CN104715490B - Navel orange image segmenting method based on adaptive step size harmony search algorithm - Google Patents
Navel orange image segmenting method based on adaptive step size harmony search algorithm Download PDFInfo
- Publication number
- CN104715490B CN104715490B CN201510163411.7A CN201510163411A CN104715490B CN 104715490 B CN104715490 B CN 104715490B CN 201510163411 A CN201510163411 A CN 201510163411A CN 104715490 B CN104715490 B CN 104715490B
- Authority
- CN
- China
- Prior art keywords
- individuality
- goes
- individual
- image
- counter
- 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.)
- Expired - Fee Related
Links
Classifications
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30188—Vegetation; Agriculture
Abstract
The invention discloses a navel orange image segmenting method based on adaptive step size harmony search algorithm, and aims at solving the shortages of slow segmenting and low segmenting precision of the traditional harmony search algorithm applied to navel orange image segmenting. The method is characterized in that search factors are generated according to the chaotic motion rules during tone adjustment in a naval orange image division process adopting the harmony search algorithm; then the difference between the optimal individual and random individual is utilized to adaptively determine the search step size, so that the search performance of the algorithm can be improved; in addition, the information of the optimal individual and the individual in the adjacent area is utilized to generate a reverse individual in the adjacent area, and the reverse individual in the adjacent area is included into the selection operation, so as to speed up the convergence of the algorithm; compared with the similar methods, the method has the advantages that the naval orange can be quickly segmented, and the segmenting precision can be improved.
Description
Technical field
The present invention relates to image segmentation field, more particularly, to a kind of navel orange based on adaptive step harmonic search algorithm
Image partition method.
Background technology
With the development of navel orange industry, the intelligent robot that navel orange is plucked gradually is being applied to navel orange production field
In.Machine vision technique is one of core technology of navel orange picking robot, and navel orange image segmentation is navel orange harvesting machine
The key technology of the NI Vision Builder for Automated Inspection part of people.Navel orange image segmentation be exactly by the navel orange image segmentation of collection become it is several not
Same part, identifies the navel orange part in image, instructs so as to provide positioning for the harvesting of navel orange picking robot.Navel orange figure
As the quality of segmentation effect often directly influences whether the harvesting precision of navel orange picking robot.
Harmonic search algorithm is a kind of intelligent optimization algorithm for proposing in recent years, and its structure is very simple, but its performance
It is very potential.Harmonic search algorithm achieves successfully application in many engineering fields, and for example, Wang Ling etc. sent out in 2011
A kind of bright method for being optimized industrial wireless sensor network deployment using harmonic search algorithm, Li Yangyang etc. were invented in 2012
A kind of utilization harmonic search algorithm simultaneously merges the multi-objective community detection method of common adjacency matrix spectrum information, according to Yu Feng etc. 2012
Year proposes the image partition method based on harmonic search algorithm and cluster analysis.Although harmonic search algorithm is led in many engineerings
Successfully application is achieved in domain, but traditional harmony searching algorithm often has splitting speed when navel orange image is split slowly,
The not high shortcoming of segmentation precision.
The content of the invention
The present invention mainly solves the technical problem existing for prior art, and for traditional harmony searching algorithm navel is applied to
Splitting speed is often there is during orange image segmentation slow, the not high shortcoming of segmentation precision, propose it is a kind of based on adaptive step and
The navel orange image partition method of sound searching algorithm.The present invention can accelerate navel orange image segmentation speed, improve segmentation precision.
