Multi-sound-source positioning method based on formation of real value weight beam in spherical harmonic domain

Abstract

The invention discloses a multi-sound-source positioning method based on formation of a real value weight beam in a spherical harmonic domain. The multi-sound-source positioning method comprises the steps that a sound field model that unit amplitude plane waves enter a spherical microphone array is firstly established; then a spherical harmonic domain noise source signal model is constructed; spherical harmonic domain real value beam forming is carried out on noise source signals received by the spherical microphone array to obtain the beam in the spherical harmonic domain; an MVDR beam when the gain of the spherical microphone array is the maximum is constructed, and the optimal real value weight at the time is obtained; finally, the peak point of the square of the real value weight in the spherical harmonic domain is calculated, and the dimensional orientation maximum estimated value point of sound sources is extracted. According to the multi-sound-source positioning method based on formation of the real value weight beam in the spherical harmonic domain, the optimum real value weight in the spherical harmonic domain is obtained by utilizing the formation of the MVDR beam, the defects of large calculation amount and high calculation complexity of a traditional method are overcome, the calculation amount is obviously lowered, all-dimensional estimation of a free space is met, and the sound field is sampled more sufficiently.

Description

technical field [0001] The invention relates to a multi-sound source localization method based on real-valued weight beamforming in the spherical harmonic domain, which can be widely used in fields such as sound source localization. Specifically, based on the minimum variance distortion-free response (MVDR) beamforming principle, combined with the conversion between the spatial domain and the spherical harmonic domain, the spatial domain sound field signal collected by the spherical microphone array is transformed into the spherical harmonic domain, and the real-valued weighted beam of the spherical harmonic domain is performed. Forming, constructing the weight design of real-valued MVDR beamforming, using the Lagrangian multiplier method to obtain the optimal real-valued weight, and finally obtaining the spatial orientation information of the sound source by searching the peak value of the square of the real-valued weight. Compared with the traditional sound source localizati...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to address the deficiencies in the prior art, and propose a multi-sound source localization method based on real-valued weighted beamforming in the spherical harmonic domain. It reduces the amount of calculation, can satisfy the omni-directional estimation of free space, and conducts more sufficient sampling of the sound field

Images