Zadanie_12_Jonasz

src/core/transformations/bin_gradient.cpp

@@ -17,8 +17,42 @@ PNM* BinarizationGradient::transform()
 
     PNM* newImage = new PNM(width, height, QImage::Format_Mono);
 
-    qDebug() << Q_FUNC_INFO << "Not implemented yet!";
+    int Gmax = 0;
+    int numerator  = 0;
+    int denominator = 0;
+    for(int x=1; x<width-1; x++){
+        for(int y=1; y<height-1; y++){
+            int t = (qRed(image->pixel(x,y)) + qGreen(image->pixel(x,y)) + qBlue(image->pixel(x,y))) / 3;
+            int Gxa = (qRed(image->pixel(x-1,y)) + qGreen(image->pixel(x-1,y)) + qBlue(image->pixel(x-1,y))) / 3;
+            int Gxb = (qRed(image->pixel(x+1,y)) + qGreen(image->pixel(x+1,y)) + qBlue(image->pixel(x+1,y))) / 3;
+            int Gya = (qRed(image->pixel(x,y+1)) + qGreen(image->pixel(x,y+1)) + qBlue(image->pixel(x,y+1))) / 3;
+            int Gyb= (qRed(image->pixel(x,y-1)) + qGreen(image->pixel(x,y-1)) + qBlue(image->pixel(x,y-1))) / 3;
 
+            int Gx = Gxa - Gxb;
+            int Gy = Gya - Gyb;
+
+            if (Gx > Gy) Gmax = Gx;
+            else Gmax = Gy;
+
+            numerator  += t * Gmax;
+            denominator += Gmax;
+
+        }
+    }
+    int T = numerator/denominator;
+    // create new image
+    for(int i=0; i<width; i++){
+        for(int j=0; j<height; j++){
+            QRgb pixel = image->pixel(i,j);
+            int tempValue = (qRed(pixel) + qGreen(pixel) + qBlue(pixel)) / 3;
+            if ( tempValue > T ){
+                newImage->setPixel(i, j, Qt::color1);
+            }
+            else {
+                newImage->setPixel(i, j, Qt::color0);
+            }
+        }
+    }
     return newImage;
 }
 

src/core/transformations/blur_gaussian.cpp

@@ -41,9 +41,8 @@ math::matrix<float> BlurGaussian::getMask(int size, Mode)
 }
 
 float BlurGaussian::getGauss(int x, int y, float sigma)
-{    
+{
 
     float gauss= exp(-(x*x+y*y)/(2*sigma*sigma))/(2*M_PI*sigma*sigma);
     return gauss;
 }
-

src/core/transformations/convolution.cpp

@@ -137,7 +137,7 @@ const math::matrix<float> Convolution::join(math::matrix<float> A, math::matrix<
 
     for (int x=0; x<size; x++)
             for (int y=0; y<size; y++)
-           {     
+           {
                 C[x][y]=A[x][y]*B[x][y];
            }
 
@@ -181,4 +181,5 @@ const math::matrix<float> Convolution::reflection(const math::matrix<float> A)
         }
 
     return C;
+
 }

src/core/transformations/corner_harris.cpp

@@ -24,9 +24,115 @@ PNM* CornerHarris::transform()
     int width  = image->width(),
         height = image->height();
 