Technical scheme:A kind of navel orange image partition method based on adaptive step harmonic search algorithm, bag
Include following steps:
Step 1, using the width navel orange image OI of camera acquisition one, is then converted into the navel orange image OI of collection
The image YI of YCrCb color spaces;
Step 2, extracts the Cb color component values of each pixel in image YI as cluster data, by the cluster numbers extracted
Store in matrix D CB according to the ranks coordinate by pixel in image YI, each element and image in matrix D CB is thus obtained
Ranks coordinate one-to-one relationship in OI between each pixel, the wherein size of matrix D CB are H × W, and the value of H is equal to
The height of image YI, the value of W is equal to the width of image YI;
Step 3, user's initiation parameter, the initiation parameter includes segmentation class number D, harmony storehouse size
Popsize, data base learning rate HMCR, reverse Size of Neighborhood NK, maximum evaluates number of times MAX_FEs;
Step 4, current evolution algebraically t=0, Evaluation: Current number of times FEs=0;
Step 5, randomly generates initial harmony storehouseWherein:Subscript i=1 ...,
Popsize, andFor harmony storehouse PtIn i-th it is individual, its random initializtion formula is:
Wherein subscript j=1 ..., D, and D represents and to divide the image into into how many classifications;It is in harmony storehouse PtIn
I-th it is individual, store D segmentation classification cluster centre, rand (0, be 1) to obey equally distributed between [0,1]
Random real number produces function;
Step 6, calculates harmony storehouse PtIn each is individualAdaptive valueWherein subscript i=1 ..., Popsize,
Calculate individualAdaptive valueMethod be:Each element DCB first in calculating matrix DCBm,nRespectively with individualityIn
The distance of the cluster centre of D segmentation classification of storage, DCBm,nIt is minimum with the distance of which cluster centre, then make DCBm,nBelong to
Which class, then all elements DCB in calculating matrix DCBm,nIt is individual with belonging to itThe distance of the cluster centre of middle segmentation classification
Sum is used as individualityAdaptive value, wherein adaptive value is more little, shows individual more outstanding, and row matrix subscript m=1 ...,
H, row subscript n=1 ..., W;
Step 7, Evaluation: Current number of times FEs=FEs+Popsize, and preserve harmony storehouse PtThe minimum individuality of middle adaptive value is
Optimum individual Bestt, make tone regulation PAR=0.01;
Step 8, using adaptive step strategy a test individuality U is producedt, and calculate test individuality UtAdaptive value F
(Ut), comprise the following steps that:
Step 8.1, makes counter j=1;
Step 8.2, randomly generates a real number r1 between [0,1], if r1 is less than data base learning rate HMCR, turns
To step 8.3, step 8.13 is otherwise gone to;
Step 8.3, randomly generates positive integer RI1 between [1, D], and makes
Step 8.4, randomly generates a real number r2 between [0,1], if r2 is less than tone regulation PAR, goes to
Step 8.5, otherwise goes to step 8.14;
Step 8.5, randomly generates a real number r3 between [0,1], if r3 is equal to 0.25,0.50 or 0.75, then again
It is regenerated until r3 is not equal to 0.25,0.50 or 0.75;
Step 8.6, makes intermediate variable TCI=r3;
Step 8.7, makes counter Ct=1;
Step 8.8, by Logistic chaos formula the value of search factor CI is calculated:
CI=4.0 × TCI × (1.0-TCI);
Step 8.9, makes intermediate variable TCI=CI;
Step 8.10, makes counter Ct=Ct+1, if counter Ct is less than or equal to 350, goes to step 8.8, no
Then go to step 8.11;
Step 8.11, orderIfMore than 255 orIt is little
In 0, then makeOtherwise keepValue is constant;
Step 8.12, goes to step 8.14;
Step 8.13, order
Step 8.14, makes counter j=j+1, if counter j is less than or equal to D, goes to step 8.2, otherwise goes to
Step 8.15;
