CN110853054B - Method for improving intuitive fuzzy clustering and extracting infrared ship by using position information - Google Patents
Method for improving intuitive fuzzy clustering and extracting infrared ship by using position information Download PDFInfo
- Publication number
- CN110853054B CN110853054B CN201911035552.5A CN201911035552A CN110853054B CN 110853054 B CN110853054 B CN 110853054B CN 201911035552 A CN201911035552 A CN 201911035552A CN 110853054 B CN110853054 B CN 110853054B
- Authority
- CN
- China
- Prior art keywords
- membership
- degree
- image
- clustering
- pixel point
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 50
- 230000011218 segmentation Effects 0.000 claims abstract description 27
- 230000000877 morphologic effect Effects 0.000 claims abstract description 7
- 238000001914 filtration Methods 0.000 claims abstract description 6
- 238000003709 image segmentation Methods 0.000 claims description 24
- 230000008569 process Effects 0.000 claims description 15
- 238000004364 calculation method Methods 0.000 claims description 12
- 239000011159 matrix material Substances 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 2
- 238000009795 derivation Methods 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 8
- 238000011156 evaluation Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 206010047571 Visual impairment Diseases 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000036039 immunity Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10048—Infrared image
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a method for extracting infrared ships by improving intuitive fuzzy clustering by using position information, which comprises the following steps: the method comprises the following steps: the original image is preprocessed by adopting morphological operation top-hat filtering, and the influence of uneven gray level in the infrared ship image on a clustering segmentation result is reduced. Step two: and clustering and segmenting the infrared image by adopting an improved intuitive fuzzy clustering algorithm. Step three: and D, performing subsequent processing on the clustering segmentation image obtained in the step two. The improved intuitive fuzzy clustering algorithm provided by the invention adds the position information of the segmented target obtained by using the iterative algorithm, and improves the effect of the intuitive fuzzy C mean value. Meanwhile, a regular term of the similarity coefficient of the evaluation area is added into the target function, and the local spatial information is fully considered. The improved algorithm can better segment the infrared ship image and has wide market prospect and application value.
Description
[ technical field ] A method for producing a semiconductor device
The invention relates to a method for extracting infrared ships by improving intuitive fuzzy clustering by using position information, in particular to a method for extracting infrared ships with uneven gray levels by improving intuitive fuzzy clustering by using position information.
[ background of the invention ]
The image segmentation is to divide an original image into a plurality of continuous, non-overlapping and meaningful areas according to the difference of the characteristics of the gray scale, the contour, the texture, the color and the like of different areas of the image. Image segmentation, as a key technology in the field of image processing and computer vision, has extremely wide applications such as medical tissue measurement, pedestrian detection, vehicle tracking, and the like. The good image segmentation result not only retains important information in the image, but also effectively reduces useless data in the image, and has very important significance. The current mainstream image segmentation methods are mainly classified into the following four categories: threshold-based segmentation methods, region-based segmentation methods, edge-based segmentation methods, and segmentation methods based on specific theories.
The infrared image reflects the temperature difference between the target and the background, has the advantages of all-weather work, capability of overcoming visual disturbance and detecting the target, long action distance, strong anti-interference capability and the like, and is widely applied to the military and civil fields. However, the infrared image has the phenomena of low pixel resolution, poor contrast, strong transitivity, blurred image edge and the like, which greatly affect the infrared image segmentation, and a general image segmentation algorithm is difficult to obtain a good segmentation result.
