Research Of Target Classification Based On Feature Extraction

Surveillance radar ascends to an altitude of several thousand meters relying on aerostat. It can detect the "low slow small" target by lookdown. Therefore, aerostat surveillance radar is often used as safeguard equipment for the city major activities. But the ground target of high-speed moving and the "low slow small" target are very similar, the conventional detection of radar can't distinguish them. So the dense cars moving on the highway surrounding the city bring great difficulties to the detection of radar.In this paper, we proposed a scheme to suppress the interference of ground moving targets to the aerostat surveillance radar, by classifying the target through the target echo. We first complete the regular processing based on the target echo of the narrowband signal, then, filter out the ground moving target through the classifier. The features or the features combinations for classifying the targets are the key of this scheme.In this paper, we extract the features and the features combinations from the time domain and frequency domain of the data of determine targets. These features and the features combinations can effectively classify the targets, such as helicopters and unmanned aerial vehicle (UAV), cars and meteorological clutter, etc. The results show that the scheme has engineering application value.