|
|
AForge.NET is a C# framework designed for developers and researchers in the fields of Computer Vision and Artificial Intelligence - image processing, neural networks, genetic algorithms, machine learning, etc.
At this point the framework is comprised of 5 main and some additional libraries:
- AForge.Imaging – a library for image processing routines and filers;
- AForge.Neuro – neural networks computation library;
- AForge.Genetic – evolution programming library;
- AForge.Vision – computer vision library;
- AForge.Machine Learning – machine learning library.
The work on the framework's improvement is in constants progress, what means that new feature and namespaces are coming constantly. To get knowledge about its progress you may track source repository's log or visit project discussion group to get the latest information about it.
The framework is provided with not only different libraries and their sources, but with many sample applications, which demonstrate the use of this framework, and with documentation help files, which are provided in HTML Help format.
In the case you found an issue in any component of the framework or you would like to request for a new feature, you may feel free to submit the issue/request on the “Issues” page.
In case you are interested in the project and would like to learn about it more or in case you would like to contribute it, you are more than welcome to participate in the project’s discussion group.
AForge.NET 1.6.2 Release
This release was made because of one important issue found in Canny Edge Detector, which was causing false detection of some edge points (the filter was providing too thick edges, marking more pixels as edges than it is actually supposed). Now the issue is resolved and this edge detector produces much more accurate results.
AForge.NET 1.6.2 Release Notes
AForge.NET 1.6.1 Release
This release of AForge.NET introduces some very good extensions to Blob Counting. Now it is possible to specify objects' sort order returned by blob counting algorithm. Also it is possible to extract not all objects using GetObjects() method, what may be an overhead, but only "interesting" objects using partial extraction of objects with GetObjectInformation() method and then ExtractBlobsImage() method to extract image or required blob.
// create blob counter and process image
BlobCounter bc = new BlobCounter( sourceImage );
// specify sort order
bc.ObjectsOrder = ObjectsOrder.Size;
// get objects' information (blobs without image)
Blob[] blobs = bc.GetObjectInformation( );
// process blobs
foreach ( Blob blob in blobs )
{
// check blob's properties
if ( blob.Rectangle.Width > 50 )
{
// the blob looks interesting, let's extract it
bc.ExtractBlobsImage( sourceImage, blob );
}
}AForge.Imaging library also introduces Extract Biggest Blob filter, which extracts the biggest blob from binary image (or from original image if it is provided).
More image processing filters:
- Histogram Equalization;
- Contrast Stretch;
- Flat Field Correction.
AForge.Neuro library finally introduces neural network saving/loading functionality.
AForge.NET 1.6.1 Release Notes
AForge.NET 1.6.0 Release
This release introduces new namespace in AForge.NET framework - AForge.Robotics.Lego. For now this namespace contains classes, which allow manipulation with Lego Mindstorms NXT device, setting/getting its motors' state, getting information about sensors' values and retrieving generic information about the NXT brick. The namespace is supposed to grow introducing new features and it is supposed that its upper namespace (AForge.Robotics) eventually will include more sub namespaces with support of more Robotics stuff.
// create an instance of communication interface for NXT device
SerialCommunication nxtCommunication = new SerialCommunication( "COM1" );
// create an instance of NXT brick
NXTBrick nxt = new NXTBrick( nxtCommunication );
// connect to the device
if ( nxt.Connect( ) == CommunicationStatus.Success )
{
// run motor A
MotorState motorState = new MotorState( );
motorState.Power = 60;
motorState.TurnRatio = 50;
motorState.Mode = MotorMode.On;
motorState.Regulation = RegulationMode.Idle;
motorState.RunState = RunState.Running;
motorState.TachoLimit = 1000;
nxt.SetMotorState( OutputPort.MotorA, motorState );
// get input value from sensor 1
InputValues inputValues;
if ( nxt.GetInputValues( InputPort.Port1, out inputValues ) == CommunicationStatus.Success )
{
// ...
}
// ...
}
AForge.Imaging introduces two new classes to calculate statistics about horizontal and vertical pixels' intensities distribution, what allows detecting which parts of an image are more intensive, filled with objects or where an object has its maximum width/height. These classes may be used as in regular image analysis, as in objects' recognition based on their statistical features.
HorizontalIntensityStatistics his = new HorizontalIntensityStatistics( sourceImage ); VerticalIntensityStatistics vis = new VerticalIntensityStatistics ( sourceImage ); // process histograms ... // to display histogram the AForge.Controls.Histogram control may be used histogramControl.Values = his.Gray.Values;
AForge.Imaging introduces Moravec corners detector implementing ICornersDetector interface. As the above image shows the use of its filter is limited to some certain cases, what is caused by some limitations of Moravec operator (anisotropic response, etc.). Other corners detection algorithms are going to be provided in future releases.
// create corner detector's instance
MoravecCornersDetector mcd = new MoravecCornersDetector( );
// process image searching for corners
Point[] corners = mcd.ProcessImage( image );
// process points
foreach ( Point corner in corners )
{
// ...
}To highlight objects' corners as it is showed on the image above, the new CornersMarker filter may be used, which can work with any specified corners detection algorithm.
// create corner detector's instance MoravecCornersDetector mcd = new MoravecCornersDetector( ); // create corner maker filter CornersMarker filter = new CornersMarker( mcd, Color.Red ); // apply the filter filter.ApplyInPlace( image );
AForge.NET 1.6.0 also introduces some more new features, like Otsu thresholding, initial implementation of Integral Image class, different bug fixed and improvements.