Difference between revisions of "Example Tracker Object"

From BoofCV
Jump to navigationJump to search
m
m
(3 intermediate revisions by the same user not shown)
Line 8: Line 8:


Example Code:
Example Code:
* [https://github.com/lessthanoptimal/BoofCV/blob/v0.17/examples/src/boofcv/examples/tracking/ExampleTrackerObjectQuad.java ExampleTrackerObjectQuad.java]
* [https://github.com/lessthanoptimal/BoofCV/blob/v0.23/examples/src/boofcv/examples/tracking/ExampleTrackerObjectQuad.java ExampleTrackerObjectQuad.java]


Concepts:
Concepts:
Line 37: Line 37:
public static void main(String[] args) {
public static void main(String[] args) {
MediaManager media = DefaultMediaManager.INSTANCE;
MediaManager media = DefaultMediaManager.INSTANCE;
String fileName = "../data/applet/tracking/track_book.mjpeg";
String fileName = UtilIO.pathExample("tracking/wildcat_robot.mjpeg");


SimpleImageSequence<ImageUInt8> video = media.openVideo(fileName, ImageType.single(ImageUInt8.class));
// Create the tracker.  Comment/Uncomment to change the tracker.
TrackerObjectQuad tracker =
FactoryTrackerObjectQuad.circulant(null, GrayU8.class);
// FactoryTrackerObjectQuad.sparseFlow(null,GrayU8.class,null);
// FactoryTrackerObjectQuad.tld(null,GrayU8.class);
// FactoryTrackerObjectQuad.meanShiftComaniciu2003(new ConfigComaniciu2003(), ImageType.pl(3, GrayU8.class));
// FactoryTrackerObjectQuad.meanShiftComaniciu2003(new ConfigComaniciu2003(true),ImageType.pl(3,GrayU8.class));


// Create the tracker.  Comment/Uncomment to change the tracker.  Mean-shift trackers have been omitted
// Mean-shift likelihood will fail in this video, but is excellent at tracking objects with
// from the list since they use color information and including color images could clutter up the example.
// a single unique color. See ExampleTrackerMeanShiftLikelihood
TrackerObjectQuad<ImageUInt8> tracker =
// FactoryTrackerObjectQuad.meanShiftLikelihood(30,5,255, MeanShiftLikelihoodType.HISTOGRAM,ImageType.pl(3,GrayU8.class));
FactoryTrackerObjectQuad.circulant(null, ImageUInt8.class);
 
// FactoryTrackerObjectQuad.sparseFlow(null,ImageUInt8.class,null);
SimpleImageSequence video = media.openVideo(fileName, tracker.getImageType());
// FactoryTrackerObjectQuad.tld(null,ImageUInt8.class);


// specify the target's initial location and initialize with the first frame
// specify the target's initial location and initialize with the first frame
Quadrilateral_F64 location = new Quadrilateral_F64(276,159,362,163,358,292,273,289);
Quadrilateral_F64 location = new Quadrilateral_F64(211.0,162.0,326.0,153.0,335.0,258.0,215.0,249.0);
ImageUInt8 frame = video.next();
ImageBase frame = video.next();
tracker.initialize(frame,location);
tracker.initialize(frame,location);


Line 58: Line 63:
gui.setBackGround((BufferedImage)video.getGuiImage());
gui.setBackGround((BufferedImage)video.getGuiImage());
gui.setTarget(location,true);
gui.setTarget(location,true);
ShowImages.showWindow(gui,"Tracking Results");
ShowImages.showWindow(gui,"Tracking Results", true);


// Track the object across each video frame and display the results
// Track the object across each video frame and display the results
long previous = 0;
while( video.hasNext() ) {
while( video.hasNext() ) {
frame = video.next();
frame = video.next();
Line 67: Line 73:


gui.setBackGround((BufferedImage) video.getGuiImage());
gui.setBackGround((BufferedImage) video.getGuiImage());
gui.setTarget(location,visible);
gui.setTarget(location, visible);
gui.repaint();
gui.repaint();


