Automatic Target Recognition Method of Flight Vehicle Based on Template Matching
Huihui Li, Qili Chen, Junfang Fan · 2021
The key factor in obtaining flight vehicle guidance information is the automatic recognition of strapdown images. When using strap-down image guidance, it is more difficult to obtain guidance information due to the influence of environment and other factors on the information collection. Aiming at the fact that deep learning algorithms cannot be directly applied to the target recognition of strapdown images, a new target recognition framework is proposed. The framework extracts features from target template images and strapdown images, and uses the extracted feature maps as the input of the deep neural network, so that the improved target recognition framework can be directly applied to the strapdown environment. Experimental results show that the proposed framework can effectively recognize strapdown images, improve recognition accuracy, and achieve precise guidance.