Design of 3-D optimal FIR filters which extract objects moving along linear trajectory

We proposo a design method of optimal FIR filter which selectively extracts the particular moving object from other moving objects and noise. Stochastic approach is applied to the problem using the information of signals and the probability distribution of velocity vectors. In the method, the frequency response of the proposed Linear Trajectory Filter (LTF) specified by a priori information of the moving object's shape and its velocity vector. In addition, we derive a general formulation of the problem for optimal filter design and its solution for any signal and noise. Through some examples, it is shown that the target object is effectively enhanced in the noisy environment.