Currently, sea-surface small targets have become the focus and difficulty of marine radar detection. The existing detection methods are limited to the use of unilateral information such as radar echo amplitude or spectrum, making it difficult to effectively detect small targets. Thus, a sea-surface small target detection method using dual-channel joint graph features (DC-JGF) is proposed in this paper. Firstly, the time-domain phase sequence and the frequencydomain amplitude sequence are extracted from the radar complex echo sequence to generate the time-frequency domain dual channel. In each channel, graphs are generated separately to provide rich correlation information. Secondly, in the time domain channel, the largest and the second-largest eigenvalues of graph Laplacian matrix are fused as the first feature to evaluate the graph density. In the frequency domain channel, by extracting non-zero elements from the diagonal of the degree matrix, the entropy value is calculated as the second feature to measure the dispersion of the vertex distribution of the graph. Then, the two features are used as detection statistics to determine whether they fall within the decision region given by the convex hull algorithm with target guidance. The detection results are obtained. Finally, experimental results using measured data demonstrate that the proposed detector can achieve robust and efficient detection performance in complex detection environments.

In recent years, millimeter-wave radars have become more and more widely used in the field of smart transportation due to their small size, light weight, fast processing speed and high detection accuracy. Millimeter-wave radars have different beam systems and their principles are different. For beamforming (BF) system millimeter-wave radars, this paper first introduces the principle of beamforming and scanning, and then designs a suitable microstrip array antenna. The radar hardware RF board is developed by cascading four chips. On this basis, the process of obtaining data for radar target detection is given. Combined with the target data and point cloud features under the BF radar, a target point cloud clustering method using the regional growth idea is proposed. Finally, the feasibility and effectiveness of this method are verified through field experiments. By comparing with the traditional DBSCAN algorithm, this method can effectively solve the problem that two targets are easily clustered into one target when the distance between two targets is close. At the same time, it can effectively reduce the processing time when the number of point clouds is large. This method provides strong support for the application of target detection, tracking and recognition of the BF radar platform, and has broad application prospects and practical value.

To address the limitations of traditional statistical models in simulating the time-frequency characteristics of sea clutter, a sea clutter data generation method based on an enhanced generative adversarial network (GAN) is proposed in this paper. The complex sea clutter is decomposed into amplitude and time-frequency components, which are separately fed into a variational autoencoder-Wasserstein generative adversarial network (VAE-WGAN) for training. The outputs are then integrated to synthesize complex signals with both amplitude and phase characteristics. To enhance the model performance, a gradient penalty mechanism is introduced to constrain the Lipschitz continuity of the discriminator, effectively mitigating the mode collapse. A self-attention module is incorporated to strengthen the model’s ability to capture localized strong scattering features, such as sea spikes, significantly improving the spatiotemporal correlation of generated signals. Experiments cover sea states 2～5, with three datasets of dimensions[64,64], [128,128], and[256,256]constructed for each sea state. Twelve cross-validation trials demonstrate that the synthetic data exhibit high consistency with measured data in amplitude distribution, normalized spectrum, temporal correlation, and time-frequency characteristics. These results validate the model’s generalization capability across varying sea states and multiscale temporal scenarios.