Multi-Feature Fusion Sea Surface Floating Small Target Detection Based on Bagging Decision Tree

In order to improve the detection performance of small targets under sea clutter background, a detection method based on bagging decision tree with multi-feature information fusion is proposed. First, the time-frequency domain characterization of the standard deviation ratio of the average range of the pulse and amplitudes (SDRARPA) is proposed to describe the discrete echo signal sharpness. Second, the frequency peak-to-average standard deviation ratio (FPASDR) is proposed to optimize the frequency peak-to-average ratio (FPAR) feature. The first two features and the average relative average amplitude (ARAA) form a three-feature information fusion. Weak target detection at sea surface using false alarm controllable bagging decision tree detector based on bagging decision tree classification method. Finally, based on the IPIX measured data, the performance of the algorithms is compared and analyzed for the radar signal data under the four polarization modes. Experimental results show that the detection method proposed in this paper obtains 98.1% and 98.8% average detection probability in 1.024 s and 2.048 s observation time, respectively. There is a significant improvement in the detection performance compared to the traditional detection methods based on convex packet detector and SVM detector.