中文摘要：

基于多地震属性、测井资料，用多变量回归算法成功地对胜利油田ken-71地区的拟测井参数做出了预测．计算采用了Daniel P．Hampson提出的将多变量回归权重系数推广为具有一定时间长度的褶积算子，使预测结果的分辨率获得提高．对该方法提高分辨率的原理做了详细的讨论，给出了该算法在胜利油田ken-71地区采用常规地面地震数据和测井数据，预测得到的目标区域的拟孔隙度参数的分布，结果显示使用该方法深度分辨率可达8～10m．

英文摘要：

The prediction of high resolution pseudo log based on multiple seismic attributes and well logs is achieved in the Ken-71 area of Shengli Oilfield. In this calculation, the multivariate regression weight coefficients are extended to convolutional operators with specific time length to improve the depth resolution, which is proposed by Daniel P. Hampson. The principle of enhanced resolution obtained by employing convolutional operator is discussed in detail. The predicted pseudo porosity by using ordinary surface seismic data and well log in this area is presented. The results show that with this approach the 8- 10 m depth resolution can be achieved.