Deep-learning-based garbage identification evolutionary learning method, device, system and medium

A technology of deep learning and learning methods, applied in the field of garbage identification, can solve the problems of insufficient accuracy, reduce the workload of maintenance personnel, and identify the limited amount of garbage, so as to solve the problems of insufficient accuracy and quantity, human-computer interaction is helpful, Improving the effect of precision and breadth

Active Publication Date: 2019-02-26
GUANGZHOU UNIVERSITY
View PDF8 Cites 57 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The first object of the present invention is to provide a garbage recognition evolutionary learning method based on deep learning, which improves the current garbage recognition technology, and at the same time uses the data with poor recognition effect as the evolutionary power to update and improve the trained garbage recognition model in real time, Make the trained garbage recognition model have the

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
  • Deep-learning-based garbage identification evolutionary learning method, device, system and medium
  • Deep-learning-based garbage identification evolutionary learning method, device, system and medium
  • Deep-learning-based garbage identification evolutionary learning method, device, system and medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0058] Machine learning is widely used in the field of object recognition, but most of its data sets rely on manpower, resulting in the trained model is always flawed, requiring high-frequency maintenance work by maintenance personnel. Therefore, this embodiment provides a garbage recognition evolutionary learning system based on deep learning. The system considers that in the field of garbage recognition, the types of garbage are changing with each passing day, and the data features that need to be extracted when the computer recognizes the target are also changeable, so the evolutionary learning system is proposed. The mechanism uses the data with poor recognition effect as the evolutionary power to update and improve the computer training model in real time to adapt to the changeable characteristics of the recognized garbage, and greatly improve the accuracy and breadth of recognition on the basis of existing technologies.

[0059] Such as figure 1 and figure 2 As shown, ...

Embodiment 2

[0077] Such as Figure 9 As shown, this embodiment implements the data processing of the information processor in the above-mentioned embodiment 1 through the server, that is, the server is directly connected to the camera and the controller. This embodiment provides a garbage recognition evolutionary learning method based on deep learning. The method Applied to the server, including the following steps:

[0078] S901. Obtain garbage sample image data.

[0079] S902. Perform preprocessing on the garbage sample image data, specifically: demeaning, normalization, PCA and whitening processing on the garbage sample image data.

[0080] S903. Using the preprocessed garbage sample image data as an input parameter of the neural network, compare it with the trained garbage recognition model, and judge whether the recognition is successful according to the comparison result, specifically:

[0081] The preprocessed garbage sample image data is used as the input parameter of the neural...

Embodiment 3

[0089] Such as Figure 10 As shown, this embodiment provides a garbage recognition evolutionary learning device based on deep learning, which includes an acquisition module 1001, a preprocessing module 1002, a first recognition module 1003, a feedback module 1004, a second recognition module 1005, a first The update module 1006 and the second update module 1007, the specific functions of each module are as follows:

[0090] The acquiring module 1001 is configured to acquire garbage sample image data.

[0091] The preprocessing module 1002 is used for preprocessing the garbage sample image data, specifically: for removing the mean value, normalizing, PCA and whitening the garbage sample image data.

[0092] The first recognition module 1003 is used to compare the preprocessed garbage sample image data with the trained garbage recognition model as the input parameter of the neural network, and judge whether the recognition is successful according to the comparison result, speci...

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 identification evolutionary learning method, a device, a system and a medium based on depth learning. The method comprsies steps of acquiring garbage sample image data; Preprocessing the garbage sample image data; The pre-processed garbage sample image data is used as the input parameter of the neural network, and is compared with the trained garbage recognition model. According to the comparison result, whether the recognition is successful or not is judged. Feeding back the corresponding information of the identified garbage to the garbage sorting and placing mechanism; The image data of garbage sample which failed to be recognized is recognized again by ResNet algorithm, the image data of garbage sample which is successfully recognized by ResNet algorithm is marked, and the corresponding garbage information is fed back to the garbage sorting and putting mechanism, and the garbage recognition model is updated. The garbage sample image data of garbagesample which failed to be recognized is identified again by ResNet algorithm. The successful garbage sample images identified by ResNet algorithm are transmitted to users or maintainers for marking and updating the garbage identification model. The invention greatly reduces the workload of the maintenance personnel and realizes the accurate classification of a large amount of garbage.

Description

technical field [0001] The invention relates to a garbage identification method, in particular to a garbage identification evolutionary learning method, device, system and storage medium based on deep learning, belonging to the technical field of garbage identification. Background technique [0002] The origin of machine learning can be traced back a long time ago, and deep learning is also within the category of machine learning. As a new field in the field of machine learning, it appeared in 2006 and its development momentum has not diminished in recent years. The structure of deep learning is compared with the traditional shallow learning institutions. Compared with the shallow structure, the structure of deep learning is more complicated, and the number of training layers of the neural network involved is more, so it is more complex than the traditional shallow structure. The learning structure can learn more abstract features of the data, and it is more sensitive to the...

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
IPC IPC(8): G06K9/62B65F1/00B65F1/10
CPCB65F1/0033B65F1/10B65F2001/008G06F18/217G06F18/241G06F18/214Y02W30/10
Inventor 刘长红杨兴鑫张宏康李文杰朱亮宇钟志鹏程健翔范俊宇黄楠陈建堂彭绍湖
Owner GUANGZHOU 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