How can find area of the dent detected?

Hi, i have already found the dent in the laminates but i need find the area of the dent

please help me on this

[Samadhan Kshirsagar]

by which method you have found dent?

[Image Analyst]

@Samadhan Kshirsagar, he gave the code below in his answer:

https://www.mathworks.com/matlabcentral/answers/153231-how-can-find-area-of-the-dent-detected#answer_152507

Did you scroll down and see it?

Answer 1

[Image Analyst]

You forgot to attach your image. All I can suggest then is to run my Image Segmentation Tutorial: http://www.mathworks.com/matlabcentral/fileexchange/25157-image-segmentation-tutorial---blobsdemo--

[Image Analyst]

Dent Detection in Car Bodies,
ICPR00(Vol IV: 775-778).
IEEE via DOI 0009
Structured light. BibRef
        TITLE = "Dent Detection in Car Bodies",
        BOOKTITLE = ICPR00,
        YEAR = "2000",
        PAGES = "Vol IV: 775-778",
        BIBSOURCE = "http://www.visionbib.com/bibliography/applicat839.html#TT113769"}

[Image Analyst]

Just call regionprops() if you've already segmented the dents

labeledImage = bwlabel(dentBinaryImage);
measurements = regionprops(labeledImage, 'Area');
allAreas = [measurements.Area]

[vijendra sn]

I want to find out area in millimeter or centimeter..how to find out.

Here is my code;

A = imread('DSC_0252.jpg'); b = rgb2gray(A);

hx = fspecial('sobel'); hy = hx'; Iy = imfilter(double(b), hy, 'circular'); Ix = imfilter(double(b), hx, 'circular');

gradmag = sqrt(Ix.^2 + Iy.^2);

k=figure,imshow(gradmag,[]); set(k, 'visible','off'); filename = 'temp_file.jpg'; saveas(k, filename);

i1 = imread(filename); delete(filename);

[x, y, rgb] = ind2sub([size(i1,1) size(i1,2) size(i1,3)], find(i1 ~= 255));

C = i1(min(x):max(x)-1,min(y):max(y)-1,:);

C = rgb2gray(C); figure,imshow(C); I = edge(C,'sobel');

I = imcrop(I,[10 7 920 725]); I = imresize(I, [3500 3700]);

f=figure,imshow(I); set(f, 'visible','off'); imwrite(I, filename);

bw = imread(filename); delete(filename);

bw1 = bwareaopen(bw,1100); bwfill=imfill(bw1,'holes'); boundary=bwboundaries(bwfill);

figure,imshow(I),hold on;

[row1,dim]=size(boundary);

for k=1:row1 x1=boundary{k}; h=plot(x1(:,2),x1(:,1),'c','LineWidth',1); end

labeledImage = bwlabel(bw1); measurements = regionprops(labeledImage, 'Area'); allAreas = [measurements.Area]

[Image Analyst]

http://www.mathworks.com/matlabcentral/answers/13205-tutorial-how-to-format-your-question-with-markup

Did you get the article yet? Are you going to?

The best method is to use profilometry to get actual topographic information. For example an instrument from GFM http://www.gfm3d.com/index.php?option=com_content&id=65%3Amikrocad&lang=en

I don't think your method is that good or robust. It depends on the size being large, so it can't measure small dents. I didn't run it but I don't see imconvhull() being called and since there is only an edge on one side of the dent, the side with the ramp won't be included. Keep working on it though. As you can imagine once you have a robust program it will probably be hundreds if not thousands of lines long.

[vijendra sn]

I don't know how to use imconvhull() as i don't have image processing tool box

k=figure,imshow(gradmag,[]);

set(k, 'visible','off');
filename  = 'temp_file.jpg';
saveas(k, filename);

i1 = imread(filename);
delete(filename);

In this small code, i am saving the image and reading the image again.I don't want to save and read as it takes too much time for processing. Is there any method to develop....

[Image Analyst]

I don't know why you want i1 to be a screenshot. What are you going to do with a screenshot of the figure? What good is that instead of the actual image that is displayed in the figure? Get the Image Processing Toolbox. You really don't want to rewrite it yourself and there are too many useful functions in there for getting boundaries, making measurements, doing size filtering and connected components labeling, displaying images, etc.

