The backpropagation ANN model with one hidden layer is employed and there are 48, 25, and 3 nodes in the input, hidden and output layers, respectively.

• 本研究构建的神经网路输入端有48个节点，隐藏层有25个节点，输出端则使用三个节点表示欲辨识的各类车种，并以倒传递网路的学习过程训练网路。