RU2774280C2 - Apparatus for preparing a beverage with capsule recognition - Google Patents

Abstract

FIELD: preparation of beverages.

SUBSTANCE: apparatus for preparing and dispensing a beverage, such as tea, coffee, hot chocolate, cold chocolate, milk, soup or baby food, containing a capsule recognition module for recognising a capsule inserted into said apparatus in the capsule recognition position, wherein the capsule recognition module comprises a camera for capturing an image of at least part of said capsule in said capsule recognition position; wherein the capsule recognition module comprises a neural network computing apparatus, wherein said neural network computing apparatus is configured to determine the type of said capsule from multiple preset capsule types based on the image of at least part of said capsule captured by said camera.

EFFECT: improvement in the dispensing of a beverage by apparatuses capable of automatically reliably identifying capsules.

24 cl, 1 tbl, 5 dwg

Description

Scope of the invention

The field of application of the invention relates to devices for preparing a drink, in which capsules with an ingredient of the prepared drink are used. The field of application of the invention relates in particular to devices for preparing a beverage using capsules, which are capable of automatically recognizing the type of capsule inserted into the device, for example to adapt the beverage preparation parameters to the recognized type of capsule.

For the purposes of this specification, the term "beverage" is intended to include any liquid substance intended for human consumption, such as tea, coffee, hot or cold chocolate, milk, soup, baby food, and the like. A "capsule" is intended to contain a pre-portioned beverage ingredient, such as a flavoring ingredient, in a package of any material, in particular airtight packaging such as plastic or aluminium, in recyclable and/or biodegradable packages of any shape and structure, including those containing ingredient soft containers or rigid cartridges. The capsule may contain an amount of an ingredient for preparing a single serving of a drink or a plurality of servings of a drink.

Prerequisites for the creation of the invention

Some beverage preparation devices use capsules that contain ingredients to be extracted or dissolved and/or ingredients that are stored and automatically dispensed in the device or otherwise added at the time of preparation of the drink. Some beverage preparation devices are equipped with filling devices which include a liquid pump, typically water, which pumps liquid from a source of water, either cold or heated by heating devices such as a thermoblock or the like.

Particularly in the field of coffee preparation, devices are widely used in which a capsule containing beverage ingredients is inserted into a brewing device.

Designs of brewing devices have been developed that facilitate the insertion of a “new” capsule and the removal of a used capsule. Generally, brewers comprise two parts with relative mobility to move from the insertion/removal position of the capsule to the position for brewing the ingredient in the capsule.

The movable part of the brewer can be manually driven as described in WO 2009/043630, WO 01/15581, WO 02/43541, WO 2010/015427, WO 2010/128109, WO 2011/144719 and WO 2012/032019. Various handle configurations are described in EP 1867260, WO 2005/004683, WO 2007/135136, WO 2008/138710, WO 2009/074550, WO 2009/074553, WO 2009/074555, WO9 2009/0745057, WO /037806, WO 2011/042400, WO 2011/042401 and WO 2011/144720. The use of such devices in beverage preparation devices is described in WO 2009/074550, WO 2011/144719, EP2014195046, EP2014195048 and EP2014195067.

The movable part of the cooking device can be driven by a motor. For example, such a system is described in EP 1 767 129. In this case, the user does not need to make physical efforts to open or close the brewing device. The brewer comprises a capsule insertion channel provided with a protective door installed in the movable part of the brewer by means of a switch to detect unwanted presence of a finger in the channel when closing and to prevent injury due to pinching. Alternative caps for the capsule insertion channel are described in WO 2012/093107 and WO 2013/127906. Various motorized systems are described in WO 2012/025258, WO 2012/025259 and WO 2013/127476.

Various systems are described in the art for enabling user interaction with such devices, for providing the device with operating instructions or for receiving feedback information from it, which, for example, are mentioned in the following sources: AT 410 377, CH 682 798, DE 44 29 353, DE 202 00 419, De 20 2006 019 039, De 2007 008 590, EP 1 448 084, EP 1 676 509, EP 08155851.2, FR 2 624 844, GB 2 397 510, US 4,377.049, US 4,458,735, US 4,554,419, US 4.767,632, US 4,954,697, US 5,312,020, US 5,335,705, US 5,372,061, US 5,375,508, US 5,645,230, US 5,685,435, US 5,731,981, US 5,836,236, US 5,959,869, US 6,182,555, US 6,354,341, US 6,759,072, US 2007/0157820, WO 97/25634, WO 99 /50172, Wo 2004/030435, Wo 2004/030438, Wo 2006/063645, Wo 2006/090183, Wo 2007/003062, Wo 2007/003990, Wo 2008/104751, Wo 2008/138710, Wo 2008/138820, Wo 2010 /003932, WO 2011/144720 and WO 2012/032019.

To improve the usability of the device, it is possible to automatically identify the capsule fed into the device and then automatically process and remove the capsule, as described in WO 2012/123440, for example.

There is still a need to improve beverage dispensing devices that automatically and reliably identify capsules.

Statement of the Invention

The invention relates to a beverage preparation device. The beverage preparation device may be intended for domestic or industrial use.

This appliance can be used to make coffee, tea, chocolate, cocoa, milk, soup, baby food, etc.

