An Improved Pole Extraction Algorithm with Low Signal-to-Noise Ratio

The poles of target are important features of the radar resonance region. The pole feature is only related to the essential properties of the target, so using pole feature to identify targets can reduce the consideration of multiple external factors in the feature extraction process. Current pole extraction algorithms tend to work under conditions of high signal-to-noise ratio, which limits the application of pole features in practical scenarios. In this paper, based on the extraction of radar target pole features using the traditional Matrix Pencil Method, an improved Matrix Pencil Method is proposed, which uses the principal component analysis method to noise reduce the target echo, and then extracts the main components of the signal to eliminate the noise interference with lower energy, and after noise reduction, the Matrix Pencil Method combined with the ordering points to identify the clustering structure density clustering algorithm is used to remove the redundant false poles and extract the effective poles of the target.