Skip to main content
ARS Home » Research » Publications at this Location » Publication #260323
Title: Visible to SWIR hyperspectral imaging for produce safety and quality evaluation

Author
item Kim, Moon
item Delwiche, Stephen - Steve
item Chao, Kuanglin - Kevin Chao
item Lefcourt, Alan
item Chan, Diane

Submitted to: Sensing and Instrumentation for Food Quality and Safety
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/2/2012
Publication Date: 2/24/2012
Citation: Kim, M.S., Delwiche, S.R., Chao, K., Lefcourt, A.M., Chan, D.E. 2012. Visible to SWIR hyperspectral imaging for produce safety and quality evaluation. Sensing and Instrumentation for Food Quality and Safety. 5(5):155-164.

Interpretive Summary: Over the past decade, our group has developed multiple generations of line-scan based hyperspectral imaging systems for visible to near-infrared (VNIR: 400-1000 nm) reflectance and recently expanded the imaging capabilities to include SWIR from 1000 nm to 1700 nm. Despite the widespread use of hyperspectral imaging techniques in the visible to near-infrared (VNIR: 400-1000 nm) for various forms of agro-food evaluation, seldom reported are the instrument artifacts that may affect the quality of image data. Furthermore, hyperspectral-based research has focused largely on the development of image processing techniques and detection aspects with minimal attention given to illustrating the underlying value of imaging with sufficient spatial resolution across the visible to short-wavelength infrared (SWIR) regions. With the use of our most recently developed VNIR and SWIR hyperspectral imaging systems, spectral and spatial attributes of apples with defects from 400 to 1700 nm are presented. In addition, we developed methods to characterize the hardware artifacts, calibrate wavelengths, and to reduce inherent detector noise. We envision that the hyperspectral imaging techniques will continue to play a significant role in agro-food sector as a critical research tools, and in further applications for rapid inspection of produce for safety and quality evaluation. This study provides valuable information to food process scientists, food technologists, and engineers who develop nondestructive sensing technologies to evaluate produce for safety and quality.

Technical Abstract: Hyperspectral imaging techniques, combining the advantages of spectroscopy and imaging, have found wider use in food quality and safety evaluation applications during the past decade. In light of the prevalent use of hyperspectral imaging techniques in the visible to near-infrared (VNIR: 400 -1000 nm) for agro-food evaluations, seldom reported are the instrument artifacts that may affect the quality of image data. Furthermore, hyperspectral-based research has focused on the development of image processing and detection aspects with minimal attention given to illustrating the underlying value of imaging with sufficient spatial resolution in the regions spanning from the visible to short-wavelength infrared (SWIR: 1000-1700). We have developed multiple generations of line-scan based hyperspectral imaging systems and expanded the imaging capabilities in the SWIR. With the use of our most recently developed VNIR and SWIR hyperspectral imaging systems, spectral and spatial attributes of apples with defects from 400 to 1700 nm are presented. In addition, we characterize the second order effect in the 800 to 1000 nm that emanates from the use of a diffraction grating in the VNIR hyperspectral imaging system. We have devised methods to perform SWIR spectral calibration and to remove the bad pixels inherent to the SWIR InGaAs focal plane array used in the imaging system. We envision that hyperspectral imaging techniques will continue to play a significant role in the agro-food sector as critical research tools, and in further applications for rapid inspection of produce safety and quality.