Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Environmental Microbial & Food Safety Laboratory » Research » Research Project #421312

Research Project: DEVELOPMENT OF SENSING AND INSTRUMENTATION TECHNOLOGIES FOR FOOD SAFETY AND SANITATION INSPECTION IN FRESH FRUIT AND VEGETABLE PROCESSING

Location: Environmental Microbial & Food Safety Laboratory

2015 Annual Report


1a. Objectives (from AD-416):
The first objective is to develop and validate multitask in-line real-time inspection technologies for small to large processors that simultaneously detect contaminants and defects of fruits and vegetables. 1a) Evaluate visible/near-infrared reflectance and fluorescence imaging techniques for whole-surface detection of fecal material, damage, defects, and spoilage artifacts on fruits and vegetables. 1b) Identify multispectral wavebands and develop detection algorithms and image segmentation procedures for whole-surface inspection of produce that can be utilized for multitask screening for safety and quality concerns. Integrate and test methods for use in in-line multitask inspection systems. 1c) Develop and evaluate methods to facilitate whole-surface line-scan imaging of fruits and vegetables for in-line inspection. 1d) Develop and evaluate two prototype multitask inspection systems, one for fruits such as apples and tomatoes and a second for leafy green vegetables such as spinach and lettuce. The second objective is to develop and validate portable optical sensing technologies for detecting the presence of chemical and biological contaminants on food products and processing surfaces. 2a) Evaluate fluorescence, reflectance, and Raman spectral and imaging technologies for use in rapid sample analysis to detect fecal contamination, organic residues, bacterial biofilms, and food adulterants. 2b) Develop and validate a portable Raman-based hyperspectral imaging platform that can be used for macro-scale imaging of food samples as large as intact fruits and vegetables. 2c) Develop and validate handheld imaging devices for contamination and sanitation inspection in processing environments. 2d) Develop and validate imaging platform for in-field detection of fecal contamination.


1b. Approach (from AD-416):
The previous project included four patents (pending/issued) for methods and technologies developed: multitask line-scan imaging inspection, macro-scale laser-induced fluorescence imaging, Raman spectral detection of melamine adulteration, and image-based portable handheld sanitation inspection devices. This new project will build upon these previous accomplishments to develop prototype devices for commercialization. Rapid line-scan imaging technologies developed during the previous project cycle will be used to construct prototype whole-surface in-line inspection systems for simultaneously detecting surface contamination and defects using a single camera. This research focuses primarily on fresh fruits and vegetables, such as leafy greens, apples, and tomatoes, and on the detection of defects and of fecal contamination (a recognized source of human pathogens associated with fresh fruits and vegetables). Two prototype whole-surface in-line inspection systems will be developed, one for flat leafy produce such as Romaine lettuce and baby spinach, and a second for round-shaped produce such as apples and tomatoes. These systems will incorporate multitasking capabilities that allow users to select desired inspection criteria, and to optimize wavelengths and thresholds to address changes in produce characteristics on-the-fly. To detect chemical and biological substances of food safety interest, and to address the needs of the fruit and vegetable industries for evaluation or inspection tools for rapid on-site or in situ assessment of food safety risks, portable NIR (1000 to 2200 nm) hyperspectral imaging and Raman hyperspectral macro-scale imaging systems will be developed and validated . These enhanced capabilities will improve the existing toolbox of available imaging technologies for addressing unforeseen biological/chemical contamination problems in a timely manner. To enhance existing survey methods in produce processing plants, a previously developed handheld imaging device for inspecting poultry processing areas will serve as the basis for the development of a similar system for inspecting produce processing surfaces. The handheld inspection devices are intended as assistive tools for human inspectors to use during off-line inspection of processing equipment surfaces. To address the industry-identified need to survey produce fields for fecal contamination, technology to detect feces in produce fields will be developed based on a previously patented laser-induced fluorescence imaging technique. The proposed field imaging platform will assist industry in addressing in-field in situ detection of fecal contamination. As an applied engineering research project, the effective outcome of this work should be commercialization of the technologies developed. Critical to this end is collaboration with industry partners. Thus, this project will continue strategic partnerships with four companies with whom Cooperative Research and Development Agreements (CRADAs) have been established.


