Intelligent garbage classification and compression device based on AI visual identification

Abstract

The invention relates to the technical field of garbage treatment, and discloses an intelligent garbage classification and compression device based on AI visual identification, the intelligent garbage classification and compression device comprises a conveying belt main body, the conveying belt main body is horizontally placed, and a collection box is arranged at the bottom of one end of the conveying belt main body; the AI visual recognition module is used for recognizing garbage and classifying the garbage, and the AI visual recognition module is located above the conveying belt body; the number of the classification boxes is four, the classification boxes are located below the conveying belt, and a compression assembly is arranged above the classification boxes; the sorting assembly is located on the side of the conveying belt body, a sorting disc is arranged on the side, away from the sorting assembly, of the conveying belt body, the sorting assembly corresponds to the sorting disc, and a connecting pipeline communicating with the classification box is arranged on the sorting disc. Unmanned, intelligent and efficient management of garbage throwing is achieved, the classification accuracy and the collection efficiency are greatly improved, and meanwhile the transportation cost and space utilization are optimized through compression treatment.

Description

Technical Field

[0001] This invention relates to the field of waste treatment technology, and in particular to an intelligent waste sorting and compression device based on AI visual recognition. Background Technology

[0003] Traditional household waste sorting methods rely heavily on manual operation, which is not only inefficient but also has an error rate of over 30%, making it difficult to meet the needs of large-scale, routine waste sorting and treatment. In addition, the sorting operation environment is harsh, posing high hygiene and safety risks and putting significant pressure on the labor protection of workers.

[0004] Most existing smart waste sorting equipment on the market can only identify and sort single types of waste, which cannot adapt to the actual usage scenarios of users disposing of mixed waste, and the applicable scenarios are obviously limited. At the same time, most devices only have basic sorting and storage functions and lack compression processing structures for loose recyclables such as plastic bottles and cardboard boxes. This results in extremely low utilization of collection bin capacity, requiring frequent collection operations, which significantly increases the manpower and transportation costs of waste collection. Relevant data shows that uncompressed loose recyclables can increase the number of waste transport vehicles by more than 30%, further driving up the overall cost of waste processing throughout the entire chain.

[0005] To address this, an intelligent waste sorting and compression device based on AI visual recognition is proposed. Summary of the Invention

[0006] The purpose of this invention is to provide an intelligent waste sorting and compression device based on AI visual recognition, which aims to solve or improve at least one of the above-mentioned technical problems.

[0007] To achieve the above objectives, the present invention provides the following solution: The present invention provides an intelligent waste sorting and compression device based on AI visual recognition, including... The conveyor belt body is placed horizontally, and a collection box is provided at the bottom of one end of the conveyor belt body; The AI ​​visual recognition module is used to identify waste and classify it. The AI ​​visual recognition module is located above the conveyor belt body. The sorting bins are provided in four places, located below the conveyor belt, and a compression assembly is provided above the sorting bins; The sorting component is located on the side of the conveyor belt body. A sorting tray is provided on the side of the conveyor belt body away from the sorting component. The sorting component is arranged correspondingly to the sorting tray. A connecting pipe communicating with the sorting box is provided on the sorting tray.

[0008] Preferably, the AI ​​visual recognition module adopts an improved SE-ResNet-50 deep learning network architecture, with a built-in algorithm model integrating channel attention mechanism, supporting real-time recognition of no less than 12 types of common household waste, with a single waste recognition response time ≤0.3s and a recognition accuracy rate ≥95%; the AI ​​visual recognition module includes a main camera unit, an infrared supplementary lighting unit, and an infrared ranging unit.

[0009] Preferably, the sorting assembly includes a mounting plate fixedly disposed on the side of the conveyor belt body, a rotating disk rotatably connected to the mounting plate, a drive motor for rotating the rotating disk fixedly mounted at the bottom of the mounting plate, two vertical plates fixedly connected to the top surface of the mounting plate, the rotating disk located between the two vertical plates, a moving plate slidably connected between the two vertical plates, the moving plate located above the rotating disk, a transverse groove formed in the middle of the moving plate, a push post fixedly connected to the side of the top surface of the rotating disk, the push post slidably located within the transverse groove, and a sorting rod fixedly connected to the side of the moving plate.

[0010] Preferably, a support plate is fixedly installed on the side of the conveyor belt body away from the mounting plate, the sorting tray passes through the support plate and is rotatably connected to the support plate, the top surface of the sorting tray and the top surface of the conveyor belt body are flush, the sorting tray has four through holes, the connecting pipe is located between the through holes and the sorting box, and the connecting pipe is a flexible hose.

