void SGMStereo::setImageSize(const png::image<png::rgb_pixel>& leftImage, const png::image<png::rgb_pixel>& rightImage) {
width_ = static_cast<int>(leftImage.get_width());
height_ = static_cast<int>(leftImage.get_height());
if (rightImage.get_width() != width_ || rightImage.get_height() != height_) {
throw std::invalid_argument("[SGMStereo::setImageSize] sizes of left and right images are different");
}
widthStep_ = width_ + 15 - (width_ - 1)%16;
}
void otsuThresholding(png::image<png::gray_pixel> & image)
{
const int numPixels = image.get_width() * image.get_height();
const size_t maxPixelValue = static_cast<size_t>(
std::numeric_limits<png::gray_pixel>::max());
// histogram generation
std::vector<unsigned int> histogram(maxPixelValue + 1, 0);
for (size_t i = 0; i < image.get_width(); ++i)
for (size_t j = 0; j < image.get_height(); ++j)
++histogram[image.get_pixel(i,j)];
// computation of the probabilities of each density level
std::vector<double> prob(maxPixelValue + 1);
for (size_t i = 0; i <= maxPixelValue; ++i) {
prob[i] = static_cast<double>(histogram[i]) / numPixels;
}
// omega & mu computation
std::vector<double> omega(maxPixelValue + 1);
std::vector<double> mu(maxPixelValue + 1);
omega[0] = prob[0];
mu[0] = 0.0;
for (size_t i = 1; i <= maxPixelValue; ++i) {
omega[i] = omega[i-1] + prob[i];
mu[i] = mu[i-1] + i * prob[i];
}
// sigma maximization
// sigma stands for inter-class variance
// and determines optimal threshold value
std::vector<double> sigma(maxPixelValue + 1);
size_t threshold = 0;
double max_sigma = 0.0;
for (size_t i = 0; i <= maxPixelValue-1; ++i) {
if (omega[i] != 0.0 && omega[i] != 1.0)
sigma[i] = square(mu[maxPixelValue] * omega[i] - mu[i])
/ (omega[i] * (1.0 - omega[i]));
else
sigma[i] = 0.0;
if (sigma[i] > max_sigma) {
max_sigma = sigma[i];
threshold = i;
}
}
// binarization
for (size_t i = 0; i < image.get_width(); ++i)
for (size_t j = 0; j < image.get_height(); ++j)
if (image.get_pixel(i,j) > threshold)
image.set_pixel(i,j, std::numeric_limits<png::gray_pixel>::max());
else
image.set_pixel(i,j, 0);
}
png::image<png::rgb_pixel_16> bumpier::model::calculate_normals(const png::image<png::gray_pixel_16>& input, double phi, space_type space) const
{
int width = input.get_width(), height = input.get_height();
png::image<png::rgb_pixel_16> output(width, height);
for (int y = 0; y < height; y++)
for (int x = 0; x < width; x++)
{
int xmin = (x + width - 1) % width,
ymin = (y + height - 1) % height,
xmax = (x + 1) % width,
ymax = (y + 1) % height;
Eigen::Vector3f normal = (position(input, x, ymax) - position(input, x, ymin)).cross(
position(input, xmax, y) - position(input, xmin, y));
if(space == tangent_space)
normal = tangentspace(Eigen::Vector2f((float)x / width, (float)y / height)) * normal;
normal.normalize();
normal = (normal.array() * 0.5 + 0.5) * 0xFFFF;
output.set_pixel(x, y, png::rgb_pixel_16(normal[0], normal[1], normal[2]));
}
return output;
}
void dig_riverbed(int i, int j, png::gray_pixel_16 depth, png::image<png::gray_pixel_16>& topomap) {
png::gray_pixel_16 cur_val,new_val,brush_depth;
int brush_radius=2;
int width = topomap.get_width();
int height = topomap.get_height();
for (int i_=i-brush_radius; i_<=i+brush_radius; i_++) {
for(int j_=j-brush_radius; j_<=j+brush_radius; j_++) {
if (i_<0||i_>width||j_<0||j_>=height) continue;
double i_dist=i_-i;
double j_dist=j_-j;
double dist=sqrt(i_dist*i_dist+j_dist*j_dist);
if(dist>brush_radius) continue;
//double brush_depth_dbl = depth - depth*sin(dist/brush_radius*M_PI/2);
double brush_depth_dbl = depth*cos(dist/brush_radius*M_PI/2);
if(brush_depth_dbl < 0) brush_depth_dbl = 0;
brush_depth = brush_depth_dbl;
cur_val = topomap.get_pixel(i_,j_);
if(brush_depth >= cur_val) new_val = 0;
else
new_val = cur_val - brush_depth;
if (new_val < cur_val)
topomap.set_pixel(i_,j_,new_val);
}
}
}
void binaryImageToNegative(png::image<png::gray_pixel> & image)
{
for (size_t i = 0; i < image.get_width(); ++i)
for (size_t j = 0; j < image.get_height(); ++j) {
image.set_pixel(i,j, (image.get_pixel(i,j) == 0 ?
