Authors : Kofi Ampomah; Owusu A. Antwi; A. F. Akwetey

Volume/Issue : Volume 9 - 2024, Issue 1 - January

Google Scholar : http://tinyurl.com/399xw4rr

Scribd : http://tinyurl.com/36k8p4dn

DOI : https://doi.org/10.5281/zenodo.10608503

Abstract : The level of coconut maturity may be measured not only by viewing the color of the shell, but also by using an acoustic recognition technique from pounding on the coconut shell. This knocking sound differentiates between Juicy, Semi - Juicy, Fleshy, Very Fleshy and mature coconut. Those with substantial knowledge and sound sensitivity to coconut knocking typically execute identifying the sound distinctive of banging on coconut. The design of a coconut maturity prediction system with acoustic frequency detection is devised to replace skilled workers. The coconut sound signal is captured using a stethoscope linked to a Max4466 Electret Microphone Amplifier. The signal is processed using an Arduino Due microcontroller. The signal processing procedure includes the following steps: converting an analog signal to a digital signal, screening the signal, and locating the signal. The signal processing procedure includes the following steps: converting an analog signal to a digital signal, screening the signal, and determining the average value of the sound signal frequency spectrum. The signal screening employs a bandpass digital filter of the type IIR (Infinite Impulse Response) with an elliptic order of 6-7. This filter is used to ensure that the signal being processed is not noise but the signal of a banging sound on a coconut. The Nave Bayes Machine Learning Classification Algorithm is used to calculate the average value of the sound signal frequency spectrum. The Naive Bayes classification approach is used for maturity prediction. The input is three average values of knocking sound frequency and coconut size, and the output is coconut maturity categorization shown on an LED screen.

Keywords : Acoustic, Distinctive, Spectrum ,Machine Learning, Elliptic Order