基于计算机视觉的垃圾分类回收平台

Garbage Classification and Recycle Platform Based on Computer Vision

随着社会的快速发展,垃圾存量急剧上升,“垃圾围城”、“垃圾围村”正日益成为困扰中国各个城市、乡村的难解之题。垃圾分类是社会进步和生态文明的标志,是人人均可参与其中来保护环境和改善环境的方式。为有效推进垃圾分类落地和提高城市管理水平,本文设计了一款基于计算机视觉目标检测算法的垃圾分类回收平台,提高分类效率。本系统提供拍照识别、图片识别等智能化功能对垃圾特征进行识别分类,提供数据集管理模块扩充垃圾类别、持续提高模型准确率,同时还提供智能图像识别API接口,如在数据预处理上采用EnlightenGAN网络进行弱光图像增强,再进行检测,供垃圾分类系统对接使用,充分将智能化技术引入生活垃圾分类中。

With the rapid development of society, the stock of garbage has risen sharply. “Garbage encircling cities” and “garbage encircling villages” are increasingly becoming difficult problems to solve in various cities and villages in China. Garbage classification is a symbol of social progress and ecological civilization, and it is a way that everyone can participate in protecting and improving the environment. In order to effectively promote the landing of garbage classification and improve the level of urban management, this paper designed a garbage classification recognition platform based on computer vision target detection algorithm to improve the classification efficiency. This system provides intelligent functions such as photo recognition and image recognition, which can identify characteristics of garbage classification to provide the data set management module, continuous expansion of junk category to improve model accuracy, and also provide an intelligent image recognition API interface, e.g. EnlightenGAN network is adopted to improve the weak light environment by using image enhancement during data preprocessing and detect again for the docking use of waste classification system, fully introducing intelligent technology into the household waste classification.