Food ingredient cooking control method, device, equipment and storage medium

Abstract

The present invention discloses a method, device, equipment and storage medium for cooking ingredient control. The method includes: acquiring a cooking ingredient image; performing image detection on the cooking ingredient image by using a pre-trained lightweight ingredient detection model to obtain the ingredient types and ingredient positions of at least one target ingredient output by the model; determining cooking parameters according to the ingredient types and ingredient positions of each target ingredient; and performing corresponding device control actions based on the cooking parameters to obtain a cooking result. The technical solution of the embodiment of the present invention improves the flexibility of the ingredient cooking process, realizes effective and accurate identification of ingredients, improves the accurate determination of cooking parameters, and thus improves the cooking quality.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular, to a method, device, equipment and storage medium for controlling food cooking. Background Art

[0002] With the progress of technology, people's living standards are constantly improving, and their requirements for diet are also getting higher and higher. However, for most non-professional chefs, they do not have rich cooking experience. Therefore, how to use modern technology means to simplify and standardize the cooking control process has become an urgent problem to be solved.

[0003] Existing cooking automation means usually require users to prepare ingredients and set cooking parameters according to preset recipes. However, this cooking method mostly relies on the cooking parameters pre-configured in the preset recipes, lacking certain flexibility. And it is unable to effectively and accurately identify ingredients, resulting in the inability to determine cooking parameters based on the actual situation of ingredients, and further leading to a decline in cooking quality. Summary of the Invention

[0004] The present invention provides a method, device, equipment and storage medium for controlling food cooking, so as to improve the flexibility of the food cooking process, realize effective and accurate identification of ingredients, improve the accurate determination of cooking parameters, and thus improve cooking quality.

[0005] According to one aspect of the present invention, there is provided a method for controlling food cooking, the method comprising:

[0006] Obtain a cooking food image;

[0007] Perform image detection on the cooking food image by using a pre-trained lightweight food detection model to obtain the food types and food positions of at least one target food output by the model;

[0008] Determine cooking parameters according to the food types and food positions of each target food;

[0009] Execute corresponding device control actions based on the cooking parameters to obtain a cooking result.

[0010] According to another aspect of the present invention, there is provided a device for controlling food cooking, the device comprising:

[0011] A food image acquisition module, configured to obtain a cooking food image;

[0012] An image detection module, configured to perform image detection on the cooking food image by using a pre-trained lightweight food detection model to obtain the food types and food positions of at least one target food output by the model;

[0013] A cooking parameter determination module, configured to determine cooking parameters according to the types and positions of the target ingredients;

[0014] A control execution module, configured to perform corresponding device control actions based on the cooking parameters to obtain a cooking result.

[0015] According to another aspect of the present invention, there is provided a cooking device, including:

[0016] At least one processor; and

[0017] A memory communicatively connected to the at least one processor; wherein,

[0018] The memory stores a computer program executable by the at least one processor, and when the computer program is executed by the at least one processor, the at least one processor is enabled to execute the ingredient cooking control method according to any embodiment of the present invention.

[0019] According to another aspect of the present invention, there is provided a computer-readable storage medium storing computer instructions for causing a processor to implement the ingredient cooking control method according to any embodiment of the present invention when executed.

[0020] In the technical solution of the present invention, by comprehensively considering the types and positions of ingredients during the determination of cooking parameters, and at the same time, performing image detection on the cooking ingredient images based on a pre-trained lightweight ingredient detection model, effective and accurate identification of ingredients is achieved, rather than completely controlling cooking based on the cooking parameters corresponding to pre-selected pre-made recipes, which improves the flexibility of the cooking process and the accurate determination of cooking parameters, thereby improving the cooking quality.

[0021] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. Description of the Drawings

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings without creative efforts based on these drawings.

[0023] Figure 1A It is a flowchart of an ingredient cooking control method provided in Embodiment 1 of the present invention;

[0024] Figure 1BIt is a schematic structural diagram of a detection network model based on the fusion of YOLO-V8 model and CBAM provided by Embodiment 1 of the present invention;

[0025] Figure 2 It is a flowchart of a method for controlling food cooking provided by Embodiment 2 of the present invention;

[0026] Figure 3 It is a flowchart of a method for controlling food cooking provided by Embodiment 3 of the present invention;

[0027] Figure 4 It is a flowchart of a method for controlling food cooking provided by Embodiment 4 of the present invention;

[0028] Figure 5 It is a schematic structural diagram of a device for controlling food cooking provided by Embodiment 5 of the present invention;

[0029] Figure 6 It is a schematic structural diagram of a cooking device for implementing the method for controlling food cooking according to the embodiments of the present invention. Detailed implementation manners

[0030] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present invention described herein can be implemented in an order different from those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0032] Embodiment 1

[0033] Figure 1AThe flowchart of a food cooking control method provided in Embodiment 1 of the present invention. This embodiment is applicable to accurately determining the cooking parameters of a cooking device to control food cooking based on the cooking parameters. This method can be executed by a food cooking control device, which can be implemented in the form of hardware and / or software, and the food cooking control device can be configured in the cooking device. As Figure 1A shown, this method can be executed by the cooking device, specifically including:

[0034] S110. Obtain a cooking food image.

[0035] S120. Use a pre-trained lightweight food detection model to perform image detection on the cooking food image to obtain the food types and food positions of at least one target food output by the model.

[0036] S130. Determine the cooking parameters according to the food types and food positions of each target food.

[0037] S140. Perform corresponding device control actions based on the cooking parameters to obtain a cooking result.

