CN101794393B - Target identification method of remote sensing image of artificial immune network based on self-adaptive PSO (Particle Swarm Optimization) - Google Patents
Target identification method of remote sensing image of artificial immune network based on self-adaptive PSO (Particle Swarm Optimization) Download PDFInfo
- Publication number
- CN101794393B CN101794393B CN2010100135373A CN201010013537A CN101794393B CN 101794393 B CN101794393 B CN 101794393B CN 2010100135373 A CN2010100135373 A CN 2010100135373A CN 201010013537 A CN201010013537 A CN 201010013537A CN 101794393 B CN101794393 B CN 101794393B
- Authority
- CN
- China
- Prior art keywords
- immunocyte
- affinity
- network
- remote sensing
- immune
- 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
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a target identification method of a remote sensing image of an artificial immune network based on a self-adaptive PSO(Particle Swarm Optimization), mainly overcoming the disadvantages of low target identification precision and low convergence speed in the traditional method. The identification method comprises the following steps of: firstly, extracting 7 invariant moment characteristics of an image target and carrying out normalization treatment on the characteristic data; secondly, setting running parameters, selecting a training sample and initializing an immune network and immune cells; thirdly, calculating the affinity degree of the immune cells and cloning the immune cells; fourthly, executing hyper-mutation operation based on the self-adaptive PSO; fifthly, selecting an immune cell with highest affinity degree and adding the immune cell into the immune network; sixthly, carrying out network inhibition operation; seventhly, judging a stop condition, turning to the eighth step eight if the condition is satisfied, and otherwise, and otherwise jumping to the third step; and eighthly, inputting characteristic values of the remote sensing images which are not used as training samples into the immune network, and judging a category attribute value of each image by the immune network. The method has the advantages of high target identification accuracy and stable target identification performance and can be used for solving the problem of target identification of a remote sensing image set.
Description
Technical field
The invention belongs to technical field of image processing, relate to the remote sensing images target identification method, this method can be used for the detection and the identification problem of remote sensing images collection.
Background technology
The remote sensing and the information processing technology thereof have played key effect in the antagonism of high-tech such as Target Recognition location, real-time follow-up, early warning, electronic countermeasure.Remote sensing image classification identification is a branch of the remote sensing images information processing technology; The Classification and Identification process is exactly that image pixel is grouped into the process of going in some type; The notion of here " class " can be certain atural object in the image; The different conditions of landforms or identical atural object, in case target assigned to some type just can be more accurately and analyze its specific nature easily.
Utilize remote sensing images to classify to the forest reserves, and, to the minerogentic condition of survey region with look for the ore deposit to have certain directive function; In marine application, can extract the shallow sea landform, carry out debating of vessel and know and classification, in addition; Can be used for the water resource investigation, have comparatively application prospects.
Obtain relative proven technique with remote sensing images and compare, the research to remote sensing images intelligence perception and decipher at present is in the primary stage.Be used to solve the remote sensing images Target Recognition at present and mainly contain two kinds of methods.
First kind of remote sensing images target identification method that is based on supporting vector machine.This method is at first extracted 7 dimension invariant moment features of remote sensing images, and the selected part training sample is input to the supporting vector machine training from each classification then, at last to residue sample discriminator.The deficiency that this method exists is that partial parameters is very big to the identifying influence, and this limited in one's ability based on statistical supporting vector machine algorithm process challenge, causes the Target Recognition result undesirable.
Second kind of remote sensing images target identification method that is based on genetic algorithm.This method is at first extracted 7 dimension invariant features of remote sensing images, and the selected part training sample is input to the genetic algorithm training from each classification then, at last to residue sample discriminator.The deficiency of this method is that the selection of initial population is very big to the recognition result influence, and the ability of searching optimum of this genetic algorithm is limited, and it is not very desirable causing recognition result.
Summary of the invention
The objective of the invention is to overcome the deficiency of above-mentioned prior art; A kind of remote sensing images target identification method of the artificial immune network based on self-adaptation PSO is proposed; It utilizes the global convergence fast of artificial immune network algorithm and population global optimization approach, designs a kind of supervised classification method that is used to improve remote sensing images Target Recognition ability.
