一种声源定位方法、装置、电子设备及存储介质
By generating audio point sequences and calculating cross-power spectra on smart home devices, the problem of inaccurate speech recognition in noisy environments is solved, enabling precise location of sound sources and accurate execution of control commands.
CN115932731BActive Publication Date: 2026-07-17QINGDAO HI-IMAGE TECH CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- QINGDAO HI-IMAGE TECH CO LTD
- Filing Date
- 2022-12-06
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
In noisy smart home environments, inaccurate voice recognition makes it difficult to accurately locate sound sources around smart home devices.
Method used
By acquiring audio frames from multiple microphones on smart home devices, an audio point sequence is generated, a fast Fourier transform is performed, the cross-power spectrum is calculated, and the directional spectrum is determined by combining the microphone spacing and sound speed. The sound source is then located when the wake word is recognized.
Benefits of technology
It improves the accuracy of voice recognition for smart home devices in noisy environments, ensuring the precision of sound source localization and the effective execution of control commands.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN115932731B_ABST
Abstract
本申请公开一种声源定位方法、装置、电子设备及存储介质，属于智能家居技术领域，该方法包括：每当获取到智能家居设备上S个麦克风新采集的音频帧时，基于新采集的音频帧和S个麦克风历史采集的音频帧，生成S个音频点序列，其中，S个麦克风被划分成至少一个麦克风对，对S个音频点序列分别进行快速傅里叶变换，得到S个频谱，基于S个频谱，确定各麦克风对的互功率谱，然后，基于各麦克风对的互功率谱、各麦克风对的麦克风间距、预设频谱范围和声速，确定智能家居设备的方位谱，若从S个音频点序列中识别到唤醒词，则基于方位谱进行声源定位，从而提供了一种对智能家居设备周围的声源进行定位的方案。
Need to check novelty before this filing date? Find Prior Art