Levinson Durbin Algorithm-An Approach for feature extraction of Red Palm Weevil

Vol. 3 No. 1

Year: 2012

Issue: Sep-Nov

Title: Levinson Durbin Algorithm-An Approach for feature extraction of Red Palm Weevil  

Author Name: Betty Martin, Vimala Juliet 

Synopsis: 

Digital audio information has become an important application of computer in the field of audio processing and pattern recognition. In pattern recognition, features are extracted from raw audio data. In this paper, an effective algorithm is proposed to extract features of the acoustic activity  of red palm weevil recorded in coconut grooves using Linear predictive coding.  The linear predictive coefficients characterize the audio content. Red palm weevil is an economic pest, the infestation of which, at an early stage can prevent total damage of the palm. The detection and infestation of this insect is possible only  by capturing the sound produced by the insect, using a reliable digital voice recorder. The sound acquired is further processed to confirm whether it belongs to the insect or not. This can be accomplished by a technique called Linear Predictive Coding. The coefficients extracted can further be matched by training data. The classification depends on the ability to accurately categorize each feature vector set corresponding to its own kind. This application is often used for sound identification or sound verification. In this article, a summary of work done to extract feature of the weevil which was later used to determine its presence on palms, has been done.