Railway track detection, which is crucial for train operational safety, faces numerous challenges such as the curved track, obstacle occlusion, and vibrations during the train’s operation. Most existing methods for railway track detection use a camera or LiDAR. However, the vision-based approach lacks essential 3D environmental information about the train, while the LiDAR-based approach tends to detect tracks of insufficient length due to the inherent limitations of LiDAR. In this study, we propose a real-time method for railway track detection and 3D fitting based on camera and LiDAR fusion sensing. Semantic segmentation of the railway track in the image is performed, followed by inverse projection to obtain 3D information of the distant railway track. Then, 3D fitting is applied to the inverse projection of the railway track for track vectorization and LiDAR railway track point segmentation. The extrinsic parameters necessary for inverse projection are continuously optimized to ensure robustness against variations in extrinsic parameters during the train’s operation. Experimental results show that the proposed method achieves desirable accuracy for railway track detection and 3D fitting with acceptable computational efficiency, and outperforms existing approaches based on LiDAR, camera, and camera–LiDAR fusion. To the best of our knowledge, our approach represents the first successful attempt to fuse camera and LiDAR data for real-time railway track detection and 3D fitting.

Detecting obstacles in the rail track area is crucial for ensuring the safe operation of trains. However, this task presents numerous challenges, including the diverse nature of intrusions, and the complexity of the driving environment. This paper presents a multimodal fusion rail-obstacle detection approach by key points processing and rail track topology reconstruction. The core idea is to leverage the rich semantic information provided by images to design algorithms for reconstructing the topological structure of railway tracks. Additionally, it combines the effective geometric information provided by LiDAR to accurately locate the railway tracks in space and to filter out intrusions within the track area. Experimental results demonstrate that our method outperforms other approaches with a longer effective working distance and superior accuracy. Furthermore, our post-processing method exhibits robustness even under extreme weather conditions.

To address the challenges of environmental impact and distribution efficiency in fresh food logistics, a segmented transshipment model involving the coordinated operation of gasoline and electric vehicles is proposed. The model minimizes total distribution costs by considering transportation, refrigeration, product damage, carbon emissions, and penalties for time window violations. The k-means++ clustering algorithm is used to determine transshipment points, while an improved adaptive multi-objective ant colony optimization algorithm (IAMACO) is employed to optimize the delivery routes for the heterogeneous fleet. The case study results show that compared to the traditional model, the segmented transshipment mode reduces the total distribution costs, carbon emissions, and time window penalty costs by 22.13%, 28.32%, and 41.08%, respectively, providing a viable solution for fresh food logistics companies to achieve sustainable and efficient growth.

Estimating 360${}^{\circ }$ depth from multiple cameras has been a challenging problem. However, existing methods often adopt a fixed-step spherical sweeping approach with densely sampled spheres and use numerous 3D convolutions in networks, which limits the speed of algorithms in practice. Additionally, obtaining high-precision depth maps of real scenes poses a challenge for the existing algorithms. In this paper, we design a cascade architecture using a dynamic spherical sweeping method that progressively refines the depth estimation from coarse to fine over multiple stages. The proposed method adaptively adjusts sweeping intervals and ranges based on the predicted depth and the uncertainty from the previous stage, resulting in a more efficient cost aggregation performance. The experimental results demonstrated that our method achieved state-of-the-art accuracy with reduced GPU memory usage and time consumption compared to the other methods. Furthermore, we illustrate that our method achieved satisfactory performance on real-world data, despite being trained on synthetic data, indicating its generalization potential and practical applicability.

Forest fire monitoring in remote sensing imagery has long relied on traditional perception models that primarily focus on detection or segmentation. However, such approaches fall short in understanding complex fire dynamics, including contextual reasoning, fire evolution description, and cross-modal interpretation. With the rise of multi-modal large language models (MLLMs), it becomes possible to move beyond low-level perception toward holistic scene understanding that jointly reasons about semantics, spatial distribution, and descriptive language. To address this gap, we introduce FireMM-IR, a multi-modal large language model tailored for pixel-level scene understanding in remote-sensing forest-fire imagery. FireMM-IR incorporates an infrared-enhanced classification module that fuses infrared and visual modalities, enabling the model to capture fire intensity and hidden ignition areas under dense smoke. Furthermore, we design a mask-generation module guided by language-conditioned segmentation tokens to produce accurate instance masks from natural-language queries. To effectively learn multi-scale fire features, a class-aware memory mechanism is introduced to maintain contextual consistency across diverse fire scenes. We also construct FireMM-Instruct, a unified corpus of 83,000 geometrically aligned RGB–IR pairs with instruction-aligned descriptions, bounding boxes, and pixel-level annotations. Extensive experiments show that FireMM-IR achieves superior performance on pixel-level segmentation and strong results on instruction-driven captioning and reasoning, while maintaining competitive performance on image-level benchmarks. These results indicate that infrared–optical fusion and instruction-aligned learning are key to physically grounded understanding of wildfire scenes.

The precise calibration of a LiDAR-camera system is a crucial prerequisite for multimodal 3D information fusion in perception systems. The accuracy and robustness of existing traditional offline calibration methods are inferior to methods based on deep learning. Meanwhile, most parameter regression-based online calibration methods directly project LiDAR data onto a specific plane, leading to information loss and perceptual limitations. A novel network, DPCalib, a dual perspective view network that mitigates the aforementioned issue, is proposed in this paper. This paper proposes a novel neural network architecture to achieve the fusion and reuse of input information. We design a feature encoder that effectively extracts features from two orthogonal views using attention mechanisms. Furthermore, we propose an effective decoder that aggregates features from two views, thereby obtaining accurate extrinsic parameter estimation outputs. The experimental results demonstrate that our approach outperforms existing SOTA methods, and the ablation experiments validate the rationality and effectiveness of our work.