神经网络方法在放射性地球物理寻找铀矿中的应用  

Application of neural network methodin radioactivity prospecting of uranium deposits

作　　者：郑勇明[1] 杨亚新[1] 余水泉 张叶[1] 陈越[1] 杨兴[1] 

机构地区：[1]东华理工大学核资源与环境教育部重点实验室,江西南昌330013 [2]中国核工业地质局,北京100013

出　　处：《铀矿冶》2009年第4期220-224,共5页Uranium Mining and Metallurgy

基　　金：国家自然科学基金资助项目(40774075);国防基础科研资助项目(A3420060145)

摘　　要：在大量采集放射性地球物理数据的基础上,将神经网络方法引入到应用放射性地球物理方法寻找铀矿的数据处理工作中。在充分分析测量数据的特点后,建立BP模型,结合工作区的地质、地球物理资料,对模型计算的结果进行解释和对比。研究结果表明,将神经网络方法应用于放射性地球物理预测铀矿找矿的远景区域工作中是可行的。On the basis of collecting a large number of radioactive geophysical data, a neural network method was introduced into the data-processing work of radioactivity prospecting of uranium deposits. After the characteristics of the measurement data were adequately analyzed, a back-propagation（BP） model was established. Combined with the geological and geophysical information of a study region, the results calculated by the model were explained and compared. The results show that it is feasible to use the neural network method for forecasting the perspective region of radioactivity prospecting of uranium deposits.

关 键 词：铀矿找矿 BP神经网络 放射性勘探数据 

分 类 号：P631.6[天文地球—地质矿产勘探] TP183[天文地球—地质学]