The preparation of a beverage typically involves mixing a plurality of beverage ingredients, such as water and milk powder, and/or extracting the beverage ingredient, such as an extract from ground coffee or tea, using water. One or more of these ingredients may be supplied as a free-flowing and/or agglomerated powder and/or in liquid form, in particular as a concentrate. Such an ingredient can be mixed with a carrier liquid or a diluent liquid, such as water, to form a beverage. Typically, a predetermined amount of beverage corresponding to a serving (eg, one serving) is prepared and dispensed at the user's request. The volume of such a serving may range from 25 to 200 ml and even up to 300 or 400 ml, for example the volume to fill a cup, depending on the type of beverage. The drinks that are prepared and served can be selected from ristretto, espresso, lungo, cappuccino, latte macchiato, latte, americano, tea, etc. In particular, the coffee machine can be configured to dispense espresso, for example, with an adjustable serving volume from 20 to 60 ml, and/or with the possibility of dispensing lungo, for example, with a serving volume in the range of 70 to 150 ml.

The apparatus of the invention has a unit for extracting a capsule containing a beverage ingredient in an extraction chamber for preparing a beverage. The block has a first part and a second part, which are made with the possibility of relative movement between a remote position for inserting a capsule and/or removing a capsule from the extraction chamber and a close position for fixing and extracting such a capsule in the extraction chamber. In a close position, the first and second parts, as a rule, limit the extraction chamber.

The capsule may comprise a capsule body, such as a substantially straight, conical, or disc-shaped body. The capsule may have a circular peripheral annular flange, such as a flexible or rigid flange, extending from a peripheral portion, such as an edge or surface, of the capsule body. The capsule may contain a flavoring ingredient for making tea, coffee, hot chocolate, cold chocolate, milk, soup, or baby food.

At least one of the first and second parts may define an ingredient receiving cavity, for example within a capsule, such as a tapered cavity, such as a conical or pyramidal cavity, or a straight cavity, such as a cylindrical, trapezoidal cavity, or disc-shaped cavity. Such a cavity may extend along an axis that is substantially collinear with the direction of relative movement of the first and second parts. Subsequently, the extraction chamber is limited on one side by such a cavity.

Another part of these first and second parts may be defined by another cavity and/or may include an extraction plate, such as a plate provided with piercing elements to open the flow surface of the capsule, or a non-penetrating plate to interact with the flow surface of the capsule, which has been previously opened or which is self-opening.

Examples of extraction units are described in WO 2008/037642 and WO 2013/026843.

At least one of these parts may have a capsule opening device, such as one or more capsule piercing devices.

The capsule may also include a self-opening mechanism. Self-opening capsules are, for example, described in CH 605 293 and WO 03/059778.

In the case of closed capsules, the first and second parts may include tools for opening the capsule, such as blades and/or tearing tool, such as a tearing profile plate, such as known among Nespresso™ devices or described in EP 0 512 470, EP 2 068 684 and WO 2014/076041, as well as in the sources cited in these documents.

At least one of the parts may have an opening for the inflow of liquid intended for mixing with the ingredient located in such a capsule.

The device includes a control unit for controlling the extraction unit in order to extract such a capsule. The control unit can be powered from the mains, for example, using an electrical cord.

The device has an outlet channel for dispensing a beverage formed by extracting such a capsule into a user's receiving vessel, such as a cup or mug, which is located in the area for receiving the receiving vessel.

The flavored beverage can be prepared by circulating (using a liquid transfer device, such as a pump) a carrier liquid, such as water, in a capsule to flavor the liquid by exposing the flavoring ingredient contained in the capsule, for example, along the direction of extraction, which may be essentially parallel to the direction of relative movement of the first and second parts, or to the longitudinal or central direction of extraction.

For example, the user's receptacle may be placed on a receptacle support to collect a beverage.

The receiving vessel support may be formed by an external placement support on which such a device is located.

The support of the receiving vessel may be formed by a support that is part of the device, for example, a movable or removable support of the device.

The area for receiving the receiving vessel may be associated with a support for the capacity of the device, designed to support such a capacity of the user under the outlet channel. The support may be: associated with a collection tray, such as a collection tray supporting the support; and/or is movable relative to the housing vertically under the outlet channel and/or at a distance under the outlet channel to enable placement of user containers of different heights under the outlet channel. Examples of suitable container supports are described in EP 0 549 887, EP 1 440 639, EP 1 731 065, EP 1 867 260, US 5,161,455, US 5,353,692, WO 2009/074557, WO 2009/074559, WO 2009/0745259, WO 2009/074559, WO 2009/113 WO 2009/113 WO 2009/113 2009/113 /154492, WO 2012/007313, WO 2013/186339, WO 2016/096705, WO 2016/096706 and WO 2016/096707.

In embodiments, the outlet channel may be attached to, or formed by, or installed on, or installed in:

- a device head that has an extended position in which the outlet port is located above the area for receiving the receiving vessel, and a retracted position in which the outlet channel is pushed inside the outer main body of the device, for example, the head of the device is moved in and out of the main body using at least one of the first and second parts or using an actuator controlled by a control unit; and/or

- a movable beverage guide that has a beverage dispensing configuration to dispense beverage into the receiving receptacle area and a beverage flow stop configuration to prevent beverage dispensing into the receptacle receiving area, for example by dripping off beverage residues from the guide through the edge of the guide into a waste receptacle, such as a beverage guide, movable between a dispensing configuration and a stop configuration by at least one of the first and second parts, or by some (or above) device head, or with using an actuator controlled by a control unit.

For example, the device is equipped with a device head as described in WO 2017/037212 and WO 2017/037215.

Examples of suitable waste receptacles for practicing the present invention are described in EP 1867260, WO 2009/074559, WO 2009/135869, WO 2010/128109, WO 2011/086087, WO 2011/086088, PCT/EP2017/050237 and WO 2017/037212.

The fluid guide may be completely hidden in the main body and/or head of the device.

Details of fluid guides suitable or adaptable for the practice of the present invention are found in WO 2006/050769, WO 2012/072758, WO 2013/127907, WO 2016/083488 and WO 2017/037212.

