Detecting the camera movement with Python and OpenCV

Detecting the movement of the camera can be achieved by getting the information from the pan and tilt. This video shows the demonstration of detecting the camera movement programmatically using the technique of optical flow.

This can be interfaced with many different applications where it is essential to keep track of the camera movements. One such example is background subtraction which is a technique used to detect moving foreground objects from a static camera. If the camera moves, the algorithm fails.

I have used the concepts from feature tracking and optical flow. Determining which features to track is the important part and I have taken features from the border region of the frame. This is for the obvious reason that the border region in a frame gives features from the background region.

This works really well but it is not a robust solution. A simpler solution is very much required.
Also, I'm waving in the video to indicate that the algorithm detects only camera movement and not foreground object motion.

Code: https://github.com/NagarajSMurthy/Auto-cam-movement-detection

––––––––––––––––––––––––––––––
We Are One by Vexento https://soundcloud.com/vexento
https://www.youtube.com/user/Vexento
Free Download / Stream: http://bit.ly/2PaIKcR
Music promoted by Audio Library https://youtu.be/Ssvu2yncgWU
––––––––––––––––––––––––––––––