Railway safety inspections, a critical component of modern transportation systems, face significant challenges from adverse weather conditions, like fog and rain, which degrade image quality and compromise inspection accuracy. To address this limitation, we propose a novel deep learning-based image dehazing algorithm optimized for outdoor railway environments. Our method integrates adaptive high-pass filtering and bilateral grid processing during the feature extraction phase to enhance detail preservation while maintaining computational efficiency. The framework uniquely combines RGB color channels with atmospheric brightness channels to disentangle environmental interference from critical structural information, ensuring balanced restoration across all spectral components. A dual-attention mechanism (channel and spatial attention modules) is incorporated during feature fusion to dynamically prioritize haze-relevant regions and suppress weather-induced artifacts. Comprehensive evaluations demonstrate the algorithm’s superior performance: On the SOTS-Outdoor benchmark, it achieves state-of-the-art PSNR (35.27) and SSIM (0.9869) scores. When tested on a specialized railway inspection dataset containing 12,840 fog-affected track images, the method attains a PSNR of 30.41 and SSIM of 0.9511, with the SSIM being marginally lower (0.0017) than DeHamer while outperforming other comparative methods in perceptual clarity. Quantitative and qualitative analyses confirm that our approach effectively restores critical infrastructure details obscured by atmospheric particles, improving defect detection accuracy by 18.6 percent compared to non-processed images in simulated inspection scenarios. This work establishes a robust solution for weather-resilient railway monitoring systems, demonstrating practical value for automated transportation safety applications.

This paper addresses the two-dimensional direction-of-arrival(2D-DOA) estimation problem using an Electromagnetic Vector Sensor (EMVS) array, where an arbitrary array geometry and the presence of unknown nonuniform noise are considered. A novel algorithm combining tensor completion and Parallel Factor Decomposition (PARAFAC) is proposed to improve the accuracy of estimation. Firstly, a covariance tensor is constructed. The impact of non-uniform noise is eliminated via removing the noise-affected elements of the covariance tensor. The noise-free covariance matrix is then recovered through tensor completion, and the target angles are extracted using PARAFAC algorithm. Additionally, Cramer-Rao Bounds (CRB) for joint 2D-DOA and polarization parameter estimation are derived, providing a theoretical basis for array design. Simulation results demonstrate that the proposed algorithm is robust to nonuniform noise and achieves high-resolution 2D-DOA estimation, outperforming traditional methods in terms of accuracy. Comparative evaluations confirm the superiority of the proposed geometry and denoising approach.

The rapid advancement of unmanned vending systems has created notable challenges in accurately detecting minor product damages, especially in complex situations involving occlusion. Current approaches often struggle to find a suitable balance between achieving high detection accuracy and meeting the necessities of real-time processing. To tackle these issues, this paper introduces PFL-YOLO, an efficient object detection model based on YOLOv11n, specifically designed for recognizing product damage. Within the backbone network, we develop a hybrid feature modeling module known as CSP-PT. By utilizing the inherent cross-stage partial feature reuse mechanism of CSP, this module reduces computational redundancy while incorporating a transformer branch to enhance semantic relationships across spatial dimensions. This structural configuration ensures that the output features maintain compactness while exhibiting enhanced representational strength. Furthermore, we present the C3K2-FM module, which features a dynamic convolution mechanism based on frequency-domain decomposition. This strategy effectively addresses the challenge of spectral homogeneity commonly observed in convolutional kernel weights used in traditional dynamic convolution. Consequently, the semantic connections among multi-scale features are strengthened, significantly improving the synergistic representation of both shallow fine details and deep semantic information. In terms of the detection head, we apply structural-level lightweight optimization and effective regression modeling techniques to diminish computational complexity and inference delay without sacrificing accuracy, thereby enhancing real-time functionalities and simplifying deployment. Experimental findings reveal that PFL-YOLO surpasses the baseline model, achieving enhancements of 2.68% in mAP@0.5 and 6.8% in mAP@0.5:0.95, while also decreasing the computational cost by 23.8%. These results confirm the effectiveness and significant practical value of the proposed approach.