public static void FindMatch(Mat modelImage, Mat observedImage, out long matchTime, out VectorOfKeyPoint modelKeyPoints, out VectorOfKeyPoint observedKeyPoints, VectorOfVectorOfDMatch matches, out Mat mask, out Mat homography)
{
int k = 2;
double uniquenessThreshold = 0.8;
double hessianThresh = 300;
Stopwatch watch;
homography = null;
modelKeyPoints = new VectorOfKeyPoint();
observedKeyPoints = new VectorOfKeyPoint();
{
//using (UMat uModelImage = modelImage.ToUMat(AccessType.Read))
//using (UMat uObservedImage = observedImage.ToUMat(AccessType.Read))
//UMat uModelImage = new UMat();
//UMat uObservedImage = new UMat();
//CvInvoke.convert
{
Brisk surfCPU = new Brisk();
//SURF surfCPU = new SURF(hessianThresh);
//extract features from the object image
UMat modelDescriptors = new UMat();
surfCPU.DetectAndCompute(modelImage, null, modelKeyPoints, modelDescriptors, false);
watch = Stopwatch.StartNew();
// extract features from the observed image
UMat observedDescriptors = new UMat();
surfCPU.DetectAndCompute(observedImage, null, observedKeyPoints, observedDescriptors, false);
BFMatcher matcher = new BFMatcher(DistanceType.L2);
matcher.Add(modelDescriptors);
matcher.KnnMatch(observedDescriptors, matches, k, null);
mask = new Mat(matches.Size, 1, DepthType.Cv8U, 1);
mask.SetTo(new MCvScalar(255));
Features2DToolbox.VoteForUniqueness(matches, uniquenessThreshold, mask);
int nonZeroCount = CvInvoke.CountNonZero(mask);
if (nonZeroCount >= 4)
{
nonZeroCount = Features2DToolbox.VoteForSizeAndOrientation(modelKeyPoints, observedKeyPoints,
matches, mask, 1.5, 20);
if (nonZeroCount >= 4)
{
homography = Features2DToolbox.GetHomographyMatrixFromMatchedFeatures(modelKeyPoints,
observedKeyPoints, matches, mask, 2);
}
}
watch.Stop();
}
}
matchTime = watch.ElapsedMilliseconds;
}
public void TestBOWKmeansTrainer2()
{
Image <Gray, byte> box = EmguAssert.LoadImage <Gray, byte>("box.png");
Brisk detector = new Brisk(30, 3, 1.0f);
VectorOfKeyPoint kpts = new VectorOfKeyPoint();
Mat descriptors = new Mat();
detector.DetectAndCompute(box, null, kpts, descriptors, false);
Mat descriptorsF = new Mat();
descriptors.ConvertTo(descriptorsF, CvEnum.DepthType.Cv32F);
//Matrix<float> descriptorsF = descriptors.Convert<float>();
BOWKMeansTrainer trainer = new BOWKMeansTrainer(100, new MCvTermCriteria(), 3, CvEnum.KMeansInitType.PPCenters);
trainer.Add(descriptorsF);
Mat vocabulary = new Mat();
trainer.Cluster(vocabulary);
BFMatcher matcher = new BFMatcher(DistanceType.L2);
BOWImgDescriptorExtractor extractor = new BOWImgDescriptorExtractor(detector, matcher);
Mat vocabularyByte = new Mat();
vocabulary.ConvertTo(vocabularyByte, CvEnum.DepthType.Cv8U);
extractor.SetVocabulary(vocabularyByte);
Mat descriptors2 = new Mat();
extractor.Compute(box, kpts, descriptors2);
}
/// <summary>
///
/// </summary>
/// <param name="imagePath"></param>
private static ObjectFeatures DetectFeatures_Brisk(Image <Bgr, byte> image)
{
try
{
Mat modelImage = image.Mat;
VectorOfKeyPoint modelKeyPoints = new VectorOfKeyPoint();
using (UMat uModelImage = modelImage.GetUMat(AccessType.Read))
{
Brisk brisk = new Brisk();
UMat descriptors = new UMat();
Debug.WriteLine("Detecting features and computing descriptors...");
brisk.DetectAndCompute(uModelImage, null, modelKeyPoints, descriptors, false);
