Presentation Type
Thesis
Department
Computer Science
Location
Walker Conference Center B
Description
The main goal of this project was to determine the best sound frequencies produced from thumping a watermelon that predict sweetness. Building on previous research, a program using MATLAB would separate the thump signal into its corresponding harmonic frequencies using the Fast Fourier Transform (FFT). These frequencies, along with the watermelon's weight and other information from the FFT spectrum, are used to determine a correlation between the recorded data and a watermelon’s sweetness level, measured by a refractometer in Brix. After testing numerous watermelons, a linear regression model was developed to predict a watermelon’s potential sweetness, with promising results. This process has been programmed into an iPhone app so that testing can be done anywhere within seconds. The hope was to make the process available to the public so that they could regularly select better watermelons before leaving the store and cutting the watermelon, which we have achieved with a new iPhone app. A future goal is to continue to determine if the same regression model will work for similar melons, like cantaloupe and honeydew, as well as bring the app to the Android platform.
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License
Dr. Thump: Predicting Watermelon Sweetness from Induced Vibrations
Walker Conference Center B
The main goal of this project was to determine the best sound frequencies produced from thumping a watermelon that predict sweetness. Building on previous research, a program using MATLAB would separate the thump signal into its corresponding harmonic frequencies using the Fast Fourier Transform (FFT). These frequencies, along with the watermelon's weight and other information from the FFT spectrum, are used to determine a correlation between the recorded data and a watermelon’s sweetness level, measured by a refractometer in Brix. After testing numerous watermelons, a linear regression model was developed to predict a watermelon’s potential sweetness, with promising results. This process has been programmed into an iPhone app so that testing can be done anywhere within seconds. The hope was to make the process available to the public so that they could regularly select better watermelons before leaving the store and cutting the watermelon, which we have achieved with a new iPhone app. A future goal is to continue to determine if the same regression model will work for similar melons, like cantaloupe and honeydew, as well as bring the app to the Android platform.