【摘要】 为了在时频域刻画裂缝等边缘信息及其多尺度特征,构造了一种新的广义S变换分频边缘检测。它借鉴了小波多尺度边缘检测的思想,利用广义S变换的时频局部谱,通过不同频率尺度滤波达到突出与定位不同尺度边缘信息的目的。模型信号处理表明:广义S变换分频边缘检测方法既能准确刻画边缘奇异信息,又能反映边缘的多尺度特征,具有与小波多尺度边缘检测相似的特点。利用这一方法对实际地震资料处理的结果表明,它可以在不同频率尺度情况下有效检测断层、裂缝发育带的几何形态和空间展布。更多还原

【Abstract】 In order to delineate fracture edge information and its multi-scale characteristic in time-frequency domain,a new multi-frequency edge detection(MFED) method based on generalized S transform(GST) is presented,which draw on the idea of wavelet multi-scale edge detection.The MFED achieves highlighting and localizing different scales edge by filtering of different frequencies based on the local time-frequency spectra of GST.The simulation of synthetic signal model demonstrates that MFED can exactly depict the various singular edges and reflect the multi-scale characteristics of the edge,which is similar to the wavelet multi-scale edge detection.MFED is applied to real seismic data processing.The results show that it can detection the fault,the geometry and distribution of fracture zone in different frequency scales.更多还原