Solid waste identification data set construction system based on convolutional neural network

Abstract

A solid waste identification data set construction system based on a convolutional neural network comprises a data set acquisition module which is used for acquiring a solid waste image set including color information and spatial information, acquiring a solid waste image set including height information, and performing synchronous matching processing on the two types of image sets, thereby obtaining a solid waste image set including color information, spatial information and height information; and the data set labeling module is used for generating a data set required by solid waste identification. Through the data set acquisition module and the data set labeling module, the waste recognition data set can be quickly and effectively constructed and expanded, so that training of the convolutional neural network model is facilitated, and the accuracy of solid waste recognition is further improved.

Description

technical field [0001] The invention relates to the technical field of data set construction, in particular to a solid waste identification data set construction system based on a convolutional neural network. Background technique [0002] With the development of urbanization, how to deal with hundreds of millions of solid waste has become a major problem. As a cutting-edge visual inspection technology, convolutional neural network can effectively detect and classify solid waste. In order to improve the detection performance of the network, it is essential to quickly obtain a large number of training data sets. However, in order to obtain high-quality data sets, the main problems are as follows: [0003] Low-quality images not only affect the training of the model, but also affect the subsequent annotation and detection of the model. [0004] The acquired images often need to be finely labeled, which is not only cumbersome, but also requires a lot of manpower. [0005] I...

AI Technical Summary

Problems solved by technology

However, the image annotation needs to mark the outline of the object to be recognized on each image. This work is not only inefficient, but also the long-term data set annotation affects the training of the convolutional neural network model, thus slowing down the production process.

Images