Equipment for the classification and selection of objects containing precious metals and its method of operation

Abstract

Provided is a facility (I) for classifying and selecting objects containing precious metals and copper. 【Solution means】The facility (I) is used to support the facility (I) on the ground, and includes a support frame (1) extending from a first end (11) to a second end (12), a conveyor belt (13) coupled to the support frame (1) and sliding in a sliding direction (V) from the first end (11) towards the second end (12) for transporting the object (O), an image acquisition and processing device (2) disposed at the first end (11) of the support frame (1), lifted relative to the conveyor belt (13), for acquiring an image of the object (O), detecting predetermined visual features of the object (O) in the acquired image, associating the category of precious metals contained in the object (O) with the visual features, classifying the object (O) based on the association, and transmitting a signal corresponding to the classification, and at least one actuator device (3) disposed on the support frame (1) downstream of the image acquisition and processing device (2), for receiving the corresponding signal from the image acquisition and processing device (2) and sorting the object (O) based on the corresponding signal. The present invention also relates to a method of operating the facility (I).

Description

Technical Field

[0001] The present invention relates to an apparatus for the classification and selection of objects containing precious metals and copper, in particular, waste electronic boards and used vehicle catalysts.

[0002] The present invention also relates to a method of operating the above apparatus.

[0003] More specifically, the present invention relates to the above-described type of apparatus designed and made for the recognition and selection of objects containing precious metals and copper, such as electronic boards and catalysts, and aims at the recovery and recycling of the precious metals and copper contained therein, but can also be used for any object containing recoverable precious metals and copper.

[0004] Hereinafter, an apparatus for the classification and selection of electronic boards and catalysts will be described, but it is obvious that it is not limited to this specific application.

Background Art

[0005] As is well known, currently, the selection of waste electronic boards is manually performed by specialized staff trained to visually inspect and group different types of electronic boards.

[0006] The purpose of the selection is to obtain a uniform portion with respect to the amount of precious metals and copper contained in the electronic board and use it for subsequent recovery and purification.

[0007] Manual selection of electronic boards requires a large number of specialized staff, and generally, in a dangerous workplace environment with a high risk of accidents, they are exposed to isolated work, there are many types of materials to be selected, and a large amount of time is required.

[0008] Furthermore, the margin of error during selection depends on the experience and concentration of each operator.

[0009] Systems used for the selection of electronic waste are also known, such as optical, color, or shape recognition technologies.

[0010] Some of these well-known systems provide the use of artificial intelligence algorithms, but can only separate electronic boards from other types of waste found inside the recycling facility.

[0011] Regarding vehicle waste catalysts, there is currently no automatic selection system, so sorting has been done manually until now, forcing operators to compare several databases showing the individual identification numbers and their values of the waste catalysts.

Summary of the Invention

Problems to be Solved by the Invention

[0012] Therefore, from the above viewpoints, an object of the present invention is to overcome the limitations not solved by the prior art by providing equipment for classifying and selecting objects containing precious metals and copper that can automatically classify the objects and automatically select them based on the precious metals contained therein.

[0013] A further object of the present invention is to provide the tools necessary for the execution of the method and the apparatus for executing the method.

Means for Solving the Problems

[0014] Accordingly, a specific object of the present invention is an equipment used to support the above-mentioned equipment on the ground, including a support frame extending from a first end to a second end, and a conveyor belt coupled to the support frame and sliding along a sliding direction from the first end to the second end, capable of transporting the object, for classifying and selecting an object containing precious metals and copper, wherein the equipment is disposed at the first end of the support frame, lifted with respect to the conveyor belt, acquires an image of the object, detects a predetermined visual feature of the object in the acquired image, associates a category of precious metals contained in the object with the visual feature, and includes an image acquisition and processing device capable of classifying the object based on the association, transmits a signal corresponding to the association, is disposed on the support frame downstream of the image acquisition and processing device, receives the signal corresponding to the image acquisition and processing device, and includes at least one actuator device capable of sorting the object based on the corresponding signal.

