CN110606292A

CN110606292A - Automatic classification garbage can based on artificial intelligence and classification method

Info

Publication number: CN110606292A
Application number: CN201910877344.3A
Authority: CN
Inventors: 蔡亦圣
Original assignee: Individual
Current assignee: Individual
Priority date: 2019-09-17
Filing date: 2019-09-17
Publication date: 2019-12-24

Abstract

The invention provides an automatic classification garbage can based on artificial intelligence, which comprises a plurality of sub garbage cans and identification equipment; the sub-garbage can comprises a can body with an opening at the upper end, a can cover covering the opening of the can body, and a driving device connected between the can body and the can cover, wherein the driving device is used for controlling the opening or closing of the can cover; the driving device is connected to the identification device; the identification equipment is used for acquiring image information of the garbage through the camera, analyzing the image information, acquiring classification categories of the garbage, and controlling corresponding driving devices to open corresponding barrel covers according to the classification categories of the garbage. This garbage bin can carry out automatic classification to rubbish, and convenient to use overcomes traditional garbage bin and need carry out the defect of classifying through the manual work.

Description

Automatic classification garbage can based on artificial intelligence and classification method

Technical Field

The invention belongs to the technical field of garbage cans, and particularly relates to an automatic classification garbage can based on artificial intelligence and a classification method.

Background

Along with the development of society, the requirements of people on living environment are higher and higher, more and more garbage is generated in life, and the garbage disposal is concerned by people more and more. Conventional waste classification is largely divided into two categories, one is recyclable waste and the other is non-recyclable waste. The recyclable garbage can be provided for people to reuse, such as paper, aluminum sheets, plastics and the like, so that the environment is protected, the waste of resources is reduced, and the utilization rate of the resources is improved. The non-recyclable garbage refers to garbage except recyclable garbage, and common garbage which is easy to decompose under natural conditions is available, such as peels, vegetable leaves, leftovers, flowers, grass, branches and leaves, and the like; and the garbage is harmful, polluted and incapable of being decomposed and reproduced for the second time, and belongs to the unrecoverable garbage. The existing garbage can is generally only marked with recoverable garbage and non-recoverable garbage, and a lot of people often throw the garbage into the garbage can directly without intending to mark on the garbage can in the actual use process of people, so that the recoverable garbage and the non-recoverable garbage are mixed together. Therefore, when the garbage is classified, the garbage is required to be manually classified and then thrown into the corresponding garbage can, which is not only troublesome, but also causes certain adverse effects on garbage recycling treatment.

Disclosure of Invention

Aiming at the defects in the prior art, the invention provides an automatic classification garbage can based on artificial intelligence and a classification method, which can automatically classify garbage, are convenient to use and overcome the defect that the conventional garbage can needs to be classified manually.

In a first aspect, the automatic classification garbage can based on artificial intelligence,

the garbage can comprises a plurality of sub garbage cans and identification equipment;

the sub-garbage can comprises a can body with an opening at the upper end, a can cover covering the opening of the can body, and a driving device connected between the can body and the can cover, wherein the driving device is used for controlling the opening or closing of the can cover; the driving device is connected to the identification device;

the identification equipment is used for acquiring image information of the garbage through the camera, analyzing the image information, acquiring classification categories of the garbage, and controlling corresponding driving devices to open corresponding barrel covers according to the classification categories of the garbage.

Preferably, the categories of the sub-trash can include wet trash, dry trash, recyclable trash, and harmful trash.

In a second aspect, a classification method for automatically classifying trash cans based on artificial intelligence, which is operated on the identification device in the trash can of the first aspect, includes the following steps:

establishing a plurality of classification categories;

respectively collecting image information of various wastes under various classification categories;

selecting a visual recognition model on a preset AI open platform, and adding the image information of the garbage and the corresponding classification category in a data set of the visual recognition model;

training the visual recognition model added with the data set to obtain a trained visual recognition model;

acquiring image information of the garbage to be classified through a camera, inputting the image information into a trained visual recognition model, and recognizing classification categories of the garbage to be classified;

and controlling the corresponding driving device to open the corresponding barrel cover according to the classification category of the garbage to be classified.

