Method for operating a cooking appliance

Abstract

The invention relates to a method for operating a cooking appliance (1), wherein the cooking appliance (1) comprises: • a cooking chamber (12) which is designed to receive food to be cooked through an access opening (13), preferably along the longitudinal axis (X), • a locking element (14) which is designed to close and release the access opening (13) of the cooking chamber (12), and • an image capture element (17) which is designed and arranged to optically capture the cooking chamber (12), preferably along the vertical axis (Z) from above, wherein the method comprises at least the following steps: • Capturing (200) the cooking chamber (12) using the image capture element (17), • Recognizing (300) the contents of the cooking chamber (12) based on a comparison (350) of the captured image data with at least two classes of predetermined image data, wherein the classified predetermined image data includes at least one class of impermissible objects and one class of cooking goods, and, • in the event of a detected impermissible object in the cooking chamber (12), at least the prevention (950) of a cooking process, wherein the comparison (350) of the captured image data with the predetermined image data is carried out using a neural network, preferably in a cloud.

Description

[0001] The invention relates to a method for operating a cooking appliance, a corresponding cooking appliance and a system therefor.

[0002] Cooking appliances are known which have a fixed, i.e., stationary, cooking chamber formed with the cooking appliance, in which the food can be placed in or on a cooking utensil and cooked with the cooking chamber closed. Such a cooking appliance can be, for example, an oven, a steamer, a combination oven with a steamer, and the like. Such cooking appliances with a fixed cooking chamber typically have a height of approximately 45 cm along the vertical axis, at least in Germany, although ovens with a height of approximately 60 cm are also common.

[0003] Such cooking appliances with a fixed cooking chamber have in common that they have an outer housing which essentially encloses the cooking appliance on the outside and protects its individual components and elements, as well as making them easier to handle together. Within the cooking appliance, an interior space is formed which represents the cooking chamber and is essentially enclosed by an inner housing, also referred to as a cooking chamber muffle or muffle. Between the inner housing and the outer housing, a housing space is formed as an intermediate space, in which functional elements of the cooking appliance, such as a control system or control unit, an electrical power supply and other components which serve the intended use of the cooking appliance can be arranged. In depth from the front, i.e. viewed from the user's perspective, the outer housing ora panel in the form of a furniture material or in the form of a control panel and the like together with the inner housing closes off the housing space so that the housing space is not accessible to the user.

[0004] The interior of the cooking appliance has a through-opening at the front in depth as an access opening, through which the interior of the cooking appliance is accessible to the user in order to insert cooking utensils into the interior of the cooking appliance as its cooking chamber and to arrange them there, as well as to remove cooking utensils from the interior of the cooking appliance after the food has been cooked. The access opening can be opened by the user by means of a closure element, for example in the form of a door that can be pivoted to the side, a flap that can be pivoted downwards, or the like, in order to access the interior of the cooking appliance, as described above, or to close the interior of the cooking appliance and carry out the cooking operation or process. Such a closure element can be closed or have a viewing window to allow the user a view into the closed interior of the cooking appliance.

[0005] Such cooking appliances with a fixed cooking chamber are usually designed as built-in appliances or as built-in kitchen appliances in order to be arranged with their outer casing in a space-saving manner and at a height that is easily accessible for the user along the vertical axis in kitchen furniture, such as in built-in kitchen cupboards, and to be flush with the surfaces of the other cooking appliances, drawers, doors and the like of the kitchen furniture towards the front of the user with their locking element, and if necessary also with their panel (see above), which can improve the visual impression for the user. However, such cooking appliances can also be designed as individual, separate appliances that can be set up and operated, in particular, on a work surface in a kitchen unit. This can simplify flexibility of use as well as procurement and disposal.

[0006] The rapid cooling function is well-known in such cooking appliances, especially ovens. This function allows the food and the cooking chamber to be cooled relatively quickly after the cooking process has ended, i.e., brought from the relatively high temperature of the preceding cooking process to a significantly lower temperature, such as the usual ambient or room temperature, relatively quickly or at an accelerated rate.

[0007] This function is useful when an automatic cooking process has ended and further cooking of the food, which would or could lead to overcooking, is to be avoided in order to maintain or maintain the desired cooking result according to the automatic cooking process. This can, in particular, prevent the food from burning.

[0008] The rapid cooling function can also be used if an automatic program is to be started immediately after a cooking process that is designed for a "cold" cooking chamber, i.e., a cooking chamber at ambient or room temperature. If the temperature of the cooking chamber does not match this initial situation, i.e., is significantly higher than ambient or room temperature, this can affect the automatic cooking process and impair the cooking result.

[0009] The "rapid cooling" function can be implemented, for example, by automatically opening the cooking chamber's closing element, such as its door or flap, a little at the end of the cooking process so that the warm or hot air from the previous cooking process can escape through the gap from the cooking chamber to the outside of the cooking appliance. This can accelerate the cooling of the cooking chamber. Furthermore, if a cooling fan of the cooking appliance is operated as part of the rapid cooling function and ambient air is drawn into the cooking chamber, this can improve the discharge of the heated air to the outside of the cooking appliance and thus accelerate the cooling of the cooking chamber. Once the ambient or room temperature is reached, the rapid cooling function can be automatically ended.

