index4log_sampling1.m

地震、测井方面matlab代码,解释的比较详细

function index=index4log_sampling1(nsamp,ntr,param)
% Function creates index vector to access log data for "random modeling"
% Creates a stepsize with a distribution approximately proportional to 1/stepsize
% and a direction which changes in such a way that 
% "mean(positive steps - negative steps)" is approximately "avinc".
% Unlike "index4log_sampling", indices that exceed "param.maxindex" are
% "reflectde back (e.g. index "index" > "param.maxindex" is changed
% to 2*param.maxindex - index.
% Written by: E. R.: February 16, 2004
% Last updated:
%
%           index=index4log_sampling1(nsamp,ntr,param)
% INPUT
% nsamp     number of rows of index matrix
% ntr       number of columns of index matrix
% param     structure with parameters to determine the way the indices are computed
%      'mininc'   minimum step from one index to the next (1 or greater)
%      'maxinc'   maximum step
%      'avinc'    average step to take
%      'maxindex'  largst index
% OUTPUT
% index     indices [1,maxindex]

%       Create a sequence of step with lengths between "mnistep" and "maxinc"
steps=ss_random_thicknesses(ntr*nsamp,1,param.mininc,param.maxinc);
steps=round(steps);

%       Revert some of the steps to achieve a average step size of "avinc"
mstep=mean(steps);
alpha=(param.avinc/mstep+1)/2;
rnd=rand(ntr*nsamp,1);
idx=rnd > alpha;
steps(idx)=-steps(idx);
index=cumsum(steps);
maxindex=param.maxindex;
temp=round(maxindex*rand(1,ntr));
index=reshape(index,nsamp,ntr)+temp(ones(nsamp,1),:);

index=mod(index(:),2*maxindex-1)+1;

idx=find(index > maxindex);

index(idx)=2*maxindex-index(idx);

index=reshape(index,nsamp,ntr);