The extraction unit may include a capsule feeder for delivering a capsule to the extraction chamber, the feeder having a capsule dispenser with a release configuration for releasing such a capsule from the feeder towards the extraction chamber and a holding configuration for holding such a capsule on distance from the extraction chamber.

The capsule dispenser may be formed by a mechanical and/or magnetic capsule closure, such as a capsule holder, for example, which is shaped to complement or match at least a portion of the external shape of such a capsule.

The capsule holder may have a capsule closure that is movable, such as rotatable and/or translational, between a position blocking movement towards the extraction chamber and a position allowing movement towards the extraction chamber.

The capsule holder may have an actuator for switching from a holding position to a release position and vice versa, such as an actuator controlled by a control unit.

The capsule holder can be formed from at least part of the first part and/or the second part of the extraction unit, for example, the surface and/or edge of the first and/or second part, preventing the capsule from moving towards the extraction chamber when the first and second parts are in proximate position and/or when they are in an intermediate position between the remote position and the proximate position. Then the capsule holder is activated simultaneously with the first and second parts of the extraction unit.

Immediately after releasing the capsule towards the extraction chamber, the capsule dispenser can be changed from the release configuration to the hold configuration so that the extraction chamber is only accessed when the capsule is to be released.

The capsule delivery device may include a conduit for directing such a capsule to the extraction chamber in a predetermined orientation of the capsule for entry into the extraction chamber, such as a conduit associated with a capsule immobilization device for immobilizing such a capsule between the first and second parts in their remote position before moving relative to their close position.

The interaction between the first and second parts (and optionally the capsule guiding channel) and the ingredient capsule may be of the type described in WO 2005/004683, WO 2007/135135, WO 2007/135136, WO 2008/037642 and WO 2013/026856.

The control unit may control the capsule dispenser to release such a capsule from the dispenser when the first and second parts are in the remote position or move to the remote position, to enter such a capsule into the extraction chamber when the first and second parts are moved back to the close position.

The control unit may control the capsule dispenser to keep such a capsule in the dispenser and away from the extraction chamber when the first and second parts are:

- in a close position or make a relative movement to this position; or

- at a remote position and are about to make a relative movement to a close position so as to leave insufficient time to place such a capsule, if it were released from the dispenser, into the extraction chamber before the first and second parts reach the close position.

The capsule dispenser may include a capsule sensor coupled to or associated with a control unit, the control unit being configured to bring the capsule dispenser into a hold or hold configuration when the capsule sensor detects the absence of such a capsule from the capsule dispenser, or in him. Examples of capsule sensors are given, for example, in WO 2012/123440, WO 2014/147128, WO 2015/173285, WO 2015/173289, WO 2015/173292, WO 2016/005352 and WO 2016/005417.

The control unit may be configured to control the actuator so that the first and second parts are moved by the actuator: from a proximate position to a remote position and from a remote position to a proximate position after a predetermined period of time has elapsed, for example, starting from the start of preparation of a drink, such as, for example, capsule detection, capsule recognition, user activation of the user interface of the device, and the like. or a combination thereof, such as a predetermined period of time in the range of 3 to 15 seconds, namely 5 to 12 seconds, such as 7 to 10 seconds.

Examples of such parts that perform relative movement by means of an actuator (for example, a motor) are described in EP 1767129, WO 2012/025258, WO 2012/025259, WO 2013/127476 and WO 2014/056641.

For example, the first part and the second part are made with the possibility of relative movement, essentially, along a straight axis by means of an actuator from a close to a remote position and/or vice versa.

The device may include a liquid supply device for supplying a liquid, such as water, to the extraction chamber, the liquid supply device being connected to and controlled by a control unit to supply such liquid to the extraction chamber and interrupt such supply automatically and/or manually by means of a user interface connected to the control unit and/or upon detection of the removal of such a receiving vessel by a detection device. For example, a fluid delivery device includes one or more of: a source of said fluid, such as a fluid tank or a fluid connector for connecting to an external fluid source; one or more liquid tubes for directing such liquid into the extraction chamber; liquid transfer devices such as a pump, such as a solenoid pump (reciprocating piston pump), or a peristaltic pump or a diaphragm pump, for pumping such liquid into the extraction chamber; and a thermal conditioner, such as a heater and/or a cooler, such as an in-line thermal conditioner, such as an in-line flow conditioner, for thermal conditioning such a fluid.

Examples of suitable sources of liquid, such as tanks or connecting devices, are described in WO 2016/005349, EP2015194020.2, PCT/EP2017/050237, as well as in the sources cited in these documents.

The thermal air conditioner may be a boiler, or a thermoblock, or an instantaneous heater (ODH), such as the ODH type described in EP 1 253 844, EP 1 380 243 and EP 1 809 151.

Examples of pumps and their applications in beverage preparation devices are described in WO 2009/150030, WO 2010/108700, WO 2011/107574 and WO 2013/098173.

The control unit may be configured to control the liquid supply device to automatically supply liquid to the extraction chamber when:

- the first and second parts have reached their close position, while the capsule is located in the extraction chamber after moving the parts from the remote to the close position in order to combine said liquid with the ingredient in the capsule and form a drink for dispensing through the outlet channel, optionally after detection with some (or the above) capsule sensor for supplying such a capsule to the block; and/or

- the first and second parts have reached their close positions, but there is no capsule in the extraction chamber to wash or clean at least part of the block and optionally the outlet channel, and the liquid supply device, for example, is configured to supply liquid with a temperature for rinsing or cleaning, which is different from the temperature of such a liquid to form a drink, for example by brewing.

In a particular embodiment, cold or chilled beverages are also provided.

