94 lines
2.1 KiB
C++
94 lines
2.1 KiB
C++
|
#include "hough.h"
|
||
|
|
#include "iostream"
|
||
|
|
#include "conversion_grayscale.h"
|
||
|
|
#include "edge_laplacian.h"
|
||
|
|
|
||
|
|
Hough::Hough(PNM* img) :
|
||
|
|
Transformation(img)
|
||
|
|
{
|
||
|
|
}
|
||
|
|
|
||
|
|
Hough::Hough(PNM* img, ImageViewer* super) :
|
||
|
|
Transformation(img, super)
|
||
|
|
{
|
||
|
|
}
|
||
|
|
|
||
|
|
PNM* Hough::transform()
|
||
|
|
{
|
||
|
|
int width = image->width();
|
||
|
|
int height = image->height();
|
||
|
|
|
||
|
|
int theta_density = getParameter("theta_density").toInt();
|
||
|
|
bool skip_edge = getParameter("skip_edge_detection").toBool();
|
||
|
|
|
||
|
|
std::cout << "skip_edge: " << skip_edge << std::endl;
|
||
|
|
int theta_size = 180 * theta_density;
|
||
|
|
double max_ro = sqrt(pow(width, 2) + pow(height, 2));
|
||
|
|
|
||
|
|
ConversionGrayscale* gray_scale = new ConversionGrayscale(image);
|
||
|
|
image = gray_scale->transform();
|
||
|
|
|
||
|
|
PNM* newImage = new PNM(theta_size, max_ro * 2 + 1, QImage::Format_Grayscale8);
|
||
|
|
|
||
|
|
int new_image_width = newImage->width();
|
||
|
|
int new_image_height = newImage->height();
|
||
|
|
|
||
|
|
math::matrix<float> hough(new_image_width, new_image_height);
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
if (skip_edge == false)
|
||
|
|
{
|
||
|
|
EdgeLaplacian *el = new EdgeLaplacian(image);
|
||
|
|
el->setParameter("size", 7);
|
||
|
|
image = el->transform();
|
||
|
|
}
|
||
|
|
|
||
|
|
for (int x = 0; x < new_image_width; x++)
|
||
|
|
{
|
||
|
|
for (int y = 0; y < new_image_height; y++)
|
||
|
|
{
|
||
|
|
hough[x][y] = 0;
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
|
||
|
|
double theta = 0.0;
|
||
|
|
double ro = 0;
|
||
|
|
float max = 0;
|
||
|
|
double max_z = 0;
|
||
|
|
|
||
|
|
for (int i = 0; i < width; i++)
|
||
|
|
{
|
||
|
|
for (int j = 0; j < height; j++)
|
||
|
|
{
|
||
|
|
if (qGray(image->pixel(i, j)) > 0)
|
||
|
|
{
|
||
|
|
for (int k = 0; k < theta_size; k++)
|
||
|
|
{
|
||
|
|
theta = (k * M_PI) / (theta_density * 180);
|
||
|
|
ro = i * cos(theta) + j * sin(theta);
|
||
|
|
hough[k][ro + max_ro]++;
|
||
|
|
}
|
||
|
|
}
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
max = hough.max();
|
||
|
|
|
||
|
|
int z;
|
||
|
|
|
||
|
|
for (int x = 0; x < new_image_width; x++)
|
||
|
|
{
|
||
|
|
for (int y = 0; y < new_image_height; y++)
|
||
|
|
{
|
||
|
|
z = hough[x][y] * (max / 255);
|
||
|
|
|
||
|
|
newImage->setPixel(x, y, QColor(z, z, z).rgb());
|
||
|
|
}
|
||
|
|
}
|
||
|
|
std::cout << "max_z: " << max_z << std::endl;
|
||
|
|
|
||
|
|
return newImage;
|
||
|
|
}
|