Tutorial OpenCV - Histogramas

Cálculo de histograma de imágenes, aplicamos la técnica de ecualización
de histograma.

Author: devcarmelo

histogramas/CMakeLists.txt

@@ -0,0 +1,10 @@
+cmake_minimum_required(VERSION 2.8)
+
+find_package( OpenCV REQUIRED )
+
+project( Histogramas )
+
+file(COPY lena.jpg DESTINATION image)
+
+add_executable( ${PROJECT_NAME} histogram.cpp )
+target_link_libraries( ${PROJECT_NAME} ${OpenCV_LIBS} )

histogramas/histogram.cpp

@@ -0,0 +1,105 @@
+#include <opencv2\opencv.hpp>
+#include <iostream>
+#include <vector>
+
+using namespace cv;
+using namespace std;
+
+void drawLine(Mat histImage, int bin_w, int i, int height, Mat b_hist, Scalar color) {
+    line(histImage,
+        Point(bin_w*(i - 1), height - cvRound(b_hist.at<float>(i - 1))),
+        Point(bin_w*(i), height - cvRound(b_hist.at<float>(i))),
+        color);
+}
+
+void drawHistogram(cv::InputArray img)
+{
+    // Dividir los canales BRG
+    vector<Mat> bgr;
+    split(img, bgr);
+
+    // Crear un histograma con 256 bin (numero de subdivisiones) uno por pixel [0..256]
+    int numbins = 256;
+
+    // Establecer rango para los canales (B, G, R)
+    float range[] = { 0, 256 };
+    const float* histRange = { range };
+
+    Mat b_hist, g_hist, r_hist;
+
+    // Calcular el histograma para cada canal
+    calcHist(&bgr[0], 1, 0, Mat(), b_hist, 1, &numbins, &histRange);
+    calcHist(&bgr[1], 1, 0, Mat(), g_hist, 1, &numbins, &histRange);
+    calcHist(&bgr[2], 1, 0, Mat(), r_hist, 1, &numbins, &histRange);
+
+    // Tamano del histograma
+    int h_width = 512;
+    int h_height = 400;
+
+    // Crear una imagen para dibujar en ella
+    Mat histImage(h_height, h_width, CV_8UC3, Scalar(20, 20, 20));
+
+    // Normalizar los histogramas 
+    normalize(b_hist, b_hist, 0, h_height, NORM_MINMAX);
+    normalize(g_hist, g_hist, 0, h_height, NORM_MINMAX);
+    normalize(r_hist, r_hist, 0, h_height, NORM_MINMAX);
+
+    int bin_width = cvRound((float)h_width / (float)numbins);
+
+    // Dibujar cada una de las lineas
+    for (int i = 1; i < numbins; i++)
+    {
+        drawLine(histImage, bin_width, i, h_height, b_hist, Scalar(255, 0, 0));
+        drawLine(histImage, bin_width, i, h_height, g_hist, Scalar(0, 255, 0));
+        drawLine(histImage, bin_width, i, h_height, r_hist, Scalar(0, 0, 255));
+    }
+
+    // Mostrar el histograma
+    imshow("Histograma BGR", histImage);
+}
+
+void equalizeColorImage(cv::InputArray src, cv::OutputArray dst)
+{
+    // Convertir BGR image a YCrCb
+    Mat ycrcb;
+    cvtColor(src, ycrcb, COLOR_BGR2YCrCb);
+
+    // Extarer cada uno de los canales
+    vector<Mat> channels;
+    split(ycrcb, channels);
+
+    // Ecualizar histograma del canal Y (luminosidad) 
+    equalizeHist(channels[0], channels[0]);
+
+    // Unir los canales nuevamente
+    merge(channels, ycrcb);
+
+    // Convertir color YCrCb a BGR
+    cvtColor(ycrcb, dst, COLOR_YCrCb2BGR);
+}
+
+int main(int, char** argv)
+{
+    cv::Mat src, dst, hist;
+
+    std::string original = "Imagen Original";
+    std::string equalized = "Imagen Equalizada";
+
+    // leer la imagen
+    src = cv::imread("image/lena.jpg", 1);
+    if (src.empty()) return -1;
+
+    equalizeColorImage(src, dst);
+    drawHistogram(src);
+
+    //// convertir en escala de grises y ecualizar histograma
+    //cv::cvtColor(src, src, cv::COLOR_BGR2GRAY);
+    //cv::equalizeHist(src, dst);
+
+    // mostrar la imagen original y ecualizada
+    cv::imshow(original, src);
+    cv::imshow(equalized, dst);
+
+    cv::waitKey(0);
+    return 0;
+}