Preferably, the acquiring image information of various types of spam under each classification category specifically includes:

reading pre-stored image information of the garbage under each classification category;

or shooting image information of the garbage under each classification category through a camera.

Preferably, after the trained visual recognition model is obtained, before the image information of the garbage to be classified is collected by the camera, the method further includes:

the method comprises the steps of collecting voice signals input by a user, analyzing the voice signals to obtain classification categories of garbage to be classified, and controlling corresponding driving devices to open corresponding barrel covers according to the classification categories of the garbage to be classified.

Preferably, the analyzing the voice signal to obtain the classification category of the garbage to be classified, and controlling the corresponding driving device to open the corresponding barrel cover according to the classification category of the garbage to be classified specifically includes:

constructing a word stock, wherein the word groups in the word stock comprise all classification categories and various garbage names under each classification category;

converting the voice signal into a text signal, and searching whether a word group in the word stock exists in the text signal;

if the searched phrase is the classification category, controlling a driving device corresponding to the classification category to open the barrel cover;

if the searched phrase is the garbage name, controlling a driving device of the classification category corresponding to the garbage name to open the barrel cover.

converting the voice signal into a text signal, and extracting a subject, a predicate and an object of the text signal according to a preset grammar rule;

judging to obtain the classification category of the garbage to be classified according to the subject, the predicate and the object;

and controlling a driving device corresponding to the classification category to open the barrel cover.

Preferably, the training of the visual recognition model after the data set is added specifically includes:

identifying the image information added with the garbage to obtain the outline of each object in the image information;

segmenting according to the outline of each object in the image information to obtain sub-images of each object;

extracting features of an object in the sub-image to obtain the dimension of the object, wherein the dimension comprises color, size and shape;

and training the visual recognition model according to the dimensions of each sub-image and the object thereof.

Preferably, after the method controls the corresponding driving device to open the corresponding barrel cover according to the classification category of the garbage to be classified, the method further includes:

if the classification type identification of the garbage to be classified in the visual identification model fails, generating failure reminding information containing the reason of the failure identification to remind a user, and starting a timer to time;

and when the timing time of the timer is up, detecting that the user does not take the garbage to be classified which is failed to be identified, and generating prompting reminding information to remind the user.

Preferably, the identifying the reason for failure includes:

when detecting that the image information of the garbage to be classified has the subimages of a plurality of objects, defining the reason of failure in identification as that the garbage contains a plurality of objects;

when an object in the image information of the garbage to be classified is identified, but the category of the object cannot be identified, defining the reason of failure identification as incomplete garbage structure;

and when the object in the image information of the garbage to be classified cannot be identified, defining the reason of failure in identification as no garbage exists.

According to the technical scheme, the automatic garbage classification can based on artificial intelligence provided by the invention has the advantages that the classification type of garbage is identified by the identification equipment, the opening of the can cover of the sub-garbage can corresponding to the classification type is controlled, a user can throw the garbage into the corresponding sub-garbage can, the garbage can automatically classify the garbage, the use is convenient, and the defect that the conventional garbage can needs to be classified manually is overcome.

The classification method of the automatic classification garbage can based on artificial intelligence provided by the invention comprises the steps of training image information of various garbage under each classification category, identifying the classification category of the garbage to be classified through a trained visual identification model, controlling the opening of a can cover of a sub-garbage can corresponding to the classification category, throwing the garbage into the corresponding sub-garbage can by a user, automatically classifying the garbage by the garbage can, being convenient to use, and overcoming the defect that the conventional garbage can needs to be classified manually.

Drawings

In order to more clearly illustrate the detailed description of the invention or the technical solutions in the prior art, the drawings that are needed in the detailed description of the invention or the prior art will be briefly described below. Throughout the drawings, like elements or portions are generally identified by like reference numerals. In the drawings, elements or portions are not necessarily drawn to scale.