The control unit may be configured to control the liquid supply device so as not to automatically supply liquid to the extraction chamber when the first and second parts have reached their close position, but there is no (for example, detected or recognized) capsule in the extraction chamber. For example, the control unit is configured to control the liquid supply device to supply liquid to the extraction chamber upon detecting an appropriate manual user input on a user interface connected to the control unit.

The control unit may have an extraction completion control program that runs automatically when the liquid supply is interrupted (for example, when a predefined extraction process is completed or an error is detected during its execution) to:

- immediate relative movement of the first and second parts to their remote position, to remove any capsule from the position between the first and second parts; or

- maintaining the first and second parts in a close position for a predetermined period of time, for example, in the range from 1 to 5 s, namely from 2 to 3 s, in order to be able to enter the request manually, for example, through a user interface connected to the control unit, to supply, by means of a liquid supply device, an additional amount of liquid to the extraction chamber, and also, in the absence of such a manually entered request within a predetermined period of time, to move the first and second parts relative to their remote position, to remove any capsule from a position between the first and the second part, for example, to remove such a capsule in the collector of used capsules formed by some (or the above) waste receptacle.

For example, before moving the first and second parts to their proximate position, the parts may remain in their remote position for a predetermined period of time, namely for a period of time in the range of 1 to 6 seconds, such as 2 to 4 seconds, to allow insertion of a new capsules between the first and second parts before moving them relative to a close position with the new capsule located in the extraction chamber to extract the new capsule.

Thus, the user may request that two (or more) beverages (eg, a double espresso) be dispensed into the same user container.

According to the invention, a device for preparing and dispensing a beverage, such as tea, coffee, hot chocolate, cold chocolate, milk, soup or baby food, includes:

- an extraction unit for extracting a capsule with a beverage ingredient to form said beverage, wherein, for example, the unit has a first part and a second part that are movable relative to a remote position for insertion and/or removal of the capsule and a close position, such as a close position , in which the first and second parts define the extraction chamber, for fixing and extracting such a capsule, while optionally at least one of these parts has a device for opening the capsule, for example one or more devices for piercing the capsule, and / or at least one of these parts has a hole for the influx of liquid intended for mixing with the ingredient in such a capsule;

- a control unit for controlling the extraction unit in order to extract such a capsule, such as a control unit, powered from the power supply, for example, using an electric cord;

- an outlet for dispensing a beverage formed by extracting such a capsule into a user's receiving receptacle, such as a cup or mug, located in a receiving receptacle accommodating area, for example on a receptacle support, such as an external accommodating support, on which such a device is located, or a device support, such as a movable or removable device support, for collecting said beverage,

a capsule recognition module for recognizing a capsule inserted into the device at the capsule recognition position, the capsule recognition module comprising a camera for capturing an image of at least a portion of the capsule at the capsule recognition position;

wherein the capsule recognition module comprises a neural network computer, the neural network computer being configured to determine a capsule type among a plurality of predefined capsule types based on an image of at least a portion of the capsule captured by the camera.

The plurality of predefined capsule types, for example, includes a capsule type corresponding to capsules unknown to the capsule recognition module.

The computing device, for example, includes a computing device such as a microcontroller, microprocessor, or any other suitable computing device, and a neural network computer program that implements the neural network when it is executed on the computing device. The neural network is, for example, a convolutional neural network.

In embodiments, the device further comprises a network interface, such as a wired or wireless network interface. In addition, the neural network computing device, for example, may be configurable and/or may be updated remotely via a network interface.

In embodiments, the capsule recognition module comprises at least one light source, such as at least one LED, for illuminating the capsule at the capsule recognition position, and a diffuser for scattering light from the at least one light source towards the capsule recognition position. The diffuser preferably forms a tapering cavity extending from the camera towards the capsule recognition position and open towards the capsule recognition position.

In accordance with the present invention, a method for configuring a neural network computing device of such a device includes:

- training the computer program of the neural network outside the computing device of the neural network by inputting several images of capsules of predetermined types until the neural network is able to correctly determine the type of each subsequent image of the capsule with a probability above a given threshold;

- copying the trained neural network computer program to the neural network device of the beverage preparation device.

In accordance with the present invention, a method for updating a neural network computing device of such a device includes:

- training the computer program of the neural network outside the computing device of the neural network by entering several images of capsules of predefined types, and the predefined types include a new, previously unknown type, until the neural network is able to correctly determine the type of each subsequent image of the capsule with probability above a given threshold value;

- copying the trained neural network computer program to the neural network device of the beverage preparation device.

The step of copying the trained neural network computer program to the neural network computing device of the device can be performed via a data network, such as the Internet, for example, from a remote server to said neural network computing device.

The present invention also relates to a neural network computer program for determining a capsule type among a plurality of predefined capsule types based on an image of at least a portion of said capsule captured by a beverage preparation device camera.

The neural network computer program preferably implements a convolutional neural network when it is executed on a computing device. In embodiments, the implementation of the convolutional neural network contains at least one convolutional layer and at least one fully connected classification layer. In embodiments, for example, a convolutional neural network comprises three convolutional layers, each followed by a pooling layer, and a two-layer fully connected classifier.

Using a properly trained neural network to recognize the type of capsule allows you to quickly determine the type of capsule.

In addition, it avoids the implementation of image comparison methods in the beverage preparation device, which require high computing power and a large storage system for storing images of samples of each type of capsule.

Brief description of graphic materials

The following is a description of the invention with reference to schematic drawings, and:

in fig. 1 is a perspective view of a device according to the invention;

in fig. 2 is a cross-sectional view of a capsule recognition module in accordance with one embodiment of the present invention;

in fig. 3 is an exploded view of the capsule recognition module shown in FIG. 2;

in fig. 4 is a cross-sectional view of a capsule recognition module according to another embodiment of the present invention;

in fig. 5 is an exploded view of the capsule recognition module shown in FIG. four.

Detailed description

In FIG. 1 shows an exemplary embodiment of a beverage preparation device 1 according to the invention for preparing and dispensing a beverage such as tea, coffee, hot chocolate, cold chocolate, milk, soup or baby food. The ingredient may be presented in the form of an ingredient capsule, for example of the type described above in the "Scope of the Invention" section.

The apparatus 1 includes an extraction unit for extracting a capsule containing a beverage ingredient to form a beverage. For example, the extraction unit consists of a first part and a second part, preferably located inside the body of the device and thus not visible in FIG. 1 that are movable relative to between a remote position for inserting and/or withdrawing the capsule and a proximate position, such as a proximate position in which the first and second parts delimit the extraction chamber for capturing and extracting the capsule. For example, at least one of the parts has a device for opening the capsule, for example, one or more devices for piercing the capsule, and/or at least one of the parts has a hole for the inflow of a liquid intended for mixing with the ingredient contained in the capsule.

The device 1 comprises a control unit, preferably located inside the body of the device, for controlling the extraction unit for extracting the capsules. The control unit can be powered from the mains, for example, in a known manner by means of an electrical cord, or from a direct current source, such as a battery, such as a car battery, or a portable battery, or a device battery.

The device 1 has an outlet 16 for dispensing beverages obtained by extracting such capsules into a user's receptacle (not shown), such as a cup or mug, located in the area 17 for receiving a receptacle for collecting the beverage.

In embodiments, outlet channel 16 is, for example, attached to, or mounted on, or formed by, head 10 of the device, with outlet channel 16 located above area 17 for receiving receptacle.

The extraction unit contains an actuator, made with the possibility of relative movement of the first and second parts between their relatively remote and near positions. The actuator is connected to the control unit of the device and is controlled in this way for the relative movement of the first and second parts.

The control unit is connected to the input device to start and/or control the extraction unit. According to the present invention, the input device includes, for example, a user interface 3, namely a touch screen.

As shown in FIG. 2, the extraction unit may include a capsule feeding device 2 for feeding a capsule into the extraction chamber and a capsule recognition unit 4 for recognizing the type of capsule fed into the extraction chamber. For example, the capsule recognition module 4 recognizes the type of capsule while the capsule is in the capsule recognition position 20 inside the capsule feeder 2 and before being fed into an extraction chamber, which is not shown in the figures and, for example, is located below the capsule recognition position 20.

Preferably, the device further comprises a capsule sensor, such as a mechanical and/or optical sensor (not shown), for detecting the presence of a capsule approaching and/or placed at the capsule recognition position 20.

The capsule delivery device 2 may have a holding means (not shown) for holding the capsule in the capsule recognition position 20 . The holding means generally has a release position for releasing the capsule from the capsule recognition position 20 towards the extraction chamber, for example by gravity, and a holding position for holding the capsule away from the extraction chamber, for example above the extraction chamber, at the recognition position 20 capsules.

The capsule delivery device 2 may have a channel for directing the capsule into the extraction chamber with a predetermined orientation of the capsule for insertion through the capsule recognition position into the extraction chamber. The channel, for example, includes guide grooves 21 into which the edge of the capsule is inserted to pass the capsule through the channel, thus maintaining the capsule in a predetermined orientation, preferably suitable for recognition by the capsule recognition module 4 and/or its introduction into the extraction chamber.

The control unit can control the holding means to release the capsule from the capsule recognition position 20 when the first and second parts of the extraction unit are in the removed position, or when they move to the removed position, to introduce the capsule into the extraction chamber when the first and second parts are moved back. to their near position. The holding means, for example, includes the surface and/or edge of the first and/or second part of the extraction unit.

The capsule recognition module 4 is preferably connected to a control unit and configured to recognize the type from among the predetermined types of the capsule located at the capsule recognition position 20 .

As described in more detail below, the capsule recognition module 4 recognizes the type of capsule by capturing an image of said capsule and passing it as input to the trained neural network computing device.

The control unit is, for example, configured to control the liquid supply device in accordance with a liquid supply program associated with the type of capsule, such as a liquid supply program with one or more adjustable liquid supply parameters selected from liquid temperature, flow rate, pressure and volume. , which(s) is(are) constant(s) or variable(s) during the extraction of the recognized capsule. For example, the type of capsule may be selected from a variety of known predetermined types of capsules capable of being extracted in the extraction chamber of the device.

The capsule recognition module 4 is preferably arranged along the channel of the feed device 2 at the height of the capsule recognition position 20 .

During use of the device according to the present invention, the following steps may be performed:

- placing the receiving vessel in the area for receiving the receiving vessel;

- introducing the capsule into the device 2 for supplying the capsule;

- recognition of the type of capsule by the capsule recognition module 4;

- relative movement of the first and second parts of the extraction unit to their remote position automatically, semi-automatically or manually;

- feeding the capsule into the extraction chamber;

- relative movement of the first and second parts to their close position in order to position the capsule in the extraction chamber;

- extracting the capsule in the extraction chamber using extraction parameters determined on the basis of the recognizable type of capsule to prepare a beverage; and

- delivery of the prepared drink through the outlet channel 16 into the receiving vessel.

According to the present invention, the capsule recognition module 4 is configured to determine the type of capsule inserted into the beverage preparation device at the capsule recognition position 20 by capturing an image of said capsule and processing said image with a neural network computer.

