Urban garbage classification and recovery method and system based on artificial intelligence
By introducing artificial intelligence technology into the garbage classification system, real-time monitoring and identification of littering behavior, combining video image traceability and points incentives, the existing garbage classification system is solved, and the problem of insufficient intelligence and high operating costs is improved. Accuracy and operational efficiency.
Patent Information
- Application Number
- CN202411278307.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-12
- Publication Date
- 2024-12-06
AI Technical Summary
The existing garbage classification system is not intelligent enough, resulting in untimely cleaning of garbage, harsh environment, serious illegal investments among residents, and misinvestment. The operator has few profit points, low user activity and high labor costs.
Adopt urban garbage classification and recycling methods based on artificial intelligence, and use intelligent garbage recycling equipment to monitor and automatically identify personnel in real time, judge and expose littering behavior, combine mixed Gaussian models and finite state machines to identify suspected garbage discarded, collect video images for traceability, Improve accuracy, and stimulate residents' classification awareness through the garbage delivery subsystem, provide points rewards, and optimize clearance and operation and maintenance management.
It improves the accuracy of garbage classification and residents' participation, reduces littering, reduces the cost of cleaning and transportation and operation and maintenance, and improves the efficiency and profit points of garbage classification operation.
Smart Images

Figure CN119079337A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of waste recycling, and in particular to an artificial intelligence-based urban waste classification and recycling method and system. Background Art
[0002] Nowadays, classified trash cans can be seen everywhere in people's daily lives. However, there are still many problems in garbage classification. First of all, the classified trash cans are not intelligent enough, which leads to untimely cleaning of garbage in the cans, poor environment around the cans, and poor experience for residents. Secondly, even though the classified trash cans in some living areas have been semi-intelligent, due to the lack of awareness of garbage classification among residents and inadequate publicity and promotion, it is still very common for residents to throw garbage randomly or throw it in the wrong place. In addition, for the operators of garbage classification, there are still problems such as few profit points, low user activity, and excessive investment in labor costs. Summary of the invention
[0003] The purpose of the present invention is to overcome the deficiencies of the prior art and provide an urban waste classification and recycling method and system based on artificial intelligence.
[0004] The objective of the present invention is achieved through the following technical solutions: In a first aspect, the present invention provides an artificial intelligence-based urban waste classification and recycling method, comprising the following steps: S1. The intelligent garbage collection equipment monitors and automatically identifies people who enter the detection range in real time, and directly logs into the garbage delivery subsystem after successful identification; S2. The personnel interact with the garbage delivery subsystem, which identifies the garbage to be delivered and controls the corresponding garbage bin door to open; S3. After the person completes the delivery, the garbage delivery subsystem displays the relevant information of the garbage delivered by the person on the screen. The intelligent garbage recycling device collects video images of the surrounding environment and identifies the video images of the surrounding environment to determine whether the person is littering. If so, the intelligent garbage recycling device displays and exposes the person littering.
[0005] Furthermore, it includes: the intelligent garbage recycling equipment monitors the garbage in the garbage bin. If the weight, height or odor level of the garbage in the bin reaches a second set threshold, the intelligent recycling equipment generates a cleaning reminder message for the corresponding garbage bin and sends it to the cleaning personnel, who then sort and process the garbage; if there is an abnormality in the intelligent garbage recycling equipment, an abnormality reminder message is generated and sent to the operation and maintenance personnel, who then process the abnormal intelligent garbage recycling equipment.
[0006] Furthermore, in step S3, the intelligent garbage collection device determines whether a person has littered, which specifically includes the following steps: S31, the intelligent garbage recycling device collects and recognizes the surrounding environment image, and models the background information of the video image of the surrounding environment through the improved mixed Gaussian model, specifically including: using the mixed Gaussian model as the initial background model, updating the initial background model according to the pixel points of the video image to obtain the final background model, and modeling the background information of the video image through the final background model; S32, judging whether the person's behavior is suspected of discarding garbage based on the difference between the outputs of the first model and the second model and the finite state machine; S33. After confirming that the person's behavior is suspected of dumping garbage, the person's trajectory is traced back to determine whether the suspected dumping of garbage is actually dumping garbage; S34. If the suspected discarded garbage is actually discarded garbage, the person who discarded the garbage will be warned and exposed according to the preset rules.
[0007] Preferably, in step S31, updating the initial background model according to the pixels of the video image to obtain the final background model comprises the following steps: S311, initializing an initial background model B (x, y), where (x, y) represents the coordinates of the pixel points of the image of the initial background model, 0≤x≤m-1, 0≤y≤n-1, m represents the image pixel width, and n represents the image pixel height; S312: For images ,like (x, y) belongs to B(x, y), which means the image The pixel with coordinates (x, y) belongs to the background of the image, otherwise it belongs to the foreground of the image. (x,y) represents the image The pixel value of the pixel point with coordinates (x, y); S313, if the image The pixel with coordinates (x, y) belongs to the background of the image. (x, y) is used as a training sample to update the initial background model B(x, y); S314, the image Update to image , repeat step S312.
