Predicting Watermelon Sweetness from Induced Vibrational Frequencies

Presentation Type

Poster

Department

Mathematics

Location

Walker Conference Center B

Description

The current project focused on gathering additional audio signals from thumping watermelons to improve the correlation between the sound produced and the predicted sugar concentration of the watermelon. The audio signals are converted to harmonic frequencies using the Fast Fourier Transform (FFT) and then combined with a few physical properties of the watermelon for processing. The resulting analysis yields coefficients for a linear regression model to predict the sweetness of a watermelon. Data was collected in MATLAB on a PC and on a developed prototype iPhone app, “Dr. Thump”.

Comments

This work is currently embargoed. It will be available May 2026.

This document is currently not available here.

Share

COinS
 
Apr 24th, 3:15 PM Apr 24th, 4:30 PM

Predicting Watermelon Sweetness from Induced Vibrational Frequencies

Walker Conference Center B

The current project focused on gathering additional audio signals from thumping watermelons to improve the correlation between the sound produced and the predicted sugar concentration of the watermelon. The audio signals are converted to harmonic frequencies using the Fast Fourier Transform (FFT) and then combined with a few physical properties of the watermelon for processing. The resulting analysis yields coefficients for a linear regression model to predict the sweetness of a watermelon. Data was collected in MATLAB on a PC and on a developed prototype iPhone app, “Dr. Thump”.