Location: Quality and Safety Assessment Research Unit
2025 Annual Report
Objectives
1. Develop imaging technologies to detect and identify plastics during poultry processing with hyperspectral imaging and artificial intelligence.
1A. Develop hyperspectral imaging technology for detection and identification of plastic foreign objects during poultry processing.
1B. Develop AI technology for enhanced detection and smart robotic removal of foreign materials in hyperspectral imagery during poultry processing.
2. Detection and identification of foodborne bacteria and toxins in poultry products with high-throughput hyperspectral microscopy and surface plasmon resonance imaging.
2A. Rapid monitoring of indicator microorganisms in poultry processing.
2B. Develop advanced hyperspectral microscope imaging (HMI) methods and system for label-free detection and identification of pathogens at the cellular level with no enrichment.
2C. Develop high-sensitive and selective immunoassay method and system for foodborne bacteria and toxin detection with surface plasmon resonance imaging.
3. Eliminate the production of semicarbazide in non-nitrofurazone treated poultry by optimization of antimicrobial treatments and/or alternative antimicrobials during processing.
4. Develop safe and effective poultry processing strategies (scalding-picking-evisceration procedures) to reduce foodborne contaminants (pathogens/chemical) and enhance the sustainability of poultry processing.
4a. Develop sustainable poultry processing using artificial intelligence (AI) technology to improve poultry food safety.
4b. Develop Internet of Things (IoT) technology with various sensing platforms and data analytics for smart poultry processing and safety.
Approach
Research on poultry safety will focus on: 1) developing and validating early, rapid, sensitive, and/or high-throughput optical sensing techniques for detecting physical and biological hazards in poultry products, and 2) eliminating semicarbazide in non-nitrofurazone treated poultry by optimization of antimicrobial treatments. In research on optical detection of physical hazards, spectroscopic and hyperspectral imaging (HSI) technologies will be developed for detection and identification of plastic foreign objects. A robot rejector and control software will be developed to eliminate foreign materials (FM) when detected by HSI. Artificial intelligence (AI) technology will be developed for enhanced detection and smart robotic removal of FM during poultry processing through the development and evaluation of customized deep learning algorithms based on hyperspectral imaging. A vision-guided smart robotic manipulator will be designed and built to remove FM by self-learning AI algorithms. To develop methods and techniques for detecting and identifying biological hazards, time-lapse image data on pure-culture indicator organisms and poultry carcass rinses from different processing locations will be collected to build a library, which will be used for on-line counting of microcolonies to build prototype systems. To detect foodborne pathogens, hyperspectral microscope imaging (HMI) methods will be developed with a spectral library of various pathogens using two HMI platforms of acousto-optical tunable filter (AOTF) and Fabry-Perot interferometer (FPI). In accordance with optimization of parameters on HMI and hypercubes, a transportable HMI system will be developed embedded with AI-based software for classification and identification. To identify foodborne bacteria and toxins, a highly-sensitive and selective immunoassay method and system will be developed using surface plasmon resonance imaging (SPRi). Microfluidic devices will be designed, simulated and fabricated for bacterial enrichment and separation. Both materials and parameters to develop a 3D printed biosensor for multiplex detection of pathogenic bacteria and toxins will be optimized and evaluated with food samples. Finally, a portable 3D printing platform for biosensor fabrication by integrating sample enrichment cartridge, biochip and SPRi detector will be developed. To develop techniques for eliminating the production of semicarbazide (SEM) in non-nitrofurazone treated poultry, a methodology for SEM analysis in chicken meat and a data library relating poultry processing conditions to SEM formation will be developed. Specifically, SEM in chicken leg quarters obtained from multiple processing facilities will be analyzed and methods to eliminate SEM production in poultry products under processing conditions will be developed.
Progress Report
Significant progress was made on developing hyperspectral imaging technology for detecting and identifying plastic foreign objects during poultry processing (Sub-objective 1A) and supporting artificial intelligence (AI) technologies for enhanced detection and smart robotic removal of foreign materials (Sub-objective 1B). Research was conducted to improve AI technology for high-performance deep learning (DL) computing, specifically targeting real-time detection and robotic removal of foreign materials in AI-powered hyperspectral imaging during poultry processing. The challenge of real-time hyperspectral imaging is capturing and processing hyperspectral images of halved chicken breast fillets at a rate of up to 0.15 sec/fillet, corresponding to a line speed of 200 birds/minute (BPM), higher than the typical U.S. line speed of 140 BPM. The research achieved an AI model inference time of approximately 0.04 sec/fillet using parallel CPU and GPU computing. These findings show there is commercial viability and feasibility of deploying a hyperspectral imaging technology with a complex AI model for poultry meat inspection. Research to demonstrate the feasibility of the real-time hyperspectral imaging technology was conducted to develop a Windows 11 application program that enables the deployment of the real-time AI technology based on hyperspectral imaging to detect foreign plastic materials.
Research was conducted to enhance the intelligence of AI technology by developing an advanced, smart identification model, called the hyperspectral vision transformer, for detecting types of foreign plastic objects using real-time hyperspectral imaging. Hyperspectral imaging has posed challenges in analyzing both the spectral and spatial properties of materials embedded in the measured images. The developed model effectively combines spatial and spectral data to accurately identify and categorize plastic contaminants found in raw chicken fillets. This capability has potential for poultry processors to trace back the sources of contamination and helps mitigate incidents involving foreign materials. This enhanced AI model was trained on a hyperspectral image dataset of 12 different plastic types in the wavelength range of 1,000 to 1,700 nm. It achieved a high classification accuracy rate of over 99% for identifying plastic types commonly found in poultry processing environments. Additionally, progress was made in developing a robotic system equipped with a delta robot, capable of moving and picking up unwanted plastic materials. Ongoing research aims to further improve these capabilities by integrating smarter self-learning methods into the technology.
Research was conducted by ARS scientists at Athens, Georgia to develop advanced hyperspectral microscope imaging (HMI) methods and systems for label-free detection and identification of pathogens at the cellular level with no enrichment (Sub-objective 2B). Different methods were developed and analyzed for improving the generalizability of DL models for hyperspectral data analysis applied to foodborne bacterial detection. Various multimodal data fusion methods using DL were tested and a set of techniques that allowed improved DL model performance on novel, unseen data sets were identified. Collaborative research trials were conducted to determine the potential for rapid Salmonella serovar classification using AI-enabled HMI with different data pre-processing approaches and to investigate source-specific variations in hyperspectral signatures of Salmonella Infantis. A mercury cadmium telluride (MCT) hyperspectral imaging system was acquired, setup, and data acquisition protocols optimized to expand the technical capabilities of the project. Experimental trials were conducted to determine the feasibility of using hyperspectral imaging for microplastic (MP) detection in soil. Using the MCT system, significant progress was made on detecting low concentrations of MPs in soil by coupling machine learning algorithms with short-wave infrared (SWIR) hyperspectral imaging to detect two types of MP - polyamide and polyethylene - with maximum particle sizes of 50 and 300 µm, respectively. This study highlighted the feasibility of using MCT hyperspectral imaging for rapid, non-invasive, and effective detection (93.8% overall accuracy achieved) of MP at concentrations as low as 0.01%.
Research was also conducted to develop highly-sensitive and selective immunoassay methods and systems for detection of foodborne bacteria and toxins (Sub-objective 2C). Collaborative research trials were conducted to test various microfluidic platforms, separation cartridge designs, and machine learning models for high-throughput, high-sensitivity bacteria sampling and analysis. Significant progress was made for the separation of foodborne bacteria using a spiral microfluidic channel where particles are separated based on size and shape. Using a viscoelastic fluid (polyethylene oxide) as the sheath flow, different shapes of bacterial cells experienced different forces in the channel, entering different outlets for separation. For effective sorting performance on matrices, the microfluidic device was designed and fabricated with 4 channels and demonstrated the ability to separate bacteria based on both size (1, 2 and 5 mm) and shape (elongated and isotropic).