3. Progress Report:
Significant progress has been made on development and validation of spectral imaging techniques and instrumentation for use on-farm or during processing operations to reduce food safety risks from contaminated fresh produce, and for detecting intentional adulteration of food ingredients. To assure comprehensive online quality and safety inspection of fruits and vegetables in processing plants, whole-surface sample presentation and imaging regimes must be considered. Because of the morphological differences between fruits and leafy greens, separate methods and processing-conveyor systems were developed for implementing whole-surface inspection of round fruits and of flat leafy green vegetables. In 2015, one of the two U.S. patents pending for the ARS whole-surface produce-imaging methods was issued. Improved prototypes of sample rotation devices and conveyor systems for round fruits and (separately) for relatively flat leafy-greens were designed and developed. For the multitask imaging module, a computer-controllable power management unit for high-power LED-based lighting optimized for online fluorescence imaging of produce was designed. The whole-surface imaging methods coupled with the patented ARS multitask imaging technology will allow thorough safety and quality inspection of round fruits and leafy greens on commercial processing lines. This progress relates to project objective 1 (development of technologies for online produce safety and quality inspection). The Raman chemical imaging technology has shown great promise as a rapid and nondestructive method for identifying, detecting, or quantifying adulterants in food ingredients, with potential for adoption by industry groups as a standard testing method. A new line laser with 40 W laser power was acquired to upgrade the patent pending line-scan Raman chemical imaging system. The upgrade allows Raman chemical imaging with an improved spectral data quality and at speeds that are three orders of magnitude higher than those of conventional Raman imaging systems. The improvements in line-scan Raman imaging will allow the use of the technique as a routine scanning tool in food industries for safety and quality applications. In addition, a new gradient temperature Raman spectroscopy (GTRS) technique was developed and a U.S. patent application for the method and device was submitted in 2015. The technique applies the precise temperature gradients utilized in differential scanning calorimetry to Raman spectroscopy. Applied to molecularly complex materials, the technique allows better elucidation of molecular-level structures useful for fingerprinting of chemical substances. This progress relates to project objective 2 (development of optical sensing technologies to detect contaminants in foods). Handheld fluorescence imaging devices have been demonstrated effective as visual aid tools for improving the efficacy of cleaning and sanitation procedures in food processing. The handheld imaging technology has been patented by ARS and licensed by a U.S. commercial partner. The first commercial version (biofilm imaging monitor) is in development and is expected to be released in 2015. Extended testing and validation of the technology are planned for use of the handheld imaging devices as visual-aid inspection tools, followed by development of standard protocols for end-users, and, ultimately, for regulatory uses. As the technology advances, new component upgrades to optimize the system for real-world use will continue. To reduce the weight of the handheld imaging device, a power management circuit board for LEDs and other electronic components was designed. In June 2015, the latest ARS version of the handheld imaging device was transferred under a Material Transfer Agreement to the U.S. Army Natick Soldier Research Development and Engineering Center. U.S. Army Public Health Command personnel serve as Department of Defense (DOD) executive agents and inspect all food and water procured by the DOD. The handheld imaging device will be tested and evaluated during audits of food preparation facilities by Veterinary Inspectors at remote overseas locations. This progress relates to project objective 2 (development of sensing technology to detect contaminants on food processing surfaces and facilities). Efforts to develop methods and instrumentation for detecting fecal material and signs of animal intrusion in produce fields prior to harvest have continued. Design and development of a semi-autonomous cart to serve as a field platform for mounting the laser-induced fluorescence imaging system was initiated. The cart will enhance the potential capabilities of the imaging system by allowing imaging at night as well as detection using time-resolved fluorescence imaging techniques, and will also reduce net operating costs by relieving the need for a human operator. Vehicle design specifications were established, and a collaboration with University of Maryland Baltimore Campus was also established. The imaging system itself was modified, including development of new mounting hardware for optics and integration of a new gated-camera, acquisition in progress, to facilitate conversion of the laboratory system to field use. This progress relates to project objective 2 (development of in-field pre-harvest technology to detect contaminants on foods).


4. Accomplishments
1. Handheld fluorescence imaging device as an aid for detection of food residues on processing surfaces. Contamination of food with pathogenic bacteria can lead to foodborne illnesses. Food processing surfaces can serve as a medium for cross-contamination if sanitization procedures are inadequate. Ensuring that these surfaces are effectively cleaned and sanitized is important for the food industry's efforts to reduce risks of foodborne illnesses and their related costs. A handheld fluorescence imaging device developed by ARS was assessed for detection of three types of food residues (spinach leaf, milk, and bovine red meat) that have been associated with foodborne illness outbreaks, on two common processing surfaces (high-density polyethylene and food-grade stainless steel). Interchangeable optical filters were selected to optimize the contrast between food residues and processing surfaces as detected using hyperspectral fluorescence imaging. The fluorescence imaging plus image analysis differentiated food residues from the processing surfaces more clearly than did human visual inspection in ambient lighting. This optical sensing device can be used over relatively large or complex surfaces of processing equipment to detect food residues, and has potential for use in the food industry as an aid for detection of specific (targeted) food residues.

