Shiyang Yan

Vehicle–infrastructure cooperative perception (VICP) extends the sensing capability of single-vehicle systems by integrating multi-source information from onboard and roadside sensors, thereby alleviating limitations in sensing range and field-of-view coverage. However, in complex urban environments, the robustness of such systems—particularly in terms of blind-spot coverage and feature representation—is severely affected by both static and dynamic occlusions, as well as distance-induced sparsity in point cloud data. To address these challenges, a 3D object detection framework incorporating point cloud feature enhancement and spatially adaptive fusion is proposed. First, to mitigate feature degradation under sparse and occluded conditions, a Redefined Squeeze-and-Excitation Network (R-SENet) attention module is integrated into the feature encoding stage. This module employs a dual-dimensional squeeze-and-excitation mechanism operating across pillars and intra-pillar points, enabling adaptive recalibration of critical geometric features. In addition, a Feature Pyramid Backbone Network (FPB-Net) is designed to improve target representation across varying distances through multi-scale feature extraction and cross-layer aggregation. Second, to address feature heterogeneity and spatial misalignment between heterogeneous sensing agents, a Spatial Adaptive Feature Fusion (SAFF) module is introduced. By explicitly encoding the origin of features and leveraging spatial attention mechanisms, the SAFF module enables dynamic weighting and complementary fusion between fine-grained vehicle-side features and globally informative roadside semantics. Extensive experiments conducted on the DAIR-V2X benchmark and a custom dataset demonstrate that the proposed approach outperforms several state-of-the-art methods. Specifically, Average Precision (AP) scores of 0.762 and 0.694 are achieved at an IoU threshold of 0.5, while AP scores of 0.617 and 0.563 are obtained at an IoU threshold of 0.7 on the two datasets, respectively. Furthermore, the proposed framework maintains real-time inference performance, highlighting its effectiveness and practical potential for real-world deployment.

Vehicle–infrastructure cooperative perception (VICP) overcomes the sensing limitations and field-of-view constraints of single-vehicle intelligence by integrating multi-source information from onboard and roadside sensors. However, in complex urban environments, system robustness—particularly regarding blind-spot coverage and feature representation—is severely compromised by occlusion (static and dynamic) and distance-induced point cloud sparsity. To address these challenges, this paper proposes a 3D object detection framework incorporating point cloud feature enhancement and spatial adaptive fusion. First, to mitigate feature degradation under sparse and occluded conditions, a Redefined-SENet (R-SENet) attention module is embedded into the feature encoding stage. This module employs a dual-dimensional squeeze-and-excitation mechanism—across pillars and intra-pillar points—to adaptively recalibrate key geometric features. Concurrently, a Feature Pyramid Backbone Network (FPB-Net) is constructed to enhance unified target modeling across varying distances via multi-scale extraction and cross-layer aggregation. Second, a Spatial Adaptive Feature Fusion (SAFF) module is introduced to resolve feature heterogeneity and spatial misalignment. By explicitly encoding feature origins and leveraging spatial attention, SAFF enables dynamic weighting and complementary fusion of fine-grained vehicle-side features and global roadside semantics. Experiments on the DAIR-V2X benchmark and a custom dataset demonstrate that the proposed method outperforms state-of-the-art approaches, achieving Average Precision (AP) scores of 0.762 and 0.694 at IoU 0.5, and 0.617 and 0.563 at IoU 0.7, respectively. Furthermore, the inference speed satisfies real-time requirements, validating the method’s effectiveness and potential for engineering deployment.