Debug.WriteLine("Computation finished!");
return(new ObjectFeatures(modelKeyPoints, descriptors));
}
}
catch (Exception e)
{
Debug.WriteLine("Exception when detecting features" + e.Message);
throw;
}
}
public void compare()
{
using (Mat des_t = new Mat())
using (Image<Gray, byte> theMaster = new Image<Gray, byte>(master))
using (Image<Gray, byte> theTest = new Image<Gray, byte>(test))
using (Image<Gray, byte> mask = new Image<Gray, byte>(theMaster.Size))
using (Mat des_m = new Mat())
using (VectorOfKeyPoint kp_m = new VectorOfKeyPoint())
using (VectorOfKeyPoint kp_t = new VectorOfKeyPoint())
{
//SURF br = new SURF(12500);
Brisk br = new Brisk(75, 8, 14.28f);
//Image<Gray, byte> backGround = new Image<Gray, byte>(theMaster.Size);
//backGround = new Image<Gray, byte>(bg);
System.Diagnostics.Stopwatch timer = System.Diagnostics.Stopwatch.StartNew();
//CvInvoke.Subtract(backGround, theMaster, mask);
//Image<Gray, byte> theMask = mask.Convert<Gray, byte>();
//Image<Gray, byte> thetMask = tmask.Convert<Gray, byte>();
//theMask = theMask.ThresholdBinary(new Gray(120), new Gray(255));
//CvInvoke.Imshow("Mask", mask.Resize(0.2, Emgu.CV.CvEnum.Inter.Nearest));
try
{
br.DetectAndCompute(theMaster, null, kp_m, des_m, false);
br.DetectAndCompute(theTest, null, kp_t, des_t, false);
br = null;
}
catch (CvException ex)
{
MessageBox.Show(ex.ToString());
throw ex;
}
finally
{
if (br != null)
br.Dispose();
}
BFMatcher bf = new BFMatcher(DistanceType.L2);
VectorOfVectorOfDMatch raw_matches = new Emgu.CV.Util.VectorOfVectorOfDMatch();
bf.Add(des_t);
try
{
bf.KnnMatch(des_m, raw_matches, 2, null);
bf = null;
}
catch (Exception e1)
{
MessageBox.Show(e1.Message);
throw;
}
MDMatch[][] dmatches_tm = raw_matches.ToArrayOfArray();
List<MDMatch> good_matches = new List<MDMatch>();
foreach (MDMatch[] m in dmatches_tm)
{
if (m[0].Distance < 0.9 * m[1].Distance && m[0].Distance > 0.7 * m[1].Distance)
good_matches.Add(m[0]);
}
List<int> qidx_t = new List<int>();
List<int> tidx_t = new List<int>();
int final_dist_t = 0;
int oct_count_t = 0;
foreach (MDMatch m in good_matches)
{
var a = kp_m[m.QueryIdx];
var b = kp_t[m.TrainIdx];
if ((Convert.ToInt32(a.Octave) == Convert.ToInt32(b.Octave))
if(Math.Abs(a.Size - b.Size)<1))
{
oct_count_t += 1;
tidx_t.Add(m.TrainIdx);
qidx_t.Add(m.QueryIdx);
}
}
for (int i = 0; i < qidx_t.Count; i++)
{
int q_base = qidx_t[i];
int t_base = tidx_t[i];
final_dist_t--;
for (int j = i; j < qidx_t.Count; j++)
{
float q = eucledianDist(kp_m[qidx_t[j]].Point, kp_m[q_base].Point);
float t = eucledianDist(kp_t[tidx_t[j]].Point, kp_t[t_base].Point);
if (Math.Abs(q - t) <= ( 0.06 * q ))
final_dist_t++;
}
}
//Image<Bgr, byte> resultImg = new Image<Bgr, byte>(theMaster.Size);
//Features2DToolbox.DrawMatches(theMaster, kp_m, theTest, kp_t, raw_matches, resultImg, new MCvScalar(0), new MCvScalar(155));
//CvInvoke.Imshow("Res", resultImg.Resize(0.25, Emgu.CV.CvEnum.Inter.Nearest));
float raw_len = dmatches_tm.Length;
float final_dist = final_dist_t;
float per = (final_dist / raw_len) * 100;
long time_taken = timer.ElapsedMilliseconds;
MessageBox.Show("Result " + per + "\nTime: " + time_taken +
"\nmKP" + kp_m.Size + "\ntKP" + kp_t.Size +
"\ngood_kp" + good_matches.Count + "\noct_cnt" + oct_count_t + "\ndist "+ final_dist);
