CN109151332A - Camera coding based on fitness function exposes optimal code word sequence search method - Google Patents
Camera coding based on fitness function exposes optimal code word sequence search method Download PDFInfo
- Publication number
- CN109151332A CN109151332A CN201810675614.8A CN201810675614A CN109151332A CN 109151332 A CN109151332 A CN 109151332A CN 201810675614 A CN201810675614 A CN 201810675614A CN 109151332 A CN109151332 A CN 109151332A
- Authority
- CN
- China
- Prior art keywords
- search
- fitness function
- exposure
- indicates
- codeword sequence
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/70—Circuitry for compensating brightness variation in the scene
- H04N23/73—Circuitry for compensating brightness variation in the scene by influencing the exposure time
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Genetics & Genomics (AREA)
- Signal Processing (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Physiology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Studio Devices (AREA)
- Image Analysis (AREA)
Abstract
Camera coding based on fitness function exposes optimal code word sequence search method, carries out global search using genetic algorithm, while introducing simulated annealing and carrying out locally optimal solution search；Construct fitness function H；The initiation parameter of set algorithm, including initial population size, individual amount, crossover probability, mutation probability, the number of iterations threshold value, simulated annealing initial temperature, equilibrium temperature, annealing ratio R atio and fitness function threshold value；When fitness function H meets threshold condition or the number of iterations i reaches the number of iterations threshold value of setting, then global search terminates, and finally obtained solution exposes optimal codeword sequence as coding.
Description
Technical field
The present invention relates to optical computing technical field of imaging, more particularly to the exposure of the coding of the camera based on fitness function is most
Excellent codeword sequence searching method.
Background technique
In moving target shooting imaging process, it is frequently encountered imageable target and imaging system within the camera exposure time
Relative motion occurs, causes imaging moving fuzzy.For conventional exposure mode, camera shutter in exposure is in opening state always
State, exposure process is equivalent to defines a box filter and image progress convolution in time-domain, and filter low pass is special
Property destroy detail of the high frequency important in scene so that motion blur image restoration become one height morbid state problem.
Coding exposure (Coded Exposure, CE) imaging technique is the novel exposure for calculating imaging field in recent years and proposing
Imaging pattern, core concept are within the camera exposure time according to the binary code sequence that is pre-designed rapidly on-off
Original box-packed filter, as high-frequency information, is become bandwidth filter with reserved graph by camera shutter.Not with conventional exposure mode
Together, the Fourier transformation frequency spectrum of the motion blur image that coding exposure image mode obtains, point spread function is free of zero point, makes
The problem of being recovered into a good state of image is obtained, so as to rapidly get a distinct image using Deconvolution Method.
Coding exposure key problem be how to choose and determine the camera shutter binary code word sequence of optimization so that
Image reserved high-frequency information as much as possible is obtained, guarantees the effect of liftering algorithm, this just needs fast and effeciently to select
Optimal codeword sequence.Existing codeword sequence choosing method often has ignored the signal-to-noise ratio factor of restored image, while code word is searched
The performance of rope algorithm can not be excellent, and algorithm search efficiency is lower, affects answering for coding exposure image technology to a certain extent
With and effect.
Summary of the invention
It is an object of the present invention to: for coding exposure problems, construct for searching for the optimal code word sequence of camera exposure
The fitness function of column provides the method for a kind of selection and the search optimal codeword sequence of camera exposure based on the fitness function.
