Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Method for realizing automatic garbage classification

A technology of automatic classification and garbage classification, applied in the direction of neural learning methods, instruments, biological neural network models, etc., can solve the problems of high labor cost, low efficiency, easy to cause errors, etc., and achieve the effect of enhancing recognition ability

Pending Publication Date: 2020-07-24
成都禧来科技有限公司
View PDF12 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a method for automatic classification of garbage, which solves the problems of manual classification, low efficiency, high labor cost, and easy to cause errors in the existing garbage classification, and also solves the problem of automatic garbage classification There is no better and more perfect method, so that the model can be trained to obtain an accuracy rate that meets the actual use requirements

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
  • Method for realizing automatic garbage classification
  • Method for realizing automatic garbage classification
  • Method for realizing automatic garbage classification

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0044] A method for realizing automatic classification of garbage, including a model training process, a garbage classification method based on the model training process, such as figure 1 As shown, the model training process includes the following steps in sequence:

[0045] S1: Put the marked garbage classification training atlas into the model preprocessor for preprocessing. The preprocessing is to increase the recognition ability of the model (Robustness);

[0046] S2: Input the preprocessed RGB three-channel image into the deep neural network. Generally speaking, the image will be converted into an expression method;

[0047] S3: In the first convolutional layer (Convolution Layer) of the deep neural network, use Zero Padding to fill the gap between the filter size (filter) and the image size;

[0048] S4: In the second sampling layer of the deep neural network, feature extraction is performed on the features obtained after the convolutional layer is extracted, so as to ...

Embodiment 2

[0057] Such as figure 2 As shown, the garbage classification method based on the model training and identification process includes the following steps in sequence:

[0058] T1: At least based on the three usage characteristics of garbage types, size characteristics and user dumping habits, a two-class (Two class classification) recognition model is trained for each two types of garbage. In the garbage classification method based on the model training process, It is to classify the garbage categories to be divided into two categories (Two class classification) in turn, and finally divide the garbage into each category;

[0059] T2: Train several two-category recognition models according to requirements to assist garbage sorting devices. Generally speaking, garbage can be divided into dry garbage, wet garbage, recyclables and hazardous garbage according to current Chinese requirements. For this situation, it can be set as Three two-class (Two class classification) recognition...

Embodiment 3

[0064] This embodiment is a further description of embodiment 2, as figure 2 , image 3 As shown, the model training process also includes a model evolution step, and the model evolution step is set such that when the recognition model encounters a photo that cannot be processed, the garbage sorting device sends the picture to the cloud through the network card, and after manual labeling , re-input into the model training process as supplementary data for learning. The photos that cannot be processed include, but are not limited to, junk images that cannot be identified by the model, or incorrectly classified images submitted manually, or images that are recognized correctly but whose recognition confidence is lower than a preset certain threshold.

[0065] The garbage classification method based on the model training and recognition process uses the two-class recognition model to ensure that only one class of features is recognized for each classification, which greatly imp...

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 method for realizing automatic garbage classification, relates to the field of garbage classification, comprises a model training process and a garbage classification methodbased on the model training process, and uses a convolution model training process to enable a training model to have a function of identifying garbage classification characteristics. The convolutional model training process comprises the steps of labeling an image set, preprocessing, converting a mode, convolution, sampling, normalization, full connection, random inactivation, outputting a resultand outputting a training result which are performed in sequence; according to the garbage classification method, a plurality of two-class recognition models are formed by using a training model, garbage passes through the two-class recognition models in sequence for recognition after being photographed, and the garbage is classified according to recognition results of the two-class recognition models passing through the garbage in sequence.

Description

technical field [0001] The invention relates to the field of garbage classification, in particular to a method for realizing automatic garbage classification. Background technique [0002] Most of the existing garbage classification work relies on manual classification, especially in residential areas and other application scenarios. If computer vision is used, automatic garbage classification can be realized, and the classification speed and accuracy can be greatly improved, thereby reducing labor costs. , with the great breakthrough of convolutional neural network in computer vision, and adding activation functions such as linear rectification function (Rectified Linear Unit, ReLU), and random deactivation (dropout), using adaptive moment estimation to optimize the model ( The use of optimization methods such as Adam: adaptive moment estimation) enables convolutional neural networks to obtain extremely high accuracy in image recognition, and intelligent garbage classificat...

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): G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F18/2413G06F18/214
Inventor 唐国凯
Owner 成都禧来科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com