Research of Garbage Cleaning Robot Based on Deep Learning

With the development of neural networks and deep learning, computer vision recognition technology has been widely applied. At present, the labor cost in the field of social cleanliness is constantly increasing, especially in some harsh working environments, and timely completion of garbage cleaning and disposal has become a difficult task. Automated garbage cleaning through robots is the most effective way to solve this problem. This design is based on Raspberry Pi for programming, using image recognition technology to accurately identify the types and locations of garbage, in order to reach the designated location for cleaning. Deep learning systems mainly focus on paper clusters, randomly classify the content in the image library according to sample and test sets, and continuously train convolutional neural network models to improve their performance. The experiment shows that this model can effectively and accurately identify paper waste and complete the grabbing process.