Aiming at the characteristics of high uncertainty, strong transitivity and the like of the infrared image, a fuzzy clustering algorithm can be adopted for image segmentation. In the fuzzy clustering algorithm, data points belong to multiple classes with different membership degrees, and more data information is reserved. The most classical Fuzzy clustering algorithm is Fuzzy C-means algorithm (FCM), Fuzzy C-means algorithm (see document J.C. Dengen. a Fuzzy image processing method related to ISODATA algorithm and Its application for Detecting Compact easy separation clustering. Proc. 1973,3(3):32-57 (J.C. Dunn.A Fuzzy relevance of the ISODATA Process and Its Use in Detecting Compact computer Well-Separated Cluster [ J ]. Journal of Cybernetics,1973 (3):32-57.) proposed by J.C. Dengen as a data clustering algorithm based on target function optimization. However, the traditional fuzzy C-means algorithm does not consider spatial information, and cannot well segment noisy images. In response to this problem, many researchers have introduced spatial information to improve the fuzzy C-means algorithm. Alhamander et al proposed FCM _ S algorithm by introducing regularization terms to account for neighborhood information (see, Muhammeri N. Alhamider, Samacheh Amani, Negium Muhammeri et al. an improved fuzzy C-means algorithm for offset field estimation and its application to MRI image segmentation. IEEE medical imaging, Vol.21. 193. 199,2002.(M.Ahmed, S.Yamany, N.Mohamed, A.Farag, and T.Moriary, "A modified fuzzy C-means algorithm for bias field estimation and segmentation of MRI data," IEEE transactions, Med.Imag., vol.21, pp.193-199,2002.); on the basis of FCM _ S, mean filtering and median filtering ideas are introduced into an algorithm by the aid of the relaxation and the tensor, FCM _ S1 and FCM _ S2 algorithms are provided, and anti-noise performance of a traditional fuzzy C mean value is improved (see the literature: relaxation and tensor. a stable kernel function-based fuzzy C mean value image segmentation algorithm combined with spatial information; IEEE system control processing journal, 34, 1907. 1916,2004. (S.Chen and D.Zhang, "Robust image segmentation FCM with spatial constraints based on new kernel-induced distance measure," IEEE Trans. Syst, Man, Cybern., vol.34, pp.1907-1916,2004.); straiios and Wasieris propose an FLICM algorithm which introduces adaptive neighborhood influence factors, avoids the problem of parameter selection, and has better robustness (see the literature: Straiios, Wasieris. A stable improved fuzzy C-means algorithm combining local spatial information. American institute of Electrical and electronics Engineers, image processing Association, 19, 1328-1337, 5.2010 (S.Krinidis and V.Chatzis, A robust fuzzy local information C-means clustering algorithm, IEEE trans.Imag.Process, 133vol.19, No.5, pp.1328-7, May 2010.); xuze water et al, which combines the intuitionistic fuzzy set theory into the FCM algorithm, proposed intuitionistic fuzzy C-means, which retained more image information and achieved good results (see: xuze water, wujie. intuitionistic fuzzy C-means algorithm. system engineering and electronics, volume 21, 580-phone 590, 2010, 8 months, (Zeshui Xu, Junjie wu. intuitionistic fuzzy C-means clustering algorithms, j.syst.eng.electron, vol.21, pp.580-590, Aug 2010)), and so on.
The above improved algorithms achieve some improvement in noise immunity and segmentation results, but still have some problems. The problems of low pixel resolution, poor contrast, blurred image edge and the like in the infrared image can seriously affect the segmentation effect of the infrared ship. Although the above algorithms all introduce neighborhood information in different ways, the position information of the segmentation target is ignored. In order to make full use of neighborhood information and position information, the invention provides a method for extracting infrared ships with uneven gray levels by improving intuitive fuzzy clustering by using the position information.
[ summary of the invention ]
The fuzzy clustering algorithm is widely applied to the field of image segmentation. However, the algorithm is sensitive to noise because neighborhood information is not considered, so that an ideal clustering segmentation result cannot be obtained for a noisy image. The infrared image is widely applied in the fields of military affairs and civilian use, but due to the fact that the infrared image has the phenomena of low pixel resolution, poor contrast, strong transition, fuzzy image edges, serious gray scale unevenness and the like, a good segmentation result cannot be obtained by applying a traditional fuzzy clustering algorithm.
The invention provides a method for extracting infrared ships by improving intuitive fuzzy clustering by utilizing position information, aiming at the problems of target segmentation and extraction in infrared images. On one hand, the method utilizes position information of a segmentation target to construct a weight coefficient, and different weight coefficients are given to different classes of the same pixel point; on the other hand, a distance regular term based on region information is added into the target function, and the regular distance term reflects the similarity between the regional gray scale and the clustering center. The improved algorithm of the invention fully considers the position information and the local area information of the segmented target, and the infrared image target segmentation effect is obviously improved.
