基于PCA-SSA-BP神经网络的烟叶分级方法  

Tobacco Leaf Grading Method Based on PCA-SSA-BP Neural Network

作　　者：杨述斌[1,2] 董春林 王锋 周敏瑞 YANG Shu-bin;DONG Chun-lin;WANG Feng;ZHOU Min-rui(School of Electrical and Information Engineering,Wuhan Institute of Technology,Wuhan 430205,China;Hubei Provincial Key Laboratory of Intelligent Robot,Wuhan 430205,China)

机构地区：[1]武汉工程大学电气信息学院,武汉430205 [2]智能机器人湖北省重点实验室,武汉430205

出　　处：《自动化与仪表》2023年第2期70-75,共6页Automation & Instrumentation

基　　金：武汉工程大学研究生教育创新基金项目(CX2021072)。

摘　　要：基于机器视觉检测的烟叶分级方法,存在识别效率低、识别稳定性差等问题。针对这些问题,该文基于主成分分析(PCA)、麻雀搜索算法(SSA)、BP神经网络提出一种烟叶分级方法,首先对采集到的烟叶图像进行分析提取17个图像特征进行PCA降维处理,以消除冗余信息,结合烟叶分级标准并平衡烟叶分级的识别正确率和分级效率确定最佳降维的维数为七维;然后,将七维特征向量作为BP神经网络的输入,利用SSA对BP网络模型的权值和阈值进行优化,完成网络模型的训练。实验结果表明,PCA-SSA-BP模型的平均识别率达到96%以上,总运行时间为97.99 s,因此该方法能有效提高小样本烟叶分级的识别效率和识别稳定性。The tobacco leaf grading method based on machine vision inspection has problems such as low recognition efficiency and poor recognition stability. To address this problem,this paper proposes a tobacco leaf grading method based on PCA,a SSA,and a BP neural network. First,the collected tobacco leaf images were analyzed to extract 17 image features for PCA dimensionality reduction to eliminate cross-redundant information,combine the tobacco leaf grading criteria,and balance the correct recognition rate and grading efficiency of tobacco leaf grading to determine the optimal dimensionality reduction of seven dimensions. Next,the 7-dimensional feature vector is used as the input of the BP neural network,and the weights and thresholds of the BP network model are optimized using SSA to complete the training of the network model. The experimental results show that the average recognition rate of the PCASSA-BP model is over 96%,and the total running time is 97.99 s. Therefore,this method can effectively improve the recognition efficiency and stability of small-sample tobacco leaf grading.

关 键 词：烟叶分级 主成分分析法 麻雀搜索算法 BP神经网络 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]