摘要
为了推进我国的环境保护，我国正大力推广垃圾分类。由于垃圾品类复杂繁多，且分类垃圾桶设计简单，大部分仅为贴有分类标签的普通垃圾桶，需要人为识别垃圾种类，因此造成垃圾分类工作效率偏低。本文提出了基于深度学习算法，将图像识别、迁移学习和嵌入式开发应用到垃圾分类当中，设计了一种智能分类垃圾桶，实现了垃圾识别、自动分拣的功能。符合当下社会环境保护和发展的形势，同时为智能制造市场填补上重要的一环，具有较强的创造性意义。
关键词:垃圾分类；深度学习；图像识别；嵌入式开发
Abstract
In order to promote China's environmental protection, China is vigorously promoting garbage classification. Due to the complexity and variety of garbage categories and the simple design of classified garbage cans, most of them are only ordinary garbage cans with classification labels, which require artificial identification of garbage types, resulting in low efficiency of garbage classification. Based on the deep learning algorithm, this paper proposes to apply image recognition, transfer learning and embedded development to garbage classification, and designs an intelligent garbage sorting bin to realize the functions of garbage recognition and automatic sorting. In line with the current situation of social environmental protection and development, and at the same time fill an important part of the intelligent manufacturing market, it has strong creative significance.
Keywords: garbage classification; deep learning; image recognition; Embedded development
目 录
引言 2
1 智能垃圾分类现状及意义 2
1.1智能垃圾分类现状 2
1.2智能垃圾分类意义 2
2系统设计 3
2.1图像采集 3
2.2压力检测模块 4
2.3语音报警模块 4
3机械设计 4
4软件设计 4
4.1卷积神经网络 4
4.2迁移学习 5
4.3模型选择与构建 5
5结果分析与性能评估 6
6结语 6