+    //Krok1 Zdef i wypelij zerami
+    math::matrix<float> Ixx(width, height);
+    math::matrix<float> Iyy(width, height);
+    math::matrix<float> Ixy(width, height);
+    math::matrix<float> CC(width, height);
+    math::matrix<float> CNS(width, height);
+
+
+    for (int x=0; x<width; x++)
+        for (int y=0; y<height; y++)
+           {
+             Ixx[x][y]=0;
+             Iyy[x][y]=0;
+             Ixy[x][y]=0;
+             CC[x][y]=0;
+             CNS[x][y]=0;
+           }
+
+    //Krok 2 skala szarosci
+    ConversionGrayscale cg(image);
+    PNM* gray_image=cg.transform();
+
+    //Krok 3 rozmyj
+    BlurGaussian bg(gray_image);
+    bg.setParameter("size",3);
+    bg.setParameter("sigma",1.6);
+    PNM* blur_image=bg.transform();
+
+    //Krok 4 sobel
+    EdgeSobel es(blur_image);
+    math::matrix<float> *Gx=es.rawHorizontalDetection();
+    math::matrix<float> *Gy=es.rawVerticalDetection();
+
+    //Krok 5 wzorek
+    for (int x=0; x<width; x++)
+        for (int y=0; y<height; y++)
+           {
+           Ixx[x][y]=(*Gx)[x][y] * (*Gx)[x][y];
+           Iyy[x][y]=(*Gy)[x][y] * (*Gy)[x][y];
+           Ixy[x][y]=(*Gx)[x][y] * (*Gy)[x][y];
+           }
+
+    //Krok 6 wzorek
+    for (int x=1; x<width-1; x++)
+        for (int y=1; y<height-1; y++)
+           {
+            float Sxx=0;
+            float Syy=0;
+            float Sxy=0;
+            math::matrix<float> H(2, 2);
+
+
+            for(int z=-1;z<=1;z++) //Znacznik sumy
+                for(int a=-1;a<=1;a++)
+                {
+                    Sxx+=Ixx[x+z][y+a]*BlurGaussian::getGauss(z,a,sigma);
+                    Syy+=Iyy[x+z][y+a]*BlurGaussian::getGauss(z,a,sigma);
+                    Sxy+=Ixy[x+z][y+a]*BlurGaussian::getGauss(z,a,sigma);
+                }
+
+            Sxx=Sxx/sigma_weight;
+            Syy=Syy/sigma_weight;
+            Sxy=Sxy/sigma_weight;
+
+
+            H[0][0]=Sxx;
+            H[0][1]=Sxy;
+            H[1][0]=Sxy;
+            H[1][1]=Syy;
+
+            float detH=H[0][0]*H[1][1]-H[0][1]*H[1][0];
+            //Wyznaczkik + suma po przekatnej kwadratowej
+            float r = detH - k_param * pow(H[0][0] + H[1][1], 2);
+
+            if(r>threshold) CC[x][y]=r;
+            }
+
+        bool search=1;
+
+        //Krok 8
+        while(search==1)
+        {
+            search=0;
+            for (int x=1; x<width-1; x++)
+                for (int y=1; y<height-1; y++)
+                {
+                 float max=CC[x][y];
+                 if(max > CC[x][y-1] && max > CC[x][y+1] && max > CC[x+1][y-1] && max > CC[x+1][y] && max > CC[x+1][y+1] && max > CC[x-1][y-1] && max > CC[x-1][y] && max > CC[x-1][y+1])
+                    CNS[x][y]=CC[x][y];
+                 else {
+                    if(CC[x][y]>0)
+                    {
+                     CNS[x][y]=0;
+                     search=1;
+                    }
+                    }
+                }
+        CC=CNS;
+        }
+
     PNM* newImage = new PNM(width, height, QImage::Format_Mono);
 
-    qDebug() << Q_FUNC_INFO << "Not implemented yet!";
+    //Krok 9
+    for (int x=1; x<width-1; x++)
+        for (int y=1; y<height-1; y++)
+        {
+            if(CC[x][y]==0) newImage->setPixel(x,y,0);
+            else newImage->setPixel(x,y,1);
+        }
 
     return newImage;
 }

src/core/transformations/edge_laplacian.cpp

@@ -10,9 +10,12 @@ EdgeLaplacian::EdgeLaplacian(PNM* img, ImageViewer* iv) :
 {
 }
 
-math::matrix<float> EdgeLaplacian::getMask(int, Mode)
+math::matrix<float> EdgeLaplacian::getMask(int size , Mode)
 {
-    int size = getParameter("size").toInt();
+
+    if( getParameter("size").toInt() > size )
+         size =getParameter("size").toInt();
+
     math::matrix<float> mask(size, size);
     int r=size/2;
 

src/core/transformations/edge_laplacian_of_gauss.cpp

@@ -12,11 +12,14 @@ EdgeLaplaceOfGauss::EdgeLaplaceOfGauss(PNM* img, ImageViewer* iv) :
 {
 }
 
-math::matrix<float> EdgeLaplaceOfGauss::getMask(int, Mode)
+math::matrix<float> EdgeLaplaceOfGauss::getMask(int size, Mode)
 {
-    size = getParameter("size").toInt();
+    if( getParameter("size").toInt() > size )
+         size =getParameter("size").toInt();
+
     double sigma = getParameter("sigma").toDouble();
 
