Edge_Drawing

Implementation of Edge Drawing Algorithm with Qt, which can produce high-quality edge segments, which are clean, well-localized and one-pixel wide, refering to the paper[1].

This project is implemented by RyuZhihao123 using Qt. Users can get egdes which are one pixel wide and clean from an input image.

Reference:

[1] Cihan Topal, Cuneyt Akinlar. 'Edge Drawing:A combined real-time edge and segment detector'.

Requirments:

ONLY NEEDS Qt. NO NEED for OpenCV etc.

Example Results:

Different Edge is showed in different colors. Some other results. all the edges are showed in the same color.

Usage:

int main(int argc, char *argv[])
{
    // Notice that choose valid paths.
    QString inputPath = "C:/Users/LiewDennis/Desktop/default.jpg";
    QString savePath  = "C:/Users/LiewDennis/Desktop/result.jpg";

    EdgeDrawing edgeDrawing;

    // Test for 'getEdgesFromImage()'
    QVector<QVector<QPoint>> edgeVector =edgeDrawing.getEdgesFromImage(QImage(inputPath),5,30,8);
    std::cout<<"Edge Amount: "<<edgeVector.size()<<std::endl;

    // Test for 'getEdgeImage()'
    QImage edgeImg = edgeDrawing.getEdgeImage(QImage(inputPath),5,30,8);
    edgeImg.save(savePath);

    return 0;
}

Others:

(1) Calling getEdgesFromImage() will return a vector of _EDGE.
QVector<QVector<QPoint>> is used to describe the edge information, which is a 2-dimentional vector. And each QVector<QPoint> is used to define one single edge.

QVector<_EDGE> getEdgesFromImage(const QImage& img,int gaussR=5,int sobelThreshold=30,int archorThreshold=60);

(2) Calling getEdgeImage() will return an edge image.

QImage getEdgeImage(const QImage &img, int gaussR=5,int sobelThreshold=30,int archorThreshold=60);