The invention adopts the following technical scheme:
first, the image is preprocessed by morphological operations. And then, clustering and segmenting the infrared image by using an improved intuitive fuzzy clustering algorithm. And finally, rejecting the non-target area through subsequent processing. The improved algorithm fully considers the position information and the local space information of the segmentation target, and the segmentation quality is obviously improved.
The technical scheme of the invention comprises the following specific steps:
the method comprises the following steps: the original image is preprocessed by adopting morphological operation top-hat filtering, and the influence of uneven gray level in the infrared ship image on a clustering segmentation result is reduced.
Step two: and clustering and segmenting the infrared image by adopting an improved intuitive fuzzy clustering algorithm.
The objective function of the improved intuitive fuzzy clustering algorithm is as follows:
wherein i is the category ordinal number, c is the category total number, k is the position coordinate of the pixel point, N is the pixel point total number, W ik Is a position weight coefficient, u ik Is the membership degree of the kth pixel point relative to the ith class, m is a fuzzy factor, x k Is the kth pixel point, v i Is the i-th class center, N k Is a neighborhood of the current pixel point, α j Is a spatial distance weight coefficient, beta j Is a neighborhood similarity coefficient, x j Is N k The pixel point in (2).
d IFS (x k ,v i ) And d IFS (x j ,v i ) For intuitive blur distance, the following is defined:
d IFS (x k ,v i )＝(μ(x k )-μ(v i )) 2 +(ν(x k )-ν(v i )) 2 +(π(x k )-π(v i )) 2
d IFS (x j ,v i )＝(μ(x j )-μ(v i )) 2 +(ν(x j )-ν(v i )) 2 +(π(x j )-π(v i )) 2
wherein, mu (x) k )、ν(x k )、π(x k ) Respectively intuitively and fuzzily centralize pixel points x k Degree of membership, degree of non-membership, and degree of hesitation; mu (x) j )、ν(x j )、π(x j ) Respectively intuitively and fuzzily centralize pixel points x j Degree of membership, degree of non-membership, and degree of hesitation; mu (v) i )、ν(v i )、π(v i ) The membership, the non-membership and the hesitation of the intuitive fuzzy centralized clustering center are respectively. Intuition fuzzy set (seeThe literature: clazimir T. atanarsoff intuitive fuzzy set and fuzzy system, 1986, Vol.20, 97-96 (Atanassov K. intuitionistic fuzzy sets [ J].Fuzzy Sets&Systems,1986,20(1):87-96.)) was introduced by Classmimel. T. atasolf in 1986, taking into account information on three aspects of membership, non-membership and hesitation. The specific calculation formula is as follows:
π(x k )＝1-μ(x k )-ν(x k )
π(x j )＝1-μ(x j )-ν(x j )
wherein x is min Is the minimum value of the image gray scale, x max λ is a constant value for the maximum value of the image gray scale.
Position weight coefficient W ik The calculation process of (2) is as follows:
wherein P is k In order to segment the position probability coefficient of the target, the total number of classes c in the present invention is set to 2, where i ═ 1 is the background class and i ═ 2 is the target class. P k The calculation process of (2) is as follows:
repeating:
if P (X) (r) ) If ε is smaller, X is deleted (r) Repeating the above steps until there is no X (r) Until deleted, then
Wherein X (r) Comprises the following steps: white spirit, the coordinate of the possible infrared Ship area obtained by the variance-Based infrared Ship positioning method proposed by the Chen Shi (see the literature: white spirit, Chen Shi. Infrared Ship image Segmentation Based on the Spatial Information Improved fuzzy C-means algorithm. IEEE control treatise journal 46, 3259-].IEEE Trans Cybern,vol.46,no.12,pp.3259-3271,Dec.2016.))，X k Is a pixel point x k The coordinates of (c).
Spatial distance weight coefficient alpha j Similarity to neighborhood coefficient β j The calculation process of (2) is as follows:
wherein d is kj Representing neighborhood pixels x j And the central pixel point x k Spatial distance of (P) j Is a neighborhood pixel point x j Position probability coefficient of (2), N j Is a neighborhood pixel point x j Neighborhood of (2), x m Is N j The pixel point in (2). Thus, neighborhood pixel x j And the center pixel point x k The smaller the spatial distance of, alpha j The larger; x is the number of m And x j The higher the similarity, P j The larger, beta j The larger. Alpha is alpha j 、β j The larger the neighborhood pixels have a greater effect on the center pixel.