2. High-throughput Raman chemical imaging-based detection of food adulterants. Food safety incidents in recent years due to milk adulteration have brought increased interest in developing rapid and accurate screening methods for authenticating milk products. ARS researchers in Beltsville, MD have developed a point-scan Raman chemical imaging technique to detect various adulterants in milk powder. One limitation of previous Raman detection methods included the long sampling time, typically measured in hours, needed to collect images using a point-scan Raman imaging system, which precluded use of the method for rapid screening for adulterants in food powders such as dry milk powder. To address this limitation, a high-throughput line-scan Raman imaging system was developed. The system uses a 785 nm line laser to project a 24 cm long excitation line on the sample surface. Raman scattering signals along the laser line are collected by a detection module consisting of a lens, a dispersive Raman imaging spectrograph, and a Charge Coupled Device (CCD) camera. A hypercube is accumulated line-by-line as a motorized table moves the samples transversely across the laser line. Compared to the point-scan system, the line-scan system is capable of imaging larger sample areas in shorter sampling times—i.e., minutes instead of hours. The developed system was used to collect Raman images from milk powder samples mixed with different concentrations of chemical adulterants. Images were created for visualizing the distribution of adulterant particles in the milk powder. The reduced scan times make the method feasible for high-throughput Raman imaging-based inspection of food powders and other materials in food processing plants.

3. Development of a three-CCD portable multispectral imaging system. ARS researchers in Beltsville, Maryland, have developed a myriad of nondestructive spectral imaging technologies for safety and quality evaluations of fruits and vegetables. Recent studies have suggested the need for imaging devices capable of multispectral imaging beyond the visible region that will allow for both quality and safety evaluations of agricultural commodities, compared to conventional multispectral imaging devices that lack flexibility in spectral waveband selectivity for such applications. A portable three-CCD camera with significant improvements over existing imaging devices was developed with a beam-splitter prism assembly to accommodate three interference filters, to use with the three CCD sensors, that can be easily changed for application-specific multispectral waveband selection in the 400 to 1000 nm region. Integrated electronic components on printed circuit boards with firmware programming were designed to enable parallel processing, synchronization, and independent control of the three CCD sensors, ensuring the transfer of data without significant delay or data loss due to buffering. The system can stream 30 frames (three single-waveband images in each frame) per second. The utility of the 3CCD camera system was demonstrated in the laboratory for detecting defect spots on apples and the imaging system can be used to enhance the current online machine vision technologies for fruit inspection.

4. Multispectral image algorithms for inspection of fecal contamination on leafy greens. Consumption of fresh leafy green vegetables (e.g., lettuce, and spinach) or fresh fruits contaminated by fecal matter can result in foodborne illnesses caused by human pathogens such as E. Coli or Salmonella. Hyperspectral fluorescence imaging with ultraviolet-A excitation was used to evaluate the feasibility of two-waveband fluorescence algorithms for the detection of bovine fecal contaminants on the abaxial and adaxial surfaces of Romaine lettuce and baby spinach leaves. Correlation analysis was used to select the most significant waveband pairs for two-band ratio and difference methods in distinguishing contaminated and uncontaminated leaf areas. Two-band ratios using bands at 665.6 nm and 680.0 nm (F665.6/F680.0) for lettuce and at 660.8 nm and 680.0 nm (F660.8/F680.0) for spinach were found to effectively differentiate all contamination spots from normal leaf areas. These methods for fecal contamination detection could be implemented for online screening of raw leafy greens using multispectral imaging with high spectral resolution for use in produce processing operations.

