Target Detection in Clutter/Interference Regions Based on Deep Feature Fusion for HFSWR

High-frequency surface wave radar (HFSWR) is of great significance for maritime detection, but in the HFSWR echo signal, ship targets are often submerged a variety clutter and interference, making it difficult to detect vessels. In this paper, we propose an intelligent detection algorithm concealed strong complex interference environments. The has two stages: preprocessing target detection. stage, faster region-based convolutional neural networks Faster R-CNN designed identify locate regions range Doppler spectrum; two-level cascade proposed. First, extremum proposed suspicious points clutter/interference regions, including real false points, quickly obtain potential positions. Second, consideration characteristics targets, lightweight extract CNN features stacked autoencoder locations. Then, fusion obtained sent extreme learning machine that acts as second-level classifier distinguish between points. Experiments show target-detection better performance vessel than current mainstream algorithms.