To improve the detection accuracy of vehicles and pedestrians in traffic scenes using object detection algorithms, this paper presents modifications, compression, and deployment of the single-stage typical algorithm YOLOv7-tiny. In the model improvement section: firstly, to address the problem of small object missed detection, shallower feature layer information is incorporated into the original feature fusion branch, forming a four-scale detection head; secondly, a Multi-Stage Feature Fusion (MSFF) module is proposed to fully integrate shallow, middle, and deep feature information to extract more comprehensive small object information. In the model compression section: the Layer-Adaptive Magnitude-based Pruning (LAMP) algorithm and the Torch-Pruning library are combined, setting different pruning rates for the improved model. In the model deployment section: the V7-tiny-P2-MSFF model, pruned by 45% using LAMP, is deployed on the embedded platform NVIDIA Jetson AGX Xavier. Experimental results show that the improved and pruned model achieves a 12.3% increase in mAP@0.5 compared to the original model, with parameter volume, computation volume, and model size reduced by 76.74%, 7.57%, and 70.94%, respectively. Moreover, the inference speed of a single image for the pruned and quantized model deployed on Xavier is 9.5 ms.
Ensuring safety on busy airport aprons remains challenging, particularly in preventing aircraft wingtip collisions. In this study, first, a simplified coordinate mapping method converts pixel detections into accurate spatial coordinates, improving aircraft position and velocity estimates. Next, an innovative dynamic warning area with a classification mechanism is introduced to enable faster responses from airport staff. Finally, this study proposes LSCD-Pose, a real-time detection network enhanced by lightweight shared modules, significantly reducing model size and computational load without sacrificing accuracy. Experiments on real airport datasets representing various apron scenarios demonstrate frame rates up to 461.7 FPS and a 90.5% reduction in model size compared with the baseline. Visualizations confirm the solution’s versatility and efficiency in effectively mitigating wingtip collisions and enhancing apron safety.
Reliable 3D scene understanding is a fundamental requirement for intelligent machines in autonomous transportation systems, as on-board perception must remain accurate and stable across diverse environments and sensing conditions. However, LiDAR point clouds acquired in real traffic scenes are often sparse and irregular, and they exhibit heterogeneous sampling patterns that hinder consistent and fine-grained semantic interpretation. To address these challenges, this paper proposes LVCA-Net, a lightweight voxel–coordinate attention framework designed for efficient LiDAR-based 3D semantic segmentation in autonomous driving scenarios. The architecture integrates (i) an anisotropic depthwise residual module for direction-aware geometric feature extraction, (ii) a hierarchical LiteDown–LiteUp pathway for multi-scale feature fusion, and (iii) a Coordinate-Guided Sparse Semantic Module that enhances spatial consistency in a cylindrical voxel space while maintaining computational sparsity. Experiments on the SemanticKITTI and nuScenes benchmarks demonstrate that LVCA-Net achieves 67.17% mean Intersection over Union (mIoU) and 91.79% overall accuracy on SemanticKITTI, as well as 77.1% mIoU on nuScenes, while maintaining real-time inference efficiency. These results indicate that LVCA-Net delivers scalable and robust 3D scene understanding with high semantic precision for LiDAR-only perception, making it well suited for deployment in autonomous vehicles and other safety-critical intelligent systems.