[0010] To further increase the user's convenience during cooking processes and to improve the cooking result and ensure its success, it is also known that cooking appliances, such as ovens or the like, communicate and, in particular, display steps and information related to the cooking process to the user, as well as independently perform steps or settings related to the cooking process. This also includes the rapid cooling described above.

[0011] This allows cooking appliances to prepare food at least partially autonomously. For example, a cooking chamber camera detects the food or the item to be cooked. The cooking chamber camera can identify the food to be cooked when the appliance is started, and the cooking appliance can then suggest a function suitable for the item.

[0012] In any case, the cooking appliance should be able to detect or receive information about the condition or status of the cooking chamber, or whether there is food or what type of food in the cooking chamber. The autonomous operation of the cooking appliance may depend on this information or the result of detecting the contents of the cooking chamber. This can ensure the safety of the cooking appliance and the surrounding area by preventing autonomous or assisted operation of the cooking appliance in certain situations or in certain cases where the contents of the cooking chamber are present, or by restricting this operation, for example, to certain permitted or non-critical functions.

[0013] WO 2022 / 010748 A1 describes a method for controlling a cooking chamber, which comprises: • Receiving a preheat command in conjunction with an expected cooking chamber condition; • Capturing a sensor measurement for the cooking chamber of an appliance; • Determining an assumed cooking chamber condition using the sensor measurement; • Determine whether the assumed cooking chamber condition matches the expected cooking chamber condition; • if the assumed condition of the cooking chamber matches the expected condition of the cooking chamber, operate the appliance according to the preheating instructions; and • If the assumed cooking chamber condition does not match the expected cooking chamber condition, the appliance will not operate according to the preheating instructions.

[0014] WO 2022 / 184483 A1 describes a method for operating a cooking appliance, which comprises at least one processing chamber with at least one opening for filling the processing chamber, at least one door for closing the opening of the processing chamber, at least one control device, and at least one safety device. The safety device checks the processing chamber for the presence of at least one living being and / or at least one potentially disruptive object. If the safety device detects the presence of at least one living being and / or at least one potentially disruptive object in the processing chamber, the control device prevents the operation of the household appliance and / or restricts the operation of the household appliance to certain operating states.

[0015] The potentially malfunctioning object is in particular an object that is unsuitable for the cooking appliance and which, for example, damages the cooking appliance. The potentially malfunctioning object can preferably also be a safety-relevant object. Such an object can, for example, be a flammable object such as a newspaper, a cuddly toy and / or an item of clothing. It can also be an electronic device and, in particular, an electronic device with an energy storage device, such as a mobile phone. In the case of a cooking appliance such as an oven, depending on the design, it can also preferably be checked, for example, whether functional cookware that is not oven-safe has been placed in the cooking compartment.

[0016] This means that the cooking appliance can also check when an operating mode is started whether the load status “Empty” or “Loaded” matches the selected operating mode and, if necessary, a warning is issued.

[0017] However, problems have arisen so far when inadmissible, i.e., prohibited, objects are placed in the cooking chamber of the cooking appliance. Since these objects are not known in advance and thus have not been trained for classification, they cannot be recognized. The cooking appliance is therefore unable to make a meaningful decision about its load status or take meaningful actions. Therefore, objects in the cooking chamber that have not been pre-trained cannot be recognized. In this case, the appliance cannot recognize the unknown object as food or other object and cannot suggest a meaningful function.

[0018] Please note that the "Loaded" and "Empty" states are not sufficient to determine whether the load condition matches the selected operating mode, as even prohibited items such as potholders would most likely be detected as "loaded." These items would be destroyed during operation and, in the worst case, could cause a fire in the appliance.

[0019] The invention addresses the problem of improving the ability of a cooking appliance to detect its load status or cooking chamber contents. This applies in particular to the resulting functions of the cooking appliance, especially the partially automatic execution of cooking processes. In any case, this should be as simple, reliable, and / or cost-effective as possible. At the very least, an alternative to such known options should be created.

[0020] According to the invention, this problem is solved by a method, a cooking appliance, and a system having the features of the independent claims. Advantageous embodiments and further developments of the invention are set forth in the following dependent claims.

[0021] Thus, the invention relates to a method for operating a cooking appliance, which can in particular be an oven or a combination appliance with an oven function.

[0022] The cooking appliance has: • a cooking chamber which is designed to receive a food to be cooked through an access opening, preferably along the longitudinal axis, • a closure element which is designed to close and open the access opening of the cooking chamber, and • an image capture element which is designed and arranged to optically capture the cooking chamber, preferably along the vertical axis from above.

[0023] The method comprises at least the following steps: • Capturing the cooking chamber using the image capture element, • Detecting the contents of the cooking chamber based on comparing the captured image data with at least two classes of predetermined image data, wherein the classified predetermined image data comprises at least one class of inadmissible objects and one class of cooking goods, and, • If an unauthorized object is detected in the cooking chamber, at least stopping a cooking process. The comparison of the captured image data with the predetermined image data is carried out using a neural network, preferably in a cloud. Neural networks can also be stored (embedded) in the cooking appliance's control system, which enables classification in the appliance without a connection to a cloud. The advantage here is that the classification function works even without networking and an internet connection. The disadvantage in this case is that improvements via remote updates or checking for faults are not possible remotely.