The technical solution adopted by the present invention is that: the camera coding based on fitness function exposes optimal code word sequence search side
Method carries out global search using genetic algorithm, while introducing simulated annealing and carrying out locally optimal solution search；Construct fitness
Function H；The initiation parameter of set algorithm, including initial population size, individual amount, crossover probability, mutation probability, iteration time
Number threshold value, simulated annealing initial temperature, equilibrium temperature, annealing ratio R atio and fitness function threshold value；
When fitness function H meets threshold condition or the number of iterations i reaches the number of iterations threshold value of setting, then the overall situation is searched
Hitch beam, finally obtained solution expose optimal codeword sequence as coding；
Detailed process includes:
When the 1st progress global search, enables i=1, i indicate global iterative number, be complete with given random initial population
Office search population, before carrying out global search each time later, i adds 1, Population Regeneration and optimal solution；I-th global search carries out
When the 1st local search, enables ij=i1, ij indicate the jth time local search during i-th global search, carry out each time
Before local search, j adds 1；
Step 1: global search calculates population and inputs corresponding fitness function Hi′；By intersect and variation obtain it is next
The population result of secondary iteration；
Step 2: it introduces simulated annealing and carries out local search:
A random perturbation is carried out in the population result obtained after intersection and variation, generates a disturbance new explanation and is calculated
It corresponds to fitness function Hi″；
Calculate fitness function Hi' and fitness function Hi" difference, obtain
△Hi=Hi′-Hi″；
If △ Hi< 0 then receives the disturbance new explanation, as the locally optimal solution of this search, and searches in the overall situation next time
It is updated when rope；
If △ Hi>=0, then it is handled according to Metropolis criterion；
Step 3: judging fitness function Hi' whether meet threshold condition or whether global iterative number i reaches and initially set
Fixed the number of iterations threshold value exits global search if having one in two conditions when meeting, and search process terminates, final
The solution arrived exposes optimal codeword sequence as coding；Otherwise, step 1 and step 2 are repeated.
Further, it is described according to Metropolis criterion carry out processing refer to: select one at random from (0,1) section
Number R, calculates following inequality:
R≤exp(-△Hi/Tij) (7)
Wherein, TijIndicate the simulated annealing temperature of jth time local search in i-th global search；
If formula (7) is set up, receive the disturbance new explanation, as the locally optimal solution of this search, and it is complete next time
It is updated when office's search, end simulation annealing local search executes step 1；
If formula (7) is invalid, give up the disturbance new explanation, according to the annealing ratio R atio of setting, calculates simulation next time
Annealing temperature Ti(j+1)=Tij* Ratio, and step 2 is repeated, continue locally optimal solution search；
As next Simulated annealing Ti(j+1)Equal to equilibrium temperature T initially setfinalWhen, end simulation annealing part
Search executes step 1.
The invention also discloses the fitness function construction methods for searching for the optimal codeword sequence of camera exposure, comprising:
(1) for encoding exposure image style of shooting, definition exposure shutter codeword sequence:
In formula, Si(i=1,2 ..., n) be binary code word sequential digit values, n indicate codeword sequence length, Si=1 indicates
Camera shutter is opened, Si=0 indicates that camera shutter is closed；The number of " 1 " is r, the number of " 0 " in the codeword sequence expression formula
For t, there is n=r+t；
(2) discrete fourier variation operation is carried out to the exposure shutter codeword sequence, calculates the first interpretational criteria factor
C1, it is expressed as follows:
C1=min (| F (S) |)
Wherein, F () indicates discrete Fourier transform operation, and min () indicates calculated minimum operation；First evaluation is quasi-
Then factor C1Indicate the minimum value of codeword sequence point spread function amplitude；
(3) the second interpretational criteria factor C is calculated2:
Wherein, var () indicates to calculate variance operation；Second interpretational criteria factor C2Indicate codeword sequence auto-correlation journey
Degree indicates that the amplitude of codeword sequence point spread function changes severe degree；
(4) the third interpretational criteria factor C in the case where encoding exposure image style of shooting is calculated3:
Wherein,Indicate the mean intensity that image is obtained in the camera unit exposure time, T indicates time for exposure, Tr/n
Real exposure time under presentation code exposure mode；Indicate that deconvolution noise factor, A are to compile
Unit matrix under code exposure short exposure mode, is one-dimensional circular matrix；Indicate that the signal during camera imaging is non-
Continuous item noise contribution is fixed value；C indicates camera constant, and CTr/n indicates signal continuous item noise contribution, with exposure
Time increase and it is increased；Third interpretational criteria factor C3The motion blur image signal-to-noise ratio of presentation code exposure mode acquisition；
(5) the 4th interpretational criteria factor C is calculated4:
C4=mean (| F (S) |/(S+1))
4th interpretational criteria factor C4Indicate the low-frequency component of acquisition image；
(6) fitness function H is constructed；
Wherein, α1, α2, α3, α4Successively indicate the first evaluation points C1, the second evaluation points C2, third evaluation points C3,
Four evaluation points C4Weight coefficient.
