Feature2D GetDetector(FeatureDetectType detectType)
{
switch (detectType)
{
case FeatureDetectType.KAZE:
return(new KAZE());
case FeatureDetectType.SIFT:
return(new SIFT());
case FeatureDetectType.SURF:
return(new SURF(300));
default:
throw new ArgumentOutOfRangeException();
}
}
public AlgorithmResult DetectFeatureMatch(
string modelName,
string observedeName,
FeatureDetectType detectType,
FeatureMatchType matchType,
int k,
double uniquenessThreshold)
{
AlgorithmResult result = new AlgorithmResult();
Image <Bgr, byte> modelImage = ImageHelper.GetImage(modelName);
Image <Bgr, byte> observedImage = ImageHelper.GetImage(observedeName);
Mat resultImage = new Mat();
// Get features from modelImage
var modelKeyPoints = new VectorOfKeyPoint();
var modelDescriptors = new UMat();
GetDetector(detectType).DetectAndCompute(
modelImage.Convert <Gray, byte>(),
null,
modelKeyPoints,
modelDescriptors,
false);
// Get features from observedImage
var observedKeyPoints = new VectorOfKeyPoint();
var observedDescriptors = new UMat();
GetDetector(detectType).DetectAndCompute(
observedImage.Convert <Gray, byte>(),
null,
observedKeyPoints,
observedDescriptors,
false);
// Perform match with selected matcher
var matches = new VectorOfVectorOfDMatch();
if (matchType == FeatureMatchType.Flann)
{
// TODO: Add parameters for GetIndexParams
var flannMatcher = new FlannBasedMatcher(new AutotunedIndexParams(), new SearchParams());
flannMatcher.Add(modelDescriptors);
flannMatcher.KnnMatch(
observedDescriptors,
matches,
k,
null);
}
else
{
// TODO: Add parameters for DistanceType
var bfMatcher = new BFMatcher(DistanceType.L2);
bfMatcher.Add(modelDescriptors);
bfMatcher.KnnMatch(
observedDescriptors,
matches,
k,
null);
}
// Find homography
Mat homography = null;
var mask = new Mat(matches.Size, 1, DepthType.Cv8U, 1);
mask.SetTo(new MCvScalar(255));
VoteForUniqueness(matches, uniquenessThreshold, mask);
// If 4 or more patches continue
int nonZeroCount = CvInvoke.CountNonZero(mask);
if (nonZeroCount >= 4)
{
// Filter for majority scale and rotation
nonZeroCount = VoteForSizeAndOrientation(
modelKeyPoints, observedKeyPoints, matches, mask, 1.5, 20);
// If 4 or more patches continue
if (nonZeroCount >= 4)
{
homography = GetHomographyMatrixFromMatchedFeatures(modelKeyPoints,
observedKeyPoints, matches, mask, 2);
}
}
// Draw the matched keypoints
DrawMatches(
modelImage, modelKeyPoints, observedImage, observedKeyPoints,
matches, resultImage, new MCvScalar(255, 255, 255), new MCvScalar(255, 255, 255), mask);
// Draw the projected region on the image
if (homography != null)
{
// Draw a rectangle along the projected model
Rectangle rect = new Rectangle(Point.Empty, modelImage.Size);
PointF[] pts = new PointF[]
{
new PointF(rect.Left, rect.Bottom),
new PointF(rect.Right, rect.Bottom),
new PointF(rect.Right, rect.Top),
new PointF(rect.Left, rect.Top)
};
// Transform the perspective of the points array based on the homography
// And get a rotated rectangle for the homography
pts = CvInvoke.PerspectiveTransform(pts, homography);
Point[] points = Array.ConvertAll(pts, Point.Round);
using (VectorOfPoint vp = new VectorOfPoint(points))
{
CvInvoke.Polylines(resultImage, vp, true, new MCvScalar(255, 0, 0, 255), 5);
}
}
result.ImageArray = ImageHelper.SetImage(resultImage.ToImage <Bgr, byte>());
return(result);
}
