High-speed target recognition system based on spectral analysis and classification algorithm

In this paper, a fast recognition system is designed to realize the fast recognition of high speed moving targets through effective analysis of the tail flame spectrum characteristics. The system is mainly composed of tracking imaging for target alignment and spectral analysis for target identification. The discriminating factor recognition algorithm is studied to realize the fast recognition of high-speed targets. The multi-band peak mean and valley mean ratio relations are studied and the effective discriminating interval is set to ensure the fast recognition of high-speed targets. The influence of the target velocity on the acquired spectrum is analyzed and the function relation between tangential velocity and radial velocity which determines the spectral shift is given. The experimental detection was carried out at a range of 0.5/1.0/2.0/4.0 km and a speed of 10-100 m/s on a small number of rocket tail flares. The results show that the larger the attenuation degree is, the larger the spectral amplitude difference is and the smaller the spectral morphology change is. In the same sampling period, the spectral distribution of different velocities showed obvious deviation, but the distribution morphology showed no change. By calculating the range of valley mean and peak mean and the inclusion relation among factors, the high speed target can be recognized effectively