The scene application that the present invention can be used for encoding the detection of exposure mode moving target and obtain, utilizes optimal codeword sequence
Coding exposure camera shutter control is carried out, helps more to retain image information using coding exposure image mode, preferably
Retain the high-frequency information in motion blur image, is conducive to promote restoration algorithm treatment effect, so that the processing knot of restoration algorithm
Fruit has more preferably clarity and visual effect.
Detailed description of the invention
Fig. 1 is the optimal code word sequence search method flow diagram based on cultural gene algorithm.
Fig. 2 is the Fourier transformation frequency domain amplitude figure for the optimal codeword sequence that cultural gene algorithm search obtains.
Fig. 3 is the optimal codeword sequence Fourier transformation frequency domain amplitude curve graph that genetic algorithm obtains.
Specific embodiment
Below by specific embodiment, and in conjunction with attached drawing, technical scheme of the present invention will be further explained in detail.
Camera coding based on fitness function exposes optimal code word sequence search method,
Global search is carried out using genetic algorithm, while introducing simulated annealing and carrying out locally optimal solution search；
Construct fitness function H；
The initiation parameter of set algorithm, including initial population size, crossover probability, mutation probability, global iterative number,
Simulated annealing initial temperature, equilibrium temperature and annealing ratio R atio；
When fitness function H meets threshold condition or the number of iterations i reaches the number of iterations threshold value of setting, then the overall situation is searched
Hitch beam, finally obtained solution expose optimal codeword sequence as coding.
The construction method of fitness function, comprising:
(1) for encoding exposure image style of shooting, definition exposure shutter codeword sequence:
In formula, Si(i=1,2 ..., n) be binary code word sequential digit values, n indicate codeword sequence length, Si=1 indicates
Camera shutter is opened, Si=0 indicates that camera shutter is closed；The number of " 1 " is r, the number of " 0 " in the codeword sequence expression formula
For t, there is n=r+t；
(2) discrete fourier variation operation is carried out to the exposure shutter codeword sequence, calculates the first interpretational criteria factor
C1, it is expressed as follows:
C1=min (| F (S) |)
Wherein, F () indicates discrete Fourier transform operation, and min () indicates calculated minimum operation；First evaluation is quasi-
Then factor C1Indicate the minimum value of codeword sequence point spread function amplitude；
(3) the second interpretational criteria factor C is calculated2:
Wherein, var () indicates to calculate variance operation；Second interpretational criteria factor C2Indicate codeword sequence auto-correlation journey
Degree indicates that the amplitude of codeword sequence point spread function changes severe degree；
(4) the third interpretational criteria factor C in the case where encoding exposure image style of shooting is calculated3:
Wherein,Indicate the mean intensity that image is obtained in the camera unit exposure time, T indicates time for exposure, Tr/n
Real exposure time under presentation code exposure mode；Indicate that deconvolution noise factor, A are to compile
Unit matrix under code exposure short exposure mode, is one-dimensional circular matrix；Indicate that the signal during camera imaging is non-
Continuous item noise contribution is fixed value；C indicates camera constant, and CTr/n indicates signal continuous item noise contribution, with exposure
Time increase and it is increased；Third interpretational criteria factor C3The motion blur image signal-to-noise ratio of presentation code exposure mode acquisition；
(5) the 4th interpretational criteria factor C is calculated4:
C4=mean (| F (S) |/(S+1))
4th interpretational criteria factor C4Indicate the low-frequency component of acquisition image；
(6) fitness function H is constructed；
Wherein, α1, α2, α3, α4Successively indicate the first evaluation points C1, the second evaluation points C2, third evaluation points C3,
Four evaluation points C4Weight coefficient.
As shown in Figure 1, enable i=1, i indicate global iterative number when the 1st progress global search, it is random first with what is given
Beginning population is global search population, and before carrying out global search each time later, i adds 1, Population Regeneration and optimal solution；I-th is complete
When office's search carries out the 1st local search, ij=i1, ij is enabled to indicate the jth time local search during i-th global search,
Before carrying out local search each time, j adds 1.