Fig. 1 is a block diagram of modules of an automatic sorting trash can according to an embodiment.

Fig. 2 is a flowchart of the classification method according to the second embodiment.

Fig. 3 is a flowchart of a speech recognition method according to the second embodiment.

Fig. 4 is a flowchart of another speech recognition method according to the second embodiment.

Fig. 5 is a flowchart of a method for training a visual recognition model according to the second embodiment.

Detailed Description

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and therefore are only examples, and the protection scope of the present invention is not limited thereby. It is to be noted that, unless otherwise specified, technical or scientific terms used herein shall have the ordinary meaning as understood by those skilled in the art to which the invention pertains.

It will be understood that the terms "comprises" and/or "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.

It is also to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in the specification of the present invention and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

As used in this specification and the appended claims, the term "if" may be interpreted contextually as "when", "upon" or "in response to a determination" or "in response to a detection". Similarly, the phrase "if it is determined" or "if a [ described condition or event ] is detected" may be interpreted contextually to mean "upon determining" or "in response to determining" or "upon detecting [ described condition or event ]" or "in response to detecting [ described condition or event ]".

The first embodiment is as follows:

an automatic classification garbage can based on artificial intelligence is disclosed, referring to figure 1,

specifically, the sub-garbage can be automatically controlled to be opened by the driving device, and when the can cover is opened, a user throws garbage into the can body. The driving device only needs to be installed on the sub-garbage can, and the can cover can be controlled to be automatically opened. Wherein the classification categories of the sub-trashcans include wet trash, dry trash, recyclable trash, and hazardous trash. Since the classification of garbage is different in each place, the number and the classification of the sub-garbage cans can be specifically set according to a specific city.

Specifically, the recognition device can be provided with a recognition area, a camera is arranged above the recognition area, and when a user places garbage in the recognition area, the camera shoots image information of the garbage. For example, when the garbage put into the identification area by the user is unused toilet paper, the identification device identifies that the garbage is recoverable garbage, and the driving device in the sub-garbage can for recovering the garbage is controlled to open the can cover.

This automatic classification garbage bin, the categorised classification of rubbish is discerned to identification equipment to the bung that the categorised classification of control corresponds sub-garbage bin is opened, and the user just can throw rubbish into the sub-garbage bin that corresponds, and this garbage bin can carry out automatic classification to rubbish, and convenient to use overcomes traditional garbage bin and need carry out the defect of classifying through the manual work.

Example two:

a classification method for automatically classifying trash cans based on artificial intelligence, which is operated on an identification device in the trash can in the first embodiment, referring to fig. 2, and comprises the following steps:

s1: establishing a plurality of classification categories;

s2: respectively collecting image information of various wastes under various classification categories;

specifically, when image information of garbage is collected, image information of each side of the garbage can be collected, for example, when image information of a cup is collected, a top view, a bottom view and a side view of the cup are collected. It is also possible to capture a picture of a partial piece of the cup, for example, if the cup is broken, image information of the complete cup cannot be captured when the cup is thrown away, so in order to improve the recognition accuracy, image information of individual parts in the object is also captured.

When the image information is collected, the image information is collected according to each classification category, for example, the image information of dry garbage is collected firstly, then the image information of wet garbage is collected, then the image information of recoverable garbage is collected, and finally the image information of harmful garbage is collected.

S3: selecting a visual recognition model on a preset AI open platform, and adding the image information of the garbage and the corresponding classification category in a data set of the visual recognition model;

specifically, the AI open platform can be a K-12 open platform. After the user logs in the AI open platform, the corresponding visual identification model can be selected in the models preset in the AI open platform according to the requirement of the user. The user can also add a new model in the AI open platform according to the self requirement, and in the process of adding the new model, the user inputs the name of the model in addition to the new model, and determines the type of the model. The classification category of the garbage may be added in the form of a tag.

