| Garbage classification is a very important research topic in modern times,and now it is mostly limited to manual classification.This article mainly studies how to use machines to classify garbage.With the enhancement of current camera methods and the enhancement of image classification technology,I consider using an image recognition technology to classify garbage images,and then approach the garbage classification.The image recognition methods used in the second stage mostly use the method of a single neural network.This paper shows that the combined model of neural network and support vector machine has good classification ability.The main research contents of the paper include the following:1.Collection,sorting and establishment of garbage image data.This article summarizes four categories and more than 40 small categories of garbage through the Shanghai Municipal Waste Disposal Regulations,and uses Google and other search engines to collect image data to build a garbage image library.2.Design and implementation of garbage classification identification model.In this paper,the image features are extracted through the transfer learning convolutional neural network,and the neural network is used to establish the model for comparison.Then,different combination models are designed.3.Comparative analysis of various combined models and single neural network models.The paper compares the advantages and disadvantages of various combined models,and finally concludes that the accuracy and training speed of some machine learning models are better than that of a single neural network model,which proves that the addition of some traditional machine learning models can improve a single neural network.This paper applies image recognition technology to garbage classification.A single neural network model and a combined model have been established.Through the comparison of the recognition effects and training time of these models,it can be concluded that the combined model of neural network and support vector machine is more suitable for use on garbage images,which will be used for real automatic garbage classification in the future.The research provides a reliable basis. |