Search process:
Step 1: global search calculates population and inputs corresponding fitness function Hi′；By intersect and variation obtain it is next
The population result of secondary iteration；
Step 2: it introduces simulated annealing and carries out local search:
A random perturbation is carried out in the population result obtained after intersection and variation, generates a disturbance new explanation and is calculated
It corresponds to fitness function Hi″；
Calculate fitness function Hi' and fitness function Hi" difference, obtain
△Hi=Hi′-Hi″；
If △ Hi< 0 then receives the disturbance new explanation, as the locally optimal solution of this search, and searches in the overall situation next time
Optimal solution is updated when rope；
If △ Hi>=0, then it is handled according to Metropolis criterion:
It selects a several R at random from (0,1) section, calculates following inequality:
R≤exp(-△Hi/Tij) (7)
Wherein, TijIndicate the simulated annealing temperature of jth time local search in i-th global search；
If formula (7) is set up, receive the disturbance new explanation, as the locally optimal solution of this search, and it is complete next time
Optimal solution is updated when office's search, end simulation annealing local search executes step 1；
If formula (7) is invalid, give up the disturbance new explanation, according to the annealing ratio R atio of setting, calculates simulation next time
Annealing temperature Ti(j+1)=Tij* Ratio, and step 2 is repeated, continue locally optimal solution search；
As next Simulated annealing Ti(j+1)Equal to equilibrium temperature T initially setfinalWhen, end simulation annealing part
Search executes step 1.
Step 3: judging fitness function Hi' whether meet threshold condition or whether global iterative number i reaches and initially set
Fixed the number of iterations threshold value exits global search if having one in two conditions when meeting, and search process terminates, final
The optimal solution arrived exposes optimal codeword sequence as coding；Otherwise, step 1 and step 2 are repeated.
Illustrate algorithm implementation procedure and effect in conjunction with a specific embodiment.First evaluation points C1, second evaluation
Factor C2, third evaluation points C3, the 4th evaluation points C4Weight coefficient setting are as follows: α1=1, α2=1, α3=0.075, α4=
0.024.The value is empirical value, it is contemplated that each evaluation points are different to the weighing factor for choosing codeword sequence, the first evaluation because
Sub- C1, the second evaluation points C2It is even more important for the determination of optimal code word, weight coefficient α1And α2Value need to be greater than α3And α4
Numerical value, in practical application can near empirical value appropriate adjustment.Searching algorithm parameter initialization setting are as follows: Population Size is set as
50, individual amount (presentation code sequence length) is set as 52, and crossover probability is set as 0.85, and mutation probability is set as 0.01, and simulation is moved back
Fiery algorithm initial temperature is set as 100, and equilibrium temperature is set as 25, and annealing ratio R atio is set as 0.95, and the number of iterations threshold value is 50
It is secondary.The value range of fitness function H is [36,38], which is empirical value, and specific value can be according to photographed scene content
With restored image quality appropriate adjustment.The fitness function threshold value chosen in the present embodiment is 36.5.It is carried out according to above-mentioned steps
Optimal codeword sequence is chosen, and Fig. 2, the Fourier transformation frequency domain amplitude figure of optimal codeword sequence are obtained.
In order to which the performance of codeword sequence determined by the method for the present invention is better described, by the method for the present invention and heredity
The codeword sequence that searching algorithm obtains compares, the obtained optimal codeword sequence Fourier transformation frequency of genetic search algorithm
Domain amplitude curve figure is as shown in Figure 3.
The minimum value and variance yields of the optimal codeword sequence frequency domain amplitude curve that two methods obtain are calculated by matlab,
The results are shown in Table 1.The evaluation result data shown in the table 1 can see, and compare genetic searching method, the method for the present invention institute is really
The minimum value of fixed optimal codeword sequence frequency domain amplitude curve is bigger, and curve variance is smaller, is more advantageous under coding exposure mode
The frequency domain information for acquiring image retains.
