Three-stage differential garbage classification treatment system

Abstract

The invention discloses a three-stage differential garbage classification treatment system, and relates to the technical field of energy conservation and environmental protection. A throwing opening is formed in the top of a shell, a classification mechanism is arranged in the shell, and the classification mechanism comprises a three-stage differential separation module, a throwing module and a visual recognition module; the three-stage differential separation module comprises a first conveying belt, a second conveying belt and a third conveying belt which are sequentially arranged from top to bottom, a dispersing roller is arranged below the throwing opening and is in transmission connection with the first conveying belt through a belt, a throwing module is arranged at one end of the third conveying belt, and a visual recognition module is arranged above the throwing module; more than two garbage cans are arranged at the bottom in the shell, and a compression mechanism is arranged in one garbage can. According to the three-stage differential garbage classification treatment system, the three-stage differential separation module is utilized, multiple kinds of garbage can be thrown at a time and accurately recognized and sorted, compared with manual throwing, the error rate is greatly reduced, and the recognition accuracy rate reaches up to 97.3%.

Description

Technical Field

[0001] The present invention relates to the technical field of energy conservation and environmental protection, and more specifically, to a three-stage differential garbage classification and treatment system. Background Art

[0002] Traditional garbage classification methods mainly rely on manual classification. That is, when residents dispose of garbage, they need to judge the type of garbage by themselves and put it into the corresponding trash can. With the development of technology, although some semi-automatic garbage classification devices have emerged, there are still many limitations in actual applications. Semi-automatic devices usually only have simple sensing functions, such as judging the approximate volume or shape of garbage through infrared sensing, and cannot accurately identify the types of garbage. Eventually, manual secondary sorting is still required.

[0003] The existing technical solutions have the following disadvantages: 1. Insufficient classification accuracy: Manual classification depends on residents' classification awareness and knowledge reserve. Different residents have different understandings of garbage classification standards, resulting in a relatively high classification error rate. A large amount of misclassified garbage needs to be manually sorted again at the garbage treatment plant, consuming a large amount of human and time costs and reducing the overall garbage classification efficiency; while semi-automatic garbage classification devices are limited by their simple sensing technology and cannot accurately identify complex garbage types. For example, some garbage may contain multiple materials at the same time, or have similar appearances but belong to different categories, which are beyond the recognition ability of semi-automatic devices, leading to inaccurate classification and affecting subsequent garbage treatment and recycling; 2. In the process of garbage disposal, the existing technology cannot provide residents with real-time classification guidance and feedback. After residents dispose of garbage, it is difficult to know whether their classification is correct, which is not conducive to improving residents' garbage classification skills and enthusiasm, nor to the popularization and promotion of garbage classification knowledge; 3. Poor adaptability: Facing the increasingly diverse and complex types of garbage, the existing garbage classification devices and methods are difficult to adapt. At the same time, the continuous emergence of new materials and packaging forms makes the traditional classification methods appear powerless in dealing with these new garbage, and cannot effectively meet the efficient and accurate requirements of modern society for garbage classification. Summary of the Invention

[0004] The technical problem to be solved by the present invention is that the existing garbage classification devices have problems of low separation accuracy and poor adaptability. In order to overcome the defects of the existing technology, the present invention proposes a three-stage differential garbage classification and treatment system. By using a three-stage differential separation module, it can realize the precise identification and sorting of multiple types of garbage in one-time delivery. Compared with manual delivery, the error rate is greatly reduced, and the recognition accuracy rate is as high as 97.3%.

[0005] To achieve this purpose, the present invention adopts the following technical solutions:

[0006] The present invention provides a three - stage differential garbage classification and treatment system, which includes a housing, a classification mechanism, a compression mechanism and a garbage bin. A feeding port is arranged at the top of the housing. A classification mechanism is arranged inside the housing. The classification mechanism includes a three - stage differential separation module, a feeding module and a visual recognition module. The three - stage differential separation module includes a first conveyor belt, a second conveyor belt and a third conveyor belt which are arranged in sequence from top to bottom. A dispersing roller is arranged below the feeding port. The dispersing roller is connected to the first conveyor belt through belt drive. One end of the third conveyor belt is provided with a feeding module, and a visual recognition module is arranged above the feeding module. Two or more garbage bins are arranged at the bottom inside the housing, and a compression mechanism is arranged in one of the garbage bins.