Technical scheme of the present invention is to regard remote sensing images collection Target Recognition problem as different classes of classification of Data problem.The best cluster centre of using the search of artificial immune network algorithm computation is as network node.Utilization is based on the quick global convergence of the artificial immune network algorithm of self-adaptation PSO, the optimum solution of search problem.Comprise that specifically process is following:
(1) extracts 7 invariant moment features of image object, and these characteristics are done normalization handle;
(2) set operational factor, select training sample, initialization immunological network and immunocyte;
(3), calculate the affinity degree of all immunocytes, clone's immunocyte with the method for statistics immunocyte to the correct recognition rata of training sample;
(4) immune cell population behind the clone is carried out following hypermutation exclusive-OR function based on self-adaptation PSO:
What 4a) immunocyte is set respectively ties up initial velocity:
V(i,j)=rand*pm*|MC(i,j)-train(k,j)|*(MC(i,j)-train(k,j))
Wherein, p
mBe the variation probability, MC (i, j) the j dimension element of i immunocyte of expression, the j dimension element of the training sample of train (k, j)) expression picked at random,
4b) carry out mutation operation, the immunocyte after obtaining making a variation based on self-adaptation PSO according to the initial velocity that is provided with:
MC′(i,j)=MC(i,j)+W(i)*V(i,j)+r
1*rand*exp(-f(i))+r
2*rand*(gBest(j)-MC(i,j))
Wherein, the affinity value of i immunocyte of f (i) expression, gBest is a global extremum, it representes the highest immunocyte of affinity in the contemporary population, r
1Be local learning parameter, r
2Be overall learning parameter, their value sum is 1, and w is self-adaptation inertia weights,
Wherein, w
Max=0.9, w
Min=0.4, f representes the affinity value of current immunocyte, f
AvgThe mean value of representing all immunocyte affinity, f
MinThe minimum value of representing all immunocyte affinity;
(5) with the method for statistics immunocyte to the correct recognition rata of training sample; The affinity degree of the immune cell population after the calculating hypermutation is different; And to this hypermutation the immune cell population after different is carried out the Immune Clone Selection operation, select the highest immunocyte of affinity and add in the immunological network;
(6) it is minimum and fail to improve the immunocyte of immunological network to the training sample correct recognition rata to select in the immunological network part affinity value, and lets their death;
(7) from immunological network, select the highest immunocyte of affinity value, and the affinity value of judging it there is no change in iteration 10 times, if do not change then continue to carry out (8), otherwise, jumped to for (3) step;
(8) will be not as the eigenwert input immunological network of the remote sensing images of training sample, immunological network is judged the category attribute value of every width of cloth image.
The present invention is because after having extracted 7 dimension invariant moment features of remote sensing images; Made subsequent treatment; Make this characteristic have rotation, translation and flexible unchangeability, thereby make the remote sensing images of the different rotary angle of same classification, different translation yardstick and different zoom size have similar eigenwert; The present invention simultaneously uses and seeks best cluster centre based on the artificial immune network algorithm of self-adaptation PSO; Because this algorithm has speed of convergence and powerful ability of searching optimum fast, thereby has obtained recognition speed and remote sensing images Target Recognition effect faster preferably.
Description of drawings
Fig. 1 is a remote sensing images Target Recognition procedure chart of the present invention;
Fig. 2 is the employed classification and Detection remote sensing images of treating of emulation experiment of the present invention.
Embodiment
With reference to Fig. 1, the present invention includes following process:
Process 1 is extracted 7 invariant moment features of image object, and these characteristics is done normalization handle.
