Nondestructive firmness prediction for apples using through-transmission ultrasound

Assessing flesh firmness is essential for determining the maturity stage and commercial shelf life of apples. However, traditional penetrometer testing is destructive and renders the sample unsellable. A research team has developed a force-monitored through-transmission (TT) ultrasound system designed to predict apple firmness completely nondestructively.

Controlled coupling force and acoustic signal analysis

Achieving consistent acoustic coupling on soft or irregular fruit surfaces without causing mechanical damage is a major technical challenge. The novel platform overcomes this bottleneck by incorporating a load-cell-monitored pressing unit that maintains a steady contact force of 15 N during measurement.

By transmitting 50 kHz ultrasonic pulses through the intact apple, the system captures wave velocity, attenuation, and key frequency-domain spectral descriptors. Changes in the tissue's elastic modulus during softening directly alter these acoustic signal characteristics.

High-accuracy modeling for postharvest sorting

Machine learning models were trained on the extracted acoustic features and optimized using hyperparameter tuning algorithms. The predictions showed strong agreement with reference destructive compression tests, achieving minimal prediction error across varying firmness levels.

This force-monitored ultrasonic approach provides a robust framework for developing online sorting lines capable of grading internal apple quality in real time, enhancing batch management in packing houses.

Source

Arisandy, Y. P., et al. (2027). Nondestructive firmness prediction for apples using force-monitored through-transmission ultrasound. Postharvest Biology and Technology, 243, 114738. https://doi.org/10.1016/j.postharvbio.2026.114738