基于时频脊-Radon变换的海面小目标检测方法

针对海杂波背景下雷达对海面弱小目标探测技术难题，提出一种基于时频脊-Radon变换的帧平滑双特征检测方法。该方法首先采用分块白化方法对海杂波进行抑制，根据抑制后海杂波单元与目标单元在时频域二维平面呈现的不同能量分布特征，对时频脊做Radon变换并在变换域特征空间中提取出峰值和频宽特征。考虑到实际雷达采用的相干脉冲数通常较少（大多情况下为64个或更少），易导致特征不稳定，进而影响海杂波与目标可分性，为此，通过多帧扫描历史数据和当前帧数据的综合应用，对特征做帧平滑处理以增强可分性。最后采用凸包分类算法，在双特征平面进行融合检测。经2级和4级海况实测数据验证，在同等参数条件下，本文检测方法相比已有基于时频三特征的检测方法、频域CFAR检测方法和分形特征检测方法具有更好的检测性能。 

A method for detecting small targets in sea surface based on Ridges-Radon Transform

Aiming at the technical difficulty of radar to detect the small targets embedded in the sea clutter, a frame smoothing dual-feature detection method based on Ridges-Radon transform was proposed. Firstly, the block whitening method is used to suppress the sea clutter. According to the different energy distribution characteristics of the sea clutter and target cell in the time-frequency plane, the Radon transform is performed on the time-frequency ridge so as to extract the peak and bandwidth features. Considering that the number of coherent pulses used by the actual radar is usually less (64 or less in most cases), it is easy to lead to the instability of the features and affect the separability of the sea clutter and targets. Therefore, the frame smoothing method which is based on the comprehensive application of multi-frame scanning historical data and current frame data is carried out to enhance the separability. Finally, the convex hull classification algorithm is used to perform fusion detection in the dual feature plane. It is verified by the measured data of level 2 and level 4 sea conditions that, under the same parameters, the proposed detection method has better detection performance than the existing detection method based on three time-frequency features、fractal feature and the CFAR detection method in frequency domain.