Label-dependent sparse representation for synthetic aperture radar target configuration recognition

ABSTRACT

Synthetic aperture radar (SAR) target images suffer from target aspect angle sensitivity. To overcome the obstacle that seriously influences recognition performance, a label-dependent sparse representation (LSR) algorithm is proposed to realize SAR target configuration recognition in the sparse domain. The label of the training sample is embedded into the sparse representation (SR) model, and dictionaries are constructed individually to eliminate disturbances. LSR is implemented according to a statistical model based on the Gaussian mixture distribution (GMD). Experiments are conducted on a wide range of moving and stationary target acquisition and recognition (MSTAR) databases. The experimental results demonstrate the effectiveness of the proposed algorithm, which outperforms other existing algorithms in terms of recognition accuracy.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the National Natural Science Foundation of China under Grant 61601274; the Fundamental Research Funds for the Central Universities under Grant GK201603089, GK201603083, and GK201603090.