codemonkey's profile - activity

I'm trying to get outlines of all of the color blobs in this image but between Canny and findContours, I'm not getting all of the contours. I attached the original image and one with the resulting contours drawn in white.

Here's my code:

// pad the image with a border of black pixels in case a hotspot is on the edge
cv::Mat matPadded = cv::Mat::zeros(cv::Size(matImage.size().width + 4, matImage.size().height + 4), CV_8UC3);
cv::Rect roi = cv::Rect(cv::Point(2, 2), matImage.size());
cv::Mat matROI = matPadded(roi);
matImage.copyTo(matROI);

cv::Mat matCanny;
cv::Canny(matPadded, matCanny, 50, 150);

// find contours
std::vector<std::vector<cv::Point>> aContours;
cv::findContours(matCanny, aContours, aHierarchy, CV_RETR_TREE, CV_CHAIN_APPROX_SIMPLE, cv::Point(0, 0));

Any tips on how I can get better results? I tried converting the image to grayscale first but it was even worse.

I'm getting this error:

error: (-215) dims <= 2 && data && (unsigned)i0 < (unsigned)size.p[0] && (unsigned)(i1*DataType<_Tp>::channels) < (unsigned)(size.p[1]*channels()) && ((((sizeof(size_t)<<28)|0x8442211) >> ((DataType<_Tp>::depth) & ((1 << 3) - 1))*4) & 15) == elemSize1()

I'm scanning through an image to determine what color the pixels are inside a polygon. I'm completely stumped as to what is causing the error. Any assistance would be great.

This is the code that is throwing the error (CVGColor is just a structure to hold RGB values):

cv::Rect boundingRect = cv::boundingRect(aPolygon);
cv::Mat matCropped = matImage(boundingRect);
std::vector<cv::Point> aPolygonTranslated;
for(size_t i=0; i<aPolygon.size(); i++)
{
  aPolygonTranslated.push_back(cv::Point(aPolygon[i].x - boundingRect.x, aPolygon[i].y - boundingRect.y));
}
for(int i=0; i<matCropped.cols; i++)
{
  for(int j=0; j<matCropped.rows; j++)
  {
    if(cv::pointPolygonTest(aPolygonTranslated, cv::Point(i, j), false) > 0)  // point is inside
    {
      CVGColor cvgPixel(matCropped.at<cv::Vec3b>(i, j)[2], matCropped.at<cv::Vec3b>(i, j)[1], matCropped.at<cv::Vec3b>(i, j)[0]);
      }
    }
}