Research to develop safe and effective poultry processing strategies to reduce foodborne contaminants and enhance the sustainability of poultry processing (Objective 4) was advanced through two separate agreements with university collaborators. Through one agreement, university collaborators conducted research on: 1) treating poultry processing wastewater for re-use in irrigation (redesigned poultryponics system and conducted 5 lettuce production trials), 2) testing alternative valorization approaches (black soldier fly larvae culture, rendering followed by yeast or algae culture, and anaerobic digestion) on solid residues from poultry processing plants, 3) the effects of broiler stunning method (electrical stunning vs. controlled atmosphere stunning) on carcass bacterial populations at different stages in the processing plant, 4) the effects of lairage temperature on blood metabolites of carbon dioxide stunned broilers, 5) the interacting effects of broiler stunning method, chilling methods, and deboning time on product quality, microbial growth, and shelf-life, and 6) the effects of poultry meat labeling related to processing and animal welfare on consumer acceptance. As part of this collaboration, ARS researchers completed trials to determine the effects of broiler processing methods (stunning/deboning) on meat biochemistry and functionality.
Through a separate agreement, collaborators conducted trials on the effects of on-farm broiler slaughter on processing efficiency (defeathering quality), bird welfare, food safety (microbial load and viscera damage), and product quality. Collaborators designed and tested two scientific devices (feather puller and environmental chamber) for future work. ARS researchers evaluated the effects of alternative processing on meat quality and functionality. Collaborators also: 1) conducted laboratory and field trials to identify interfering factors and potential mitigation strategies on the peracetic acid sensor they developed for poultry processing applications, and 2) continued further testing on an alternative in-line, rotational carcass chilling system.
Poultry litter moisture levels are known to impact bird health and food safety. Research to explore the potential for developing a microwave-sensing technology for measuring poultry litter moisture content was conducted. Initial trials used a laboratory vector network analyzer (VNA) to measure the dielectric properties of fresh poultry litter samples (rice hulls, peanut shells, and pine shavings) at room temperature from 2-18 GHz. Measurements were taken at three different sample densities varying in moisture content from 10 to 32%. Linearity was observed between dielectric properties and moisture content and correlations were observed between moisture content and calibration functions. Sample water activity was also measured and showed similarly promising results. In preparation for future trials on “used” poultry litter, a mobile microwave measurement station was developed so that such measurements could be taken in biosafety level 2 (BSL-2) animal and laboratory facilities. Custom antenna brackets were designed, 3D printed and mounted to a cart to hold horn-lens antennas. Custom software was written to facilitate microwave measurements with a VNA remotely from an external laptop.
During FY2025 laboratory research tools to increase the throughput and accuracy of several microbiological laboratory procedures were developed, tested, and implemented. A five-piece guide frame system that allows a rapid, easy, and accurate transfer of a 96-well microtiter plate assay to a large-volume petri dish with solid agar was developed. Once the design of the system was finalized, the systems were 3D printed in-house and tested in the laboratory. Testing demonstrated that usage of the guide frame system could reduce the time necessary for this routine microbiology lab procedure from 4 h to 30 minutes and substantially reduce the amount of lab supplies needed. To facilitate laboratory research on the effects of adding amendments (acidifiers) to poultry litter, a 3D printed applicator tool was designed and tested. The tool allows for uniform application of amendments to the surface of poultry litter samples in testing jars. Applicators with different hole sizes (3, 4, and 5-mm) and depths were produced and tested for their ability to provide uniform amendment applications. Four custom applicators (5-mm holes, 3-cm depth) were produced and utilized by ARS scientists to perform studies on litter.
Accomplishments
1. Hyperspectral image-based deep learning technology for real-time detection and identification of plastic foreign materials in poultry processing. Plastic foreign material contamination in poultry meat products poses a food safety risk and can lead to product recalls. ARS researchers in Athens, Georgia, have developed a method to improve the real-time inference performance of a deep learning model for hyperspectral imaging to detect foreign materials in poultry processing. This enhancement can process data at processing speeds of up to 750 birds per minute, which is significantly greater than the typical rate of 140 birds per minute. Additionally, a novel artificial intelligence technology, called a hyperspectral vision transformer, was developed that achieves over 99% classification accuracy for 12 types of plastic contaminants on the surface of broiler breast fillets. These advancements greatly enhance the feasibility of implementing hyperspectral imaging technologies for foreign material detection in poultry products.
