How I can make automatic program to detect the spots!

Question

Khaled Al-Faleh 2017-3-12

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/329455-how-i-can-make-automatic-program-to-detect-the-spots

评论： Khaled Al-Faleh 2017-3-13

Hi everyone, I have a problem which is I have a small parking and I need to detect the spots automatic rather than manually, in the figure below this my parking and I did a program to take image from my camera and cut it to get each spot alone but the problem if I move the camera I have to change the Coordinates see the code..

data1 = imread('data_car_01.png');
data1_gray = rgb2gray(data1);
A1 = imcrop(data1_gray,[140 0 240 300]);
subplot(3,3,2);
imshow(A1);
% % % Second Spot Picture 1
B1 = imcrop(data1_gray,[370 0 230 300]);
subplot(3,3,3);
imshow(B1);
% % % Third Spot Picture 1
C1 = imcrop(data1_gray,[590 0 235 300]);
subplot(3,3,4);
imshow(C1);
% % % Forth Spot Picture 1
D1 = imcrop(data1_gray,[820 0 500 300]);
subplot(3,3,4);
imshow(D1);

so how I can make it in automatic way..

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Khaled Al-Faleh 2017-3-12

data_car_01.png

here the image

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

KSSV 2017-3-13

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/329455-how-i-can-make-automatic-program-to-detect-the-spots#answer_258407

在 MATLAB Online 中打开

data = imread('data_car_01.png');
[r, c, n] = size(data);
% Convert to monochrome.
redPlane = data(:, :, 1);
greenPlane = data(:, :, 2);
bluePlane = data(:, :, 3);
% Find the standard deviation of each color channel.
redStdDev = std(single(redPlane(:)));
greenStdDev = std(single(greenPlane(:)));
blueStdDev = std(single(bluePlane(:)));
% Take the color channel with the highest contrast.
% Transfer it into a monochrome image.  This will be the one that we use.
if redStdDev >= greenStdDev && redStdDev >= blueStdDev
    % Red has most contrast - use that channel.
    monoImage = single(redPlane);
elseif greenStdDev >= redStdDev && greenStdDev >= blueStdDev
    % Green has most contrast - use that channel.
    monoImage = single(greenPlane);
else
    % Blue has most contrast - use that channel.
    monoImage = single(bluePlane);
end
maxValue = max(max(monoImage));
minValue = min(min(monoImage));
monoImage = uint8(255 * (single(monoImage) - minValue) / (maxValue - minValue));
% Threshold to get the walls.  This will also sharpen up blurry, fuzzy wall edges.
thresholdValue = uint8((maxValue + minValue) / 2);
binaryImage = 255 * (monoImage < thresholdValue);
I = binaryImage ;
I(I==0) = 100 ;
I(I==255) = 0 ;
[L,num] = bwlabel(I,4) ;
 vislabels(L)

Use vislabels from the link: https://in.mathworks.com/matlabcentral/fileexchange/19665-visualize-output-of-bwlabel/content/vislabels.m. You can get your region of interest using L==1. L==2 etc.

Please note that the above code is taken from https://in.mathworks.com/matlabcentral/fileexchange/27175-maze-solution by Image Analyst.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Khaled Al-Faleh 2017-3-13

thanks for your answering but sir I want to cut every spot alone in above code it determine the lines and its good but I need each spot alone save in different variable did you get the idea please ??

请先登录，再进行评论。