【摘要】 地震属性图像中有许多点的梯度振幅值比较大,而这些点在特定的应用领域中并不都是边缘点,若直接利用这些地震属性进行储层预测就会降低预测精度。针对此问题,本文首先利用井资料确定砂泥岩分布范围,再采用Sobel算子进行边缘检测,求出振幅梯度幅值阈值,并依此阈值作为识别砂岩边界的标准。为了检验上述边缘检测的效果,利用正演模型模拟结果与原始地震剖面进行对比,发现两者结果吻合很好,说明上述边缘检测是有效的。更多还原

【Abstract】 As the gradient amplitude values for some points in seismic attribute image are relatively large and not all these points are the marginal point in specific application area,the precision of the prediction may be decreased if such seismic attributes are applied directly to the reservoir prediction.Aimed at this problem,the paper first determined the distribution range of sand and shale by using the well data,and then implemented the edge detection using the Sobel operator to derive the threshold of the gradient amplitude and finally the threshold was used as a standard to identify the boundary of sand and shale.To verifying the effect of the above method,a comparison between the forward modeling result and the real seismic section was carried out.It can be seen from results that they matched well each other,showing the effectiveness of this edge detection technique.更多还原