A garbage classification and recognition system based on vision

Abstract

The present invention discloses a garbage classification and identification system based on vision, which relates to the field of garbage disposal, and includes a collection module, a control module, a server module, an execution module, a display module and a carrier module. In the present invention, the control module receives the pictures collected by the collection module, processes the pictures through a cross-platform computer vision library, and returns the pictures to the server module, obtains the classification results through the training model, and returns the processed results to the control module, controls the reduction motor and the delivery component in the execution module to work, the output shaft at the lower end of the reduction motor drives the gear to rotate, so that the garbage identification and classification mechanism drives the garbage to move horizontally, and at the same time, the reduction motor drives the optical sensor, the industrial camera and the near-infrared spectrum sensor to rotate after the transmission, and after the rotation sets the angle, the three perform subsequent collection, realize the multi-angle and multi-mode information collection while realizing the garbage transportation, obtains comprehensive and clear garbage feature information, and improves the accuracy of garbage identification and classification.

Description

Technical Field

[0001] The present invention relates to the technical field of garbage disposal, and in particular to a garbage classification and identification system based on vision. Background Art

[0002] With the rapid development of the economy, the amount of garbage generated is increasing, and the unreasonable disposal of garbage has caused serious damage to resources and the environment. For example, a large amount of recyclable garbage is not effectively recycled, but mixed with other garbage, resulting in a waste of resources; hazardous waste is not properly handled, which will pollute the soil, water and air and endanger people's health. Therefore, reasonable garbage classification is a key measure to solve resource and environmental problems.

[0003] Manual sorting is the most traditional way of sorting garbage, and it is also the most commonly used method. This method is inefficient. Also, due to the wide variety and different forms of garbage, workers may make mistakes in sorting due to fatigue, lack of knowledge, etc.

[0004] And some methods are to use artificial intelligence, computer vision and other technologies to automatically determine the type of garbage and achieve automatic classification.

[0005] For example, Chinese patent CN215625174U discloses an automated garbage classification and identification system based on machine vision, including a manipulator for moving a clamp, a camera for collecting garbage image information, a clamp for moving the garbage, and a trash can for placing the garbage. The clamp is mounted on the manipulator, and the trash can is arranged on one side of the manipulator. The trash can includes a barrel body and a barrel cover, one side of the barrel cover is hingedly connected to the open side of the upper end of the barrel body by a hinge, and a barrel cover opening and closing mechanism for opening and closing the barrel cover is arranged on the barrel body.

[0006] However, in real life, most communities are only equipped with basic facilities such as classified trash cans and trash compartments. These facilities are set up and labeled according to categories such as recyclables, hazardous waste, wet garbage, and other garbage, which is convenient for residents to classify and place them, but there is still a problem of misclassification, and the existing automatic garbage classification and identification system based on machine vision cannot be applied to the community garbage compartment, where the position and angle of the collection camera are fixed, and it is impossible to obtain comprehensive and clear information on the appearance characteristics of the garbage. For different forms of the same garbage and two forms of similar materials, it is very easy to have incomplete collection and lead to identification and classification errors: Take newspapers as an example. A complete newspaper is easily identified as recyclables for collection, but when the newspaper is crumpled into a ball, it is likely to be mistakenly identified as a stone and identified as other garbage for collection; ceramic fragments and glass fragments are very similar in appearance, both are hard and brittle materials, glass is recyclable garbage, and ceramics are other garbage. In the process of garbage classification and identification, they are easily confused, so there is also a problem of misclassification. Summary of the invention

[0007] The purpose of the present invention is to provide a vision-based garbage classification and identification system to solve the problem that the existing automatic garbage classification and identification system based on machine vision proposed in the above background technology cannot be applied to community garbage rooms, wherein the position and angle of the acquisition camera are fixed, and it is impossible to obtain comprehensive and clear garbage appearance feature information. For different forms of the same garbage and two garbages with similar forms and materials, it is very easy to have incomplete collection and lead to identification and classification errors.

[0008] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a garbage classification and identification system based on vision, comprising a collection module, a control module, an execution module, a display module and a carrier module;

[0009] The collection module is used to take pictures of the garbage and collect information, and specifically includes an optical sensor, an industrial camera and a near-infrared spectrum sensor;

[0010] The control module uses the Jetson Nano development board as the main control unit, transmits the garbage photos taken by the acquisition module to the server module, and the server module uses the training model to identify and classify the garbage photos, and the results are returned to the control module. The control module controls the execution module to classify and place the garbage;

[0011] The execution module includes a garbage identification and classification mechanism and a garbage bin. The garbage identification and classification mechanism includes a slide rail assembly, a slide seat assembly, a delivery assembly, a viewing angle rotation assembly, a reduction motor and gears. The delivery assembly is used to receive and deliver the garbage. The optical sensor, the industrial camera and the near-infrared spectrum sensor are all located above the delivery assembly and are installed at intervals below the viewing angle rotation assembly. The viewing angle rotation assembly and the gears are respectively connected to the upper and lower ends of the output shaft of the reduction motor. The reduction motor is fixedly connected between two adjacent groups of the slide seat assemblies. The delivery assembly is slidably connected between two groups of the slide rail assemblies through multiple groups of the slide seat assemblies to drive the garbage to move. At the same time, the viewing angle rotation assembly drives the optical sensor, the industrial camera and the near-infrared spectrum sensor to move in an arc to collect information on the garbage from multiple angles and in multiple ways.

[0012] Preferably, the display module is used for human-computer interaction, and specifically includes a touch display screen;

[0013] The carrier module includes a garbage compartment, the garbage identification and classification mechanism is fixedly installed in the upper middle part of the inner cavity of the garbage compartment, a plurality of groups of garbage bins arranged side by side are arranged in the lower part of the inner cavity of the garbage compartment and are located obliquely below the garbage identification and classification mechanism, and the touch display screen is installed below the feeding port on the outside of the garbage compartment.

[0014] Preferably, the viewing angle rotation assembly includes a top plate, a concave wheel, a sprocket ring, a connecting frame, a chain and a driving sprocket, the optical sensor, the industrial camera and the near-infrared spectrum sensor are spaced apart and fixedly installed on the lower edge of the sprocket ring, a plurality of groups of the concave wheels are circularly distributed and rotatably connected to the upper side of the sprocket ring, and the middle groove of the concave wheel is rollingly engaged with the inner wall of the circular hole in the middle of the top plate.

[0015] Preferably, the sprocket ring is transmission connected to the driving sprocket through the chain, the driving sprocket is rotatably mounted on the lower side of the top plate, an extension rod is fixedly mounted on the middle part of the driving sprocket, and the extension rod is fixedly connected to the upper end of the output shaft of the reduction motor.

[0016] Preferably, the slide rail assembly includes a guide rail and a rack, the rack is fixedly connected to one side of one set of the guide rails, and the gear is meshed with the rack, and the two sets of the guide rails are horizontally fixedly installed in the inner cavity of the garbage compartment.

[0017] Preferably, the top plate is fixedly connected to the middle upper part of one group of the slide seat assemblies through the connecting frame, and the slide seat assembly includes a suspension beam, a side wheel member and a roller member, two groups of the roller members are rotatably installed on both ends of the suspension beam in a vertical posture, four groups of the side wheel members are rotatably installed on both sides of the two ends of the suspension beam in a horizontal posture, and one group of the roller members is located in the middle position of the two groups of the side wheel members at the same end.

[0018] Preferably, the side wheel member and the suspension beam are both rollingly connected in the inner groove of the guide rail, the reduction motor is fixedly installed in the middle position of the two groups of suspension beams, and the delivery assembly is located above the suspension beam.

[0019] Preferably, the delivery assembly includes a servo motor, a conveyor roller, a support frame and a conveyor belt. The support frame is fixedly connected to the middle part of the upper side of the two groups of suspension beams, and the two groups of conveyor rollers are rotatably installed on both sides of the support frame.

[0020] Preferably, the conveyor belt is drivingly connected to the outer annular surfaces of the two groups of conveyor rollers, the servo motor is fixedly installed on the outside of the support frame, and the output end of the servo motor is drivingly connected to one group of the conveyor rollers.

[0021] Preferably, the training model is a residual convolutional neural network model composed of multiple residual network units, and the multiple residual network units are directly connected across layers.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] 1. In the present invention, the control module receives the pictures collected by the optical sensor, industrial camera and near-infrared spectrum sensor in the acquisition module, processes the pictures through the cross-platform computer vision library, and returns them to the server module, classifies the pictures through the training model, obtains the classification results, and returns the processed results to the control module, and controls the reduction motor and the delivery component in the execution module to start working according to the returned results. After the reduction motor is running, the output shaft at its lower end drives the gear to rotate. After the gear engages the rack, the three sets of slide seat components obtain power, and then the garbage identification and classification mechanism drives the garbage to move horizontally along the guide rail. At the same time, the reduction motor drives the viewing angle rotation component to rotate, thereby driving the optical sensor, industrial camera and near-infrared spectrum sensor installed below to rotate. After rotating to a set angle, the optical sensor, industrial camera and near-infrared spectrum sensor perform a second collection, realizing multi-angle and multi-mode information collection while realizing garbage transportation, obtaining comprehensive and clear garbage feature information, and improving the accuracy of garbage identification and classification.

[0024] 2. In the present invention, after the reduction motor is running, its lower end output shaft drives the gear to rotate, and the gear meshes with the rack with a fixed position, so that the gear moves along the teeth of the rack, so that the three sets of slide seat assemblies obtain power and drive the delivery assembly to move, and the upper end output shaft of the reduction motor drives the active sprocket to rotate through the extension rod. After the chain transmission, the sprocket ring rotates in the top plate hole through multiple sets of concave wheels, and drives the optical sensor, industrial camera and near-infrared spectrum sensor installed below to rotate. A set of reduction motors drives the delivery assembly to move and the sprocket ring to rotate at the same time, realizing one source for two purposes, ensuring the synchronous movement of the two, and improving the coordination of the garbage identification and classification mechanism.

[0025] 3. In the present invention, three groups of slide seat assemblies obtain power, and the side wheels and rollers at both ends of the suspension beam roll in the inner grooves of the guide rails, and then the garbage identification and classification mechanism drives the garbage to move horizontally along the guide rails. After the garbage identification and classification are determined, the control module controls the delivery assembly to move to the position of the designated garbage bin by controlling the reduction motor. At this time, the servo motor drives the transmission roller connected to it to rotate. With the cooperation of another transmission roller, the conveyor belt transports the garbage on it into the designated garbage bin to complete the delivery after garbage classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a flow chart of a garbage classification and identification system based on vision according to the present invention;

[0027] Figure 2 It is a three-dimensional schematic diagram of the overall structure of a garbage classification and identification system based on vision of the present invention;

[0028] Figure 3 It is a three-dimensional schematic diagram of the overall structure of a garbage classification and identification system based on vision in the present invention;

[0029] Figure 4 A top view of the internal structure of a garbage classification and identification system based on vision according to the present invention;

[0030] Figure 5 It is a three-dimensional structural schematic diagram of the positional relationship of a garbage identification and classification mechanism in a garbage classification and recognition system based on vision of the present invention;

[0031] Figure 6 It is a three-dimensional structural schematic diagram of a garbage identification and classification mechanism in a garbage classification and recognition system based on vision of the present invention;

[0032] Figure 7 It is a first three-dimensional structural schematic diagram of specific details of a garbage identification and classification mechanism in a garbage classification and identification system based on vision of the present invention;

[0033] Figure 8 A second three-dimensional structural schematic diagram of specific details of a garbage identification and classification mechanism in a garbage classification and identification system based on vision according to the present invention;

[0034] Fig. 9 It is a three-dimensional structural schematic diagram of a viewing angle rotation component in a garbage classification and identification system based on vision of the present invention;

[0035] Fig.10 The present invention is a schematic diagram of the process of bin placement in a vision-based garbage classification and identification system.

