GA-BP神经网络对片烟结构的预测研究

doi:10.16180/j.cnki.issn1007-7820.2022.06.006

摘要/Abstract

摘要：

针对烟叶复烤厂打叶过程中片烟结构难以预测的问题,文中提出了一种基于MATLAB图像处理的GA-BP神经网络预测模型。对于烟叶分类问题,基于获取的烟叶图片,利用MATLAB软件对图片做预处理,提取衡量片烟结构的主要特征变量,并利用行业标准与聚类分析算法对数据进行分类。通过统计学的标准数学方法,构建了遗传算法优化的BP神经网络预测模型对主要影响参数进行预测优化。研究结果表明,文中所提方法预测精度较高,预测极差均小于0.059,可有效解决打叶过程中片烟的预测问题。

关键词: MATLAB, 图像处理, 聚类分析, 留一交叉验证, BP神经网络, 遗传算法, GA-BP神经网络, 二值化

Abstract:

In view of the problem that it is difficult to predict the structure of tobacco during the threshing process of redrying plant, a GA-BP neural network prediction model based on MATLAB image processing is proposed. For the classification of tobacco leaf, the obtained tobacco leaf images are preprocessed using of the MATLAB software. Then, the main characteristic variables that measure the structure of the tobacco are extracted, and industry standards and cluster analysis algorithms are used to classify the data. Through the standard mathematical method of statistics, the BP neural network prediction model optimized by genetic algorithm is constructed to predict and optimize the main influencing parameters. The research results show that the method proposed in this study has high prediction accuracy, and the prediction range is less than 0.059, which indicates that the method can effectively solve the problem of prediction of slices of tobacco in the process of threshing.

Key words: MATLAB, image processing, cluster analysis, leave-one-out-cross-validation, BP neural network, genetic algorithm, GA-BP neural network, binarization

中图分类号:

TP29

张崇崇,黄亚宇. GA-BP神经网络对片烟结构的预测研究[J]. 电子科技, 2022, 35(6): 35-42.

ZHANG Chongchong,HUANG Yayu. A GA-BP Neural Network for Predicting the Structure of Leaf Tobacco[J]. Electronic Science and Technology, 2022, 35(6): 35-42.

