YCNNet: Road Target Recognition Method by Fusion of LiDAR and Thermal Infrared Camera

In unmanned systems, LiDAR and image sensors are extensively used for perception, with target recognition particularly dependent on image sensors. However, general image sensors struggle to reliably recognize targets in complex nighttime road scenes. To address this challenge, this paper proposes a target detection network(YCNNet) based on thermal infrared camera and LiDAR fusion, which utilizes improved YOLOv11-CG and CenterPoint-SEPB to detect thermal infrared images and point clouds respectively, and employs NMS matching to achieve decision-level fusion of these multimodal sensors, effectively improving the robustness of night road target recognition. To validate its performance, we created a novel joint dataset, YCNdataset, comprising paired point clouds and thermal infrared images. After ablation and comparative experiments, the results show that the proposed method can reliably identify road targets under night conditions, and achieves good recognition results, with an mAP of 82.34%.