5. Determination of optimal bands for detecting fecal contamination on spinach leaves. Ensuring food safety in the production of fresh produce for human consumption is a global issue and must be addressed to decrease the incidence of foodborne illnesses and the associated costs. Hyperspectral fluorescence imaging using violet (405 nm) excitation was coupled with multivariate image analysis techniques and evaluated for the detection of fecal contaminants on spinach leaves. Fluorescence images were acquired spanning 464 to 800 nm for fecal contamination on both fresh leaves (pre-storage) and leaves subjected to a 27-day period of controlled storage. During the storage period, fluorescence emission profiles of the spinach leaves were monitored; peak emission blue-shifts were observed to accompany a color change from green to green-yellow-brown. A detection model developed using partial least squares discriminant analysis correctly detected fecal contamination on 100% of the fresh green spinach leaves, with 19% false positives for the non-fresh post-storage leaves. A wavelength ratio technique using four wavebands (680, 688, 703 and 723 nm) successfully identified 100% of fecal contaminations on both fresh and non-fresh leaves. These methods have potential for implementation in on-line fluorescence imaging inspection of fresh produce for fecal contaminant detection to help reduce the risk of foodborne illnesses.


Review Publications
Qin, J., Chao, K., Cho, B., Kim, M.S. 2015. High-throughput Raman chemical imaging for rapid evaluation of food safety and quality. Transactions of the ASABE. 57:1783-1792.

Schmidt, W.F., Broadhurst, C., Qin, J., Lee, H., Nguyen, J.K., Chao, K., Hapeman, C.J., Shelton, D.R., Kim, M.S. 2015. Temperature dependent Raman spectroscopy of melamine and structural analogs in milk powder. Applied Spectroscopy. 669:398-406.

Lee, H., Everard, C., Kang, S., Cho, B., Chao, K., Chan, D.E., Kim, M.S. 2014. Multispectral fluorescence imaging for detection of bovine feces on Romaine lettuce and baby spinach leaves. Biosystems Engineering. 127:125-134.

Lefcourt, A.M., Beck, E., Lo, Y., Kim, M.S. 2015. Using hyperspectral fluorescence spectra of deli commodities to select wavelengths for surveying deli food contact surfaces. Journal of Biosystems Engineering. 40(2):145-152.

Beck, E., Lefcourt, A.M., Kim, M.S., Lo, M. 2015. Use of a portable fluorescence imaging device to facilitate cleaning of deli slicers. Food Control. 51:256-262.

Lee, H., Park, S.H., Boh, S.H., Lim, J., Kim, M.S. 2014. Development of a portable 3CCD camera system for multispectral imaging of biological samples. Sensors. 14:20262-20273.

Kandpal, L.M., Lee, S., Kim, M.S., Bae, H., Cho, B. 2015. Short wave infrared (SW-IR) hyperspectral imaging technique for examination of aflatoxin B_1 on corn kernels. Food Control. 51:171-176.

Baek, I., Kim, M.S., Lee, H., Cho, B. 2014. Optimal fluorescence waveband determination for detecting defect cherry tomatoes using fluorescence excitation-emission matrix. Sensors. 14:21483-21496.

Lohumi, S., Lee, S., Lee, W., Kim, M.S., Mo, C., Bae, H., Cho, B. 2014. Detection of starch adulteration in onion powder by FT-NIR and FT-IR spectroscopy. Journal of Agricultural and Food Chemistry. 62:9246-9251.

Rao, X., Yang, C., Ying, Y., Kim, M.S., Chan, D.E., Chao, K. 2014. Differentiation of deciduous-calyx Korla fragrant pears using NIR hyperspectral imaging analysis. Transactions of the ASABE. 57:1875-1883.

Garrido-Novell, C., Perez-Marin, D., Guerrero-Ginel, E., Kim, M.S., Garrido-Varo, A. 2015. Quantification and spatial characterization of moisture and NaCl content of Iberian dry-cured ham slices using NIR hyperspectral imaging. Journal of Food Engineering. 153:117-123.

Lee, H., Kim, M.S., Jeong, D., Delwiche, S.R., Chao, K., Cho, B. 2014. Detection of cracks on tomatoes using hyperspectral near-infrared reflectance imaging system. Journal of Food Engineering. 14(10):18837-18850.

Le, H.D., Kim, M.S., Hwang, J., Yang, Y., U-Thainual, P., Kang, J.U., Kim, D. 2014. An average enumeration method of hyperspectral imaging data for quantitative evaluation of medical device surface contamination. Journal of Biomedical Optics. 5:3613-3627.

Lim, J., Mo, C., Kim, G., Kang, S., Lee, K., Kim, M.S., Moon, J. 2014. Non-destructive and rapid prediction of moisture content in red pepper (Capsicum annuum L.) powder using near-infrared spectroscopy and a partial least squares regression model. Journal of Biosystems Engineering. 39:184-193.