[0007] In a preferred technical solution of the present invention, the feeding module includes a support beam, a first servo motor, a second servo motor and a garbage shovel. The support beam is horizontally arranged below the third conveyor belt. A first servo motor is arranged on the support beam. The power output shaft of the first servo motor is connected to a second servo motor, and the power output shaft of the second servo motor is connected to a garbage shovel. And the power output shafts of the first servo motor and the second servo motor are perpendicular to each other.

[0008] In a preferred technical solution of the present invention, the visual recognition module includes a camera, a main controller, a display screen and a voice broadcaster. The camera is arranged above the garbage shovel. The main controller is fixedly arranged on the top of the housing. The display screen and the voice broadcaster are integrated on the main controller. And the camera, the display screen and the voice broadcaster are all electrically connected to the main controller.

[0009] In a preferred technical solution of the present invention, an ultrasonic sensor is further arranged inside the garbage bin, and the ultrasonic sensor is electrically connected to the main controller.

[0010] In a preferred technical solution of the present invention, the compression mechanism includes an electric push rod, a fixed compression plate, a movable compression plate, a first spring and a conical tip. The fixed compression plate and the movable compression plate are both vertically arranged inside the garbage bin, and are connected by a first spring between the fixed compression plate and the movable compression plate. The electric push rod is arranged on the housing, and the output end of the electric push rod is connected to the fixed compression plate. A conical tip is further arranged on the fixed compression plate, and a through hole matching the conical tip is arranged on the movable compression plate.

[0011] In a preferred technical solution of the present invention, a liquid collection bin is further arranged at the bottom inside the garbage bin. A hook is arranged on the liquid collection bin, and the hook is hung on the garbage bin. A filter plate is arranged at the top of the liquid collection bin, and filter holes are arranged on the filter plate.

[0012] In a preferred technical solution of the present invention, baffles are arranged on the first conveyor belt, the second conveyor belt and the third conveyor belt.

[0013] In a preferred technical solution of the present invention, universal wheels are arranged at the bottom of the housing.

[0014] The beneficial effects of the present invention are as follows:

[0015] 1. By using the three-stage differential separation module, it is possible to achieve the precise identification and sorting of multiple types of garbage with a single input. Compared with manual input, the error rate is greatly reduced, and the recognition accuracy rate is as high as 97.3%.

[0016] 2. By adding inclined baffles to the first conveyor belt, the second conveyor belt, and the third conveyor belt, it effectively prevents the garbage from getting stuck during the conveying process and enables it to be smoothly transported into the garbage shovel.

[0017] 3. The double servo motors are used to drive the garbage shovel for discharging, effectively realizing the classification of garbage.

[0018] 4. The display screen shows the garbage categories in real time, which helps to improve the awareness of garbage classification of the depositors.

[0019] 5. The ultrasonic sensor can detect the full load of the trash can, and then control the compression mechanism to compress the recyclable garbage, effectively reducing the volume. At the same time, during the compression process, the solid-liquid separation of the garbage is realized, effectively solving the problem of liquid leakage during compression. At the same time, the design of the hook is conducive to the later treatment of the liquid and solid generated by compression, and saves volume at the same time.

[0020] 6. The device is driven by universal wheels. Through the voice call module, the device will perform path planning and move autonomously, which is convenient for discharging. Brief Description of the Drawings

[0021] Figure 1 is a three-dimensional structural schematic diagram of the three-stage differential garbage classification and treatment system provided by the specific embodiment of the present invention;

[0022] Figure 2 is Figure 1 a structural schematic diagram in the main viewing direction;

[0023] Figure 3 is a structural schematic diagram of the visual recognition module;

[0024] Figure 4 is a structural schematic diagram of the compression mechanism;

[0025] Figure 5 is the F1 score-confidence curve of the visual recognition module for different garbage categories;

[0026] Figure 6 is the precision-confidence curve of the model prediction of the visual recognition module at different confidence levels;

[0027] Figure 7 is the precision-recall curve and mAP@0.5 of the visual recognition module.

