Package dlife.vision

Classes and interfaces for image processing and computer/robot vision.

See:
Description

Class Summary

BufferedImageWithOverlay	A BufferedImageWithOverlay allows a BufferedImage to be drawn while allowing things like text, lines, rectangles and ovals to be overlaid in front of it.
Camera	Base class for creating cameras for different image sources.
FilterManager	The FilterManager maintains the chain of Filters that are being applied to the video stream of a Camera device.
VideoDisplay	A VideoDisplay is a wrapper for two BufferedImageWithOverlay objects, one for the raw video and one for the filtered video.
VideoSource	A VideoSource provides BufferedImages captured from some video source and methods for controlling the flow of video (e.g.
VideoSourceGUI	This class provides GUI components for controlling a VideoSource (e.g.
VisionEventDescriptors	This is a holding place for constants that describe the different types of events that generate notifications of Observers in the dlife.vision package.

Exception Summary

VideoSourceConnectionException	Exception that is thrown when an error/exception occurs when attempting to connect to or disconnect from a VideoSource.
VideoSourceCreationFailedException	Exception that is thrown when the construction of a VideoSource fails.

Package dlife.vision Description

Classes and interfaces for image processing and computer/robot vision.

Overview:

The dLife vision system provides support for processing images from a variety of sources including single images, sequence of images and V4L devices. Users will interact with the vision system through Camera and Filter objects. Camerea's provide an image or stream of images to be processed. Filters process the images (e.g. blur, blobify, etc) and possibly provide information about the image(s) (e.g. blob locations).

Quick Start with Image Files

1. Complete the Basic dLife Installation.
2. Launch the ControlCenter.
3. Select dlife.vision.cameras.VisionPseudoRobot in the "Robot" drop down.
4. Click "Controller" navigate to and choose dLifeDist/examples/dlife/vision/CokeCanTracker.class.
5. Click "Start Up".
6. Select the dLifeDist/examples/dlife/vision/SampleImages/CokeCan directory in the file chooser dialog that appears and click "Connect"
7. Click "Run". In the "Raw Video" window a cyan circle should indicate the location of the Coke can. In the "Filtered Video" window the Coke can will appear blurry, will be colored yellow and will have a white rectangle around it.

Quick Start with Video For Linux (V4L)

1. Complete the Basic dLife Installation.
2. Install the VisionServer on a linux box.
3. Attach a V4L compatible device (e.g. capture card and robot camera) to the linux box.
   You can use XawTV or VLC to test your V4L device. These are readily available for most linux distributions.
4. Start the VisionServer on the linux box.
5. Select dlife.vision.cameras.VisionPseudoRobot in the "Robot" drop down of the ControlCenter.
6. Click "Controller"
7. Navigate to and choose dLifeDist/examples/dlife/vision/CokeCanTrackerV4L.class in the File Chooser dialog that appears.
8. Click "Start Up".
9. Enter the connection information for your VisionServer in the dialog that appears and click "Connect"
10. Choose and configure your V4L device in the dialog that appears and click "Connect"
11. Click "Run".
    When run this controller tries to place a cyan circle at the center of the Coke can in the "Raw Video" window. How well it is able to do this depends upon the lighting and the quality of the video camera used. If the controller is not finding the Coke can you can try the following:
    1. Double click the "Camera" tab to place it in its own window.
    2. Click the "Filters" tab.
    3. Click the "Match" tab within the "Filters" tab.
    4. Clear the match filter.
    5. Pause the video.
    6. Click and drag on the Coke can in the "Filtered Video" window to select new colors to match.
    7. Click on the "Code" tab within the "Filters" tab.
    8. Copy the code for the MatchFilter to the CokeCanTrackerV4L.java file, replacing the MatchFilter that was there.
    9. Recompile CokeCanTrackerV4L.
    10. Shutdown and restart the controller.

Additional Resources

• The dlife.vision.cameras package provides Camera classes for a variety of image sources. This documentation also describes the vision system's user interface.
• The dlife.vision.filter package provides a variety of filters that can be used to process images.
• The examples.dlife.vision package contains working examples that use the dLife vision system.
• The dlife.vision.server package contains information about installing and using the dLife VisionServer to stream video from Video For Linux (V4L) devices.

Developer Information:

NEED TO WRITE THIS