1 coded sequence frequency domain amplitude curve evaluation comparison result of table
| Evaluation index | Amplitude curve minimum value | Amplitude curve variance yields |
| The method of the present invention | 0.89 | 4.73 |
| Genetic searching method | 0.09 | 6.04 |
Embodiment described above is a kind of preferable scheme of the invention, not makees limit in any form to the present invention
System, there are also other variants and remodeling on the premise of not exceeding the technical scheme recorded in the claims.
Claims (3)
1. the fitness function construction method for searching for the optimal codeword sequence of camera exposure characterized by comprising
(1) for encoding exposure image style of shooting, definition exposure shutter codeword sequence:
In formula, Si(i=1,2 ..., n) be binary code word sequential digit values, n indicate codeword sequence length, Si=1 indicates camera
Shutter opening, Si=0 indicates that camera shutter is closed；The number of " 1 " is r in the codeword sequence expression formula, and the number of " 0 " is t,
There is n=r+t；
(2) discrete fourier variation operation is carried out to the exposure shutter codeword sequence, calculates the first interpretational criteria factor C1:
C1=min (| F (S) |)
Wherein, F () indicates discrete Fourier transform operation, and min () indicates calculated minimum operation；First interpretational criteria because
Sub- C1Indicate the minimum value of codeword sequence point spread function amplitude；
(3) the second interpretational criteria factor C is calculated2:
Wherein, var () indicates to calculate variance operation；Second interpretational criteria factor C2It indicates codeword sequence auto-correlation degree, indicates
The amplitude of codeword sequence point spread function changes severe degree；
(4) the third interpretational criteria factor C in the case where encoding exposure image style of shooting is calculated3:
Wherein,Indicate the mean intensity that image is obtained in the camera unit exposure time；
T indicates time for exposure, the real exposure time under Tr/n presentation code exposure mode；
It indicates that deconvolution noise factor, A are the unit matrix under coding exposure short exposure mode, is one-dimensional circular matrix；
It indicates the irrelevant item noise contribution of signal during camera imaging, is fixed value；
C indicates that camera constant, CTr/n indicate signal continuous item noise contribution, increased with time for exposure increase；
Third interpretational criteria factor C3The motion blur image signal-to-noise ratio of presentation code exposure mode acquisition；
(5) the 4th interpretational criteria factor C is calculated4:
C4=mean (| F (S) |/(S+1))
4th interpretational criteria factor C4Indicate the low-frequency component of acquisition image；
(6) fitness function H is constructed；
Wherein, α1, α2, α3, α4Successively indicate the first evaluation points C1, the second evaluation points C2, third evaluation points C3, the 4th comment
Valence factor C4Weight coefficient.
2. the camera coding based on fitness function exposes optimal code word sequence search method, which is characterized in that calculated using heredity
Method carries out global search, while introducing simulated annealing and carrying out locally optimal solution search；Construct fitness function H；
The initiation parameter of set algorithm, including initial population size, individual amount, crossover probability, mutation probability, the number of iterations
Threshold value, simulated annealing initial temperature, equilibrium temperature, annealing ratio R atio and fitness function threshold value；
When fitness function H meets threshold condition or the number of iterations i reaches the number of iterations threshold value of setting, then global search knot
Beam, finally obtained solution expose optimal codeword sequence as coding；
Detailed process includes:
When the 1st progress global search, enables i=1, i indicate global iterative number, searched using given random initial population as the overall situation
Rope population, before carrying out global search every time later, i adds 1, Population Regeneration and optimal solution；
When i-th global search carries out the 1st local search, ij=i1, ij is enabled to indicate the jth during i-th global search
Secondary local search, before carrying out local search each time, j adds 1；
Step 1: global search calculates population and inputs corresponding fitness function Hi′；By intersecting and variation is changed next time
The population result in generation；
Step 2: it introduces simulated annealing and carries out local search:
A random perturbation is carried out in the population result obtained after intersection and variation, generates a disturbance new explanation and to calculate its right
Answer fitness function Hi″；
Calculate fitness function Hi' and fitness function Hi" difference, obtain
△Hi=Hi′-Hi″；
If △ Hi< 0 then receives the disturbance new explanation, as the locally optimal solution of this search, and more in global search next time
New optimal solution；
If △ Hi>=0, then it is handled according to Metropolis criterion；
Step 3: judging fitness function Hi' whether meet threshold condition or global iterative number i whether reach it is initially set repeatedly
For frequency threshold value, if having one in two conditions when meeting, exit global search, search process terminates, it is finally obtained most
Excellent solution exposes optimal codeword sequence as coding；Otherwise, step 1 and step 2 are repeated.