[0036] In the figure: 1. garbage compartment; 2. touch screen; 3. garbage identification and classification mechanism; 31. slide rail assembly; 311. guide rail; 312. rack; 32. slide seat assembly; 321. suspension beam; 322. side wheel member; 323. roller member; 33. delivery assembly; 331. servo motor; 332. conveyor roller; 333. support frame; 334. conveyor belt; 34. optical sensor; 35. industrial camera; 36. viewing angle rotation assembly; 361. top plate; 362. concave wheel; 363. sprocket ring; 364. connecting frame; 365. chain; 366. driving sprocket; 37. near-infrared spectrum sensor; 38. reduction motor; 381. extension rod; 39. gear; 4. garbage can. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the implementation regulations described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] Example 1: Reference Figure 1-Figure 10 As shown: a garbage classification and identification system based on vision, including a collection module, a control module, an execution module, a display module and a carrier module. The collection module is used to take pictures of the garbage put in and collect information, and specifically includes an optical sensor 34, an industrial camera 35 and a near-infrared spectrum sensor 37. The control module uses the JetsonNano development board as the main control unit, and transmits the garbage photos taken by the collection module to the server module. The server module uses the training model to identify and classify the garbage photos, and the results are returned to the control module. The control module controls the execution module to classify the garbage;

[0039] The training model is a residual convolutional neural network model ResNet-50, which is composed of multiple residual network units. Multiple residual network units are directly connected across layers, which allows the neural network to increase the network depth by stacking these residual network units, effectively solving the gradient instability problem and improving the network accuracy. The 50-layer network structure in the residual convolutional neural network model ResNet-50 makes its accuracy and speed higher than other networks, which can well meet the system's needs for the recognition and classification of a large number of junk images;

[0040] The execution module includes a garbage identification and classification mechanism 3 and a garbage bin 4. The garbage identification and classification mechanism 3 includes a slide rail assembly 31, a slide assembly 32, a delivery assembly 33, a viewing angle rotation assembly 36, a reduction motor 38 and a gear 39. The delivery assembly 33 is used to receive and deliver the garbage. The optical sensor 34, the industrial camera 35 and the near-infrared spectrum sensor 37 are all located above the delivery assembly 33 and are installed at intervals below the viewing angle rotation assembly 36. The viewing angle rotation assembly 36 and the gear 39 are respectively connected to the upper and lower ends of the output shaft of the reduction motor 38. The reduction motor 38 is fixedly connected between two adjacent sets of slide assemblies 32. The delivery assembly 33 slides through multiple sets of slide assemblies 32. It is connected between two sets of slide rail assemblies 31 and is used to drive the movement of garbage. The slide rail assembly 31 includes a guide rail 311 and a rack 312. At the same time, the viewing angle rotation assembly 36 drives the optical sensor 34, the industrial camera 35 and the near-infrared spectrum sensor 37 to move in an arc to collect information on the garbage from multiple angles and in multiple ways. The display module is used for human-computer interaction, specifically including a touch display screen 2. The carrier module includes a garbage compartment 1. The garbage identification and classification mechanism 3 is fixedly installed in the middle and upper part of the inner cavity of the garbage compartment 1. A plurality of groups of garbage bins 4 arranged side by side are arranged in the lower part of the inner cavity of the garbage compartment 1 and are located obliquely below the garbage identification and classification mechanism 3. The touch display screen 2 is installed below the feeding port on the outside of the garbage compartment 1.

