Font Size: a A A

Research And Implementation Of Voiceprint Recognition Algorithm In Military Calls

Posted on:2015-10-07Degree:MasterType:Thesis
Country:ChinaCandidate:L LiuFull Text:PDF
GTID:2272330482954518Subject:Circuits and Systems
Abstract/Summary:PDF Full Text Request
In recent years, with the development of pattern recognition, image processing and information sensing, biological recognition technology has a broad application prospect and great practical value. In military actions, application of voiceprint recognition technology, through confirming the identity of people who give out orders, can effectively avoid the enemy issue false orders and disturb our military actions by using our phone channels to camouflage my commander.This thesis is aimed to research how to confirm the identity of the other side based on voiceprint recognition algorithm, and then establish a voiceprint recognition system. This thesis mainly completes the following work:(1) Research the feature extraction method of voice signal, it focus on the research of the Mixing characteristic coefficient, including the Linear Prediction Coefficient(LPC), Linear Prediction Cestrum Coefficient(LPCC), Mel Frequency Cepstral Coefficient(MFCC), and the differential MEL cepstrum coefficient(△MFCC). In the meanwhile, this thesis conducts some relevant experiments, comparing their ability to identify the voice. It focuses on the research of the Mixing characteristic coefficient, the portfolio analysis of the LPCC, MFCC and AMFCC. This paper puts forward the coefficient of LPC-MFCC(including First order difference) and the coefficient of LPCC+MFCC(including △MFCC), and then conducts the experiments about their recognition capability based on the computer and random background telephone channel voice. At last this subject chooses the LPC-MFCC as the characteristic coefficient of the voiceprint recognition system.(2) Research the method of voice denoising based on the wavelet transform. It does some experiments about the denoising ability of different types of wavelet. Finally, it does some experiments about the Dbl wavelet algorithm’s denoising ability under different signal to noise ratio.(3) Under the military application background, it writes out the voiceprint recognition system of military communication program based on MATLAB software. It also designs the simulation interface, using the graphical user interface in MATLAB.(4) Using ICETEK-DM6437-B as the hardware development platform, this thesis achieves the basic function of the system and tests the performance of the system.The experiment shows that the recognition rate of the system designed in this thesis is beyond 80% and it basically meets the design expectations. Because the model is not complicated and the recognition speed is fast, it has a certain practical value.
Keywords/Search Tags:Voiceprint recognition, Noise enviroment, Feature extraction, Telephone channel
PDF Full Text Request
Related items