Skip to main content
ARS Home » Midwest Area » East Lansing, Michigan » Sugarbeet and Bean Research » Research » Publications at this Location » Publication #242362

Title: A Hyperspectral Imaging System for Quality Detection of Pickles

item ARIANA, DIWAN - Michigan State University
item Lu, Renfu

Submitted to: Meeting Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 8/1/2009
Publication Date: 9/10/2009
Citation: Ariana, D., Lu, R. 2009. A Hyperspectral Imaging System for Quality Detection of Pickles. In: Zhang, Q., editor. Proceedings of the 4th IFAC International Workshop on Bio-Robotics, Information Technology, and Intelligent Control for Bioproduction Systems, September 10-11, 2009, Champaign, Illinois. Paper No. 801.

Interpretive Summary:

Technical Abstract: A hyperspectral imaging system in simultaneous reflectance (400-675 nm) and transmittance (675-1000 nm) modes was developed for detection of hollow or bloater damage on whole pickles. Hyperspectral reflectance and transmittance images were acquired from normal and bloated whole pickle samples collected from a commercial pickle processing plant. Principal component analysis was applied to the hyperspectral images of the pickle samples, and the second principal component score images were used for defect detection by means of image segmentation method. An overall classification accuracy of 86% was achieved when the transmittance images for 675-1000 nm were used, compared with the manual classification accuracy of 70% and the reflectance imaging classification accuracy of 63%. Hyperspectral transmittance imaging is much more effective for detecting internal defect in whole pickles than reflectance imaging for the visible region. With further improvement, the hyperspectral imaging system can meet the need of detecting bloated pickles in a commercial plant setting.