2. Identification of foodborne bacteria in unseen environments using artificial intelligence-based methods. Developing sensitive and selective technologies for detecting foodborne bacteria is challenging due to the complexity of food matrices and heterogeneous bacterial populations. Artificial intelligence techniques show great potential for detecting bacteria, but robust bacterial identification in real-world environments remains challenging due to distributional shifts between training and real-world data. To tackle this problem, ARS researchers in Athens, Georgia, developed a model, called an uncertainty-aware multimodal ensemble, using hyperspectral microscope imaging and deep learning. This model, trained on pure-culture bacterial data, achieved 89% accuracy on unseen mixed-culture bacterial data, significantly outperforming the 51% accuracy of common baseline methods. These advancements will facilitate the further development of technologies for detecting pathogens in food matrices.
3. Guide frame system to adapt microtiter plate assay to solid media. Within industrial, regulatory, and research microbiology laboratories, many procedures require high-throughput assessment of small-volume inoculation. Such assessments are often performed on 96-well microtiter plates. However, for bacteria species that do not yield a color or cloudiness change in the broth, contents of the microtiter plates are typically transferred to plate cultures with solid media for detection. The common practice of transferring a microtiter plate assay to a plate culture is time-consuming and often inaccurate. ARS researchers in Athens, Georgia, developed a novel 3D printed guide frame system to adapt a 96-well microtiter plate assay to a large volume petri dish with solid media. The guide frame system can reduce the time necessary for this routine microbiology laboratory procedure from 4 h to 30 minutes while ensuring a high degree of accuracy and substantially reducing the amount of lab supplies needed. The device can be 3D printed for less than $3 and the 3D printing script for the guide frame system was made available on NIH 3D, an open platform to provide wide and easy access to this device.
Review Publications
Wei, C., Wang, W., Jlabo, Y., Yoon, S.C., Ni, X., Wang, X., Song, Z., Hu, Y. 2025. Detection of camellia oil adulteration with excitation-emission matrix fluorescence spectra and machine learning. Journal of Food Composition and Analysis. https://doi.org/10.1007/s10895-025-04229-7.
Guo, X., Wang, W., Jia, B., Ni, X., Zhuang, H., Yoon, S.C., Gold, S.E., Pokoo-Aikins, A., Mitchell, T.R., Bowker, B.C., Ye, J. 2025. Detection of aflatoxin B1 content and revelation of its dynamic accumulation process using visible/near-infrared hyperspectral and microscopic imaging. Food Microbiology. https://doi.org/10.1016/j.ijfoodmicro.2025.111065.
Jessup, A., Hayden, M., An, J., Jones, J., Barahona-Dominguez, L., Dees, J., Cho, S. 2025. Influence of electrical- and gas-stunned broilers on sensory characteristics and consumer acceptance during chilled storage. Journal of Food Science. 90(3),e70076. https://doi.org/10.1111/1750-3841.70076.
Khan, Z., Yoon, S.C., Bhandarkare, S.M. 2025. Deep learning model compression and hardware acceleration for high-performance foreign material detection on poultry meat using NIR hyperspectral imaging. Sensors. 25(3):970. https://doi.org/10.3390/s25030970.
Arthur, W., Morgan, Z., Inskeep, A.E., Browne, C., Wells, D.E., Bourassa, D.V., Higgins, B.T. 2025. Assessing nitrogen recovery in Poultryponics for hydroponic lettuce production using treated poultry processing wastewater for increased nitrogen neutrality. Bioresource Technology. 422: 132227. https://doi.org/10.1016/j.biortech.2025.132227.
