Save 360 arrays in a cell

Hello,

My programs loads 360 images (laser stripe) and from each image gets the 0 crossing of all rows. My issue is that after processing all images, im not able to save the 0 crossing arrays into a cell so I can post process them. Here my code:

clc; clear all; close all;
%% CARGAR LAS IMAGENES (load images)

for k = 1:360
  jpgFilename = sprintf('%d.jpg', k);
  fullFileName = fullfile('C:\MASTER\S2\percepcion\laser360\imagenes', jpgFilename);
  if exist(fullFileName, 'file')
    imageData = imread(fullFileName );
    imageData=imrotate(imageData,90); 
    
 % se binariza la imagen (binarization)
    img_o=rgb2gray(imageData);
    mask=(img_o(:,:) > 50);
    imageData = bsxfun(@times, img_o, cast(mask, 'like', img_o));
%     figure()
%     imshow(imageData)
%     title('Binarizada')

% 
[rows, columns, numberOfColorChannels] = size(imageData);

img_peaks = NaN(rows, 1);

img_puntos = zeros(rows, columns);
Xp = [];
Yp = [];
Xn=[];
Yn=[];
m=[];
n=[];
zerocross=[];
coefficients=[];
lineaslaser=cell(k,1);

%se busca la linea del 0 crossing 

for icross=1:rows
    
    img_puntos(icross,:)=imageData(icross,:);
    %filtrado
    filtrado=sgolayfilt(double(img_puntos(icross,:)),3,33 );
    for ks=1:length(filtrado)
        if filtrado(ks)<0
       filtrado(ks)=0;
        end
    end
    %derivada (gradient)
    derivada=gradient(filtrado);
    %maximo y minimo
    
    [Yp(icross),Xp(icross)]=max(derivada);
    [Yn(icross),Xn(icross)]=min(derivada);
    
     %coeficientes de la recta (line coefs)
     m(icross)= (Yp(icross)-Yn(icross))/(Xp(icross)-Xn(icross));
     n(icross)= (Yp(icross)-m(icross)*Xp(icross));

     % 0 crossing
     zerocross(icross)= -n(icross)/m(icross);
     img_peaks(icross) = zerocross(icross);
end

axis=1:1920;
img_peaks((1:10),1)=NaN;
img_peaks((1500:1920),1)=NaN;

% save 360 lines in a cell
lineaslaser{k}=img_peaks(k);

% figure()
% plot(axis,img_peaks)
% title('plot perfil de la pieza (0 crossing)')

  else
    warningMessage = sprintf('Warning: image file does not exist:\n%s', fullFileName);
    uiwait(warndlg(warningMessage));
  end
end

[DGM]

What exactly is the problem? Should I guess?

You're saving a single element from a numeric vector into a cell vector. Are you expecting to be saving the entire vector?

lineaslaser{k}=img_peaks(k);

Considering that you do this right before that assignment, the first 10 entries in lineaslaser will be scalar NaNs.

img_peaks((1:10),1)=NaN;

But none of that really matters, because every time the loop cycles, you delete the entire array:

lineaslaser=cell(k,1);

If you know the size of the arrays needed, preallocate them as the required size instead of growing them in a loop.

Xp = [];
Yp = [];
Xn=[];
Yn=[];
m=[];
n=[];

[mikel lasa]

Yes, I want to save an entire vector in a cell for each loop. There is my problem, for each loop a vector is generated, so my aim is to create a variable that saves 360 vectors. I thought that cell is maybe the right way to do that.

Apart from that:

img_peaks((1:10),1)=NaN;

The first 10 elements are NaN because each image has some noise, I delete that noise this way. Its just, for cleaning.

[DGM]

Okay. If you move the preallocation outside the main loop and set that line to

lineaslaser{k}=img_peaks;

does it help?

Answer 1

[mikel lasa]

Edit:

I have achieved what I was looking for. As you said I deleted this line:

lineaslaser=cell(k,1);

and I changed my this line:

lineaslaser{k}=img_peaks(k);

for this one:

lineaslaser{k}=img_peaks(:,1);

now works, I have a 360 element cell where each cell is a 1920x1 vector. Thanks for your help!