阅读说明：本技术 一种基于迭代学习控制的生活垃圾智能识别与自动分类系统及方法 (Intelligent household garbage identification and automatic classification system and method based on iterative learning control ) 是由 酆雨舟 陈昱昊 佟雨馡 周浩然 董思嘉 于 2019-08-27 设计创作，主要内容包括：本发明一种具有识别与分类的垃圾桶,它包括分类识别模块和垃圾分类桶；分类识别模块包括电源模块、微控制器、分类识别模块、语音控制模块和信息显示模块。分类识别模块感应垃圾投递手势,采集垃圾的照片并识别分类；语音控制模块采集语音和按键控制信息；信息显示模块显示垃圾状态信息并语音提醒；垃圾分类桶包括控制信息接收器和多个分类的分类垃圾子桶,每个分类垃圾子桶包括垃圾分桶、分桶盖板及分类垃圾袋,实现不同类别垃圾的分类收储。本发明兼顾投递手势、垃圾图像和语音控制信息,能够综合并集成手势、照片和语音以及按键控制信息,实现垃圾的智能识别和自动分类。(The invention relates to a garbage can with identification and classification functions, which comprises a classification identification module and a garbage classification can; the classification recognition module comprises a power supply module, a microcontroller, a classification recognition module, a voice control module and an information display module. The classification recognition module senses a rubbish delivery gesture, collects a picture of rubbish and recognizes classification; the voice control module collects voice and key control information; the information display module displays the garbage state information and carries out voice reminding; the garbage classification can comprises a control information receiver and a plurality of classified classification garbage sub-cans, each classification garbage sub-can comprises a garbage sub-can body, a sub-can cover plate and a classification garbage bag, and classification collection and storage of different types of garbage are achieved. The invention gives consideration to delivery gestures, junk images and voice control information, can integrate and integrate gestures, photos, voice and key control information, and realizes intelligent identification and automatic classification of junk.)

1. The system is characterized in that the intelligent recognition and automatic classification system comprises a garbage recognition device (1) and a garbage classification barrel (2), the garbage recognition device (1) is arranged at the upper end of the garbage classification barrel (2), the garbage recognition device (1) is connected with the garbage classification barrel (2) in a sensing manner, and the method comprises the following steps:

the classification recognition apparatus (1) includes: the device comprises a power supply module (101), a microcontroller (102), an induction acquisition module (103), a voice control module (104) and an information display module (105), wherein the power supply module (101) provides electric energy for the classification recognition device (1), and the induction acquisition module (103), the voice control module (104) and the information display module (105) are respectively connected with a port of the microcontroller (102); the intelligent garbage collection and classification system is characterized in that a firmware program of a pre-trained intelligent garbage recognition and classification model is arranged in the microcontroller (102), the microcontroller (102) initializes the induction collection module (103), the voice control module (104) and the information display module (105), the induction collection module (103) collects garbage delivery gestures and entity photos and transmits the collected garbage delivery gestures and the collected entity photos to the microcontroller (102), the voice control module (104) collects voice and key control information of a garbage deliverer and transmits the collected voice and key control information to the microcontroller (102), and the microcontroller (102) receives the voice and key control information transmitted by the voice control module (104) and judges classification categories corresponding to the voice and the keys;

the garbage classification barrel (2) comprises a classification information receiver (201), a plurality of classification garbage sub-barrels (202) and corresponding classification garbage bags (203), the garbage classification barrel (2) is provided with at least three classification garbage sub-barrels (202), the classification garbage sub-barrels (202) are provided with LED display screens (501), each classification garbage sub-barrel (202) comprises a barrel cover (6), a barrel body (7) and a distance sensor (204), the barrel covers can be opened and closed, the distance sensors (204) are arranged in the garbage classification barrels (2), the collected garbage information is updated and displayed on the information display module (105) in real time, the classification information receiver (201) is in wireless connection with the classification garbage sub-barrels (202), the classification information receiver (201) sends clear garbage classification information, and the classification information receiver (201) controls automatic opening and closing of the barrel covers of the classification garbage sub-barrels (202) corresponding to classification according to the received garbage classification type information.