[0028] In the figure:

[0029] 1. Housing; 11. Feeding port; 12. Universal wheel; 2. Classification mechanism; 21. Three-stage differential separation module; 211. Dispersion roller; 212. Belt; 213. First conveyor belt; 214. Second conveyor belt; 215. Third conveyor belt; 216. Baffle; 22. Feeding module; 221. Support beam; 222. First servo; 223. Second servo; 224. Garbage shovel; 23. Visual recognition module; 231. Camera; 232. Main controller; 233. Display screen; 234. Voice announcer; 235. Ultrasonic sensor; 3. Compression mechanism; 31. Electric push rod; 32. Fixed compression plate; 33. Movable compression plate; 331. Through hole; 34. First spring; 35. Conical tip; 36. Liquid collection bin; 37. Hook; 38. Filter plate; 4. Garbage can. Detailed implementation manners

[0030] The technical solution of the present invention will be further described below in conjunction with the accompanying drawings and through specific implementation manners.

[0031] As Figures 1-7As shown, a three-stage differential garbage classification and processing system is provided in the embodiment, including a housing 1, a classification mechanism 2, a compression mechanism 3 and a garbage can 4. A delivery port 11 is provided at the top of the housing 1, and a classification mechanism 2 is provided in the housing 1. The classification mechanism 2 includes a three-stage differential separation module 21, a delivery module 22 and a visual recognition module 23. The three-stage differential separation module 21 includes a first conveyor belt 213, a second conveyor belt 214 and a third conveyor belt 215 arranged in sequence from top to bottom. A dispersion roller 211 is provided below the delivery port 11, and the dispersion roller 211 is connected to the first conveyor belt 213 through a belt 212. A delivery module 22 is provided at one end of the third conveyor belt 215, and a visual recognition module 23 is provided above the delivery module 22. More than two garbage cans 4 are provided at the bottom of the housing 1, and a compression mechanism 3 is provided in one of the garbage cans 4. In this embodiment, the housing 1 is a rectangular box structure. One end of the first conveyor belt 213 is aligned with the position of the delivery port 11, and the user can deliver garbage to the housing 1 through the delivery port 11 at the top. The classification mechanism 2 is used to classify the garbage put in by the user and put it into the corresponding garbage bin 4. The compression mechanism 3 is used to compact the garbage in the garbage bin 4 to improve the space utilization rate. There are four garbage bins 4, which are used to hold different types of garbage (such as recyclable garbage, kitchen waste, other garbage and hazardous waste) for subsequent recycling and treatment. The three-stage differential separation module 21 can utilize the speed difference between the three conveyor belts to evenly disperse the garbage, so as to facilitate the visual recognition module 23 to identify and classify. Among them, the dispersion blades are evenly arranged on the dispersion roller 211, and the dispersion roller 211 is located above one end of the first conveyor belt 213, which can initially disperse the garbage thrown by the user and prevent the garbage from accumulating on the first conveyor belt 213; the speeds of the first conveyor belt 213, the second conveyor belt 214 and the third conveyor belt 215 are successively increased, and the speed difference is utilized to make the garbage evenly dispersed on the second conveyor belt 214 and the third conveyor belt 215, so as to ensure that the distance between adjacent garbage is large enough, so as to facilitate the visual recognition module 23 to accurately identify the type of garbage, and greatly reduce the error rate compared with manual delivery, and the recognition accuracy rate is as high as 97.3%. The delivery module 22 is located below one end of the third conveyor belt 215, and is used to receive the garbage from the third conveyor belt 215, and deliver the garbage to the corresponding garbage bin 4 according to the recognition result of the visual recognition module 23. The three-stage differential separation module 21, the delivery module 22 and the visual recognition module 23 are electrically connected. In addition, the components used in this embodiment are all commercially available and will not be described in detail here.

[0032] Specifically, the delivery module 22 includes a support beam 221, a first servo 222, a second servo 223, and a garbage shovel 224. The support beam 221 is horizontally arranged below the third conveyor belt 215. A first servo 222 is arranged on the support beam 221. The power output shaft of the first servo 222 is connected to a second servo 223, and the power output shaft of the second servo 223 is connected to a garbage shovel 224. Moreover, the power output shafts of the first servo 222 and the second servo 223 are perpendicular to each other. In this embodiment, the support beam 221 is horizontally arranged between the garbage bin 4 and the third conveyor belt 215, and both ends of the support beam 221 are fixedly installed on the inner side wall of the housing 1. The power output shaft of the first servo 222 is arranged vertically upward, used to drive the garbage shovel 224 to rotate in the horizontal plane, so that the garbage in the garbage shovel 224 moves to directly above the corresponding garbage bin 4. The power output shaft of the second servo 223 is arranged horizontally, used to drive the garbage shovel 224 to rotate in the vertical plane, so as to pour the garbage into the corresponding garbage bin 4. And the first servo 222 and the second servo 223 cooperate with each other to realize automatic classification and delivery of garbage. The garbage shovel 224 has a triangular prism structure. In the initial state, the opening of the garbage shovel 224 faces upward and is aligned with one end of the third conveyor belt 215, so that the garbage can smoothly fall into the garbage shovel 224.