[0041] In this embodiment, a single or similar garbage is placed on the delivery component 33 through the feeding port outside the garbage compartment 1, and the button in the touch display screen 2 is clicked. At this time, the optical sensor 34, the industrial camera 35 and the near-infrared spectrum sensor 37 are awakened as acquisition modules, and the first image acquisition is performed on the garbage on the delivery component 33, wherein the optical sensor 34 collects the color characteristics, glossiness and transparency of the garbage, the industrial camera 35 collects the shape, size information, texture characteristics, identification and label information of the garbage, and the near-infrared spectrum sensor 37 collects the garbage material category information through the specific absorption and reflection spectra of different substances in the near-infrared band. A variety of methods are combined to jointly determine the type of garbage. The Jetson Nano development board as the main control unit in the control module receives the pictures collected by the acquisition module, processes the pictures through the cross-platform computer vision library, and returns the processed pictures to the server module. The server module classifies the pictures through the training model-residual convolutional neural network model ResNet-50 to obtain the classification results, and returns the processed results to the Jetson Nano development board. Nano development board controls the reduction motor 38 and delivery component 33 in the execution module to start working according to the returned result. After the reduction motor 38 is running, its lower output shaft drives the gear 39 to rotate. After the gear 39 engages the rack 312, the three sets of slide seat components 32 obtain power, and then the garbage identification and classification mechanism 3 drives the garbage to move horizontally along the guide rail 311. At the same time, the reduction motor 38 drives the viewing angle rotation component 36 to rotate, thereby driving the optical sensor 34, industrial camera 35 and near-infrared spectrum sensor 37 installed below to rotate. After rotating to a set angle, the optical sensor 34, industrial camera 35 and near-infrared spectrum sensor 37 perform a second collection, realizing multi-angle and multi-mode information collection while transporting garbage, obtaining comprehensive and clear garbage feature information, and improving the accuracy of garbage identification and classification. According to the majority principle, the result with the most identification and classification times in multiple collections is used as the final output result. The control module controls the delivery component 33 to move to the position of the designated garbage bin 4 to deliver garbage, so as to minimize garbage classification errors, and finally display it outward through the touch screen 2 in the display module.

[0042] Embodiment 2: According to Figure 6-Figure 9As shown, the viewing angle rotating assembly 36 includes a top plate 361, a middle concave wheel 362, a sprocket ring 363, a connecting frame 364, a chain 365 and a driving sprocket 366. The optical sensor 34, the industrial camera 35 and the near-infrared spectrum sensor 37 are distributed at intervals and fixedly installed on the lower edge of the sprocket ring 363. Multiple groups of middle concave wheels 362 are distributed in a circle and rotatably connected to the upper side of the sprocket ring 363. The middle groove of the middle concave wheel 362 is rollingly clamped on the inner wall of the circular hole in the middle of the top plate 361. The sprocket ring 363 is connected to the inner wall of the circular hole in the middle of the top plate 361 through The chain 365 is transmission connected to the driving sprocket 366, and the driving sprocket 366 is rotatably installed on the lower side of the top plate 361. An extension rod 381 is fixedly installed in the middle of the driving sprocket 366, and the extension rod 381 is fixedly connected to the upper end of the output shaft of the reduction motor 38. The slide rail assembly 31 includes a guide rail 311 and a rack 312. The rack 312 is fixedly connected to one side of one set of guide rails 311, and the gear 39 is meshed with the rack 312. The two sets of guide rails 311 are horizontally fixedly installed in the inner cavity of the garbage compartment 1.

[0043] In this embodiment, after the reduction motor 38 is running, its lower end output shaft drives the gear 39 to rotate, and the gear 39 meshes with the rack 312 installed on one side of the guide rail 311. The position of the rack 312 is fixed, so that the gear 39 moves along the teeth of the rack 312 and applies force to the reduction motor 38, so that the three sets of slide seat assemblies 32 obtain power and move horizontally along the inner groove of the guide rail 311, and the upper end output shaft of the reduction motor 38 drives the driving sprocket 366 to rotate on the lower side of the top plate 361 through the extension rod 381, and the chain 365 is transmitted to the lower side of the top plate 361. After the movement, the sprocket ring 363 and the multiple sets of concave wheels 362 on its upper side obtain a rotational force, and the concave wheels 362 are stuck on the edge of the hole of the top plate 361, so that the sprocket ring 363 rotates in the hole of the top plate 361, and drives the optical sensor 34, industrial camera 35 and near-infrared spectrum sensor 37 installed below to rotate. After rotating to a set angle, the optical sensor 34, industrial camera 35 and near-infrared spectrum sensor 37 perform a second collection, and identify and classify the garbage pictures according to the first collection method, and then perform multiple collections;

