利用多种模式识别法预测廊固拗陷含油砂体

Predicting oil sand body in the Langgu depression by several pattern recognition methods

通过分析地震资料，最大限度地提取各种特征参数，确定各种地震参数及其分布规律，再使用综合参数法、判别分析法、自组织学习人工神经网络和有监督学习人工神经网络等多种技术对廊固拗陷已知油藏J砂体进行了预测验证，并对B砂体和C砂体从正反两方面进行了成功预测。

Various characteristic parameters were uttermostly derived from seisrnic data analysis. Then various seismic parameters and the distributions were determined definitely. The prediction and verification of known J oil'sand body in the Langgu depression have been done successfully by making use of several methods such as integration, discrimination, unsupervised artificial neural network and supervised artificial neural network. What is more, forward and inverse predictions of B and C sand bodies have been achieved successfully.