAM.
Statistics with histogram module 7 obtains the original histogram of each frame image data in real time, obtains the lower limit gray scale of its correspondence respectively further according to rectangular histogram
Thresholding LTH.
The acquiring method of lower limit gray scale threshold value is: set a threshold grayscale greyTH and lower limit gray scale threshold ratio low_Rate, wherein threshold
Value gray scale greyTH is the integer arbitrarily chosen in the range of gradation of image, and lower limit gray scale threshold ratio low_Rate is the arbitrary small number of 0 to 1.
The rectangular histogram of each two field picture of real-time statistics, obtaining less than threshold grayscale greyTH number of pixels summation is cnt_th, then lower limit gray scale thresholding LTH
For: count from zero the pixel grey scale at cnt_th × low_Rate.
Image detection module 9 by the correlation data that frame image data and the correlation module 5 of Real-time Collection calculate judge acquisition target relative to
The state of acquisition window, including: acquisition window collects effective acquisition target, acquisition window does not collects effective acquisition target and acquisition target and adopting
Collect states such as moving on window.
Such as in slice fingerprint sensor, finger detected, be and effective acquisition target detected.
Lock frame state control module 8 judges that in the correlation calculations of the every frame of correlation module 5 this frame is the need of lock frame or release lock frame, output lock
Control frame signal selects module 3 and the 2nd RAM to select module 4 to a RAM.Lock frame state control module 8 judges that this frame is the need of lock frame
Or the foundation of release lock frame is: if two two field pictures do not reach the threshold of regulation at acquisition target relative to the displacement on the main moving direction of acquisition window
Value, then abandon present frame, uses next frame image replace present frame and participate in correlation calculations with template frame, and template frame now is currently to lock frame
That frame;If the displacement between two two field pictures has reached the threshold value of regulation, then export the template frame currently carrying out correlation calculations, then present frame is turned
Change template frame lock into, carry out next round correlation calculations.The empirical value obtained is tested according to the threshold value of described regulation.A described RAM choosing
Select module 3 by the incoming correlation module of selected data 5 in present frame RAM and template frame RAM;Described selected data is a RAM choosing
Select module 3 and select to put into the frame image data of the incoming frame RAM in RAM storage array 6.
After each frame data have judged image detection, lock frame state control module 8 is started working.Lock frame state control module 8 is according to dependency mould
The lower limit gray scale thresholding LTH that offset data, the image sensing signal of image detection module offer and the statistics with histogram module that block calculates provides,
Draw and whether update lock frame and lock frame count.The condition updating lock frame is: if the image sensing signal that image detection module 9 provides is acquisition window
Do not collect effective acquisition target, update lock frame the most all the time, and reset lock frame count.If the image sensing signal that image detection module 9 provides
Collect effective acquisition target for acquisition window but acquisition target does not starts mobile, then judge the difference of two frame lower limit gray scale thresholdings before and after statistics with histogram
Whether value reaches threshold value;Or two frame data at acquisition target relative to the amount of offsetting on the main moving direction of acquisition window, i.e. correlation module calculate
Go out level or vertical offset is not zero;Maybe when locking the threshold value that frame count reaches regulation, it is determined that present frame forces to update lock frame, and resets lock frame meter
Number.Collect effective acquisition target if the image sensing signal that image detection module provides is acquisition window and starts mobile, and now horizontal and vertical
During the threshold component that side-play amount sum reduces more than the increase with lock frame count, update lock frame, and reset lock frame count.
It is illustrated in figure 2 each frame image data sequencing schematic diagram on collecting flowchart, the wherein frame number at interval between template frame and present frame
Depend on that locking frame judges, if current template frame and present frame meet release lock frame requirement, the frame number between rear pattern plate frame and the present frame of release lock frame
Being 0, if current template frame and present frame are unsatisfactory for release lock frame requirement, the frame number between rear pattern plate frame and the present frame of release lock frame is more than or equal to
1.Therefore template frame is probably the view data gathered a long time ago and is latched, and present frame carries out correlation calculations.