S4: training the visual recognition model added with the data set to obtain a trained visual recognition model;

in particular, the trained visual recognition model can also be tested. After the test is finished, when the visual recognition model needs to be actually applied, a Python programming tool can be used for obtaining the url link of the application interface, and then the api link is modified into a trained and obtained api link.

S5: acquiring image information of the garbage to be classified through a camera, inputting the image information into a trained visual recognition model, and recognizing classification categories of the garbage to be classified;

s6: and controlling the corresponding driving device to open the corresponding barrel cover according to the classification category of the garbage to be classified.

The classification method comprises the steps of firstly training image information of various kinds of garbage under each classification category, identifying the classification category of the garbage to be classified through a trained visual recognition model, controlling a barrel cover of a sub-garbage barrel corresponding to the classification category to be opened, throwing the garbage into the corresponding sub-garbage barrel by a user, automatically classifying the garbage by the garbage barrel, being convenient to use, and overcoming the defect that the conventional garbage barrel needs to be classified manually.

Specifically, when image information of spam is collected, pre-stored image information of spam, such as pictures from websites or pictures obtained by historical shooting, can be read. The images of the garbage can be shot in real time through the camera, in order to guarantee the accuracy of identification, 50% of image information can be acquired by adopting the two methods, and 50-100 images can be acquired by garbage of each classification type.

Specifically, the method can also perform garbage classification according to the voice signal input by the user, so that the user can control the opening of the cover of the corresponding sub-garbage can through voice. Specifically, the method recognizes the voice information of the user in the following two ways.

1) Referring to fig. 3, analyzing the voice signal to obtain classification categories of the garbage to be classified, and controlling the corresponding driving device to open the corresponding barrel cover according to the classification categories of the garbage to be classified specifically includes:

s11: constructing a word stock, wherein the word groups in the word stock comprise all classification categories and various garbage names under each classification category;

specifically, the constructed lexicon includes classification categories (such as dry garbage, wet garbage, recyclable garbage, harmful garbage and the like) and various garbage names (such as paper towels, batteries, leftovers, glass, plastics, old clothes and the like), so that the classification categories can be identified through keywords.

S12: converting the voice signal into a text signal, and searching whether a word group in the word stock exists in the text signal;

s13: if the searched phrase is the classification category, controlling a driving device corresponding to the classification category to open the barrel cover;

specifically, if the user inputs 'i want to throw dry garbage', the method detects that the converted text information contains the keyword 'dry garbage', so that the barrel cover corresponding to the sub-garbage barrel corresponding to the dry garbage is opened.

S14: if the searched phrase is the garbage name, controlling a driving device of the classification category corresponding to the garbage name to open the barrel cover.

Specifically, if the user inputs 'i want to throw away the glass cup', the method detects that the converted text information contains the keyword 'glass', so that the cover corresponding to the sub-trash can in the classification type (recyclable trash) corresponding to the glass is opened.

2) Referring to fig. 4, analyzing the voice signal to obtain the classification category of the garbage to be classified, and controlling the corresponding driving device to open the corresponding barrel cover according to the classification category of the garbage to be classified specifically includes:

s21: converting the voice signal into a text signal, and extracting a subject, a predicate and an object of the text signal according to a preset grammar rule;

s22: judging to obtain the classification category of the garbage to be classified according to the subject, the predicate and the object;

s23: and controlling a driving device corresponding to the classification category to open the barrel cover.

Specifically, the method identifies the subject, the predicate and the object in the converted text signal, analyzes the classification type of the garbage to be classified through the subject, the predicate and the object, for example, a user inputs "i want to throw away the glass water cup", the subject is "i", the predicate is "throw away", and the object is "glass water cup", the method analyzes that the user wants to throw away the garbage, the thrown away garbage is a glass water cup, and therefore the barrel cover corresponding to the sub-garbage barrel in the classification type (recyclable garbage) corresponding to the glass is opened.