2. The trash can with identification and classification of claim 1, wherein: the sensing acquisition module (103) comprises an infrared sensor (301), a camera (302), an illumination sensor (303) and an LED auxiliary illuminating lamp (304), the infrared sensor (301) senses a rubbish delivery gesture, the camera (302) acquires rubbish photos, and when the illumination sensor (303) senses that light is dark, the LED auxiliary illuminating lamp (304) is turned on and acquired information is transmitted to the microcontroller (102); the voice control module (104) comprises a sound pickup (401), a voice recognition processor (402) and a plurality of classification keys (403), wherein the voice recognition processor (402) analyzes and recognizes voice classification information collected by the sound pickup (401) and corresponding garbage classification information input by the classification keys (403); the information display module (105) comprises an LED display screen (501) and a voice synthesis module (502), the LED display screen (501) displays the state information of classified garbage, the voice synthesis module (502) simulates human voice to carry out voice reminding, and the microcontroller (102) acquires, identifies and analyzes the collected information to determine the classification of the garbage.

3. The trash can with identification and classification of claim 1, wherein: the microcontroller (102) processes, compares and analyzes the garbage delivery gesture, the entity photo, the voice and the key control information, automatically and intelligently identifies the corresponding garbage classification, and sends the garbage classification information to the classification information receiver (201) of the garbage classification barrel (2).

4. The trash can with identification and classification of claim 1, wherein: the classified information receiver (201) of the garbage classified bin (2) receives the garbage classified information, controls the bin cover of the classified garbage sub-bin (202) corresponding to the classification to be opened and closed, and realizes intelligent opening and closing of the bin cover during classified garbage delivery; the distance sensor (204) in the garbage classification barrel (2) monitors the garbage height in the classification garbage sub-barrel (202) in real time, updates and displays the collected garbage information on an LED display screen (501) of the information display module (105) in real time, and carries out voice prompt.

5. An intelligent identification and automatic classification method of household garbage based on iterative learning control is characterized in that: the method for realizing intelligent garbage identification and automatic garbage classification comprises the following steps:

a. after a power supply module (101) of the classification identification module (1) is powered on, a microcontroller (102) initializes an induction acquisition module (103), a voice control module (104) and an information display module (105) and monitors monitoring information and running states of the modules in real time;

b. the sensing acquisition module (103) acquires the garbage delivery gesture and the entity photo and transmits the acquired garbage delivery gesture and the acquired entity photo to the microcontroller (102);

c. the voice control module (104) collects the voice and key control information of the garbage deliverer and transmits the collected voice and key control information to the microcontroller (102);

d. the microcontroller (102) processes, compares and analyzes the garbage delivery gesture, the entity photo, the voice and the key control information, automatically and intelligently identifies corresponding garbage classification, sends the garbage classification information to a classification information receiver (201) of the garbage classification barrel (2), and iteratively updates the gesture, the voice and the control information into a classification image library as marks of the garbage entity photo;

e. a classification information receiver (201) of the garbage classification barrel (2) receives the garbage classification information, controls the opening and closing of a barrel cover of a classification garbage sub-barrel (202) corresponding to classification, and realizes intelligent opening and closing of the barrel cover during classified garbage delivery;

f. the distance sensor (204) in the garbage classification barrel (2) monitors the garbage height in the classification garbage sub-barrel (202) in real time, updates and displays the collected garbage information on an LED display screen (501) of the information display module (105) in real time, and carries out voice prompt.

6. The process of the inductive pick module (103) of claim 5 for picking up spam gestures and physical photos, comprising the steps of:

2a, waiting for an infrared sensor (301) in an induction acquisition module (103) to scan in real time and capture the infrared change condition of a garbage throwing area;

2b, judging whether the infrared change sensed by the infrared sensor (301) is a garbage delivery gesture, if so, executing the step 2c, and if not, executing the step 2 a;

2c, judging whether the illumination intensity monitored by the illumination sensor (303) in real time is too dark, and if so, turning on an LED auxiliary illuminating lamp (304);

2d, the camera (302) takes and captures the garbage delivery gesture and the entity picture and transmits the captured picture to the microcontroller (102).