Step 8.15, calculates test individuality UtAdaptive value F (Ut), go to step 9;
Step 9, using neighborhood reverse strategy test individuality U is producedtThe reverse individuality NOU of neighborhoodt, and it is reverse to calculate neighborhood
Individual NOUtAdaptive value F (NOUt), comprise the following steps that:
Step 9.1, makes in current harmony storehouse and is designated as BestI under adaptive value optimum individual;
Step 9.2, makes counter j=1;
Step 9.3, if counter j is less than or equal to D, goes to step 9.4, otherwise goes to step 9.11;
Step 9.4, makes counter i=(BestI-NK+Popsize) %Popsize, wherein % represent that complementation is accorded with;
Step 9.5, makes neighborhood lower boundThe neighborhood upper bound
Step 9.6, makes counter i=(1+BestI-K+Popsize) %Popsize, makes counter mt=1;
Step 9.7, if mt is less than or equal to 2 × NK, goes to step 9.8, otherwise goes to step 9.10;
Step 9.8, makes neighborhood lower boundAnd make the neighborhood upper bound
To take minimum value function, max is to take max function to wherein min;
Step 9.9, makes counter i=(i+1) %Popsize, and makes counter mt=mt+1, goes to step 9.7;
Step 9.10, makes counter j=j+1, goes to step 9.3;
Step 9.11, makes counter j=1, and reverse factor of n r=rand of neighborhood (0,1);
Step 9.12, if counter j is less than or equal to D, goes to step 9.13, otherwise goes to step 9.15;
Step 9.13, order
Step 9.14, makes counter j=j+1, goes to step 9.12;
Step 9.15, calculates neighborhood reversely individuality NOUtAdaptive value, go to step 10;
Step 10, makes BUtTo test individuality UtAnd its reverse individuality NOU of neighborhoodtAdaptive value reckling between the two;
Step 11, to make and be designated as WorstI under the worst individuality of adaptive value in current harmony storehouse;
Step 12, the as follows worst individuality in current harmony storehouseWith individualitySelect between the two
Go out more excellent individual into harmony storehouse of future generation:
Step 13, Evaluation: Current number of times FEs=FEs+2 preserves harmony storehouse PtThe minimum individuality of middle adaptive value is optimum
Body Bestt;
Step 14, tone regulation
Step 15, current evolution algebraically t=t+1;
Step 16, repeat step 8 to step 15 is until Evaluation: Current number of times FEs reaches end, implementation procedure after MAX_FEs
In the optimum individual Best that obtainstFor the cluster centre of D segmentation classification, the cluster centre of classification is split using D for obtaining
All elements in matrix D CB are classified, after the classification of all elements in DCB is determined, in recycling matrix D CB
Ranks coordinate one-to-one relationship in each element and image OI between each pixel, is carried out to each pixel in image OI
Classification, that is, obtain the result of final segmentation.
The invention has the advantages that:The present invention is splitting the tone adjustment process of navel orange image using harmonic search algorithm
It is middle that search factor is produced according to chaotic motion rule, and adaptively determined using the difference between optimum individual and random individual
Step-size in search, so as to strengthen the search performance of algorithm;In addition, it is anti-to produce neighborhood using the individual information of optimum individual and its neighborhood
To individuality, and the reverse individuality of neighborhood is fused to during selection operation, to accelerate convergence of algorithm speed;With congenic method phase
Than the present invention can accelerate the splitting speed of navel orange image, improve segmentation precision.
Description of the drawings
Fig. 1 is navel orange image to be split.
Fig. 2 is with navel orange image after present invention segmentation.
Specific embodiment
Below by embodiment, and accompanying drawing is combined, technical scheme is described in further detail.
