Skip to main content
ARS Home » Midwest Area » East Lansing, Michigan » Sugarbeet and Bean Research » Research » Publications at this Location » Publication #316180

Title: Nondestructive detection of chilling injury in cucumber fruit using hyperspectral imaging with feature selection and supervised classification

item CEN, HAIYAN - Michigan State University
item Lu, Renfu
item ZHU, QIBING - Jiangnan University
item MENDOZA, FERNANDO - Michigan State University

Submitted to: Postharvest Biology and Technology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/1/2015
Publication Date: 9/19/2015
Citation: Cen, H., Lu, R., Zhu, Q., Mendoza, F. 2015. Nondestructive detection of chilling injury in cucumber fruit using hyperspectral imaging with feature selection and supervised classification. Postharvest Biology and Technology. 111:325-361.

Interpretive Summary: Cucumbers are susceptible to physiological disorder resulting from postharvest storage at low temperatures. Chilling injury symptoms in cucumbers are difficult to detect at the early stage, because they are often not visible at the surface of the fruit. In this research, a hyperspectral imaging system was used to acquire hyperspectral reflectance (for the visible region of 500-675 nm) and transmittance (for the region of 675-1,000 nm) images from pickling cucumbers. One hundred thirty normal ‘Excursion’ pickling cucumbers first underwent chilling treatments at 4 degrees Celsius and 95% relative humidity for 2, 5 and 8 days, respectively, followed with 3 and 6 days of post-chilling storage at room temperature to induce different degrees of chilling injury in the cucumbers. The test cucumbers were images for each chilling treatment and post-chilling storage date, and then classified into three grades of normal, lightly chilling injury and severely chilling injury by visible inspection of each sliced sample. Three mathematical methods (i.e., mutual information feature selection, max-relevance min-redundancy, and sequential forward selection) were used to select the optimal wavebands and two-waveband ratio combinations, followed by three classification algorithms (naïve Bayes, support vector machine, and k-nearest neighbor), for chilling injury classification. Most useful wavebands were found from the transmittance images in the near-infrared region. Both spectral and image feature extraction methods produced similar classification results. The best classification method yielded overall accuracies of 100% for two-class classification (i.e., normal and chilling injury) and 91.6% for three-class classification. This research demonstrated the potential of hyperspectral imaging technique, and the effectiveness of using optimal selected waveband ratios, for detection of chilling injury in cucumbers. The optimal wavebands and their ratio combinations can be applied for online inspection of chilling injury in cucumbers.

Technical Abstract: Chilling injury, as a physiological disorder in cucumbers, occurs after the fruit has been subjected to low temperatures. It is thus desirable to detect chilling injury at early stages and/or remove chilling injured cucumbers during sorting and grading. This research was aimed to apply hyperspectral imaging technique, combined with feature selection methods and supervised classification algorithms, to detect chilling injury in cucumbers. Hyperspectral reflectance (500-675 nm) and transmittance (675-1,000 nm) images for normal and chilling injured cucumbers were acquired, using an in-house developed online hyperspectral imaging system. Three feature selection methods including mutual information feature selection (MIFS), max-relevance min-redundancy (MRMR), and sequential forward selection (SFS) were used and compared for optimal wavebands selection. Supervised classifications with naïve Bayes, support vector machine (SVM), and k-nearest neighbor were then implemented for the two-class (i.e., normal and chilling) and three-class (i.e., normal, lightly chilling, and severely chilling) classifications based on the spectral and image analysis at selected two-band ratios. It was found that the majority of the optimal wavebands selected by MIFS, MRMR, and SFS for both two-class and three-class classifications were from the spectral transmittance images in the short-near infrared region. The SFS feature selection method together with the SVM classifier resulted in the best overall classification accuracy of 100% and the relative area under the receiver operating characteristic curve value of 1 for the two-class classification, and the overall accuracy of 90.5% for the three-class classification, based on the spectral analysis. The classification results based on the textural features (first-order statistics and second-order statistics features) extracted from the optimal two-band ratio images were comparable to those achieved using the spectral features, with the best overall accuracies of 100% and 91.6% for the two-class and the three class classifications, respectively. These results demonstrated the potential of hyperspectral imaging technique for online detection of chilling injury in cucumbers.