Shanmugam, S.R., Schorer, R., Arthur, W., Drabold, E., Rudar, M., Higgins, B.T. 2024. Upcycling nutrients from poultry slaughterhouse wastewater through cultivation of the nutritional yeast, Yarrowia lipolytica. Journal of Environmental Chemical Engineering. 13(1): 115245. https://doi.org/10.1016/j.jece.2024.115245.
Arthur, W., Akplah, C., Drabold, E., Manjankattil, S., Smith, J., Wells, D.E., Bourassa, D.V., Higgins, B.T. 2025. Dosing salmonella into poultryponics: fate of salmonella during treatment of poultry processing wastewater and irrigation of hydroponic lettuce. Journal of Environmental Management. 377: 124559. https://doi.org/10.1016/j.jenvman.2025.124559.
Riggs, M., Hauck, R., Baker-Cook, B., Osborne, R., Pal, A., Bethonico Terra, M., Sims, G., Urrutia, A., Orellana-Galindo, L., Reina, M., Devillena, J., Bourassa, D. 2022. Meat quality of broiler chickens processed using electrical and controlled atmosphere stunning systems. Poultry Science. https://doi.org/10.1016/j.psj.2022.102422.
Wellington, A., Morgan, Z., Reina Antillon, M., Drabold, E., Wells, D.E., Bourassa, D.V., Wang, Q., Higgins, B.T. 2024. Pilot-scale evaluation of poultryponics: insights into nitrogen utilization and food pathogen dynamics. ACS Environmental Science & Technology Water. 4(9):3964-3975. https://doi.org/10.1021/acsestwater.4c00262.
Drabold, E.T., Sakhakarmy, M., Shanmugam, S.R., Adhikari, S., Arthur, W., Rudar, M., Boersma, M., Wang, Q., Higgins, B.T. 2025. Thermal hydrolysis of poultry byproducts for the production of microbial media. Nature Scientific Reports. 15: 6107. https://doi.org/10.1038/s41598-025-90411-7.
Eady, M.B., Park, B. 2016. Classification of Salmonella Enterica serotypes with selective bands using visible/NIR hyperspectral imaging. Journal of Microscopy. 263(1):10-19.
Gamble, G.R., Park, B., Yoon, S.C., Lawrence, K.C. 2016. Effect of sample preparation on the discrimination of bacterial isolates cultured in liquid nutrient media using laser induced breakdown spectroscopy. Applied Spectroscopy. 70(3):494-504.
Chen, J., Park, B., Huang, Y., Zhao, Y., Kwon, Y. 2017. Label-free SERS detection of Salmonella Typhimurium on DNA aptamer modified AgNR substrates. Journal of Food Measurement and Characterization. 11: 1773-1779.
Chen, H., Shin, T., Park, B., Ro, K.S., Jeong, C., Jeon, H., Tan, P. 2025. Accurately detecting low concentrations of microplastics in soils using short-wave infrared hyperspectral imaging. Soil & Environmental Health. 3. https://doi.org/10.1016/j.seh.2025.100157.
Seo, Y., Park, B., Hinton Jr, A., Yoon, S.C., Lawrence, K.C., Gamble, G.R. 2016. Identification of staphylococcus species with hyperspectral microscope imaging and classification algrorithms. Journal of Food Measurement and Characterization. 10(2):253-263.
Wang, W., Ni, X., Lawrence, K.C., Yoon, S.C., Heitschmidt, G.W., Feldner, P.W. 2015. Feasibility of detecting Aflatoxin B1 in single maize kernels using hyperspectral imaging. Journal of Food Engineering. Volume 166, Pages 182–192 (2015).
Wei, W., Lawrence, K.C., Ni, X., Yoon, S.C., Heitschmidt, G.W., Feldner, P.W. 2014. Near-infrared hyperspectral imaging for detecting Aflatoxin B1 of maize kernels. Food Control. Volume 51, Pages 347-355, May 2015.
Yoon, S.C., Shin, T., Heitschmidt, G.W., Lawrence, K.C., Park, B., Gamble, G.R. 2019. Hyperspectral image recovery using a color camera for detecting colonies of foodborne pathogens on agar plate. Journal of Biosystems Engineering. https://doi.org/10.1007/s42853-019-00024-y.