基于人工神经网络和生物光子学的小麦籽粒状态分类

摘要

In this paper,biophoton analytical technology (BPAT) was combined with artificial neural network related algorithm to recognize different states of wheat kernels.Taking wheat kernels soaked in water for different time and wheat kernels before and after removing germs as the study objects,we acquired the biophoton emission of the test samples by using a customized ultraweak bioluminescence detector.Then,we classified the measured biophoton emission data of the wheat kernels by a BP (Back Propagation) algorithm of the artificial neural network.The results showed that the recognition rate of germ-free wheat was above 90％,and the recognition rate of new and old wheat was above 70％.%将生物光子分析技术与人工神经网络相关算法相结合,识别小麦籽粒的不同状态.以经过不同时间的水浸泡处理和去胚芽前后的小麦籽粒为研究对象,通过定制的生物超微弱发光测试仪获取试验样品的生物光子辐射.利用人工神经网络中的误差反向传播(Back Propagation,BP)算法对测量的小麦籽粒生物光子辐射数据进行分类研究.结果显示,该方法对有无胚芽小麦的识别率均达90％以上,对于新旧小麦的识别率也在70％以上.