3. the camera coding based on fitness function exposes optimal code word sequence search method as claimed in claim 2, special
Sign is, described to carry out processing according to Metropolis criterion and refer to: a several R to be selected at random from (0,1) section, under calculating
State inequality:
R≤exp(-ΔHi/Tij) (7)
Wherein, TijIndicate the simulated annealing temperature of jth time local search in i-th global search；
If formula (7) is set up, receive the disturbance new explanation, as the locally optimal solution of this search, and is searched in the overall situation next time
Optimal solution is updated when rope, end simulation annealing local search executes step 1；
If formula (7) is invalid, give up the disturbance new explanation, according to the annealing ratio R atio of setting, calculates next simulated annealing
Temperature Ti(j+1)=Tij* Ratio, and step 2 is repeated, continue locally optimal solution search；
As next Simulated annealing Ti(j+1)Equal to equilibrium temperature T initially setfinalWhen, end simulation annealing local search,
Execute step 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810675614.8A CN109151332B (en) | 2018-06-26 | 2018-06-26 | Camera coding exposure optimal codeword sequence searching method based on fitness function |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810675614.8A CN109151332B (en) | 2018-06-26 | 2018-06-26 | Camera coding exposure optimal codeword sequence searching method based on fitness function |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109151332A true CN109151332A (en) | 2019-01-04 |
| CN109151332B CN109151332B (en) | 2020-07-28 |
Family
ID=64802321
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810675614.8A Active CN109151332B (en) | 2018-06-26 | 2018-06-26 | Camera coding exposure optimal codeword sequence searching method based on fitness function |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109151332B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110097509A (en) * | 2019-03-26 | 2019-08-06 | 杭州电子科技大学 | A kind of restored method of local motion blur image |
| CN112286211A (en) * | 2020-12-28 | 2021-01-29 | 上海交大智邦科技有限公司 | Environment modeling and AGV path planning method for irregular layout workshop |
| CN112989745A (en) * | 2021-02-20 | 2021-06-18 | 南京工程学院 | Thermal design method for optimal arrangement of PCB electronic components |
| CN113534042A (en) * | 2021-07-14 | 2021-10-22 | 深圳市微能信息科技有限公司 | TDOA (time difference of arrival) positioning algorithm based on improved reverse learning cuckoo search |
| CN117114144A (en) * | 2023-10-24 | 2023-11-24 | 青岛农业大学 | Rice salt and alkali resistance prediction method and system based on artificial intelligence |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070258706A1 (en) * | 2006-05-08 | 2007-11-08 | Ramesh Raskar | Method for deblurring images using optimized temporal coding patterns |
| US20100321509A1 (en) * | 2009-06-18 | 2010-12-23 | Canon Kabushiki Kaisha | Image processing apparatus and method thereof |
| CN103595923A (en) * | 2013-11-04 | 2014-02-19 | 西北核技术研究所 | Double-camera high-speed imaging system and method based on code exposure |
| CN105139394A (en) * | 2015-08-19 | 2015-12-09 | 杭州电子科技大学 | Noise image quality evaluation method combining reconstruction with noise scatter histograms |
-
2018
- 2018-06-26 CN CN201810675614.8A patent/CN109151332B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070258706A1 (en) * | 2006-05-08 | 2007-11-08 | Ramesh Raskar | Method for deblurring images using optimized temporal coding patterns |
| US20100321509A1 (en) * | 2009-06-18 | 2010-12-23 | Canon Kabushiki Kaisha | Image processing apparatus and method thereof |
| CN103595923A (en) * | 2013-11-04 | 2014-02-19 | 西北核技术研究所 | Double-camera high-speed imaging system and method based on code exposure |
| CN105139394A (en) * | 2015-08-19 | 2015-12-09 | 杭州电子科技大学 | Noise image quality evaluation method combining reconstruction with noise scatter histograms |
Non-Patent Citations (4)
| Title |
|---|
| JEON H G,LEE J Y, ET AL: "Generating Fluttering Patterns with Low Autocorrelation for Coded Exposure Imaging", 《INTERNATIONAL JOURNAL OF COMPUTER VISION》 * |
