Object Tracking

Object tracking is an essential part of many applications including pedestrian avoidance, security and surveillance, and augmented reality. This examples shows motion-based tracking of moving people in a video from a stationary camera.

Object Detection and Counting

Video processing can be used to detect and count objects that move in video sequences. In this case study, scientists in Australia are using video footage to estimate the wildlife population of waterbirds.

Video Processing in 4 Easy Steps

Step 2. Displaying the Video

There are two methods for displaying video in MATLAB:

• deployableVideoPlayer: Efficiently view a series of video frames
• implay: Launch the Video Viewer app for viewing videos

Step 3. Processing the Video

A video is a sequence of individual video frames, or images. This means an algorithm designed to perform edge detection on an image can be quickly converted to perform edge detection on a video.

Step 4. Writing the Video

After processing, you can write each frame of a video back to a file. You can create a video file with the function:

>> vid_w = VideoWriter('newfile.avi');
>> open(vid_w)

The variable vid_w can accumulate new frames to create a video.

A Complete MATLAB Example

Putting all the components together, let’s run through a complete example to show the steps of reading, displaying, processing, and writing video:

%% Read and process a video into MATLAB
 
% Setup: create Video Reader and Writer
videoFileReader = VideoReader('tilted_face.avi');
myVideo = VideoWriter('myFile.avi');

% Setup: create deployable video player and face detector
depVideoPlayer = vision.DeployableVideoPlayer;
faceDetector = vision.CascadeObjectDetector();
open(myVideo);

%% Detect faces in each frame
while hasFrame(videoFileReader)

	% read video frame
	videoFrame = readFrame(videoFileReader);
	% process frame
bbox = faceDetector(videoFrame);
videoFrame = insertShape(videoFrame, 'Rectangle', bbox); 

	% Display video frame to screen
	depVideoPlayer(videoFrame);

	% Write frame to final video file
	writeVideo(myVideo, videoFrame);
	pause(1/videoFileReader.FrameRate);

end
close(myVideo)

You can download this code on MATLAB Central.