Speech endpoint detecting method and device

Abstract

The invention belongs to the field of video monitoring, and provides a speech endpoint detecting method and device. The method comprises the following steps: sampling data of an input speech signal, and preprocessing the sampled speech signal; adding a Hamming window to the preprocessed speech signal for framing and recording as Rn (n is more than 0 and less than or equal to N), wherein N is the total number of frames; calculating a frequency spectrum information entropy of the n-th speech signal; and determining the frame as a speech frame if the frequency spectrum information entropy of the n-th speech signal is more than a set threshold value, and otherwise, determining the frame as a non-speech frame. The method applies the frequency spectrum entropy as a characteristic for distinguishing a speech frame from a non-speech frame, can effectively distinguish speech frames from non-speech frames, and has a good detection effect for low signal to noise ratio environments, so the defects that the traditional frequency spectrum entropy-based algorithm only considers the frequency spectrum information of the current frame, the noise frequency spectrum information entropy greatly fluctuates in a non-stationary noise environment, and the difficulty of threshold value selection is increased can be overcome.

Description

technical field [0001] The invention belongs to the field of video monitoring, in particular to a voice endpoint detection method and device. Background technique [0002] At present, in real-time video monitoring, a pickup is used to pick up abnormal sounds in the monitoring scene, thereby adjusting the optical axis of the camera to point to the abnormal sound, and real-time monitoring of abnormal events can be realized. Since the omnidirectional pickup can pick up sounds in all directions, it can effectively solve the disadvantages of traditional video surveillance that abnormal events cannot be quickly captured due to abnormal events occurring in the blind area of ​​the surveillance camera's field of view. In video surveillance, using a pickup to pick up abnormal sounds in the surveillance scene, the most critical first step is the voice endpoint detection technology. [0003] Traditional endpoint detection methods, such as short-term energy, zero-crossing rate algorithm...

AI Technical Summary

Problems solved by technology

In a low signal-to-noise ratio or non-stationary environment, the short-term energy of speech is easily confused with noise, and the zero-crossing rate is easy to distinguish unvoiced and noise, but it is difficult to distinguish voiced and noise. The short-term zero energy product method can improve the endpoint to a certain extent. The robustness of detection, but the anti-noise performance of short-term zero-energy product characteristic parameters is not as good as that of information entropy. To some extent, spectral entropy has certain robustness to noise, but when the signal-to-noise ratio decreases, although the shape of It remains unchanged, but the spectral entropy decreases, and the traditional method based on spectral entropy only considers the spectral information of the current frame. In the non-stationary noise environment, the noise spectral information entropy fluctuates in a large range, which brings difficulties to the threshold selection.

Images