Camera-to-coin distance finder and comparing their values

Question

Batuhan Istanbullu 2022-12-9

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1874767-camera-to-coin-distance-finder-and-comparing-their-values

评论： Batuhan Istanbullu 2023-2-5

Hey,

I'm just wondering that is there any chance to do that analysis which to my problem it is Image Segmentation Tutorial analysis of finding its circular are I mean diameter of any circle of the same object in different pictures?, like same coin but its getting away from the camera and in order to detect that distance I just want it to detect the radius change and make a scale for it. Is there any chance that I can do this with a similar code like yours?, Please look at my two examples; I want to calculate their area or radius and compare them with each other so that I can detect the change and find the distance irl with a scale. It would mean a lot if you just answer me back.

Best Regards,

Batuhan

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Image Analyst 2022-12-10

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1874767-camera-to-coin-distance-finder-and-comparing-their-values#answer_1124587

Just measure the width of a known thing, like the square or circle, in each image. You can then make up a look up table of working distance (scene-to-lens distance) vs. size in pixels. From that you can fit a function to get you the mm per pixel for any distance. Then when you plug in the known distance, you can measure the size in pixels and report the size in mm.

What was the working distance for each of the two images?

22 个评论
显示 20更早的评论隐藏 20更早的评论

Image Analyst 2022-12-11

在 MATLAB Online 中打开

IMG_4461.JPG

The problem is you're not really thinking about what the algorithm needs to do and you're just randomly tossing in useless operations, like drawing a line and converting to gray scale. Also, you're not displaying the intermediate images so you don't really know what you did last and what you need to do next.

Here, try this:

rgb = imread('IMG_4461.jpg');
% Display image.
subplot(2, 2, 1);
imshow(rgb);
impixelinfo;
title('Original RGB Image')
[r, g, b] = imsplit(rgb);
% Display image.
subplot(2, 2, 2);
imshow(b);
impixelinfo;
title('Blue Channel')
% Threshold blue channel, or use the Color Thresholder app if your fingers are going to be in there.
mask = b < 140;
% Display image.
subplot(2, 2, 3);
imshow(mask);
impixelinfo;
title('Initial Mask Image')
% Get rid of dark surround.
mask = imclearborder(mask);
% If you know there is only one blob, take the largest one.
mask = bwareafilt(mask, 1);
% Display image.
subplot(2, 2, 4);
imshow(mask);
impixelinfo;
title('Final Mask Image')
% Get area
props = regionprops(mask, 'Area', 'EquivDiameter');
area = props.Area
diameter = props.EquivDiameter

area =

235361

diameter =

547.422078919377

If you want you could also ask for Centroid and then use viscircles to put the circle at the detected location in the overlay above the image.

Now I don't know what this distance (paper-to-lens distance) this was (Looks like more than 50 mm) but then you'd have to do this for several distances to be able to create a formula that gave the lateral spatial resolution factor (in mm per pixel).

Batuhan Istanbullu 2022-12-11

Well, sir you have no idea what you have just did. Thank you so much! Also do you mind If I ask something else too?, if you don't mind; the below code gives me the exact positions of the red images that I wanted to take into account, Please look at the images that I've just uploaded. I did found the exact positions of the things but I wanted to find their areas. Can I get them from workspace?

I = imread('IMG_4461.jpg');

J = imrotate(I,180);

imshow(J);

% imtool (rgb) ;

d = drawline;

pos = d.Position;

diffPos = diff(pos);

diameter = hypot(diffPos(1),diffPos(2));

gray_image = rgb2gray(J);

imshow(gray_image)

imshow(J);

% Now to track red objects in real time

% we have to subtract the red component

% from the grayscale image to extract the red components in the image.

diff_im = imsubtract(J(:,:,1), rgb2gray(J));

%Use a median filter to filter out noise

diff_im = medfilt2(diff_im, [3 3]);

% Convert the resulting grayscale image into a binary image.

diff_im = im2bw(diff_im,0.18);

% Remove all those pixels less than 11000px

diff_im = bwareaopen(diff_im,11000);

% Label all the connected components in the image.

bw = bwlabel(diff_im, 8);

% Here we do the image blob analysis.