[0011] Preferably, a toothed ring is fixedly connected to the bottom periphery of the sorting tray, a rotating motor is fixedly connected to the bottom surface of the support plate, a control gear is fixedly connected to the output shaft of the rotating motor, and the control gear meshes with the toothed ring; the rotating motor and the AI ​​visual recognition module are respectively electrically connected to a controller.

[0012] Preferably, the compression assembly includes a horizontal plate fixedly disposed below the conveyor belt body, a plurality of linear push rod motors are fixedly connected to the horizontal plate, and a pressure plate is fixedly connected to the output end of the linear push rod motors, the pressure plate being disposed corresponding to the sorting box.

[0013] Preferably, a height sensor is fixedly connected inside the sorting box, and the height sensor and the linear actuator motor are electrically connected to the controller.

[0014] The present invention discloses the following technical effects: This invention significantly improves sorting and identification performance; it adopts an improved SE-ResNet-50 algorithm combined with a dual-camera structure with infrared supplementary lighting, which can adapt to complex lighting environments. The identification response time for 12 types of household waste is ≤0.3 seconds, the sorting accuracy rate for single waste is ≥98.7%, and the sorting accuracy rate for mixed waste is ≥92.3%. The sorting mechanism has a sorting efficiency of 15 pieces / minute, which is suitable for the actual use scenario of mixed disposal.

[0015] Space utilization and processing costs are significantly optimized; the compression module can achieve a compression rate of up to 82% for fluffy recyclables, increasing the space utilization of the collection bin by more than 4 times, reducing the frequency of collection by more than 30%, and significantly reducing collection costs; the built-in force feedback foreign object detection system can effectively avoid equipment damage and improve operational safety.

[0016] The equipment boasts outstanding stability and comprehensive benefits; its modular design reduces the difficulty of operation and maintenance, and its dual-redundant servo motor design ensures a sorting fault tolerance rate of ≥99%. The overall power consumption of the equipment is ≤150W, which is 25% lower than that of similar equipment. Each unit can process ≥500kg of waste per day, which can reduce carbon emissions in the entire waste treatment chain and contribute to the construction of "zero-waste cities". Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the intelligent waste sorting and compression device of the present invention; Figure 2 This is a top view of the intelligent waste sorting and compression device of the present invention; Figure 3 This is a schematic diagram of the distribution of the sorting boxes in this invention; The components include: 1. Conveyor belt body; 2. Collection box; 3. AI visual recognition module; 4. Sorting box; 5. Sorting tray; 6. Connecting pipe; 7. Mounting plate; 8. Rotary disc; 9. Vertical plate; 10. Motion plate; 11. Horizontal groove; 12. Actuating column; 13. Sorting rod; 14. Support plate; 15. Through hole; 16. Control gear; 17. Horizontal plate; 18. Linear push rod motor; and 19. Pressure plate. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] Reference Figures 1-3 This invention provides an intelligent waste sorting and compression device based on AI visual recognition, including... Conveyor belt body 1, the conveyor belt body 1 is placed horizontally, and a collection box 2 is provided at the bottom of one end of the conveyor belt body 1; AI visual recognition module 3 is used to identify waste and classify waste. AI visual recognition module 3 is located above the conveyor belt body 1. There are four sorting bins 4. The sorting bins 4 are located below the conveyor belt and a compression component is installed above the sorting bins 4. The sorting component is located on the side of the conveyor belt body 1. A sorting tray 5 is provided on the side of the conveyor belt body 1 away from the sorting component. The sorting component is provided in correspondence with the sorting tray 5. A connecting pipe 6 connected to the sorting box 4 is provided on the sorting tray 5.

[0022] By integrating AI visual recognition, automatic sorting, rotation alignment, and intelligent compression control, the system solves the problems of low efficiency, large errors, poor environment, and the inability of existing equipment to adapt to mixed waste disposal and low storage space utilization in traditional manual sorting, thus achieving automated, intelligent, and efficient classification and compression of household waste.

[0023] Visual recognition combined with infrared ranging improves the reliability of classification in complex scenarios.

[0024] The scheme was further optimized with a sorting tray diameter of 400mm and a crank length of 15mm.

[0025] Furthermore, the training database for the AI ​​visual recognition module 3 contains over 100,000 multi-angle images of garbage, with a labeling accuracy of 99.2%. This provides data support for high-speed, high-accuracy recognition, ensuring stable output of classification results even in complex scenarios.

