Analysis And Research Of Giant Panda Individual Identification System Based On Voiceprint

Giant panda is a national first-class protected animal.It is very precious.China has set up several giant panda habitats and research centers to strengthen the research and protection of giant pandas.At present,the domestication protection of giant pandas in the wild is at a critical stage.The Chengdu Giant Panda Breeding Research Base needs a non-contact individual identification method as an extension and supplement to other conventional identification methods to better identify the giant pandas individually.Management protection.Voiceprint feature extraction is a key part of the sound-based giant panda individual identification system,which is of great significance to system function realization and performance improvement.Through comprehensive comparison of various common methods for the classification of giant pandas,combined with the technical characteristics and wide application of voiceprint recognition,this thesis analyzes and discusses the individual recognition system based on the giant panda voiceprint.The main research contents are as follows:Firstly,this paper introduces the significance and value of individual identification of giant pandas,and the research status of voiceprint recognition technology at home and abroad,and introduced the basic theory and related technical analysis of giant panda individual identification system,and briefly summarized the overall framework and process of the system.Explain that the standard of system performance evaluation and the calculation method of recognition accuracy are studied.Secondly,introduces and analyzes the short average energy characteristics,Mel-Frequency Cepstral Coefficients and linear prediction coefficient LPC,and then further study the MFCC and its second-order difference and LPC(linear prediction coefficient)The characteristics are combined,and the specific optimization and improvement of the recognition effect are verified and analyzed through specific experiments.Thirdly,a method for studying the characteristics of giant panda's voiceprint based on Mel's composite cepstrum coefficient is proposed.Combining the characteristics and frequency distribution of giant pandas,the mel filter bank in MFCC is supplemented and optimized for the characteristics of mel filter in MFCC and its performance at low high frequency resolution.The Meer composite frequency cepstrum coefficient is proposed.The CMFCC(Composite mel frequency cepstrum coefficient)feature improves the high-frequency resolution of the system's extracted features on the giant panda's sound.Through specific experiments,the improvement of the system's recognition effect is verified,which is in line with theoretical expectations.Finally,designing and implementing the giant panda individual identification system.Firstly,the requirements analysis of the system is carried out in combination with the requirements background and requirements description.Then the overall structure and main function modules of the system are designed.Then the implementation of the individual identification system is carried out.The collection process and processing method of the giant panda sound data under this topic are introduced.And display in tabular form,then analyze and implement the system flow and main functions,and finally perform system testing and overall experimental analysis.