基于YOLOv3与ResNet50算法的智能垃圾分类系统

Intelligent Garbage Classification System Based on YOLOv3 and ResNet50 Algorithm

垃圾分类对于我们日常生活来说意义重大,它不仅体现的是环境问题,更是资源的节约和绿色的生活方式,更加代表和体现了社会的文明水平,我国各地近些年也相继推行了垃圾分类的政策。2020年,新冠疫情席卷全球,对我们的日常生活造成了非常大的冲击,病毒与细菌的感染不只是通过呼吸,而且可以通过间接的接触来传染,这就要求了我们在生活中应该尽量避免去接触公共的设施和物品,而垃圾回收点恰恰是人员来往较多,很容易造成接触感染。对此,设计了一种疫情之下的垃圾智能分类与识别系统。设计的总体思想是采用Arm公司的EAIDK-310开发板,同时配备罗技C270摄像头采集装置对实际生活中的常见垃圾进行识别与分类,然后通过摄像头采集装置采集到垃圾的图像,结合深度学习模型自动检测、识别垃圾的类别,从而可以根据类别直接进行垃圾的分类丢弃。模型是基于YOLOv3在复杂环境下垃圾检测的基础上完成,构建ResNet50网络,对垃圾的类别进行训练并识别。通过该智能垃圾分类系统,可以提高投放垃圾效率,大大减少人员在垃圾站附近的聚集并做到无接触的垃圾丢弃,从而能够有效避免因为接触而造成的交叉感染的风险,符合当下疫情期间的社会需求。

Garbage classification is of great significance to our daily life.It not only reflects the environmental problems,but also the conservation of resources and the green way of life.It also represents and reflects the level of civilization of the society.In recent years,the policy of garbage classification has been successively implemented in all parts of our country.In 2020,the new crown worldwide epidemic,to our daily life caused a very big impact,viral and bacterial infections is not only through the breath,and can be transmitted through indirect contact,the requires that in our life,we should try to avoid to contact the public facilities and items,and garbage collection point on the exchange of personnel is more,it is easy to cause contact infection.Therefore,an intelligent waste classification and identification system under the epidemic situation is designed.Design of the overall idea is using the Arm of the company's EAIDK-310 development board,at the same time equipped with logitech C270 camera collection equipment to the real life of common in recognition and classification of garbage,and then through the camera image sampling device to waste,combined with deep learning model automatic detection and identification of rubbish categories.Thus,garbage can be classified and discarded directly according to categories.The model is based on the completion of YOLOv3 garbage detection in a complex environment.ResNet50 network is built to train and identify the categories of garbage.The intelligent garbage sorting system can improve the efficiency of garbage delivery,greatly reduce the gathering of people near garbage stations and achieve non-contact garbage disposal,so as to effectively avoid the risk of cross-infection caused by contact and meet the social needs during the current epidemic.