[0024] According to the invention, inadmissible objects that are neither food nor suitable for carrying out a cooking process can be detected, and in such cases, the cooking process can be prevented from taking place in the cooking chamber. This can increase the safety of the cooking process.

[0025] In this process, exemplary prohibited objects whose accidental presence in the cooking chamber is expected or cannot be ruled out are used in advance based on their image data to create a class of prohibited objects. Existing image data of prohibited objects as well as image data from the cooking chamber with the image acquisition unit located there are preferably used to train the neural network. Common cooked items are also classified or trained.

[0026] During operation, particularly at the start of operation and / or when the closure element is closed, image data from the cooking chamber can be captured and made available to the neural network, allowing the neural network to classify the current view of the cooking chamber and at least distinguish between food being cooked and prohibited objects. In the latter case, the execution of a cooking process can be prevented or blocked, in particular without preventing or restricting other functions of the cooking appliance.

[0027] What makes this combination of features unique compared to the state of the art is that, thanks to the use of a neural network that has also been trained on a class of prohibited objects, even unknown prohibited objects can be detected and correctly classified in this way. This is advantageous because the manufacturer of the cooking appliance cannot predict or even conceive which prohibited objects might actually be accidentally or intentionally placed in the cooking chamber by the user.

[0028] The image capture element can, in particular, be a camera for capturing a wide area of ​​the cooking chamber, preferably with depth information. The optically captured image data can then be processed using known image processing and pattern recognition methods to identify the item being cooked, or various items being cooked, as well as inadmissible objects from the optically captured image data.

[0029] According to one aspect of the invention, if an unauthorized object is detected in the cooking chamber, a warning is issued indicating the unauthorized object. This alerts the user to this situation so that they can take corrective action. The warning can be provided directly by the cooking appliance acoustically, visually, and / or haptically. Additionally or alternatively, this can also be provided indirectly, for example, via a mobile device such as a user's smartphone.

[0030] According to a further aspect of the invention, if the captured image data cannot be assigned to a class of predetermined image data, at least one cooking process is also prevented. Thus, the potentially unsafe operation of a cooking process of the cooking appliance can be prevented even if the contents or condition of the cooking chamber cannot be reliably detected or determined from the captured image data as described above, since in this case, an unauthorized object in the cooking chamber cannot be ruled out.

[0031] According to a further aspect of the invention, if the captured image data cannot be assigned to a class of predetermined image data, a warning is output indicating unknown contents of the cooking chamber. This can also alert the user to this situation in order to remedy the situation. In this case, the warning can also be provided directly by the cooking appliance acoustically, visually, and / or haptically. Additionally or alternatively, this can also be provided indirectly, for example, via a mobile device such as a user's smartphone.

[0032] According to a further aspect of the invention, if the acquired image data cannot be assigned to a class of predetermined image data, the detection of the cooking chamber and the detection of the contents of the cooking chamber are repeated. This can reduce the probability that insufficient image data will prevent classification.

[0033] According to a further aspect of the invention, in the event that the comparison of the captured image data with the two classes of predetermined image data cannot be carried out, at least one cooking process is also prevented. In this case, technical problems are assumed that prevent the classification. This can be, for example, a missing data connection to the neural network, particularly when using a cloud. This can also be the case if the user is faster in filling the cooking chamber and operating the cooking appliance to start a cooking process than the cooking appliance executes the method according to the invention. In any case, a potentially unsafe cooking process can also be prevented in this case.

[0034] According to a further aspect of the invention, if the comparison of the captured image data with the two classes of predetermined image data cannot be performed, a notification is output indicating that the comparison of the captured image data cannot be performed. This also makes it possible to alert the user to this situation in order to remedy the situation. In this case, the notification can also be provided directly by the cooking appliance acoustically, visually, and / or haptically. Additionally or alternatively, this can also be provided indirectly, for example, via a mobile device such as a user's smartphone.

[0035] According to a further aspect of the invention, at least the further step takes place before dispensing: • Capturing a user input of the cooking process, wherein the method comprises at least the further step after outputting: • Recording a confirmation and / or change of the user input of the cooking process and • Operating the cooking chamber.

[0036] Thus, before each notification is issued, as described above, an input for the cooking process can be recorded from the user. This input allows the user to adjust or change the cooking process, in particular, to select and / or parameterize an operating mode. This can be done using buttons, a touchscreen, voice input, gestures, and the like on the cooking appliance itself, or also using a mobile device and the like.

[0037] In any case, the user can first adjust the cooking process before receiving one of the previously described instructions.

[0038] The user can now respond to the respective warning, for example, by removing an unauthorized object from the cooking chamber, checking the cooking chamber itself, and so on. During this time, the cooking process remains blocked.

[0039] However, the user can override the inhibition of a cooking process by making the final decision to have the cooking appliance carry out the cooking process settings, possibly after making changes. This will then result in the cooking chamber operating as desired.

