Research Project: Advancing Technologies for Grain Trait Measurement and Storage Preservation

Location: Stored Product Insect and Engineering Research

2023 Annual Report

Objectives
OBJECTIVE 1: Improve stored grain management, technology and processing practices to maintain grain end-use quality by controlling or eliminating adverse storage environments, insect infestations. Sub-objective 1A: Develop an insect monitoring and identification device for behavioral study and pest management in food facility environments. Sub-objective 1B: Increase efficacy of fumigation of milled and whole grain products through improved monitoring and modeling of fumigant applications. Sub-objective 1C: Increase efficacy of insecticidal aerosol applications in grain processing facilities based on measurement and modeling of droplet distribution and deposition. OBJECTIVE 2: Resolve existing issues and develop new technologies and techniques to rapidly and accurately evaluate intrinsic grain and seed quality to improve breeding efficiency, marketability, end-product use and environmental influences. Sub-objective 2A: Develop imaging methods for the detection of hard vitreous amber color (HVAC) of Durum wheat seeds as a replacement for manual wheat inspection. Sub-objective 2B: Selecting maize seeds for breeding programs using single seed near infrared spectroscopy (NIR) to improve hybrid development.

Approach
United States farmers annually (2016-2018) grow 562 million metric tons of corn, soybeans, wheat, sorghum and other grains to supply the nation and the world with food, animal feed and biofuels. Our project goal is to improve U.S. grain quality and international competitiveness through the application of engineering principles to rapidly measure grain traits and to maintain grain and grain-based product quality after harvest. We propose to develop unique instrumented systems to rapidly measure quality or compositional traits for breeders when selecting traits for varietal development. We also propose to develop technology to detect and control insects and maintain product quality during handling, processing and storage. This research will lead to expedited development of varieties and hybrids by breeders; better systems and information for storage management by farmers and processors, resulting in better profitability and production efficiency, less waste and increased food availability using fewer resources.

Progress Report
ARS researchers at Manhattan Kansas, continued work on improvement to stored grain management technology by focusing on processing practices to maintain grain end-use quality and by controlling or eliminating adverse storage environments and insect infestations under Objective 1. Under Objective 2, work continued to develop new technologies and techniques that rapidly and accurately evaluate intrinsic grain and seed traits to improve breeding efficiency which has direct links to marketability, end-product use, and environmental influences. Under Sub-objective 1A, Artificial Intelligence (AI) was used to identify several stored product insect species from images, with very good accuracy, providing a first step towards creating autonomous systems to monitor stored grain and food warehouses. Grain insect monitoring equipment being developed by an industry collaborator will be used to collect images and train insect AI identification models. Development of this technology will provide commercial grain storage operators enhanced detection and prevention measures for insect control as well as development in general food storage preservation. Sub-objective 1B We explored fumigation efficacy in two areas. Firstly, fumigation distribution inside moving railcars was quantified. Railcars are an often-overlooked study area but provide broad mobility for insect pests to travel and contaminate facilities. Secondly, fumigant distribution in grain structures was studied using Computational Fluid Dynamic (CFD) models with predictions compared with experimental measurements. Alternatives to gaseous fumigation using aerosols were studied under Sub-Objective 1C. Methods to measure spray distribution droplet size were developed and used inside various structures using different delivery methods. Modelling of these applications using CFD was also completed. These results will help commercial applicators develop good operating procedures for the pest control industry and better understanding of the factors affecting the fumigation process. Knowledge of the total amount of stored grain in the U.S. is critical for marketing and food security. Estimates of this amount are dependent on grain properties and time in storage. Grain test weight is one of these properties and was studied under Sub-objective 1D. Discrete Element Modelling (DEM) was used to estimate test weight versus lab test weight allowing a better understanding of grain properties which affect test weight the most and ultimately will provide better estimates of total grain-in-storage. Development of imaging methods for the detection of hard vitreous amber color (HVAC) of Durum wheat seeds, Sub-objective 2A, was completed by our research engineers. Back-lit transmission images were used and provided good results when compared with the standard visual method used for grading. The method has great potential to help in the commercial grading of Durum wheat. This has set the groundwork for looking at the characteristics of other wheat types by this method. We continued work toward more rapid development of corn hybrids under Sub-objective 2B. Selection of maize haploid seeds for breeding programs made advances using single seed near infrared spectroscopy (NIR). The instrumentation designed and developed to do this was installed with two collaborators in Iowa and Florida. Preliminary samples were used to develop prediction models to measure oil content differences between haploid and hybrid seeds. Methods to implement a standard operating procedure for sorting were explored and are being evaluated as a new crop becomes available in 2023. This work has the potential to significantly reduce the time and cost for hybrid development by replacing much of the human visual inspection currently used for sorting.

