Egg Quality Prediction Using Dielectric and Visual Properties Based on Artificial Neural Network
Egg Quality Prediction Using Dielectric and Visual Properties Based on Artificial Neural Network
复制标题
DOI：
10.1007/s12161-014-9948-x
复制
发表时间：
2015-03-01
影响因子：
2.9
通讯作者：
Alimardani, Reza
中科院分区：
文献类型：
--
作者：
Soltani, Mahmoud;Omid, Mahmoud;Alimardani, Reza
近年来，快速、无损的食品质量检测技术得到了发展。本研究旨在设计和开发一种基于无线电频率（40 kHz至20 MHz）介电技术、机器视觉和人工神经网络（ANN）技术的卵子鉴定系统。研究了哈夫单位、蛋黄指数、蛋黄/蛋白比和蛋黄质量作为鸡蛋的品质因子。对所设计的电子器件进行了标定和评价，以预测上述参数。人工神经网络验证的哈夫单位、蛋黄指数、蛋黄/蛋白和蛋黄重的决定系数R(2)值分别为0.998、0.998、0.998和0.994。在评价模型中，哈夫单位、蛋黄指数、蛋黄/蛋白和蛋黄重的平均绝对百分比误差（mape）分别为5.41、6.84、8.79和4.24%。评价结果表明，所设计的装置可可靠地用于鸡蛋品质指标的预测。
Recently, rapid and nondestructive technologies have been developed in the qualification of food products. This research aimed to design and develop an egg qualifying system based on dielectric technology in the range of radio frequency (40 kHz to 20 MHz), machine vision, and artificial neural network (ANN) techniques. Haugh unit, yolk index, yolk/albumen ratio, and yolk weight were studied as quality factors of egg. The designed electronic device was calibrated and evaluated for prediction of the mentioned parameters. The coefficient of determination (R (2)) values in the validation of the ANN were 0.998, 0.998, 0.998, and 0.994 for the Haugh unit, yolk index, yolk/albumen, and yolk weight, respectively. In evaluation mode, the mean absolute percent errors (MAPEs) obtained were 5.41, 6.84, 8.79, and 4.24 % for the Haugh unit, yolk index, yolk/albumen, and yolk weight, respectively. Results of the evaluation showed the designed device can be confidently used in the prediction of egg quality indices.