% We get a set of properties for each labeled region.

stats = regionprops(bw, 'BoundingBox', 'Centroid');

% Display the image

imshow(J)

hold on

%This is a loop to bound the red objects in a rectangular box.

for object = 1:length(stats)

bb = stats(object).BoundingBox;

bc = stats(object).Centroid;

rectangle('Position',bb,'EdgeColor','r','LineWidth',2)

plot(bc(1),bc(2), '-m+')

a=text(bc(1)+15,bc(2), strcat('X: ', num2str(round(bc(1))), ' Y: ', num2str(round(bc(2)))));

set(a, 'FontName', 'Arial', 'FontWeight', 'bold', 'FontSize', 12, 'Color', 'yellow');

end

hold off

Image Analyst 2022-12-11

Seems like you ignored everything I did and threw in useless stuff again. OK, whatever. That's not how I'd do it but you can do whatever you want.

I don't understand the grammar of "which area or diameter is taking into account the final mask one". What I am doing is taking the largest blob only and getting the area of that and the "Equivalent Circular Diameter", in other words the diameter of a circle that has the same area as the blob. It's appropriate for roughly circular things, not for squares and rectangles though. The area and diameter have units of pixels because we have not specified the field of view or real world size of the blobs. That is what you were going to do. You were going to put the camera at one distance, measure the diameter, then move the camera to a different distance and measure the diameter again, then at a different distance, and a different one, until you have maybe 5 or 10 lens-to-scene distances. So then you'd have a "pixelsPerMm" spatial calibration factor for every distance. Then you can fit those to a polynomial and be able to get the pixelsPerMm for any distance.

Batuhan Istanbullu 2023-2-5

Dear Image Analyst,

I have been using the same code that you have shared with me but for whatever reason I can't find the largest blob in the picture. Can you help me where I'm doing wrong?

rgb = imread('A3.jpeg');

J = imrotate(rgb,180);

subplot(2, 2, 1);

imshow(J);

impixelinfo;

title('Original RGB Image')

[r, g, b] = imsplit(J);

subplot(2, 2, 2);

imshow(r);

impixelinfo;

title('Blue Channel')

mask =r <110 ;

subplot(2, 2, 3);

imshow(mask);

impixelinfo;

title('Initial Mask Image')

mask = imclearborder(mask);

mask = bwareafilt(mask, 1);

subplot(2, 2, 4);

imshow(mask);

impixelinfo;

title('Final Mask Image')

% Get area

props = regionprops(mask, 'Area', 'EquivDiameter', 'BoundingBox', 'Centroid');

area = props.Area;

diameter = props.EquivDiameter;

hold on

for object = 1:length(props)

bb = props(object).BoundingBox;

bc = props(object).Centroid;

rectangle('Position',bb,'EdgeColor','r','LineWidth',2)

plot(bc(1),bc(2), '-m+')

a=text(bc(1)+15,bc(2), strcat('X: ', num2str(round(bc(1))), ' Y: ', num2str(round(bc(2)))));

set(a, 'FontName', 'Arial', 'FontWeight', 'bold', 'FontSize', 12, 'Color', 'yellow');

end

Image Analyst 2023-2-5

在 MATLAB Online 中打开

This should do it:

mask = bwareafilt(mask, 1);
subplot(2, 2, 4);
imshow(mask);

If all you see is black, then there is nothing in the original mask image -- no blobs at all.

Batuhan Istanbullu 2023-2-5

Thank you for your help!

请先登录，再进行评论。

Camera-to-coin distance finder and comparing their values

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

采纳的回答

22 个评论
显示 20更早的评论隐藏 20更早的评论

更多回答（0 个）

另请参阅

类别

标签

产品

版本

Community Treasure Hunt

Camera-to-coin distance finder and comparing their values

0 个评论 显示 -2更早的评论隐藏 -2更早的评论

采纳的回答

22 个评论 显示 20更早的评论隐藏 20更早的评论

更多回答（0 个）

另请参阅

类别

标签

产品

版本

Community Treasure Hunt

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

22 个评论
显示 20更早的评论隐藏 20更早的评论