针对电动汽车充电过程中底盘热异常识别中预警滞后与误报率高的问题，文章提出温度、温升速率与空间梯度多特征融合的快速识别方法。该方法构建三维特征体系，建立分级、多条件耦合识别规则，实现对局部热点与突发升温的早期精准辨识。实际车位试验验证表明，所提方法可将平均预警提前时间缩短至22 s，将识别准确率提升至 96%，并凭借空间梯度特征将误报率显著降低至 3% 以下，具有明确的工程应用价值。

In response to the problems of delayed early warning and high false alarm rate in the identification of abnormal temperature rise of the charging chassis of electric vehicles, this paper proposes a rapid identification method that integrates multiple features of temperature, temperature rise rate and spatial gradient. This method constructs a threedimensional feature system and establishes a hierarchical and multi-condition coupled identification rule to achieve early and accurate identification of local hotspots and sudden temperature rises. The actual parking space experiment verification shows that the proposed method can shorten the average warning time to 22 seconds, increase the identification accuracy to 96%, and significantly reduce the false alarm rate to less than 3% by relying on the spatial gradient feature, which has clear engineering application value.