基于Mel倒谱系数和矢量量化的昆虫声音自动鉴别

为了给生产单位害虫管理的普通技术人员提供简便易操作的昆虫种类鉴别方法, 本研究把人类语音识别领域的先进技术应用于昆虫识别, 提出了一种新颖的昆虫声音自动鉴别方法, 用声音参数化技术为昆虫声纹识别设计了一种简单易行的方案。声音信号经过预处理、分段得到一系列的声音样本, 从声音样本提取Mel倒谱系数（MFCC）, 并用Linde-Buzo-Gray（LBG）算法对提取的MFCC进行矢量量化(VQ), 所得码字作为声音样本的特征模型。特征参数之间的匹配用搜索最近邻的方法实现。本文方法在包含70种昆虫声音的库中进行了试验, 取得了超过96%的识别率和理想的时间性能。试验结果证明了该方法的有效性。

Automatic acoustical identification of insects based on MFCC and VQ

This study aims to provide general technicians who manage pests in production with a convenient way to recognize insects. A simple and viable scheme to identify insect voiceprints automatically is introduced using a sound parameterization technique that dominates speaker recognition technology. The acoustic signal was preprocessed and segmented into a series of sound samples. Mel-frequency cepstrum coefficient (MFCC) was extracted from the sound sample, and a feature model was trained using Linde-Buzo-Gray algorithm to generate vector quantization (VQ) codebook from above MFCC. The matching for a test sample was completed by finding the nearest neighbour in all the VQ codebooks. The method was tested in a database with acoustic samples of 70 different insect sounds. The recognition rate above 96% was obtained, and an ideal time performance was also achieved. The test results proved the efficiency of the proposed method.