[0015] Furthermore, according to the present invention, the image acquisition and processing device includes a cybernetic logic unit capable of supporting a neural network that detects a predetermined visual feature of the object in the acquired image, associates a category of precious metals contained in the object with the predetermined visual feature, and classifies the object (O) based on the association.

[0016] Furthermore, according to the present invention, the image acquisition and processing device includes at least one lighting element capable of illuminating the object, at least one detection element capable of acquiring an image of the object, and a cybernetic logic unit operably connected to the at least one lighting element and the at least one detection element.

[0017] According to the present invention, the at least one lighting element is a light emitting diode (LED), and the detection element is a video camera or a CMOS optical sensor.

[0018] Preferably, according to the present invention, the image acquisition and processing device includes a graphic interface capable of showing the operator the result of the classification of the object and changing the classification parameters.

[0019] Furthermore, according to the present invention, the equipment can include a plurality of slides arranged at the second end of the conveyor belt, and since the actuator device includes at least one piston capable of moving from a retracted rest position relative to the conveyor belt to an extended working position, it can contact the objects arranged on the conveyor belt and convey them to one of the plurality of slides based on the corresponding signals.

[0020] Furthermore, according to the present invention, the equipment can include a plurality of slides arranged at the second end of the conveyor belt, and the actuator device is coupled to the support frame and includes a support element lifted relative to the conveyor belt and gripping and conveying means coupled to the support element for picking up the objects arranged on the conveyor belt and conveying them to one of the plurality of slides based on the corresponding signals.

[0021] A further object of the present invention is an operating method of equipment including the steps of acquiring an image of the object by the image acquisition and processing device, detecting predetermined visual features of the object in the acquired image by the control logic unit, associating the category of precious metal present in the object with the predetermined visual features, classifying the object based on the association and transmitting a signal corresponding to the classification, and sorting the object based on the corresponding signal by the actuator device.

[0022] Furthermore, according to the present invention, the detection step includes a sub-step of adjusting the image acquired in the acquisition step, a sub-step of extracting predetermined visual features by a trained convolutional filter, and a sub-step of classifying an object by a neural network.

[0023] Furthermore, according to the present invention, the method includes a step of training the neural network, a sub-step of determining an image dataset for training the neural network, a sub-step of selecting images based on their quality, and a sub-step of associating each image selected in the selection step with a category determined based on the precious metal and copper included in the object (O) of the image.

Brief Description of the Drawings

[0024] Here, based on a preferred embodiment of the present invention, in particular, with reference to the accompanying drawings, the present invention will be described as a non-limiting example.

[0025]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Embodiments for Carrying Out the Invention

[0026] In each drawing, similar parts are denoted by the same reference numerals.

[0027] Referring to FIG. 1, the equipment I for the classification and selection of the object O containing precious metals and copper basically includes a support frame 1 for supporting the equipment I on the ground, an image acquisition and processing device 2 arranged on the support frame 1, and an actuator device 3 arranged on the support frame 1 downstream of the image acquisition and processing device 2.

[0028] The support frame 1 extends from a first end 11 to a second end 12.

[0029] The support frame 1 includes a plurality of support legs G.

[0030] The support frame 1 includes a plurality of outputs or slides S arranged downstream of the image acquisition and processing device 2.

[0031] The sorted object O is conveyed from the equipment I to a collection container (not shown for the purpose of the present invention) via the slide S.

[0032] The equipment I includes a conveyor belt 13 that is coupled to the support frame 1 and slides along a sliding direction V from the first end 11 to the second end 12, and can convey the object O within the equipment I.

[0033] The object O is a waste electronic board containing precious metals such as gold, silver, palladium, rhodium, and non-precious metals such as copper.

[0034] The object O is also a catalyst mainly containing precious metals such as palladium, rhodium, and platinum.

[0035] Without departing from the scope of protection of the present invention, the object O may be other known objects containing precious metals and copper.