[0040] According to a further aspect of the invention, at least the further step takes place before dispensing: • Capturing user input of the cooking process and • Operating the cooking chamber.

[0041] Similar to the procedure described above, in this case a message is first issued and then the user makes the final selection or determination as to how the cooking appliance should be operated, as described above.

[0042] According to a further aspect of the invention, if a food item is detected in the cooking chamber, at least one operating mode is offered to the user, preferably and its parameterization based on the detected food item, and confirmation or change of the offered operating mode, preferably and its parameterization, is detected by the user. Thus, the user can be offered or displayed, in particular by means of an operating and display element of the cooking appliance and / or by means of a mobile terminal of the user, an operating mode of the cooking appliance for carrying out a cooking process that is suitable for the detected food item, which the user can at least confirm in order to initiate execution. Preferably, the user can change or modify the operating mode and additionally or alternatively parameterize the operating mode, for example with regard to the cooking temperature.

[0043] This can support or relieve the user in carrying out the cooking process, as well as improve the quality of the cooking process through appropriate suggestions from the cooking appliance. In any case, the user must take action to actually carry out the cooking process, which can ensure safe operation.

[0044] According to a further aspect of the invention, the classified predetermined image data comprise at least one class of an empty cooking chamber and the method comprises the further steps: • in the event that the captured image data can be assigned to the class of an empty cooking chamber, offering at least one operating mode, preferably a self-cleaning function, preferably a pyrolysis, preferably and its parameterization, based on the detected empty cooking chamber for the user and • Recording of a confirmation or change of the offered operating mode, preferably and its parameterization, by the user.

[0045] This allows the invention to offer or suggest a suitable operating mode even when the cooking chamber is empty. This could, for example, involve performing a self-cleaning function of the cooking chamber, for example, using pyrolysis or wet-chemical cleaning, or even heating the cooking chamber to evaporate residual water or moisture.

[0046] According to a further aspect of the invention, the acquired image data is compared with at least two classes of predetermined image data based on a predetermined probability per class. This can represent a concrete implementation possibility.

[0047] According to a further aspect of the invention, the predetermined probability of the class of inadmissible objects is lower than the predetermined probability of the class of cooked items. In this way, the decision regarding an object classified as inadmissible in the cooking chamber can be made more easily, i.e., it can occur more frequently, which can promote safe operation.

[0048] Preferably, the predetermined probability of the class of prohibited objects is greater than 50% and / or the predetermined probability of the class of cooked goods is greater than 90%. This could represent a concrete implementation option.

[0049] According to a further aspect of the invention, the predetermined probability of the empty cooking chamber class is greater than the predetermined probability of the prohibited object class. Thus, a cooking chamber is only detected as empty if there is a high probability that this is actually the case. This reflects the risk that a cooking chamber incorrectly detected as empty would, for example, be released for a self-cleaning function such as pyrolysis, which operates at a comparatively high temperature and can thus lead to particularly severe damage if an prohibited object or even food were present in the cooking chamber.

[0050] Preferably, the predetermined probability of the empty cooking chamber class is greater than 95%. This could represent a concrete implementation possibility.

[0051] According to a further aspect of the invention, at least the following step is carried out before the detection: • Starting the cooking appliance, with the detection occurring in response to the start-up. Thus, the first detection can occur when the cooking appliance starts operating, with the start of operation being when the cooking appliance is switched on or when it transitions from a rest or standby mode.

[0052] According to a further aspect of the invention, at least the following step is carried out before the detection: • Closing the closure element, with detection occurring in response to the closure. In this case, closure can be detected, as described in more detail below. This closure can be interpreted as a possible change in the contents of the cooking chamber, which can trigger the corresponding steps as described above.

[0053] According to a further aspect of the invention, the cooking appliance, preferably the control unit of the cooking appliance, is designed and configured to detect the opening and / or closing of the closure element. This can be achieved, in particular, by detecting the movement of the closure element, which can enable a simple and direct conclusion of this kind. Additionally or alternatively, this detection can be achieved by detecting a change in the temperature and / or air flow of the cooking chamber, since this change can be attributed to a temporarily open closure element.

[0054] In any case, the opening and / or closing can be carried out sufficiently clearly so that a gap-like opening of the closure element by the user in order to be able to view the food in the cooking chamber can be ruled out, which would not sufficiently free the access opening of the cooking chamber for removing the food from the cooking chamber.

[0055] According to a further aspect of the invention, the cooking appliance, preferably the control unit of the cooking appliance, is designed and configured to detect the opening and / or closing of the closure element from a sensor-detected movement of the closure element. For this purpose, an angle sensor of the closure element can be used, for example. This allows the corresponding properties and advantages to be implemented in concrete terms, as already described above.

[0056] According to a further aspect of the invention, the cooking appliance, preferably the control unit of the cooking appliance, is designed and configured to detect the opening and / or closing of the closure element from a sensor-detected temperature change in the cooking chamber. This can be done, for example, using a temperature sensor in the cooking chamber. This allows the corresponding properties and advantages to be implemented in concrete terms, as already described above.