[0008] Preferably, the step S32 includes: using a first learning rate To train the current background model, get the first model , the first model Output the first binary image , using the second learning rate To train the current background model, get the second model , the second model Output the second binary image ,in , , .
[0009] Preferably, the step S32 comprises: To calculate the pixel i of the original image through the first model And the second model The pixel value after , , Indicates that the pixel i of the original image passes through the first model After the pixel, Indicates that the pixel i of the original image passes through the second model The pixels after that are respectively passed through the first model And the second model Detect pixel i and determine the type of pixel i, including: If the first model Detect pixel i as background, the second model If pixel i is detected as background, pixel i is determined to be a background pixel; If the first model Detect pixel i as foreground, the second model If pixel i is detected as background, pixel i is determined to be a suddenly exposed background pixel; If the first model Detect pixel i as background, the second model If pixel i is detected as a foreground, pixel i is determined to be a candidate static foreground pixel; If the first model Detect pixel i as foreground, the second model If pixel i is detected as a foreground, pixel i is determined to be a moving foreground pixel.
[0010] Preferably, the step S32 includes: if the pixel point i is a moving foreground pixel, using a finite state machine to detect whether it is suspected of discarded garbage, specifically including: The moving foreground pixel is used as the trigger state. If the object stops and maintains the first model Detect pixel i as background, second model Detect that pixel i is in the foreground state. When the state is maintained for a preset time , it is judged as suspected garbage disposal.
[0011] Preferably, the step S33 includes the following sub-steps: S331, when it is determined at time t that garbage is suspected to be discarded and there is no moving object within the range D around the suspected garbage, trace back to the first frame and assume that Time is thrown away by garbage, ; S332, in Draw a time window on the image of the frame , in the time and space window Above The center point p of the discarded garbage in the frame image is the center of the circle, defining the time and space range , where r is the radius of the circle centered on p on the current image, is the time range; S333, based on the human body width and height estimation and human body detection algorithm, in the spatiotemporal window If multiple people are detected, the person closest to the center point p is considered to be the one who discarded the garbage. The center point of this person is recorded as ; S334, then Draw a time window for the center of the circle , whose time and space range is , reused in The color features of the people discarding garbage detected in Find the location of the person who discards the garbage within the time and space range, and record the center point of the person who discards the garbage at this time as , and Draw a time window for the center of the circle ; S335, repeat steps S331-S334 until the time of repeating steps S331-S334 exceeds t, and the process is completed. Track the trajectory of people who discard garbage within the scope and trace the source of littering.
[0012] Preferably, step S34 includes: after the tracing of the litterer is completed, warning the litterer includes: if the time of littering exceeds 30 seconds and the distance between the litterer and the litterer exceeds 3m, a warning is issued through voice to urge the litterer to throw the garbage correctly into the trash can; if the litterer ignores the warning, the litterer is marked in the background.
[0013] In a second aspect, the present invention provides an urban waste classification and recycling system based on artificial intelligence, which is used to implement any one of the above-mentioned urban waste classification and recycling methods based on artificial intelligence, including: Intelligent garbage collection equipment, used to collect and store garbage delivered by personnel, and transmit information of garbage delivery subsystem to employee end, supervision platform and server; sensor module is used to detect abnormal information, image acquisition module is used to collect images of garbage delivered by users, and transmit images to server for identification; information acquisition unit is used to collect interaction information of personnel and image information of garbage, and transmit interaction information of personnel and image information of garbage to server for identification; garbage delivery subsystem is used for personnel to deliver garbage; The staff end is used to receive the information transmitted by the intelligent waste recycling equipment, and the staff clears and operates the intelligent waste recycling equipment according to the information; The supervision platform is used to receive information transmitted by the intelligent garbage recycling equipment, monitor, analyze and display garbage placement information and equipment information in various regions, so that management personnel can command and dispatch; it is also used to determine whether the user has misplaced or randomly placed garbage based on whether the image of the garbage dropped by the user collected by the image acquisition module corresponds to the type of garbage bin, and expose the user who has misplaced or randomly placed garbage; The server is provided with a garbage classification identification module to identify the received image information and voice information of the residents, as well as the image information of the garbage, obtain an identification result, and transmit the identification result to the intelligent garbage recycling device.
[0014] The beneficial effects of the present invention are: 1) The present invention adopts the first learning rate To train the current background model, get the first model , using the second learning rate To train the current background model, get the second model , using the first model And the second model The difference in the output can be used to determine whether the stationary foreground object is an object that first moves and then becomes stationary, that is, whether it is suspected discarded garbage, thus avoiding misidentification.
[0015] 2) The present invention combines the trajectory tracing of the person who discarded the garbage to further determine whether the suspected discarded garbage is the real discarded garbage, rather than directly judging it as the real discarded garbage, which can avoid misidentification and improve the accuracy of identifying littering behavior. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 This is a flow chart of an urban waste classification and recycling method based on artificial intelligence according to an embodiment of the present invention; Figure 2 A flowchart of an urban garbage classification and recycling method based on artificial intelligence for judging whether a person has littered in an embodiment of the present invention; Figure 3 The present invention is a schematic structural diagram of an urban waste classification and recycling system based on artificial intelligence. DETAILED DESCRIPTION
[0017] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.
