Removing blobs from a binary image

Question

I have a binary image which contain few blobs.

I want to remove blobs which are less than a certain area.

Can any one suggest me a way?

I am using Open-CV. i did dilation and erosion to get those blobs. so i need something different to remove the reaming blobs which are less than a certain area.

1. run conncomp labeling algorithm. 2. Throw out areas with size smaller than treshold. Another variant is N-times erosion and after - N-times dilation (but the higher N the worse restavration of blobs) — Eddy_Em
– Eddy_Em, Commented Jun 7, 2013 at 18:38

Alexey · Accepted Answer · 2013-06-07 19:29:49Z

3

You can do something like this:

// your input binary image
// assuming that blob pixels have positive values, zero otherwise
Mat binary_image; 

// threashold specifying minimum area of a blob
double threshold = 100;

vector<vector<Point>> contours;
vector<Vec4i> hierarchy;
vector<int> small_blobs;
double contour_area;
Mat temp_image;

// find all contours in the binary image
binary_image.copyTo(temp_image);
findContours(temp_image, contours, hierarchy, CV_RETR_CCOMP,
                                                  CV_CHAIN_APPROX_SIMPLE);

// Find indices of contours whose area is less than `threshold` 
if ( !contours_all.empty()) {
    for (size_t i=0; i<contours.size(); ++i) {
        contour_area = contourArea(contours_all[i]) ;
        if ( contour_area < threshold)
            small_blobs.push_back(i);
    }
}

// fill-in all small contours with zeros
for (size_t i=0; i < small_blobs.size(); ++i) {
    drawContours(binary_image, contours, small_blobs[i], cv::Scalar(0), 
                                                 CV_FILLED, 8);
}

edited Jun 7, 2013 at 19:29

answered Jun 7, 2013 at 19:04

Alexey

5,9989 gold badges46 silver badges81 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Teh Sunn Liu Over a year ago

Can somebody please explain me how did contours_all came in the if condition? and what it contains. I couldnt understand

user2307268 · Accepted Answer · 2015-05-04 15:09:43Z

0

            //src_gray is your image that we threshold
            threshold(src_gray, threshold_output, NULL, 255, THRESH_OTSU);
            /// Find contours
            findContours(threshold_output, contours, hierarchy, RETR_TREE, CHAIN_APPROX_SIMPLE, Point(0, 0));
            /// Approximate contours
            vector<Rect> boundRect(contours.size());
            for (unsigned int i = 0; i < contours.size(); i++)
            {   //identify bounding box
                boundRect[i] = boundingRect(contours[i]);
            }
            for (unsigned int i = 0; i < contours.size(); i++)
            {

                if ((boundRect[i].area() < //enter your area here))
                {
                    src_gray(boundRect[i])=//set it to whatever value you want;
                }
            }

Well give this a try...

answered May 4, 2015 at 15:09

user2307268

1 Comment

Priyank Pathak Over a year ago

add more details here

Collectives™ on Stack Overflow

Removing blobs from a binary image

2 Answers 2

1 Comment

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related