Oil spill detection techniques using satellite images in gulf of mexico region ijaerdv05i0266156

Page 3

International Journal of Advance Engineering and Research Development (IJAERD) Volume 5, Issue 02, February-2018, e-ISSN: 2348 - 4470, print-ISSN: 2348-6406 produces difference images using information of pixels. Log value used to find out intensity for images and it covers larger đ?‘Ľ area. Let us considered iteration x =1 with standard deviation đ?œŽđ?‘&#x;1 đ?‘Žđ?‘›đ?‘‘ mean đ?œ‡1đ?‘&#x; values. Now find out energy function đ??¸đ?‘?đ?‘ž đ?‘? with iteration ‘x’. Now apply Gibbs expression, to find out prior probability (đ?œ‹đ?‘?đ?‘ž ) x exp(−Epq ) đ?‘? đ?œ‹đ?‘?đ?‘ž = x x ) exp −Epq + exp(−Epq đ?‘? đ?‘? Now (đ?‘?đ?‘&#x; ) determine conditional probability then generate the distance matrix (đ?‘‘đ?‘?đ?‘ž ) for the given input image. đ?‘?2 đ?‘Śđ?‘— đ?‘Śđ?‘— − đ?œ‹đ?‘&#x; 1 đ?‘?đ?‘&#x;đ?‘? đ?‘? đ?‘? = đ?‘’đ?‘Ľđ?‘? − đ?‘? đ?œ‡đ?‘&#x; , đ?œŽđ?‘&#x; 2(đ?œŽđ?‘&#x;đ?‘? )2 đ?œŽđ?‘&#x; 2đ?œ‹ đ?‘? đ?‘‘đ?‘?đ?‘ž = −1đ?‘› đ?‘?đ?‘&#x;đ?‘?

đ?‘Śđ?‘— đ?‘? đ?œ‡đ?‘&#x; , đ?œŽđ?‘&#x;đ?‘?

đ?‘? Now compute objective function- đ?‘‘đ?‘?đ?‘ž = −1đ?‘› đ?‘?đ?‘&#x;đ?‘? đ?‘? đ?‘?−1 đ??˝đ?‘?đ?‘ž − đ??˝đ?‘?đ?‘ž ≤đ?›ż

đ?‘Śđ?‘— đ?œ‡ đ?‘&#x;đ?‘? ,đ?œŽđ?‘&#x;đ?‘?

,

đ?‘?+1 Now determine Membership matrix {đ?œ‡đ?‘?đ?‘ž } đ?‘? đ?‘? đ?œ‹ exp(−đ?‘‘ ) đ?‘?đ?‘ž đ?‘?đ?‘ž đ?‘?+1 đ?œ‡đ?‘?đ?‘ž = đ?‘? đ?‘? đ?‘? exp(−đ?‘‘ đ?‘? ) đ?œ‹đ?‘?đ?‘ž exp −đ?‘‘đ?‘?đ?‘ž + đ?œ‹đ?‘?đ?‘ž đ?‘?đ?‘ž đ?‘?+1 Update deviation and mean values đ?œŽđ?‘&#x; and đ?œ‡đ?‘&#x;đ?‘?+1 respectively, c=c+1 đ?‘? đ?‘Ľ (đ?œ‡đ?‘?đ?‘ž đ?‘Śđ?‘— ) đ?‘? đ?‘Ľ (đ?œ‡đ?‘?đ?‘ž )

đ?œ‡đ?‘&#x;đ?‘?+1 =

đ?œŽđ?‘&#x;đ?‘?+1 =

B.

đ?‘Ľ

đ?‘? (đ?‘Ś − đ?œ‡ đ?‘? +1 )2 đ?‘˘đ?‘?đ?‘ž đ?‘— đ?‘&#x; đ?‘? ) (đ?‘˘ đ?‘Ľ đ?‘?đ?‘ž

Fuzzy c-Mean

It is a clustering technique, which allows one object belongs to two or more object in clusters form. Similarly object or data will be placed in one place and other place. This method improves frequently using pattern recognition. Minimization object function 2 , đ?‘? đ?‘› Jx= đ?‘€ 1≤n<∞ đ?‘?=1 đ?‘ž=1 đ?‘˘đ?‘?đ?‘ž || xp – cq || Whereas m, belongs to real number which is greater than 1, upq is a degree of membership, p is a dimensional measured data, Cq is a cluster center. || * || it denotes similarity between center and measured data. In fuzzy c- mean clustering, with the help of iterative optimization objective function, fuzzy partitioning can be done, using update membership đ?‘˘đ?‘?đ?‘ž and center of cluster đ?‘?đ?‘? . đ?‘˘đ?‘?đ?‘ž =

1 2 | đ?‘Ľ đ?‘? −đ?‘? đ?‘ž | đ?‘? đ?‘› −1 đ?‘–=1 ||đ?‘Ľ đ?‘? −đ?‘? || . đ?‘–

, cpq=

đ?‘€ đ?‘˘đ?‘› đ?‘Ľ đ?‘? =1 đ?‘?đ?‘ž đ?‘? đ?‘€ đ?‘˘đ?‘› đ?‘ž =1 đ?‘?đ?‘ž

(đ?‘–+1)

Iteration stops when đ?‘šđ?‘Žđ?‘Ľđ?‘?đ?‘ž đ?‘˘đ?‘?đ?‘ž

(đ?‘–)

− đ?‘˘đ?‘?đ?‘ž < Es

E denotes terminal point between 0 and 1 for kth iteration; it also covers local minimum pointđ??˝đ?‘Ľ . Following steps are 1.

First initialize process U so it denoted as [upq]matrix, u[0]

2.

Now calculate center of vectors c(i)=[cq] with u(i) , cpq=

đ?‘€ đ?‘˘đ?‘› đ?‘Ľ đ?‘? =1 đ?‘?đ?‘ž đ?‘? đ?‘€ đ?‘˘đ?‘› đ?‘ž =1 đ?‘?đ?‘ž

3. Update the process by using u(i), u(i+1) 1 Therefore Upq= 2 | đ?‘Ľ −đ?‘? | đ?‘? đ?‘ž đ?‘? đ?‘› −1 đ?‘–=1 ||đ?‘Ľ đ?‘? −đ?‘? || . đ?‘–

@IJAERD-2018, All rights Reserved

1012

Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.