[0036] The equipment I includes an image acquisition and processing device 2 coupled to the support frame 1.

[0037] Specifically, the image acquisition and processing device 2 is disposed at the first end 11 and is lifted with respect to the conveyor belt 13.

[0038] The image acquisition and processing device 2 can detect precious metals and copper contained in the object O, classify the object O based on the detection, and transmit a signal corresponding to the classification.

[0039] Specifically, the image acquisition and processing device 2 can detect an image of the object O.

[0040] More specifically, referring to FIG. 2, the image acquisition and processing device 2 includes at least one illumination element 21 and at least one detection element 22.

[0041] The at least one illumination element 21 is of the LED (light emitting diode) type or the like.

[0042] In the present embodiment, the image acquisition and processing device 2 includes two illumination elements, but it is also possible to provide a different number of illumination elements.

[0043] The at least one detection element 22 is a video camera or a CMOS type optical sensor.

[0044] Since the at least one detection element 22 is disposed at a predetermined distance from the conveyor belt 13, an image of the object O can be detected from different angles.

[0045] The image acquisition and processing device 2 includes an optical sensor or a photodiode (not shown) that can detect the presence of the object O on the conveyor belt 13 in the image acquisition and processing device 2 and activate the illumination element 21 and the video camera 22 to acquire an image.

[0046] The image acquisition and processing device 2 includes a control logic unit U operably connected to the video camera and the lighting element.

[0047] The logic control unit U includes a memory unit in which the data and operation parameters of the facility I and the images detected by the video camera are stored.

[0048] The logic control unit U can process the images (details will be described later).

[0049] Furthermore, the logic control unit U can transmit a signal corresponding to the image processing.

[0050] Specifically, in the memory unit, the unit category of the object O is stored based on the amount and type of precious metals and copper present therein.

[0051] In the present invention, specifically, the five categories of precious metals present in the object O are predetermined, and copper is added thereto.

[0052] The association between the image of the object O and the categories of precious metals and copper contained therein is then predefined and stored in the memory unit.

[0053] The association is obtained by experimental tests.

[0054] In the present invention, the association mode between the image of the object O and the categories of precious metals and copper is constituted by a look-up table.

[0055] However, without departing from the protection scope of the present invention, the association mode may be different.

[0056] Furthermore, the logic control unit U includes a processing device capable of implementing a neural network.

[0057] Without departing from the scope of protection of the present invention, the neural network may be included in a remote device such as the cloud.

[0058] The neural network is trained by associating an image of the object O with categories based on the precious metals and copper present in the object O.

[0059] In use, the neural network can recognize an image of the object O and associate the category based on the precious metals and copper present in the object O (details will be described later).

[0060] The image acquisition and processing device 2 further includes a graphic interface or a display D operably connected to the logic control unit U.

[0061] An operator can communicate with the facility I through the display D.

[0062] For example, the operator can display the classification result of the object O.

[0063] Furthermore, the operator can change the classification level of the object O through the display.

[0064] Specifically, the operator can determine which category or combination of categories of the object O to select.

[0065] The actuator device 3 is disposed downstream of the image acquisition and processing device 2, substantially on the support frame 1 of the second end 12.

[0066] The actuator device 3 receives the signal transmitted by the logic control unit U by a programmable logic controller (PLC) unit, wirelessly, or via Bluetooth (registered trademark).

[0067] In the first embodiment, the actuator device 3 includes at least one piston 31.

[0068] The actuator device 3 receives the signal transmitted by the logic control unit U via the PLC unit, and controls the piston 31 to sort the objects O and convey them to one of the plurality of slides S.

[0069] The actuator device 3 may also include additional pistons to optimize sorting.

[0070] In the above-described embodiment, the facility I includes a second piston 32.

[0071] Since the piston 31 can move from a retracted rest position relative to the conveyor belt 13 to a drawn-out working position, it can contact the objects O arranged on the conveyor belt 13 and convey them to one of the plurality of slides S.

