Dimension Reduction of Bispectrum Using Singular Value Decomposition
Shunjun Wu · Journal of Astronautics · 2007
For target recognition based on radar high resolution range profile(HRRP),bispectrum is an efficient recognition feature,but its high dimension restrict its application in realistic.In this paper,bispectrum matrix of HRRP was decomposed by singular value decomposition,the singular values and singular vectors were used as the recognition features.Compared with selected bispectrum feature,they had a better recognition performance and needn't be computed again when there was a new target was added into the templates.Experiments based on measured data testify the efficiency of new feature.