Saturday, 1 February 2014

Motion Detection 1

Motion Detection- part 1

Hello everybody. This post is dedicated to motion detection. Firstly I wanna thank Andrew Krillov for creating a wonderful library named Aforge.NET framework. In my project, I didn't use the predefined motion detection algorithms but developed one of my own. 

I am a normal second year student of B-Tech studying Electronics and Communication Engineering in Siliguri Institute of Technology under West Bengal University of Technology. I started this as a part of a mega project of "Automated Elephant Tracker". Well that's another story. While gong in depth of the later project I realized that I need to understand the basic logic of
1: Motion Detection
2: Face Detection
3: Face Recognition.




Firstly I will start with motion detection. Motion Detection is analysed or detected by calculating the difference between two frames. My first approach was too simple.

1: Make the first frame as the background frame
2: Make the current frame as the background frame.
3: Compute the difference between the two.
4: If the detection is beyond any certain level, motion is detected.

The above algorithm was implemented using Aforge.NET framework and was developed as a Windows form application. The code diplays the motioned areas in red but if the frame is moved, then the algorithm fails.

Advantages

1: Simple code
2: Easy to implement
3: Less complexity

Disadvantages

1: Doesn't works accurately
2: Algorithm may fail in certain cases.

The next part will be out by next week.