"Real Time Tracking Algorithm On Low Cost Fpga" consists of two parts: Part A: A system that streams video to DE2 board and displays them on a CRT/LCD screen Part B: Implementation of a video tracking algorithm on FPGA both in hardware and software.
Identifying movement in a video stream is basically done by comparing two consecutive frames. In our project we subtracted every two consecutive frame. The result of this kind of calculation is an image array where each pixel value reflects the difference between the two frames.
Assuming:
- Only one object is moving in the picture.
- A small area of the picture is covered by this object.
- The rest of the picture remains unchanged.
Calculating the 2D mean will retrieve the coordinates of the center of mass of the moving object.