Multiclass 3-D aircraft identification and orientation estimation using multilayer feedforward neural network

Multilayer neural networks using the modified backpropagation learning algorithm are applied to achieve identification and orientation estimation of different classes of aircraft in a variety of 3-D orientations. 2-D distortion-invariant (L, Phi ) feature space was introduced for describing an aircraft image and used as the input of the neural network classifier. The optimum structure of the neural network was studied to obtain a high-performance classifier, and the reliability measure of the designed neural network classifier is introduced.<
>