1.1) have rotation and translation invariance for making eigenwert, press the invariant moment features of following formulas Extraction image object:
M
1=(u
20+u
02)
M
3=(u
30-3u
12)
2+(3u
21-u
03)
2
M
4=(u
30+u
12)
2+(u
21+u
03)
2
M
5=(u
30+u
12)(u
30-3u
12)[(u
30+u
12)
2-3(u
21+u
03)
2]
+(3u
21-u
03)(u
21+u
03)[3(u
30+u
12)
2-(u
21+u
03)
2]
M
6=(u
20-u
02)[(u
30+u
12)
2-(u
21+u
03)
2]+4u
11(u
30+u
12)(u
21+u
03)
M
7=(3u
21-u
03)(u
30+u
12)[(u
30+u
12)
2-3(u
21+u
03)
2]
-(u
30-3u
12)(u
21+u
03)[3(u
30+u
12)
2-(u
21+u
03)
2]
Wherein, M
1~M
7Represent 7 invariant moment features respectively, u
Pq(p+q) central moment of presentation video, p and q are integer, p, q ∈ [0,3], the value of central moment does
The horizontal ordinate of x remarked pixel wherein, the ordinate of y remarked pixel, the mean value of all pixel horizontal ordinates in the x presentation video, y representes the mean value of all ordinates, f (x, y) expression is positioned at (x, y) gray-scale value of locational pixel;
1.2) have the yardstick unchangeability for making eigenwert, invariant moment features is handled by following formula, but be not limited to these formula:
M′
1=M
1/P;
M′
2=M
2/r
2;
M′
3=M
3/r
3;
M′
4=M
4/r
2;
M′
5=M
5/r
6;
M′
6=M
6/r
4;
M′
7=M
7/r
6;
Wherein, M '
1~M '
7Be the eigenwert after handling, P is the number of pixels that image comprised, r=(u
20+ u
02);
1.3) characteristic of extracting is done the normalization processing.
Process 2 is set operational factor, selects training sample, initialization immunological network and immunocyte.The operational factor of setting comprises scale num, the variation Probability p of initial immunocyte
m, clone scale pclone, data dimension N and class categories count C; Training sample is picked at random from the sample of each classification; Initial immune cell population is with B (k) expression, and wherein, k is the algebraically of evolution, k=0, and initial immunological network is empty.
Process 3 is calculated the affinity of immunocyte, and is cloned.According to formula
Calculate the affinity value of immunocyte, wherein, B
iRepresent i immunocyte, K is the number of all categories training sample, and C is the classification number of classification, Pr
IjExpression B
iCorrect identification number to j class training sample; Grand scale pclone and affinity f (B
i) size is directly proportional.
Process 4 is carried out the hypermutation exclusive-OR function based on self-adaptation PSO.Immune cell population behind the clone is carried out following hypermutation exclusive-OR function based on self-adaptation PSO:
What 4a) immunocyte is set respectively ties up initial velocity:
V(i,j)=rand*pm*|MC(i,j)-train(k,j)|*(MC(i,j)-train(k,j))
Wherein, p
mBe the variation probability, MC (i, j) the j dimension element of i immunocyte of expression, train (k, j) the j dimension element of the training sample of expression picked at random,
4b) carry out mutation operation, the immunocyte after obtaining making a variation based on self-adaptation PSO according to the initial velocity that is provided with:
MC′(i,j)=MC(i,j)+W(i)*V(i,j)+r
1*rand*exp(-f(i))+r
2*rand*(gBest(j)-MC(i,j))
Wherein, the affinity value of i immunocyte of f (i) expression, gBest is a global extremum, it representes the highest immunocyte of affinity in the contemporary population, r
1Be local learning parameter, r
2Be overall learning parameter, their value sum is 1, and w is self-adaptation inertia weights,
Wherein, w
Max=0.9, w
Min=0.4, f representes the affinity value of current immunocyte, f
AvgThe mean value of representing all immunocyte affinity, f
MinThe minimum value of representing all immunocyte affinity;
Process 5 is selected the highest immunocyte of affinity and is added immunological network to.According to formula
Calculate the affinity degree value of the immune cell population of hypermutation after different, and the immune cell population after different is carried out the Immune Clone Selection operation to this hypermutation, selects the highest immunocyte of affinity and adds in the immunological network;
Process 6, network suppresses operation.It is minimum and fail to improve the immunocyte of immunological network to the training sample correct recognition rata to select in the immunological network part affinity value, and lets their death;
Process 7 is judged stop condition.From immunological network, select the highest immunocyte of affinity value, and the affinity value of judging it there is no change in iteration 10 times, if do not change then the immunological network training is accomplished, otherwise, continue the training immunological network.
