Acoustics‐Driven Artificial Intelligence Framework for Non‐Destructive Quality Estimation of Fruits: A Comprehensive Review

ABSTRACT Quality estimation of agro‐food products is a top priority for producers, suppliers, and consumers. Non‐destructive techniques, particularly acoustic vibration methodology, have emerged as promising tools for the pre‐harvest and post‐harvest quality evaluation of agricultural products. This technique is effective in assessing the stiffness or firmness of fruits based on the mass and resonance frequency of the sample. This review explores the principles and applications of acoustic vibration methods in evaluating the quality of agro‐food products. Impact and forced excitation methods induce sample vibration, while contact and non‐contact sensors detect the resulting acoustic signals. These signals are processed to extract features that determine quality parameters. The review highlights the integration of artificial neural networks (ANN) and machine learning algorithms to enhance these assessments' accuracy, reliability, and efficiency. The acoustic vibration technique has the capability to evaluate several internal quality parameters such as texture, firmness, crunchiness, and crispiness. It is also useful for defect detection, sorting, grading, and assessing the ripeness of fruits and vegetables. Incorporating ANN and machine learning algorithms reduces human error and improves system reliability and speed. Additionally, this technique can indirectly measure soluble solid content, titratable acidity, and water activity in various commodities. Acoustic vibration technology holds tremendous potential for online quality detection of food products in an industrial context. However, further research is needed to develop cost‐effective and accurate devices to make this technology accessible to the agro‐food industry.