BoofMiscOps.pause(20);
// shoot for a specific frame rate
long time = System.currentTimeMillis();
BoofMiscOps.pause(Math.max(0,80-(time-previous)));
previous = time;
}
}
}
}
}
}
</syntaxhighlight>
</syntaxhighlight>

Revision as of 05:48, 28 March 2016

In this example several different trackers can be used to track the object in the video. All of the trackers in this example use texture information (see mean-shift example for color based trackers) to localize and track the object. No one single tracker can track all the video sequences provided with BoofCV or in general. You should try out all the different video sequences and see which trackers work best when.

Example Code:

Concepts:

  • Object tracking
  • Local trackers
  • Re-detection

Relevant Applets:

Example Code

/**
 * Demonstration on how to use the high level {@link TrackerObjectQuad} interface for tracking objects in a
 * video sequence.  This interface allows the target to be specified using an arbitrary quadrilateral.  Specific
 * implementations might not support that shape, so they instead will track an approximation of it.  The
 * interface also allows information on target visibility to be returned.  As is usually the case, tracker
 * specific information is lost in the high level interface and you should consider using the trackers
 * directly if more control is needed.
 *
 * This is an active area of research and all of the trackers eventually diverge given a long enough sequence.
 *
 * @author Peter Abeles
 */
public class ExampleTrackerObjectQuad {

	public static void main(String[] args) {
		MediaManager media = DefaultMediaManager.INSTANCE;
		String fileName = UtilIO.pathExample("tracking/wildcat_robot.mjpeg");

		// Create the tracker.  Comment/Uncomment to change the tracker.
		TrackerObjectQuad tracker =
				FactoryTrackerObjectQuad.circulant(null, GrayU8.class);
//				FactoryTrackerObjectQuad.sparseFlow(null,GrayU8.class,null);
//				FactoryTrackerObjectQuad.tld(null,GrayU8.class);
//				FactoryTrackerObjectQuad.meanShiftComaniciu2003(new ConfigComaniciu2003(), ImageType.pl(3, GrayU8.class));
//				FactoryTrackerObjectQuad.meanShiftComaniciu2003(new ConfigComaniciu2003(true),ImageType.pl(3,GrayU8.class));

				// Mean-shift likelihood will fail in this video, but is excellent at tracking objects with
				// a single unique color.  See ExampleTrackerMeanShiftLikelihood
//				FactoryTrackerObjectQuad.meanShiftLikelihood(30,5,255, MeanShiftLikelihoodType.HISTOGRAM,ImageType.pl(3,GrayU8.class));

		SimpleImageSequence video = media.openVideo(fileName, tracker.getImageType());

		// specify the target's initial location and initialize with the first frame
		Quadrilateral_F64 location = new Quadrilateral_F64(211.0,162.0,326.0,153.0,335.0,258.0,215.0,249.0);
		ImageBase frame = video.next();
		tracker.initialize(frame,location);

		// For displaying the results
		TrackerObjectQuadPanel gui = new TrackerObjectQuadPanel(null);
		gui.setPreferredSize(new Dimension(frame.getWidth(),frame.getHeight()));
		gui.setBackGround((BufferedImage)video.getGuiImage());
		gui.setTarget(location,true);
		ShowImages.showWindow(gui,"Tracking Results", true);

		// Track the object across each video frame and display the results
		long previous = 0;
		while( video.hasNext() ) {
			frame = video.next();

			boolean visible = tracker.process(frame,location);

			gui.setBackGround((BufferedImage) video.getGuiImage());
			gui.setTarget(location, visible);
			gui.repaint();

			// shoot for a specific frame rate
			long time = System.currentTimeMillis();
			BoofMiscOps.pause(Math.max(0,80-(time-previous)));
			previous = time;
		}
	}
}