Skip to main content
ARS Home » Midwest Area » East Lansing, Michigan » Sugarbeet and Bean Research » Research » Publications at this Location » Publication #227183

Title: Internal defect detection of whole pickles using hyperspectral reflectance and transmittance imaging

Author
item ARIANA, D - MICHIGAN ST UNIVERSITY
item Lu, Renfu

Submitted to: ASABE Annual International Meeting
Publication Type: Proceedings
Publication Acceptance Date: 5/20/2008
Publication Date: 6/30/2008
Citation: Ariana, D.P., Lu, R. 2008. Internal defect detection of whole pickles using hyperspectral reflectance and transmittance imaging. ASABE Annual International Meeting. Paper No. 084266.

Interpretive Summary: Good, normal brined cucumbers are usually packed as high-value pickles in whole, half or quarter size, whereas bloated or defective pickles are processed into lower value products such as relish (finely cut or finely chopped pickles). Bloated pickles are characterized by soft tissue and the presence of hollow center. Since bloating damage is largely hidden inside pickles, it is difficult to detect with machine vision systems that are primarily designed for inspecting external or surface characteristics. Currently, bloated pickles are separated from normal ones by manually inspecting individual pickles moving on the conveyor belts. This inspection method is labor intensive and unreliable due to speed demand and inspector fatigue. Therefore, development of an effective inspection system would be valuable to the pickling industry. In this research, we tested and evaluated an online prototype of hyperspectral imaging, a technique that acquires both spectral and spatial information simultaneously, for detection of defective whole pickles from normal pickles. A classification algorithm was developed to segregate normal pickles from defective or bloated pickles. The algorithm achieved 91% classification accuracy for defective pickles and 86% overall classification accuracy for both normal and defective pickles, which were better than the overall classification accuracy of 70% achieved by manual grading. The hyperspectral imaging technique is useful for internal defect detection of pickles. With further improvement in hardware and software, the technology can meet the need for quality inspection of pickles and pickled products in a commercial plant setting.

Technical Abstract: Hyperspectral imaging technique in simultaneous reflectance and transmittance modes was investigated for detection of hollow or bloater damage on whole pickles that was caused by mechanical injury during harvesting and handling or developed during the brining process. Normal and bloated pickle samples were collected from a commercial pickle processing plant. Hyperspectral images were acquired from the pickle samples using an on-line hyperspectral reflectance and transmittance imaging system covering the spectral range of 400-1000 nm. 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. The system achieved 91% classification accuracy for defective pickles and 86% overall classification accuracy for both normal and defective pickles, compared with the overall classification accuracy rate of 70% achieved by the human inspectors. Transmittance images at 675-1000 nm were much more effective for internal defect detection compared to reflectance images for the visible region of 500-675 nm. With further improvement, the hyperspectral imaging system could meet the need of detecting bloated pickles in a commercial plant setting.