[0044] A group of reduction motors 38 simultaneously drives the delivery assembly 33 to move and the sprocket ring 363 to rotate, achieving dual purpose with one source, ensuring the synchronous movement of the two, and improving the coordination of the garbage identification and classification mechanism 3.

[0045] Embodiment 3: According to Figure 6-Figure 9As shown, the top plate 361 is fixedly connected to the middle part of the upper side of one group of slide assemblies 32 through a connecting frame 364. The slide assembly 32 includes a suspension beam 321, a side wheel member 322 and a roller member 323. Two groups of roller members 323 are respectively rotatably installed at the two ends of the suspension beam 321 in a vertical posture. Four groups of side wheel members 322 are respectively rotatably installed on both sides of the two ends of the suspension beam 321 in a horizontal posture. One group of roller members 323 is located in the middle position of the two groups of side wheel members 322 at the same end. The side wheel member 322 and the suspension beam 321 are both rollingly connected in the inner groove of the guide rail 311. The reduction motor 38 is fixedly installed The delivery assembly 33 is located above the suspension beams 321 in the middle position of the two groups of suspension beams 321. The delivery assembly 33 includes a servo motor 331, a conveying roller 332, a support frame 333 and a conveyor belt 334. The support frame 333 is fixedly connected to the middle of the upper side of the two groups of suspension beams 321. The two groups of conveying rollers 332 are rotatably installed on both sides of the support frame 333 respectively. The conveyor belt 334 is transmission-connected to the outer annular surfaces of the two groups of conveying rollers 332. The servo motor 331 is fixedly installed on the outer side of the support frame 333, and the output end of the servo motor 331 is transmission-connected to one of the groups of conveying rollers 332.

[0046] In this embodiment, the three groups of slide seat assemblies 32 obtain power, and the side wheel members 322 and roller members 323 at both ends of the suspension beam 321 roll in the inner groove of the guide rail 311, and then the garbage identification and classification mechanism 3 drives the garbage to move horizontally along the guide rail 311, and the suspension beam 321 plays a connecting and supporting role. After the garbage identification and classification are determined, the control module controls the delivery assembly 33 to move to the position of the designated garbage bin 4 by controlling the reduction motor 38. At this time, the servo motor 331 drives the conveying roller 332 connected thereto to rotate. With the cooperation of another conveying roller 332, the conveyor belt 334 transports the garbage thereon into the designated garbage bin 4 to complete the garbage delivery.

[0047] The use method and working principle of the device: The garbage compartment 1 is used as a carrier module and installed in a specific area of ​​the community. The interior of the garbage compartment 1 is divided into an identification and classification area for installing a garbage identification and classification mechanism 3, a storage area for accommodating multiple garbage bins 4 placed side by side, and an installation area for a host and auxiliary equipment related to the system;

[0048] Residents put single or similar garbage on the delivery component 33 through the feeding port outside the garbage compartment 1, and click the button on the touch screen 2. At this time, the optical sensor 34, the industrial camera 35 and the near-infrared spectrum sensor 37 are awakened as the acquisition module, and the garbage on the conveyor belt 334 is first imaged. A variety of methods are combined to jointly determine the type of garbage. The Jetson Nano development board as the main control unit in the control module receives the image collected by the acquisition module, processes the image through the cross-platform computer vision library, and returns the processed image to the server module. The server module classifies the image through the training model-residual convolutional neural network model ResNet-50, obtains the classification result, and returns the processed result to the Jetson Nano development board; the Jetson Nano development board controls the reduction motor 38 and the servo motor 331 in the execution module to start working according to the returned result;

