Abstract: Accurate analysis of fruit and vegetable quality is of great significance for ensuring food safety, improving consumer satisfaction, and promoting the sustainable development of the fruit and vegetable industry. Machine vision technology has been widely used in the fruit and vegetable industry in recent years. Traditional machine learning algorithms often have limitations when dealing with large amounts of complex image data generated by machine vision, and their performance cannot meet the actual needs. The integration of machine vision and deep learning algorithms enables efficient analysis and processing of complex fruit and vegetable images. The fruit and vegetable quality detection system based on machine vision and deep learning has achieved remarkable results in practical applications, providing strong technical support to the intelligent upgrading of the fruit and vegetable industry. This review summarizes recent progress on machine vision based on deep learning in fruit and vegetable quality analysis. It discusses the current challenges facing this field and future development trends with respect to the construction of public datasets, the development of lightweight models and 3D sensing devices, multimodal fusion, model interpretability, the development of portable and miniaturized devices, and the construction of a full-industry-chain intelligent fruit and vegetable management system empowered by the Internet of Things (IoT) and blockchain technology. These efforts are expected to promote the technological upgrading and collaborative innovation of the fruit and vegetable industry.

Key words: fruits and vegetables; machine vision; deep learning; convolutional neural networks; freshness