How to find centroid of moving object in a video?

Question

BHAVIN JOISAR 2020-1-3

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/498770-how-to-find-centroid-of-moving-object-in-a-video

评论： MD MOSTAKIM ALAM SOHAG 2021-1-18

Hello, I am tryng to find velocity of a moving ball using Histogram tracking method, where I am able to track the ball successfully. Now I want to find the velocity of the moving ball, to do that first I have to find the centroids of the ball. So when I use function regionprops() it traces centroids of entire Frame, where as I only want to find the centroids of the ball which is been tracked. And after that I want to calculate the velocity of the same ball which is been tracked.

So how do I mark the centroid of only one object in entire frame?

2 个评论
显示无隐藏无

KALYAN ACHARJYA 2020-1-3

As you already extracted the frame by frame, then see this answer, might be helpful.

BHAVIN JOISAR 2020-1-4

Hello Sir, I tried many ways in which I can only detect the centroid of the tracked object , I have even used some of the methods which are discussed in the link you shared. But again I am getting centroids of entire frame.

Can you suggest any function or method using which I can only get centroids of the tracked object.

I even tried using opticalFlow to avoid finding centroids, but I cannot get the velocity of the moving object which is my final task.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Image Analyst 2020-1-4

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/498770-how-to-find-centroid-of-moving-object-in-a-video#answer_408653

编辑：Image Analyst 2020-1-4

track_color_in_video.zip

You need to segment your objects in the frame first. Then call regionprops() to get the centroid of each blob. If blobs enter or exit the scene, or overlap each other, then you'll have to have special code to handle those situations. See my Image Segmentation Tutorial and attached demo.

11 个评论
显示 9更早的评论隐藏 9更早的评论

Image Analyst 2021-1-2

在 MATLAB Online 中打开

For you, with what looks like a computer graphics disc with 255 outside of it, you'd want to find where any of the pixels is not 255, like

coloredMask = any(thisFrame < 255, 3);
coloredMask = bwareafilt(mask, 1); % Take largest area blob.

You wouldn't need to do all that stuff about HSV color space (the majority of the loop).

% Read one frame at a time, and find specified color.
for k = 1 : numberOfFrames	
	% Read one frame
	thisFrame=read(videoObject,k);
	hImage=subplot(3, 4, 1);
	% Display it.
	imshow(thisFrame);
	axis on;
	caption = sprintf('Original RGB image, frame #%d 0f %d', k, numberOfFrames);
	title(caption, 'FontSize', fontSize);
	drawnow;
	
	if k == 1
		% Enlarge figure to full screen.
		set(gcf, 'Units', 'Normalized', 'OuterPosition', [0 0 1 1]);
		% Give a name to the title bar.
		set(gcf, 'Name', 'Demo by ImageAnalyst', 'NumberTitle', 'Off') 
		hCheckbox = uicontrol('Style','checkbox',... 
			'Units', 'Normalized',...
			'String', 'Finish Now',... 
			'Value',0,'Position', [.2 .96 .4 .05], ...
			'FontSize', 14);
	end
	
	% Overall color mask.
        coloredMask = any(thisFrame < 255, 3);
        coloredMask = bwareafilt(mask, 1); % Take largest area blob.
	subplot(3, 4, 9);
	imshow(coloredMask, []);
	axis on;
	title('Colored Blob Mask', 'FontSize', fontSize);
	drawnow;
    

etc.

Image Analyst 2021-1-3

在 MATLAB Online 中打开

I'm sure you figured it out by now, but try running my version and see how it goes:

% Demo to track colored spot.  Finds and annotates centroid and bounding box of spot.
% Modify thresholds to detect different colors.
clc;    % Clear the command window.
close all;  % Close all figures (except those of imtool.)
imtool close all;  % Close all imtool figures if you have the Image Processing Toolbox.
clear;  % Erase all existing variables. Or clearvars if you want.
workspace;  % Make sure the workspace panel is showing.
format long g;
format compact;
fontSize = 20;
% Specify input video file name.
folder = pwd;
baseFileName = 'Drop-Wall - P1bar-Tw200C-U6m_s-D50um - Front.avi';
fullFileName = fullfile(folder, baseFileName);
% Check if the video file actually exists in the current folder or on the search path.
if ~exist(fullFileName, 'file')
	% File doesn't exist -- didn't find it there.  Check the search path for it.
	fullFileNameOnSearchPath = baseFileName; % No path this time.
	if ~exist(fullFileNameOnSearchPath, 'file')
		% Still didn't find it.  Alert user.
		errorMessage = sprintf('Error: %s does not exist in the search path folders.', fullFileName);
		uiwait(warndlg(errorMessage));
		return;
	end
end
% Instantiate a video reader object for this video.
videoObject = VideoReader(fullFileName);
% Setup other parameters
numberOfFrames = videoObject.NumberOfFrame;
x = zeros(1, numberOfFrames);
y = zeros(1, numberOfFrames);
% Read one frame at a time, and find specified color.
for k = 1 : numberOfFrames
	% Read one frame
	thisFrame=read(videoObject,k);
	hImage=subplot(2, 2, 1);
	% Display it.
	imshow(thisFrame);
	axis on;
	caption = sprintf('Video Frame #%d of %d', k, numberOfFrames);
	title(caption, 'FontSize', fontSize);
	drawnow;
	impixelinfo;
	
