Easy but not basic background subtraction in opencv?

Question

I have to do almost anything in Effective Background Subtraction with OpenCV (subtracting the background from the foreground with color, with the exception of the camera, not the video file). The problem is that in this topic there is no explanation of the background subtraction phase itself.

I watched the official openCV book on the Internet, and a simple Frame Differencing is not enough for what I need. I tried to understand the more complex Averaging background method , but I get lost after the cvAcc frames to get the average value: /

If anyone could help me, I would really appreciate it.

Thanks!

EDIT with the code I have:

Amount

cvCvtScale( currentFrame, currentFloat, 1, 0 ); if(totalFrames == 0) cvCopy(currentFloat, sum); else cvAcc(currentFloat, sum); 

average

 cvConvertScale( sum, imgBG, (double)(1.0/totalFrames) ); 

adapted background (with alpha 0.05 in #define)

 cvRunningAvg(currentFrame, imgBG, alpha); 

Creating the final image with foregrond only (far from ideal!)

 void createForeground(IplImage* imgDif,IplImage * currentFrame) { cvCvtColor(imgDif, grayFinal, CV_RGB2GRAY); cvSmooth(grayFinal, grayFinal); cvThreshold(grayFinal, grayFinal, 40, 255, CV_THRESH_BINARY); unsigned char *greyData= reinterpret_cast<unsigned char *>(grayFinal->imageData); unsigned char *currentData= reinterpret_cast<unsigned char *>(currentFrame->imageData); unsigned char *fgData= reinterpret_cast<unsigned char *>(currentFrame->imageData); int i=0; for(int j=0 ; j<(grayFinal->width*grayFinal->height) ; j++) { if(greyData[j]==0) { fgData[i]=0; fgData[i+1]=0; fgData[i+2]=0; i=i+3; } else { fgData[i]= currentData[i]; fgData[i+1]= currentData[i+1]; fgData[i+2]= currentData[i+2]; i=i+3; } } cvSetData( imgFG , fgData , imgFG->width*imgFG->nChannels); } 

PROBLEM NOW!

The biggest problem now is that when I have a small arrow somewhere in the picture, after I hold my hand “on top” of it for several seconds, when I remove it, the light keeps in the foreground a lot of time .. any help on this?

Answer 1

Simple background subtraction methods (pre-selection methods).

1. Let's say if your video has a static background (that is, small changes in the background have an almost constant background), then consider the N number of frames and average it. Then U can get a background image, i.e. Img_BG .

  Img_BG = (1/N)*sum(framesFrom1 to N); 

2. If your video with any backlight changes from time to time, your background image is updated as follows, it is called launch avg.

  Img_ApdatedBG =(1- alpha)*Img_BG+(alpha)*CurrentFrame; Img_BG = Img_ApdatedBG; 

alpha is the weight giving the current frame, usually between 0.05 and 0.1. This method uses less memory compared to method 1.

3.You can calculate the background image as the calculation of median of N frames.

Answer 2

You can try this implementation of background subtraction with a mixture of Gaussian . Wiki explains the OpenCV features used at every step.

This is a reliable method of subtracting the background, which relates to light changes and objects going in and out of the scene.

Hope this helps.