[0033] Specifically, the visual recognition module 23 includes a camera 231, a main controller 232, a display screen 233, and a voice broadcaster 234. The camera 231 is arranged above the garbage shovel 224. The main controller 232 is fixedly arranged on the top of the housing 1. The display screen 233 and the voice broadcaster 234 are integrated on the main controller 232. Moreover, the camera 231, the display screen 233, and the voice broadcaster 234 are all electrically connected to the main controller 232. In this embodiment, the camera 231 is used to capture the garbage image in the garbage shovel 224 and send the image information to the main controller 232 for analysis and processing. The model of the main controller 232 is preferably a CanMV-K230 RISC-V chip, which has characteristics such as high precision, low latency, high performance, ultra-low power consumption, and fast startup. An identification model is preset in the main controller 232, which can accurately identify the types of garbage in the image, and then control the delivery module 22 to deliver it to the corresponding garbage bin 4. The set display screen 233 can synchronously present the types of garbage, enhancing the transparency of delivery and operation feedback. For example, kitchen waste - carrot. At the same time, the set voice broadcaster 234 can broadcast the recognition result, thus meeting the usage needs of elderly or low-educated users and reducing the usage threshold.

[0034] Specifically, an ultrasonic sensor 235 is further provided in the trash can 4, and the ultrasonic sensor 235 is electrically connected to the main controller 232. In this embodiment, the provided ultrasonic sensor 235 can obtain the remaining capacity in the trash can 4 by detecting the accumulation height of the garbage in the trash can 4, and display the remaining capacity information on the display screen 233. And through the dual visual and auditory prompts, the usability and management efficiency of the waste sorting facilities are significantly improved.

[0035] Specifically, the compression mechanism 3 includes an electric push rod 31, a fixed compression plate 32, a movable compression plate 33, a first spring 34 and a conical tip 35. The fixed compression plate 32 and the movable compression plate 33 are both vertically arranged in the trash can 4, and the fixed compression plate 32 and the movable compression plate 33 are connected by the first spring 34. The electric push rod 31 is arranged on the housing 1, and the output end of the electric push rod 31 is connected to the fixed compression plate 32. A conical tip 35 is further arranged on the fixed compression plate 32, and a through hole 331 matching the conical tip 35 is arranged on the movable compression plate 33. In this embodiment, the main body of the electric push rod 31 is located in the groove at the bottom of the adjacent trash can 4, and can push the fixed compression plate 32 and the movable compression plate 33 to slide left and right in the trash can 4 to realize the compression of the garbage and improve the space utilization rate. The fixed compression plate 32 and the movable compression plate 33 are both of flat plate structures, and in the initial state, the fixed compression plate 32 and the movable compression plate 33 are located at one side inside the trash can 4. Two or more conical tips 35 are evenly arranged on the side of the fixed compression plate 32 close to the movable compression plate 33, and during the compression process, the conical tips 35 can pass through the through hole 331 to puncture plastic paper boxes and the like in the trash can 4 for exhaust treatment, so as to realize solid-liquid separation while compressing the garbage and prevent odors such as putrefaction. Two or more first springs 34 are provided. When the plastic is stuck on the conical tip 35, the first spring 34 can provide an outward force to the garbage to separate it from the conical tip 35, which is beneficial to the treatment of the compressed garbage and effectively improves the safety of people when handling garbage, preventing scratches by spines and the like. The number of the conical tips 35 is equal to that of the through holes 331 and their positions correspond one by one.

[0036] Specifically, a liquid collection bin 36 is further provided at the inner bottom of the trash can 4. A hook 37 is provided on the liquid collection bin 36, and the hook 37 is hung on the trash can 4. A filter plate 38 is provided at the top of the liquid collection bin 36, and filter holes are provided on the filter plate 38. In this embodiment, the liquid collection bin 36 is used to collect the liquid after solid-liquid separation of the garbage. The provided hook 37 facilitates lifting the liquid collection bin 36 from the trash can 4 for subsequent processing. The provided filter plate 38 can ensure that the sewage passes through smoothly and falls into the liquid collection bin 36, while the solid impurities are blocked in the trash can 4. A sewage discharge port is further provided at the bottom of one side of the trash can 4, which can discharge the collected sewage in it without removing the liquid collection bin 36, facilitating use. More than two filter holes are evenly provided.

