Page Banner

United States Department of Agriculture

Agricultural Research Service

Research Project: TECHNOLOGIES FOR ASSESSING AND GRADING QUALITY AND CONDITION OF CUCUMBERS AND TREE FRUITS

Location: Sugarbeet and Bean Research

Title: VISIBLE/NEAR-INFRARED HYPERSPECTRAL TRANSMITTANCE IMAGING FOR DETECTION OF INTERNAL MECHANICAL INJURY IN PICKLING CUCUMBERS

Authors
item Ariana, Diwan - USDA/FAS
item Lu, Renfu

Submitted to: ASABE Annual International Meeting
Publication Type: Proceedings
Publication Acceptance Date: May 13, 2006
Publication Date: July 9, 2006
Citation: Ariana, D.P., Lu, R. 2006. Visible/near-infrared hyperspectral transmittance imaging for detection of internal mechanical injury in pickling cucumbers. ASABE Annual International Meeting. Paper No. 063039.

Interpretive Summary: Pickling cucumbers are susceptible to damage during growth, harvest, transport, and processing, which could result in defective fruit. Cucumbers with external defects such as broken, smashed, and cuts are easy to identify and can be removed during post harvest sorting and grading. However, internal defects such as carpel separation or hollow centers are difficult to detect by human inspectors or machine vision systems without cutting open the fruit, which prohibits evaluation of individual cucumbers. Hence, a rapid and nondestructive method to evaluate internal defects would be valuable for the pickling cucumber industry. This research investigated a hyperspectral imaging technique in transmittance mode for detecting internal mechanical injury in pickling cucumbers resulting from dropping or rolling, which occurs frequently during mechanical harvesting and postharvest handling. Hyperspectral imaging is a technique that combines the main features of imaging and spectroscopy to acquire from the product item both spatial and spectral information simultaneously. Hyperspectral transmittance images were acquired from normal and defective cucumbers over a period of six days after they were harvested. Image processing algorithms were developed for detecting defective cucumbers from normal cucumbers. The hyperspectral imaging system correctly detected 86% of the defective cucumbers and achieved an overall classification accuracy of 92%. The technique can be implemented for online sorting and grading of pickling cucumbers. This would help to assure the quality and consistency of final pickled products, which are estimated to be worth more than five billion dollars for the U.S. pickle industry.

Technical Abstract: Internal product quality is an important aspect in the quality control and assurance of pickled products. A rapid and non-destructive method for internal defect detection would be of value to the pickling cucumber industry. A hyperspectral transmittance imaging technique was developed to detect internal mechanical injury in the form of carpel suture separation or hollow cucumbers that resulted from dropping and rolling under weight. Hyperspectral transmittance images were acquired from both normal and defective cucumbers at five times over a period of six days after harvest. Partial least squares discriminant analysis (PLS-DA) and hyperspectral image thresholding methods were used to classify cucumber samples into defective or normal classes. Transmittance spectra of defective and normal cucumbers were similar in shape but higher in magnitude for the defective cucumbers. Cucumbers, both normal and defective, had good transmittance in the near-infrared range of 700-950 nm, but they had poor or low transmittance in the visible range (450-700 nm). The hyperspectral image thresholding method resulted in higher classification accuracies compared to PLS-DA. The image thresholding method correctly detected 86% of the defective cucumbers and achieved an overall classification accuracy of 92%. The hyperspectral transmittance imaging technique has the potential for rapid detection of internal mechanical injury in pickling cucumbers.

Last Modified: 9/1/2014
Footer Content Back to Top of Page