[0026] Further optimizing the solution, the AI ​​visual recognition module 3 adopts an improved SE-ResNet-50 deep learning network architecture, with a built-in algorithm model integrating channel attention mechanism. It supports real-time recognition of no less than 12 types of common household waste, with a single waste recognition response time of ≤0.3s and a recognition accuracy of ≥95%. The AI ​​visual recognition module 3 includes a main camera unit, an infrared supplementary lighting unit, and an infrared ranging unit. It can still achieve clear imaging in complex lighting environments such as dimness, backlighting, and strong light, improving the environmental adaptability of the device.

[0027] The scheme is further optimized. The sorting component includes a mounting plate 7 fixedly installed on the side of the conveyor belt body 1. A rotating disk 8 is rotatably connected to the mounting plate 7. A drive motor that drives the rotating disk 8 to rotate is fixedly installed at the bottom of the mounting plate 7. Two vertical plates 9 are fixedly connected to the top surface of the mounting plate 7. The rotating disk 8 is located between the two vertical plates 9. A moving plate 10 is slidably connected between the two vertical plates 9. The moving plate 10 is located above the rotating disk 8. A horizontal groove 11 is opened in the middle of the moving plate 10. A push column 12 is fixedly connected to the side of the top surface of the rotating disk 8. The push column 12 slides in the horizontal groove 11. A sorting rod 13 is fixedly connected to the side of the moving plate 10. During operation, the drive motor drives the rotating disk 8 to rotate. The push column 12 moves in a circular motion with the rotating disk 8 and slides in the horizontal groove 11, pushing the moving plate 10 to move in a straight line along the vertical plates 9. This, in turn, drives the sorting rod 13 to complete the pushing action. The structure is simple and reliable, the movement is stable, and continuous and stable high-speed sorting can be achieved. The sorting efficiency is high and the failure rate is low.

[0028] The design is further optimized by fixing a support plate 14 on the side of the conveyor belt body 1 away from the mounting plate 7. The sorting disc 5 passes through the support plate 14 and is rotatably connected to it. The top surface of the sorting disc 5 is flush with the top surface of the conveyor belt body 1. Four through holes 15 are provided on the sorting disc 5. A connecting pipe 6, which is a flexible hose, is located between the through holes 15 and the sorting box 4. The sorting disc 5 can adapt to the rotation angle of the sorting disc 5 to ensure smooth garbage falling without leakage or blockage.

[0029] The scheme is further optimized. A toothed ring is fixedly connected to the bottom periphery of the sorting tray 5, and a rotating motor is fixedly connected to the bottom surface of the support plate 14. The output shaft of the rotating motor is fixedly connected to a control gear 16, which meshes with the toothed ring. The rotating motor and the AI ​​vision recognition module 3 are electrically connected to a controller. The AI ​​vision recognition module 3 transmits the classification signal to the controller, which controls the rotating motor to drive the control gear 16 to rotate the toothed ring and the sorting tray 5, so that the through holes 15 of the corresponding categories are accurately aligned with the sorting position. This achieves rapid and accurate alignment of the sorting tray 5. Combined with the pushing action of the sorting rod 13, the waste is accurately put into the bin, significantly reducing the sorting error rate.

[0030] The scheme was further optimized. The compression component includes a horizontal plate 17 fixedly installed below the conveyor belt body 1. Several linear push rod motors 18 are fixedly connected to the horizontal plate 17. The output end of the linear push rod motors 18 is fixedly connected to a pressure plate 19, and the pressure plate 19 is correspondingly set with the sorting bin 4. The sorting accuracy rate for single waste is 98.7%, and the sorting accuracy rate for mixed waste is 92.3%.

[0031] The compression rate of the plastic bottle is 82%, the compression rate of the cardboard box is 75%, and the compression time for a single compression is 4.5 seconds.

[0032] Further optimization of the scheme involves a height sensor fixedly connected inside sorting bin 4. The height sensor and linear actuator motor 18 are electrically connected to the controller. The height sensor is an ultrasonic sensor, which automatically initiates compression when a threshold (70%) is reached. The intelligent control system's main control chip is an NVIDIA Jetson NX (21 TOPS computing power) + an Arduino Mega 2560 (sensor signal processing). A dynamic optimization algorithm adjusts the sorting strategy in real time based on pressure sensor data to reduce the error rate.