As it is shown on figure 3, under release lock frame condition, deposit template frame RAM and become depositing output frame RAM, the template frame that output is released;Deposit
Put present frame RAM to become depositing template frame RAM, become the comparison displacement template of correlation calculations next time;Deposit incoming frame RAM to become depositing
Put present frame RAM, and template frame carries out correlation calculations;Deposit output frame RAM to become depositing incoming frame RAM.This procedural order repeating query turns
Change.
Under lock frame condition, the RAM depositing present frame and the RAM depositing incoming frame exchanges.I.e. present frame RAM becomes incoming frame RAM,
I.e. abandon present frame, incoming frame is become present frame and does computing.
The mode of structure ASIC solving image deformation that the present invention provides achieves that high speed, low cost, amount of calculation be little and the structure of streamlined.
Utilize the image that the method is spliced, owing to arithmetic accuracy is high, error is little, can significantly correct image offset, image deformation, short image, image
The phenomenons such as hangover, lay a solid foundation for subsequent treatment;It is especially suitable for the popularization of fingerprint class commercial product.
Those skilled in the art is under conditions of the spirit and scope of the present invention determined without departing from claims, it is also possible to carry out above content
Various amendments.Therefore the scope of the present invention is not limited in above explanation, but determined by the scope of claims.
Claims (6)
1. solve the structure of image deformation, it is characterised in that include that two field picture provides module, data transmission module, a RAM
Module, the 2nd RAM is selected to select module, correlation module, RAM storage array, statistics with histogram module, lock frame shape
State control module, image detection module;
The image that offset data that described lock frame state control module calculates according to correlation module, image detection module provide
The lower limit gray scale threshold value that detection signal, statistics with histogram module provide, it is judged that this frame is the need of lock frame or release lock frame, defeated
Going out to lock control frame signal selects module and the 2nd RAM to select module to a RAM;
Described lock frame state control module judges that this frame the need of the foundation of lock frame or release lock frame is: if two two field pictures are gathering
Object does not reaches the threshold value of regulation relative to the displacement on the main moving direction of acquisition window, then abandon present frame, use next
Two field picture replaces present frame and template frame to carry out correlation calculations, and template frame is that frame currently locking frame;If between two two field pictures
Displacement reached regulation threshold value, then export the template frame currently carrying out correlation calculations, present frame be converted into template frame lock
Fixed, carry out next round correlation calculations;
A described RAM selects module by the selected data incoming dependency mould in present frame RAM and template frame RAM
Block;Described selected data is that a RAM selects module to select to put into the frame figure of the incoming frame RAM in RAM storage array
As data;
Described RAM storage array is made up of four or more RAM;Being assigned as of described RAM storage array:
At least currently carry out the current frame image data of correlation calculations with a RAM storage, be defined as present frame
RAM;At least currently carry out the template frame image data of correlation calculations with a RAM storage, be defined as template frame
RAM;The frame image data at least currently gathered with a RAM storage, is defined as incoming frame RAM;At least with one
The view data that individual RAM storage current release lock frame spreads out of, is defined as output frame RAM;Also comprise a RAM storage
Current relevance matrix, is defined as correlation matrix RAM;
The storage rule of described RAM storage array is:
Under release lock frame condition, deposit template frame RAM and become depositing output frame RAM, the template frame that output is released;
Deposit present frame RAM to become depositing template frame RAM, become the displacement template of the comparison of correlation calculations next time;Deposit
Incoming frame RAM becomes depositing present frame RAM, and template frame carries out correlation calculations;Deposit output frame RAM to become depositing
Put incoming frame RAM;
Under lock frame condition, the RAM depositing present frame and the RAM depositing incoming frame exchanges, and is become by present frame RAM
Incoming frame RAM, and abandon present frame, incoming frame is become present frame and does computing.
