Skip to main content
ARS Home » Southeast Area » Athens, Georgia » U.S. National Poultry Research Center » Quality Safety and Assessment Research » Research » Publications at this Location » Publication #379061

Research Project: Assessment of Quality Attributes of Poultry Products, Grain, Seed, Nuts, and Feed

Location: Quality Safety and Assessment Research

Title: Determination of foreign material content in uncleaned peanuts by microwave measurements and machine learning techniques

Author
item JULRAT, SAKOL - Oak Ridge Institute For Science And Education (ORISE)
item Trabelsi, Samir

Submitted to: Journal of Microwave Power and Electromagnetic Energy
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/30/2021
Publication Date: 10/23/2021
Citation: Julrat, S., Trabelsi, S. 2021. Determination of foreign material content in uncleaned peanuts by microwave measurements and machine learning techniques. Journal of Microwave Power and Electromagnetic Energy. https://doi.org/10.1080/08327823.2021.1993047.
DOI: https://doi.org/10.1080/08327823.2021.1993047

Interpretive Summary: Among the parameters used in peanut grading is foreign material content. Currently this done manually which is labor intensive and time consuming. In this paper, a low-cost microwave sensor was built and tested in the laboratory and in the field for rapid and nondestructive determination of in-shell peanuts moisture content and foreign material content. The microwave sensor comprised two subsystems one for measurements on cleaned peanut pods and one for measurement on the bulk uncleaned peanut sample which is a mixture of peanut pods and foreign materials (shells, sticks, peanut raisins, stones, ...etc). Measurements on a handful of cleaned peanuts provided the in-shell peanuts moisture content. The latter along with the measured dielectric properties of the bulk sample provided the foreign material content. For this purpose two algorithms were used: The linear regression algorithm and the artificial neural network algorithm. Results show that moisture content of peanut pods can be predicted with a standard error of calibration (SEC) of 0.58%. For foreign material content, use of the regression model showed it can be predicted with a standard error of performance (SEP) of 2.4%. However, a better performance was obtained when the artificial neural network was used with an SEP of 1.4%. The microwave sensor can be used at peanut buying points to determine rapidly and nondestructively peanut pods' moisture content and foreign material content before sending the sample to the grading room. This will result in significant labor and time reduction and thus in significant cost savings. Use of the sensor in the field, before even taking the peanut to the buying point, may save the growers costs associated with transportation, cleaning, and drying.

Technical Abstract: Foreign material content determination in uncleaned peanuts based on dielectric properties and bulk density measurements by microwave techniques is presented in this paper. A microwave free-space transmission technique was used at 10 GHz for these measurements. Two subsystems for measuring the dielectric properties of cleaned unshelled peanuts (nine-peanut pods) and uncleaned peanuts (container size: 12.1 cm x 21 cm x 20.5 cm) were developed. The moisture content of the cleaned unshelled peanuts was determined from the nine-peanut pods measurement for the refence moisture content and used for foreign material determination algorithms. The dielectric properties and bulk density measurements from the uncleaned peanut sample (container size: 12.1 cm x 21 cm x 20.5 cm) are related to the foreign material content. These parameters were supplied to the machine learning algorithms, linear regression technique and artificial neural network algorithms. Results for the artificial neural network algorithm showed the best estimate of foreign material content (standard error of performance 1.36%) compared to the linear regression algorithm.