[0072] Referring to FIG. 5, in the second embodiment, the facility I differs from the above description only in the structure of the actuator device 3'.

[0073] The actuator device 3' includes a support element 30' coupled to the support frame 1 and lifted relative to the conveyor belt 13, and gripping and conveying means 31' coupled to the support element 30'.

[0074] The actuator device 3' receives the signal transmitted by the logic control unit U via the PLC unit, and controls the gripping and conveying means 31' to sort the objects O and convey them to one of the plurality of slides S.

[0075] Referring to the block diagram shown in FIG. 3, the operation of the above-mentioned equipment I for the classification and selection of the object O containing precious metals and copper is carried out as follows.

[0076] In the initial step 100, the equipment I is started up.

[0077] In the positioning step 110, the object O is placed on the conveyor belt 13 that slides along the sliding direction V until it enters the image acquisition and processing device 2.

[0078] In the presence confirmation step 120, the diode confirms the presence of the object O on the conveyor belt 13.

[0079] If the object O is detected, the acquisition step 130 is executed, and an image of the object O is acquired by the video camera 22.

[0080] On the contrary, if the object O is not detected, the operation confirmation step 140 of the equipment I is executed.

[0081] If the equipment I is on, the presence confirmation step 120 of the object O is executed again. If the equipment I is off, this method is terminated.

[0082] Following the acquisition step 130, a series of pre-adjustment and processing steps 150 of the detected image are executed.

[0083] Specifically, in the first step 151, the identification of the object O in the image is maximized by cropping and resizing the acquired image of the object O.

[0084] In the second step 152, the image cropped and resized in the first step 151 is processed by a trained convolutional filter to extract the visual features required for classification.

[0085] The visual features are automatically extracted by a convolutional network based on the input dataset.

[0086] In the third step 153, based on the criteria obtained during training, the neural network classifies the image obtained from the processing of the previous step.

[0087] Specifically, referring to FIG. 4, the training step 200 is performed by the following sub-steps.

[0088] In the first sub-step 201, an image dataset for training the neural network is determined and stored in the memory unit of the cybernetic logic unit U.

[0089] In the second sub-step 202, the quality of the data subset is selected. That is, based on the quality, specifically, based on the ability to faithfully represent the categories to which they belong, the images are selected.

[0090] In the third sub-step 204, each detected image is associated with a category determined according to predetermined laboratory data based on the precious metals and copper contained in the object O of the image.

[0091] In the fourth sub-step 205, the neural network is trained for classification.

[0092] After executing the third classification step 153, the mechanical selection step 160 is executed.

[0093] Specifically, in the first sub-step 161, the object O is conveyed to the second end 12 by the conveyor belt 13.

[0094] In the second sub-step 162, the pistons 31, 32 sort the object O based on the classification.

[0095] Specifically, the pistons 31, 32 convey the object O toward the slide associated with the classification of the object O.

[0096] In fact, the category is associated with each of the plurality of slides S based on the precious metals and copper present in the object O.

[0097] The operator can set which type of object O to convey to which slide according to his own needs.

Industrial Applicability

[0098] As is clear from the above description, the equipment which is the object of the present invention can separate the electronic board and the catalyst based on the content of precious metals and copper present therein, so that the width of the error related to the recognition by the artificial intelligence algorithm can be significantly reduced.

[0099] In this way, the health risks associated with the long-term handling and repeated selection operations of waste materials by the operator are eliminated or at least reduced.

[0100] Furthermore, unlike the operator who requires an instruction stage for training to recognize the card associated with the metal content based on the appearance, this equipment does not require an initial period for full operation.

[0101] Although the present invention has been described by way of non-limiting examples based on preferred embodiments, it is clear that those skilled in the art may make changes and / or variations without departing from the scope of protection as defined by the appended claims.