The structure of solution image deformation the most according to claim 1, it is characterised in that described two field picture provides module to carry
For each two field picture collected, and module is selected to select to put in RAM storage array by a RAM two field picture
Incoming frame RAM;
Described data transmission module transmission correlation data, offset data, two field picture;
A described RAM selects module to select to put into the incoming frame RAM in RAM storage array by the frame image data of reception
In, and by the incoming correlation module of selected data in present frame RAM and template frame RAM;Described 2nd RAM selects
Output frame RAM is read address, reading data and reading enable and processes and transfer by module;Described statistics with histogram module is asked in real time
Go out the original histogram of each frame image data, obtain its corresponding lower limit gray scale thresholding further according to rectangular histogram respectively;
The correlation data that described image detection module is calculated by frame image data and the correlation module of Real-time Collection judges
Acquisition target is relative to the state of acquisition window.
The structure of solution image deformation the most according to claim 1, it is characterised in that this frame of described judgement the need of
The foundation of lock frame or release lock frame also includes: if the image sensing signal that image detection module provides is that acquisition window has not collected
Effect acquisition target, updates lock frame the most all the time, and resets lock frame count;
If the image sensing signal that image detection module provides is that acquisition window collects effective acquisition target but acquisition target is not opened
Begin mobile, then judge whether the difference of two frame lower limit gray scale thresholdings reaches threshold value before and after statistics with histogram;Or two frame data adopting
Collection object is relative to the amount of offsetting on the main moving direction of acquisition window;Maybe when locking the threshold value that frame count reaches regulation, it is determined that
Present frame is forced to update lock frame, and resets lock frame count;
If the image sensing signal that image detection module provides is that acquisition window collects effective acquisition target and acquisition target starts
Mobile, and when now horizontal and vertical side-play amount sum is more than the threshold component reduced to lock the increase of frame count, update lock frame,
And reset lock frame count.
The structure of solution image deformation the most according to claim 1, it is characterised in that described module frame and present frame it
Being separated with frame number between, the frame number at its interval depends on that locking frame judges;If current template frame and present frame meet release lock frame bar part,
Frame number between rear pattern plate frame and the present frame of release lock frame is 0;If current template frame and present frame are unsatisfactory for release, lock frame is wanted
Asking, the frame number between rear pattern plate frame and the present frame of release lock frame is more than or equal to 1.
5. according to the structure solving image deformation described in claim 1 or 4, it is characterised in that described correlation module root
The template frame transmitted according to the RAM selection module received and present frame calculate the dependency of two two field pictures;If described template frame
Be with reference to that motionless frame in two two field pictures, the most described present frame be in two two field pictures with reference to template frame move up and down another
One two field picture;According to template frame and the coverage condition of present frame, choose the part or all of data of two two field picture overlay areas, will
Data one to one in two two field picture overlay areas ask squared difference sum to obtain the correlation operation under current coverage condition
Result;
Two two field pictures correlation operation result composition correlation matrix under all coverage conditions;Described correlation matrix is by being correlated with
Property module export the storage of RAM storage array and offset tracking block, and calculated two two field pictures by offset tracking block
Offset data.
