Best way not to have to copy a Mat back into an OutputArray

outputarray

asked 2013-10-02 10:10:18 -0600

changelog
16 ●2 ●5

I have this code that basically does a "dumb" background subtraction on two frames.

void FrameDifferenceBGS::operator()(cv::InputArray _image, cv::OutputArray _fgmask, double learningRate)
{
  cv::Mat img_input = _image.getMat();

  if(img_input.empty())
    return;

  _fgmask.create(img_input.size(), CV_8U);
  cv::Mat img_foreground = _fgmask.getMat();

  if(img_input_prev.empty())
  {
    img_input.copyTo(img_input_prev);
    return;
  }

  cv::absdiff(img_input_prev, img_input, img_foreground);

  if(img_foreground.channels() == 3)
    cv::cvtColor(img_foreground, img_foreground, CV_BGR2GRAY);

  if(enableThreshold)
    cv::threshold(img_foreground, img_foreground, threshold, 255, cv::THRESH_BINARY);

  if(showOutput)
    cv::imshow("Frame Difference", img_foreground);

  img_input.copyTo(img_input_prev);
  img_foreground.copyTo(_fgmask);
  firstTime = false;
}

If I don't add img_foreground.copyTo(_fgmask) in the end, the output array isn't updated with the result of img_foreground, resulting on a black image when this is called.

What am I doing wrong / should be doing here?

edit retag flag offensive close merge delete

add a comment

3 answers

Sort by » oldest newest most voted

answered 2013-10-03 02:15:00 -0600

Michael Burdinov
4638 ●6 ●33 ●86

I think that the "Best way not to have to copy a Mat back into an OutputArray" is not using OutputArray in the first place. InputArray and OutputArray are used by OpenCV functions that can work with various different types. For example output can be either vector<point> or Mat. OpenCV users can simply pass Mat to such function. You shouldn't use those InputArray or OutputArray unless you have some kind of polymorphic behavior that you must implement. It seems to me that this is not the case.

edit flag offensive delete link

Comments

I'm doing it this way because I need it to behave the same way as any other instance of BackgroundSubtractor.

changelog ( 2013-10-03 02:54:48 -0600 )edit

I see. Then you should take @Moster advice and use _image and _fgmask directly. There no need to create Mat headers for them. As far as I know you can pass OutputArray to parameter that expect InputArray without any problem.

Also you don't need to call create() function, OpenCV will do this for you whenever needed.

Michael Burdinov ( 2013-10-03 04:40:15 -0600 )edit

add a comment

answered 2013-10-02 11:10:45 -0600

Moster
1716 ●5 ●24

updated 2013-10-02 11:20:49 -0600

I guess you are giving your function 2 Mat objects as an input, right?

Wouldnt it work, if you just use your outputArray in the last function that manipulates the img_foreground like this:

cv::threshold(img_foreground, _fgmask.getMat(), threshold, 255, cv::THRESH_BINARY);

edit: just using _fgmask instead of _fgmask.getMat() should also work. I havent used cv::OutputArray yet, but I dont see why not.

edit flag offensive delete link

add a comment

answered 2019-11-16 04:31:36 -0600

Freeman
61 ●6 ●2

The more efficient approach is not to copy a Mat back into an OutputArray output, but to get a reference to the (previously sized) matrix and populate it:

cv::Mat& A = output.getMatRef();

edit flag offensive delete link

add a comment

Best way not to have to copy a Mat back into an OutputArray

3 answers

Comments

Links

Question Tools

Stats

Related questions

Best way not to have to copy a Mat back into an OutputArray edit

3 answers

Comments

Links

Question Tools

Stats

Related questions

Best way not to have to copy a Mat back into an OutputArray