Tracking moving object using KLT

Following is code to track corner points using klt algorithm.

I detect corner points using harris corner detector which also include some background points but i want to track only moving objects in the video and place box around them. Can i use blob and gaussian mixture model?

 videoFileReader = vision.VideoFileReader('CrowdVideo.avi'));
 videoPlayer = vision.VideoPlayer('Position',[100,100,680,520]);
 release(videoFileReader);
 release(videoPlayer);
 objectFrame = step(videoFileReader);

 %Detect interest points in the object region
%for rgb image
points = detectHarrisFeatures(rgb2gray(objectFrame));

 %Display the detected points
pointImage = insertMarker(objectFrame, points.Location, '+', 'Color', 'white');
figure;
imshow(pointImage);
title('Detected Interest points');

 %Creater tracker object
tracker = vision.PointTracker('MaxBidirectionalError',1);

 %Initialize tracker
initialize(tracker, points.Location, objectFrame);

 %read, track, display point and results in each frame
while ~isDone(videoFileReader)
    frame = step(videoFileReader);
    [points, validity] = step(tracker, frame);
    out = insertMarker(frame, points(validity,:),'+');
    step(videoPlayer, out);
end

 release(videoPlayer);
release(videoFileReader);
  end

Answer 1

[Mohamed Ata]

yes, blob analysis is the best way for your purpose