| LIRONG HE, GUANGMANG CU，ELTA: "The optimal code searching method with an improved criterion of coded exposure for remote sensing image restoration", 《OPTICS COMMUNICATIONS》 * |
| 何丽蓉: "基于编码曝光和颤振探测的遥感图像快速复原方法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 黄魁华: "信息集成中编码曝光技术优化研究", 《中国博士学位论文全文数据库 信息科技辑》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110097509A (en) * | 2019-03-26 | 2019-08-06 | 杭州电子科技大学 | A kind of restored method of local motion blur image |
| CN112286211A (en) * | 2020-12-28 | 2021-01-29 | 上海交大智邦科技有限公司 | Environment modeling and AGV path planning method for irregular layout workshop |
| CN112989745A (en) * | 2021-02-20 | 2021-06-18 | 南京工程学院 | Thermal design method for optimal arrangement of PCB electronic components |
| CN113534042A (en) * | 2021-07-14 | 2021-10-22 | 深圳市微能信息科技有限公司 | TDOA (time difference of arrival) positioning algorithm based on improved reverse learning cuckoo search |
| CN117114144A (en) * | 2023-10-24 | 2023-11-24 | 青岛农业大学 | Rice salt and alkali resistance prediction method and system based on artificial intelligence |
| CN117114144B (en) * | 2023-10-24 | 2024-01-26 | 青岛农业大学 | Rice salt and alkali resistance prediction method and system based on artificial intelligence |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109151332B (en) | 2020-07-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109151332A (en) | Camera coding based on fitness function exposes optimal code word sequence search method | |
| CN109408731B (en) | Multi-target recommendation method, multi-target recommendation model generation method and device | |
| Upadhyay et al. | A new design method based on firefly algorithm for IIR system identification problem | |
| CN108596243B (en) | Eye movement gaze prediction method based on hierarchical gaze view and conditional random field | |
| CN114385376B (en) | Client selection method for federal learning of lower edge side of heterogeneous data | |
| CN113312925A (en) | Remote sensing image text generation and optimization method based on self-reinforcement learning | |
| CN111694977A (en) | Vehicle image retrieval method based on data enhancement | |
| CN113887397A (en) | Classification method and classification system of electrophysiological signals based on ocean predator algorithm | |
| CN116822593A (en) | Large-scale pre-training language model compression method based on hardware perception | |
| Wang et al. | A Convolutional Neural Network Pruning Method Based On Attention Mechanism. | |
| KR102120150B1 (en) | Learning method and learning device for variational interference using neural network and test method and test device for variational interference using the same | |
| CN110851959A (en) | Wind speed interval prediction method integrating deep learning and quantile regression | |
| Kwasigroch et al. | Deep neural network architecture search using network morphism | |
| CN113780522A (en) | Tokamak plasma major rupture prediction algorithm based on deep neural network | |
| CN112487933A (en) | Radar waveform identification method and system based on automatic deep learning | |
| CN117371320A (en) | Photonic crystal fiber prediction method and system based on reinforced gray wolf optimization algorithm | |
| KR102110316B1 (en) | Method and device for variational interference using neural network | |
| CN116883364A (en) | Apple leaf disease identification method based on CNN and Transformer | |
| CN116977661A (en) | Data processing method, device, equipment, storage medium and program product | |
| CN116824140A (en) | Small sample segmentation method for test scene non-mask supervision | |
| CN109711460A (en) | The acquisition methods and device of initial cluster center | |
| CN112015854B (en) | Heterogeneous data attribute association method based on self-organizing mapping neural network | |
| CN113920159A (en) | Infrared aerial small target tracking method based on full convolution twin network | |
| Hu et al. | A new method of small target detection based on neural network | |
| Wong et al. | A Comparison of Six Convolutional Neural Networks for Weapon Categorization |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |