Hyperspectral imaging technology for quality and safety evaluation of horticultural products: A review and celebration of the past 20-year progress

Abstract

In the past 20 years, hyperspectral imaging has been widely investigated as an emerging, promising technology for evaluating quality and safety of horticultural products. This technology has originated from remote sensing and joins the domains of machine vision and point spectroscopy to provide superior image segmentation for the detection of defects and contaminations, and to map the chemical composition. Thanks to the advancements in instrumentation and data analysis in the past two decades, hyperspectral imaging technology has evolved into a powerful nondestructive inspection tool and the scope of applications in postharvest quality and safety evaluation has expanded tremendously. In this article, different imaging modes (reflectance, transmittance, fluorescence and Raman) and their combinations, and the potential for real-time acquisition of hyperspectral images at industry relevant speeds are first discussed in terms of their advantages and disadvantages. Next reviewed are different data processing/analysis methods and associated steps from data pre-processing over the spectral and spatial domains to the actual model building and performance evaluation. An overview is then given of hyperspectral imaging applications for external quality and defect evaluation, internal quality and maturity assessment, and food safety detection of 2 horticultural products. Finally, a brief discussion is presented on the challenges and opportunities in future development and application of hyperspectral imaging technology in food quality and safety evaluation of horticultural products.

Quantitative analysis and visualization of moisture and anthocyanins content in purple sweet potato by Vis–NIR hyperspectral imaging

“…This poor modeling outcome can be due to the fact that the spectral variables in the full‐range spectra mostly includes noise from the environment, leading to uncertainty and low prediction potential of the target‐detection models (Zhou et al., 2020), a phenomenon that often involves the need for image preprocessing routines and optimal wavelength selection. Coupled with the large volumes of image data at tens to hundreds of wavelengths (Lu et al., 2020), these other limitations make it difficult to set up real‐time inspection systems with HSI system. And according to ElMasry et al.…”

Linking Tissue Damage to Hyperspectral Reflectance for Non-Invasive Monitoring of Apple Fruit in Orchards

“…Anthocyanins responsible for red color of apples increase their consumer acceptance [33] serving also for the protection against UV and excessive sun irradiation [34,35]. In the case of fruit, the information about the fruit carried by the reflected light also translates into the perceived fruit quality (appearance, ripeness, presence of damage) [1,18,21]. Non-destructive assessment of pigments in plant in general as well as in apple fruit is complicated by overlapping light absorption by individual pigments and non-linear relationship of reflectance vs. pigment content, especially in the bands of strong absorption [2,17].…”

“…In the past two decades, it has evolved into a powerful noninvasive inspection tool; the scope of applications in postharvest quality and safety evaluation using HRI has expanded tremendously [1,18]. Still, overwhelming majority of the studies on application of Vis-NIR spectroscopy is focused on postharvest applications where the basics of the spectroscopic and chemometric theory are established and a broad range of commercially available instrumentation for packing lines is available [21]. However, the studies dedicated to the spectral imaging of fruit in the field and interpretation of reflectance images of fruit recorded under ambient conditions are much scarcer.…”

“…There were also doubts that IH can be helpful for monitoring ecophysiological responses even under controlled conditions [20]. These uncertainties and limitations stem particularly from insufficient understanding of the relationships between biochemistry and architecture features of plant tissues and their measured optical properties [21]. Solving these problems are central to confident interpretation of reflectance images, development of computer vision and robotic systems for automated fruit harvesting and grading capable of working in orchard [5,18].…”

Linking Tissue Damage to Hyperspectral Reflectance for Non-Invasive Monitoring of Apple Fruit in Orchards

“…In the past two decades, the HRI technology has evolved into a powerful noninvasive inspection tool. However, the overwhelming majority of the studies on application of Vis-NIR spectroscopy is focused on postharvest applications where the basics of the spectroscopic and chemometric theory are established and a broad range of commercially available instrumentation for packing lines is available [ 25 ]. The studies dedicated to the spectral imaging of fruits in the field and interpretation of reflectance images of fruits recorded under ambient conditions are much scarcer.…”