Development of Neural Network Recognition Model for Optical Illusion Images

Abstract

In the battlefield, tank armored vehicles have high mobility, strong defense and stable transportation capabilities, and play a role that cannot be ignored. However, its appearance and camouflage pattern design often have the camouflage effect of geometric optical illusion. The good recognition effect can ensure the safety of soldiers, and also avoid the loss of life and property of the people by accidentally hitting the surrounding houses. Therefore, how to quickly and effectively identify its appearance has become an urgent problem to be solved in the military. In this paper, the outline of the vehicle is disassembled into 12 most representative basic shapes for analysis, using image preprocessing technology, invariant moment feature extraction, and developing a BPNN(back propagation neural network) recognition model, which effectively improves the recognition rate of tank armored vehicles over above 95%.