图/表 9

图1

图2

图3

图4

表1

图5

图6

图7

表2

参考文献 19

[1]	杨丹, 赵海滨, 龙哲, 等. MATLAB图像处理实例详解[M]. 北京: 清华大学出版社, 2013.
	Yang Dan, Zhao Haibin, Long Zhe, et al. Detailed explanation of MATLAB image processing example[M]. Beijing: Tsinghua University Press, 2013.
[2]	张泽旭. 神经网络控制与MATLAB仿真[M]. 哈尔滨: 哈尔滨工业大学出版社, 2011.
	Zhang Zexu. Neural network control and MATLAB simulation[M]. Harbin: Harbin Institute of Technology Press, 2011.
[3]	MATLAB中文论坛. MATLAB神经网络30个案例分析[M]. 北京: 北京航空航天大学出版社, 2010.
	MATLAB Chinese Forum. Analysis of 30 cases of MATLAB neural network[M]. Beijing: Beihang University Press, 2010.
[4]	赖安卿, 刘云鹏, 付尧明, 等. 振荡燃烧火焰图像处理[J]. 燃烧科学与技术, 2020, 26(1):10-17.
	Lai Anqing, Liu Yunpeng, Fu Yaoming, et al. Image processing of combustion oscillating flame[J]. Journal of Combustion Science and Technology, 2020, 26(1):10-17.
[5]	杨玮, 兰红, 李民赞, 等. 基于图像处理和GBRT模型的表土层土壤容重预测[J]. 农业机械学报, 2020, 51(9):193-200.
	Yang Wei, Lan Hong, Li Minzan, et al. Prediction of top soil layer bulk density based on image processing and gradient boosting regression tree model[J]. Transactions of the Chinese Society for Agricultural Machinery, 2020, 51(9):193-200.
[6]	刘慧力, 贾洪雷, 王刚, 等. 基于深度学习与图像处理的玉米茎秆识别方法与试验[J]. 农业机械学报, 2020, 51(4):207-215.
	Liu Huili, Jia Honglei, Wang Gang, et al. Method and experiment of maize(Zea Mays L.) stems recognition based on deep learning and image processing[J]. Transactions of the Chinese Society for Agricultural Machinery, 2020, 51(4):207-215.
[7]	郁浩, 张荣福, 程金光, 等. 基于图像处理的钢包标牌识别方法[J]. 电子科技, 2016, 29(1):94-97.
	Yu Hao, Zhang Rongfu, Cheng Jinguang, et al. Ladle license plate recognition based on image processing[J]. Electronic Science and Technology, 2016, 29(1):94-97.
[8]	Gonzalez R C, Woods R E, Eddins S L. 数字图像处理[M]. 阮秋琦,译. 北京: 电子工业出版社, 2005.
	Gonzalez R C, Woods R E, Eddins S L. Digital image processing[M].Translated by Ruan Qiuqi. Beijing: Electronic Industry Press, 2005.
[9]	陈朋. 烟叶纹理表面的视觉检测技术研究与应用[D]. 贵州: 贵州大学, 2017.
	Chen Peng. Research and application of visual inspection technology for texture of tabacco surface[D]. Guizhou: Guizhou University, 2017.
[10]	李伟, 张长胜, 冯广, 等. 基于图像处理的指针式表盘自动读取系统设计[J]. 电子科技, 2018, 31(6):24-26.
	Li Wei, Zhang Changsheng, Feng Guang, et al. The design of automatic system read pointer based on image processing[J]. Electronic Science and Technology, 2018, 31(6):24-26.
[11]	蒙西, 乔俊飞, 韩红桂. 基于类脑模块化神经网络的污水处理过程关键出水参数软测量[J]. 自动化学报, 2019, 45(5):906-919.
	Meng Xi, Qiao Junfei, Han Honggui. Soft measurement of key effluent parameters in wastewater treatment process using brain-like modular neural network[J]. Acta Automatica Sinica, 2019, 45(5):906-919.
[12]	辛焕平. MATLAB R2017a模式识别与智能计算[M]. 北京: 电子工业出版社, 2018.
	Xin Huanping. MATLAB R2017a pattern recognition and intelligent computing[M]. Beijing: Electronic Industry Press,2018.
[13]	Zhao Y F, Fu Z G. On-Line monitoring of coal calorific value based on GA-BP[J]. Advances in Energy and Power Engineering, 2016, 4(4):103-110. doi: 10.12677/AEPE.2016.44014
[14]	吕升, 郭迎清, 孙浩. 基于GA-AANN神经网络的SDQ算法的航空发动机传感器数据预处理[J]. 推进技术, 2018, 39(5):1142-1150.
	Lü Sheng, Guo Yingqing, Sun Hao. Aero-engine sensor data preprocessing based on GA-AANN neural network SDQ algorithm[J]. Journal of Propulsion Technology, 2018, 39(5):1142-1150.
[15]	许永, 李超, 杨乾栩, 等. 基于BP神经网络建立初烤烟叶主流烟气中氢氰酸释放量的预测模型[J]. 云南农业大学学报(自然科学版), 2018, 33(1):72-78.
	Xu Yong, Li Chao, Yang Qianxu, et al. Establishment of Predicting model of hydrogen cyanide contents in flue-cured tobacco mainstream smoke based on BP neural network[J]. Journal of Yunnan Agricultural University(Natural Sciences), 2018, 33(1):72-78.
[16]	秦兴, 高晓琪, 陈滨. 基于压缩卷积神经网络的图像超分辨率算法[J]. 电子科技, 2020, 33(5):1-8.
	Qin Xing, Gao Xiaoqi, Chen Bin. Image super-resolution algorithm based on squeezenet convolution neural network[J]. Electronic Science and Technology, 2020, 33(5):1-8.
[17]	都月, 孟晓辰, 祝连庆. 遗传算法和神经网络的重叠光谱解析[J]. 光谱学与光谱分析, 2020, 40(7):2066-2072.
	Du Yue, Meng Xiaochen, Zhu Lianqing. Overlapping spectrum analysis based on genetic algorithm and BP neural network[J]. Spectroscopy and Spectral Analysis, 2020, 40(7):2066-2072.
[18]	顾迎捷, 桂小林, 李德福, 等. 基于神经网络的机器阅读理解综述[J]. 软件学报, 2020, 31(7):2095-2126.
	Gu Yingjie, Gui Xiaolin, Li Defu, et al. Survey of machine reading comprehension based on neural network[J]. Journal of Software, 2020, 31(7):2095-2126.
[19]	雷宇, 李孝安. 一种基于GA-BP学习的神经网络推理机制的研究[J]. 电子设计工程, 2013, 21(12):34-36.
	Lei Yu, Li Xiaoan. Study on GA-BP-based neural network inference mechanism[J]. Electronic Design Engineering, 2013, 21(12):34-36.

		训练样本	验证集	测试集	极差
BP	所有特征	98.294%	97.824%	97.126%	59.000×10^-3
BP	理想特征	98.897%	98.555%	98.973%	17.461×10^-3
GA-BP	所有特征	98.781%	98.346%	98.753%	4.060×10^-3
GA-BP	理想特征	99.851%	99.843%	99.836%	4.733×10^-3