Extraction of feature from the acoustic activity of RPW using MFCC

Abstract

RPW is a destructive pest which infests palms of age ranging from 5–20 yrs. Control measure if adopted timely on infestation can lead to saving of palm, else ends up in outright death of palm. RPW usually infests palm in its younger stage while adult causes damage through feeding. The burrowing of the larva into palm heart causes more mortality. Major symptoms such as crown loss or leaf wilt are usually visible only after palm has become infected. By the time these symptoms are observed the damage is usually high to kill the tree. Hence in order to detect the presence of weevil even at the early stage the sound of larvae burrowing and chewing are recorded using digital voice recorders. The recorded sound file is further analyzed. In sound processing, Mel frequency cepstrum is a representation of short term power spectrum of a sound based on a linear cosine transform of a log power spectrum on a nonlinear Mel scale of frequency. MFCC's have been used for feature extraction. In frequency domain, the most often used feature extraction is Mel frequency cepstral coding using Mel scale.