esteban.da's profile - activity

Error building opencv with sfm module Hy, I have download the opencv and opencv_contrib sources from git and I'm trying

Hy,

I'm trying to create a delaunay triangulation from a set of point 3f. The 2 first coordinates are the coordinates in the plan and the third one is the value of my data at that point (the goal is to create isolines).

I have tried to use SubDiv2D but the problem is that it take point 2f and return a list of Vec6f and so I have to find the association of each point2f with my point3f. Is there a way to get the indice from the vector of input instead of the coordinates ?

Thanks

Hello, I'm trying to use the contourArea fonction in ordre to remove the "too small" shapes but when I'm running the above code I'm obtaining this king of results : "contour 1 color 125 area 1" or more generally "countour n color x area n". So si there someone having an idea about why I obtain the number of the contour instead of the real area delimitate by the contour ?

Thank you all by advance

#include <opencv2/core/core.hpp>
#include <opencv2/highgui/highgui.hpp>
#include <opencv2/imgproc/imgproc.hpp>
#include <iostream>
#include <stdio.h>
#include <stdlib.h>

using namespace cv;
using namespace std;

void drawStuff(int, void*);
void showInputWindow();
void showCannyWindow();
void showContourWindow();

int lowThreshold;
int const max_lowThreshold = 100;
const char* window_name = "Edge Map";

int thresh = 40;
int max_thresh = 120;
Mat img_rgb,img_gray,img_bw,canny_output,drawing;

int main(int argc, char *argv[]){
    img_rgb  = imread(argv[1]);
    blur( img_rgb, img_rgb, Size(3,3) );
    namedWindow( window_name, WINDOW_AUTOSIZE );
    cv::createTrackbar( "Min Threshold:", window_name, &lowThreshold, max_lowThreshold, drawStuff );

    drawStuff(0,0);
    cv::waitKey(0);
}

void drawStuff(int, void*){
    vector<vector<Point> > contours;
    vector<Vec4i> hierarchy;

    Canny( img_rgb, canny_output, lowThreshold, lowThreshold*2, 3 );
    cv::dilate(canny_output, canny_output, cv::Mat(), cv::Point(-1,-1));
    showCannyWindow();

    findContours( canny_output, contours, hierarchy, CV_RETR_TREE, CV_CHAIN_APPROX_SIMPLE, Point(0, 0) );
    drawing = Mat::zeros( canny_output.size(), CV_8UC3 );

    vector<Point> approxShape;
    vector<double> areas;
    for(size_t i = 0; i < contours.size(); i++){
            approxPolyDP(contours[i], approxShape, arcLength(Mat(contours[i]), true)*0.04, true);
        areas.push_back(contourArea(contours[i],false));
            drawContours(drawing, contours, i, Scalar((i+1)*10%255, (i+1)*10%255, (i+1)*10%255), CV_FILLED);   // fill one gray per contour
        printf("contour %d color %d area%d\n",i,(i+i)*10%255,areas[i]);
    }
    //cv::fastNlMeansDenoising(img_rgb,contours,3,7,21);

    showContourWindow();
}

void showInputWindow(){
    cv::namedWindow("InputImage");
    cv::imshow("InputImage",img_rgb);
}

void showCannyWindow(){
    cv::namedWindow("Canny");
    cv::imshow("Canny",canny_output);
}
void showContourWindow(){
    cv::namedWindow("Fill");
    cv::imshow("Fill",drawing);
}