The beverage preparation device typically extracts different types of capsules to prepare different drinks and/or different types of drinks. Various types of capsules capable of being extracted in an extraction chamber, for example, correspond to different ingredients contained therein and/or different conditioning conditions for the ingredients.

In embodiments, each type of capsule corresponds to a particular type of coffee that differs from the coffee contained in other types of capsules, for example, but not exclusively, in one or more of origin, roast, grind, amount contained in the capsule, and/or content caffeine. Alternatively, or as a combination, the different types of capsules extractable in the beverage preparation device 1 correspond to the ingredients for the preparation of various beverages such as, for example, coffee, milk, soup, infant formula, tea, cold drinks, and the like. .

Preferably, each type of capsule is associated with a specific aspect of the capsule, such as a color or combination of colors, specific symbols and/or patterns formed on the surface of the capsule, and the like, which allows, for example, the user to visually distinguish between different types of capsules. The neural network device of the capsule recognition module 4 comprises a neural network program previously trained to recognize the type of capsule based on a digital image of at least a portion of said capsule. The neural network device typically includes a data processor configured to execute the neural network program.

The device 1 may be configured to extract each capsule using preparation parameters specific to the particular type of capsule. For example, the preparation parameters include one or more of: carrier liquid temperature, carrier liquid volume, extraction time, carrier liquid pressure, carrier liquid type, series of sequential preparation steps, and the like. The preparation parameters used for each type of capsules made with the possibility of extraction in the device 1 are preferably stored in an internal or external data storage device connected or connected to the control unit and/or to the capsule recognition module 4. Appropriate preparation parameters for a particular capsule are selected based on the type of capsule determined by the capsule recognition module 4 and are used in the control unit to control the extraction of the recognized capsule.

The device 1 can also be configured to store and/or transmit to an external terminal and/or server information about the type of each capsule extracted in the device for, for example, monitoring the consumption of capsules in the device 1.

As shown in FIG. 2, the capsule recognition module 4 comprises a light source, such as one or more white LEDs 40 or any other suitable light source, preferably with a known and defined spectrum, a camera 41, such as a CCD camera, and a neural network computing device, not shown in the figures, for example, in the form of a microcontroller, microprocessor, or other appropriate computing device configured to execute the neural network computer program. The capsule recognition module 4 preferably further comprises a controller, not shown in the figures, such as, but not limited to, an application specific integrated circuit (ASIC) or programmable microcontroller, to control the light source, the camera 41 and the neural network computer. Alternatively, the neural network computing device and the capsule recognition module controller are the same component or are made as a single unit as a single component. The controller of the capsule recognition module 4, for example, is configured to turn on and off the light source and/or receive and process signals from the camera 41. The light source, the camera 41, the neural network computing device and the module controller are preferably attached, for example, soldered, to the electronic board 42, typically a printed circuit board (PCB), which provides the necessary power and data connection and/or in-circuit connection in a known manner. The controller is preferably connected to and controlled by the control unit of the device 1.

Preferably, the capsule recognition module 4 also comprises a light guide 43 and a diffuser 44 for directing the light emitted by the light source towards the capsule recognition position 20 and/or for limiting the amount of light received by the camera 41, preferably to the amount of light reflected by the capsule located at position 20 capsule detection to avoid detecting excessive light such as ambient light.

The light source is preferably a distributed light source. In embodiments, the light source, for example, includes a plurality of LEDs 40, which are preferably evenly distributed, for example, around the camera 41. In the illustrated embodiments, the light source, for example, includes four LEDs 40 distributed around the camera 41, with only two of them being schematically depicted. in fig. 2. However, within the scope of the present invention, it is possible to use other types of distributed light sources and/or with a different distribution, such as, for example, an annular light source surrounding the camera 41.

The light guide 43, for example, is made in the form of a substantially transparent cover associated with the light source and/or the camera 41, for example at least partially enclosing him/her. The lid has, for example, holes or other guiding means for directing light through the diffuser 44 to and away from the capsule recognition position 20. In the embodiment shown schematically in FIG. 2, light guide 43, for example, includes light guide protrusions 430, each protrusion 430 extending from light source LED 40 toward diffuser 44 to direct light emitted from LED 40 towards diffuser 44.

The light guide 43 and/or the diffuser 44 are preferably designed in such a way as to prevent the light emitted by the light source from directly reaching the recognition position 20 of the capsule. In the embodiment shown in FIG. 2, the light guide protrusions 430, for example, are designed in such a way that they direct most of the directed light into the diffuser 44 to the area of the inner wall of the diffuser where it will be reflected, therefore in a direction transverse to the longitudinal axis of the diffuser 44 passing from the camera 41 to the recognition position. 20 capsules. The ends of the light guide protrusions 430 on the side of the diffuser 44, for example, are beveled at an angle of 45° such that most of the directed light is reflected on the beveled surface and directed towards the diffuser 44 at an angle substantially perpendicular to the longitudinal axis of the diffuser 44.

In the example shown, the diffuser 44 preferably provides a light color, such as white, and the opaque element forms a truncated conical cavity, open towards the recognition position 20 of the capsule. However, within the scope of the present invention, the diffuser may take other forms. However, the diffuser preferably forms a tapering cavity extending from the chamber 41 to the capsule recognition position, preferably open in the direction of the capsule recognition position.

The diffuser 44 is preferably configured to prevent reflections from the capsule located at the capsule recognition position 20 and/or in the capsule recognition module, which could lead to erroneous illumination of the capsule and hence its erroneous recognition. The inner surface of the diffuser 44, for example, is textured to improve the dispersion of light passing from the light source to the recognition position 20 of the capsule.