7. The voice control module (104) of claim 5 collecting voice and key control processes of a spammer, comprising the steps of:

3a, waiting for a sound pickup (401) on the voice control module (104) to receive voice information;

waiting for a plurality of classification keys (403) on the voice control module (104) to receive key control information;

3c, judging whether the voice and the key information are definite garbage classification information, if so, executing the step 3d, otherwise, executing the step 3 a;

3d, the voice control module (104) sends explicit garbage classification information to the microcontroller (102).

8. The microcontroller (102) processing, comparing and analyzing spam delivery gestures, physical photos, voice and key control information process of claim 5 comprising the steps of:

4a, the microcontroller (102) receives the garbage delivery gesture and the entity photo transmitted by the induction acquisition module (103), calls the intelligent recognition firmware program of the garbage delivery gesture and the entity classification, and recognizes the delivery gesture in the photo and the classification category corresponding to the garbage;

4b, the microcontroller (102) receives the voice and key control information transmitted by the voice control module (104) and judges classification categories corresponding to the voice and the keys;

4c, judging and comparing whether the gesture, classification, voice and garbage classification information controlled by the keys are consistent, if so, executing the step 4e, and if not, executing the step 4 d;

confirming the garbage classification category according to the priority of 'key press > voice > gesture > photo', marking the garbage classification category with high priority as a classification label of the collected photo, iteratively updating parameters of the intelligent garbage recognition classification model, and realizing iterative learning control of the classification model;

sending the clear spam classification information to the classification information receiver (201).

9. The classification information receiver (201) according to claim 5, which receives the garbage classification information and controls the intelligent opening and closing process of the bin cover of the corresponding classified garbage sub-bin (202), comprising the following steps:

5a, after a classification information receiver (201) of the garbage classification barrel (2) receives the clear garbage classification information, opening a barrel cover of a classification garbage sub-barrel (202) corresponding to classification;

and 5b, when the infrared sensor of the garbage sorting sub-barrel (202) senses infrared change, judging whether the garbage delivery is finished, and if so, closing the barrel cover of the garbage sorting sub-barrel (202).

10. The process of real-time updating and displaying the garbage information collected by the garbage classification can (2) to the LED display screen (501) of the information display module (105) according to claim 5, comprising the following steps:

a distance sensor (204) of the garbage sorting bin (2) monitors the height of the garbage in the sorting garbage sub-bin (202) in real time and sends collected data to a microprocessor (102);

and 6b, summarizing the height information of the garbage in each garbage classification barrel by the microprocessor (102), displaying the state information of the garbage in the barrel on an LED display screen (501) in a chart mode, and simulating voice to prompt by a voice synthesis module (502).

11. The iteratively updated smart garbage recognition classification model parameters of claim 8, wherein the process comprises: according to the method, collected garbage delivery entity photos are shot by a camera, corresponding classification categories are obtained through a pre-trained intelligent garbage recognition classification model, the obtained classification categories are compared with collected gesture, voice and key control input information, parameters of the intelligent garbage recognition classification model are iteratively corrected through compared error information, and iterative updating is continuously repeated in such a way, so that the next intelligent recognition classification result is more consistent with the actual classification category, namely, specific characteristics of household garbage generated by residents are recorded, tracked and repeatedly learned, and the purpose of gradually improving recognition classification precision is achieved.

Technical Field

The invention relates to the field of intelligent environmental protection of household garbage, in particular to a system and a method for intelligently identifying and automatically classifying the household garbage based on iterative learning control.

Background

Along with the gradual improvement of the living standard of urban people, more and more domestic garbage are generated, and the domestic garbage becomes one of important pollution sources of cities. Although classification trash cans are arranged in a plurality of communities at present, as residents do not pay attention to or are difficult to correctly distinguish trash classification in daily life, the mixing rate of household trash is always high, and the traditional classification trash can becomes a decoration.

However, the body of the mandatory classification rules is still limited to the public institutions and related enterprises, and the residents who generate most domestic garbage have no daily habit of forming garbage classification and lack the common knowledge of related garbage classification.

The household garbage generated by residents has the typical characteristics of wide sources and scattered positions, the household garbage can be generated in places where the residents live, different residents can generate different types of garbage in life, and the same garbage can generate different conditions after being treated by different residents. Aiming at the situations, the invention collects and analyzes the specific characteristics of domestic garbage generated by each resident, designs and develops a domestic garbage intelligent identification and automatic classification system and method based on iterative learning control.

