Voice recognition method and device, storage medium and air conditioner

Abstract

The invention discloses a voice recognition method and device, a storage medium and an air conditioner. The method comprises the following steps: acquiring a first voice data; adjusting an acquisitionstate of a second voice data according to the first voice data, and acquiring the second voice data based on the adjusted acquisition state; performing far field voice recognition on the acquired second voice data by utilizing a preset far field voice recognition model so as to obtain semantic information corresponding to the second voice data. According to the scheme in the invention, the problem that the far field voice recognition effect is poor by utilizing a deep learning method or microphone array method for eliminating reverberation and noise in the far field voice data can be solved,and the effect of improving the far field voice recognition effect is achieved.

Description

technical field

[0001] The invention belongs to the technical field of voice control, and in particular relates to a voice recognition method, device, storage medium and air conditioner, in particular to a microwave radar-based far-field voice recognition method, device, storage medium and air conditioner. Background technique

[0002] Speech recognition technology is currently a relatively mature way of human-computer interaction. From the initial near-field speech recognition of handheld devices, such as Sirfi speech recognition and various voice assistants, to now, the application of speech recognition has been completed to intelligent hardware, The extension of home appliances, robots and other fields. However, the new human-computer interaction method has more stringent requirements on hardware, software, algorithms, etc., especially the far-field speech recognition technology faces huge challenges.

[0003] With the continuous development of smart home systems, smart ...

Images