[0018] For example, the present invention provides an artificial intelligence-based urban waste classification and recycling method, the flow chart is as follows: Figure 1 As shown, the following steps are included: S1, the intelligent garbage recycling equipment monitors in real time whether there are residents entering the detection range. If so, the intelligent garbage recycling equipment automatically identifies the identity of the residents and directly logs into the garbage delivery subsystem after successful identification; when someone enters the detection range of the intelligent garbage recycling equipment, the identity of the residents is automatically identified and logged in quickly. The residents can log in to the garbage delivery subsystem without performing additional operations, which is convenient and fast.
[0019] S2, the resident interacts with the garbage delivery subsystem. After the garbage delivery subsystem identifies the type of garbage that the resident wants to deliver, it controls the garbage bin door corresponding to the garbage type to open. Residents can choose the interaction method with the garbage delivery system on their own. Residents can click on the type of garbage to be delivered on the screen of the intelligent garbage recycling device, and the garbage bin door of the corresponding type will automatically open, and will automatically close after the resident completes the delivery. Residents can say the type of garbage to be delivered through voice interaction, and the garbage bin door of the corresponding type will automatically open, and will automatically close after the resident completes the delivery. If the resident does not know the type of garbage to be delivered, the garbage delivery subsystem can assist in identification. Residents can conduct voice questions and answers with the garbage delivery subsystem, and the garbage delivery subsystem will answer the type of garbage to be delivered and automatically open the corresponding door.
[0020] For example, if a resident wants to throw away a banana peel but does not know what type of garbage it belongs to, he or she can have a conversation with the garbage delivery subsystem. For example, if the resident asks the garbage delivery subsystem "I want to throw away a banana peel", the garbage delivery subsystem recognizes the voice of the receiving resident and outputs the recognition result "Banana peels belong to kitchen waste, and the kitchen waste bin door is being opened for you", and controls the kitchen waste bin door to open, and automatically closes the door after the resident completes the delivery. Residents can place the garbage to be delivered under the camera of the intelligent garbage recycling device, and the intelligent garbage recycling device will prompt the resident with the image recognition result through voice or text, and automatically open the corresponding bin door.
[0021] For residents who do not have enough knowledge about garbage classification, they can interact with the garbage delivery subsystem to help them identify the types of garbage. This can not only increase their knowledge of garbage classification and enhance their awareness of garbage classification, but also effectively reduce the occurrence of littering.
[0022] S3, after the residents complete the delivery, the garbage delivery subsystem automatically displays the relevant information of the user's garbage delivery on the screen. If the delivered garbage is non-recyclable garbage, the screen displays the weight of the delivered garbage; if the delivered garbage is recyclable garbage, the screen displays the weight of the delivered garbage and the corresponding reward points; by delivering recyclables, residents can accumulate corresponding points to exchange for cash or goods, etc., which effectively stimulates the enthusiasm of users to participate in garbage sorting. After the residents complete the delivery, the camera of the intelligent garbage recycling equipment will collect the environment image around the equipment, and identify the environment image to determine whether the residents have uncivilized behaviors such as littering. If so, the intelligent garbage recycling equipment will display and expose the uncivilized behavior. If the delivered garbage is non-recyclable garbage, the screen displays the weight of the delivered garbage. By monitoring and exposing uncivilized behaviors such as littering, the occurrence of uncivilized phenomena such as littering can be reduced, thereby enhancing residents' awareness of garbage sorting.
[0023] S4, the internal monitoring device of the intelligent garbage recycling equipment monitors the garbage in the garbage bin. If the weight, height or odor level of the garbage in the bin reaches a second set threshold, the intelligent recycling equipment will send a reminder message of the garbage bin removal to the removal personnel, and the removal personnel will classify and process different types of garbage; if there is an equipment abnormality in the garbage in the bin, the abnormal prompt message will be sent to the operation and maintenance personnel, and the operation and maintenance personnel will perform operation and maintenance on the abnormal equipment.
[0024] Specifically, if the weight, height or odor level of the garbage in the bin reaches a second set threshold, the intelligent recycling device sends a garbage bin removal reminder message to the removal personnel, and the removal personnel classify and process different types of garbage, specifically including: The sensor inside the intelligent recycling device monitors the weight, height or odor level of the garbage in the barrel. If the weight or height reaches a first set threshold, the compression module inside the intelligent recycling device compresses the garbage in the barrel. If the odor level reaches the first set threshold, the cleaning module inside the intelligent recycling device disinfects and deodorizes the garbage in the barrel. When the weight, height or odor of the compressed garbage reaches a second set threshold, the intelligent recycling device sends the garbage bin removal prompt information to the user terminal of the platform user with removal authority, and the platform user takes the order. If the order is successful, the platform user goes to the intelligent recycling device to be removed for removal according to the optimal route and time limit generated by the user terminal. The removal prompt information includes: removal equipment information, garbage removal information; If no one takes the order, the garbage bin removal reminder information will be sent to the employee terminal of the removal personnel. After receiving the removal reminder information at the employee terminal, the removal personnel will go to the smart recycling equipment to be removed for removal according to the optimal route and time limit generated by the employee terminal.
