How to calculate the pixels occured by white color in binary image

After converting the image to binary image and my primary aim is count the area covered by the white color and the black color in the binary image. For Example, we have performed edge detection on the image and we got an output image in binary image. I want to calculate the area of the edge in the image which are in white.

Answer 1

[Image Analyst]

To find the white and black pixels:

numWhitePixels = sum(binaryImage(:));
numBlackPixels = sum(~binaryImage(:));

To get the number of perimeter pixels of the white regions

perimImage = bwperim(binaryImage);
numPerimeterPixels = sum(perimImage(:));

An alternate way to get the area of the white is to use bwarea():

% A weighted sum of white pixels that depends on shape of perimeter.
area = bwarea(binaryImage);

[ABDUL WAHID MOHAMMED]

"sum(binaryImage(:));" what is the syntax for input in this command? sum(filename(?:?)); ?

[KSSV]

Your 2D matrix representing binaryimage. This comes using imread.

[Image Analyst]

You said "After converting the image to binary image...." so I assume you already have a binary image that you got from doing something like

grayImage = imread(filename);
binaryImage = grayImage > 200; % or some threshold.

Now, you may have called it something different, like mask or BW. What did you call it? Just replace my "binaryImage" with whatever you named your image.

[ABDUL WAHID MOHAMMED]

After converting the gray image to binary, How can we supervise a part of image which is in white to black color

I have attached image which is a binary image. I want to supervise the highlighted part image which is in red to the green line. I want to supervise by changing white color to black.

[Image Analyst]

Google: "Supervise: observe and direct the work of (someone)."

Explain your definition of supervise as it relates to doing something to an image, because I have no idea.

[Walter Roberson]

I think "supervise" might mean "control" here.

Unfortunately the image did not get attached.

[Image Analyst]

So, "After converting the gray image to binary" the black and white pixels are what you have. If you want to further process them, some of the main functions for doing that are bwareaopen(), bwmorph(), bwareafilt(), bwpropfilt(), imdilate(), imerode(), bwlabel(), and maybe some others. They all do different things to the image of course.

[ABDUL WAHID MOHAMMED]

My aim is to change a part(a small part nothing else) of binary image which is in white color to the black.I have attached an image in which red area shows the part which need to be changed it the green line

[Image Analyst]

No image was attached. To change a white pixel to black set it to false.

[ABDUL WAHID MOHAMMED]

Here is the link where I have asked the question:

https://www.mathworks.com/matlabcentral/answers/332678-how-can-we-change-particular-part-of-binary-image-which-is-in-white-color-to-black-color?s_tid=email_ans_new_ans_ans_h#answer_260988

I have tried it with bwmorph but it does not satisfy the result.Thanks for you help.

[Asmalina Mohamed Saat]

to Image analyst.. this coding;

numWhitePixels = sum(binaryImage(:));

numBlackPixels = sum(~binaryImage(:));

if i want to know the black pixel and white pixel separately, can i use the above coding? and if i want to know to sum of black pixel and white pixel, what is the coding can i use?

[Image Analyst]

To get the sum you use the "+" operator, or use numel() to count the pixels in the whole image

totalPixels = numWhitePixels + numBlackPixels;

or

totalPixels = numel(binaryImage);

[Asmalina Mohamed Saat]

thanks for your answer... but how i want to make sure that the coding numWhitePixels = sum(binaryImage(:)); is calculate only the white region only?

[Image Analyst]

Yes, that will count the white pixels ONLY, not the black ones. nnz means "number of non-zeros". White pixels are non-zero (they're 1 or "true"), and black pixels are zero (0 or "false) so they are not counted by nnz().

[Asmalina Mohamed Saat]

okay..thanks

and if I want to calculate the area of white region can I use this coding area = bwarea(numwhitepixel);?? andwhat is the unit of that area?

[Image Analyst]

You use bwarea() on a binary image, not a scalar number that is the count of how many pixels there are.

% Method 1
area = nnz(binaryImage);
% Method 2
area = bwarea(binaryImage);

The units are pixels. It uses a different algorithm for computing area. See it's documentation. It's not simply a pixel count but weights the edge pixels according to the shape of the boundary locally.

If you want to report the area in real world units like square centimeters, you'll have to get a spatial calibration factor to convert linear pixels to cm, and areas to cm^2. See my attached spatial calibration demo.

Answer 2

[Walter Roberson]

nnz(TheBinaryImage)

or

sum(TheBinaryImage(:))