The diffuser 44 optionally includes openings and/or protrusions that block and/or deflect light rays emitted by the light source in the direction of the capsule recognition position 20. Alternatively, or as a combination, the light diffuser is made of a transparent or translucent material, and the light from the light source is directed, for example, by means of a light guide through the walls of the light diffuser, which leads to the refraction of light with distribution from the wall of the light diffuser into the cavity in front of the recognition position of the capsule.

In FIG. 4 schematically shows an embodiment of a light diffuser 44 having projections 440 formed on its inner surface to prevent light, such as light guided by light guide 43 and exiting light guide 43 through light guide projections 430, from directly reaching capsule recognition position 20.

In the embodiment shown, the diffuser 44 forms a semi-ellipsoidal cavity open towards the capsule recognition position 20, with the camera 41 preferably located at the end opposite the capsule recognition position 41.

In the illustrated embodiment, the protrusions 440 at least partially enclose the ends of the light guide protrusions 430 of the light guide 43. The protrusions 440 are designed to block light refracted from the beveled ends of the light guide protrusions 430 in a direction proximal to the longitudinal axis of the diffuser 44, to avoid direct hit of this light to the capsule recognition position 20, i. e. without initially reflecting at least once from the inner wall of the diffuser 44. However, the protruding elements 440 of the diffuser 44 preferably have an opening configured to transmit light refracted towards the inner wall of the diffuser 44 into the diffuser cavity.

Optionally, the device 1 comprises a capsule detector for detecting a capsule located at or approaching the capsule recognition position 20 . The capsule detector, for example, is contained in the capsule recognition module 4, preferably attached, for example soldered, to the electronic board 42. However, other arrangements of the capsule detector are possible within the scope of the present invention. The capsule detector may be of any suitable type, such as a presence and/or motion detector, such as an infrared (IR) detector, an inductive and/or resistive detector, a mechanical switching element, and the like. The capsule detector is controlled, for example, by the controller of the capsule recognition module 4 or directly by the control unit of the device.

In embodiments, when a capsule approaches the capsule dispenser 2 and/or is placed in the capsule recognition position 20, the capsule detector detects the presence of the capsule and sends an appropriate signal to the controller and/or device control unit, which activates the light source 40 to illuminate the at least a portion of the capsule surface located at the capsule recognition position 20. The camera 41 is in turn activated to capture an image of at least a portion of the illuminated capsule. The captured digital image is then transmitted to and processed by the neural network computing device to determine the capsule type. In other embodiments, for example, if the device does not include a capsule detector, the capsule recognition module 4, in particular the light source 40 and the camera 41, is activated, for example, by activating the user interface of the device, for example, by activating the drink selector.

Preferably, camera 41 is a CCD sensor or any suitable camera that provides a digital image of at least a portion of the capsule, such as a color or grayscale image. The image is transmitted as input to the neural network computing device, which runs the neural network computer program using the specified input data to determine the type of capsule, and the parameters of the neural network, in particular its synaptic weights and deviations, are set in advance during training of the neural network, implemented using a neural network computer program.

The neural network for determining the type of capsule in accordance with the present invention is trained manually and/or automatically, for example, on a training stand containing a capsule feeder for feeding capsules of various types to one or more capsule recognition positions, each capsule recognition position on the training stand provided with a camera to acquire an image of at least a portion of each capsule fed into the respective capsule recognition position. Preferably, the illumination and image capture conditions at the capsule recognition positions on the training stand are similar to those at the capsule recognition position on the beverage preparation device. The resulting images are then transmitted as input to a computing device running the neural network computer program of the present invention. Program output, i.e. neural network, for each image is compared with the actual type of the corresponding reference capsule, and the result of the comparison is passed back to the program so that it corrects the parameters of the neural network, in particular, the weights and / or deviations of one or more synaptic connections.

After reaching the required rate of successful recognition, the neural network is preferably considered sufficiently trained, and the neural network program with corrected parameters can be, for example, copied and loaded into neural network computing devices embedded or to be embedded in beverage preparation devices according to the present invention.

The neural network computer program of the present invention can be updated to recognize a new type of capsule with a new training session in accordance with the scheme described above, wherein the capsules include the new type of capsules. After the neural network remembers the new type of capsule, the computer program with the updated parameters can be copied and loaded into the computing devices of the neural network of functional devices to update their capsule recognition module to a newer version. Optionally, one capsule type out of the various recognizable capsule types corresponds to capsules unknown to the device.

Training of the neural network computer program according to the present invention is preferably carried out on a training stand outside of the beverage preparation apparatus. However, the imaging conditions of the exemplary capsules are preferably similar or even identical to those specified in the device.

An example of a suitable architecture for a neural network implemented with the neural network computer program of the present invention is shown in the table below.

Table 1. An example of a suitable neural network architecture

Level type In. Ex. Input Size Options Actions one Convolution one 32 1x128x128 320 9652608 2 ReLU 32 32 1x126x126 - 508032 3 Subsampling with maximum value determination 32 32 1x126x126 - 508032 four Convolution 32 32 1x63x63 9248 68704544 5 ReLU 32 32 1x61x61 - 119072 6 Subsampling with maximum value determination 32 32 1x61x61 - 119072 7 Convolution 32 32 1x30x30 9248 14475776 eight ReLU 32 32 1x28x28 - 25088 9 Subsampling with maximum value determination 32 32 1x28x28 - 25088 ten Linear one one 6272 401408 802880 eleven ReLU one one 64 - 64 12 Linear one one 64 704 1419 Total 421003 94941675

According to this example, the neural network is a convolutional neural network containing twelve levels, with the first convolutional level having an input size of 128x128x1. Thus, a neural network typically expects a grayscale image of 128 pixels by 128 pixels as input and performs an 11-channel classification. A neural network, for example, has five core layers: three convolutional layers followed by two fully connected layers. At each convolutional level, 3x3x1 program code fragments are used with a step of 1 and without filling, with 32 feature maps being output. Each convolution is followed by a ReLU non-linearity and a subsampling operation to determine the maximum value. The output of the feature extraction component is then passed to a two-level fully coupled classifier, such as the LogSoftMax classifier. The hidden layer of the classifier, for example, is built with sixty-four neurons and also uses the ReLU activation function.