[0033] During operation, waste is continuously conveyed by the conveyor belt body 1. Upon passing through the AI ​​visual recognition module 3, it is quickly identified and its category determined. Based on the recognition result, the controller controls the sorting tray 5 to rotate to the corresponding through-hole 15, subsequently driving the sorting component to move. The sorting rod 13 pushes the waste into the corresponding through-hole 15 of the sorting tray 5, and it falls into the corresponding sorting bin 4 through the connecting pipe 6. A height sensor monitors the waste height in real time; once a threshold is reached, the compression component automatically starts to compact the waste. Unidentifiable waste ultimately falls into the collection bin 2 for centralized processing. The entire process is fully automated, featuring fast recognition, accurate sorting, strong compression, and large storage capacity, significantly improving the efficiency and economy of waste sorting and processing.

[0034] In the description of this invention, it should be understood that the terms "longitudinal", "lateral", "up", "down", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this invention, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention.

[0035] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. For those skilled in the art, other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. An intelligent waste sorting and compression device based on AI visual recognition, characterized in that: include The conveyor belt body (1) is placed horizontally, and a collection box (2) is provided at the bottom of one end of the conveyor belt body (1). AI visual recognition module (3), the AI ​​visual recognition module (3) is used to identify garbage and classify garbage, the AI ​​visual recognition module (3) is located above the conveyor belt body (1); A sorting box (4) is provided, and four sorting boxes (4) are provided. The sorting boxes (4) are located below the conveyor belt, and a compression component is provided above the sorting boxes (4). The sorting component is located on the side of the conveyor belt body (1). A sorting tray (5) is provided on the side of the conveyor belt body (1) away from the sorting component. The sorting component is provided in correspondence with the sorting tray (5). A connecting pipe (6) connected to the sorting box (4) is provided on the sorting tray (5).

2. The intelligent waste sorting and compression device based on AI visual recognition according to claim 1, characterized in that: The AI ​​visual recognition module (3) adopts an improved SE-ResNet-50 deep learning network architecture, with an algorithm model that integrates channel attention mechanism, and supports real-time recognition of no less than 12 types of common household waste. The single waste recognition response time is ≤0.3s and the recognition accuracy is ≥95%. The AI ​​visual recognition module (3) includes a main camera unit, an infrared supplementary light unit and an infrared ranging unit.

3. The intelligent waste sorting and compression device based on AI visual recognition according to claim 1, characterized in that: The sorting assembly includes a mounting plate (7) fixedly installed on the side of the conveyor belt body (1). A rotating disk (8) is rotatably connected to the mounting plate (7). A drive motor that drives the rotating disk (8) to rotate is fixedly installed at the bottom of the mounting plate (7). Two vertical plates (9) are fixedly connected to the top surface of the mounting plate (7). The rotating disk (8) is located between the two vertical plates (9). A moving plate (10) is slidably connected between the two vertical plates (9). The moving plate (10) is located above the rotating disk (8). A horizontal groove (11) is opened in the middle of the moving plate (10). A push column (12) is fixedly connected to the side of the top surface of the rotating disk (8). The push column (12) is slidably located in the horizontal groove (11). A sorting rod (13) is fixedly connected to the side of the moving plate (10).

4. The intelligent waste sorting and compression device based on AI visual recognition according to claim 3, characterized in that: A support plate (14) is fixedly installed on the side of the conveyor belt body (1) away from the mounting plate (7). The sorting tray (5) passes through the support plate (14) and is rotatably connected to the support plate (14). The top surface of the sorting tray (5), the top surface of the sorting tray (5) and the top surface of the conveyor belt body (1) are flush. Four through holes (15) are opened on the sorting tray (5). The connecting pipe (6) is located between the through holes (15) and the sorting box (4). The connecting pipe (6) is a flexible hose.

5. The intelligent waste sorting and compression device based on AI visual recognition according to claim 4, characterized in that: The sorting tray (5) is fixedly connected to a toothed ring on the bottom periphery, and the support plate (14) is fixedly connected to a rotating motor on the bottom surface. The output shaft of the rotating motor is fixedly connected to a control gear (16), and the control gear (16) meshes with the toothed ring. The rotating motor and the AI ​​visual recognition module (3) are respectively electrically connected to a controller.

6. The intelligent waste sorting and compression device based on AI visual recognition according to claim 5, characterized in that: The compression assembly includes a horizontal plate (17) fixedly installed below the conveyor belt body (1). Several linear push rod motors (18) are fixedly connected to the horizontal plate (17). A pressure plate (19) is fixedly connected to the output end of the linear push rod motor (18). The pressure plate (19) is correspondingly set with the sorting box (4).

7. The intelligent waste sorting and compression device based on AI visual recognition according to claim 6, characterized in that: A height sensor is fixedly connected inside the sorting box (4), and the height sensor and the linear push rod motor (18) are electrically connected to the controller.

Images