基于改进卷积神经网络的红枣缺陷识别  

Research on jujube defect recognition method based on improved convolution neural network

作　　者：张忠志 薛欢庆[2] 范广玲[3] ZHANG Zhong-zhi;XUE Huan-qing;FAN Guang-ling(Geely University of China,Chengdu,Sichuan 641423,China;Daqing Normal University,Daqing,Heilongjiang 163712,China;Northeast Petroleum University,Daqing,Heilongjiang 163311,China)

机构地区：[1]吉利学院,四川成都641423 [2]大庆师范学院,黑龙江大庆163712 [3]东北石油大学,黑龙江大庆163311

出　　处：《食品与机械》2021年第8期158-162,192,共6页Food and Machinery

基　　金：黑龙江省教育科学规划“十四五”规划重点课题(编号:JJB1421006)。

摘　　要：目的:建立一种基于改进的卷积神经网络的红枣缺陷自动识别方法。方法:采用双分支卷积神经网络结构,分支1结合迁移学习策略进行预训练,分支2基于轻量级网络融合特征图提取红枣图像中的特征信息。通过对比实验验证了该方法的优越性。结果:与改进前相比,改进后的缺陷识别方法优化了卷积神经网络的结构,检测准确率进一步提高,从96.02%提高到99.50%。结论:该方法提高了网络学习速度和收敛速度,具有较好的分类识别效果。Objective:An automatic identification method of jujube defects based on improved convolution neural network was established.Methods:Using the dual branch convolution neural network structure,branch 1 combined with the transfer learning strategy for pre training,analysis 2 extracted the feature information from the jujube image based on the lightweight network fusion feature map.The superiority of this method was verified by comparative experiments.Results:Compared with the improvement before,the improved defect recognition method optimized the structure of the convolutional neural network,and the detection accuracy was further improved,from 96.02%to 99.50%.Conclusion:This method improved the network learning speed and convergence speed,and had good classification and recognition effect.

关 键 词：红枣缺陷 卷积神经网络 自动识别 迁移学习策略 轻量级网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP391.41[自动化与计算机技术—控制科学与工程] S665.1[农业科学—果树学]