摘要
首次提出用人工神经网络理论处理磁测资料，进行磁异常场分区的定量计算方法。并将该方法应用于南海研究。该方法将研究区域划分为若干统计单元，确定各统计单元中的特征变量、选择标准模型，并对BP神经网络进行训练。用学习成功的网络识别、分析整个研究区的磁异常场，从而将南海的磁异常场划分为12个特征区。
A method of quantitative calculation is pressnted for processing the data ofmagnetic measuement to divide a magnetic anomaly field using artifical neural nets.This method is successfully applied to the South China Sea. The study region is divided into statistical elements, The controlling variables for elements is chosen.The standard modles are selected and used to train the B-P network. The network that has been magnetfic anomaly field of the study region. Thus the magnetic field of the South China Sea is divided into 12 characteristic region.
关键词
南海
神经网络
磁异常场
定量分区
Artifical neural netrwork
Divide area of magnetic anomaly the South China Sea