	if k == 1
		% Enlarge figure to full screen.
		g = gcf;
		g.Units = 'Normalized';
		g.WindowState = 'maximized';
		% Give a name to the title bar.
		set(gcf, 'Name', 'Demo by ImageAnalyst', 'NumberTitle', 'Off')
		% Create a checkbox where they can finish early if they want, by clicking on it.
		hCheckbox = uicontrol('Style','checkbox',...
			'Units', 'Normalized',...
			'String', 'Finish Now',...
			'Value',0,'Position', [.45 .77 .1 .05], ...
			'FontSize', 20, 'FontWeight', 'bold');
	end
	
	% Overall color mask is anything that is not white or black.
	threshold = 250;
	spotMask = any(thisFrame < threshold, 3);
	% There is, for some reason, a black line at line 510.  Erase that.
	spotMask(510, :) = false;
	% Because of very bad compression artifacts, there are holes.  We need to fill those:
	spotMask = imfill(spotMask, 'holes');
	% Take largest area blob.
	spotMask = bwareafilt(spotMask, 1); 
	
	% Display mask
	subplot(2, 2, 2);
	hold off;
	imshow(spotMask, []);
	axis on;
	hold on;
	title('Colored Droplet Mask', 'FontSize', fontSize);
	
	[labeledImage, numberOfRegions] = bwlabel(spotMask);
	if numberOfRegions >= 1
		props = regionprops(labeledImage, 'BoundingBox', 'Centroid');
		% Delete old texts and rectangles
		if exist('hRect', 'var')
			delete(hRect);
		end
		if exist('hText', 'var')
			delete(hText);
		end
		
		caption = sprintf('%d Droplets found in frame #%d 0f %d', numberOfRegions, k, numberOfFrames);
		title(caption, 'FontSize', fontSize);
		drawnow;
		
		% This is a loop to bound the colored objects in a rectangular box.
		for r = 1 : numberOfRegions
			% Find location for this droplet.
			thisBB = props(r).BoundingBox;
			thisCentroid = props(r).Centroid;
			hRect(r) = rectangle('Position', thisBB, 'EdgeColor', 'r', 'LineWidth', 2);
			hSpot = plot(thisCentroid(1), thisCentroid(2), 'r+', 'MarkerSize', 50, 'LineWidth', 2);
			hText(r) = text(thisBB(1), thisBB(2)-20, strcat('X: ', num2str(round(thisCentroid(1))), '    Y: ', num2str(round(thisCentroid(2)))));
			set(hText(r), 'FontName', 'Arial', 'FontWeight', 'bold', 'FontSize', 12, 'Color', 'yellow');
		end
		hold off
		drawnow;
		
		% Remember this centroid
		x(k) = thisCentroid(1);
		y(k) = thisCentroid(2);
		subplot(2, 2, 3:4);
		plot(x, 'r-', 'MarkerSize', 50, 'LineWidth', 2);
		hold on;
		plot(y, 'b-', 'MarkerSize', 50, 'LineWidth', 2);
		hold off;
		grid on;
		title('Red = x, Blue = y', 'FontSize', fontSize);
		xlabel('Frame number', 'FontSize', fontSize);
	end
	
	% See if they want to bail out
	if get(hCheckbox, 'Value')
		% Finish now checkbox is checked.
		delete(hCheckbox);
		break;
	end
end
msgbox('Done with demo.');

Image Analyst 2021-1-17

Save all the images to memory or disk, then use montage() or imtile().

Start your own question if you still have trouble, rather than here in BHAVIN's question where he'll continue to get emails about activity on his thread.

MD MOSTAKIM ALAM SOHAG 2021-1-18