By utilizing a Lagrange multiplier method to differentiate the objective function, the membership degree u can be deduced ik And a clustering center v i Degree of membership mu (v) i ) V (v) non-membership degree v (v) i ) Hesitation degree pi (v) i ) The iterative formula is:
the specific process of the improved intuitive fuzzy clustering algorithm is as follows:
(1) calculating each pixel point x of the image k Degree of membership mu (x) k ) Non-membership v (x) k ) And a degree of hesitation pi (x) k )。
(2) And setting a category total number c, a fuzzy factor m, a maximum iteration number T and an iteration termination threshold epsilon.
(3) And randomly initializing a membership matrix U.
(4) A position weight coefficient matrix W is calculated.
(5) Updating the clustering center v i Degree of membership mu (v) i ) V (v) non-membership degree v (v) i ) Hesitation degree pi (v) i ) And each element U of the membership matrix U ik 。
(6) If U (t+1) -U (t) If | < epsilon or the iteration times exceed the maximum iteration times T, stopping the iteration; otherwise, returning to the substep (5).
(7) And (5) defuzzification is carried out, and the infrared ship image segmentation is completed.
Step three: and (3) performing subsequent processing on the clustering segmentation image obtained in the step two, wherein the subsequent processing comprises the following contents:
(1) regions with connected component areas smaller than a certain threshold (e.g., 20) are rejected.
(2) And calculating the ratio of the aspect ratio to the upper area and the lower area of each connected domain, and removing the undesirable areas. The aspect ratio of the ship is set to 0.5 to 6, and the ratio of the upper and lower areas is set to 0 to 1.
(3) Regions having a region area smaller than a certain threshold value when being connected with image boundaries are rejected. The threshold is set at 95% of the sum of all connected domain areas.
The invention has the advantages and effects that: the traditional fuzzy C-means algorithm does not consider target position information and spatial neighborhood information, and cannot effectively segment noisy images. Uneven gray scale is a common phenomenon in infrared images, areas with uneven gray scale can be regarded as noise, and the traditional fuzzy C-means algorithm cannot achieve an ideal segmentation effect. The improved intuitive fuzzy clustering algorithm provided by the invention adds the position information of the segmented target obtained by using the iterative algorithm, and improves the effect of the intuitive fuzzy C mean value. Meanwhile, a regular term of the similarity coefficient of the evaluation area is added into the target function, and the local spatial information is fully considered. The improved algorithm can better segment the infrared ship image and has wide market prospect and application value.
[ description of the drawings ]
FIG. 1 is a schematic diagram of an algorithm for extracting infrared ships by using position information to improve intuitive fuzzy clustering.
FIG. 2a is an original image pre-processed by morphological operations according to the present invention.
FIG. 2b is a diagram of the preprocessing result of the present invention using morphological operations to preprocess an image.
Fig. 3a is an original image applied to infrared ship image segmentation according to the present invention.
Fig. 3b is an original image applied to infrared ship image segmentation of the present invention.
Fig. 3c is a clustering result graph applied to infrared ship image segmentation in the present invention.
FIG. 3d is a clustering result graph applied to the infrared ship image segmentation of the present invention.
Fig. 4a is a clustering result graph after the application of the present invention and the subsequent processing of infrared ship image segmentation.
Fig. 4b is a clustering result graph after the application of the present invention and the subsequent processing of infrared ship image segmentation.
[ detailed description ] embodiments
In order to better understand the technical scheme of the invention, the following description is further provided for the embodiment of the invention with reference to the attached drawings.
The schematic diagram of the invention is shown in fig. 1, and the specific implementation steps are as follows:
the method comprises the following steps: the original image is preprocessed by adopting morphological operation top-hat filtering, and the influence of uneven gray level in the infrared ship image on a clustering segmentation result is reduced. Fig. 2a is an original, and fig. 2b is an image after preprocessing.
Step two: and (3) clustering and segmenting the infrared image by adopting an improved intuitive fuzzy clustering algorithm.
