Res. Agr. Eng., 2020, 66(1):1-7 | DOI: 10.17221/13/2019-RAE

Dielectric technique combined with artificial neural network and support vector regression in moisture content prediction of oliveOriginal Paper

Mahdi Rashvand¹, Mahmoud Soltani Firouz^*,2

¹ Department of Machine Design and Mechatronics, Institute of Mechanics, Iranian Research Organization for Science and Technology, Tehran, Iran

² Department of Mechanics of Agricultural Machinery, Faculty of Agricultural Engineering and Technology, University of Tehran, Karaj, Iran

Olives are one of the most important agriculture crops in the world, which are harvested in different stages of growth for various uses. One of the ways to detect the adequate time to process the olives is to determine their moisture content. In this study, to determine the moisture content of olives, a dielectric technique was used in seven periods of harvesting and three different varieties of olive including Oily, Mary and Fishemi. The dielectric properties of the olive fruits were measured using an electronic device in the range of 0.1-30 MHz. Artificial Neural Network (ANN) and Support Vector Regression (SVR) methods were applied to develop the prediction models by using the obtained data acquired by the system. The best results (R = 0.999 and MSE = 0.014) were obtained by the ANN model with a topology of 384-12-1 (384 features in the input vector, 12 neurons in the hidden layer and 1 output). The results obtained indicated the acceptable accuracy of the dielectric technique combined with the ANN model.

Keywords: capacitive sensor; data mining; estimation; quality factor

Published: March 31, 2020  Show citation