基于深度学习的列车运行环境感知关键算法研究综述

陈智超,杨杰,李凡,冯志成

Review on deep learning-based key algorithm for train running environment perception

Zhichao CHEN,Jie YANG,Fan LI,Zhicheng FENG

表 4 OSDaR23数据集中BEVFusion的多模态检测实验结果

Tab.4 Experimental results of multimodal detection of BEVFusion in OSDaR23 dataset

模型 模态 mAP/% 行人 接触网杆 信号杆 道路车辆 止冲挡 BEVFusion C 28.76 0.01 4.66 20.06 16.53 +TF C 32.29 0.29 8.30 27.83 25.43 BEVFusion L 79.99 90.33 75.63 59.57 82.26 +TF L 85.56 90.99 81.32 65.85 85.20 +TF+TA-GTP L 86.94 90.72 80.10 67.84 85.46 BEVFusion L+C 86.79 88.85 73.36 64.87 83.83 +TF L+C 87.25 91.57 69.98 66.40 83.46