The objective function of the improved intuitive fuzzy clustering algorithm is as follows:
wherein i is the category ordinal number, c is the category total number, k is the position coordinate of the pixel point, N is the pixel point total number, W ik Is a position weight coefficient, u ik Is the membership degree of the kth pixel point relative to the ith class, m is a fuzzy factor, x k Is the k-th pixel point, v i Is the class i center, N k Is a neighborhood of the current pixel point, α j Is a spatial distance weight coefficient, beta j Is a neighborhood similarity coefficient, x j Is N k The pixel point in (2).
d IFS (x k ,v i ) And d IFS (x j ,v i ) For intuitive blur distance, the following is defined:
d IFS (x k ,v i )＝(μ(x k )-μ(v i )) 2 +(ν(x k )-ν(v i )) 2 +(π(x k )-π(v i )) 2
d IFS (x j ,v i )＝(μ(x j )-μ(v i )) 2 +(ν(x j )-ν(v i )) 2 +(π(x j )-π(v i )) 2
wherein, mu (x) k )、ν(x k )、π(x k ) Respectively intuitively and fuzzily centralize pixel points x k Degree of membership, degree of non-membership, and degree of hesitation; mu (x) j )、ν(x j )、π(x j ) Respectively intuitively and fuzzily centralize pixel points x j Degree of membership, degree of non-membership, and degree of hesitation; mu (v) i )、ν(v i )、π(v i ) The membership, the non-membership and the hesitation of the intuitive fuzzy centralized clustering center are respectively. Intuitive fuzzy sets (see Classmil. T. attaxasofv. intuitive fuzzy sets and fuzzy systems,1986, Vol.20, 97-96 (Atanassov K. intuitionistic fuzzy sets [ J ] J].Fuzzy Sets&Systems,1986,20(1):87-96.)) was introduced by Classmimel. T. atasolf in 1986, taking into account information on three aspects of membership, non-membership and hesitation. The specific calculation formula is as follows:
π(x k )＝1-μ(x k )-ν(x k )
π(x j )＝1-μ(x j )-ν(x j )
wherein x min Is the minimum value of image gray scale, x max λ is a constant value for the maximum value of the image gray scale.
Position weight coefficient W ik The calculation process of (2) is as follows:
wherein P is k The position probability coefficient of the segmentation target is shown, wherein i-1 is a background class, and i-2 is a target class. P k The calculation process of (c) is as follows:
repeating:
if P (X) (r) ) If ε is smaller, X is deleted (r) Repeating the above steps until there is no X (r) Until deleted, then
Wherein X (r) Comprises the following steps: white spirit, the coordinate of the possible infrared Ship area obtained by the variance-Based infrared Ship positioning method proposed by the Chen Shi (see the literature: white spirit, Chen Shi. Infrared Ship image Segmentation Based on the Spatial Information Improved fuzzy C-means algorithm. IEEE control treatise journal 46, 3259-].IEEE Trans Cybern,vol.46,no.12,pp.3259-3271,Dec.2016.))，X k Is a pixel point x k The coordinates of (a).
Spatial distance weight coefficient alpha j Similarity to neighborhood coefficient β j The calculation process of (c) is as follows:
wherein d is kj Representing neighborhood pixels x j And the central pixel point x k Spatial distance of (P) j Is a neighborhood pixel point x j Position probability coefficient of (2), N j Is a neighborhood pixel point x j Neighborhood of (2), x m Is N j The pixel point in (2). Thus, neighborhood pixel x j And the central pixel point x k The smaller the spatial distance of, alpha j The larger; x is the number of m And x j The higher the similarity, P j The larger, beta j The larger. Alpha is alpha j 、β j The larger the neighborhood pixels have a greater effect on the center pixel.
By utilizing a Lagrange multiplier method to differentiate the objective function, the membership degree u can be deduced ik And a clustering center v i Degree of membership mu (v) i ) V (v) non-membership degree v (v) i ) Hesitation degree pi (v) i ) The iterative formula is:
the specific process of the improved intuitive fuzzy clustering comprises the following steps:
(1) calculating each pixel point x of the image k Degree of membership mu (x) k ) Non-membership v (x) k ) And a degree of hesitation pi (x) k )。
(2) Setting a category total number c, wherein the category total number c is set to be 2 in the embodiment of the invention; a blurring factor m; a maximum number of iterations T and an iteration termination threshold epsilon.
(3) And randomly initializing a membership matrix U.
(4) A position weight coefficient matrix W is calculated.
(5) Updating the clustering center v i Degree of membership μ (v) i ) V (v) non-membership degree v (v) i ) Hesitation degree pi (v) i ) And each element U of the membership matrix U ik 。
(6) If U (t+1) -U (t) If | < epsilon or the iteration times exceed the maximum iteration times T, stopping the iteration; otherwise, returning to the substep (5).
(7) And (5) defuzzification is carried out, and the infrared ship image segmentation is completed.
Fig. 3a and 3b are original images for infrared ship image segmentation of the present invention, and fig. 3c and 3d are clustering result graphs for infrared image segmentation of the present invention. As can be seen from fig. 3c and 3d, the clustered and segmented image may contain other regions, and a subsequent process is required to remove non-target regions.
Step three: and (3) performing subsequent processing on the clustering segmentation image obtained in the step two, wherein the subsequent processing comprises the following contents:
(1) and eliminating the area of the connected domain smaller than 20.
(2) And calculating the ratio of the aspect ratio to the upper area and the lower area of each connected domain, and removing the undesirable areas. The aspect ratio of the ship is set to 0.5 to 6, and the ratio of the upper and lower areas is set to 0 to 1.
(3) Regions having a region area smaller than a certain threshold value when being connected with image boundaries are rejected. The threshold is set at 95% of the sum of all connected domain areas.
Fig. 4a and 4b are cluster segmentation images after subsequent processing. In contrast to fig. 3c and 3d, the subsequent processing eliminates non-target areas.
Claims (2)
1. A method for improving intuitive fuzzy clustering and extracting infrared ships by using position information is characterized by comprising the following steps: the method comprises the following specific steps:
the method comprises the following steps: preprocessing an original image by adopting morphological operation top-hat filtering, and reducing the influence of gray unevenness in an infrared ship image on a clustering segmentation result;
step two: clustering and segmenting the infrared image by adopting an improved intuitive fuzzy clustering algorithm; the specific process is as follows:
(1) calculating each pixel point x of the image k Degree of membership mu (x) k ) Non-membership v (x) k ) And a degree of hesitation pi (x) k )；
(2) Setting a category total number c, a fuzzy factor m, a maximum iteration number T and an iteration termination threshold epsilon;
(3) randomly initializing a membership matrix U;
(4) calculating a position weight coefficient matrix W;
(5) updating the clustering center v i Degree of membership mu (v) i ) V (v) non-membership degree v (v) i ) Hesitation degree pi (v) i ) And each element U of the membership matrix U ik ；
(6) If U (t+1) -U (t) If | < epsilon or the iteration times exceed the maximum iteration times T, stopping the iteration; otherwise, returning to the substep (5);
(7) defuzzification is carried out, and infrared ship image segmentation is completed;
step three: performing subsequent processing on the clustering segmentation image obtained in the step two;
the improved intuitive fuzzy clustering algorithm comprises the following specific steps:
the objective function of the improved intuitive fuzzy clustering algorithm is as follows:
wherein i is the category ordinal number, c is the category total number, k is the position coordinate of the pixel point, N is the pixel point total number, W ik Is a position weight coefficient, u ik Is the membership degree of the kth pixel point relative to the ith class, m is a fuzzy factor, x k Is the k-th pixel point, v i Is the class i center, N k Is a neighborhood of the current pixel point, α j Is a spatial distance weight coefficient, beta j Is a neighborhood similarity coefficient, x j Is N k The pixel point in (1);
d IFS (x k ,v i ) And d IFS (x j ,v i ) For intuitive blur distance, the following is defined:
d IFS (x k ,v i )＝(μ(x k )-μ(v i )) 2 +(ν(x k )-ν(v i )) 2 +(π(x k )-π(v i )) 2
d IFS (x j ,v i )＝(μ(x j )-μ(v i )) 2 +(ν(x j )-ν(v i )) 2 +(π(x j )-π(v i )) 2
