Reading a sequence of image frames (bmp) and storing it as a 4-D array

I need to read a sequence of image frames (bmp) and store it as a 4-D array (mXnX3Xp) where p=no. of frames. When I try to do this using an array named 'image', I don't get an error but get an 1-D array instead. Checking size(image) returns pX1.

My code for reading a single frame is:

A=imread('cropped_sequence0000.bmp'); %All images are named as cropped_sequence0000 to 0999
B=A(:,:,3); %Extract blue channel only
C=size(B);
d=C(:,2); %Find width of each frame
h=C(:,1); %Find height of each frame

if mod(d,2)==0 %Take half-width for analysis
 f=d/2;
else
 f=(d+1)/2;
end

if mod(h,2)==0 %Take half-height for analysis
 i=h/2;
else
 i=(h+1)/2;
end

G=B([i:h],f); %Read desired area of interest
P=size(G);
q=P(:,1);
J=[1:q];

[k,l]=min(G(:)); %Find extrema of blue pixels
[m,n]=max(G(:));

plot(J,G)

Now, I read the entire directory as follows:

for i=1:999
image{i}=imread(sprintf('cropped_sequence%.4d.bmp',i));
end

I expected image to be a 4-D array but it turns out to be an 1-D array of size 999. I want to have a 4-D array that reads all the individual frame information by reading each frame in tandem as in the single frame code above. Finally, I need to plot the locations of the extrema of blue pixels versus frame number.

[Rena Berman]

(Answers dev) restored question

Answer 1

[Image Analyst]

What are you doing plotting G - the blue channel image??? To create a 3D image you use cat():

if imageNumber = 0
  image4D = G; % First image - nothing to concatenate onto yet.
else
  image4D = cat(3, image4D, G); % Concatenate an image onto the 4D array.
end

[Image Analyst]

Well that's better because you're not plotting the image but there are still a number of problems. For one (the most serious), you're overwriting max and min() functions when you do this:

min=minimum(:,2); %location of the maxima and not the pixel value itself
max=maximum(:,2);

Secondly you're not indexing maximum and minimum with q so you're overwriting the came value at every single iteration of q.

Try something like this

totalNumberOfFrames = size(image4D, 4);
minLocations = zeros(1, totalNumberOfFrames);
maxLocations = zeros(1, totalNumberOfFrames);
for frameNumber = 1 : totalNumberOfFrames 
  % Extract blue channel for this frame.
  blue = image4D(:,:,3, frameNumber);
  % Extract line segment for this blue channel.
  centerLine = blue(i::h, f); % Get line from row i to row h at column f.
  % Find the index where the min and max are.
  % Be sure to add back in the offset i!!!
  % If max was at i, min will be at element 1 so need to add i-1 to it.
  [~, minLocations(frameNumber)] = min(centerLine) + i - 1;
  [~, maxLocations(frameNumber)] = min(centerLine) + i - 1;
end
f = 1 : totalNumberOfFrames;
plot(f, minLocations, 'r-', 'LineWidth', 3);
hold on;
plot(f, maxLocations, 'b-', 'LineWidth', 3);
legend('Min', 'Max');
grid on;

[Walter Roberson]

j = 0;
fname = sprintf('cropped_sequence%.4d.bmp',j);
image4d = imread(fname);
image4d(:,:,:,601) = 0;
for j = 1 : 600
  fname = sprintf('cropped_sequence%.4d.bmp',j);
  image4d(:,:,:,j+1) = imread(fname);
end

[Image Analyst]

I called it G because that's what you called your image in your original code. But you're not using that anyway. You're now calling the image temp_img, and that's fine. It looks like your latest code should construct the 4D image properly once you change image4d into image4D - MATLAB is case sensitive. And change this line:

mat_plot(j)=temp_img;

to

mat_plot = temp_img;

or just get rid of mat_plot altogether and use only temp_img.

[Walter Roberson]

And change "for j = 0" to "if j == 0"