Claims

1. Equipment (I) for the classification and selection of an object (O) containing a precious metal and copper, A support frame (1) used to support the equipment (I) on the ground and extending from a first end (11) to a second end (12), A conveyor belt (13) coupled to the support frame (1) and sliding along a sliding direction (V) from the first end (11) towards the second end (12), capable of transporting the object (O), Including, the equipment (I) being, Disposed at the first end (11) of the support frame (1), lifted with respect to the conveyor belt (13), Obtaining an image of the object (O), Detecting predetermined visual features of the object (O) in the obtained image, Associating the category of the precious metal contained in the object (O) with the visual features, Classifying the object (O) based on the association, Transmitting a signal corresponding to the classification, Including an image acquisition and processing device (2) capable of performing the above, Disposed on the support frame (1) downstream of the image acquisition and processing device (2), receiving the corresponding signal from the image acquisition and processing device (2), and including at least one actuator device (3, 3') capable of sorting the object (O) based on the corresponding signal. The equipment (I) is characterized by this.

2. The image acquisition and processing device (2) being, Detecting predetermined visual features of the object (O) in the obtained image, Associating the category of the precious metal contained in the object (O) with the predetermined visual features, Classifying the object (O) based on the association, The equipment (I) according to the above claim, characterized by including a logic control unit (U) capable of supporting a neural network capable of performing the above.

3. The image acquisition and processing device (2) being, At least one lighting element (21) capable of illuminating the object (O), At least one detection element (22) capable of obtaining an image of the object (O), A logic control unit (U) operably connected to the at least one lighting element (21) and the at least one detection element (22), The equipment (I) according to the above claim, characterized by including the above.

4. The equipment (I) according to the claim, wherein the at least one lighting element (21) is an LED (light emitting diode), and the detection element (22) is a video camera or a CMOS optical sensor.

5. The equipment (I) according to any one of the preceding claims, wherein the image acquisition and processing device (2) includes a graphic interface (D) capable of showing the result of the classification of the object (O) to an operator and changing classification parameters.

6. Including a plurality of slides (5) arranged at the second end (12) of the conveyor belt (13), The equipment (I) according to any one of the preceding claims, wherein the actuator device (3) includes at least one piston (31) capable of moving from a rest position retracted with respect to the conveyor belt (13) to a working position pulled out, contacting the object (O) arranged on the conveyor belt (13), and conveying them to one of the plurality of slides (S) based on the corresponding signal.

7. Including a plurality of slides (S) arranged at the second end (12) of the conveyor belt (13), The equipment (I) according to any one of claims 1 to 5, wherein the actuator device (3') is coupled to the support frame (1), includes a support element (30') lifted with respect to the conveyor belt (13), and gripping and conveying means (31') coupled to the support element (30') and capable of picking up the object (O) arranged on the conveyor belt (13) and conveying them to one of the plurality of slides (S) based on the corresponding signal.

8. A step (130) of acquiring an image of the object (O) by the image acquisition and processing device (2); A step (150) of detecting a predetermined visual feature of the object (O) in the acquired image by the logic control unit (U); A step (152) of associating the category of the precious metal included in the object (O) with the predetermined visual feature; A step (153) of classifying the object (O) based on the association and transmitting a signal corresponding to the classification; A step (160) of sorting the object (O) by the actuator device (3, 3') based on the corresponding signal; The operation method of the facility (I) according to any one of claims 1 to 7, including

9. wherein the detection step (150) comprises a sub-step (151) of adjusting the image obtained in the acquisition step (130); a sub-step (152) of extracting the predetermined visual features by a trained convolutional filter; a sub-step (153) of classifying the object (O) by a neural network; The method according to the claim, characterized by including

10. including a step (200) of training the neural network, a sub-step (201) of determining an image dataset for training the neural network; a sub-step (202) of selecting the images based on the quality of those images; a sub-step (203) of associating a specific category based on each image selected in the selection step (202) and the precious metal and copper included in the object (O) of the image; The method according to the claim, characterized by including

Images