Through the training of above process 2~7 realizations to immunological network.
Process 8 with not as the eigenwert input immunological network of the remote sensing images of training sample, is judged the category attribute value of every width of cloth image by immunological network.
Effect of the present invention can further specify through following emulation experiment:
1. simulated conditions:
In order to verify superiority based on the artificial immune network Classifying Method in Remote Sensing Image of self-adaptation PSO; We with its with based on the Classifying Method in Remote Sensing Image of artificial immune network, and compare through emulation experiment based on the artificial immune network Classifying Method in Remote Sensing Image of PSO.In the emulation experiment algorithm parameter be provided with just the same: variation Probability p m=0.5; Clone's scale pclone=10; Maximum iteration time is 100, stop condition for when the affinity of optimum individual be 1 or the affinity of optimum individual do not change with interior in 10 generations.
At CPU is to use MATLAB to carry out emulation in core2 2.4HZ, internal memory 2G, the WINDOWS XP system.
2. emulation content:
Actual measurement remote sensing images in the selection accompanying drawing 2 are as test pattern, and every width of cloth image only comprises target and background.The entire image collection comprises all types of target different rotary angle, different scale and incomplete image totally 1064 width of cloth, aircraft class 608 width of cloth wherein, naval vessel class 456 width of cloth.Through range estimation, wherein the aircraft class is divided into 9 types, and the naval vessel class is divided into 4 types; Training sample is chosen as: aircraft class 160, naval vessel class are 120, and remaining is a test sample book.
3. analysis of simulation result:
Table 1 has provided the Target Recognition result of each method, and wherein, AINC representes the remote sensing images target identification method based on artificial immune network, and PSOAINC representes the artificial immune network remote sensing images target identification method based on PSO.
Three kinds of distinct methods of table 1 are to the identification of targets result
Target identification method | Target Recognition precision/% | Standard variance | Working time |
AINC | 91.2 | 14.2 | 1.22 |
PSOAINC | 91.5 | 14.8 | 1.04 |
The inventive method | 92.7 | 7.7 | 1 |
Can find out that from table 1 the inventive method all is optimum on all 3 evaluation indexes (Target Recognition precision, standard variance and working time), wherein working time be with working time of the present invention as a unit interval.Therefore, have higher Target Recognition precision, more stable target identification performance and speed of convergence faster based on the artificial immune network remote sensing images target identification method of self-adaptation PSO.
Claims (2)
1. artificial immune network remote sensing images target identification method based on self-adaptation PSO comprises following process:
(1) extract 7 invariant moment features of image object, and these characteristics done normalization handle:
1a), press the invariant moment features of following formulas Extraction image object for making eigenwert have rotation and translation invariance:
M
1=(u
20+u
02)
M
3=(u
30-3u
12)
2+(3u
21-u
03)
2
M
4=(u
30+u
12)
2+(u
21+u
03)
2
M
5=(u
30+u
12)(u
30-3u
12)[(u
30+u
12)
2-3(u
21+u
03)
2]
+(3u
21-u
03)(u
21+u
03)[3(u
30+u
12)
2-(u
21+u
03)
2]
M
6=(u
20-u
02)[(u
30+u
12)
2-(u
21+u
03)
2]+4u
11(u
30+u
12)(u
21+u
03)
M
7=(3u
21-u
03)(u
30+u
12)[(u
30+u
12)
2-3(u
21+u
03)
2]
-(u
30-3u
12)(u
21+u
03)[3(u
30+u
12)
2-(u
21+u
03)
2]
Wherein, M
1~M
7Represent 7 invariant moment features respectively, u
Pq(p+q) central moment of presentation video, p and q are integer, p, q ∈ [0,3], the value of central moment does