[0049] After the reduction motor 38 is running, its lower output shaft drives the gear 39 to rotate, and the gear 39 meshes with the rack 312 installed on one side of the guide rail 311. The position of the rack 312 is fixed, so that the gear 39 moves along the teeth of the rack 312 and applies force to the reduction motor 38, so that the three sets of slide seat assemblies 32 obtain power, and the side wheel parts 322 and roller parts 323 at both ends of the suspension beam 321 roll in the inner groove of the guide rail 311, and then the garbage identification and classification mechanism 3 drives the garbage to move horizontally along the guide rail 311;

[0050] At the same time, the output shaft at the upper end of the reduction motor 38 drives the driving sprocket 366 to rotate at the lower side of the top plate 361 through the extension rod 381. After being transmitted by the chain 365, the sprocket ring 363 and the multiple groups of concave wheels 362 on its upper side obtain the rotation force, and the concave wheels 362 are stuck on the edge of the hole of the top plate 361. As a result, the sprocket ring 363 rotates at the hole of the top plate 361, and drives the optical sensor 34, industrial camera 35 and near-infrared spectrum sensor 37 installed below to rotate. After rotating to a set angle, the optical sensor 34, industrial camera 35 and near-infrared spectrum sensor 37 perform a second collection, and identify and classify the garbage pictures according to the first collection method, and the subsequent collection is carried out in the same way;

[0051] This enables multi-angle and multi-mode information collection while garbage is being transported, obtaining comprehensive and clear garbage feature information and improving the accuracy of garbage identification and classification;

[0052] According to the majority principle, the result with the most identical identification and classification times in multiple collections is used as the final output result. Different output results correspond to different trash cans 4. The control module controls the delivery component 33 to move to the position of the designated trash can 4. At this time, the servo motor 331 drives the conveyor roller 332 connected thereto to rotate. With the cooperation of another conveyor roller 332, the conveyor belt 334 transports the garbage thereon into the designated trash can 4, thereby completing the identification, classification and delivery of the garbage once, and displaying it to the residents on the touch display screen 2. The same operation can be performed for subsequent garbage.

[0053] In actual use, this embodiment can also be installed in the workshop of a garbage disposal enterprise after appropriate changes, and cooperate with conveying mechanisms, robotic arms and other mechanisms to further identify and classify the conveyed garbage, which is beneficial to improving the efficiency of garbage disposal.

[0054] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A garbage classification and identification system based on vision, characterized in that: It includes an acquisition module, a control module, a server module, an execution module, a display module and a carrier module; The collection module is used to take photos of the put-in garbage and collect information, and specifically includes an optical sensor (34), an industrial camera (35) and a near-infrared spectrum sensor (37); The control module transmits the garbage photos taken by the acquisition module to the server module, and the server module uses the training model to identify and classify the garbage photos, and the results are returned to the control module, and the control module controls the execution module to classify and place the garbage; The execution module comprises a garbage identification and classification mechanism (3) and a garbage can (4); the garbage identification and classification mechanism (3) comprises a slide rail assembly (31), a slide seat assembly (32), a delivery assembly (33), a viewing angle rotation assembly (36), a reduction motor (38) and a gear (39); the delivery assembly (33) is used to receive and deliver the garbage; the optical sensor (34), the industrial camera (35) and the near-infrared spectrum sensor (37) are all located above the delivery assembly (33) and are installed at intervals below the viewing angle rotation assembly (36); the viewing angle rotation assembly The gear (36) and the gear (39) are respectively connected to the upper and lower ends of the output shaft of the reduction motor (38); the reduction motor (38) is fixedly connected between two adjacent groups of the slide seat assemblies (32); the delivery assembly (33) is slidably connected between two groups of the slide rail assemblies (31) through multiple groups of the slide seat assemblies (32) to drive the movement of garbage; at the same time, the viewing angle rotation assembly (36) drives the optical sensor (34), the industrial camera (35) and the near-infrared spectrum sensor (37) to move in an arc, so as to collect information on the garbage from multiple angles and in multiple ways; The viewing angle rotation assembly (36) comprises a top plate (361), a concave wheel (362), a sprocket ring (363), a connecting frame (364), a chain (365) and a driving sprocket (366); the optical sensor (34), the industrial camera (35) and the near-infrared spectrum sensor (37) are spaced apart and fixedly mounted on the lower edge of the sprocket ring (363); a plurality of groups of the concave wheels (362) are circularly distributed and rotatably connected to the upper side of the sprocket ring (363); and the groove in the middle of the concave wheel (362) is rollingly engaged with the inner wall of the circular hole in the middle of the top plate (361).