Patent No. 201810365225.5 discloses a working method of a multifunctional trash can machine for automatic trash classification, which is characterized in that an infrared emission tube, an infrared reflection receiving tube and an infrared projection receiving tube are arranged in a trash detection bin at the front side of a can body, and the trash such as paper, metal, plastic, glass and the like is identified and classified according to the difference of absorption and reflection capacities of trash substances to 950nm near infrared light. However, it can only identify one type of garbage, such as one of paper, metal, plastic and glass, and if these various types of garbage need to be recovered separately, several garbage cans must be installed, so it is difficult to meet the classification and identification requirements of various types, different forms and mixed nature of domestic garbage.

Patent No. CN201811422605.4 discloses a control device and method for an intelligent trash can, and the intelligent trash can, wherein the working mode and dynamic path planning of the trash can are controlled by obtaining voice control information, and a convolutional neural network is used to model and classify a large number of samples in a picture database. However, the picture library is not expanded after being established, and cannot be calibrated and updated according to the specific domestic garbage condition generated by the resident user, so that the picture library is difficult to identify and classify with the specific garbage feature generated by the user.

Patent number CN201910191797.0 discloses a multi-angle garbage classification cloud platform based on deep learning, acquires the rubbish picture through front end image acquisition module, conveys the rubbish image to garbage classification cloud platform through wireless high-speed transmission module, discerns and classifies. However, the wireless high-speed transmission module and the cloud platform need to be maintained for a long time, the 4G network transmission module included in the cloud platform needs to consume a large amount of 4G network traffic for transmitting the garbage image, and high wireless network communication cost is generated.

Patent No. 201910312315.2 an intelligent classification garbage bin based on machine vision adopts the identification control system on the center rotary platform, after discerning the rubbish in the garbage bin, and control center selective shipment platform is rotatory for the rotatory top that divides the bucket to the rubbish of the rubbish kind of discerning of discernment bucket, and control discernment platform selective shipment makes the opening of center rotary platform open. The central rotating platform is used as a garbage delivery and identification platform, so that the volume and the weight of delivered garbage are limited, the interior light is dark due to the sealing property of the identification platform, and the quality of images acquired by the central rotating platform is difficult to guarantee although a light source module is adopted, so that the accuracy of garbage image identification and classification can be greatly reduced.

Disclosure of Invention

The invention aims to solve the technical problem of providing a system and a method for collecting, processing, integrating and analyzing delivery gestures, garbage images, voices and key control information when residents deliver garbage, fully utilizing the collected gestures, voices and key control input information, further collecting and labeling garbage entity photos, recording, tracking and repeatedly learning garbage characteristics generated by different resident groups, and realizing accurate recognition and automatic classification of household garbage.

The iterative learning control labels the garbage entity pictures by using control input information such as voice, gestures and keys during garbage delivery, and generates a garbage classification image library with user characteristics by repeatedly labeling the garbage delivery pictures and classification information so as to improve the accuracy of intelligent garbage identification and automatic classification.

In order to solve the technical problems, the invention provides an intelligent household garbage identification and automatic classification system and method based on iterative learning control, which comprises a garbage identification device and a garbage classification barrel, wherein the garbage identification device is arranged at the upper end of the garbage classification barrel and is connected with the garbage classification barrel in an induction way, and the system comprises:

the classification recognition device includes: the system comprises a power supply module, a microcontroller, an induction acquisition module, a voice control module and an information display module, wherein the power supply module supplies electric energy to a classification recognition module, and the induction acquisition module, the voice control module and the information display module are respectively connected with a port of the microcontroller; the intelligent garbage delivery gesture and entity classification recognition system is characterized in that a garbage delivery gesture and an entity classification intelligent recognition firmware program are arranged in the microcontroller, the microcontroller initializes an induction acquisition module, a voice control module and an information display module, the induction acquisition module acquires the garbage delivery gesture and an entity photo and transmits the acquired garbage delivery gesture and the acquired entity photo to the microcontroller, the voice control module acquires voice and key control information of a garbage deliverer and transmits the acquired voice and key control information to the microcontroller, and the microcontroller receives the voice and key control information transmitted by the voice control module and judges classification categories corresponding to the voice and the keys;