The horizontal ordinate of x remarked pixel wherein, the ordinate of y remarked pixel,
The mean value of all pixel horizontal ordinates in the presentation video,
The mean value of representing all ordinates, and f (x, y) expression is positioned at (x, y) gray-scale value of locational pixel;
1b), invariant moment features is handled by following formula for making eigenwert have the yardstick unchangeability:
M′
1=M
1/P;
M′
2=M
2/r
2;
M′
3=M
3/r
3;
M′
4=M
4/r
2;
M′
5=M
5/r
6;
M′
6=M
6/r
4;
M′
7=M
7/r
6;
Wherein, M '
1~M '
7Be the eigenwert after handling, P is the number of pixels that image comprised, r=(u
20+ u
02);
(2) set operational factor, select training sample, initialization immunological network and immunocyte, described operational factor comprises that the dimension N and the class categories of the scale num of initial immunocyte, the Probability p that makes a variation m, clone's scale pclone, data counted C; Training sample is picked at random from the sample of each classification; Initial immune cell population is with B (k) expression, and wherein, k is the algebraically of evolution, k=0, and initial immunological network is empty;
(3) through the correct recognition rata of statistics immunocyte to training sample, calculate the affinity of all immunocytes, clone's immunocyte,
The affinity of said all immunocytes of calculating is according to formula
Calculate, wherein, B
iRepresent i immunocyte, K is the number of all categories training sample, and C is the classification number of classification, P
RijExpression B
iCorrect identification number to j class training sample;
(4) immune cell population behind the clone is carried out following hypermutation exclusive-OR function based on self-adaptation PSO:
What 4a) immunocyte is set respectively ties up initial velocity:
V(i,j)=rand*p
m*|MC(i,j)-train(k,j)|*(MC(i,j)-train(k,j))
Wherein, p
mBe the variation probability, MC (i, j) the j dimension element of i immunocyte of expression, train (k, j) the j dimension element of the training sample of expression picked at random,
4b) carry out mutation operation, the immunocyte after obtaining making a variation based on self-adaptation PSO according to the initial velocity that is provided with:
MC′(i,j)=MC(i,j)+w(i)*V(i,j)+r
1*rand*exp(-f(i))+r
2*rand*(gBest(j)-MC(i,j))
Wherein, the affinity value of i immunocyte of f (i) expression, gBest is a global extremum, it representes the highest immunocyte of affinity in the contemporary population, r
1Be local learning parameter, r
2Be overall learning parameter, their value sum is 1, and w is self-adaptation inertia weights,
Wherein, w
Max=0.9, w
Min=0.4, f representes the affinity value of current immunocyte, f
AvgThe mean value of representing all immunocyte affinity, f
MinThe minimum value of representing all immunocyte affinity;
(5) through the correct recognition rata of statistics immunocyte to training sample; The affinity of the immune cell population after the calculating hypermutation is different; And to this hypermutation the immune cell population after different is carried out the Immune Clone Selection operation, select the highest immunocyte of affinity and add in the immunological network;
(6) it is minimum and fail to improve the immunocyte of immunological network to the training sample correct recognition rata to select in the immunological network part affinity value, and lets their death;
(7) from immunological network, select the highest immunocyte of affinity value, and the affinity value of judging it there is no change in iteration 10 times, if do not change then continue to carry out (8), otherwise, jumped to for (3) step;
(8) will be not as the eigenwert input immunological network of the remote sensing images of training sample, immunological network is judged the category attribute value of every width of cloth image.
