基于局部投影和边缘检测的自动分层模型

AUTOMATIC LAYER SEPARATION MODEL BASED ON THE LOCAL PROJECTION AND EDGE DETECTION METHODS

笔者将局部投影算法和边缘检测运用于测井曲线的自动分层。首先将测井时间序列进行相空间重构,在高维相空间中用局部投影算法来分离测井信号中的由非地层因素引起的噪声;再由边缘检测确定出隶属度高的数据点,根据损失函数确定出可能的分界点集,最后用最优分割法实现自动分层。通过对测井时间序列的处理,得到较满意的结果。

In this paper local projection and edge detection methods were used for automatic layer separation. First, phase space reconstruction was adopted for log time series, and then noise caused by non-stratigraphic factors was reduced by local projection. Edge pro- jection was utilized to study data of high-degree membership and to detect possible layer boundary according to the loss function. Final- ly, the model was constructed based on local projection and edge detection. Its practicability and feasibility are demonstrated clearly by log time series data obtained in west Sichuan Province.