Target velocity identification using 3-D Matched Filter with Nelder-Mead optimization

This paper discusses the use of the 3-D Matched Filter to find the velocities of targets that are traveling through an image sequence. The 3-D Matched Filter is a Fast Fourier Transform-based filter that is especially effective at reducing the random noise relative to the energy of the moving target. By testing a number of velocity vectors, it is possible to determine the velocity of a target traveling through a very noisy image sequence. Using an exhaustive search to find the velocity vector is possible, but the length of time required to do this makes it impractical. Because of this, the use of an optimization technique to find the velocity vector can yield a massive reduction in time required to find the target´s velocity.