/*
Description:
Read the license plate
Input:
//Rectified license plate image containing six characters
Int16Image labeledRectifiedPlateImage
BlobMatcher_Int16 matcher //initialized blobmatcher
ClassLexicon lexicon //initialized lexicon
Output:
//Result by the blob matcher
ref LicensePlate result
//Result by the lexicon
ref LicensePlate lexiconResult
Return:
//six characters found
bool
*/
public static bool MatchPlate(Int16Image binaryCharacterImage, BlobMatcher_Int16 matcher,
ClassLexicon lexicon, ref LicensePlate result, ref LicensePlate lexiconResult, bool dilate)
{
// NIEUW
// 2de optie voor aanroep als eerste low confidence levert
if (dilate)
{
Int16Image temp = new Int16Image();
VisionLab.Dilation(binaryCharacterImage, temp, new Mask_Int32(3, 3, 1));
binaryCharacterImage = new Int16Image(temp);
temp.Dispose();
}
if (VisionLab.LabelBlobs(binaryCharacterImage, Connected.EightConnected) != 6)
return false;
//Calculate dimensions and locations of all characters/blobs.
vector_BlobAnalyse ba_vec = new vector_BlobAnalyse();
ba_vec.Add(BlobAnalyse.BA_TopLeft);
ba_vec.Add(BlobAnalyse.BA_Height);
ba_vec.Add(BlobAnalyse.BA_Width);
vector_Blob returnBlobs = new vector_Blob();
VisionLab.BlobAnalysis(binaryCharacterImage, VisionLab.VectorToSet_BlobAnalyse(ba_vec), VisionLab.MaxPixel(binaryCharacterImage), returnBlobs, SortOrder.SortDown, BlobAnalyse.BA_TopLeft, UseXOrY.UseX);
ba_vec.Dispose();
Int16Image binaryCharacter = new Int16Image();
//Create data structure for lexicon.
vector_vector_LetterMatch match = new vector_vector_LetterMatch();
// NIEUW
// Change the matcher params
matcher.ChangeParams(60, 10, 64, 0);
//Process each character/blob.
foreach (Blob b in returnBlobs)
{
//Cut out character
VisionLab.ROI(binaryCharacterImage, binaryCharacter, b.TopLeft(), new HeightWidth(b.Height(), b.Width()));
//Convert ROI result to binary
VisionLab.ClipPixelValue(binaryCharacter, 0, 1);
//Calculate character's classification for all classes.
vector_PatternMatchResult returnMatches = new vector_PatternMatchResult();
float conf = matcher.AllMatches(binaryCharacter, (float)-0.5, (float)0.5, returnMatches);
float err = returnMatches[0].error;
int id = returnMatches[0].id;
string chr = matcher.PatternName(id);
// NIEUW
// If error to big decrease the confidence
if(err > 0.20f)
conf -= 0.2f;
//Fill datastructure for lexicon.
match.Add(VisionLabEx.PatternMatchResultToLetterMatch(returnMatches));
//Store best match in result
result.characters.Add(
new LicenseCharacter(
chr,
err,
conf,
// NIEUW
// Extra param: The middle of a character
// (used for matching patterns)
b.TopLeft().x + ((b.TopRight().x - b.TopLeft().x)/2) ,
// NIEUW
// All other results that we're found
// So we can switch between em
returnMatches
));
}
//Validate match with lexicon.
vector_int bestWord = new vector_int();
lexiconResult.confidence = lexicon.FindBestWord(match, bestWord, Optimize.OptimizeForMinimum);
for (int i = 0; i < bestWord.Count; i++)
{
string character = matcher.PatternName(bestWord[i]);
//Store lexicon result
lexiconResult.characters.Add(new LicenseCharacter(character));
}
// NIEUW
// Create the best match with the aid of the pattern matcher
result.CalculateBestMatch(matcher);
binaryCharacter.Dispose();
returnBlobs.Dispose();
match.Dispose();
bestWord.Dispose();
GC.Collect();
return true;
}
/*
* Description:
* Read the license plate
* Input:
* //Rectified license plate image containing six characters
* Int32Image labeledRectifiedPlateImage
* BlobMatcher_Int16 matcher //initialized blobmatcher
* ClassLexicon lexicon //initialized lexicon
* Output:
* //Result by the blob matcher
* ref LicensePlate result
* //Result by the lexicon
* ref LicensePlate lexiconResult
* Return:
* //six characters found
* bool
*/
public static bool MatchPlate(Int32Image binaryCharacterImage, BlobMatcher_Int16 matcher, ClassLexicon lexicon, ref LicensePlate result, ref LicensePlate lexiconResult)
{
//Check if 6 characters/blobs have been found and label image.
if (VisionLab.LabelBlobs(binaryCharacterImage, Connected.EightConnected) != 6)
{
return(false);
}
//Calculate dimensions and locations of all characters/blobs.
vector_BlobAnalyse ba_vec = new vector_BlobAnalyse();
ba_vec.Add(BlobAnalyse.BA_TopLeft);
ba_vec.Add(BlobAnalyse.BA_Height);
ba_vec.Add(BlobAnalyse.BA_Width);
vector_Blob returnBlobs = new vector_Blob();
VisionLab.BlobAnalysis(binaryCharacterImage, VisionLab.VectorToSet_BlobAnalyse(ba_vec), VisionLab.MaxPixel(binaryCharacterImage), returnBlobs, SortOrder.SortDown, BlobAnalyse.BA_TopLeft, UseXOrY.UseX);
ba_vec.Dispose();
Int32Image binaryCharacter = new Int32Image();
Int16Image binaryCharacter16 = new Int16Image();
//Create data structure for lexicon.