wherein, mu (x) k )、ν(x k )、π(x k ) Respectively intuitively and fuzzily centralize pixel points x k Degree of membership, degree of non-membership, and degree of hesitation; mu (x) j )、ν(x j )、π(x j ) Respectively intuitively and fuzzily centralize pixel points x j Degree of membership, degree of non-membership, and degree of hesitation; mu (v) i )、ν(v i )、π(v i ) Membership, non-membership and hesitation of an intuitive fuzzy centralized clustering center respectively; the specific calculation formula is as follows:
π(x k )＝1-μ(x k )-ν(x k )
π(x j )＝1-μ(x j )-ν(x j )
wherein x min For minimizing image gray scaleValue, x max Is the maximum value of the image gray scale, and lambda is a constant;
position weight coefficient W ik The calculation process of (2) is as follows:
wherein P is k Setting the total number of classes c as 2 for dividing the position probability coefficient of the target, wherein i-1 is a background class, and i-2 is a target class; p k The calculation process of (2) is as follows:
repeating:
if P (X) (r) ) If ε is smaller, X is deleted (r) Repeating the above steps until there is no X (r) Until deleted, then
Wherein X (r) As coordinates, X, of the infrared ship's probable region k Is a pixel point x k The coordinates of (a);
spatial distance weight coefficient alpha j Similarity to neighborhood coefficient β j The calculation process of (2) is as follows:
wherein d is kj Representing neighborhood pixels x j And the central pixel point x k Spatial distance of (P) j Is a neighborhood pixel point x j Position probability coefficient of (2), N j Is a neighborhood pixel point x j Neighborhood of (2), x m Is N j The pixel point in (1); thus, neighborhood pixel x j And the central pixel point x k The smaller the spatial distance of, alpha j The larger; x is the number of m And x j The higher the similarity, P j The larger, beta j The larger; alpha is alpha j 、β j The larger the neighborhood pixel has to be, the larger the influence of the neighborhood pixel on the center pixel is;
utilizing Lagrange multiplier method to conduct derivation on target function and deduce membership degree u ik And a clustering center v i Degree of membership mu (v) i ) V (v) non-membership degree v (v) i ) Hesitation degree pi (v) i ) The iterative formula is:
the subsequent treatment comprises the following steps:
s31, eliminating the area of the connected domain smaller than a certain threshold;
s32, calculating the ratio of the length-width ratio to the upper area and the lower area of each connected domain, and eliminating the regions which are not qualified;
and S33, eliminating the area of the region which is smaller than a certain threshold value when the area is connected with the image boundary.
2. The method for improving the intuitive fuzzy clustering extraction of infrared ships according to claim 1, wherein the method comprises the following steps: the threshold value in the step S31 is 20, the length-width ratio range of the ship is set to be 0.5-6, and the ratio range of the upper area and the lower area is set to be 0-1; the threshold value in step S33 is set to 95% of the sum of all the connected component areas.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911035552.5A CN110853054B (en) | 2019-10-29 | 2019-10-29 | Method for improving intuitive fuzzy clustering and extracting infrared ship by using position information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911035552.5A CN110853054B (en) | 2019-10-29 | 2019-10-29 | Method for improving intuitive fuzzy clustering and extracting infrared ship by using position information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110853054A CN110853054A (en) | 2020-02-28 |
| CN110853054B true CN110853054B (en) | 2022-08-16 |
Family
ID=69598351
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911035552.5A Active CN110853054B (en) | 2019-10-29 | 2019-10-29 | Method for improving intuitive fuzzy clustering and extracting infrared ship by using position information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110853054B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111539945B (en) * | 2020-04-28 | 2022-09-13 | 西华大学 | Insulator image segmentation method for improving PCM clustering |
| CN111524181B (en) * | 2020-04-28 | 2023-03-24 | 陕西科技大学 | Automatic measurement method for porous material holes based on scanning electron microscope image segmentation |