2. the artificial immune network remote sensing images target identification method based on self-adaptation PSO according to claim 1 is wherein cloned scale pclone and affinity f (B
i) size is directly proportional.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2010100135373A CN101794393B (en) | 2010-01-04 | 2010-01-04 | Target identification method of remote sensing image of artificial immune network based on self-adaptive PSO (Particle Swarm Optimization) |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2010100135373A CN101794393B (en) | 2010-01-04 | 2010-01-04 | Target identification method of remote sensing image of artificial immune network based on self-adaptive PSO (Particle Swarm Optimization) |
Publications (2)
Publication Number | Publication Date |
---|---|
CN101794393A CN101794393A (en) | 2010-08-04 |
CN101794393B true CN101794393B (en) | 2012-02-08 |
Family
ID=42587073
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2010100135373A Expired - Fee Related CN101794393B (en) | 2010-01-04 | 2010-01-04 | Target identification method of remote sensing image of artificial immune network based on self-adaptive PSO (Particle Swarm Optimization) |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN101794393B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107066833A (en) * | 2016-11-25 | 2017-08-18 | 安徽奥里奥克科技股份有限公司 | Immunological network categorizing system and method with cell differentiation is oriented to based on fictitious force |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101478534A (en) * | 2008-12-02 | 2009-07-08 | 广东海洋大学 | Network exception detecting method based on artificial immunity principle |
-
2010
- 2010-01-04 CN CN2010100135373A patent/CN101794393B/en not_active Expired - Fee Related
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101478534A (en) * | 2008-12-02 | 2009-07-08 | 广东海洋大学 | Network exception detecting method based on artificial immunity principle |
Non-Patent Citations (5)
Title |
---|
Ai-ling Chen, et al..An Effective Hybrid Optimization Algorithm based on Self-adaptive Particle Swarm Optimization Algorithm and Artificial Immune Clone Algorithm.《Fourth International Conference on Natural Computation》.2008,第129-132页. * |
Bo Liu, et al..Improved particle swarm optimization combined with chaos.《Chaos, Solitons and Fractals》.2005,第25卷第1261–1271页. * |
JP特开2006-134267A 2006.05.25 |
Ruochen Liu,et al..A new artificial immune network classifier for SAR image.《Proc. of SPIE,MIPPR 2009:Pattern Recognition and Computer Vision》.2009,第7496卷第74960W-1至8页. * |
徐俊.免疫智能在优化中研究及其在数字图像中应用.《中国优秀硕士学位论文全文数据库》.2009,(第03期),全文. * |
Also Published As
Publication number | Publication date |
---|---|
CN101794393A (en) | 2010-08-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhao et al. | ApLeaf: An efficient android-based plant leaf identification system | |
CN101551809B (en) | Search method of SAR images classified based on Gauss hybrid model | |
CN105844283B (en) | Method, image search method and the device of image classification ownership for identification | |
CN103810704B (en) | Based on support vector machine and the SAR image change detection of discriminative random fields | |
Zhang et al. | A GANs-based deep learning framework for automatic subsurface object recognition from ground penetrating radar data | |
CN106354735A (en) | Image target searching method and device | |
CN106844739B (en) | Remote sensing image change information retrieval method based on neural network collaborative training | |
CN103824093B (en) | It is a kind of based on KFDA and SVM SAR image target's feature-extraction and recognition methods | |
CN104063719A (en) | Method and device for pedestrian detection based on depth convolutional network | |
CN104537647A (en) | Target detection method and device | |
CN104392228A (en) | Unmanned aerial vehicle image target class detection method based on conditional random field model | |
CN105320764B (en) | A kind of 3D model retrieval method and its retrieval device based on the slow feature of increment | |
CN109753874A (en) | A kind of low slow small classification of radar targets method based on machine learning | |
CN105160351B (en) | Semi-supervised hyperspectral classification method based on anchor point sparse graph | |
CN110472572A (en) | The quick identification and classification method of naval target under a kind of complex environment | |
CN106203483A (en) | A kind of zero sample image sorting technique of multi-modal mapping method of being correlated with based on semanteme | |
CN111311702B (en) | Image generation and identification module and method based on BlockGAN | |
CN104392459A (en) | Infrared image segmentation method based on improved FCM (fuzzy C-means) and mean drift | |
CN102930291B (en) | Automatic K adjacent local search heredity clustering method for graphic image | |
CN104966075A (en) | Face recognition method and system based on two-dimensional discriminant features | |
CN106570537A (en) | Random forest model selection method based on confusion matrix | |
CN107480627B (en) | Behavior recognition method and device, storage medium and processor | |
Ahmad et al. | Hybrid dense network with attention mechanism for hyperspectral image classification | |
CN104463210B (en) | Classification of Polarimetric SAR Image method based on object-oriented and spectral clustering | |
CN112084860A (en) | Target object detection method and device and thermal power plant detection method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | 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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120208 Termination date: 20180104 |
|
CF01 | Termination of patent right due to non-payment of annual fee |