[0057] According to a further aspect of the invention, the cooking appliance, preferably the control unit of the cooking appliance, is designed and configured to detect the opening and / or closing of the closure element from a sensor-detected change in the air flow in the cooking chamber. This can be done, for example, using an air flow sensor of a fan in the cooking chamber. This allows the corresponding properties and advantages to be implemented in concrete terms, as already described above.

[0058] According to a further aspect of the invention, the optical detection by the image capture element occurs in response to a detected closure of the closure element. This can minimize the effort required for optical detection and only allow the cooking chamber to be viewed using optical data after a change in the presence or absence of the food has actually occurred due to the opening and subsequent closing.

[0059] The invention also relates to a cooking appliance with a cooking chamber designed to receive a cooking item through an access opening, preferably along the longitudinal axis, with a closure element designed to close and open the access opening of the cooking chamber, and with an image capture element designed and arranged to optically capture the cooking chamber, preferably along the vertical axis from above. The cooking appliance is characterized in that the cooking appliance, preferably a control unit of the cooking appliance, is designed and configured to carry out the corresponding steps of a method as described above. This makes it possible to provide a cooking appliance for implementing the method according to the invention and a system according to the invention, as described below.

[0060] The invention further relates to a system comprising a cooking appliance as described above and a neural network, preferably in a cloud, for carrying out a method as described above. Thus, a system for implementing the method according to the invention can be provided.

[0061] In other words, the invention is based on the idea that by detecting and classifying non-permissible objects, an effective warning can be given before commissioning and an error or destruction of the objects can be prevented.

[0062] According to the invention, in addition to image data of food and empty cooking chambers, image data of prohibited objects from the oven or kitchen context can also be used to specifically train a neural network. For this purpose, a neural network in a cloud can typically be used to classify the individual images from the cooking chamber, i.e., to assign the captured images to one of the following classes, for example: • Loading (food), • empty (tray, grill or cooking chamber floor is visible), • prohibited items, • unknown (state cannot be assigned to any of the previous classes).

[0063] Additional classes are also possible.

[0064] For the class of prohibited or impermissible objects, specific image data of possible kitchen-related cases must be trained. These include, for example, potholders, wooden boards, bags, boxes, cake tins, cookie cutters, and kitchen utensils such as rolling pins and soup ladles.

[0065] This can be used or implemented according to the invention as follows: • Device start and door opening The loading status can be determined with an image of the cooking chamber contents every time the appliance is started and the door is opened / closed. Depending on the load condition, the correct operating mode or the two most likely operating modes can then be suggested. • Start of an operating mode If an operating mode is started manually, the load status is queried or determined. Then it is checked whether the current load condition is permissible for the selected operating mode or not. If not, a message appears or the operating mode cannot be started. Examples of operating modes include: convection, top and bottom heat, steam cooking, microwave cooking, and combinations thereof. Pyrolysis, rapid cooling, keeping warm, reheating, and similar special functions are also explicitly included.

[0066] For each loading condition, i.e. for each class, a recognition probability can be defined which must be achieved in order for a cooking chamber image and thus the current load to be assigned.

[0067] If this detection probability is not achieved, another cooking chamber image can be taken in the first step and the classification repeated.

[0068] If the detection probability remains below the threshold, the class can be determined as unknown.

[0069] This has the advantage that the consequences of a misclassification can be taken into account. For example, the pyrolysis process, which requires an empty cooking chamber, is the most critical in the event of an error, but can be easily trained due to the few variations of an empty cooking chamber. Inadmissible objects, on the other hand, are difficult to train and recognize due to the large number of objects and variations possible. In contrast, an incorrect classification as a "non-admissible object" is generally uncritical, as no cooking process is initiated; rather, an error message can be displayed instead. Thus, in the event of a misclassification, the user simply unnecessarily checks the cooking chamber contents.

[0070] An embodiment of the invention is shown purely schematically in the drawings and is described in more detail below. It shows Fig. 1 a perspective view of a cooking appliance according to the invention in the form of an oven with an open access opening, viewed diagonally from the front above; Fig. 2 a longitudinal section through the cooking appliance according to the invention of Fig. 1 with access opening slightly open; Fig. 3 a flow diagram of a method according to the invention; and Fig. 4-6 Representations of exemplary loading states of the cooking chamber according to the method according to the invention.

[0071] The Fig. 1 and Fig. 2 are viewed in Cartesian coordinates. There is a longitudinal axis X, which can also be referred to as depth X or length X. Perpendicular to the longitudinal axis X extends a transverse axis Y, which can also be referred to as width Y. Perpendicular to both the longitudinal axis X and the transverse axis Y extends a vertical axis Z, which can also be referred to as height Z and corresponds to the direction of gravity. The longitudinal axis X and the transverse axis Y together form the horizontal X, Y, which can also be referred to as the horizontal plane X, Y.

[0072] Fig. 1 shows a perspective view of a cooking appliance 1 according to the invention in the form of an oven 1 with open access opening 13 from diagonally above the front. Fig. 2 shows a longitudinal section through the cooking appliance 1 according to the invention of Fig. 1 with access opening slightly open 13.

