Question

我使用以下代码校准我的相机。但我没有得到任何适当的输出。我想在实验前校准我的GoPro Hero4 Black相机。

我首先使用了检查板，但我找不到相关的代码，所以我正在使用这个设计。我没有计算机视觉工具箱。所以我只使用图像处理工具箱。

 clear all;
    close all;
    clc


    tp=0.2; % tp is the time duration between the images shown
    [I]=ChargementSeqIm('Go','.bmp',5,0);
    % 5 is the total number of images (PP are already processed)
        figure(1);
        for c=1:5
            imshow(I(:,:,c),[]);colorbar;title(sprintf('Image number n %d',c));pause(tp);
        end

    %% part II.1 thresholding and noise reduction
    % II.1. Applied threshold (0.8) in order to get a BW image
    figure(2);
    for c=1:5
        bw(:,:,c)=im2bw(uint8(I(:,:,c)),0.8); %thresholding (0.8)
        imshow(bw(:,:,c),[]);colorbar;title(sprintf('Image number n %d',c));pause(tp);
    end
    % Applied morphological operations in order to remove the noise in the image
    se = strel('disk',4);
    figure(3);
    PTIS=bw;
    for c=1:5
        PTIS(:,:,c)=imclose(PTIS(:,:,c),se);
        imshow(PTIS(:,:,c),[]);colorbar;title(sprintf('Image number n %d',c));pause(tp);
    end

    %% II.2. Obtain the labels of the image objects
    figure(3);
    imgLabel=PTIS;
    for c=1:5
        L(:,:,c)=bwlabel(PTIS(:,:,c));
        imshow(L(:,:,c),[]);colormap(hsv);colorbar;title(sprintf('Image number n %d',c));pause(tp);
    end

    %% II.3. Detect the centroids of each object and locate it. Finally display
    figure(4);
    for c=1:5
        objCnt=regionprops(L(:,:,c),'Centroid');
        centres(:,:,c) = cat(1, objCnt.Centroid);
        imshow(PTIS(:,:,c),[]);colorbar;
        hold on; plot(centres(:,1,c),centres(:,2,c),'+');
        plot(centres(:,1,c),centres(:,2,c),'m');title('Images of patterns + detected centroids');pause(tp);
    end

    %% Task - III : Arrange a matrix of 3D coordinates and 
    % a matrix of 2D coordinates of the projected points in the image.

    %% III.2. Build Matrix uivi
    %figure(5)

    uivi=[]
    for c = 1:5
            uivi = [uivi;centres(:,:,c)];

    end
    % --> TODO <-- %
    %% III.3. Build Matrix XiYiZi (use the information from the file: information_calibration_pattern.txt)
    xyz=[];
    k=1;
    z=[0 17 30 43 57];% z vale from the .txt file
    for c=1:5
        for d=0:7
            for e=0:5
                XYZ(k,:)=[d*17,e*17,z(c)];
                k=k+1;
            end
        end
    end


    figure(5);
    plot3(XYZ(:,1),-XYZ(:,2),XYZ(:,3),'+');hold on;plot3(XYZ(:,1),-XYZ(:,2),XYZ(:,3),'m');xlabel('X(mm)');ylabel('-Y(mm)');zlabel('Z(mm)');title('Amers dans le repère mire (Xi,Yi,Zi)');
    % Build and save matrix XYZuv to compute the calibration matrix
    XYZuv=[XYZ,uivi];
    save ('XYZuv.mat', 'XYZuv');

    %% IV. Matrix calibration computation

    % IV.1 Matrix calibration computation Call functions createBC() and createM() to compute the calibration matrix M.
    [B,C]=createBC(XYZuv);
    M=createM(B,C);

    % IV.2 Check the contains and display your comment accordingly
     Check = sqrt(M(3,1)^2+M(3,2)^2+M(3,3)^2);
     fprintf('Faugeras and Toscani constraint is used')

    %% V. Validity of the calibration matrix

    % V.1. Re-projection
    % Create homogeneous coordinates
    XYZ1=[XYZ';ones(1,size(XYZ,1))];

    % Re-project and scale appropriately
    u_v=[M*XYZ1] 
    % --> TODO <-- %
    u_v(:,1)=u_v(:,1)./u_v(:,3);
    u_v(:,2)=u_v(:,2)./u_v(:,3);
    u_v(:,3)=u_v(:,3)./u_v(:,3);

    %% V.2. Display
    uivia=uivi'; % Actual 2D coordinates
    uivir=u_v;
    N=48;


    %figure;
    for c=1:5
        imshow(I(:,:,c),[]);colorbar;
        hold on; plot(uivia(1,1+(c-1)*N:c*N),uivia(2,1+(c-1)*N:c*N),'+');
        plot(uivir(1,1+(c-1)*N:c*N),uivir(2,1+(c-1)*N:c*N),'*m');title('Images of patterns + amers reprojetés + detected landmarks');pause(0.5)
    end


    figure (6);
    imshow(I(:,:,1),[]);colorbar;
    for c=1:5
        hold on; plot(uivia(1,1+(c-1)*N:c*N),uivia(2,1+(c-1)*N:c*N),'+');
        plot(uivir(1,1+(c-1)*N:c*N),uivir(2,1+(c-1)*N:c*N),'*m');title('Images of patterns + detected landmarks');pause(0.5)
    end

    %% V.3. Compute the mean difference and the standard deviation of the distribution of differences and save it

    ecarts=[];
    % Compute the re-projection error / distance
    first=uivia';
    second=uivir(1:2,:)';
    distance=first-second;
    ecarts=sqrt((distance(:,1)).^2+(distance(:,2)).^2);


    figure (7);
    hist(ecarts,24);title('Histogram of differences')

    % compute average and standard deviation of the error/distance
    average = mean(ecarts);
    standard = std(ecarts);

    %% V.4. Intrinsic and extrinsic parameters
    IntrinsicExtrinsicParameters

如何校准此图像

0 个答案: