Low cost measurement setup based on a piezoelectric microphone for estimating apple bruising using Shannon entropy

• A new measurement system to estimate the bruise volume of apples was introduced.
• The proposed system based on piezoelectric microphone is very cheap and robust.
• Shannon entropy signal obtained from acoustic impact signal.
• Pulse-width time and entropy peak are extracted from Shannon entropy signal.
• These features are efficiently used for estimating apple bruise volume.

This study presents a low cost measurement setup for estimating apple bruise volume during drop impact tests. The proposed measurement system consists of three main stages namely recording impact sound, signal processing, and estimating bruise volume. In the recording stage, the impact sound occurred when striking the apple on the surface was recorded by a piezoelectric microphone into computer. The impact sound was converted into Shannon entropy signals based on the Kernel density approach, and two measurable quantities namely entropy peak and pulse-width time were extracted from Shannon entropy signals in the signal processing stage. In the bruise volume estimating stage, the regression curves that map entropy peak and pulse-width features into bruise volume are estimated using measurements data. To evaluate the performance of the proposed method, we conducted several drop experiments with six different drop heights and two different temperatures (1 and 19 °C). Experimental results showed that bruise volume was exponentially proportional to the entropy peak and it was inversely proportional to pulse-width time as power functions, and all the curves were fitted with coefficients of determination of more than 0.98. Moreover, it was found that colder apples developed less bruise volume then warmer ones.