2

I have a 12 hours of videos with 9 channel that I need to review.

Is there a software that would allow me to know where in the video there was movement or at least anything to help in the process?

I don't need to edit the video just to identify when there was someone in the cameras angles. (I would assume that movement is the best way.)

1
  • What operating system?
    – meisyal
    Commented Nov 5, 2016 at 4:21

4 Answers 4

1

You can run your video through a motion detection & tracking process using python & OpenCV as in this tutorial.

  • Both are Free, Gratis & Open Source.
  • Both are cross platform and will run on Windows, OS-X, Linux or even Raspberry Pi platforms.
  • The referenced tutorial uses just 90 lines of code & no compiler.

The tutorial does not give an output of the times that the event(s) occurred but if you use:

# Find OpenCV version
(major_ver, minor_ver, subminor_ver) = (cv2.__version__).split('.')

if int(major_ver)  < 3 :
    fps = camera.get(cv2.cv.CV_CAP_PROP_FPS){0}".format(fps)
else :
    fps = camera.get(cv2.CAP_PROP_FPS)

Note that I have used camera in the above to agree with the tutorial but getting fps will not work from an actual web camera see here.

Used once the file is opened to get the frames per second and then simply keep a count of the frames processed you can then compute the time point within the video and print that out with an appropriate message, and/or save it to a text file, whenever the status changes to occupied or unoccupied.

2
  • 2
    I did that tutorial, and while it does the hardest part, it does not go as far as telling you "Movement detected at 5:41, 27:45, 3:47:56" for instance. So without a bit of development the user would still have to stay for 12 hours staring at the computer.
    – Nicolas Raoul
    Commented Nov 4, 2016 at 6:59
  • 1
    @NicolasRaoul - added some details on the changes needed. Commented Nov 4, 2016 at 7:26
1

If you're looking to develop an application to handle this, I would recommend checking out the LEADTOOLS Multimedia SDK (Windows-only) to implement this type of task. It supports C, C++, .NET and will work with other languages that support COM objects. This SDK also includes a DirectShow Motion Detection filter. When the filter detects motion, it invokes a callback to provide the area within the video the motion was detected. Using this callback, you could also find out when in the video this occurred.

The filter has options for:

  • changing sensitivity
  • object size
  • enabling/disabling viewing the rectangles where motion has been detected
  • a region of interest
  • and more

With this filter, it would be possible to not only run it while capturing, but also after the recording has occurred as you've indicated you need to do.

Disclaimer: I am an employee of the company that wrote this library.

0
0

This may not be an ideal solution but just a suggestion. Using a software like ManyCam you can emulate a web cam, then using a software like iSpy, you could add your video feed to your emulated webcam in ManyCam and then add that emulated webcam to iSpy and use iSpy's motion detection. Manycam may have the ability to fast forward but it seems it is limited in video feeds you can add unless you purchase.

0

I spy is aiding me in a similar task, it is freeware and you can analyze cameras and videos for motion or change with a connected vlc-player. You don't necessarily need ManyCam for it. It has a lot of possibilities to adjust settings like sensitivity, color of objects and also define zones.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.