[0073] A cooking appliance 1 according to the invention is thus considered using the example of an oven 1. The oven 1 has an outer housing 10, which can also be referred to as an outer housing 10 and which essentially closes off or encloses the oven 1 to the outside. The outer housing 10 closes off along the longitudinal axis X towards the front, i.e. towards the user, in the upper area with a panel 10a containing operating and display elements (not shown). An inner housing 11, which can also be referred to as an inner housing 11, is arranged within the outer housing 10. The inner housing 11 closes off with the outer housing 10 along the longitudinal axis X to the rear, along the vertical axis Z downwards and upwards, and along the transverse axis Y on both sides, so that an intermediate space 12a is formed between the outer housing 10 and the inner housing 11, which intermediate space 12a can also be referred to as the housing space 12a. This intermediate space 12a represents the interior of the oven 1.

[0074] The inner housing 11 also essentially encloses an interior space 12 in which a cooking process can be carried out. For this purpose, corresponding elements for heating the interior space 12 (not shown) are arranged in the intermediate space 12a or in the cooking chamber 12. Accordingly, the interior space 12 can also be referred to as the cooking chamber 12.

[0075] The cooking chamber 12 is accessible to the user from the front along the longitudinal axis X through an access opening 13, which is located below the panel 10a. The access opening 13 can be closed and opened by means of a closure element 14 in the form of a flap 13. For this purpose, the flap 14 is connected to the outer casing 10 of the oven 1 by means of at least one hinge (not designated) so as to be pivotable upwards about the transverse axis Y, so that the flap 14 can be pivoted upwards away from the user to close the access opening 13, and can be pivoted downwards towards the user to release the access opening 13 and to access the cooking chamber 12, see Fig. 1.

[0076] The cooking chamber 12 has a side wall 12a on each side along the transverse axis Y, on each of which a removable support grid 15 is arranged. A telescopic rail 16 is removably attached to each support grid 15 at the same height along the vertical axis Z, which together form a pair of telescopic rails 16. The telescopic rails 16 are each designed in three parts along the longitudinal axis X (not shown), so that a front rail element is connected to a rear rail element of the telescopic rail 16 by means of a middle rail element, wherein the rear rail element is held fixedly and removably on the respective support grid 15.

[0077] A cooking utensil containing food, such as a baking tray (not shown), can be placed by the user onto the pair of telescopic rails 16 or their front rail elements in order to be pushed into and pulled out of the cooking chamber 12 along the longitudinal axis X. This can make it easier for the user to insert the baking tray containing food into the cooking chamber 12 and to remove it from the cooking chamber 12.

[0078] The closure element 14 or the flap 14 has a closure element sensor 14a in the form of an angle sensor 14a to detect the position of the flap 14. An image capture element 17 in the form of a camera 17 is arranged on the ceiling (not labeled) of the inner housing 11 and is directed downwards into the cooking chamber 12 essentially along the vertical direction Z in order to optically capture the cooking chamber 12 as completely as possible from above. Furthermore, a fan 18 in the form of a blower 18 is provided to convey a cold air flow through a fan duct 18a in the form of a blower duct 18a into the cooking chamber 12.

[0079] The cooking appliance 1 or the oven 1 further comprises a control unit 19 for operating the cooking appliance 1 or the oven 1 or its functions. This includes the function of rapid cooling with subsequent keeping warm. The cooking appliance 1 or the oven 1 is connected via a preferably wireless data connection to a neural network in a cloud (not shown). The cooking appliance 1 or the oven 1 according to the invention and the neural network form a system according to the invention, which can execute a method according to the invention as follows, see flowchart of the Fig. 3: The cooking appliance 1 is started 000. In response to the start 000, the cooking chamber 12 is detected 200 by means of the image detection element 17. Alternatively, during the course of operation of the cooking appliance 1, the closure element 14 can also be closed 100 and the detection 200 can occur in response to the closing 100. In any case, the detection 200 occurs when the contents of the cooking chamber 12 are unknown or may have changed.

[0080] For this purpose, the cooking appliance 1 or the oven 1 or its control unit 19 can detect the opening and closing of the flap 14. The opening and closing of the flap 14 can be detected from a sensor-detected movement of the flap 14 by means of the closure element sensor 14a, from a sensor-detected temperature change of the cooking chamber 12 by means of a corresponding temperature sensor (not shown), and / or from a sensor-detected air flow change of the air flow of the fan 18 by means of a corresponding air flow sensor (not shown).

[0081] The contents of the cooking chamber 12 are detected 300 based on a comparison 350 of the captured image data with at least two classes of predetermined image data, wherein the classified predetermined image data comprises at least one class of prohibited objects and one class of cooked items. If a prohibited object is detected in the cooking chamber 12, at least one inhibition 950 of a cooking process occurs, so that the detected prohibited object, such as a potholder or the like, cannot be damaged or destroyed by the intended cooking process, which also applies to the cooking appliance 1 and its surroundings.

[0082] The comparison of the captured image data with the predetermined image data is carried out using a neural network, e.g., in the cloud. The predetermined image data was previously trained, or the neural network was trained accordingly, so that even unknown, inadmissible objects cannot be identified but can be classified, which also enables their classification.