Referring to fig. 5, the training of the visual recognition model after the data set is added specifically includes:

s31: identifying the image information added with the garbage to obtain the outline of each object in the image information;

s32: segmenting according to the outline of each object in the image information to obtain sub-images of each object;

s33: extracting features of an object in the sub-image to obtain the dimension of the object, wherein the dimension comprises color, size and shape;

s34: and training the visual recognition model according to the dimensions of each sub-image and the object thereof.

Specifically, in order to improve the accuracy of the garbage classification, the method can also perform recognition after garbage segmentation, that is, the method first obtains the contour of each object in the image information, and segments the image to obtain the sub-image of each object. For example, if garbage thrown by a user contains apple pits, paper towels and water cups, the image is segmented to obtain three sub-images: apple pits, paper towels and water cups. And then, sequentially extracting the features of the three images, and finally training according to the images and the extracted dimensions.

specifically, when spam identification fails, the user may be alerted to the reason for the failure. Therefore, the user can take the garbage to be classified which fails to be identified away in time and then reorganize the garbage to be classified for identification.

Specifically, after the method reminds the user of the identification failure, the user still does not take the garbage to be classified which is failed in the identification, and then prompt reminding information is generated to prompt the user to take the garbage to be classified which is failed in the identification quickly.

Preferably, the identifying the reason for failure includes:

specifically, if the recognition is failed due to the existence of multiple objects in the image of the garbage to be classified, or if the multiple objects in the image do not belong to the same classification category, the failure is caused by the fact that multiple objects are contained in the garbage.

specifically, if only images of a part of the object in the images of the garbage to be classified, such as a small part of a fruit slice, fail to be identified by the method due to too few image features of the part of the object, the failure reason is that the garbage structure of the user is incomplete and cannot be identified.

Specifically, if there is nothing in the image information of the garbage to be classified, the user is reminded that there is no garbage, and please re-identify the garbage.

For the sake of brief description, the method provided by the embodiment of the present invention may refer to the corresponding contents in the foregoing product embodiments.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; such modifications and substitutions do not depart from the spirit and scope of the present invention, and they should be construed as being included in the following claims and description.

Claims

1. An automatic classification garbage can based on artificial intelligence is characterized in that,

2. The automated classification trash can based on artificial intelligence of claim 1,

the categories of the sub-trash can include wet trash, dry trash, recyclable trash, and harmful trash.

3. A sorting method for automatically sorting trash cans based on artificial intelligence, operating on an identification device in a trash can according to claim 1 or 2, comprising the steps of:

establishing a plurality of classification categories;

4. The method for classifying trash cans automatically based on artificial intelligence as claimed in claim 3, wherein the collecting image information of trash under each classification category specifically comprises:

5. The method for classifying trash can automatically classified based on artificial intelligence as claimed in claim 3, further comprising, after the obtaining of the trained visual recognition model and before the capturing of the image information of trash to be classified by the camera:

6. The method for classifying trash cans automatically classified based on artificial intelligence as claimed in claim 5, wherein the step of analyzing the voice signal to obtain the classification category of the trash to be classified, and the step of controlling the corresponding driving device to open the corresponding can lid according to the classification category of the trash to be classified specifically comprises:

7. The method for classifying trash cans automatically classified based on artificial intelligence as claimed in claim 5, wherein the step of analyzing the voice signal to obtain the classification category of the trash to be classified, and the step of controlling the corresponding driving device to open the corresponding can lid according to the classification category of the trash to be classified specifically comprises:

8. The method for classifying trash cans automatically based on artificial intelligence as claimed in claim 3, wherein the training of the visual recognition model after the data set is added specifically comprises:

9. The method for classifying trash cans automatically classified based on artificial intelligence as claimed in claim 8, further comprising, after the step of controlling the corresponding driving device to open the corresponding can lid according to the classification category of the trash to be classified:

10. The method of claim 9, wherein the identifying the cause of failure comprises: