Xiong Xiao-jun1, He Zhen-hua1, Zhao Ming-jin2, Huang De-ji1
1. Chengdu University of Technology,Chengdy City,Sichuan Province,610059,China;
2. Hengtai-Aipu Technical Co.,Ltd of Petroleum Exploration and Development,Beijing City,100089,China
Abstract：In order to overcome the shortcoming that there is large influence of noise on the conventional non-continuous detection-based fractural prediction method and by introducing 2-D edge-preserved smoothing technique,the paper presented a new generalized Hilbert transform-based fractural detection method (GHT),which first carried out edge-preserved smoothing processing of amplitude slice by extracting amplitude slice of 3-D poststack seismic data in time domain,then implemented windowing Hilbert transform in frequency domain by optimizing window and order,and finally using the amplitude information of analytic signal to predict fracture. The computing results of practical data showed the GHT-based fractural detection method presented by the paper is characteristics having not only strong noise-suppressed capacity but also higher precision of fractural detection,which is able to effectively predict distribution feature of structural fracture and fracture formed by karst pore and cave,the detected results and analytic data of conventional stress field are coincident to drilling data.