the garbage classification can comprises a classification information receiver, a plurality of classification garbage sub-cans and corresponding classification garbage bags, the garbage classification can is provided with at least three classification garbage sub-cans, the classification garbage sub-cans are provided with LED display screens, the classification garbage sub-cans comprise openable and closable cans covers, cans and distance sensors, the distance sensors are arranged in the garbage classification can and update and display collected garbage information to an information display module in real time, the classification information receiver is in wireless connection with the classification garbage sub-cans, the classification information receiver sends clear garbage classification information, and the classification information receiver controls automatic opening and closing of the cans of the classification garbage sub-cans corresponding to classification according to the received garbage classification information.

As a further improvement of the invention, the sensing and collecting module comprises an infrared sensor, a camera, an illumination sensor and an LED auxiliary illuminating lamp, wherein the infrared sensor senses a rubbish delivery gesture, the camera collects rubbish photos, and when the illumination sensor senses dark light, the LED auxiliary illuminating lamp is turned on and collected information is transmitted to the microcontroller; the voice control module comprises a sound pick-up, a voice recognition processor and a plurality of classification keys, wherein the voice recognition processor analyzes and recognizes voice classification information collected by the sound pick-up and corresponding garbage classification information input by the classification keys; the information display module comprises an LED display screen and a voice synthesis module, the LED display screen displays the state information of classified garbage, and the voice synthesis module simulates human voice to carry out voice reminding. The microcontroller acquires, identifies and analyzes the collected information to determine the classification to which the garbage belongs.

As a further improvement of the invention, the microcontroller processes, compares and analyzes the garbage delivery gesture, the entity photo, the voice and the key control information, automatically and intelligently identifies the corresponding garbage classification, and sends the garbage classification information to the classification information receiver of the garbage classification bin.

As a further improvement of the invention, the classification information receiver of the garbage classification can receives the garbage classification information, controls the opening and closing of the can cover of the classification garbage sub-can corresponding to the classification, and realizes the intelligent opening and closing of the can cover during the delivery of the classification garbage; and a distance sensor in the garbage classification barrel monitors the garbage height in the classification garbage sub-barrel in real time, updates and displays the collected garbage information on an LED display screen of the information display module in real time, and carries out voice prompt.

Compared with the prior art, the invention has the following advantages:

the intelligent identification and automatic classification system for household garbage comprises: the classification recognition module device is a garbage classification barrel; the classification recognition device comprises a power supply module, a microcontroller, a classification recognition module, a voice control module and an information display module. The classification recognition module senses a rubbish delivery gesture, collects a picture of rubbish and recognizes classification; the voice control module collects voice and key control information; the information display module displays the garbage state information and carries out voice reminding; the garbage classification can comprises a control information receiver and a plurality of classified classification garbage sub-cans, each classification garbage sub-can comprises a garbage sub-can body, a garbage sub-can cover plate and a classification garbage bag, and classification collection and storage of different types of garbage are achieved. The invention gives consideration to delivery gestures, junk images and voice control information, can integrate and integrate gestures, photos, voice and key control information, and realizes accurate intelligent identification and automatic classification of junk.

Drawings

FIG. 1 is a schematic view of the garbage can with identification and classification according to the present invention.

FIG. 2 is a block diagram of the present invention with a classification recognition device for recognizing and classifying trash cans.

FIG. 3 is a detailed block diagram of the present invention with a classification recognition device for recognizing and classifying trash cans.

FIG. 4 is a block diagram of the garbage classification can of the present invention with the function of identifying and classifying the garbage.

FIG. 5 is a flow chart of the present invention for implementing intelligent garbage recognition and automatic garbage classification.

FIG. 6 is a flow chart of the invention for collecting garbage delivery gestures and physical photographs by the collection module.

FIG. 7 is a flow chart of the voice control module collecting the voice and key control of the spammer according to the present invention.

FIG. 8 is a flow chart of the microcontroller processing, comparing and analyzing spam delivery gestures, physical photos, voice and key control information.

Fig. 9 is a flow chart illustrating the process of controlling the intelligent opening and closing of the bin covers of the classified trash bins corresponding to the classification according to the present invention.

Fig. 10 is a flowchart illustrating real-time update and display of garbage information collected by the garbage can on the LED display screen according to the present invention.

