Development of high-speed dual-camera system for batch screening of aflatoxin contamination of corn using multispectral fluorescence imaging

Authors: HAN, DEOK - Mississippi State University; YAO, HAIBO - Mississippi State University; HRUSKA, ZUZANA - Mississippi State University; KINCAID, RUSSELL - Mississippi State University; Rajasekaran, Kanniah - Rajah; Bhatnagar, Deepak

Submitted to: Transactions of the ASABE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/7/2019
Publication Date: 1/7/2019
Citation: Han, D., Yao, H., Hruska, Z., Kincaid, R., Rajasekaran, K., Bhatnagar, D. 2019. Development of high-speed dual-camera system for batch screening of aflatoxin contamination of corn using multispectral fluorescence imaging. Transactions of the ASABE. 62(2):381-391. https://doi.org/10.13031/trans.13125.
DOI: https://doi.org/10.13031/trans.13125

Interpretive Summary: Hyperspectral imaging reveals abundant information and has been a successful research tool throughout many areas including the military, environmental research, and agricultural research. However, the method has limitations in processing speed. The above mentioned studies have been centered on single kernel based studies or on small samples of 25g. Significant amount of time is needed to acquire and process the images for a large sample, in addition to addressing the problem of a huge data storage requirement. Grain elevators need to unload, on average, one 50,000-pound truckload every two minutes. Current chemical and optical methods of aflatoxin detection are unable to meet the screening requirement. Therefore, a high speed batch screening system with reliable accuracy is needed. The objective of the current study was to develop an imaging system to detect aflatoxin contaminated corn kernels, and to test the feasibility of the system with high sensitivity and specificity along with fast processing time. Our approach was to use multispectral images to cover large size samples with high speed, concurrently. To develop a high speed multispectral screening system for a large sample size, two high performance cameras were used in conjunction with dual UV excitation sources and novel detection software (AflaView™) was developed to collect fluorescence images of each corn kernel sample weighing 200g. In order to achieve optimum results with minimal screening time, four screenings per sample were used in this study. Screened samples were chemically analyzed for aflatoxin content. All samples were shuffled between imaging repetitions. Performance of the system was illustrated on different kinds of corn samples including laboratory inoculated kernels, and some commercial samples from grain elevators and farmers around the US. Applications of this method for practical screening of corn kernels for aflatoxin contamination is described in this paper.

Technical Abstract: Aflatoxins are fungal toxins produced by Aspergillus flavus. Food and feed crops get contaminated with carcinogenic aflatoxins resulting in economic losses as well as potentially serious health issues. Grain elevators need to unload, on average, one 50,000-pound truckload every two minutes. Current sampling based analytical chemistry methods for aflatoxin detection cannot meet this type of large throughput screening requirements. Therefore, a high speed, batch screening system with reliable accuracy is necessary. To develop a high speed multispectral screening system, two high performance cameras in conjunction with dual UV excitation sources and novel image processing software were utilized to collect fluorescence images of corn samples. Each camera simultaneously captures a single band fluorescence image (436 nm and 532 nm) from corn kernel samples, and the detection software processes the images to automatically detect contaminated kernels by using a normalized difference fluorescence index (NDFI). The system was tested with various commercial samples collected from different locations of the USA and laboratory samples that were prepared through artificial field inoculation. Each sample was imaged/screened four times, and the screened samples were chemically analyzed for aflatoxin content. All sample were shuffled between imaging repetitions to increase the likelihood of screening both the germ and endosperm sides of every kernel. Processing time for each screening was less than one second. The sensitivity and specificity were measured over given thresholds for NDFI, and the results were promising. High sensitivity (0.987) and specificity (0.96) were achieved for the laboratory samples, and high sensitivity (0.75 ~ 1) and somewhat lower specificity were obtained from commercial samples.