[0025] Optionally, residents can also make an appointment for door-to-door recycling on the user side. The system will send the door-to-door recycling information to the user side of the platform user with the transportation authority, and the platform user will grab the order. If the order is successful, the platform user will go to the resident's home for door-to-door recycling according to the optimal route and time limit generated by the user side. If no one grabs the order, the door-to-door recycling information will be sent to the employee side of the transportation personnel. After the transportation personnel receive the transportation prompt information on the employee side, they will go to the resident's home for door-to-door recycling according to the optimal route and time limit generated by the employee side. The door-to-door recycling information includes: resident location information, garbage collection information, and appointment time information.
[0026] After completing the cleaning work, the cleaning personnel / platform users with cleaning authority upload the cleaning completion information on the employee terminal / user terminal. The cleaning completion information includes the garbage type, garbage weight, cleaning time, and images of the cleaning process.
[0027] For users who meet the removal conditions, they can also complete the removal task, which can not only reduce the pressure of removal work, improve the removal efficiency, but also effectively reduce the cost of removal personnel and equipment. At the same time, providing appointment door-to-door recycling services can bring convenience to residents' lives and increase their enthusiasm for garbage sorting.
[0028] Among them, if there is an abnormality in the equipment of the garbage in the box, the abnormal prompt information is sent to the operation and maintenance personnel, and the operation and maintenance personnel perform operation and maintenance on the abnormal equipment, specifically including: the sensor inside the intelligent recycling equipment detects that the intelligent recycling equipment is abnormal, and then sends the abnormal prompt information to the employee end of the operation and maintenance personnel. The abnormal prompt information includes: equipment information and abnormal information. After receiving the abnormal prompt information, the operation and maintenance personnel go to the abnormal intelligent recycling equipment for equipment maintenance according to the optimal route and limited time generated by the employee end. After completing the maintenance work, the operation and maintenance personnel upload the operation and maintenance completion information on the employee end. The operation and maintenance completion information includes operation and maintenance information, operation and maintenance time, and operation and maintenance process images.
[0029] Exemplarily, the haulage personnel transport the recyclable garbage to a secondary sorting center, which performs image recognition on the recyclable garbage, and after identifying different types of recyclable garbage, performs secondary sorting on the different types of recyclable garbage.
[0030] Specifically, after the recyclable garbage is transported to the secondary sorting center, it is placed on the conveying unit of the secondary sorting center, and the classification and identification unit of the secondary sorting center identifies the category of the recyclable garbage. After identifying the category of the recyclable garbage, the sorting unit of the secondary sorting center sorts the identified recyclable garbage into a containing device of the corresponding category. A weighing unit is provided at the bottom of the containing device, and the weighing unit weighs the recyclable garbage in the containing device. When the weight reaches the third set threshold, the compression unit compacts the recyclable garbage in the containing device. When the weight reaches the fourth set threshold, it is automatically packaged.
[0031] By performing secondary sorting of recyclable garbage through an intelligent secondary sorting center, labor costs can be effectively reduced and the work efficiency of secondary sorting can be improved. Different types of recyclable garbage after secondary sorting can be treated differently, and efficient resource recovery and utilization can also be achieved. The cleaning personnel transport the old clothes sorted out by the secondary sorting center to the old clothes processing center for cleaning and packaging. The cleaned old clothes are classified and packaged, and sold to old clothes recycling organizations, thereby realizing resource reuse and obtaining benefits, and effectively reducing the problem of few profit points for garbage sorting operators.
[0032] For example, in an artificial intelligence-based urban garbage classification and recycling method provided by the present invention, a flowchart for determining whether a person has littered is as follows: Figure 2 As shown, the following steps are included: S31, the intelligent garbage recycling device collects and recognizes the surrounding environment image, and models the background information of the video image of the surrounding environment through the improved mixed Gaussian model, specifically including: using the mixed Gaussian model as the initial background model, updating the initial background model according to the pixel points of the video image to obtain the final background model, and modeling the background information of the video image through the final background model; S32, judging whether the person's behavior is suspected of discarding garbage based on the difference between the outputs of the first model and the second model and the finite state machine; S33. After confirming that the person's behavior is suspected of dumping garbage, the person's trajectory is traced back to determine whether the suspected dumping of garbage is actually dumping garbage; S34. If the suspected discarded garbage is actually discarded garbage, the person who discarded the garbage will be warned and exposed according to the preset rules.
