Research Advances in Hyperspectral Imaging Technology for Fruit Quality Assessment

Hyperspectral imaging (HSI) is an advanced sensing technique that simultaneously acquires high-resolution spatial data and continuous spectral information, enabling non-destructive, real-time evaluation of both external and internal fruit quality attributes. Despite its widespread application in agricultural product assessment, comprehensive reviews specifically addressing fruit quality evaluation using HSI are limited. This paper presents a comprehensive review of recent advancements in the application of HSI technology for fruit quality detection.

This paper provides a comprehensive review from three key dimensions: scenario adaptability, technological evolution trends, and industrial implementation bottlenecks, with a further analysis of the research outlook in HSI applications for fruit quality assessment. Specifically, by employing non-destructive and rapid spectral imaging techniques, HSI has markedly enhanced the accuracy of assessing various quality parameters, including external appearance, surface defects, internal quality (such as sugar content, acidity, and moisture), and ripeness. Furthermore, significant progress has been achieved in utilizing HSI for disease detection, variety classification, and origin traceability, thereby providing robust technical support for fruit quality control and supply chain management. In addition, bibliometric analysis is utilized to identify key research areas and emerging trends in the application of HSI technology for fruit quality assessment.

Future research should focus on optimizing spectral dimensionality reduction techniques to enhance both the efficiency and accuracy of models. Transfer learning and incremental learning approaches should also be explored to improve the models' ability to generalize across various scenarios and fruit types. In parallel, developing lightweight system hardware and strengthening edge processing capabilities will be essential for enabling the practical deployment of HSI technology in real-world applications. Integrating lightweight deep learning networks and acceleration modules will support real-time inference, enhancing processing speed and facilitating faster data analysis. It is also crucial to establish standardized systems and protocols to promote the sharing of research findings and ensure broader application across different industries. Additionally, incorporating multimodal technologies, such as thermal imaging, gas sensors, and visual data, will improve the accuracy and robustness of detection platforms. This integration will allow for more precise and comprehensive assessments of fruit quality, further advancing the digitalization and intelligent application of HSI technology.