Pouyan Ebrahimi Lialekol - Petroleum Exploration Department, Faulty of Mining Engineering, Sahand University of Technology, Sahand New Town, Iran
Mirali Hamim Oskouei - Petroleum Exploration Department, Faulty of Mining Engineering, Sahand University of Technology, Sahand New Town, Iran
Ehsan Barzgar - Petroleum Exploration Department, Faulty of Mining Engineering, Sahand University of Technology, Sahand New Town, Iran
S. Milad Razavi - Petroleum Exploration Department, Amirkabir University of Technology, Tehran, Iran

[توضیح سیویلیکا: اطلاعات مقاله طبق درخواست 34740 حذف گردید.]In this study, comparing results of Image Processing, Neural Network (NN) and Gas Chimney methods distinguishes open and closed fractures. Aftera precise processing, Dip-Median and Diffusion Filters were combined together to cover their limitations. Then using appropriate attributes,fractures were highlighted. After that seismic data were classified to Fault-Yes and Fault-No groups. Then, after introducing the computed attributes and classified groups to NN, whole data were considered. After extracting the resultant images, the fractures were highlighted using Laplacian filterand Principal Component Analysis (PCA) to increase precision of images from fracture point of view and reduce all vector spaces to principal components.