}
}
public string compareAll()
{
using (Mat des_t = new Mat())
using (Image<Gray, byte> theMaster = new Image<Gray, byte>(master))
using (Image<Gray, byte> theTest = new Image<Gray, byte>(test))
using (Image<Gray, byte> mask = new Image<Gray, byte>(theMaster.Size))
using (Mat des_m = new Mat())
using (VectorOfKeyPoint kp_m = new VectorOfKeyPoint())
using (VectorOfKeyPoint kp_t = new VectorOfKeyPoint())
{
//SURF br = new SURF(10000, 8, 4);
Brisk br = new Brisk(80, 7, 4f);
//Image<Gray, byte> backGround = new Image<Gray, byte>(theMaster.Size);
//backGround = new Image<Gray, byte>(bg);
System.Diagnostics.Stopwatch timer = System.Diagnostics.Stopwatch.StartNew();
//CvInvoke.Subtract(backGround, theMaster, mask);
//Image<Gray, byte> theMask = mask.Convert<Gray, byte>();
//Image<Gray, byte> thetMask = tmask.Convert<Gray, byte>();
//theMask = theMask.ThresholdBinary(new Gray(120), new Gray(255));
//CvInvoke.Imshow("Mask", mask.Resize(0.2, Emgu.CV.CvEnum.Inter.Nearest));
try
{
br.DetectAndCompute(theMaster, null, kp_m, des_m, false);
br.DetectAndCompute(theTest, null, kp_t, des_t, false);
br = null;
}
catch (CvException ex)
{
MessageBox.Show(ex.ToString());
throw ex;
}
finally
{
if (br != null)
br.Dispose();
}
BFMatcher bf = new BFMatcher(DistanceType.L2);
VectorOfVectorOfDMatch raw_matches = new Emgu.CV.Util.VectorOfVectorOfDMatch();
bf.Add(des_t);
try
{
bf.KnnMatch(des_m, raw_matches, 2, null);
bf = null;
}
catch (Exception e1)
{
MessageBox.Show(e1.Message);
throw;
}
MDMatch[][] dmatches_tm = raw_matches.ToArrayOfArray();
List<MDMatch> good_matches = new List<MDMatch>();
foreach (MDMatch[] m in dmatches_tm)
{
if (m[0].Distance < 0.8 * m[1].Distance && m[0].Distance > 0.7 * m[1].Distance)
good_matches.Add(m[0]);
}
List<int> qidx_t = new List<int>();
List<int> tidx_t = new List<int>();
int final_dist_t = 0;
int oct_count_t = 0;
foreach (MDMatch m in good_matches)
{
var a = kp_m[m.QueryIdx];
var b = kp_t[m.TrainIdx];
//if(Convert.ToInt32(a.Octave) == Convert.ToInt32(b.Octave))
//if (Math.Abs(a.Size - b.Size) < 1f)
{
oct_count_t += 1;
tidx_t.Add(m.TrainIdx);
qidx_t.Add(m.QueryIdx);
}
}
for (int i = 0; i < qidx_t.Count; i++)
{
int q_base = qidx_t[i];
int t_base = tidx_t[i];
for (int j = i+1; j < qidx_t.Count; j++)
{
float q = eucledianDist(kp_m[qidx_t[j]].Point, kp_m[q_base].Point);
float t = eucledianDist(kp_t[tidx_t[j]].Point, kp_t[t_base].Point);
if (Math.Abs(q - t) <= (0.005f * q))
final_dist_t++;
}
}
//Image<Bgr, byte> resultImg = new Image<Bgr, byte>(theMaster.Size);
//Features2DToolbox.DrawMatches(theMaster, kp_m, theTest, kp_t, raw_matches, resultImg, new MCvScalar(0), new MCvScalar(155));
//CvInvoke.Imshow("Res", resultImg.Resize(0.3, Emgu.CV.CvEnum.Inter.Nearest));
float raw_len = raw_matches.Size;
float final_dist = final_dist_t;
float per = (final_dist / raw_len) * 100;
float rat = 2000f * (final_dist_t / (float)oct_count_t) / raw_len;
float frac = (final_dist / oct_count_t);
float f = (final_dist * oct_count_t) / (raw_len); //result
long time_taken = timer.ElapsedMilliseconds;
var name = test.Substring(35).Split('.')[0];
string result = string.Format("{0},{1},{2},{3},{4},{6},{7}", name, raw_len, oct_count_t, final_dist, per, rat, frac, f);
return result;
}
}