[0083] In the event of a detected impermissible object in the cooking chamber 12, an indication 800a, 800b of the detected impermissible object in the cooking chamber 12 is also output to the user, which can be done by means of the display and control element of the panel 10a.

[0084] In the event that the captured image data cannot be assigned to a class of predetermined image data, at least a cooking process is prevented 950, since the cooking chamber cannot be classified as "safe" and, in case of doubt, an unsafe state of the cooking chamber must be assumed. In this case, too, an indication 850a, 800b of an unknown content of the cooking chamber 12 is output. Furthermore, the detection 200 of the cooking chamber 12 and the recognition 300 of the content of the cooking chamber 12 are repeated 400 in order to still be able to perform a successful classification if possible.

[0085] In the event that the comparison 350 of the captured image data with the two classes of predetermined image data cannot be performed, for example, due to a missing data connection from the cooking appliance 1 to the neural network in the cloud, at least one cooking process is also stopped 950. In this case, too, an indication 900a, 900b indicating that the comparison 350 of the captured image data cannot be performed is output to the user.

[0086] It may be the case that the user is already making settings for the cooking process on the cooking appliance 1 while the contents of the cooking chamber 12 are being detected 300 or wants to do so immediately afterwards.

[0087] Thus, on the one hand, the case may arise that a user input of the cooking process is detected 700a and subsequently output 800a, 850a, 900a of a notification, as previously described. Meanwhile, the cooking process is stopped 950. However, since it is up to the user to start the cooking process despite a notification, for example, after checking the cooking chamber 12 themselves, a confirmation and / or change of the user input of the cooking process by the user can be detected 720, which then leads to the operation 750 of the cooking chamber 12.

[0088] On the other hand, the output 800a, 850a, 900a of an indication can also take place first, so that the user can then achieve operation 750 of the cooking chamber 12 by detecting 700b a user input of the cooking process.

[0089] In any case, if a food item is detected in the cooking chamber 12, at least one operating mode, preferably and its parameterization, is offered 500 to the user based on the detected food item. Furthermore, a confirmation or change of the offered operating mode, preferably and its parameterization, by the user is detected 550, which also leads to operation 750 of the cooking chamber 12.

[0090] In addition, the classified predetermined image data can comprise at least one class of an empty cooking chamber 12, so that if the captured image data can be assigned to the class of an empty cooking chamber 12, an offering 600 of at least one operating mode, preferably a self-cleaning function such as preferably pyrolysis, and its parameterization can be performed for the user based on the detected empty cooking chamber 12. In this case, too, a detection 650 of a confirmation or change of the offered operating mode, preferably and its parameterization, by the user can occur, which likewise leads to an operation 750 of the cooking chamber 12. In this way, operating modes can also be offered which are to be carried out when the cooking chamber is empty.

[0091] In any case, the comparison 350 of the acquired image data with at least two classes of predetermined image data is performed based on a predetermined probability per class. The predetermined probability of the class of impermissible objects is lower than the predetermined probability of the class of cooked goods, wherein preferably the predetermined probability of the class of impermissible objects is greater than 50% and / or the predetermined probability of the class of cooked goods is greater than 90%.

[0092] Furthermore, if a classification of the empty cooking chamber 12 is performed as described above, the predetermined probability of the class of the empty cooking chamber 12 is greater than the predetermined probability of the class of prohibited objects, wherein the predetermined probability of the class of the empty cooking chamber 12 is preferably greater than 95%. This allows the respective risk of the classification to be taken into account.

[0093] The Fig. 4-6 show representations of exemplary loading states of the cooking chamber 1 according to the method according to the invention. List of reference symbols (part of the description) X Longitudinal axis; Depth; Length Y transverse axis; width Z vertical axis; height X, Y Horizontal; horizontal plane 1 cooking appliance; oven 10 outer casing; outer casing 10a aperture 11 inner casing; inner casing 12 Interior; cooking chamber 12a Side walls of the cooking chamber 12 12b Intermediate space; housing space 13 Access opening 14 Closure element; flap 14a Closure element sensor; angle sensor 15 support grids 16 telescopic rails 17 Image capture element; camera 18 fans; blowers 18a Fan duct; blower duct 19 Control unit 000 Starting the cooking appliance 1 100 Closing the locking element 14 200 Detecting the cooking chamber 12 300 Detecting the contents of the cooking chamber 12 350 Comparing the acquired image data with classes of predetermined image data 400 Repeating the capture 200 and the recognition 300 500 Offering an operating mode 550 Recording of confirmation or change of the offered operating mode by user 600 Offering an operating mode, preferably a self-cleaning function, preferably pyrolysis 650 Recording of confirmation or change of the offered operating mode by user 700a, 700b Capturing a user input of the cooking process 720 Recording a confirmation and / or change of the user input of the cooking process 750 Operating the cooking chamber 12 800a, 800b Output of a notification about a detected illegal object in the cooking chamber 12 850a, 850b Output of an indication of an unknown content of the cooking chamber 12 900a, 900b Output of an indication that comparison is not possible 350 950 Stopping a cooking process QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] WO 2022 / 010748 A1

[0013] WO 2022 / 184483 A1

[0014]

