
Ukazuje si od 1919 roku
3b'12
Organ Stowarzyszenia Elektryków Polskich Wydawnictwo SIGMA-NOT Sp. z o.o.
Yucheng LIU1, Yubin LIU2
School of Electrical & Information Engineering, Chongqing University of Science & Technology (1), Computer Science School, Panzhihua
University (2)
A New Image Enhancement Based on the Fuzzy C-Means Clustering
Abstract. The enhancement of the infrared dim small target image is a very important pretreatment in automatic recognition of target and infrared target tracking system. The paper proposed a new image enhancement algorithm based on the Fuzzy C-Means clustering. The algorithm conducted cluster analysis on the pixel and gray of the infrared image and increased the image gray level difference between the various objects so as to achieve the enhanced purpose for infrared small target image. The experimental results showed that this algorithm is able to enhance small target image to the maximum extent with ensuring no loss of the target information
Streszczenie. W artykule za[proponowano nowy algorytm poprawy jako ci obrazu otrzymanego w podczerwieni bazuj cy na klastrach rozmytych typu C. Algorytm analizuje piksle i odcień szaro ci obrazu i rozszerza poziom ró nic szaro ci. Symulacje potwierdzi y przydatno algorytmu dla ma ych zbiorów. (Nowy algorytm poprawy obrazu w podczerwieni bazuj cy na rozmytych klastrach typu C).
Keywords: infrared image; image enhancement; algorithm; fuzzy C-means clustering S owa kluczowe: obraz w podczerwieni, jako obrazu.
Introduction
Due to the constraints of the sensitivity, resolution, noise and other features of infrared imaging devices and the existence of radiation and other factors affect between the objectives and the surrounding environment, the image acquired by the Infrared imaging system is not as rich level as the visible light image. Infrared images showed more images noisy, low contrast. Especially in the infrared small target, they often were submerged in a large background, which brought great difficulties to the identification and extraction of the target. So it is necessary to conduct the enhancement of the infrared images obtained. As the edge gradient of the target and the background was not large and the edge ambiguity was great, in general, the enhancement effect of a variety of the image enhancement methods used in the past was not obvious.
Based on the gray distribution feature of the infrared dim small target image, the paper proposed a new image enhancement algorithm based on the Fuzzy C-Means clustering. The algorithm conducts first the cluster analysis of the infrared image, gives then each category gray value so as to increase the edge gradient of the target.
(1)
Jm U,V (uik)mdik2(xk,vi)
k 1i 1
nc
Here, U={uik} is for the membership matrix, V={v1, v2, … , vc} is for the set of C cluster center points, m∈ [1,∞] is for the weighted index. When m=1, the fuzzy clustering retrogrades as the hard C-means clustering. The reference [3, 4] showed that the best choice range of m is [1.5, 2.5]. Usually m = 2 is the ideal value.
The distance from the k-th sample to the i-th class center is defined as: (2)
2dik(xk,vi) ||xk vi|| (xk vi)TA(xk vi)
Here, A is positive definite matrix P×P, when A is the unit
matrix, the distance is the Euclidean distance. Fuzzy C-means clustering optimizes the objective function as shown in the formula (1) through repeated iterations, which performs the following steps:
(1) Initialize the cluster centers V={v1, v2, … , vc} ; (2) Calculate the membership matrix: (3)
Fuzzy C-means clustering algorithm
Fuzzy C-means clustering is an unsupervised statistical method. It does not require training samples. By means of the iterative, it optimizes the objective function used to represent the image pixel similarity with the C class center and obtains maximum value, resulting in optimal clustering [1, 2]. The basic idea of the fuzzy C-means clustering algorithm is as follows: Class the data set X={x1, x2, … , xn}∈Rpn into C classes. If set the membership of any samples xk in the X for the i-th class as uik, then classification results can be a fuzzy membership matrix U = {uik}∈Rcn, Fuzzy C-Means clustering is achieved by minimizing the objective function Jm (U, V) about the membership matrix U and cluster center V:
uik [ (
j 1
c
djk(xk,vj)2/(m 1)
djk(xk,vj)
1,k 1,2, c
(3) Update cluster centers and calculate Jm(U,V);
(4)
vi
(u
k 1nk 1
n
ik
)mxk
,i 1,2, c
ik
(u
)m
(4) Select the appropriate matrix norm (such as, Euclidean distance), compared Jm(k) with Jm(k+1), if∥Jm(k+1)- Jm(k)∥<ε th
en stop iteration; otherwise k = k +1, go to (2).
PRZEGL D ELEKTROTECHNICZNY (Electrical Review), ISSN 0033-2097, R. 88 NR 3b/2012 1
A New Image Enhancement ... 暂无评价 4页 2财富值 Improving fuzzy c-...based on fuzzy c-means clustering in which some membership grades are given...
New two-dimensional fuzzy C-means clustering algorithm for image segmentation_专业资料。JCn.ot i.eho.20)58287.etSuhUnvTcni(081:8~8 DoI0.O/17—0~...
scale intervals and fix on the points, a Remote Sensing Image Enhancement Algorithm Based on Weighting Fuzzy C-Means Clustering (RSIE-WFCM) is proposed....
A New Image Enhancement ... 暂无评价 4页 2...(hard) c-means and fuzzy c-means clustering ...algorithm with a new Gaussian function based ...
Remote Sensing Image Classification Based on Improved Fuzzy c-Means_专业资料 暂无评价
which for the effect of a sensor, atmosphere etc, an remote sensing image enhancement method based on weighting fuzzy c-means clustering (RSIE-WFCM) ...
A New Image Enhancement ... 暂无评价 4页 2财富值 Improving fuzzy c-... Clustering based on gene... 2页 免费 Parallel K-Means Cluster... 6页...
的图像分割方法 ESFCM (Edge-based Semi-Fuzzy C-means Clustering) 该方法利用...Based on the inherent character that the fuzzy part in an image should be...
a 2-dimension Fuzzy C-Means (2DFCM) clustering based algorithm to segment ...[C]//International Conference on Image Processing.New York, USA: IEEE, ...
Fuzzy C-Means Clustering Course Project Presentation...? ? Image and speech enhancement Edge detection ...[PalPB97] Resubstitution errors based on the ...
热门文档