[0037] Specifically, baffles 216 are provided on the first conveyor belt 213, the second conveyor belt 214 and the third conveyor belt 215. In this embodiment, the baffles 216 are arranged obliquely, which can effectively prevent the garbage from getting stuck during the conveying process and enable it to be smoothly transported to the feeding module 22.

[0038] Specifically, universal wheels 12 are provided at the bottom of the housing 1. In this embodiment, the housing 1 is driven by the universal wheels 12. A voice call module is also integrated on the main controller 232, which can recognize the user's voice. Through the voice call module, the device will perform path planning and move autonomously, facilitating the user to put garbage.

[0039] This embodiment also provides a usage method of the three-stage differential garbage classification and treatment system, including the following steps:

[0040] S1. Garbage feeding and primary conveying: The user feeds the garbage through the feeding port 11 above the housing 1, and the garbage then falls onto the first conveyor belt 213, thus completing the initial collection of the garbage and introducing the garbage into the classification system to prepare for subsequent classification operations;

[0041] S2. Primary differential sorting: The garbage is conveyed from the first conveyor belt 213 to the second conveyor belt 214. Due to the different volumes and masses of the garbage itself and the randomness during feeding, the distribution of the garbage on the conveyor belt is different. At the same time, there is a speed difference between the first conveyor belt 213 and the second conveyor belt 214, which causes a displacement difference of different garbage during the process of being conveyed from the first conveyor belt 213 to the second conveyor belt 214, thereby initially sorting the garbage and making the different types of garbage start to show a difference in position during the conveying process, laying a foundation for further precise sorting;

[0042] S3. Secondary differential sorting: The garbage is conveyed to the third conveyor belt 215 by the second conveyor belt 214, and the above-mentioned conveyor belt differential sorting steps are repeated to further increase the distance between the garbage, thereby effectively expanding the spacing between different garbage, making the distribution of the garbage more discrete, and facilitating subsequent identification and classification operations;

[0043] S4. Garbage identification and type display: The garbage is conveyed to the garbage shovel 224 by the third conveyor belt 215. At this time, the garbage is identified by the visual recognition module 23. At the same time, the display screen 233 on one side of the feeding port 11 displays the type of the garbage being identified in real time, so as to accurately judge the type of the garbage and timely show it to the user, improving the transparency of garbage classification and the user's classification awareness;

[0044] S5. Garbage is put into the corresponding area: According to the recognition result of the visual recognition module 23, first, the first servo 222 drives the garbage shovel 224 to rotate above the corresponding trash can 4, and then the second servo 223 drives the garbage shovel 224 to rotate to accurately pour the garbage into the corresponding trash can 4, so as to realize the classified collection of garbage;

[0045] S6. Detection of full trash can: The ultrasonic sensor 235 monitors the garbage loading amount in the trash can 4 in real time (detection accuracy ±2 cm), and the display screen 233 analyzes the recognition result in real time, dynamically displaying the garbage type (such as "recyclables - plastic bottles") and the remaining capacity percentage in the trash can 4 (such as "remaining space: 65%"); when it is detected that the trash can is full (capacity ≥ 90%), the display screen 233 synchronously displays a red warning icon and gives a voice prompt "The trash can is about to be full, please clean it in time";

[0046] S7. Compressing and processing garbage: When the ultrasonic sensor 235 detects that a certain trash can 4 is full of garbage, if it is hazardous waste, kitchen waste or other waste, it will be fed back to the display screen 233 to remind to clean the garbage of the independent category in time, and if it is recyclable waste (such as paper boxes, plastic bottles, etc.), the compression mechanism 3 will be triggered to compress the garbage to reduce the volume and improve the cycle of cleaning the trash can 4.