+
     math::matrix<float> mask(size, size);
     int r=size/2;
 

src/core/transformations/edge_sobel.cpp

@@ -51,7 +51,7 @@ math::matrix<float>* EdgeSobel::rawHorizontalDetection()
         for (int y = 0; y < height; y++)
         {
             math::matrix<float> temp = getWindow(x, y, 3, LChannel, NullEdge);
-            x_gradient[x][y]  = Convolution::sum(Convolution::join(g_x, temp));
+            (*x_gradient)(x, y)  = Convolution::sum(Convolution::join(g_x, temp));
 
         }
     }
@@ -71,7 +71,7 @@ math::matrix<float>* EdgeSobel::rawVerticalDetection()
                 for (int y = 0; y < height; y++)
                 {
                     math::matrix<float> temp = getWindow(x, y, 3, LChannel, NullEdge);
-                    y_gradient[x][y] = Convolution::sum(Convolution::join(g_x, temp));
+                  (*y_gradient)(x, y) = Convolution::sum(Convolution::join(g_x, temp));
                 }
             }
 

src/core/transformations/hough.cpp

@@ -17,7 +17,50 @@ PNM* Hough::transform()
 {   
     int thetaDensity = getParameter("theta_density").toInt();
 
-    qDebug() << Q_FUNC_INFO << "Not implemented yet!";
+     int width = image->width();
+     int height = image->height();
+     PNM* tempImage = new PNM(width, height, QImage::Format_Indexed8);
+     tempImage = ConversionGrayscale(image).transform();
 
-    return 0;
+     if (!getParameter("skip_edge_detection").toBool()) {
+         tempImage = EdgeLaplacian(tempImage).transform();
+     }
+
+     int pmax =int(sqrt(height*height + width*width));
+     int thetaSize= 180 * thetaDensity;
+     int lenght =  pmax * 2 + 1;
+     double theta, p;
+     int max = 0;
+
+     math::matrix<int> hough(thetaSize,lenght);
+
+     for (int i = 0; i < thetaSize; i++) {
+         for (int j = 0 ; j < lenght ; j++) {
+             hough[i][j] = 0;
+         }
+     }
+
+     for (int i = 0; i < width; i++) {
+         for (int j = 0; j < height; j++) {
+             if (qGray(tempImage->pixel(i, j)) > 0) {
+                 for (int k = 0; k < thetaSize; k++) {
+                     theta = (k * 3.14) / (180 * thetaDensity);
+                     p = i*cos(theta) + j*sin(theta);
+                     hough[k][p+pmax]++;
+                     if (hough[k][p+pmax] > max) {
+                         max = hough[k][p+pmax];
+                     }
+                 }
+             }
+         }
+     }
+
+     PNM* newImage = new PNM(thetaSize,lenght, QImage::Format_Grayscale8);
+     for (int i = 0; i < thetaSize; i++) {
+         for (int j = 0; j < lenght; j++) {
+             int value = int(((hough[i][j] * 255)/ max));
+             newImage->setPixel(i, j, QColor(value,value,value).rgb());
+         }
+     }
+     return newImage;
 }

src/core/transformations/hough_lines.cpp

@@ -24,7 +24,34 @@ PNM* HoughLines::transform()
 
     PNM* newImage = new PNM(image->copy());
 
-    qDebug() << Q_FUNC_INFO << "Not implemented yet!";
-
+    // step 1
+    EdgeLaplacian* edgeLA = new EdgeLaplacian(image);
+    edgeLA->setParameter("size", 3);
+    PNM* tempImage = edgeLA->transform();
+    // step 2
+    BinarizationGradient* binarizationGradient = new BinarizationGradient(tempImage);
+    PNM* binImage = binarizationGradient->transform();
+    // step 3
+    Hough* hough = new Hough(binImage);
+    hough->setParameter("theta_density", 3);
+    hough->setParameter("skip_edge_detection", true);
+    tempImage = hough->transform();
+    // step 4
+    QPainter p(newImage);
+    p.setPen(Qt::red);
+    for(int i=0; i < tempImage->width(); i++) {
+        for(int j=0; j < tempImage->height(); j++) {
+            if(qGray(tempImage->pixel(i, j)) > threshold) {
+                double rtheta = ((double)i/3.0)*M_PI/180.0;
+                int rrho = j - tempImage->height()/2;
+                p.drawLine(0, round(rrho/sin(rtheta)), newImage->width()-1, round((rrho - (newImage->width()-1)*cos(rtheta))/sin(rtheta)));
+            }
+            if (!drawWholeLines == true){
+                if (qRed(binImage->pixel(i, j)) == 0){
+                    newImage->setPixel(i, j, image->pixel(i, j));
+                }
+            }
+        }
+    }
     return newImage;
 }