vector_vector_LetterMatch match = new vector_vector_LetterMatch();
//Process each character/blob.
foreach (Blob b in returnBlobs)
{
//Cut out character
VisionLab.ROI(binaryCharacterImage, binaryCharacter, b.TopLeft(), new HeightWidth(b.Height(), b.Width()));
//Convert ROI result to binary
VisionLab.ClipPixelValue(binaryCharacter, 0, 1);
//Calculate character's classification for all classes.
vector_PatternMatchResult returnMatches = new vector_PatternMatchResult();
VisionLab.Convert(binaryCharacter, binaryCharacter16);
float conf = matcher.AllMatches(binaryCharacter16, (float)-0.5, (float)0.5, returnMatches);
float err = returnMatches[0].error;
int id = returnMatches[0].id;
string chr = matcher.PatternName(id);
//Fill datastructure for lexicon.
match.Add(VisionLabEx.PatternMatchResultToLetterMatch(returnMatches));
//Store best match in result
result.characters.Add(new LicenseCharacter(chr, err, conf));
}
//Validate match with lexicon.
vector_int bestWord = new vector_int();
lexiconResult.confidence = lexicon.FindBestWord(match, bestWord, Optimize.OptimizeForMinimum);
for (int i = 0; i < bestWord.Count; i++)
{
string character = matcher.PatternName(bestWord[i]);
//Store lexicon result
lexiconResult.characters.Add(new LicenseCharacter(character));
}
binaryCharacter.Dispose();
binaryCharacter16.Dispose();
returnBlobs.Dispose();
match.Dispose();
bestWord.Dispose();
//GC.Collect();
return(true);
}
/*
* Description:
* Read the license plate
* Input:
* //Rectified license plate image containing six characters
* Int16Image labeledRectifiedPlateImage
* BlobMatcher_Int16 matcher //initialized blobmatcher
* ClassLexicon lexicon //initialized lexicon
* Output:
* //Result by the blob matcher
* ref LicensePlate result
* //Result by the lexicon
* ref LicensePlate lexiconResult
* Return:
* //six characters found
* bool
*/
public static bool MatchPlate(Int16Image binaryCharacterImage, BlobMatcher_Int16 matcher,
ClassLexicon lexicon, ref LicensePlate result, ref LicensePlate lexiconResult, bool dilate)
{
// NIEUW
// 2de optie voor aanroep als eerste low confidence levert
if (dilate)
{
Int16Image temp = new Int16Image();
VisionLab.Dilation(binaryCharacterImage, temp, new Mask_Int32(3, 3, 1));
binaryCharacterImage = new Int16Image(temp);
temp.Dispose();
}
if (VisionLab.LabelBlobs(binaryCharacterImage, Connected.EightConnected) != 6)
{
return(false);
}
//Calculate dimensions and locations of all characters/blobs.
vector_BlobAnalyse ba_vec = new vector_BlobAnalyse();
ba_vec.Add(BlobAnalyse.BA_TopLeft);
ba_vec.Add(BlobAnalyse.BA_Height);
ba_vec.Add(BlobAnalyse.BA_Width);
vector_Blob returnBlobs = new vector_Blob();
VisionLab.BlobAnalysis(binaryCharacterImage, VisionLab.VectorToSet_BlobAnalyse(ba_vec), VisionLab.MaxPixel(binaryCharacterImage), returnBlobs, SortOrder.SortDown, BlobAnalyse.BA_TopLeft, UseXOrY.UseX);
ba_vec.Dispose();
Int16Image binaryCharacter = new Int16Image();
//Create data structure for lexicon.
vector_vector_LetterMatch match = new vector_vector_LetterMatch();
// NIEUW
// Change the matcher params
matcher.ChangeParams(60, 10, 64, 0);
//Process each character/blob.
foreach (Blob b in returnBlobs)
{
//Cut out character
VisionLab.ROI(binaryCharacterImage, binaryCharacter, b.TopLeft(), new HeightWidth(b.Height(), b.Width()));
//Convert ROI result to binary
VisionLab.ClipPixelValue(binaryCharacter, 0, 1);
//Calculate character's classification for all classes.
vector_PatternMatchResult returnMatches = new vector_PatternMatchResult();
float conf = matcher.AllMatches(binaryCharacter, (float)-0.5, (float)0.5, returnMatches);
float err = returnMatches[0].error;
int id = returnMatches[0].id;
string chr = matcher.PatternName(id);
// NIEUW
// If error to big decrease the confidence
if (err > 0.20f)
{
conf -= 0.2f;
}
//Fill datastructure for lexicon.
match.Add(VisionLabEx.PatternMatchResultToLetterMatch(returnMatches));
//Store best match in result
result.characters.Add(
new LicenseCharacter(
chr,
err,
conf,
// NIEUW
// Extra param: The middle of a character
// (used for matching patterns)
b.TopLeft().x + ((b.TopRight().x - b.TopLeft().x) / 2),
// NIEUW
// All other results that we're found
// So we can switch between em
returnMatches
));
}
//Validate match with lexicon.
vector_int bestWord = new vector_int();
lexiconResult.confidence = lexicon.FindBestWord(match, bestWord, Optimize.OptimizeForMinimum);
for (int i = 0; i < bestWord.Count; i++)
{
string character = matcher.PatternName(bestWord[i]);
//Store lexicon result
lexiconResult.characters.Add(new LicenseCharacter(character));
}
// NIEUW
// Create the best match with the aid of the pattern matcher
result.CalculateBestMatch(matcher);
binaryCharacter.Dispose();
returnBlobs.Dispose();
match.Dispose();
bestWord.Dispose();
GC.Collect();
return(true);
}
