Submitted to: Applied Engineering in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: August 29, 2005
Publication Date: January 1, 2006
Citation: Liu, Y., Chen, Y.R., Wang, C.-Y., Chan, D.C., Kim, M.S. 2006. Development of hyperspectral imaging technique for the detection of chilling injury in cucumbers; spectral and image analysis. Applied Eng. in Agric. 22:(1):101-111. Interpretive Summary: Exposure to low temperature environments during the storage and transportation process can induce chilling injury in cucumbers. Symptoms of chilling injury develop very rapidly, and affected areas might become sites for further fungal decay and bacterial infection after a few days at warmer temperatures. Such bacterial pathogens could be transmitted to humans by consumption of uncooked or mishandled cucumbers. Hence, one of the greatest concerns in the vegetable industry is to detect chilling injury, ultimately to provide safe and high quality cucumber products for consumers. Existing methods in cucumber safety/quality assessment, such as human visual inspection, can provide reliable information about chilling injured portions but are time consuming and unsuitable for on-line application. The development of fast, non-destructive, accurate, and on-line/at-line techniques is desired. Hyperspectral imaging spectroscopy could form the basis for such a technique due to its capacity for non-invasive sampling of large areas at higher speeds. This paper explored the fundamental spectral features of good cucumbers and chilling injured cucumber skins, and reported the determination of characteristic bands in the visible and NIR regions for chilling damaged cucumbers from region of interest (ROI) spectra. A number of classification methods for the discrimination of good cucumber skins from those of chilling injured ones were developed. The results revealed that both a dual-band ratio algorithm (R811 nm / R756 nm) and a principal component analysis (PCA) model from a narrow 733-848 nm spectral region can detect chilling injured skins with over 90% accuracy. We then utilized these findings to analyze the original hyperspectral images for the identification of chilling injury in cucumbers. The results revealed the great potential of the dual-band algorithm in hyperspectral imaging analysis for the detection of chilling injury in cucumbers. This information is useful to vegetable packers, retailers, and researchers who are interested in applying both visible/NIR and imaging spectroscopy based safety/quality grading or classifying.
Technical Abstract: Hyperspectral images of cucumbers were acquired before and during cold storage treatment as well as during subsequent room temperature (RT) storage to explore their potential use for the detection of chilling induced damage in whole cucumbers. Region of interest (ROI) spectral features of chilling injured areas, resulting from cold storage treatments at 0°C or 5 °C, showed a reduction in reflectance intensity during multi-day post-chilling periods of RT storage. Large spectral differences between good-smooth skins and chilling injured skins occurred in the 700-850 nm visible/NIR region. A number of data processing methods, including simple spectral band algorithms and principal component analysis (PCA), were attempted to discriminate the ROI spectra of good cucumber skins from those of chilling injured skins. Results revealed that using either a dual-band ratio algorithm (Q811/756) or a PCA model from a narrow spectral region of 733-848 nm could detect chilling injured skins with a success rate of over 90%. Furthermore, the dual-band algorithm was applied to the analysis of images of cucumbers at different stages of chilling damage, and the resultant images showed more correct identification of chilling injured spots than PCA method. The results also suggested that chilling injury was relatively difficult to detect at the stage of the first 0-2 days of post-chilling RT storage, due to insignificant manifestation of chilling induced symptoms.