std::numeric_limits<png::gray_pixel>::max() : 0));
}
}
void
generate_image(png::image< pixel >& image)
{
typedef png::pixel_traits< pixel > traits;
size_t colors = 1 << traits::get_bit_depth();
size_t size = colors / 2;
image.resize(size, size);
png::palette palette(colors);
for (size_t c = 0; c < colors; ++c)
{
palette[c] = png::color(c * 255 / colors,
(colors - c - 1) * 255 / colors,
c * 255 / colors);
}
image.set_palette(palette);
for (size_t j = 0; j < image.get_height(); ++j)
{
for (size_t i = 0; i < image.get_width(); ++i)
{
image.set_pixel(i, j, i + j);
}
}
}
void makeSegmentBoundaryImage(const png::image<png::rgb_pixel>& inputImage,
const png::image<png::gray_pixel_16>& segmentImage,
std::vector< std::vector<int> >& boundaryLabels,
png::image<png::rgb_pixel>& segmentBoundaryImage)
{
int width = static_cast<int>(inputImage.get_width());
int height = static_cast<int>(inputImage.get_height());
int boundaryTotal = static_cast<int>(boundaryLabels.size());
segmentBoundaryImage.resize(width, height);
for (int y = 0; y < height; ++y) {
for (int x = 0; x < width; ++x) {
segmentBoundaryImage.set_pixel(x, y, inputImage.get_pixel(x, y));
}
}
int boundaryWidth = 2;
for (int y = 0; y < height - 1; ++y) {
for (int x = 0; x < width - 1; ++x) {
int pixelLabelIndex = segmentImage.get_pixel(x, y);
if (segmentImage.get_pixel(x + 1, y) != pixelLabelIndex) {
for (int w = 0; w < boundaryWidth - 1; ++w) {
if (x - w >= 0) segmentBoundaryImage.set_pixel(x - w, y, png::rgb_pixel(128, 128, 128));
}
for (int w = 1; w < boundaryWidth; ++w) {
if (x + w < width) segmentBoundaryImage.set_pixel(x + w, y, png::rgb_pixel(128, 128, 128));
}
}
if (segmentImage.get_pixel(x, y + 1) != pixelLabelIndex) {
for (int w = 0; w < boundaryWidth - 1; ++w) {
if (y - w >= 0) segmentBoundaryImage.set_pixel(x, y - w, png::rgb_pixel(128, 128, 128));
}
for (int w = 1; w < boundaryWidth; ++w) {
if (y + w < height) segmentBoundaryImage.set_pixel(x, y + w, png::rgb_pixel(128, 128, 128));
}
}
}
}
boundaryWidth = 7;
for (int y = 0; y < height - 1; ++y) {
for (int x = 0; x < width - 1; ++x) {
int pixelLabelIndex = segmentImage.get_pixel(x, y);
if (segmentImage.get_pixel(x + 1, y) != pixelLabelIndex) {
png::rgb_pixel negativeSideColor, positiveSideColor;
int pixelBoundaryIndex = -1;
for (int boundaryIndex = 0; boundaryIndex < boundaryTotal; ++boundaryIndex) {
if ((boundaryLabels[boundaryIndex][0] == pixelLabelIndex && boundaryLabels[boundaryIndex][1] == segmentImage.get_pixel(x + 1, y))
|| (boundaryLabels[boundaryIndex][0] == segmentImage.get_pixel(x + 1, y) && boundaryLabels[boundaryIndex][1] == pixelLabelIndex))
{
pixelBoundaryIndex = boundaryIndex;
break;
}
}
if (boundaryLabels[pixelBoundaryIndex][2] == 3) continue;