Accomplishments
1. Artificial intelligence used for image-based identification of stored product insects. Monitoring stored product insect pests is a common practice in managing stored grain and ensuring grain quality from storage to sale. However, the current manual sampling and monitoring methods used in large grain storage and food production facilities are time-consuming, labor-intensive, and require expertise for accurate species identification, all of which incur significant expenses for the facility. A major advance in addressing these challenges has been made by USDA-ARS scientists from Manhattan, Kansas. They have developed image-based identification using deep learning methods of artificial intelligence (AI) for five common stored grain insect species: the lesser grain borer, rusty grain beetle, red flour beetle, rice weevil, and saw-toothed grain beetle.Identification is highly reliable, achieving at least 96% accuracy for all species. The specific insect features found to be most useful for reliable species identification were the differences in the shapes of the insect body and head. In contrast, differences in antennae, legs, and snouts were less critical for accurate identification. By employing this image-based species identification, the bottlenecks associated with previous methods are bypassed, leading to quicker response times for pest population detection and ultimately resulting in reduced damage and economic losses. This work is part of a broader effort to develop camera-based systems for automated monitoring of other pests in warehouses, flour mills, and general food storage facilities. By providing an improved method to identify insect species, this development opens the possibility of more efficient pest management. This research has led to the discussion with a company producing image-based insect monitoring devices to co-develop insect identification using AI technology, showcasing the potential of AI to improve pest management.

Review Publications
Mendoza, Q.A., Pordesimo, L.O., Nielsen, M.L., Armstrong, P.R., Campbell, J.F. 2023. Application of machine learning for insect monitoring in grain facilities. Artificial Intelligence. 4:348-360. https://doi.org/10.3390/ai4010017.
Gokhan, H., Armstrong, P.R., Mendoza, P.D., Seabourn, B.W. 2022. Compositional analysis in sorghum (Sorghum bicolor) NIR spectral techniques based on mean spectra from single seeds. Frontiers in Plant Science. 13:995328. https://doi.org/10.3389/fpls.2022.995328.
Mendoza, P.D., Armstrong, P.R., Peiris, K.H., Siliveru, K., Bean, S.R., Pordesimo, L.O. 2023. Prediction of sorghum oil and kernel weight using near-infrared hyperspectral imaging. Cereal Chemistry. 100(3):775-783. https://doi.org/10.1002/cche.10656.
Hacisalihoglu, G., Armstrong, P.R. 2023. Crop seed phenomics: Enabling nondestructive phenotyping approaches for characterization of functional and quality traits. Plants. 12(5):1177. https://doi.org/10.3390/plants12051177.
Boac, J., Casada, M.E., Pordesimo, L.O., Petingco, M., Maghirang, R., Harner III, J. 2023. Evaluation of particle models of corn kernels for discrete element method simulation of shelled corn mass flow. Smart Agricultural Technology. 4. Article 100197. https://doi.org/10.1016/j.atech.2023.100197.
Morrison III, W.R., Brabec, D.L., Bruce, A., Arthur, F.H., Athanassiou, C.G. 2023. Immediate and delayed movement of resistant and susceptible adults of Tribolium castaneum (Herbst) (Coleoptera: Tenebrionidae) after short exposures to phosphine. Pest Management Science. 79:2006-2074. https://doi.org/10.1002/ps.7383.
Brabec, D.L., Kaloudis, E., Athanassiou, C., Campbell, J.F., Agrafioti, P., Scheff, D.S., Bantas, S., Sotiroudas, V. 2022. Fumigation monitoring and modeling of hopper-bottom railcars loaded with corn grits. Journal of Biosystems Engineering. 47:358-369. https://doi.org/10.1007/s42853-022-00148-8.
Pordesimo, L., Igathinathane, C., Holt, G. 2023. Hammer milling switchgrass from weathered bales. Industrial Crops and Products. 197. Article 116647. https://doi.org/10.1016/j.indcrop.2023.116647.
Al-Bakri, A., Al-Amery, M., Su, K., Anderson, H., Geneve, R., Crocker, M., Teets, N., Armstrong, P.R., Kachroo, P., Hildebrand, D. 2023. Assessment of oil quantification methods for high oil seeds. Analytical Chemistry. 50. Article 102715. https://doi.org/10.1016/j.bcab.2023.102715.
Pordesimo, L.O., Igathinathane, C., Bevans, B.D., Holzgraefe, D.P. 2023. Potential of dimensional measurements of individual pellets for evaluating feed pellet quality. Applied Engineering in Agriculture. 38(5):777-785. https://doi.org/10.13031/aea.14845.
Pordesimo, L.O., Casada, M.E., McNeill, S.G. 2023. On farm storage of grain crops. Smart Agricultural Technology. 17:1-13. https://doi.org/10.1007/978-3-030-89123-7_122-1.