2. The visual-based garbage classification and identification system according to claim 1, characterized in that: The display module is used for human-computer interaction and specifically comprises a touch display screen (2); The carrier module comprises a garbage compartment (1), the garbage identification and classification mechanism (3) is fixedly mounted in the middle and upper part of the inner cavity of the garbage compartment (1), a plurality of groups of garbage bins (4) arranged side by side are arranged in the lower part of the inner cavity of the garbage compartment (1) and are located obliquely below the garbage identification and classification mechanism (3), and the touch display screen (2) is mounted below the outer feeding port of the garbage compartment (1).

3. The visual-based garbage classification and identification system according to claim 1 is characterized in that: The sprocket ring (363) is transmission-connected to the driving sprocket (366) via the chain (365); the driving sprocket (366) is rotatably mounted on the lower side of the top plate (361); an extension rod (381) is fixedly mounted on the middle of the driving sprocket (366); the extension rod (381) is fixedly connected to the upper end of the output shaft of the reduction motor (38).

4. The visual-based garbage classification and identification system according to claim 3 is characterized by: The slide rail assembly (31) comprises a guide rail (311) and a rack (312); the rack (312) is fixedly connected to one side of one set of the guide rails (311), and the gear (39) is meshed with the rack (312); the two sets of guide rails (311) are horizontally fixedly installed in the inner cavity of the garbage compartment (1).

5. The visual-based garbage classification and identification system according to claim 4 is characterized in that: The top plate (361) is fixedly connected to the middle part of the upper side of one group of the slide assembly (32) through the connecting frame (364); the slide assembly (32) comprises a suspension beam (321), a side wheel member (322) and a roller member (323); two groups of the roller members (323) are respectively rotatably mounted on the two ends of the suspension beam (321) in a vertical posture; four groups of the side wheel members (322) are respectively rotatably mounted on both sides of the two ends of the suspension beam (321) in a horizontal posture; and one group of the roller members (323) is located in the middle of the two groups of the side wheel members (322) at the same end.

6. The visual-based garbage classification and identification system according to claim 5, characterized in that: The side wheel member (322) and the suspension beam (321) are both rollingly connected in the inner groove of the guide rail (311), the reduction motor (38) is fixedly mounted in the middle position of the two groups of suspension beams (321), and the delivery assembly (33) is located above the suspension beam (321).

7. The visual-based garbage classification and identification system according to claim 6 is characterized by: The delivery assembly (33) comprises a servo motor (331), a conveying roller (332), a support frame (333) and a conveying belt (334); the support frame (333) is fixedly connected to the middle part of the upper side of the two groups of suspension beams (321); and the two groups of conveying rollers (332) are rotatably mounted on both sides of the support frame (333).

8. The visual-based garbage classification and identification system according to claim 7, characterized in that: The conveyor belt (334) is drivingly connected to the outer annular surfaces of the two groups of conveyor rollers (332), the servo motor (331) is fixedly mounted on the outside of the support frame (333), and the output end of the servo motor (331) is drivingly connected to one group of the conveyor rollers (332).

9. The visual-based garbage classification and identification system according to claim 1, characterized in that: The training model is a residual convolutional neural network model composed of multiple residual network units, and the multiple residual network units are directly connected across layers.

Images