FIG. 11 is a flowchart illustrating iterative updating of garbage intelligent recognition classification model parameters according to the present invention.

Wherein: 1-classification recognition module, 2-garbage classification barrel, 101-power module, 102-microcontroller, 103-induction acquisition module, 104-voice control module, 105-information display module, 201-classification information receiver, 202-classification garbage sub-barrel, 203-garbage bag, 301-infrared sensor, 302-camera, 303-illumination sensor, 304-LED auxiliary illuminating lamp, 401-sound pick-up, 402-voice recognition processor, 403-classification key, 501-LED display screen, and 502-voice synthesis module.

Detailed Description

As shown in fig. 1-3, an intelligent household garbage identification and automatic classification system and method based on iterative learning control, the intelligent identification and classification system includes: a classification and identification device 1 and a garbage classification barrel 2.

The classification recognition apparatus 1 includes: the system comprises a power supply module 101, a microcontroller 102, an induction acquisition module 103, a voice control module 104 and an information display module 105.

The power supply module 101 supplies electric energy to the classification recognition device 1; the induction acquisition module 103, the voice control module 104 and the information display module 105 are respectively connected with a port of the microcontroller 102; the microcontroller 102 is internally provided with a firmware program of a trained intelligent garbage recognition and classification model; the sensing acquisition module 103 comprises an infrared sensor 301, a camera 302, an illumination sensor 303 and an LED auxiliary illuminating lamp 304, wherein the infrared sensor 301 senses a rubbish delivery gesture, the camera 302 acquires rubbish photos, and when the illumination sensor 303 senses that light is dark, the LED auxiliary illuminating lamp 304 is turned on and transmits acquired information to the microcontroller 102; the voice control module 104 comprises a sound pickup 401, a voice recognition processor 402 and a plurality of classification keys 403, wherein the voice recognition processor 402 analyzes and recognizes voice classification information collected by the sound pickup 401 and corresponding garbage classification information input by the plurality of classification keys 403; the information display module 105 comprises an LED display screen 501 and a voice synthesis module 502, the LED display screen 501 displays the state information of classified garbage, and the voice synthesis module 502 simulates human voice to carry out voice reminding. The microcontroller 102 acquires, identifies and analyzes the collected information to determine the classification to which the spam belongs.

The trash sorting bin 2 includes a sorting information receiver 201 and a plurality of sorting trash sub-bins 202 and corresponding sorting trash bags 203. Each classified garbage sub-can 202 comprises an openable and closable can cover, a can body and a distance sensor, and the classified information receiver 201 controls the automatic opening and closing of the can cover of the corresponding classified garbage sub-can 202 according to the received garbage classification type information; the classification garbage sub-barrels 202 and the garbage bags 203 of different types are distinguished by different colors, and the garbage bags of each color store one type of classified garbage, are not mixed with other types of classified garbage any more, do not need to be sorted again, and realize the classification storage of different types of garbage.

As shown in fig. 5, the method for implementing intelligent garbage identification and automatic garbage classification includes the following steps:

a. after the power module 101 of the classification recognition module 1 is powered on, the microcontroller 102 initializes the induction acquisition module 103, the voice control module 104 and the information display module 105, and monitors monitoring information and running states of the modules in real time;

b. the sensing acquisition module 103 acquires the garbage delivery gesture and the entity photo, and transmits the acquired garbage delivery gesture and the acquired entity photo to the microcontroller 102;

c. the voice control module 104 collects the voice and key control information of the garbage deliverer and transmits the collected voice and key control information to the microcontroller 102;

d. the microcontroller 102 processes, compares and analyzes the garbage delivery gesture, the entity photo, the voice and the key control information, automatically and intelligently identifies the corresponding garbage classification, and sends the garbage classification information to the classification information receiver 201 of the garbage classification can 2;

e. the classification information receiver 201 of the garbage classification barrel 2 receives the garbage classification information, controls the opening and closing of the barrel cover of the classification garbage sub-barrel (202) corresponding to the classification, and realizes the intelligent opening and closing of the barrel cover during the delivery of the classification garbage;

f. the distance sensor 204 in the garbage sorting bin 2 monitors the height of the garbage in the sorting garbage sub-bin 202 in real time, updates and displays the collected garbage information on the LED display screen 501 of the information display module 105 in real time, and carries out voice prompt.