The present embodiment is based on navel orange image Fig. 1 to be split, and the specific implementation step of the present invention is as follows:
Step 1, using the width navel orange image OI of camera acquisition one as shown in figure 1, and then turning the navel orange image OI of collection
Change the image YI for YCrCb color spaces into;
Step 2, extracts the Cb color component values of each pixel in image YI as cluster data, by the cluster numbers extracted
Store in matrix D CB according to the ranks coordinate by pixel in image YI, each element and image in matrix D CB is thus obtained
Ranks coordinate one-to-one relationship in OI between each pixel, the wherein size of matrix D CB are H × W, and the value of H is equal to
The height of image YI, the value of W is equal to the width of image YI;
Step 3, user's initiation parameter, the initiation parameter includes segmentation class number D, harmony storehouse size
Popsize=20, data base learning rate HMCR=0.9, reverse Size of Neighborhood NK=3, maximum evaluates number of times MAX_FEs=200;
Step 4, current evolution algebraically t=0, Evaluation: Current number of times FEs=0;
Step 5, randomly generates initial harmony storehouseWherein:Subscript i=1 ...,
Popsize, andFor harmony storehouse PtIn i-th it is individual, its random initializtion formula is:
Wherein subscript j=1 ..., D, and D represents and to divide the image into into how many classifications;It is in harmony storehouse PtIn
I-th it is individual, store D segmentation classification cluster centre, rand (0, be 1) to obey equally distributed between [0,1]
Random real number produces function;
Step 6, calculates harmony storehouse PtIn each is individualAdaptive valueWherein subscript i=1 ..., Popsize,
Calculate individualAdaptive valueMethod be:Each element DCB first in calculating matrix DCBm,nRespectively with individualityIn
The distance of the cluster centre of D segmentation classification of storage, DCBm,nIt is minimum with the distance of which cluster centre, then make DCBm,nBelong to
Which class, then all elements DCB in calculating matrix DCBm,nIt is individual with belonging to itThe distance of the cluster centre of middle segmentation classification
Sum is used as individualityAdaptive value, wherein adaptive value is more little, shows individual more outstanding, and row matrix subscript m=1 ...,
H, row subscript n=1 ..., W;
Step 7, Evaluation: Current number of times FEs=FEs+Popsize, and preserve harmony storehouse PtThe minimum individuality of middle adaptive value is
Optimum individual Bestt, make tone regulation PAR=0.01;
Step 8, using adaptive step strategy the test individuality U of is producedt, and calculate test individuality UtAdaptive value F
(Ut), comprise the following steps that:
Step 8.1, makes counter j=1;
Step 8.2, randomly generates a real number r1 between [0,1], if r1 is less than data base learning rate HMCR, turns
To step 8.3, step 8.13 is otherwise gone to;
Step 8.3, randomly generates positive integer RI1 between [1, D], and makes
Step 8.4, randomly generates a real number r2 between [0,1], if r2 is less than tone regulation PAR, goes to
Step 8.5, otherwise goes to step 8.14;
Step 8.5, randomly generates a real number r3 between [0,1], if r3 is equal to 0.25,0.50 or 0.75, then again
It is regenerated until r3 is not equal to 0.25,0.50 or 0.75;
Step 8.6, makes intermediate variable TCI=r3;
Step 8.7, makes counter Ct=1;
Step 8.8, by Logistic chaos formula the value of search factor CI is calculated:
CI=4.0 × TCI × (1.0-TCI);
Step 8.9, makes intermediate variable TCI=CI;
Step 8.10, makes counter Ct=Ct+1, if counter Ct is less than or equal to 350, goes to step 8.8, no
Then go to step 8.11;
Step 8.11, orderIfMore than 255 orIt is little
In 0, then makeOtherwise keepValue is constant;
Step 8.12, goes to step 8.14;
Step 8.13, order
Step 8.14, makes counter j=j+1, if counter j is less than or equal to D, goes to step 8.2, otherwise goes to
Step 8.15;
Step 8.15, calculates test individuality UtAdaptive value F (Ut), go to step 9;
Step 9, using neighborhood reverse strategy test individuality U is producedtThe reverse individuality NOU of neighborhoodt, and it is reverse to calculate neighborhood
Individual NOUtAdaptive value F (NOUt), comprise the following steps that:
Step 9.1, makes in current harmony storehouse and is designated as BestI under adaptive value optimum individual;
Step 9.2, makes counter j=1;
Step 9.3, if counter j is less than or equal to D, goes to step 9.4, otherwise goes to step 9.11;
Step 9.4, makes counter i=(BestI-NK+Popsize) %Popsize, wherein % represent that complementation is accorded with;
Step 9.5, makes neighborhood lower boundThe neighborhood upper bound
Step 9.6, makes counter i=(1+BestI-K+Popsize) %Popsize, makes counter mt=1;
Step 9.7, if mt is less than or equal to 2 × NK, goes to step 9.8, otherwise goes to step 9.10;
Step 9.8, makes neighborhood lower boundAnd make the neighborhood upper bound
To take minimum value function, max is to take max function to wherein min;
Step 9.9, makes counter i=(i+1) %Popsize, and makes counter mt=mt+1, goes to step 9.7;
Step 9.10, makes counter j=j+1, goes to step 9.3;
Step 9.11, makes counter j=1, and reverse factor of n r=rand of neighborhood (0,1);
Step 9.12, if counter j is less than or equal to D, goes to step 9.13, otherwise goes to step 9.15;
Step 9.13, order
Step 9.14, makes counter j=j+1, goes to step 9.12;
Step 9.15, calculates neighborhood reversely individuality NOUtAdaptive value, go to step 10;
Step 10, makes BUtTo test individuality UtAnd its reverse individuality NOU of neighborhoodtAdaptive value reckling between the two;
Step 11, to make and be designated as WorstI under the worst individuality of adaptive value in current harmony storehouse;
Step 12, the as follows worst individuality in current harmony storehouseWith individualitySelect between the two
Go out more excellent individual into harmony storehouse of future generation:
Step 13, Evaluation: Current number of times FEs=FEs+2 preserves harmony storehouse PtThe minimum individuality of middle adaptive value is optimum
Body Bestt;
Step 14, tone regulation
Step 15, current evolution algebraically t=t+1;
Step 16, repeat step 8 to step 15 is until Evaluation: Current number of times FEs reaches end, implementation procedure after MAX_FEs
In the optimum individual Best that obtainstFor the cluster centre of D segmentation classification, the cluster centre of classification is split using D for obtaining
All elements in matrix D CB are classified, after the classification of all elements in DCB is determined, in recycling matrix D CB
Ranks coordinate one-to-one relationship in each element and image OI between each pixel, is carried out to each pixel in image OI
Classification, that is, obtain the result that gathered navel orange image is finally split.
Specific embodiment described herein is only explanation for example spiritual to the present invention.Technology neck belonging to of the invention
The technical staff in domain can be made various modifications to described specific embodiment or supplement or replaced using similar mode
Generation, but without departing from the spiritual of the present invention or surmount scope defined in appended claims.
Claims (1)
1. a kind of navel orange image partition method based on adaptive step harmonic search algorithm, is characterized in that:Comprise the following steps:
Step 1, using the width navel orange image OI of camera acquisition one, is then converted into YCrCb colors by the navel orange image OI of collection
The image YI of color space;
Step 2, extracts the Cb color component values of each pixel in image YI as cluster data, by the cluster data for extracting by
Ranks coordinate of the pixel in image YI is stored in matrix D CB, is thus obtained in matrix D CB in each element and image OI
Ranks coordinate one-to-one relationship between each pixel, the wherein size of matrix D CB are H × W, and the value of H is equal to image
The height of YI, the value of W is equal to the width of image YI;
Step 3, user's initiation parameter, the initiation parameter includes segmentation class number D, harmony storehouse size Popsize, note
Recall storehouse learning rate HMCR, reverse Size of Neighborhood NK, maximum evaluates number of times MAX_FEs;
Step 4, current evolution algebraically t=0, Evaluation: Current number of times FEs=0;
Step 5, randomly generates initial harmony storehouseWherein:Subscript i=1 ...,
Popsize, andFor harmony storehouse PtIn i-th it is individual, its random initializtion formula is:
Wherein subscript j=1 ..., D, and D represents and to divide the image into into how many classifications;It is in harmony storehouse PtIn i-th
Individuality, store D segmentation classification cluster centre, rand (0, be 1) that equally distributed random reality is obeyed between [0,1]
Number produces function;
Step 6, calculates harmony storehouse PtIn each is individualAdaptive valueWherein subscript i=1 ..., Popsize, calculate
It is individualAdaptive valueMethod be:Each element DCB first in calculating matrix DCBm,nRespectively with individualityIn deposit
The distance of the cluster centre of D segmentation classification of storage, DCBm,nIt is minimum with the distance of which cluster centre, then make DCBm,nWhich belongs to
Individual class, then all elements DCB in calculating matrix DCBm,nIt is individual with belonging to itIt is middle segmentation classification cluster centre distance it
With as individualityAdaptive value, wherein adaptive value is more little, shows individual more outstanding, and row matrix subscript m=1 ..., H,
Row subscript n=1 ..., W;
Step 7, Evaluation: Current number of times FEs=FEs+Popsize, and preserve harmony storehouse PtThe minimum individuality of middle adaptive value is optimum
Individual Bestt, make tone regulation PAR=0.01;
Step 8, using adaptive step strategy a test individuality U is producedt, and calculate test individuality UtAdaptive value F (Ut),
Comprise the following steps that:
Step 8.1, makes counter j=1;
Step 8.2, randomly generates a real number r1 between [0,1], if r1 is less than data base learning rate HMCR, goes to step
Rapid 8.3, otherwise go to step 8.13;
Step 8.3, randomly generates positive integer RI1 between [1, D], and makes
Step 8.4, randomly generates a real number r2 between [0,1], if r2 is less than tone regulation PAR, goes to step
8.5, otherwise go to step 8.14;
Step 8.5, randomly generates a real number r3 between [0,1], if r3 is equal to 0.25,0.50 or 0.75, then again again
It is produced until r3 is not equal to 0.25,0.50 or 0.75;
Step 8.6, makes intermediate variable TCI=r3;
Step 8.7, makes counter Ct=1;
Step 8.8, by Logistic chaos formula the value of search factor CI is calculated:
CI=4.0 × TCI × (1.0-TCI);
Step 8.9, makes intermediate variable TCI=CI;
Step 8.10, makes counter Ct=Ct+1, if counter Ct is less than or equal to 350, goes to step 8.8, otherwise turns
To step 8.11;
Step 8.11, orderIfMore than 255 orLess than 0,
Then makeOtherwise keepValue is constant;
Step 8.12, goes to step 8.14;
Step 8.13, order
Step 8.14, makes counter j=j+1, if counter j is less than or equal to D, goes to step 8.2, otherwise goes to step
8.15;
Step 8.15, calculates test individuality UtAdaptive value F (Ut), go to step 9;
Step 9, using neighborhood reverse strategy test individuality U is producedtThe reverse individuality NOU of neighborhoodt, and it is reversely individual to calculate neighborhood
NOUtAdaptive value F (NOUt), comprise the following steps that:
Step 9.1, makes in current harmony storehouse and is designated as BestI under adaptive value optimum individual;
Step 9.2, makes counter j=1;
Step 9.3, if counter j is less than or equal to D, goes to step 9.4, otherwise goes to step 9.11;
Step 9.4, makes counter i=(BestI-NK+Popsize) %Popsize, wherein % represent that complementation is accorded with;
Step 9.5, makes neighborhood lower boundThe neighborhood upper bound
Step 9.6, makes counter i=(1+BestI-K+Popsize) %Popsize, makes counter mt=1;
Step 9.7, if mt is less than or equal to 2 × NK, goes to step 9.8, otherwise goes to step 9.10;
Step 9.8, makes neighborhood lower boundAnd make the neighborhood upper boundWherein
To take minimum value function, max is to take max function to min;
Step 9.9, makes counter i=(i+1) %Popsize, and makes counter mt=mt+1, goes to step 9.7;
Step 9.10, makes counter j=j+1, goes to step 9.3;
Step 9.11, makes counter j=1, and reverse factor of n r=rand of neighborhood (0,1);
Step 9.12, if counter j is less than or equal to D, goes to step 9.13, otherwise goes to step 9.15;
Step 9.13, order
Step 9.14, makes counter j=j+1, goes to step 9.12;
Step 9.15, calculates neighborhood reversely individuality NOUtAdaptive value, go to step 10;
Step 10, makes BUtTo test individuality UtAnd its reverse individuality NOU of neighborhoodtAdaptive value reckling between the two;
Step 11, to make and be designated as WorstI under the worst individuality of adaptive value in current harmony storehouse;
Step 12, the as follows worst individuality in current harmony storehouseWith individualitySelect between the two more
It is excellent individual into harmony storehouse of future generation:
Step 13, Evaluation: Current number of times FEs=FEs+2 preserves harmony storehouse PtThe minimum individuality of middle adaptive value is optimum individual
Bestt;
Step 14, tone regulation
Step 15, current evolution algebraically t=t+1;
Step 16, repeat step 8 to step 15 reaches up to Evaluation: Current number of times FEs and terminate after MAX_FEs, in implementation procedure
The optimum individual Best for arrivingtFor the cluster centre of D segmentation classification, the cluster centre of classification is split to square using D for obtaining
All elements in battle array DCB are classified, and after the classification of all elements in DCB is determined, recycle each in matrix D CB
Ranks coordinate one-to-one relationship in element and image OI between each pixel, is carried out point to each pixel in image OI
Class, that is, obtain the result of final segmentation.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510163411.7A CN104715490B (en) | 2015-04-09 | 2015-04-09 | Navel orange image segmenting method based on adaptive step size harmony search algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510163411.7A CN104715490B (en) | 2015-04-09 | 2015-04-09 | Navel orange image segmenting method based on adaptive step size harmony search algorithm |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104715490A CN104715490A (en) | 2015-06-17 |
CN104715490B true CN104715490B (en) | 2017-05-03 |
Family
ID=53414784
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510163411.7A Expired - Fee Related CN104715490B (en) | 2015-04-09 | 2015-04-09 | Navel orange image segmenting method based on adaptive step size harmony search algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104715490B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107300609A (en) * | 2017-07-15 | 2017-10-27 | 江西理工大学 | The water body chlorophyll alpha flexible measurement method searched for based on adaptability harmony |
Families Citing this family (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105069225B (en) * | 2015-08-07 | 2017-09-29 | 江西理工大学 | Using the OPTIMIZATION DESIGN OF VIBRATING SCREEN method of gaussian sum sonar surveillance system rope |
CN105046712B (en) * | 2015-08-07 | 2017-06-30 | 江西理工大学 | Based on the circle detection method that adaptability difference of Gaussian develops |
CN106023162B (en) * | 2016-05-12 | 2018-06-05 | 江西理工大学 | Strawberry image dividing method based on Integrated Strategy harmonic search algorithm |
CN105930685B (en) * | 2016-06-27 | 2018-05-15 | 江西理工大学 | The rare-earth mining area underground water ammonia nitrogen concentration Forecasting Methodology of Gauss artificial bee colony optimization |
CN106570873B (en) * | 2016-11-08 | 2019-10-25 | 江苏大学 | A kind of medical image cutting method |
CN106898009B (en) * | 2017-03-18 | 2019-05-14 | 江西理工大学 | Improve the multi-threshold image segmentation method of reversed harmony chess game optimization |
CN107145942B (en) * | 2017-05-24 | 2020-04-10 | 江西理工大学 | Soft measurement method for nitrate nitrogen of farmland water body in harmony search optimization |
CN107194935B (en) * | 2017-05-24 | 2020-06-05 | 江西理工大学 | Navel orange image segmentation method based on reverse harmony search |
CN107392911A (en) * | 2017-07-15 | 2017-11-24 | 江西理工大学 | The shaddock image partition method of application adaptability gaussian sum sonar surveillance system rope |
CN111061871B (en) * | 2019-11-26 | 2022-02-22 | 广东奥博信息产业股份有限公司 | Method for analyzing tendency of government and enterprise service text |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101661613A (en) * | 2009-08-27 | 2010-03-03 | 北京交通大学 | Image restoration method based on image segmentation, and system therefor |
CN104317721A (en) * | 2014-11-12 | 2015-01-28 | 大连交通大学 | Regression test case selection method based on improved harmony search algorithm |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7272258B2 (en) * | 2003-01-29 | 2007-09-18 | Ricoh Co., Ltd. | Reformatting documents using document analysis information |
-
2015
- 2015-04-09 CN CN201510163411.7A patent/CN104715490B/en not_active Expired - Fee Related
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101661613A (en) * | 2009-08-27 | 2010-03-03 | 北京交通大学 | Image restoration method based on image segmentation, and system therefor |
CN104317721A (en) * | 2014-11-12 | 2015-01-28 | 大连交通大学 | Regression test case selection method based on improved harmony search algorithm |
Non-Patent Citations (1)
Title |
---|
和声搜索算法改进与应用;周雅兰,黄韬;《计算机科学》;20140630;52-56,75 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107300609A (en) * | 2017-07-15 | 2017-10-27 | 江西理工大学 | The water body chlorophyll alpha flexible measurement method searched for based on adaptability harmony |
Also Published As
Publication number | Publication date |
---|---|
CN104715490A (en) | 2015-06-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104715490B (en) | Navel orange image segmenting method based on adaptive step size harmony search algorithm | |
CN106023162B (en) | Strawberry image dividing method based on Integrated Strategy harmonic search algorithm | |
CN106960214A (en) | Object identification method based on image | |
CN103366180A (en) | Cell image segmentation method based on automatic feature learning | |
CN108304826A (en) | Facial expression recognizing method based on convolutional neural networks | |
CN105574215A (en) | Instance-level image search method based on multiple layers of feature representations | |
CN107784657A (en) | A kind of unmanned aerial vehicle remote sensing image partition method based on color space classification | |
CN105678278A (en) | Scene recognition method based on single-hidden-layer neural network | |
CN106570192A (en) | Deep learning-based multi-view image retrieval method | |
CN106022392A (en) | Deep neural network sample automatic accepting and rejecting training method | |
CN104809737B (en) | Grapefruit image segmentation method based on double-strategy harmony search algorithm | |
CN112819096B (en) | Construction method of fossil image classification model based on composite convolutional neural network | |
CN109886267A (en) | A kind of soft image conspicuousness detection method based on optimal feature selection | |
CN104504007A (en) | Method and system for acquiring similarity degree of images | |
CN102170516B (en) | Color space transition method based on fuzzy logic and neural network | |
CN104361135A (en) | Image search method | |
CN104050680A (en) | Image segmentation method based on iteration self-organization and multi-agent inheritance clustering algorithm | |
CN103700118B (en) | Based on the moving target detection method of pulse coupled neural network | |
CN109523015A (en) | Image processing method in a kind of neural network | |
CN104463122A (en) | Seal recognition method based on PCNN | |
CN104778683B (en) | A kind of multi-modality images dividing method based on Functional Mapping | |
CN109948653A (en) | A kind of image similarity judgment method, electronic equipment and storage medium | |
Cai et al. | Cloud classification of satellite image based on convolutional neural networks | |
CN109636741A (en) | A kind of image denoising processing method | |
CN108428234A (en) | Interactive Segmentation performance optimization method based on image segmentation result evaluation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
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: 20170503 |
|
CF01 | Termination of patent right due to non-payment of annual fee |