The structure of solution image deformation the most according to claim 1, it is characterised in that described correlation data includes phase
The closing property maximum of matrix, minima and coordinate thereof;Described offset data includes the level participating in two two field pictures of computing and hangs down
Nogata to side-play amount, speed both horizontally and vertically, speed weight both horizontally and vertically.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210345140.3A CN103679623B (en) | 2012-09-18 | 2012-09-18 | Solve the structure of image deformation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210345140.3A CN103679623B (en) | 2012-09-18 | 2012-09-18 | Solve the structure of image deformation |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103679623A CN103679623A (en) | 2014-03-26 |
CN103679623B true CN103679623B (en) | 2016-09-28 |
Family
ID=50317082
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201210345140.3A Active CN103679623B (en) | 2012-09-18 | 2012-09-18 | Solve the structure of image deformation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103679623B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104883623B (en) * | 2015-04-30 | 2018-12-18 | 北京小鸟看看科技有限公司 | A kind of head-mounted display video control method and circuit |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101226589A (en) * | 2007-01-18 | 2008-07-23 | 中国科学院自动化研究所 | Method for detecting living body fingerprint based on thin plate spline deformation model |
US7609865B2 (en) * | 2004-11-08 | 2009-10-27 | Biomagnetics | 3D fingerprint and palm print data model and capture devices using multi structured lights and cameras |
CN101782965A (en) * | 2010-02-11 | 2010-07-21 | 上海点佰趣信息科技有限公司 | Method for treating deformed fingerprint image |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7809211B2 (en) * | 2005-11-17 | 2010-10-05 | Upek, Inc. | Image normalization for computed image construction |
-
2012
- 2012-09-18 CN CN201210345140.3A patent/CN103679623B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7609865B2 (en) * | 2004-11-08 | 2009-10-27 | Biomagnetics | 3D fingerprint and palm print data model and capture devices using multi structured lights and cameras |
CN101226589A (en) * | 2007-01-18 | 2008-07-23 | 中国科学院自动化研究所 | Method for detecting living body fingerprint based on thin plate spline deformation model |
CN101782965A (en) * | 2010-02-11 | 2010-07-21 | 上海点佰趣信息科技有限公司 | Method for treating deformed fingerprint image |
Non-Patent Citations (2)
Title |
---|
Estimating Fingerprint Deformation;Arun Ross等;《Proc. of International Conference on Biometric Authentication》;20040731;第3072卷;第249-255页 * |
利用去除形变的方法进行指纹匹配;毛克明等;《计算机科学与探索》;20090515;第3卷(第3期);第247-256页 * |
Also Published As
Publication number | Publication date |
---|---|
CN103679623A (en) | 2014-03-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106780620B (en) | Table tennis motion trail identification, positioning and tracking system and method | |
CN104282020B (en) | A kind of vehicle speed detection method based on target trajectory | |
CN108648161B (en) | Binocular vision obstacle detection system and method of asymmetric kernel convolution neural network | |
EP2426642B1 (en) | Method, device and system for motion detection | |
Jin et al. | Corn plant sensing using real‐time stereo vision | |
CN111563915B (en) | KCF target tracking method integrating motion information detection and Radon transformation | |
CN112837303A (en) | Defect detection method, device, equipment and medium for mold monitoring | |
CN104299248B (en) | Utilize method of the forward-looking sonar image to multiple dynamic object motion predictions under water | |
CN110110797B (en) | Water surface target training set automatic acquisition method based on multi-sensor fusion | |
CN101324958A (en) | Method and apparatus for tracking object | |
CN108320306A (en) | Merge the video target tracking method of TLD and KCF | |
CN108521554A (en) | Large scene multi-target cooperative tracking method, intelligent monitor system, traffic system | |
CN107967692A (en) | A kind of target following optimization method based on tracking study detection | |
CN107197121A (en) | A kind of electronic image stabilization method based on on-board equipment | |
CN114979489A (en) | Gyroscope-based heavy equipment production scene video monitoring and image stabilizing method and system | |
CN107423709A (en) | A kind of object detection method for merging visible ray and far infrared | |
CN113724297A (en) | Event camera-based tracking method | |
CN107016701B (en) | A kind of measurement method and device of corn kernel filling rate | |
CN108710879A (en) | A kind of pedestrian candidate region generation method based on Grid Clustering Algorithm | |
CN103679623B (en) | Solve the structure of image deformation | |
CN102163335B (en) | Multi-camera network structure parameter self-calibration method without inter-camera feature point matching | |
CN108447072A (en) | A kind of image partition method and system of crystal grain | |
CN104820972B (en) | A kind of infrared image ME noise remove methods based on in-orbit statistic of classification | |
Ma et al. | An improved kernel correlation filter for occlusion target tracking | |
CN116777956A (en) | Moving target screening method based on multi-scale track management |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20170420 Address after: 610041 high tech Zone, Chengdu province Tianfu Avenue North Section of the new century global center, No. 1700 E3-1-1603 Patentee after: Hefei formula Electronic Technology Co., Ltd. Address before: 610041 Sichuan, Chengdu high tech Tianfu Avenue South extension of the high-tech incubator Park, building A-E-3, No. 1 Patentee before: Chengdu Finchos Electron Co., Ltd. |