| CN112052870B (en) * | 2020-07-20 | 2024-04-16 | 武汉罗布科技有限公司 | Ship magnetic field classification method based on discrete Frenchet distance |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108062757A (en) * | 2018-01-05 | 2018-05-22 | 北京航空航天大学 | It is a kind of to utilize the method for improving Intuitionistic Fuzzy Clustering algorithm extraction infrared target |
| CN108198193A (en) * | 2018-01-16 | 2018-06-22 | 北京航空航天大学 | It is a kind of to utilize the method for improving Intuitionistic Fuzzy Clustering algorithm segmentation infrared ship image |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107977952A (en) * | 2016-10-21 | 2018-05-01 | 冯原 | Medical image cutting method and device |
-
2019
- 2019-10-29 CN CN201911035552.5A patent/CN110853054B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108062757A (en) * | 2018-01-05 | 2018-05-22 | 北京航空航天大学 | It is a kind of to utilize the method for improving Intuitionistic Fuzzy Clustering algorithm extraction infrared target |
| CN108198193A (en) * | 2018-01-16 | 2018-06-22 | 北京航空航天大学 | It is a kind of to utilize the method for improving Intuitionistic Fuzzy Clustering algorithm segmentation infrared ship image |
Non-Patent Citations (3)
| Title |
|---|
| Xiangzhi Bai 等.Infrared Ship Target Segmentation Based on Spatial Information Improved FCM.《IEEE TRANSACTIONS ON CYBERNETICS》.2015, * |
| 湛西羊 等.融入局部信息的直觉模糊核聚类图像分割算法.《信号处理》.2017,第33卷(第3期), * |
| 白相志 等.新型Top-hat变换及其在红外小目标检测中的应用.《数据采集与处理》.2009,第24卷(第5期), * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110853054A (en) | 2020-02-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108062757B (en) | Method for extracting infrared target by using improved intuitionistic fuzzy clustering algorithm | |
| CN110853054B (en) | Method for improving intuitive fuzzy clustering and extracting infrared ship by using position information | |
| Al-Amri et al. | Image segmentation by using edge detection | |
| CN108198193B (en) | Method for segmenting infrared ship image by using improved intuitionistic fuzzy clustering algorithm | |
| Zhang et al. | Novel fuzzy clustering algorithm with variable multi-pixel fitting spatial information for image segmentation | |
| Sulaiman et al. | Denoising-based clustering algorithms for segmentation of low level salt-and-pepper noise-corrupted images | |
| CN107590427B (en) | Method for detecting abnormal events of surveillance video based on space-time interest point noise reduction | |
| Beevi et al. | A robust segmentation approach for noisy medical images using fuzzy clustering with spatial probability | |
| Zhao et al. | A novel Neutrosophic image segmentation based on improved fuzzy C-means algorithm (NIS-IFCM) | |
| Lorette et al. | Fully unsupervised fuzzy clustering with entropy criterion | |
| Pan et al. | Single-image dehazing via dark channel prior and adaptive threshold | |
| Zhengzhou et al. | Gray-scale edge detection and image segmentation algorithm based on mean shift | |
| Devikar et al. | Segmentation of images using histogram based FCM clustering algorithm and spatial probability | |
| Irshad et al. | Image fusion using computational intelligence: A survey | |
| Liu et al. | An optional gauss filter image denoising method based on difference image fast fuzzy clustering | |
| Ye et al. | Improved edge detection algorithm of high-resolution remote sensing images based on fast guided filter | |
| CN112465837B (en) | Image segmentation method for sparse subspace fuzzy clustering by utilizing spatial information constraint | |
| Vignesh et al. | Performance and Analysis of Edge detection using FPGA Implementation | |
| CN110264417B (en) | Local motion fuzzy area automatic detection and extraction method based on hierarchical model | |
| Liu et al. | Automatic Lung Parenchyma Segmentation of CT Images Based on Matrix Grey Incidence. | |
| Kannan et al. | Fuzzy clustering Approach in segmentation of T1-T2 brain MRI | |
| Panda et al. | Edge preserving image fusion using intensity variation approach | |
| Xu | Research on blurred edge information segmentation of image based on computer vision | |
| Dhanalakshmi et al. | Image processing using Modified Multiple kernel fuzzy c-means clustering (MMKFCM) technique | |
| Huang et al. | Multi-focus image fusion combined with cnn and algebraic multi-grid method |
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 |