Learning OpenCV, 2nd Edition

cap >> image;

if( !image.data ) exit(0);

backgroundDiff( image, mask );

// A simple visualization is to write to the red channel

//

cv::split( image, Igray );

Igray[2] = cv::max( mask, Igray[2] );

cv::merge( Igray, image );

cv::imshow( "Example9_2", image );

if( cv::waitKey(7) == 0x20 ) break;

}

exit(0);

}

We’ve just seen a simple method of learning background scenes and segmenting foreground objects. It will

work well only with scenes that do not contain moving background components (it would fail with a waving

curtain or waving trees). It also assumes that the lighting remains fairly constant (as in indoor static

scenes). You can look ahead to

X

Figure 9-6X to check the performance of this averaging method.

12BAccumulating Means, Variances, and Covariances

The averaging background method just described made use the accumulation operator,

cv::Mat::operator+=() to do what was essentially the simplest possible thing: to sum up a bunch of

data that we could then normalize into an average. The average is a convenient statistical quantity for a lot

of reasons, of course, but one often overlooked advantage of the average is the fact that it can be computed

incrementally in this way.

F

7

F This means that we can do processing on line without needing to accumulate all

of the data before analyzing. We will now consider a slightly more sophisticated model, which can also ...