Phase demodulation is the core of fringe projection profilometry systems. However, current U-Net-based phase demodulation approaches demonstrate deficiencies in global context propagation, adversely affecting wrapped phase extraction precision. To this end, this paper proposes a deep-learning-based model for single-shot wrapped phase extraction, named the full-scale connection and attention enhancement network (SEC-UNet3+). The network mitigates the limitations of the traditional U-Net architecture by introducing cross-layer full-scale connection and a feature integration module in the decoder, enabling efficient interaction between shallow detail features and deep semantic features. Unlike the skip connection strategy within the same-layer in U-Net, cross-layer full-scale connection can enhance the feature utilization. Additionally, a skip connection is embedded between the feature mapping layer and the output transformation layer in the squeeze and excitation module, preventing information loss during the feature calibration process. Compared to the U-Net model, the proposed method achieves an approximately 5% to 15% reduction in both the mean squared error and mean absolute error for phase extraction. The experimental results confirm that SEC-UNet3+ outperforms traditional Fourier transform and mainstream U-Net-based approaches in phase demodulation accuracy, proving particularly effective for single-shot wrapped phase retrieval in dynamic scenarios.

We present a reconfigured lateral-shearing short-coherence digital holographic microscopy (DHM) using dual-plate beam splitters (BSs) in this paper. An adjustable air gap is formed by the back and front surfaces of the first and second plate BSs. By changing the distance of this air gap properly, interference fringes can easily be formed in a short-coherence light source. Moreover, the orientation and carrier frequency of the fringes can be modulated by altering the wedge angel of the air gap. The validity and stability of the proposed system are demonstrated through the optical experimental results. Owing to its compact design and flexible adjustment capabilities, this system shows great promise for the development of highly stable optical measurement and imaging instruments in the practical applications.

Small ship detection in synthetic aperture radar (SAR) imagery is challenged by speckle noise, strong scale variation, and complex sea clutter, especially under practical efficiency requirements for SAR-based maritime surveillance. We propose SAR-LiteDETR, a lightweight transformer-based detector tailored for maritime SAR scenes. SAR-LiteDETR includes three lightweight plug-and-play modules: PConv-Blocks, an Attention-Guided Feature Pyramid Network (AG-FPN), and a Channel-wise Multi-Branch Fusion Block (CMBF). These three modules improve backbone efficiency and small-target sensitivity, stabilize multi-scale feature alignment under clutter, and adaptively fuse cross-level features to suppress sea clutter and alleviate feature dilution, respectively, while keeping computational cost low. Experiments on the SSDD, HRSID, and LS-SSDD datasets show that SAR-LiteDETR consistently outperforms state-of-the-art and baseline detectors. In particular, it achieves $AP_{S}$ and precision of 69.6%/97.0% on SSDD and 54.2%/93.5% on HRSID, respectively, while requiring only 42.1 GFLOPs. The results indicate that SAR-LiteDETR provides a favorable balance between detection accuracy and efficiency for near real-time SAR-based maritime surveillance.