detecting only slow moving objects using opencv

Question

I am using OpenCV for motion detection and using back ground subtraction algo for this. I got the following code from net.

cv::Mat frame;
    cv::Mat back;
    cv::Mat fore;
    cv::VideoCapture cap(0);
    bg.nmixtures = 3;
    bg.bShadowDetection = false;
const int history = 5;
cv::BackgroundSubtractorMOG2 bg (history,nmixtures,bShadowDetection);

    std::vector<std::vector<cv::Point> > contours;

    cv::namedWindow("Frame");
    cv::namedWindow("Background");

    for(;;)
    {
        cap >> frame;
        bg.operator ()(frame,fore);
        bg.getBackgroundImage(back);
        cv::erode(fore,fore,cv::Mat());
        cv::dilate(fore,fore,cv::Mat());
        cv::findContours(fore,contours,CV_RETR_EXTERNAL,CV_CHAIN_APPROX_NONE);
        cv::drawContours(frame,contours,-1,cv::Scalar(0,0,255),2);
        cv::imshow("Frame",frame);
        cv::imshow("Background",back);
        if(cv::waitKey(30) >= 0) break;
    }

So can i set a threshold so that if the change in new and old frame is more than the threshold then dont do anything. Or may be some other algorithm that should suit my situation of capturing only slow moving object.

score 0 · Answer 1 · answered Apr 26 '14 at 13:19

0

You can change the history value (increase it) in the mixture of gaussian model if you wish to detect slow moving objects.

answered Apr 26 '14 at 13:19

Bharat

2,139
2
16
35

I tried that actually , but in that case it captures both slow and fast movement. – user2588495 Apr 26 '14 at 15:05
You can compute for two thresholds and subtract the two to get the motion for only the slow one in that case. – Bharat Apr 26 '14 at 20:45

Andrey Smorodov · Answer 2 · 2014-04-26T16:28:46.260

0

You can try to use moving average of frames instead of using every frame as input for BG subtraction. Or use moving average to output of BG subtraction, then binarize by thresholding.

See addWeighted and moving Average (see Cumulative moving average).

Integration will reduce influence of fast changes.

edited Apr 26 '14 at 16:28

answered Apr 26 '14 at 16:11

Andrey Smorodov

10,649
2
35
42

detecting only slow moving objects using opencv

2 Answers2