The error "Matrix dimensions must agree"

1 回表示 (過去 30 日間)
Ashbub
Ashbub 2018 年 4 月 17 日
コメント済み: Ashbub 2018 年 4 月 17 日
Hi all
I am getting this error
"Matrix dimensions must agree.
Error in TwoDTensorRev4 (line 106)
TbinAreaSum= double(TbinAreaSum)+double(binAreaSum);"
Here is the code and the input file is attached as well. Please suggest. Thank you.
s=load('C20-28D-150kPa-voidForFabricStudyXZ.mat');
fieldname=fieldnames(s)
m=s.Avizo_C20_28D_150kPa_voidForFabricStudyXZ_mat;
sm=squeeze(m);
sm = bwareaopen(sm,2,4);
szdim=size(sm,3);
slices=szdim;
slices=200;
pieslice=25;
indcs=zeros(size(sm,1),size(sm,2));
porientd=zeros(slices,1);
num=zeros(slices,1);
perea=zeros(slices,1) ;
avgperea=zeros(slices,1);
edges=zeros(1,pieslice,1);
sumareapd=zeros(slices,1);
TbinAreaSum=zeros(pieslice-1,1);
for jj=1:slices
[L2,num(jj,1)] = bwlabel(sm(:,:,jj));
areap=zeros(num(jj,1),1);
areapd=zeros(2.*num(jj,1),1);
areapactual=zeros(num(jj,1),1);
binAreaSum=zeros(pieslice-1,1);
perip=zeros(num(jj,1));
vector=zeros(num(jj,1),2);
vectord=zeros(2.*num(jj,1),2);
cent=zeros(num(jj,1),1);
hbin=zeros(1,2.*num(jj,1));
thetap=zeros(1,2.*num(jj,1));
sumperea(jj)=0;
for ii=1:num(jj,1)
areao=regionprops((L2==ii),'area');
perio=regionprops((L2==ii),'perimeter');
areap(ii,1)=areao.Area./(areao.Area);
areapactual(ii,1)=areao.Area;
perip(ii,1)=perio.Perimeter;
perea(ii,jj)=perip(ii,1)./areapactual(ii,1);
coeff=pca(indcs);
vector(ii,1)=coeff(2,1);
vector(ii,2)=coeff(1,1);
sumperea(jj)=sumperea(jj)+perea(ii,jj);
end
sumareapd(jj,1)=2.*sum(areap(:,1));
avgperea(jj,1)=sumperea(jj)./num(jj);
vectord=[vector;-vector];
areapd=[areap;areap];
for mm=1:2.*num(jj,1)
thetap(mm)=atan2d(vectord(mm,2),vectord(mm,1));
if (thetap(mm) < 0)
thetap(mm)=360+thetap(mm);
else
thetap(mm)=thetap(mm);
end;
end;
for nn = 1:pieslice
edges(1,nn)=(nn-1).*pi./((pieslice-1)./2.0);
end
hbin = discretize(round(deg2rad(thetap),10),round(edges,10));
binAreaSum=accumarray(hbin',areapd);
TbinAreaSum= TbinAreaSum+binAreaSum;
end
  2 件のコメント
KSSV
KSSV 2018 年 4 月 17 日
Getting different error on running the code:
Undefined function or variable 'indcs'.
Error in Junk (line 38)
coeff=pca(indcs);
Ashbub
Ashbub 2018 年 4 月 17 日
Thankyou. I have defined the indcs. Please have a look.

サインインしてコメントする。

採用された回答

Walter Roberson
Walter Roberson 2018 年 4 月 17 日
pieslice=25;
TbinAreaSum=zeros(pieslice-1,1);

so TbinAreaSum is (25-1)=24 by 1. You then do not change the contents until the last line where you try

TbinAreaSum= TbinAreaSum+binAreaSum; 

binAreaSum was created as

binAreaSum=zeros(pieslice-1,1);

which is also 24 x 1. But you write all over it, in the second last line,

binAreaSum=accumarray(hbin',areapd); 

so binAreaSum will only be 24 x 1 if hbin' happens to have 24 unique values in it.

Now you create hbin as

hbin = discretize(round(deg2rad(thetap),10),round(edges,10));

after

for nn = 1:pieslice                       
edges(1,nn)=(nn-1).*pi./((pieslice-1)./2.0);
end

so edges has 25 entries -- but it is not immediately clear that round(edges,10) happens to have 24 unique entries. It might be 25 unique entries; it might be less.

You need to put in a breakpoint before the accumarray and check length(unique(hbin))

  1 件のコメント
Ashbub
Ashbub 2018 年 4 月 17 日
Thanks a lot. You've hit the right spot. Used histcounts instead of discretize and made few changes to remove the need for using accumarray as well. It did the trick.

サインインしてコメントする。

その他の回答 (0 件)

カテゴリ

Help Center および File ExchangeLogical についてさらに検索

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by