Ballistic Missile Target Recognition Based on Improved Bispectra Feature

According to the characteristic of ballistic missile target recognition, the reduced-dimensional bispectra methods that are applied in high-resolution range profile recognition are studied, and an improved selected bispectra method based on extended Fisher class separability discriminant measure is proposed. The improved selected bispectra features are evaluated using computer simulation experiment. Compared to conventional selected bispectra and integrated bispectra methods, the improved selected bispectra method is highly discriminative and provides proper recognition of high-resolution radar ballistic missile target range profiles.