[0033] Specifically, in step S31, updating the initial background model according to the pixels of the video image to obtain the final background model includes the following steps: S311, initializing an initial background model B (x, y), where (x, y) represents the coordinates of the pixel points of the image of the initial background model, 0≤x≤m-1, 0≤y≤n-1, m represents the image pixel width, and n represents the image pixel height; S312: For images ,like (x, y) belongs to B(x, y), which means the image The pixel with coordinates (x, y) belongs to the background of the image, otherwise it belongs to the foreground of the image. (x,y) represents the image The pixel value of the pixel point with coordinates (x, y); S313, if the image The pixel with coordinates (x, y) belongs to the background of the image. (x, y) is used as a training sample to update the initial background model B(x, y); S314, the image Update to image , repeat step S312.
[0034] Specifically, the step S32 includes: using a first learning rate To train the current background model, get the first model , the first model Output the first binary image , using the second learning rate To train the current background model, get the second model , the second model Output the second binary image ,in , , The larger the learning rate of the mixed Gaussian model, the slower the mixed Gaussian model adapts to new samples, the slower the update speed of the mixed Gaussian model, and the less sensitive it is to static foreground pixels. In other words, the foreground object can only be judged as the background when it stays for a long time. The smaller the learning rate of the mixed Gaussian model, the faster the mixed Gaussian model adapts to new samples, the faster the update speed of the mixed Gaussian model, and the more sensitive it is to static foreground pixels. In other words, when the foreground object stays for a short time, it can be judged as the background.
[0035] Specifically, the step S32 includes: To calculate the pixel i of the original image through the first model And the second model The pixel value after , , Indicates that the pixel i of the original image passes through the first model After the pixel, Indicates that the pixel i of the original image passes through the second model The pixels after that are respectively passed through the first model And the second model Detect pixel i and determine the type of pixel i, including: If the first model Detect pixel i as background, the second model If pixel i is detected as background, pixel i is determined to be a background pixel; If the first model Detect pixel i as foreground, the second model If pixel i is detected as background, pixel i is determined to be a suddenly exposed background pixel; If the first model Detect pixel i as background, the second model If pixel i is detected as a foreground, pixel i is determined to be a candidate static foreground pixel; If the first model Detect pixel i as foreground, the second model If pixel i is detected as a foreground, pixel i is determined to be a moving foreground pixel.
[0036] Preferably, the step S32 includes: if the pixel point i is a moving foreground pixel, using a finite state machine to detect whether it is suspected of discarded garbage, specifically including: The moving foreground pixel is used as the trigger state. If the object stops and maintains the first model Detect pixel i as background, second model Detect that pixel i is in the foreground state. When the state is maintained for a preset time , it is judged as suspected garbage disposal.
[0037] Specifically, after determining whether the current stationary foreground object is an object that moves first and then becomes stationary, that is, whether it is suspected discarded garbage, it is combined with the trajectory tracing of the person who discarded it to further determine whether the suspected discarded garbage is real discarded garbage, rather than directly determining it as real discarded garbage. This can avoid misidentification, because the current suspected discarded garbage may be someone temporarily placing something here, and then returning to take it away or delivering it to the trash can. If the person who discarded the suspected discarded garbage does not return after leaving the suspected discarded garbage, the suspected discarded garbage can be determined to be real discarded garbage. Therefore, the use of the above technical means can improve the accuracy of identifying littering behavior, and the step S33 includes the following sub-steps: S331, when it is determined at time t that garbage is suspected to be discarded and there is no moving object within the range D around the suspected garbage, trace back to the first frame and assume that Time is thrown away by garbage, ; S332, in Draw a time window on the image of the frame , in the time and space window Above The center point p of the discarded garbage in the frame image is the center of the circle, defining the time and space range , where r is the radius of the circle centered on p on the current image, For time range ; S333, based on the human body width and height estimation and human body detection algorithm, in the spatiotemporal window If multiple people are detected, the person closest to the center point p is considered to be the one who discarded the garbage. The center point of this person is recorded as ; S334, then Draw a time window for the center of the circle , whose time and space range is , reused in The color features of the people discarding garbage detected in Find the location of the person who discards the garbage within the time and space range, and record the center point of the person who discards the garbage at this time as , and Draw a time window for the center of the circle ; S335, repeat steps S331-S334 until the time of repeating steps S331-S334 exceeds t, and the process is completed. Track the trajectory of people who discard garbage within the scope and trace the source of littering.
[0038] Specifically, step S34 includes: after the tracing of the litterer is completed, warning the litterer includes: if the time of littering exceeds 30 seconds and the distance between the litterer and the litterer exceeds 3m, a warning is issued through voice to urge the litterer to throw the garbage correctly into the trash can; if the litterer ignores the warning, the litterer is marked in the background.
[0039] Exemplarily, the present invention also includes an artificial intelligence-based urban waste classification and recycling system, the structural diagram of which is shown in FIG. Figure 3 As shown, including: Intelligent waste recycling equipment, user end, employee end, supervision platform, server, secondary sorting center, and old clothes processing center.
[0040] Intelligent garbage recycling equipment is used to collect and store garbage delivered by residents, and transmit information from the garbage delivery subsystem to the user end, employee end, supervision platform, and server.
[0041] The intelligent garbage recycling equipment includes: a garbage bin, an information collection unit, an information transmission unit, an advertisement delivery unit, a garbage delivery subsystem, a screen, and a speaker.
[0042] The garbage bin body comprises a plurality of garbage bins for containing different types of garbage, a sensor module, a compression module and an image acquisition module.
[0043] The sensor module is arranged in the box, including: a weight sensor, a height sensor, a temperature sensor, a humidity sensor, a smoke sensor, an odor sensor, a water immersion sensor, an overflow sensor, etc., which is used to detect whether there is abnormal information such as weight, height, temperature, humidity, smoke, odor, water immersion, overflow, etc. in the garbage in the box.
[0044] The compression module is used to compress the garbage when the overflow sensor detects that the garbage in the box has reached a preset height or weight, so as to buy more time for the garbage removal personnel.
[0045] The image acquisition module collects images of the garbage thrown by the user in the bin and transmits the images to the server for identification to determine whether the type of garbage thrown by the user corresponds to the type of the garbage bin.
[0046] The information collection unit is used to collect the image information, voice information and image information of the residents and the image information of the garbage, and transmit the image information, voice information and image information of the residents and the garbage to the server through the information transmission unit for identification.
[0047] The advertisement delivery unit is used to deliver merchant advertisements; the garbage delivery subsystem is used for residents to deliver garbage; the screen is used to interact with users; and the loudspeaker is used to play voice information.
[0048] The user end is used to receive the information transmitted by the intelligent garbage recycling equipment. Users can find the garbage recycling station they want to deliver according to the information of the intelligent garbage recycling equipment. They can also check their points information after the garbage delivery is completed and redeem points. Users can use points to spend recharge, exchange offline goods, exchange cash, collect prizes, charitable donations, etc., to improve the current situation of single points acquisition and use methods. The user end also includes functions such as environmental protection consultation, dynamic publishing, and classified knowledge, so that users can keep abreast of community environmental protection activities and surrounding dynamics, and increase user usage. Residents can also make appointments for door-to-door recycling through the user end. For platform users who meet the conditions for transportation, they can also grab orders for transportation through the user end. After completing the transportation task, rewards are returned to the platform users who meet the transportation conditions through the user end. The user end is also used to publish advertising information.
[0049] The staff end is used to receive the information transmitted by the garbage recycling station, and the staff clears and operates the garbage recycling station according to the information of the garbage recycling station. The users include clearing personnel and operation and maintenance personnel, and different permissions are given to these two types of personnel. For clearing personnel, they can receive information about overflowing and odor in the garbage bin on the staff end, clear the garbage bin within the specified time according to the equipment information, and receive door-to-door recycling appointment information from residents and collect the garbage at the door within the specified time, and upload the type, time, and weight information of the cleared garbage on the staff end. View the relevant information and current status of the equipment, inspect the garbage collection status of each point and report it, and receive the work assignment of the superior supervisor. It can also be used to record the clearing work trajectory of the clearing personnel. For operation and maintenance personnel, they can receive abnormal operation alarm information of the equipment, perform operation and maintenance within the specified time according to the equipment information, inspect whether the equipment at each point is normal and report it, and receive the work assignment of the superior supervisor. It can also be used to record the operation and maintenance work trajectory of the operation and maintenance personnel.
[0050] The supervision platform is used to receive the information transmitted by the intelligent garbage recycling equipment, monitor, analyze and display the garbage delivery information, equipment information, user information and employee information, secondary sorting center and used clothing processing center in various levels of regions, so that managers can command and dispatch. The supervision platform can also compare the type, time and weight information of the garbage to be transported uploaded by the transporter with the type, time and weight information of the garbage to be transported recorded by the intelligent garbage recycling equipment to determine whether the garbage transportation information uploaded by the transporter is true and accurate. The supervision platform can also compare the images of the transporter before and after the transport collected by the image acquisition module to determine whether the transporter's transportation work records are true and accurate. The supervision platform can also determine whether the user has misplaced or randomly thrown garbage based on whether the image of the garbage delivered by the user collected by the image acquisition module corresponds to the type of garbage bin, and punish the user who misplaced or randomly threw garbage, including penalty measures such as deducting points. The supervision platform includes: a monitoring module, an analysis module, and a large visualization screen. Among them, the monitoring module is used to obtain the information transmitted by the intelligent garbage recycling equipment; the analysis module is used to analyze the information transmitted by the intelligent garbage recycling equipment; the large visual screen is used to display the data of the monitoring module and the analysis module for management personnel to command and dispatch.
[0051] The server, on which the garbage classification recognition model is deployed, recognizes the received image information, voice information of the residents, and image information of the garbage to obtain a recognition result, and transmits the recognition result to the intelligent garbage recycling device.
[0052] The secondary sorting center includes a conveying unit, a weighing unit, a classification and identification unit, a sorting unit, and a compression unit. The conveying unit is used to convey the recyclable garbage transported to the secondary sorting center. The classification and identification module is used to classify and identify the recyclable garbage. The sorting module is used to sort different types of recyclable garbage into corresponding storage devices. The weighing unit is used to weigh each type of recyclable garbage after sorting. The compression unit is used to compact the recyclable garbage after sorting.
[0053] After being transported to the secondary sorting center, the recyclable garbage is placed on the conveying unit. The classification identification unit and the sorting unit are arranged above the conveying unit. After identifying the category of the recyclable garbage, the classification identification unit controls the sorting unit to sort the identified recyclable garbage into the corresponding category of the containing device. A weighing unit is provided at the bottom of the containing device. The weighing unit weighs the recyclable garbage in the containing device, and the compression unit compacts the recyclable garbage therein. The used clothes processing center includes a cleaning unit and a packaging unit. The cleaning unit is used to clean and disinfect used clothes. The packaging unit is used to automatically package and process the clothes after cleaning and disinfection.
[0054] From the perspective of the supervision platform, managers can intuitively see the capacity of the smart trash cans and the location information of employees on the website, so as to assign employees to handle the garbage in a more timely and convenient manner. When the garbage is thrown into the trash can, it will be automatically classified and the data will be transmitted to the management end. Managers can better formulate plans based on the big data view to respond to the country's garbage classification policy.
[0055] From the employee's perspective, employees can directly view the area they are responsible for, the status of the trash cans and their own location information on their mobile phones, turning a patrol work method into a work method that improves efficiency and reduces unnecessary labor.
[0056] From the user's perspective, when users open the WeChat mini program, they can make an appointment for door-to-door recycling, learn about their points information, environmental information, learn environmental knowledge, use points to redeem goods, etc. This can not only enhance residents' awareness of garbage classification, but also increase their enthusiasm for protecting the environment.
[0057] The above is only a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the concept described herein through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art do not deviate from the spirit and scope of the present invention, and should be within the scope of protection of the claims attached to the present invention.
Claims
1. A method for urban waste classification and recycling based on artificial intelligence, characterized in that: The following steps are involved: S1. The intelligent garbage collection equipment monitors and automatically identifies people who enter the detection range in real time, and directly logs into the garbage delivery subsystem after successful identification; S2. The personnel interact with the garbage delivery subsystem, which identifies the garbage to be delivered and controls the corresponding garbage bin door to open; S3. After the person completes the delivery, the garbage delivery subsystem displays the relevant information of the garbage delivered by the person on the screen. The intelligent garbage recycling device collects video images of the surrounding environment and identifies the video images of the surrounding environment to determine whether the person is littering. If so, the intelligent garbage recycling device displays and exposes the person littering.
2. The method for urban waste classification and recycling based on artificial intelligence according to claim 1 is characterized in that: include: The intelligent garbage recycling device monitors the garbage in the garbage bin. If the weight, height or odor level of the garbage in the bin reaches a second set threshold, the intelligent recycling device generates a corresponding garbage bin cleaning reminder message and sends it to the cleaning personnel, who then classify and process the garbage. If there is an abnormal situation in the intelligent garbage recycling equipment, an abnormal prompt message will be generated and sent to the operation and maintenance personnel, who will handle the abnormal intelligent garbage recycling equipment.
3. The method for urban waste classification and recycling based on artificial intelligence according to claim 1 is characterized by: In step S3, the intelligent garbage collection device determines whether a person litters or not, specifically comprising the following steps: S31, the intelligent garbage recycling device collects and recognizes the surrounding environment image, and models the background information of the video image of the surrounding environment through the improved mixed Gaussian model, specifically including: using the mixed Gaussian model as the initial background model, updating the initial background model according to the pixel points of the video image to obtain the final background model, and modeling the background information of the video image through the final background model; S32, judging whether the person's behavior is suspected of discarding garbage based on the difference between the outputs of the first model and the second model and the finite state machine; S33. After confirming that the person's behavior is suspected of dumping garbage, the person's trajectory is traced back to determine whether the suspected dumping of garbage is actually dumping garbage; S34. If the suspected discarded garbage is actually discarded garbage, the person who discarded the garbage will be warned and exposed according to the preset rules.
4. The method for urban waste classification and recycling based on artificial intelligence according to claim 3 is characterized in that: In step S31, the initial background model is updated according to the pixels of the video image to obtain the final background model, which includes the following steps: S311, initializing an initial background model B (x, y), where (x, y) represents the coordinates of the pixel points of the image of the initial background model, 0≤x≤m-1, 0≤y≤n-1, m represents the image pixel width, and n represents the image pixel height; S312: For images ,like (x, y) belongs to B(x, y), which means the image The pixel with coordinates (x, y) belongs to the background of the image, otherwise it belongs to the foreground of the image. (x,y) represents the image The pixel value of the pixel point with coordinates (x, y); S313, if the image The pixel with coordinates (x, y) belongs to the background of the image. (x, y) is used as a training sample to update the initial background model B(x, y); S314, the image Update to image , repeat step S312.
5. The method for urban waste classification and recycling based on artificial intelligence according to claim 4 is characterized in that: The step S32 includes: using a first learning rate To train the current background model, get the first model , the first model Output the first binary image , using the second learning rate To train the current background model, get the second model , the second model Output the second binary image ,in , , .
6. The method for urban waste classification and recycling based on artificial intelligence according to claim 5 is characterized in that: The step S32 comprises: To calculate the pixel i of the original image through the first model And the second model The pixel value after , , Indicates that the pixel i of the original image passes through the first model After the pixel, Indicates that the pixel i of the original image passes through the second model The pixels after that are respectively passed through the first model And the second model Detect pixel i and determine the type of pixel i, including: If the first model Detect pixel i as background, the second model If pixel i is detected as background, pixel i is determined to be a background pixel; If the first model Detect pixel i as foreground, the second model If pixel i is detected as background, pixel i is determined to be a suddenly exposed background pixel; If the first model Detect pixel i as background, the second model If pixel i is detected as a foreground, pixel i is determined to be a candidate static foreground pixel; If the first model Detect pixel i as foreground, the second model If pixel i is detected as the foreground, then pixel i is determined to be a moving foreground pixel.
7. The method for urban waste classification and recycling based on artificial intelligence according to claim 6 is characterized in that: The step S32 includes: if the pixel point i is a moving foreground pixel, using a finite state machine to detect whether it is suspected of discarded garbage, specifically including: The moving foreground pixel is used as the trigger state. If the object stops and maintains the first model Detect pixel i as background, second model Detect that pixel i is in the foreground state. When the state is maintained for a preset time , it is judged as suspected garbage disposal.
8. The method for urban waste classification and recycling based on artificial intelligence according to claim 7 is characterized in that: The step S33 includes the following sub-steps: S331, when it is determined at time t that garbage is suspected to be discarded and there is no moving object within the range D around the suspected garbage, trace back to the first frame and assume that Time is thrown away by garbage, ; S332, in Draw a time window on the image of the frame , in the time and space window Above The center point p of the discarded garbage in the frame image is the center of the circle, defining the time and space range , where r is the radius of the circle centered on p on the current image, is the time range; S333, based on the human body width and height estimation and human body detection algorithm, in the spatiotemporal window If multiple people are detected, the person closest to the center point p is considered to be the one who discarded the garbage. The center point of this person is recorded as ; S334, then Draw a time window for the center of the circle , whose time and space range is , reused in The color features of the people discarding garbage detected in Find the location of the person who discards the garbage within the time and space range, and record the center point of the person who discards the garbage at this time as , and Draw a time window for the center of the circle ; S335, repeat steps S331-S334 until the time of repeating steps S331-S334 exceeds t, and the process is completed. Track the trajectory of people who discard garbage within the scope and trace the source of littering.
9. The method for urban waste classification and recycling based on artificial intelligence according to claim 8 is characterized in that: The step S34 includes: after the tracing of the litterer is completed, warning the litterer includes: if the time of littering exceeds 30 seconds and the distance between the litterer and the litterer exceeds 3m, a warning is issued through voice to urge the litterer to throw the garbage into the trash can correctly; if the litterer ignores the warning, the litterer is marked in the background.
10. An artificial intelligence-based urban waste classification and recycling system, used to implement an artificial intelligence-based urban waste classification and recycling method according to any one of claims 1 to 9, characterized in that: include: Intelligent garbage collection equipment is used to collect and store garbage delivered by personnel, and transmit information of the garbage delivery subsystem to the employee end, supervision platform, and server; The sensor module is used to detect abnormal information, and the image acquisition module is used to collect images of garbage delivered by users and transmit the images to the server for identification; An information collection unit, used to collect interaction information of personnel and image information of garbage, and transmit the interaction information of personnel and image information of garbage to a server for identification; Garbage delivery subsystem, used for personnel to deliver garbage; The staff end is used to receive the information transmitted by the intelligent waste recycling equipment, and the staff clears and operates the intelligent waste recycling equipment according to the information; A supervision platform is used to receive information transmitted by the intelligent garbage recycling equipment, monitor, analyze and display garbage placement information and equipment information in various areas, so that management personnel can conduct command and dispatch; It is also used to determine whether the user has misplaced or randomly placed the garbage according to whether the image of the garbage dropped by the user collected by the image acquisition module corresponds to the type of the garbage bin, and to expose the user who has misplaced or randomly placed the garbage; The server is provided with a garbage classification identification module to identify the received image information and voice information of the residents, as well as the image information of the garbage, obtain an identification result, and transmit the identification result to the intelligent garbage recycling device.
Citation Information
Patent Citations
Intelligent trash can
CN104058202A
Remnant tracing detection and alarm method
CN111062273A
Illegal garbage discarding behavior detection method and device and application
CN113435419A
Intelligent garbage treatment method and system and storage medium
CN117088015A
Intelligent garbage classification and collection equipment and system
CN211365766U
