Abstract
In order to solve the problem of low signal-to-noise ratio of target detection under strong sea clutter, an effective marine target detection algorithm based on YOLOv3 network is proposed. According to the characteristics of radar echo data, after data pre-processing the RGB image suitable for the detection network is constructed, and the number and aspect ratio of target anchor boxes are extracted by K-means clustering algorithm, and the network parameters are adjusted to obtain the optimized target detection model through multi-scale training, so as to detect targets in the sea clutter, and the target position is marked by the anchor boxes, provide effective information characteristics for the tracking method. The experimental results on the radar measured data show that the proposed method can suppress a large number of clutter and detect small and weak targets accurately, taking into account the speed and accuracy of detection, and verify the robustness of the algorithm. This approach simplifies the radar signal processing, reduces the computational complexity, and lays a foundation for the construction of intelligent radar.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call