Tracer l'Histogramme d'une Image couleur en Image avec Opencv

La OpenCV 2.3 documentation présente un nouveau tutoriel sur tracer un histogramme d'une image couleur.

Voir vz::plot_histogram(), écrit en C++. https://www.ccoderun.ca/programming/doxygen/vz_imagination/namespacevz.html#a4cf3b5cd26b8cc8ac28fdbb8355f56a5

Dans le cas où le lien est mort, le code ressemble à ceci:

//"src" is the cv::Mat input image
const size_t number_of_channels = src.channels();
const cv::Scalar background_colour(0,0,0);
std::vector<cv::Mat> split;
cv::split(src, split);
const int       height          = 480;
const int       width           = 640;
const int       histogram_size  = 256; //the number of "bins"
const float     range[]         = {0, 256}; //upper bound is exclusive, meaning 0-255
const float *   ranges          = {range};
const bool      uniform         = true;
const bool      accumulate      = false;
cv::Mat         mask;
//prepare the destination image
const int margin        = 3;
const int min_y         = margin;
const int max_y         = height - margin;
const int thickness     = 1;
const int line_type     = cv::LINE_AA;
const float bin_width   = static_cast<float>(width) / static_cast<float>(histogram_size);
cv::Mat dst(height, width, CV_8UC3, background_colour); //create the output image, starting with a pure colour
cv::Scalar colours[] =
{
{255, 0, 0},    //blue
{0, 255, 0},    //green
{0, 0, 255}     //red
};
if (number_of_channels == 1)
{
//for greyscale images, we only have a single colour channel, so ignore
//the RGB colour definitions and use either black or white for the histogram
colours[0] = (background_colour == cv::Scalar(0, 0, 0)) ?
cv::Scalar(255, 255, 255) :
cv::Scalar(0, 0, 0);
}
//iterate through all the channels in this image
for (size_t idx=0; idx < split.size(); idx++)
{
const cv::Scalar colour = colours[idx % 3];
cv::Mat & m = split[idx];
cv::Mat histogram;
cv::calcHist(&m, 1, 0, mask, histogram, 1, &histogram_size, &ranges, uniform, accumulate);
cv::normalize(histogram, histogram, 0, dst.rows, cv::NORM_MINMAX);
for (int i = 1; i < histogram_size; i++)
{
const int x1 = std::round(bin_width * (i - 1));
const int x2 = std::round(bin_width * (i - 0));
const int y1 = std::clamp(height - static_cast<int>(std::round(histogram.at<float>(i - 1))), min_y, max_y);
const int y2 = std::clamp(height - static_cast<int>(std::round(histogram.at<float>(i - 0))), min_y, max_y);
cv::line(dst, cv::Point(x1, y1), cv::Point(x2, y2), colour, thickness, line_type);
}
}
//histogram is in "dst"

Résultant histogrammes devrait ressembler à quelque chose comme ceci:

J'ai appris de ces, ils pourraient aider ->

http://docs.opencv.org/doc/tutorials/imgproc/histograms/histogram_calculation/histogram_calculation.html

et

http://docs.opencv.org/doc/tutorials/imgproc/histograms/histogram_comparison/histogram_comparison.html