Claims

[1] Method for operating a cooking appliance (1), wherein the cooking appliance (1) comprises: • a cooking chamber (12) which is designed to receive a cooking product through an access opening (13), preferably along the longitudinal axis (X), • a closure element (14) which is designed to close and open the access opening (13) of the cooking chamber (12), and • an image capture element (17) which is designed and arranged to optically capture the cooking chamber (12), preferably along the vertical axis (Z) from above, the method comprising at least the following steps: • Detecting (200) the cooking chamber (12) by means of the image capture element (17), • Detecting (300) the contents of the cooking chamber (12) based on comparing (350) the captured image data with at least two classes of predetermined image data, wherein the classified predetermined image data comprises at least one class of inadmissible objects and one class of cooking goods, and, • in case of a detected inadmissible object in the cooking chamber (12), at least preventing (950) a cooking process, wherein the comparison (350) of the acquired image data with the predetermined image data is carried out by means of a neural network, preferably in a cloud. [2] Method according to claim 1, comprising at least the further step: • in the event of a detected impermissible object in the cooking chamber (12), outputting (800a, 800b) an indication of the detected impermissible object in the cooking chamber (12). [3] Method according to claim 1 or 2, comprising at least the further step: • in the event that the captured image data cannot be assigned to a class of predetermined image data, also at least preventing (950) a cooking process. [4] Method according to claim 3, comprising at least the further step: • in the event that the captured image data cannot be assigned to a class of predetermined image data, outputting (850a, 850b) an indication of an unknown content of the cooking chamber (12). [5] Method according to claim 3 or 4, comprising at least the further step: • in the event that the captured image data cannot be assigned to a class of predetermined image data, repeating (400) the capture (200) of the cooking chamber (12) and the recognition (300) of the contents of the cooking chamber (12). [6] Method according to one of the preceding claims with at least the further step: • in the event that the comparison (350) of the captured image data with the two classes of predetermined image data cannot be carried out, also at least preventing (950) a cooking process. [7] Method according to claim 6, comprising at least the further step: • in the event that the comparison (350) of the acquired image data with the two classes of predetermined image data cannot be carried out, outputting (900a, 900b) an indication of the impossibility of comparing (350) the acquired image data. [8] Method according to one of claims 2 to 7 with at least the further step before outputting (800a, 850a, 900a): • detecting (700a) a user input of the cooking process, wherein the method comprises at least the further step after the output (800a, 850a, 900a): • Detecting (720) a confirmation and / or change of the user input of the cooking process and • Operating (750) the cooking chamber. [9] Method according to one of claims 2 to 7 with at least the further step after outputting (800b, 850b, 900b): • Detecting (700b) a user input of the cooking process and • Operating (750) the cooking chamber. [10] Method according to one of the preceding claims with at least the further steps: • in the case of a detected cooking product in the cooking chamber (12), offering (500) at least one operating mode, preferably and its parameterization, based on the detected cooking product for the user and • Recording (550) a confirmation or change of the offered operating mode, preferably and its parameterization, by the user. [11] Method according to one of the preceding claims, wherein the classified predetermined image data comprises at least one class of an empty cooking chamber (12), with at least the following steps: • in the event that the captured image data can be assigned to the class of an empty cooking chamber (12), offering (600) at least one operating mode, preferably a self-cleaning function, preferably a pyrolysis, preferably and its parameterization, based on the detected empty cooking chamber (12) for the user and • Recording (650) a confirmation or change of the offered operating mode, preferably and its parameterization, by the user. [12] Method according to one of the preceding claims, wherein the comparing (350) of the acquired image data with at least two classes of predetermined image data is carried out based on a predetermined probability per class. [13] Method according to claim 12, where the predetermined probability of the class of inadmissible objects is lower than the predetermined probability of the class of cooked goods, wherein preferably the predetermined probability of the class of inadmissible objects is greater than 50% and / or the predetermined probability of the class of cooked goods is greater than 90%. [14] Method according to one of claims 9 to 12 and according to claim 13, wherein the predetermined probability of the class of the empty cooking chamber (12) is greater than the predetermined probability of the class of inadmissible objects, wherein preferably the predetermined probability of the class of the empty cooking chamber (12) is greater than 95%. [15] Method according to one of the preceding claims with at least the preceding step before the detection (200): • Starting (000) of the cooking appliance (1), wherein the detection (200) occurs in response to the starting (000). [16] Method according to one of the preceding claims with at least the preceding step before the detection (200): • Closing (100) the closure element (14), wherein the detection (200) occurs in response to the closing (100). [17] Cooking appliance (1) with a cooking chamber (12) which is designed to receive a cooking product through an access opening (13), preferably along the longitudinal axis (X), with a closure element (14) which is designed to close and open the access opening (13) of the cooking chamber (12), and with an image capture element (17) which is designed and arranged to optically capture the cooking chamber (12), preferably along the vertical axis (Z) from above, characterized by , that the cooking appliance (1), preferably a control unit (19) of the cooking appliance (1), is designed and configured to carry out the corresponding steps of a method according to one of the preceding claims. [18] System with a cooking appliance (1) according to claim 17 and a neural network, preferably in a cloud, for carrying out a method according to one of claims 1 to 16.

Images