[0047] In step S4, the recognition model used by the visual recognition module 23 is the "lajifenlei.pt" model trained based on the YOLOv5 framework (the structure of the "lajifenlei.pt" model is the same as that of the YOLOv5 framework). It is converted to the ONNX format through export.py, and the original detection head structure is retained to ensure integrity. The NNCASE toolchain is used to convert the ONNX model to the kmodel format dedicated to K230, and INT8 quantization and channel pruning are completed synchronously. The model volume is further compressed and the accuracy loss is less than 1%. The entire process does not rely on the cloud, supports local deployment, and ensures data privacy and real-time performance. Moreover, for the problems of uneven quantity of various label data and constant fill light environment inside the trash can in the original dataset (including 12,000 pictures), four types of geometric transformation methods (namely Affine affine transformation, BBoxSafeRandomCrop safe random cropping, GridDistortion grid distortion, Perspective perspective transformation) and two types of pixel transformation data augmentation methods (namely GaussNoise Gaussian noise and ISONoise camera sensor noise) are used, effectively expanding the data volume of the corresponding dataset. At the same time, redundant weather enhancement (such as rain and fog) is avoided, the anti-interference ability of the model is improved, and subsequent HSV color enhancement (such as hsv_s = 0.7) can be extended to handle the stain coverage scenario, further improving the robustness in complex environments. In addition, to solve the problems of high cost and insufficient samples in real garbage data annotation, data can be artificially expanded (such as garbage images with different angles, brightness, and occlusions). The enhanced dataset covers more extreme scenarios (such as tilted plastic bottles and partially occluded carrots), forcing the model to learn more robust features (such as contours and textures) rather than relying on a single visual cue (such as color at a fixed angle), and combined with the adaptive anchor box mechanism of YOLOv5, the adaptability of the model to scale and position changes is further improved.

[0048] The visual recognition module 23 in this embodiment has the following advantages:

[0049] 1. Low-cost and high-performance visual recognition solution: The CanMV-K230 RISC-V chip is adopted, which integrates a dual-core XuanTie C908 CPU (main frequency 1.6GHz) and the third-generation AI KPU acceleration module, supporting multi-precision AI computing power. It is equipped with a lightweight Yolov5 model optimized by NNCASE quantization and pruning. The model volume is reduced by 60%, the operation speed is increased by 40%, and the power consumption is lower than 500mW, meeting the requirements of 7×24-hour edge deployment, and taking into account both cost control and high-performance recognition requirements.

[0050] 2. Multi-dimensional and high-precision classification ability: The "lajifenlei.pt" model trained based on a large-scale data model (YOLOv5 model) performs excellently in 11 sub-categories such as hazardous waste (batteries, expired medicines), recyclables (plastic bottles, metal cans), kitchen waste (potatoes, radishes), and other waste (bricks, tiles): ① The accuracy rate reaches 97.3%, significantly higher than traditional rule matching or single-sensor solutions; ② F1 recall rate: 0.980 for batteries, 0.989 for plastic bottles, and both paper cups and metal cans exceed 0.995 (as Figure 7 shown), effectively reducing missed detections and misjudgments; ③ Strong anti-interference ability: Through the above four types of geometric transformation methods and two types of pixel transformation data augmentation strategies, the robustness of the model to complex scenarios such as stain coverage and angle deflection is increased by 30%. As Figure 7 shown, the average mAP@0.5 for all categories reaches 0.989. The recognition accuracy of regular-shaped objects such as paper and metal cans exceeds 0.995. After actual measurement and verification, the ultrasonic sensor has a full-load detection accuracy rate of 98.7% and a false alarm rate < 1.3% in environments such as humidity and light changes.

[0051] 3. Intelligent full-load detection and dynamic warning: Integrate an ultrasonic full-load detection module to achieve real-time quantitative monitoring of the trash can capacity (detection range 0.1 - 5m, accuracy ±2cm). When the garbage loading reaches 90%, the system automatically triggers a red warning on the screen and a voice prompt, and simultaneously generates an operation and maintenance work order (optional IoT module for uploading to the management platform), solving problems such as "high cost of manual inspection" and "risk of full-load overflow" of traditional trash cans, and transforming facility management from "passive response" to "active prevention".

[0052] 4. Real-time interaction and user experience optimization: The display screen simultaneously presents the type of garbage and the status inside the trash can (such as "kitchen waste - carrots, remaining space: 40%"), enhancing the transparency of waste disposal and operation feedback. For elderly or low-educated users, voice broadcast of recognition results is supported (such as "What you have put in is recyclables - plastic bottles"), reducing the usage threshold; the full-load warning function significantly improves the usability and management efficiency of waste sorting facilities through dual visual and auditory prompts.