As shown in fig. 6, the process of collecting the garbage delivery gesture and the entity photo by the sensing collection module 103 includes the following steps:

2a, waiting for the infrared sensor 301 in the induction acquisition module 103 to scan in real time and capture the infrared change condition of the garbage throwing area;

2b, judging whether the infrared change sensed by the infrared sensor 301 is a garbage delivery gesture, if so, executing the step 2c, and if not, executing the step 2 a;

2c, judging whether the illumination intensity monitored by the illumination sensor 303 in real time is too dark, and if so, turning on the LED auxiliary illuminating lamp 304;

2d, the camera 302 shoots and collects the garbage delivery gesture and the entity photo, and transmits the collected photo to the microcontroller 102;

as shown in fig. 7, the voice control module 104 collects the voice of the spammer and the key control process, and includes the following steps:

3a, waiting for the sound pickup 401 on the voice control module 104 to receive voice information;

3b, waiting for a plurality of classification keys 403 on the voice control module 104 to receive key control information;

3c, judging whether the voice and the key information are definite garbage classification information, if so, executing the step 3d, otherwise, executing the step 3 a;

3d, the voice control module 104 sends explicit garbage classification information to the microcontroller 102;

as shown in fig. 8, the microcontroller 102 processes, compares and analyzes the spam delivery gesture, the entity photo, the voice and the key control information, comprising the following steps:

4a, the microcontroller 102 receives the garbage delivery gesture and the entity photo transmitted by the induction acquisition module 103, calls the intelligent recognition firmware program of the garbage delivery gesture and the entity classification, and recognizes the delivery gesture in the photo and the classification category corresponding to the garbage;

4b, the microcontroller 102 receives the voice and key control information transmitted by the voice control module 104, and judges classification categories corresponding to the voice and the keys;

4c, judging and comparing whether the gesture, classification, voice and garbage classification information controlled by the keys are consistent, if so, executing the step 4e, and if not, executing the step 4 d;

confirming the garbage classification category according to the priority of 'key press > voice > gesture > photo', marking the garbage classification category with high priority as a classification label of the collected photo, and iteratively updating parameters of a pre-trained garbage intelligent recognition classification model to realize iterative learning control of the classification model;

4e, sending clear garbage classification information to the classification information receiver 201;

as shown in fig. 9, the classification information receiver 201 receives the garbage classification information and controls the intelligent opening and closing process of the can lid of the classification garbage sub-can 202 corresponding to the classification, including the following steps:

5a, after the classification information receiver 201 of the garbage classification can 2 receives the clear garbage classification information, opening the can cover of the classification garbage sub-can 202 of the corresponding classification;

5b, when the infrared sensor of the classified garbage sub-bucket 202 senses infrared change, judging whether the garbage delivery is finished, and if so, closing the bucket cover of the classified garbage sub-bucket 202;

as shown in fig. 10, the process of updating and displaying the garbage information collected by the garbage sorting bin 2 on the LED display screen 501 of the information display module 105 in real time includes the following steps:

6a, a distance sensor 204 of the garbage sorting bin 2 monitors the height of the garbage in the sorting garbage sub-bin 202 in real time and sends collected data to the microprocessor 102;

and 6b, the microprocessor 102 collects the height information of the garbage in each garbage classification barrel, displays the state information of the garbage in the barrel on an LED display screen 501 in a chart mode, and simulates human voice to carry out voice prompt through the voice synthesis module 502.

As shown in fig. 11, the process of iteratively updating the parameters of the intelligent garbage recognition classification model is as follows: according to the method, collected garbage delivery entity photos are shot by a camera, corresponding classification categories are obtained through a pre-trained intelligent garbage recognition classification model, the obtained classification categories are compared with collected gesture, voice and key control input information, parameters of the intelligent garbage recognition classification model are iteratively corrected through compared error information, and iterative updating is continuously repeated in such a way, so that the next intelligent recognition classification result is more consistent with the actual classification category, namely, specific characteristics of household garbage generated by residents are recorded, tracked and repeatedly learned, and the purpose of gradually improving recognition classification precision is achieved.