基于声学频谱-时域信息融合的噪声环境中应急车辆检测

打开文本图片集

中图分类号：TP391.4 文献标识码：A DOI：10.3969/j.issn.1674-8484.2025.04.003

Abstract：An in-vehicle detection method was proposed based on the fusion of spectral and temporal featurestodetecttheexternal emergency vehicle sirens during high-speed driving.The input audio signal was transformedusing the fastFourier transform，andits log-Mel spectrogram wascomputed toextractspectral features.Aconvolutional neural network was used to model theraw waveform inthe time domain，yielding temporal features.A coordinate attntion mechanism was used to fuse and enhance thespectral and the temporal representations.The fused features were subsequently fed intoaclasifier forfinal detection.The experiments were conducted on both publicand real-recorded datasets.The resultsshow that on the LSADEVSRN dataset，the proposed method achieves an AUC （area under the receiver operating characteristic curve） score of 98.92% ，with representing an improvement of 14.88% compared to using temporal features alone，and 2.52% compared to using spectral features alone.These results confirm the effectiveness of the fusion strategy， with a high robustness particularly under noisy conditions.

Key words：automotive safety; siren detection; emergency vehicles; sound eventdetection; feature fusion

随着计算机技术和传感器技术的快速发展，汽车智能化技术逐渐成为汽车领域的发展主流[。（剩余12184字）