A statistical method of recognizing small-sized images of three-dimensional objects

Abstract

In this paper a statistical approach to the recognition of small-sized images of three-dimensional objects is proposed. The problem of identifying the parameters of the statistical model for the recognition of vehicles on the terrain is formulated and solved. The basis for constructing the recognition system is a training sample. The resulting formula dependencies allow us to proceed to the solution of the problem of recognizing similar objects that are observed under different angles at other points different from those used in the identification phase.