Speech reconstruction method based on millimeter wave radar phase ranging

Abstract

The invention belongs to the field of voice reconstruction, and in particular relates to a voice reconstruction method based on millimeter-wave radar phase ranging, comprising the following steps: S1 using a discrete Fourier transform method to obtain a high-resolution one-dimensional range image sequence of a sound source target; S2 One-dimensional range image sequence peak point detection based on unit average constant false alarm detector; S3 Peak point phase information extraction and sound source audio information reconstruction. The beneficial effects obtained by the present invention are: high-quality speech information reconstruction can be realized through the present invention, and in a multi-sound source environment, the radar using the millimeter wave frequency band can actively detect the sound source target and accurately obtain the one-dimensional distance of the sound source target Image, and detect the phase information of the peak point, and then obtain the vibration information of the sound source target, and reconstruct the audio information of the sound source target, effectively eliminating the interference of other sound sources.

Description

technical field [0001] The invention belongs to the field of voice reconstruction, in particular to a voice reconstruction method based on millimeter-wave radar phase ranging. Background technique [0002] In a real acoustic environment, the target voice is easily disturbed by human voices and background sounds. No matter in daily life or in the military field, there is an urgent need for a method of speech restoration under noise background. Speech reconstruction is to capture the speech content of interest through various sensors to restore the original speech signal to the greatest extent. [0003] At present, the commonly used sensors for picking up sound are various microphone arrays. The principle of the microphone is that the vibration of the sound is transmitted to the diaphragm of the microphone, and the magnet inside is pushed to form a changing current to realize the transformation from physical signal to electrical signal. However, the microphone array cannot s...

AI Technical Summary

Problems solved by technology

[0005] The problem to be solved by the present invention is that the traditional microphone array can only passively receive the superposition information of multiple sound sources at the array elements, and cannot separate the independent sound source information

[0006] The main idea of ​​the present invention is to propose a voice reconstruction method based on millimeter-wave radar phase ranging for the problem that the microphone array cannot separate multiple sound source information

Images