int main (int argc, char const *argv[])
{
sourceImage = cvLoadImage("source.bmp", CV_LOAD_IMAGE_ANYDEPTH | CV_LOAD_IMAGE_ANYCOLOR);
objectImage = cvLoadImage("source.bmp", CV_LOAD_IMAGE_ANYDEPTH | CV_LOAD_IMAGE_GRAYSCALE);
templateImage = cvLoadImage("template.bmp", CV_LOAD_IMAGE_ANYDEPTH | CV_LOAD_IMAGE_GRAYSCALE);
cvNamedWindow(windowName, CV_WINDOW_AUTOSIZE);
cvCreateTrackbar(trackbarName, windowName, &level, 255, on_change);
while (1) {
int key = cvWaitKey(10);
templateMatching();
if ( key == 'q' ) {
break;
}
}
cvDestroyWindow(windowName);
return 0;
}
void on_change(int value) {
templateMatching();
}
void CTldAlgorithm::perform(CVideoLoader &loader, std::unique_ptr<CMeasuredData> &data)
{
cv::Mat templ = CPatternController::getInstance().getImgVec().begin()->second;
CPointContainer points;
tld::TLD tracker;
tracker.trackerEnabled = trackerEnabled;
tracker.alternating = alternating;
tracker.learningEnabled = learningEnabled;
bool isInitialized = false;
cv::Rect rect;
cv::Mat frame, frameGray;
int frameCount = 1;
std::chrono::time_point<std::chrono::system_clock> start, end;
while(true)
{
frame = loader.getNextFrame();
if(true == frame.empty())
{
break;
}
start = std::chrono::system_clock::now();
if(false == isInitialized)
{
isInitialized = true;
rect = templateMatching(templ, frame, 0);
frameGray = frame.clone();
cv::cvtColor(frame, frameGray, CV_BGR2GRAY);
tracker.detectorCascade->imgWidth = frameGray.cols;
tracker.detectorCascade->imgHeight = frameGray.rows;
tracker.detectorCascade->imgWidthStep = frameGray.step;
tracker.detectorCascade->varianceFilter->enabled = varianceFilterEnabled;
tracker.detectorCascade->ensembleClassifier->enabled = ensembleClassifierEnabled;
tracker.detectorCascade->nnClassifier->enabled = nnClassifierEnabled;
tracker.detectorCascade->nnClassifier->thetaTP = thetaTP;
tracker.detectorCascade->nnClassifier->thetaFP = thetaFP;
tracker.detectorCascade->useShift = useShift;
tracker.detectorCascade->shift = shift;
tracker.detectorCascade->minScale = minScale;
tracker.detectorCascade->maxScale = maxScale;
tracker.detectorCascade->minSize = minSize;
tracker.detectorCascade->numTrees = numTrees;
tracker.selectObject(frameGray, &rect);
}
if (1 == frameCount % 2)
{
tracker.processImage(frame);
}
if (tracker.currBB != NULL) {
if(nullptr != data)
{
points.addPoint(tracker.currBB->tl(),tracker.currBB->br());
data->addPredictPoints(tracker.currBB->tl(),tracker.currBB->br());
}
cv::rectangle(frame,tracker.currBB->tl(),tracker.currBB->br(),cv::Scalar(255, 0, 0),2, 8);
}
else if(nullptr != data)
{
// points.addPoint(cv::Point(-1, -1), cv::Point(-1, -1));
data->addPredictPoints(cv::Point(-1, -1), cv::Point(-1, -1));
}
end = std::chrono::system_clock::now();
std::chrono::duration<double> elapsedTime = end-start;
if(nullptr != data)
{
data->addComputeTime( elapsedTime.count() );
}
cv::imshow(m_winName, frame);
if(true == loader.interval(20)) break;
frameCount = (frameCount + 1) % 2;
}
//points.toFile(m_winName);
tracker.release();
}