Preferably, before the input image data is transmitted by the neural network of the present invention, it is subjected to a three-stage pre-processing. In the first stage of pre-processing, the data is preferably determined in the region of zero by subtracting the average activation calculated over the entire training set for each pixel. In the second preprocessing step, the data is separated using the standard global deviation of the data set to normalize the range of input values. In the third and final pre-processing step, local contrast normalization is performed using a thirteen x thirteen Gaussian window with weighting for gain at the edges of the image.

However, other neural network architectures and/or data pre-processing are possible within the scope of the present invention to provide reliable capsule type recognition based on a digital image of a capsule or at least a portion thereof.

Claims (33)

1. A device for preparing and dispensing a drink, comprising: - an extraction unit for extracting a capsule with a beverage ingredient to form said beverage; - a control unit for controlling the extraction unit in order to extract such a capsule; - an outlet channel for dispensing said beverage, formed by extracting such a capsule, into the user's receiving vessel located in the area for receiving the receiving vessel, a capsule recognition module for recognizing a capsule inserted into said device at a capsule recognition position, the capsule recognition module comprising a camera for capturing an image of at least a portion of said capsule at said capsule recognition position; characterized in that the capsule recognition module comprises a neural network computing device, wherein said neural network computing device is configured to determine the type of said capsule among a plurality of predefined capsule types based on an image of at least a part of said capsule captured by said camera. 2. The device according to the preceding claim, wherein said extraction unit has a first part and a second part that are movable relative to a remote position for insertion and/or removal of the capsule and a close position, such as a close position in which the first and second parts limit the extraction chamber to fix and extract such a capsule. 3. Apparatus according to the preceding claim, wherein at least one of said parts has a capsule opening device, such as one or more capsule piercing devices, and/or at least one of said parts has an opening for the inflow of liquid to be mixed with the ingredient contained in such a capsule. 4. Apparatus according to any one of the preceding claims, wherein said control unit is powered by a mains power supply by means of an electrical cord. 5. An apparatus according to any one of the preceding claims, wherein said receiving receptacle accommodating area is a receptacle support which is an external accommodating support on which the device is located, or a movable or removable device support for collecting said beverage. 6. An apparatus according to any one of the preceding claims, wherein said plurality of predefined capsule types comprises a capsule type corresponding to capsules unknown to said capsule recognition module. 7. An apparatus according to any one of the preceding claims, wherein said neural network computing device includes a computing device, such as a microcontroller, microprocessor, or any other suitable computing device, and a neural network computer program that implements the neural network when executed on said computing device. . 8. The apparatus of the preceding claim, wherein said neural network is a convolutional neural network. 9. The device according to any one of the preceding claims, further comprising a network interface. 10. The apparatus of the preceding claim, wherein said network interface is a wired or wireless network interface. 11. The apparatus of the preceding claim, wherein said neural network computing device is configurable and/or can be updated remotely via said network interface. 12. An apparatus according to any one of the preceding claims, wherein said capsule recognition module comprises at least one light source for illuminating the capsule at said capsule recognition position and a diffuser for scattering light from said at least one light source towards said recognition position capsules. 13. The apparatus of the preceding claim, wherein the at least one light source is at least one LED. 14. An apparatus according to claim 12 or 13, wherein said diffuser forms a tapering cavity extending from said chamber to said capsule recognition position and open towards said capsule recognition position. 15. A device according to any one of the preceding claims, wherein said beverage is tea, coffee, hot chocolate, cold chocolate, milk, soup or baby food. 16. A device according to any one of the preceding claims, wherein said user vessel is a cup or mug. 17. A method for configuring a neural network computing device of a device according to any one of the preceding claims, comprising: training the neural network computer program outside said neural network computing device by inputting several capsule images of said predefined types until said neural network is able to correctly determine the type of each successive capsule image with a probability higher than a given threshold value; - copying said trained neural network computer program to said neural network computing device of said beverage preparation device. 18. The method for updating the computing device of the neural network of the device according to any one of the preceding claims, including: - training a neural network computer program outside of said neural network computing device by inputting multiple images of capsules of said predefined types, said predefined types including a new, previously unknown type, until said neural network is able to correctly determine the type of each subsequent imaging of the capsule with a probability above a predetermined threshold; - copying said trained neural network computer program to said neural network computing device of said beverage preparation device. 19. The method according to the preceding claim, wherein said copying step is performed via a data network. 20. The method according to the preceding claim, wherein said data communication network is an Internet network, and said copying step is performed from a remote server to said neural network computing device. 21. A method according to any one of the preceding claims, wherein the neural network computer program is for determining a capsule type among a plurality of predefined capsule types based on an image of at least a portion of said capsule captured by a beverage preparation device camera. 22. The method of the preceding claim, wherein the neural network computer program implements the convolutional neural network when said neural network computer program is executed on said neural network computing device. 23. The method of the preceding claim, wherein said convolutional neural network comprises at least one convolutional layer and at least one fully connected classification layer. 24. The method of claim 22, wherein said convolutional neural network comprises three convolutional layers, each followed by a pooling layer, and a two-layer fully connected classifier.

Images