Radar Target Recognition Architecture with Automatic Sample Annotation Capability

Radar data acquisition constrained by uncertainties requires that the radar target recognition system needs to have dynamic learning capability. As a key, automatic sample annotation directly restricts the practicalization of radar target recognition technology, which is mainly based on the supervised learning method. Aiming at the automatic sample annotation problem in radar target recognition based on high resolution range profiles, a radar target recognition architecture with automatic sample annotation capability is proposed by combining traditional target recognition technology with deep learning technology, which can effectively improve the robustness of radar target recognition system. The experimental results based on the measured data show that the proposed architecture can efficiently complete the automatic annotation of data samples required for training radar target recognition models, accelerate the application of radar target recognition under the condition of insufficient data, support the online learning of radar target recognition, and have strong engineering practicability.