[0053] 5. Technical integration and scenario adaptation capabilities: ① Multi-sensor collaboration: The visual recognition module is fused with the data of ultrasonic sensors in real time, which not only ensures the classification accuracy but also realizes the monitoring of equipment status, laying a hardware foundation for the subsequent access of weight sensors, odor sensors, etc.; ② Edge computing deployment: The K230 module supports offline operation without relying on the cloud and can still work stably in scenarios with weak network coverage such as communities and parks. Local data processing ensures privacy and security; ③ Model scalability: Based on the open framework of Yolov5, it supports the subsequent addition of new waste categories (such as express packaging and electronic waste), and the model can be quickly iterated through transfer learning to adapt to different regional waste classification standards.

[0054] The present invention is described through preferred embodiments. Those skilled in the art know that various changes or equivalent replacements can be made to these features and embodiments without departing from the spirit and scope of the present invention. The present invention is not limited by the specific embodiments disclosed herein, and other embodiments falling within the scope of the claims of this application belong to the protection scope of the present invention.

Claims

1. Three-stage differential garbage classification and treatment system, characterized in that: It includes a housing (1), a sorting mechanism (2), a compression mechanism (3) and a trash can (4). A feeding opening (11) is provided at the top of the housing (1). A sorting mechanism (2) is provided inside the housing (1). The sorting mechanism (2) includes a three-stage differential separation module (21), a feeding module (22) and a visual recognition module (23). The three-stage differential separation module (21) includes a first conveyor belt (213), a second conveyor belt (214) and a third conveyor belt (215) arranged in sequence from top to bottom. A dispersion roller (211) is provided below the feeding opening (11). The dispersion roller (211) is in transmission connection with the first conveyor belt (213) through a belt (212). One end of the third conveyor belt (215) is provided with a feeding module (22). A visual recognition module (23) is provided above the feeding module (22). Two or more trash cans (4) are provided at the bottom inside the housing (1). A compression mechanism (3) is provided in one of the trash cans (4).

2. The three - stage differential garbage classification and treatment system according to claim 1, characterized in that: The feeding module (22) includes a support beam (221), a first servo motor (222), a second servo motor (223) and a trash shovel (224). The support beam (221) is horizontally arranged below the third conveyor belt (215). A first servo motor (222) is provided on the support beam (221). The power output shaft of the first servo motor (222) is connected to a second servo motor (223). The power output shaft of the second servo motor (223) is connected to a trash shovel (224). And the power output shafts of the first servo motor (222) and the second servo motor (223) are perpendicular to each other.

3. The three-stage differential garbage classification and treatment system according to claim 2, wherein: The visual recognition module (23) includes a camera (231), a main controller (232), a display screen (233) and a voice announcer (234). The camera (231) is provided above the trash shovel (224). The main controller (232) is fixedly arranged on the top of the housing (1). A display screen (233) and a voice announcer (234) are integrated on the main controller (232). And the camera (231), the display screen (233) and the voice announcer (234) are all electrically connected to the main controller (232).

4. The three-stage differential garbage classification and treatment system according to claim 3, characterized in that: An ultrasonic sensor (235) is further provided inside the trash can (4). The ultrasonic sensor (235) is electrically connected to the main controller (232).

5. The three - stage differential garbage classification and treatment system according to claim 1, wherein: The compression mechanism (3) includes an electric push rod (31), a fixed compression plate (32), a movable compression plate (33), a first spring (34) and a conical tip (35). The fixed compression plate (32) and the movable compression plate (33) are both vertically arranged inside the trash can (4). And the fixed compression plate (32) and the movable compression plate (33) are connected by a first spring (34). The electric push rod (31) is provided on the housing (1). And the output end of the electric push rod (31) is connected to the fixed compression plate (32). A conical tip (35) is further provided on the fixed compression plate (32). A through hole (331) matching with the conical tip (35) is provided on the movable compression plate (33).

6. The three-stage differential garbage classification and treatment system according to claim 5, wherein: A liquid collection bin (36) is further provided at the inner bottom of the trash can (4). A hook (37) is provided on the liquid collection bin (36), and the hook (37) is hung on the trash can (4). A filter plate (38) is provided at the top of the liquid collection bin (36), and filter holes are provided on the filter plate (38).

7. The three-stage differential garbage classification and treatment system according to claim 1, characterized in that: Baffles (216) are provided on the first conveyor belt (213), the second conveyor belt (214), and the third conveyor belt (215).

8. The three - stage differential garbage classification and treatment system according to claim 1, characterized in that: Universal wheels (12) are provided at the bottom of the housing (1).

Images