Garbage classification method based on lightweight convolutional neural network

A convolutional neural network and garbage classification technology, applied in the field of garbage classification, can solve problems such as low model complexity and high classification accuracy at the same time, and achieve the effect of low complexity and high classification accuracy

Active Publication Date: 2020-05-08
QIQIHAR UNIVERSITY
View PDF5 Cites 13 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve the problem that the existing methods cannot have both low model complexity and high classification accuracy, and propose a garbage classification method based on lightweight convolutional neural network

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Garbage classification method based on lightweight convolutional neural network
  • Garbage classification method based on lightweight convolutional neural network
  • Garbage classification method based on lightweight convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0032] Specific implementation mode one: as figure 1 As shown, a kind of garbage classification method based on lightweight convolutional neural network described in this embodiment comprises the following steps:

[0033] Step 1. Obtain the original garbage image data set, and perform data enhancement on the acquired image to obtain the image data set after data enhancement;

[0034] Step 2, preprocessing the image data set after data enhancement to obtain the preprocessed image data set;

[0035]Step 3, constructing a convolutional neural network model, training the constructed convolutional neural network on the preprocessed image data set, and obtaining a trained convolutional neural network model;

[0036] The structure of the convolutional neural network model is: starting from the input end of the convolutional neural network model, the convolutional neural network model sequentially includes an input layer, a first sub-convolution unit, a second sub-convolution unit, a...

specific Embodiment approach 2

[0052] Specific embodiment 2: The difference between this embodiment and specific embodiment 1 is: in the step 1, data enhancement is performed on the acquired image to obtain a data-enhanced image data set. The specific process is as follows:

[0053] Carry out data enhancement on the acquired image, the method of data enhancement is: set the angle range of the random rotation of the image to 0°~20°, the offset coefficient of the image in the horizontal and vertical direction is 0.2, and the filling method is adjacent filling ;

[0054] After the data augmentation is completed, the image dataset after data augmentation is obtained.

[0055] The data set TrashNet used in the present invention only contains 2527 pieces of garbage images, including cardboard, glass, metal, paper, plastic, and 6 categories of garbage, with a total of 2527 pieces of RGB three-color images. Among them: 403 pieces of cardboard, 501 pieces of glass, 410 pieces of metal, 594 pieces of paper, 482 piec...

specific Embodiment approach 3

[0057] Embodiment 3: The difference between this embodiment and Embodiment 1 is that in the step 2, the image data set after data enhancement is preprocessed to obtain the preprocessed image data set. The specific process is as follows:

[0058] The size of each image in the image data set after data enhancement is uniformly scaled to the size of (224,224), and each value in the generated matrix is ​​multiplied by 1 / 255, so that each value is between 0 and 1, Obtain the preprocessed image dataset.

[0059] For data with relatively large values ​​or heterogeneous data (for example, one eigenvalue of the data is in the range of 0 to 1, and the other eigenvalue is in the range of 100 to 200), it is not appropriate to input it into the neural network, so Doing so may result in large gradient updates, which in turn will cause the network to fail to converge. In order to make the learning of the network easier, the input data should have the following characteristics: ①Small value:...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a garbage classification method based on a lightweight convolutional neural network, and belongs to the technical field of garbage classification. The method solves the problemthat an existing method cannot have low model complexity and high classification precision at the same time. According to the method, a feature extraction layer is divided into nine parts, wherein convolution of each part adopts a method of combining depth separable convolution and common convolution, convolution kernels alternately adopt 1 * 1 and 3 * 3 in size, and batch normalization processing is carried out on a convolution result of each time. Different from a common ReLU activation function and a Flatten connection layer, the model provided by the invention adopts Leaky ReLU as the activation function and a global average pooling layer as the connection layer. Experimental results show that after the network is trained and tested on a TrashNet data set, the accuracy of 93.02% is obtained, the classification precision is high, the complexity of the model is low, and the classification precision and the complexity of the model can be considered at the same time. The method can beapplied to intelligent garbage classification.

Description

technical field [0001] The invention belongs to the technical field of garbage classification, and in particular relates to a garbage classification method based on a lightweight convolutional neural network. Background technique [0002] In recent years, the rapid economic growth has promoted urbanization, but with the development of urbanization, the problem of environmental pollution has also become increasingly serious. Rapid industrial and economic development consumes a lot of resources and increases municipal solid waste. At present, due to the dumping of household garbage, more than 500 million square meters of land in China have been encroached on, and nearly two-thirds of large and medium-sized cities have been swallowed up by garbage. There is no doubt that recycling municipal solid waste is of great significance to alleviate environmental pollution and excessive consumption of resources. However, how to effectively carry out garbage recycling is a major problem...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06K9/62G06N3/04G06N3/08G06Q50/26
CPCG06N3/082G06Q50/26G06V20/10G06N3/045G06F18/2415
Inventor 石翠萍王涛李静辉靳展王天毅
Owner QIQIHAR UNIVERSITY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap