Waste management system and method

The method employs machine learning to classify waste items and provide disposal instructions, addressing unclear recycling rules and damaged item challenges, thereby improving waste management efficiency and user compliance.

GB2630671APending Publication Date: 2024-12-04LANDAWARE LTD
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
GB2024004896
Authority / Receiving Office
GB · GB
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-04-06
Filing Date
2024-04-05
Publication Date
2024-12-04

AI Technical Summary

Technical Problem

Current waste management systems face challenges in educating users on proper recycling due to unclear disposal instructions, varying recycling rules by location, and difficulties in identifying waste items, especially when they are damaged or lack barcodes, leading to incorrect disposal and inefficiencies in resource allocation.

Method used

A method using machine learning to classify waste items through image comparison with a database, providing users with accurate disposal instructions, and utilizing a server or local database to determine suitable receptacles, with features like NFC and QR code reading, and user reward systems to promote correct disposal practices.

Benefits of technology

Enhances user education on proper waste disposal, reduces incorrect recycling, optimizes waste receptacle usage, and encourages responsible behavior through accurate identification and efficient resource management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

There is provided a system and method of instructing a user how to dispose of a waste item. The location of a waste receptacle is identified 402, and an image of the waste item is captured 404 with a user device. The waste item is classified 408, either locally or by a server using a database, and disposal information is displayed 412 to a user. The system includes remote devices determining their location and a server. The server downloads one of a plurality of different image databases to each remote device in response to the location of the remote device, to enable classification of a waste item. A method also includes identifying a waste receptacle with a user device, determining at least a class of waste items that may be placed in the waste receptacle, and informing the user, of waste items that may be placed in the receptacle. A method of determining a schedule for emptying of a plurality of waste receptacles by building and applying a model is also provided.
Need to check novelty before this filing date? Find Prior Art

Description

The present invention relates to waste management systems and methods, in particular using machine learning. Waste items are an increasing problem in terms of environmental damage but recycling efforts are not yet making an adequate impact to protect our oceans and other areas. One problem is that people are unaware of whether something can be recycled and, if so, how. While much packaging includes disposal instructions these can be hard to interpret and the ability to recycle a particular item can vary in different locations. This complexity can drive users to either discard items in general waste or to use an inappropriate recycling channel. The problem is compounded when the waste item comprises litter. These items may be torn, crushed, dirty, incomplete or otherwise damaged. Even if the recycling instructions for the item are intact, those removing the litter may be reluctant to try to read them and / or touch the waste item for any longer than absolutely necessary. Some items of litter have no recycling instructions at all. Similarly, it is not possible to identify the waste item from a barcode or QR code (or similar) either because the item does not carry such a code or it has been damaged so as to be unreadable. It is an object of the present invention to ameliorate these difficulties. According to a first aspect of the present invention, there is provided a method of instructing a user how to dispose of a waste item, the method comprising: determining a location of a waste receptacle, capturing an image of a waste item, classifying the waste item, and conveying disposal information to a user. In one embodiment, determining the location of a waste receptacle comprises determining the location of a user device. This provides an accurate location but may not be possible due to lack of GPS signal or appropriate hardware in the user device. Accordingly, in another embodiment determining the location of a waste receptacle comprises capturing information associated with the waste receptacle. Preferably capturing information associated with the waste receptacle comprises reading a barcode, a QR code or Near Field Communications (NFC) data associated with the receptacle. In one embodiment, the steps of capturing an image of a waste item and classifying the waste item preferably comprise an instant image comparison with a database. In an alternative embodiment, classifying the waste item comprises comparing an image of the waste item with a plurality of images in a database. The database is preferably generated by machine learning (ML). The database may be located at a server in which case the user device transmits at least one image to the server, the server conducts the classifying step and transmits disposal information to a user device. By using the power of a server, a more accurate classification can usually be performed. However, if communications to the server are not possible, then the database located at a user device may be used. The database at the user device is preferably downloaded from a server. However, due to limitations of the user device, the database is usually smaller than that at a server. In one embodiment the database at the user device is selected at the server from a plurality of different databases in response to location of the user device. This ensures that the database at the user device comprises the most appropriate elements for a given location (e.g. a region, country, local government area etc.) In a simple example, the disposal information comprises a message confirming that the waste item may be placed in the waste receptacle although more detailed information is preferred. Conveying disposal information to a user preferably comprises displaying disposal information on a user device. In certain embodiments, an image of the waste receptacle is transmitted to the server. In such cases, the method may further comprise determining whether the receptacle should be emptied and notifying the appropriate operator. In some embodiments, a user reward is allocated in response to disposal of one or more waste items. In some embodiments, the method further comprises adding the waste item to a shopping list maintained in respect of a user and that shopping list may be maintained at the user device. Items are preferably added to the shopping list only when the user device is in a particular location, e.g. at the user's home. This will prompt the user to replace items that they use at home rather than list items of litter that the user has disposed of. According to a second aspect of the present invention there is provided a method of instructing a user how to dispose of a waste item, the method comprising: identifying a waste receptacle with a user device, determining at least a class of waste items that may be placed in the waste receptacle, and informing the user, via the user device, of at least some waste items that may be placed in the receptacle. This provides a quick and simple way of ensuring that waste items are placed in the correct waste or recycling channel. In one embodiment, determining at least a class of waste items is performed locally by the user device. This provides the user with a fast answer that doesn't rely upon a connection with a server. However, a server may be able to make a more accurate determination of the class of waste items so in another embodiment determining at least a class of waste items is performed by a remote server in response to a transmission received from the user device. A local determination is preferably performed by comparison with a local database and, further preferably, that local database has been downloaded from a server. A local database is necessarily limited in size so the local database is preferably selected from a plurality of candidate databases in response to the location of the user device. In one embodiment, the step of identifying the waste receptacle is performed in response to the location of the user device. More preferably, identifying the waste receptacle comprises consulting a database linking a plurality of locations with classes of waste items. However, the location of the device may not be available, for example due to lack of a GPS signal in a building or a lack of a GPS receiver in a user device. In another embodiment, therefore, identifying the waste receptacle comprises comparing an image of the waste receptacle with a plurality of images stored in a database. According to a third aspect of the present invention, there is provided a method of determining a schedule for emptying of a plurality of waste receptacles, comprising: storing a plurality of item disposals for each waste receptacle, building a model of item disposals for each waste receptacle over time, applying the model to determine a time at which each receptacle will require emptying, and instructing that the receptacle should be emptied at a particular time. This can maximise the use of valuable resources for those having responsibility for a waste receptacle by ensuring that it is not emptied too early or too late. According to a fourth aspect of the present invention, there is provided a waste management system comprising a server and a plurality of remote devices in communication with the server, each remote device being arranged to identify its location to the server, the server being arranged to download one of a plurality of different databases to each remote device in response to the location of the remote device, the remote device further being arranged to compare an image of a waste item with images from the image database to determine a class of a waste item, and the remote device further being arranged to inform a user of a suitable receptacle for the waste item. According to a fifth aspect of the present invention, there is provided a method of instructing a user how to dispose of a waste item comprising: determining a class of the waste item from a plurality of classes by comparing a characteristic of the waste item with a plurality of characteristics in a database, identifying a suitable receptacle for the class of waste item, informing a user of the identity of the suitable receptacle, and recording the disposal in a database. Preferably determining a class of the waste item includes capturing at least one image of the waste item and comparing that at least one image with a plurality of images in the database. Many people have a smartphone that is able to capture an image and transmit that image to a remote server which minimises the effort on the part of the user. Further preferably, determining a class of the waste item includes deriving a confidence measure of the determined class. This allows for different treatment of items for which the confidence level is low. For example, where the confidence measure is below a predetermined threshold, the method further comprises requesting further information from the user to improve the accuracy of the identified class. The further information may comprise a further image of the item and / or the further information may comprise recycling information displayed on the item. Preferably the plurality of classes comprise at least general waste, paper, plastics and metals. In a preferred embodiment, determining a class of the waste item comprises determining a plurality of classes of the waste item. Further preferably, informing the user may further comprise informing the user how to separate portions of the waste item, which portions belong to different classes. In preferred embodiments, the database is preferably derived from a machine learning algorithm. This may permit the database to be generated and / or updated with no explicit input from the database owner. Preferably, identifying a suitable receptacle comprises comparing an image of at least one receptacle with a plurality of images in a database. This allows a user to correctly identify a receptacle in an unfamiliar place or where the user is unable to read the instructions on the receptacle(s). Alternatively, or in addition, identifying a suitable receptacle may comprise determining a location of the user and deriving a suitable receptacle from a database of receptacles. That receptacle may comprise the closest suitable receptacle to the user to minimise inconvenience to the user or the suitable receptacle may comprise the closest suitable receptacle to a user's expected routing, which may be known from a navigation app on the user device. Preferably recording the disposal in a database comprises storing an award in respect of the user to maintain the user's motivation and one such award comprises entry in a competition. According to a sixth aspect of the present invention, there is provided a method of instructing a user how to dispose of a waste item comprising: identifying a suitable receptacle for the class of waste item, determining a location of the user and deriving a suitable receptacle from a database of receptacles, informing a user of the identity of the suitable receptacle, and recording the disposal in a database. When identifying a suitable receptacle comprises determining a location of the user and deriving a suitable receptacle, that receptacle may comprise the closest suitable receptacle to the user to minimise inconvenience to the user or the suitable receptacle may comprise the closest suitable receptacle to a user's expected routing, which may be known from a navigation app on the user device. In preferred embodiments of the invention, the determination of the class of the discarded item is by comparing an image with a plurality of images in a database. In preferred embodiments of the invention, the user may be assisted to determine which receptacle to use for disposal of the waste item. One possibility is that the user's location is determined using known techniques such as cell tower triangulation, WiFi IDs, GPS, A-GPS and so on. By maintaining a database of the location of waste and recycling receptacles, the user can be advised of the closest or most convenient receptacle. Training of the image database my occur in at least two different manners. In a first technique, an image capture device captures images of a plurality of items known (or suspected) to belong to a certain class. These items may have been manually sorted such as those from a particular recycling receptacle. Conveniently, the training process may be conducted at a recycling or waste management facility where items are likely to comprise a stream travelling along a conveyor belt. One or more cameras may be arranged around the conveyor belt to capture images of the items. These images can then be added to the database by way of the machine learning algorithm. As is known, such a machine learning algorithm will be trained on a large number of items (or rather their respective images) so that inappropriate items in a particular stream do not cause errors in the database. In some preferred embodiments, the waste item may comprise portions that fall into different classes such as food containers having a non-recyclable lid or a teabag comprising cardboard, a metal staple and organic matter. Such embodiments of the invention are arranged to inform the user of the identity of the suitable receptacles and, if necessary, how to separate the different portions of the item. Some elements of the present invention may be deployed at recycling centres to sort waste into different classes. This process is complex and requires enormous investment in machinery as well as staff to identify failings in that machinery. This results in significant expense which is mitigated by aspects of the present invention where waste items are correctly identified at source. This applies both the public and private receptacles ("kerbside pick up") for waste items. The present invention will now be described, by way of example, with reference to the accompanying drawings, in which: Figure 1 shows a communications system to which embodiments of the present invention may be applied, Figure 2 shows a flow chart of a method according to an embodiment of the present invention, Figure 3 shows a flow chart of a method according to another embodiment of the present invention, Figure 4 shows a logic diagram of a set of interactions between a user device and a server according to an embodiment of the present invention, Figure 5 shows a logic diagram of another set of interactions between a user device and a server according to an embodiment of the present invention, Figure 6 shows a flow chart of an ancillary process for use with embodiments of the present invention, Figure 7 shows a flow chart of another process according to an embodiment of the present invention, Figure 8 shows a flow chart of another process according to an embodiment of the present invention, Figure 9 shows a block schematic diagram of a waste management system according to an embodiment of the present invention, and Figure 10 shows a logic diagram of the operation of the waste management system shown in Figure 9. Figure 1 shows a communications system 100 to which embodiments of the present invention may be applied. A waste item 102 may be analysed by a user device 104 which is in communication with a cellular network 114. The network is connected, via internet 116 to a server 118 including a database 120. The user device comprises a camera 106, near field communication (NFC) capability 108, a location system such as GPS and a display 112. The user device may comprise a mobile phone, a tablet or similar device. The device may not necessarily include NFC capability and GPS functionality as different embodiments of the present invention operate using different hardware. Likewise, the location of the user device is only utilised in certain embodiments and, where the user location is exploited, it may be determined by other means such as the detection of local WiFi™ networks and / or triangulation from cellular network towers. The user device 104 may communicate with the server 118 using other communications technology, including but not limited to WiFi™ and wired network connections. Figure 2 shows a flow chart 200 illustrating an embodiment of the present invention. The functions described may be carried out by elements of the communications system as shown in Figure 1 or by other suitable hardware. The skilled person will appreciate that some functionality may reside in different pieces of hardware in different embodiments. For ease of description, some basic and inherent steps are omitted, for example unlocking steps, user verification steps, encryption steps and so on. The method starts at 202 and at step 204 an image of the waste device is taken, for example by the camera on the user device. Further images or even a short video recording could be captured, if desired. At step 206 the image (or images) are transmitted to a server such as that shown in Figure 1 (item 118). At step 208 the server makes a comparison between the image(s) and those stored in a database to determine a class and, optionally, a confidence rating for the determined class (step 210). The classes may comprise general waste, glass, plastics, aluminium and so on. The determination of the class may be made, for example, by comparing a trade mark on packaging. The optional confidence rating will generally be higher for an image of a complete item rather than a fragment. At step 212 it is determined whether the confidence is above a threshold and, if so, the class is transmitted to the user device at step 214. At step 216 the user device displays the identity of a suitable receptacle to the user and at step 218 the server records the disposal (associated with the particular user) in a database. At optional step 220 the server determines whether the user is now entitled to a reward and, if so, the reward is processed at step 224. The reward may comprise a message sent to the user's device such as "Congratulations! You have reached your weekly target of x pieces of litter." Alternatively, or in addition, the reward may comprise a monetary award or an entry in a prize draw. Whether or not the user is entitled to a reward, the process ends at 234. Returning to step 212, if the confidence rating of the class determined for the waste item is not above the predetermined threshold then a message is transmitted to the user device at step 226 and displayed to the user at step 228. This message may be "We cannot identify this waste item, please send further information". The user may at this point do some or all of the following: Capture one or more further images of the waste item Study recycling information on the waste item Make a decision on the appropriate receptacle for disposal At step 230 this further information is transmitted to the server and at step 232, the server updates the image database to include the at least one image of the waste item and the class determined from the further information. In this way, the database is updated by users to accommodate new products that are not yet identified in the database. In fact, the database may be entirely populated by such user information (i.e. with no pre-population of the database before public use of the method). Although this will be frustrating for early users of the system, the database will likely be populated very quickly, at least for commonly-found waste items. Processing then continues to step 218 and proceeds as described above. Since the confidence rating is optional, in some embodiments the steps 212 and 226 to 232 may be omitted and processing proceeds directly from step 210 to step 214. The principle of confidence may extend to the training steps 226 to 232 to enhance the resilience of the database. Whether deliberately or otherwise, users may make errors in their identification of the class of waste items that could not be identified by the server. When the database is updated at step 232 with an image and a class associated with that image, a confidence measure may also be stored in respect of the association. Factors that would increase this confidence measure might include: image is taken in a building (implying that it is not litter and the user is better able to determine what to do with the item) information from the user agrees with that received from other users user has provided information in the past that correlates with that from other users image(s) is / are clear and well lit Factors that might reduce the confidence measure might include: image is taken outside (implying that it is litter and the user is unfamiliar with the item) information from the user does not agree with that of other users user has provided information in the past that fails to correlate with that received from other users image(s) is / are not clear In a further refinement to the described embodiment, a waste item containing two or more portions belonging to different classes may be presented by the user. Such items may include a hand soap dispenser in which a bottle is recyclable (as plastics) but the handpump is not recyclable. Another example would be food packaging where a container is recyclable but a film lid is not. Further examples might be a glass jar with a metal lid or a teabag comprising cardboard, a metal staple and organic matter. In such cases of a waste item having two or more portions belonging to different classes, the process of Figure 2 would vary as follows: The comparison step 208 would generate two or more classes and an explanation for the user (e.g. "The film must be separated from the box. The box is to be placed in the plastics receptacle and the film is to be placed in the general waste receptacle") Step 214 would transmit both (or more) classes to the user device as well as the user explanation Step 216 would additionally display the user explanation Step 218 may record two (or more) disposals on the user's account The at least one image captured of the waste item may be analysed before transmission to the server to ensure that it meets certain requirements, for example that the entire waste item is included in the image and that the image is reasonably in-focus. The at least one image captured of the waste item may also be manipulated before comparison with the plurality of images in the database. For example, it may be compressed at the user device to reduce the communication overhead and / or modified at the user device or the server to facilitate the comparison with the database of images. The comparison may be conducted using any suitable search algorithm such as a depth-first search, breadth first search or a graph search. The embodiment of Figure 2 may be implemented in a number of different ways, such as by App Clips for loS devices or Instant Apps for Android devices. Alternatively, the user may access the service via a web browser (i.e. without downloading any apps) with certain data such as preferences and rewards stored in cookies. These variations may also be applied to the emodiments described below. Figure 3 shows a flow chart 300 of a further embodiment of the invention which may be integrated with one or more of the previously-described embodiments or may be operated in a stand-alone manner. For example, the method of Figure 3 may replace step 216 in Figure 2. A problem can arise, when a user is away from their home or workplace, that they do not know the location of a suitable receptacle for a waste item. In addition, recycling rules can vary from country to country and even from place to place. Two possible scenarios are considered: The user knows the class of the waste item but does not know where to dispose of it (this embodiment operates in a stand-alone mode), and The user knows neither the class of the waste item nor where to dispose of it (combine this embodiment with one or more previously-described embodiments) The process starts at 302 and at step 304 the question is raised from the user device as to where a receptacle may be found. This may be by way of an explicit request from the user or implied from the fact that the user has identified a waste item to the server. At step 306 the location of the user is determined using known techniques such as GPS, WiFi™ location, cell tower triangulation and so on. The server then determines, at step 308, the location of one or more suitable receptacles for the class or classes of the waste item. At step 310 the user is informed of the location of such receptacle(s), for example by identifying the location on a mapping application on the user device. The user may also be informed of how to travel to the location in known manner. In a further refinement, the user may be informed of a suitable location that imposes the least inconvenience to his or her journey. In this case, the user is following route directions in a mapping application on their device and a suitable receptacle on, or close to, that route can be identified. The appropriate receptacle may still not clear, however. Typically, recycling facilities include several receptacles and it may not be clear to the user which receptacle to use. It may be that the user is in a new location with unfamiliar recycling rules or the user may be unable to read the script or language identifying the receptacles. At step 312 an image of the facility is captured by the user and at step 314 it is uploaded to the server. At step 316, the sever determines whether the receptacle in the image is the correct one or which of several receptacles is the appropriate one. A warning may be presented to the user is they appear to be at risk of using the wrong recycling channel. The warning may comprise a visual and / or audible warning. The user may, of course, already be at a recycling facility and know the class of the item to be disposed of. However, the user may still not know which receptacle is appropriate due to differing rules in different jurisdictions or simply because he or she cannot read a foreign language. In this case, the initial steps 304 to 310 shown in Figure 3 are not required. A further possible refinement of the method shown in Figure 3 would be to identify a suitable receptacle which the user will pass, or pass close to, on his or her current journey. Such as journey could be determined from a navigation application on the user's device. For example, if the user is walking towards a known destination using their device to navigate, they can be informed as to the location of a suitable receptacle that they will pass en route. Alternatively, the navigation app may be instructed to modify the user's route ("add stop" feature) to include a stop at a suitable receptacle with minimal disruption to their journey. Figure 4 shows a logic diagram 400 of a client / server embodiment of the present invention. At step 402 a user device identifies the location of a waste receptacle or bin. This can be accomplished in a variety of different ways. The location of the user device may be determined in known manner such as using GPS, proximity to WiFi™ networks and so on. The location may equally be determined from a postcode (zipcode). The waste receptacle may be identified from the location by the user device or the location uploaded to a server as will be described subsequently. Alternatively, the waste receptacle may be identified from a Near Field Communication device associated with the receptacle, from an image of a QR code, barcode or alphanumeric code displayed on the receptacle. In some circumstances both a location and an image will be captured by the user device. At 404 the user device captures an image of the item to be discarded such as by the user taking a photo of the item. At 406 the image of the item and the location of the user device are uploaded to a server. Optionally, both the location of the user device and an image of the waste receptacle are also uploaded to the server. At 408, the server classifies the item to be discarded by comparing the image of the item with a plurality of images in a database. In order to speed up this classification, the item may be identified as a certain type of item such as a cup or the material may be inferred from the colour or brand. Comparisons with the database can then be limited to comparisons with images of that type of item, that colour and so on. The classification may also take account of a trademark or brand which is visible on the item. The server provides a classification for the item and, preferably, also a confidence measure for that classification. As previously described, the database has preferably been built using machine learning and, if the confidence measure is above a threshold, the server may add the new image of the item to the database together with the classification. Where the user device has not been able to identify the waste receptacle, the server further performs this step. At 410, feedback is provided to the user device. This feedback may range from the simple message such as "Yes, you may discard the item in that bin" or a warning "no, that is not the correct recycling channel" to further instructions for the disposal of the item, particularly for items such as vapes, batteries, recycling and so on. This is particularly appropriate for large items such as domestic appliances, furniture, display screens and so on. The feedback may include a reference to other digital content, for example by way of a web address, article, video, blog post, audio file etc. The feedback may include suggesting an alternative brand or product to the user. At 412, the feedback is conveyed to the user, for example by displaying a message on the screen of the user device. When the user device has uploaded a picture of the waste receptacle or bin, at optional step 414, the server determines whether the bin is full or overflowing. If the bin needs to be emptied then the server may notify the owner of the bin of this (provided that a mechanism exists to do so). At step 416 the server determines whether the user has earned a reward by discarding the item. Such a reward may be the result of the user hitting a target for waste disposal and / or entry into a random draw and so on. At step 418 the outcome of the reward determination is transmitted to the user device and at step 420 this is conveyed to the user. At optional step 422, the user device adds the discarded item to a shopping list held on the device or elsewhere, so that the item may be replaced. Figure 5 shows a logic diagram 500 for another embodiment, similar to that of Figure 4, where the user device performs item classification locally. Many of the steps are identical to those shown in Figure 4 and may be varied in the same manner. At step 502 the location of the user device is determined and at step 504 a photo is taken of the item. Instead of sending the photo of the item to the server, the user device performs the classification locally at step 506 and displays the outcome to the user at step 508. The local classification is performed in the same manner as the remote classification, i.e. by comparison with a database but the size of the database in the user device is necessarily limited by memory, processing power and / or battery limitations. While tests have demonstrated good product recognition from a database of only 800 images, in order to improve the reliability of recognition, the local database is preferably selected to differ from device to device as discussed below with reference to Figures 9 and 10. At step 510, the photo of the item is uploaded to the server and at step 512, the server determines whether the user is entitled to a reward and appropriate feedback is provided to the user device at step 514. The user device conveys this information to a user at step 516. At step 518, provided that the user device has uploaded a photo of the receptacle, the server determines whether the receptacle needs to be emptied and provides the appropriate notification. The user device may also add the item to a shopping list as previously described. The embodiment shown in Figure 5 has the advantage that the user device does not need to be in communication with the server to provide the user with disposal information for the item. This could occur if the user device is not within range of a cellular network tower or the server is offline for some reason. Steps 510 to 518 can then be performed later when the user device is able to communicate with the server. In embodiments without a reward feature, these steps may be omitted entirely. In a variant of the embodiment shown in Figure 5, steps 504 and 506 can be replaced with a single step of instant recognition and classification. The user simply points a device camera at the item to be disposed of and the user device identifies the item almost instantly. At step 510, instead of uploading a photo, the user device confirms disposal of an item to the server. This confirmation could be a simple count of one item or identify the class and / or brand and / or material of the item. Two hybrid modes of operation can also be envisaged. In the first mode, a user device defaults to requesting an item classification from the server unless it is unable to communicate with it, in which case a classification is performed locally. This has the advantage of using the (generally) more accurate server-based classification where possible. In the second mode, a user device defaults to making the item classification locally unless it is unable to do so (or can only do so with a low level of confidence), in which case it uploads a photo of the item to the server for classification. This has the advantage of reducing communication and sever overhead. Further steps may be incorporated in the processes described with reference to Figures 4 and 5. Figure 6 shows a short routine 600 that may be executed by the user device at the start of those two processes. The routine starts at 602 and proceeds to step 604 at which the receptacle or bin is scanned. The bin may be provided with an identifier such as a URL, a barcode, QR code, RFID tag for Near Field Communication and so on. Any means of identifying the bin may be used. The bin preferably carries a sign to indicate that a user may earn rewards for disposing of litter, if they download an app to their device or visit the identified website. The user device interprets the identifier as referring to a downloadable application. At step 606 the user device determines whether the app is already installed on the device. If not, it is downloaded at step 608 and the routine ends at step 610. If the app is already installed, then step 608 is omitted. The routine shown in Figure 6 thus allows a user to join the reward scheme and start to earn rewards immediately, even if they have not previously downloaded the app. While embodiments of the present invention have been described with reference to identifying a waste receptacle and an item and informing a user whether or not the item may be discarded in the receptacle, a more straightforward embodiment is shown at 700 in Figure 7. In this embodiment, a user identifies a receptacle or bin with their device at step 704. Step 706 determines what class of waste items maybe placed in that particular bin and step 708 informs the user of the class. The routine ends at 710. This embodiment provides a simpler alternative to a user in that he or she does not have to take a photo or point their device at a waste item. The user may be informed at step 708 in a variety of different ways. For example, the user may simply be informed of a class of items, for example: soft plastics. The user may also be given examples to improve their understanding, such as: carrier bags, film from containers, chocolate wrappers. In preferred embodiments, step 708 comprises displaying one or more images of suitable items to the user. This will usually allow the user to understand the range of suitable items more quickly, especially if they do not speak the local language. In further preferred embodiments, the user may select from multiple images presented on screen to obtain further examples of suitable items. The step 706 may be performed locally or at a remote server as previously described. Where the step is performed locally, the size of the database may be limited, whether that is a database linking locations with the class(es) of suitable waste items or an image database for identifying a receptacle from its appearance. To maximise the efficiency of either database, it is preferably selected in response to the location of the user device. In this way a location database will only need to cover a certain geographical area and an image database will only need to include images of waste receptacles that are present in a certain country or region. The server may be arranged to download the appropriate local databases to user devices in response to their location. The receptacle may be capable of receiving multiple classes of items, for example: plastic bottles, food wrapping, aluminium cans & disposable cutlery. Figure 8 shows a flow chart of a method (which may be executed at a server) according to an embodiment of the invention whereby an artificial intelligence arrangement is used to schedule emptying of waste receptacles or bins. The method starts at 802 and at step 804 the disposal of a waste item is logged in respect of each of the bins that the system is responsible for. The system is, of course, only aware of waste items disposed of by users of its own app. Waste items will be placed in the bin by others so the system does not have total visibility of waste items placed in any particular bin. Consequently, a database is built over time at 806 that is responsive to (some) items placed in each bin and also information provided by those responsible for the bin whether it was emptied too early (not full) or too late (overflowing). A pattern for usage of the bin will necessarily emerge. This could be affected by days of the week - a bin at a leisure attraction will likely be filled more quickly at the weekend. It could also be affected by the weather - a bin at a beach will likely fill more quickly on a sunny day. By comparing, in real or near-real time, the known usage of the bin (collected from users via their apps) with previous usage of the bin, the system will thus be able to accurately predict when a receptacle will need to be emptied at step 808. Instructions to the appropriate personnel are provided at 810 and the method ends at 812. Figure 9 shows a block diagram 900 of a waste management system according to another embodiment of the present invention. A server 902 is connected via internet 904 and cellular networks 906, 908 to a first plurality of user devices 910 and a second plurality of user devices 912. The user devices 910 are located in a first area Al while the user devices 912 are located in a different area A2. The rules and practices of disposal of waste items or the identity of waste items differ between the two areas. For example, area Al may be managed by a particular local government while the area A2 is managed by a different local government with a different waste recycling policy. As another example, area Al may be located in a first country while area A2 is located in a different country having different brands and therefore different litter to identify. Since the memory and processing power of the user devices is limited, the system is arranged to provide the most appropriate image database to each of the devices 910, 912. This ensures good accuracy at identifying waste items locally at the device while minimising the memory and processing requirements. The server is responsible for downloading a database to each of the user devices 910, 912 as described with reference to Figure 10. Figure 10 shows a logic diagram 1000 of the operations performed by the system of Figure 9. At step 1002 the server enquires the location of the devices 910 and 912. At step 1004, the device 910 determines its location and at 1006 reports it to the server. At step 1008, the device 912 determines its location and at 1010 reports it to the server. The server, at step 1012 determines the appropriate database for the remote device 910 and at 1014 transmits the database to the device 910. The device 910 stores the database at step 1016. The server, at step 1018 determines the appropriate database for the remote device 912 and at 1020 transmits the database to the device 912. The device 912 stores the database at step 1022. The remote user devices 910 and 912 are thus provided with appropriate and efficient image databases to perform the waste item recognition described with reference to Figure 5. Further refinements to the method of Figure 10 can be applied, such as a query step to determine whether the remote devices 910, 912 already have the appropriate databases installed. This would save bandwidth and processing overhead when a remote device already has the correct database installed. Where refinements and modifications to embodiments of the present invention have been described in the context of one embodiment, the skilled reader will appreciate that these refinements and modifications can be applied equally to other embodiments. In other aspects, the present invention is defined by the following clauses: 29. A method of instructing a user how to dispose of a waste item comprising: determining a class of the waste item from a plurality of classes by comparing a characteristic of the waste item with a plurality of characteristics in a database, identifying a suitable receptacle for the class of waste item, informing a user of the identity of the suitable receptacle, and recording the disposal in a database. 30. A method according to clause 29, wherein determining a class of the waste item includes capturing at least one image of the waste item and comparing that at least one image with a plurality of images in the database. 31. A method according to any one of the clauses 29 to 30, wherein determining a class of the waste item includes deriving a confidence measure of the determined class. 32. A method according to clause 31, wherein if the confidence measure is below a predetermined threshold, the method further comprises requesting further information from the user. 33. A method according to clause 32, wherein the further information includes a further image of the item. 34. A method according to any one of the clauses 32 or 33, wherein the further information includes recycling information displayed on the item. 35. A method according to any one of the clauses 29 to 34, wherein the plurality of classes comprise at least general waste, paper, plastics and metals. 36. A method according to any one of the clauses 29 to 35, wherein determining a class of the waste item comprises determining a plurality of classes of the waste item. 37. A method according to clause 36, wherein informing the user may further comprise informing the user how to separate portions of the waste item, which portions belong to different classes. 38. A method according to any one of the clauses 29 to 37, wherein the database is derived from a machine learning algorithm. 39. A method according to any one of the clauses 29 to 38, wherein identifying a suitable receptacle comprises comparing an image of at least one receptacle with a plurality of images in a database. 40. A method according to any one of the clauses 29 to 39, wherein identifying a suitable receptacle comprises determining a location of the user and deriving a suitable receptacle from a database of receptacles. 41. A method according to clause 40, wherein the suitable receptacle comprises the closest suitable 5 receptacle to the user. 42. A method according to clause 40, wherein the suitable receptacle comprises the closest suitable receptacle to a user's expected routing. 43. A method according to any one of the clauses 29 to 42, wherein recording the disposal in a database comprises storing an award in respect of the user. 10   44. A method according to clause 43, wherein the award comprises entry in a competition. 45. A method of instructing a user how to dispose of a waste item comprising: identifying a suitable receptacle for the class of waste item, determining a location of the user and deriving a suitable receptacle from a database of receptacles, informing a user of the identity of the suitable receptacle, and recording the disposal in a database. 15   46. A method according to clause 45, wherein the suitable receptacle comprises the closest suitable receptacle to the user. 47. A method according to clause 45, wherein the suitable receptacle comprises the closest suitable receptacle to a user's expected routing.

Claims

1. A method of instructing a user how to dispose of a waste item, the method comprising:determining a location of a waste receptacle,capturing an image of a waste item,classifying the waste item,conveying disposal information to a user.

2. A method as claimed in claim 1, wherein determining the location of a waste receptacle comprises determining the location of a user device.

3. A method as claimed in claim 1, wherein determining the location of a waste receptacle comprises capturing information associated with the waste receptacle.

4. A method as claimed in any one of the claims 1 to 3, wherein capturing an image of a waste item and classifying the waste item comprises an instant image comparison with a database.

5. A method as claimed in any one of the claims 1 to 3, wherein classifying the waste item comprises comparing an image of the waste item with a plurality of images in a database.

6. A method as claimed in claim 5, wherein the database is generated by machine learning.

7. A method as claimed in claim 5 or claim 6, wherein the database is located at a server, theuser device transmits at least one image to the server, the server conducts the classifying step and transmits disposal information to a user device.

8. A method as claimed in claim 5 or claim 6, wherein the database is located at a user device.

9. A method as claimed in claim 8, wherein the database at the user device is downloaded froma server.

10. A method as claimed in claim 9, wherein the database is selected at the server from a plurality of different databases in response to location of the user device.

11. A method as claimed in any one of the claims 1 to 10, wherein the disposal information comprises a message confirming that the waste item may be placed in the waste receptacle.

12. A method as claimed in any one of the claims 1 to 11, wherein conveying disposal information to a user comprises displaying disposal information on a user device.

13. A method as claimed in any one of the claims 1 to 12, further comprising determining whether the receptacle should be emptied and notifying the appropriate operator.

14. A method as claimed in any one of the claims 1 to 13, wherein a user reward is allocated in response to disposal of one or more waste items.

15. A method as claimed in any one of the claims 1 to 14, further comprising adding the waste item to a shopping list maintained in respect of a user.

16. A method as claimed in claim 15, wherein the shopping list is maintained at the user device.

17. A method as claimed in claim 15 or claim 16, wherein the waste item is added to a shoppinglist only is response to the user device being in a particular location.

18. A method of instructing a user how to dispose of a waste item, the method comprising:identifying a waste receptacle with a user device,determining at least a class of waste items that may be placed in the waste receptacle, andinforming the user, via the user device, of at least some waste items that may be placed in the receptacle.

19. A method as claimed in claim 18, wherein determining at least a class of waste items is performed by the user device.

20. A method as claimed in claim 18, wherein determining at least a class of waste items is performed by a remote server in response to a message received from the user device.

21. A method as claimed in claim 19, wherein determining the class of waste items is performed by comparison with a local database.

22. A method as claimed in claim 21, wherein the local database is downloaded from a server.

23. A method as claimed in claim 21 or claim 22, wherein the local database is selected from a plurality of candidate databases in response to the location of the user device.

24. A method as claimed in any one of the claims 18 to 23, wherein identifying the waste receptacle is performed in response to the location of the user device.

25. A method as claimed in claim 24, wherein identifying the waste receptacle comprises consulting a database linking a plurality of locations with classes of waste items.

26. A method as claimed in any one of the claims 18 to 25, wherein identifying the waste receptacle comprises comparing an image of the waste receptacle with a plurality of images stored in a database.

27. A method of determining a schedule for emptying of a plurality of waste receptacles, comprising:Storing a plurality of item disposals for each waste receptacle,building a model of item disposals for each waste receptacle over time,applying the model to determine a time at which each receptacle will require emptying, and instructing that the receptacle should be emptied at a particular time.

28. A waste management system comprising a server and a plurality of remote devices in communication with the server,each remote device being arranged to identify its location to the server,the server being arranged to download one of a plurality of different image databases to each remote device in response to the location of the remote device,the remote device further being arranged to compare an image of a waste item with images from the image database to determine a class of a waste item, andthe remote device further being arranged to inform a user of a suitable receptacle for the waste item.

Citation Information

Patent Citations

  • Garbage classification method, terminal and computer readable storage medium

    CN110436081A

  • Intelligent garbage can

    CN112278639A

  • Garbage disposal support apparatus and garbage disposal support method

    US20230019417A1