else if (boundaryLabels[pixelBoundaryIndex][2] == 2) {
negativeSideColor.red = 0; negativeSideColor.green = 225; negativeSideColor.blue = 0;
positiveSideColor.red = 0; positiveSideColor.green = 225; positiveSideColor.blue = 0;
} else if (pixelLabelIndex == boundaryLabels[pixelBoundaryIndex][boundaryLabels[pixelBoundaryIndex][2]]) {
negativeSideColor.red = 225; negativeSideColor.green = 0; negativeSideColor.blue = 0;
positiveSideColor.red = 0; positiveSideColor.green = 0; positiveSideColor.blue = 225;
} else {
negativeSideColor.red = 0; negativeSideColor.green = 0; negativeSideColor.blue = 225;
positiveSideColor.red = 225; positiveSideColor.green = 0; positiveSideColor.blue = 0;
}
for (int w = 0; w < boundaryWidth - 1; ++w) {
if (x - w >= 0) segmentBoundaryImage.set_pixel(x - w, y, negativeSideColor);
}
for (int w = 1; w < boundaryWidth; ++w) {
if (x + w < width) segmentBoundaryImage.set_pixel(x + w, y, positiveSideColor);
}
}
if (segmentImage.get_pixel(x, y + 1) != pixelLabelIndex) {
png::rgb_pixel negativeSideColor, positiveSideColor;
int pixelBoundaryIndex = -1;
for (int boundaryIndex = 0; boundaryIndex < boundaryTotal; ++boundaryIndex) {
if ((boundaryLabels[boundaryIndex][0] == pixelLabelIndex && boundaryLabels[boundaryIndex][1] == segmentImage.get_pixel(x, y + 1))
|| (boundaryLabels[boundaryIndex][0] == segmentImage.get_pixel(x, y + 1) && boundaryLabels[boundaryIndex][1] == pixelLabelIndex))
{
pixelBoundaryIndex = boundaryIndex;
break;
}
}
if (boundaryLabels[pixelBoundaryIndex][2] == 3) continue;
else if (boundaryLabels[pixelBoundaryIndex][2] == 2) {
negativeSideColor.red = 0; negativeSideColor.green = 225; negativeSideColor.blue = 0;
positiveSideColor.red = 0; positiveSideColor.green = 225; positiveSideColor.blue = 0;
} else if (pixelLabelIndex == boundaryLabels[pixelBoundaryIndex][boundaryLabels[pixelBoundaryIndex][2]]) {
negativeSideColor.red = 225; negativeSideColor.green = 0; negativeSideColor.blue = 0;
positiveSideColor.red = 0; positiveSideColor.green = 0; positiveSideColor.blue = 225;
} else {
negativeSideColor.red = 0; negativeSideColor.green = 0; negativeSideColor.blue = 225;
positiveSideColor.red = 225; positiveSideColor.green = 0; positiveSideColor.blue = 0;
}
for (int w = 0; w < boundaryWidth - 1; ++w) {
if (y - w >= 0) segmentBoundaryImage.set_pixel(x, y - w, negativeSideColor);
}
for (int w = 1; w < boundaryWidth; ++w) {
if (y+ w < height) segmentBoundaryImage.set_pixel(x, y + w, positiveSideColor);
}
}
}
}
}
Eigen::Vector3f bumpier::model::position(const png::image<png::gray_pixel_16>& image, int x, int y) const
{
return position(Eigen::Vector2f((float)x / image.get_width(), (float)y / image.get_height()), (float)image.get_pixel(x, y) / 0xFFFF);
}