| The geophysical prospecting,the abbreviation physical prospecting,is the essential method to study the geologic structure and solution of prospecting exploration question.It takes each kind of physical property and so on rock and ore density,magnetism,electric properties,elasticity,radioactivity differences as research foundation.And,the gravitational exploration,the magnetic prospecting are two main physical prospecting methods used presently.One of region heavy magnetism data processing's goals is through to region geophysics field each method processing and the analysis,the inference studies in the region the structure break distribution.This process may through the analytic continuation, the horizontal direction derivative and so on,carry on the synthesis interpretation with other geophysics material to obtain each kind of related depth portion structure and the break distributed information willfully.Then according to the geologic body between differences,we can explain the inference existence faulted structure and the ring structure. To the region heavy magnetism data takeoff as well as in the processing process,the artificial intelligence algorithm is applied massively in which,solves such as judgment spool thread related and so on non-linear problems.At present,the artificial intelligence model applied quite.maturely is the BP artificial intelligence network model.This study has studied the traditional BP artificial intelligence network model in heavy magnetism data processing application,and summarized traditional BP algorithm some flaws and the insufficiency,to other artificial intelligence model's research comparison,has made the improvement to the traditional BP artificial intelligence model,this new method final experiment certificate effective overcame the BP algorithm easily to fall into the partial minimum point,thus caused the network training defeat's shortcoming,moreover the BP network convergence rate was slow,the training time long question was also improved.
|
PDF Full Text Request