形成一個64×64矩陣的巨陣列

Question

我正在使用matlab進行視頻鏡頭邊界檢測。這裏是我的代碼片段

clc; 
clear all; 

[file path] = uigetfile('*.avi','Pick a Video');% this line opens a pop up window so that you can choose your video for further processing  
video = VideoReader([path file]);% the video chosen through UI get stored in video object 
numOfFrames = video.NumberOfFrames;% this will be a number (total number of frames in your chosen video) 
%allFrames = read(video);% catching all the frames from the video in object allFrames 
for i=1:1:numOfFrames 
    try 
     lFrame = read(video, i);%frames(:, :, :, i); 
    catch 
     break; 
    end 
    lRFrame = lFrame(:,:,1); 
    lGFrame = lFrame(:,:,2); 
    lBFrame = lFrame(:,:,3); 
    lGray = 0.299*lRFrame + 0.587*lGFrame + 0.114*lBFrame; 
    grayImage = imresize(lGray, [256 256]); 

    % step 2 of the algorithm begins from here % 
    meanFilterFunction = @(theBlockStructure) mean2(theBlockStructure.data(:)) * ones(1,1, class(theBlockStructure.data)); 

    blockSize = [4 4]; 
    blockyImage4 = blockproc(single(grayImage), blockSize, meanFilterFunction); 
end 

我得到一個64×64圖像中blockyImage4但因爲它是在一個循環中，每次刷新值本身很下一次迭代。

我想要一個可以存儲（可能是所有）圖像（矩陣）的圖像數組。我應該可以訪問它們。如果所有圖像都不能存儲在數組中，那麼如何只存儲連續迭代的兩幅圖像，因爲我需要「i」和「i + 1」圖像來找出它們之間的差異。請幫忙

Answer 1

想要在Matlab中存儲3D（x，y，time）數據嗎？無論是在單元陣列還是3D陣列中。對於陣列，不要忘記預先分配所需的大小以避免性能下降。

因此，要麼：

blockyimages = cell(numOfFrames, 1); 
for i=1:1:numOfFrames 
    ... 
    blockyimages{i} = blockproc(single(grayImage), blockSize, meanFilterFunction); 
end 
% access the blocky image of the ith frame with blockyimages{i} 
% you could even do blockyimages = cat(3, blockyimages{:}) to convert to a 3d matrix 

或

blockyimages = zeros(64, 64, numOfFrames); % better frame sizes/block sizes 
for i=1:1:numOfFrames 
    ... 
    blockyimages(:, :, i) = blockproc(single(grayImage), blockSize, meanFilterFunction); 
end 
% access the blocky image of the ith frame with blockyimages(:, :, i) 

所以，如果你選擇了第二個選項，以後要再做一次與blockyimages的東西，只是讓通過循環第二遍。例如，如果你想計算連續blockyimages的差異。

diffblockyimages = zeros(64, 64, numOfFrames - 1); 
for i = 1 : numOfFrames - 1 
    framei = blockyimages(:, :, i); 
    frameiplusone = blockyimages(:, :, i+1); 
    diffblockyimages(:, :, i) = frameiplusone - framei; 
end 

雖然同樣可以在差別的特殊情況下，已經實現也爲

diffblockyimages = diff(blockyimages, 1, 3); 

Answer 2

由於我的評論被阻止我提出可能的解決方案：
測試，對我來說

工作正常

file_path = uigetfile('*.avi','Pick a Video');% this line opens a pop up window so that you can choose your video for further processing  
video = VideoReader(file_path);% the video chosen through UI get stored in video object 
numOfFrames = video.NumberOfFrames;% this will be a number (total number of frames in your chosen video) 
%allFrames = read(video);% catching all the frames from the video in object allFrames 

all_frames = cell(1,numOfFrames); 
for i=1:numOfFrames %default step is always 1 
    try 
     lFrame = read(video, i);%frames(:, :, :, i); 
    catch 
     break; 
    end 
    lGray = rgb2gray(lFrame); 
    grayImage = imresize(lGray, [256 256]); 
all_frames{1,i} = grayImage; 

end 

現在你有all_frames數組中的所有幀 訪問很簡單：

img1 = all_frames{1,1}; imshow(img1); 
img2 = all_frames{1,2}; imshow(img2);