[vijendra sn]

Hi,

I am not getting the correct functions. can u tell me, how to code for above small code

[Image Analyst]

Isn't "allAreas" the areas that you're looking for? Remember to do

bw1 = bwconvull(bw1);

like I suggested before you label the image.

[vijendra sn]

i am getting error for this code:

A = imread('DSC_0977.jpg');
B = rgb2gray(A);
hx = fspecial('sobel');
hy = hx';
Iy = imfilter(double(B), hy, 'circular');
Ix = imfilter(double(B), hx, 'circular');

gradmag = sqrt(Ix.^2 + Iy.^2);
k=figure(1),imshow(gradmag,[]);
set(k, 'visible','off');
filename  = 'temp_file.jpg';
saveas(k, filename);

i1 = imread(filename);
delete(filename);

[x, y, rgb] = ind2sub([size(i1,1) size(i1,2) size(i1,3)], find(i1 ~= 255));

C = i1(min(x):max(x)-1,min(y):max(y)-1,:);

C = rgb2gray(C);
I = edge(C,'sobel');

I = imcrop(I,[10 7 920 725]);
I = imresize(I, [3250 3700]);

f=figure(3),imshow(I);
set(f, 'visible','off');
imwrite(I, filename);

bw = imread(filename);
delete(filename);

bw1 = bwareaopen(bw,1500);
bwfill = imfill(bw1,'holes');
boundary = bwboundaries(bwfill);

figure(2),imshow(I),hold on;
[row1,dim]=size(boundary);

for k=1:row1
    x1=boundary{k};
    h=plot(x1(:,2),x1(:,1),'c','LineWidth',1);
end
bw1 = bwconvull(bw1);
labeledImage = bwlabel(bw1);
measurements = regionprops(labeledImage,'Area');
allAreas = [measurements.Area];

ERROR IS: Undefined function or method 'bwconvull' for input arguments of type 'logical'

[Image Analyst]

Well how old is your version? bwconvhull() is several years old but if you have an antique version, then you may not have that function and should upgrade.

[vijendra sn]

My Matlab version is 7.10.0.499(2010a).

[Image Analyst]

Please upgrade. What happens when you say "which bwconvhull"?

Answer 2

[vijendra sn]

Hi,

i used bwconvhull function...how to analyze this function but still i couldn't get how to find area of the these dent detected. There are 9 dent area detected.

How to find the area of the 1 pixel in image

close all;
clear all;
clc;

A = imread('DSC_0977.jpg');
B = rgb2gray(A);

hx = fspecial('sobel');
hy = hx';
Iy = imfilter(double(B), hy, 'circular');
Ix = imfilter(double(B), hx, 'circular');

gradmag = sqrt(Ix.^2 + Iy.^2);
k=figure(1),imshow(gradmag,[]);
set(k, 'visible','off');
filename  = 'temp_file.jpg';
saveas(k, filename);

i1 = imread(filename);
delete(filename);

[x, y, rgb] = ind2sub([size(i1,1) size(i1,2) size(i1,3)], find(i1 ~= 255));

C = i1(min(x):max(x)-1,min(y):max(y)-1,:);

C = rgb2gray(C);
I = edge(C,'sobel');

I = imcrop(I,[10 7 920 725]);
I = imresize(I,[3250 3500]);

f=figure(3),imshow(I);
set(f, 'visible','off');
imwrite(I, filename);

bw = imread(filename);
delete(filename);

bw1 = bwareaopen(bw,750);
bwfill = imfill(bw1,'holes');
boundary = bwboundaries(bwfill);

figure(2),imshow(I),hold on;
[row1,dim]=size(boundary);

for k = 1:row1
    x1= boundary{k};
    h = plot(x1(:,2),x1(:,1),'c','LineWidth',1);
    bw2 = bwconvhull(boundary{k});
    [label,Total]=bwlabel(bw2);
    measurements = regionprops(label,'Area');
    allareas(k,1)= [measurements.Area];
end