BP神经网络在烟梗长短梗率检测中的应用  

The Application of BP Neural Network in the Determination of the Stalk Length and Stem Rate

作　　者：崔云月 管一弘[1] 孙娜 王端生 杨雄飞 黄岗 CUI Yun-yue;GUAN Yi-hong;SUN Na;WANG Duan-sheng;YANG Xiong-fei;HUANG Gang(School of Science,Kunming University of Science and Technology;Kunming Julin Technology Company Limited,Yunnan 650000,China)

机构地区：[1]昆明理工大学理学院 [2]昆明聚林科技有限公司,云南昆明650000

出　　处：《软件导刊》2021年第2期63-67,共5页Software Guide

摘　　要：为了提高打叶复烤加工中长梗率和短梗率的检测效率,提出一种基于BP神经网络的长短梗率检测方法。该方法首先采用阈值分割法得到目标物的二值图像;然后利用旋转法求取图像中烟梗的最小外接矩形,从而识别出长梗和短梗;最后以烟梗的灰度级占比作为神经网络输入,以烟梗质量作为神经网络输出,建立烟梗质量拟合模型拟合长短梗的质量,从而得到烟梗的长梗率和短梗率。结果表明,该方法平均相对误差为3.91%,满足实际检测指标要求,具有一定应用价值。In order to improve the detection efficiency of long stem rate and short stem rate in leaf beating and redrying process,a detection method of long and short stem rate based on BP neural network is proposed.Firstly,the threshold segmentation method is used to obtain the binary image of the object.Then the minimum external rectangle of the tobacco stalk in the image is obtained by using the rotation method,and the long stalk and the short stalk are identified.Finally,the gray scale proportion of the tobacco stalk is used as the input of the neural network,and the quality of the tobacco stalk is used as the output of the neural network.The quality fitting model of the tobacco stalk is established to fit the quality of the long and short stalks,so as to obtain the long stem rate and short stem rate.The results show that the average relative error of this method is 3.91%,which meets the requirements of actual test index and has certain application value.

关 键 词：长短梗率 图像处理 BP神经网络 质量拟合 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]