改进型MobileNetv2模型在智能垃圾分类系统的应用研究

Application of improved MobileNetv2 model in intelligent garbage sorting system

基于图像处理技术设计了智能垃圾分类系统。模型利用MobileNetv2网络具有的优异特征提取和分类能力实现图像特征提取,通过迁移学习方法,将提取到的特征迁移到目标网络上,防止训练过程中因数据增强和数据扩充产生的过拟合现象,提高了模型准确率,实验表明该垃圾分类模型具有更高的垃圾处理效率[1]。

This paper designs an intelligent garbage sorting system based on image processing technology.The model uses the excellent feature extraction and classification ability of MobileNetv2 network to realize image feature extraction.Through the transfer learning method,the extracted features are transferred to the target network to prevent the over fitting phenomenon caused by data enhancement and data expansion in the training process,and the accuracy of the model is improved.Experiments show that the garbage sorting model has high garbage treatment efficiency.