[0038] Among them, the cooking food image can be obtained by the image acquisition device deployed in the cooking device. The image acquisition device is located directly above the cavity of the cooking device and is used to perform image acquisition on the food placed directly below its own position when detecting the placement action of the food plate. For example, the cooking device can be a steam oven, and the image acquisition device can be a camera; the image acquisition device can remain in the standby mode under normal conditions, and when detecting the food placement action, the image acquisition device performs image acquisition on the food placed in the cooking device to obtain the cooking food image.

[0039] Among them, the lightweight food detection model can be pre-trained by relevant technical personnel and pre-deployed in the cooking device; the model of the lightweight food cooking model occupies less memory and has relatively low detection accuracy. The lightweight food detection model is used to predict the food types and food positions in the cooking food image.

[0040] Specifically, input the cooking food image collected by the image acquisition device into the pre-trained lightweight food detection model, and the lightweight food detection model predicts the food types and food positions in the cooking food image to obtain the food types and food positions of at least one target food output by the model.

[0041] Determine cooking parameters based on the type and location of the target food ingredient. Specifically, the cooking parameters may include cooking temperature, cooking humidity, cooking time, and the heating power of the heating tube, etc. Different food ingredient types may correspond to different cooking temperatures, cooking humidities, and cooking times, and their mapping relationships can be preset by relevant technicians according to actual needs. Different food ingredient locations correspond to different heating powers of the heating tubes, and their mapping relationships can be preset by relevant technicians according to actual needs. It should be noted that the food ingredient location is the relative position of the food ingredient with respect to the internal cavity of the cooking device. A plurality of heating tubes at different positions are deployed around the cavity of the cooking device. According to the food ingredient location, the heating powers of different heating tubes can be adaptively adjusted, and the mapping relationship between the specific food ingredient location and the heating power of the specific heating tube can be preset according to actual experience values or experimental values.

[0042] The cooking device performs corresponding device control actions based on the determined cooking parameters, obtains a cooking result, and presents the cooking result to the initiator of the food ingredient cooking.

[0043] The technical solution of the present invention realizes effective and accurate identification of food ingredients by comprehensively considering the food ingredient type and food ingredient location during the process of determining cooking parameters, and at the same time, performing image detection on the cooking food ingredient image based on a pre-trained lightweight food ingredient detection model, rather than completely controlling cooking based on the cooking parameters corresponding to pre-selected pre-made recipes, improving the flexibility of the cooking process and the accurate determination of cooking parameters, thereby improving the cooking quality.

[0044] This embodiment also provides a model training method for the lightweight food ingredient detection model, including: obtaining food ingredient sample images of different food ingredients in different cooking devices, and annotating the food ingredient types and food ingredient locations in the food ingredient sample images to obtain food ingredient sample images with labels of standard food ingredient types and standard food ingredient locations; using the food ingredient sample images to perform model training on a pre-constructed detection network model to obtain a trained target food ingredient detection model; performing model quantization processing on the target food ingredient detection model to obtain a lightweight food ingredient detection model.

[0045] To ensure the richness and diversity of the food ingredient sample images, and thus ensure that the trained lightweight food ingredient detection model can take into account the image detection of different food ingredients under different types of cooking devices, during the construction process of the food ingredient sample images, the same and different food ingredient images under different cooking devices are selected as the food ingredient sample images.

[0046] Label the ingredient positions and ingredient types in the ingredient sample image to generate sample labels. Among them, the ingredient positions are reflected in the form of detection frames in the ingredient sample image, including the relative position coordinates of the detection frames. Manual, automated, or semi-automated labeling means can be used, and the label value of the obtained ingredient sample image is the sample ground truth.

[0047] Exemplarily, input the ingredient sample image labeled with the standard ingredient type and the standard ingredient position into the pre-constructed detection network model to obtain the predicted ingredient type and the predicted ingredient position output by the model. Among them, the detection network model can be a YOLO series model (a type of deep learning model for real-time object detection) and SSD (Single Shot MultiBox Detector), etc.

[0048] Determine the loss value based on the standard ingredient type, the standard ingredient position, the predicted ingredient type, and the predicted ingredient position, based on a preset loss function; perform model training on the detection network model according to the loss value until the model training end condition is met to obtain the trained target ingredient detection model. Among them, the model training end condition can be that the loss value tends to be stable, or the loss value reaches the set loss threshold, or the number of iterations reaches the set iteration threshold, etc., and this embodiment does not limit this.

[0049] Perform quantization processing on the target ingredient detection model, specifically, it can be model pruning, post-training quantization, or low-rank decomposition, etc., so as to reduce the storage size, computational amount, and memory usage of the model, while trying to maintain the performance of the model to obtain a lightweight ingredient detection model.

[0050] To further improve the recognition accuracy of cooking ingredient images, this embodiment uses a combination of the YOLO-V8 model and CBAM (Convolutional Block Attention Module) to obtain the detection network model.

[0051] In an alternative embodiment, the model structure of the detection network model includes a YOLO-V8 network module and a convolutional block attention module (CBAM). The YOLO-V8 network module includes a backbone network, a neck network, and a detection head. As Figure 1B shown in the schematic diagram of the structure of a detection network model based on the fusion of the YOLO-V8 model and CBAM. The input end of the convolutional block attention module (CBAM) is connected to the output end of the backbone network; the output end of the convolutional block attention module (CBAM) is connected to the input end of the neck network.

[0052] Correspondingly, a detection network model pre-constructed with food ingredient sample images is used for model training to obtain a trained target food ingredient detection model, including: inputting the food ingredient sample images into the backbone network of the YOLO-V8 network module for feature extraction to obtain multi-scale feature maps; inputting the multi-scale feature maps into the convolutional block attention module for feature map weighting processing to obtain weighted feature maps; inputting the weighted feature maps into the neck network of the YOLO-V8 network module for feature map fusion processing to obtain fused feature maps; inputting the fused feature maps into the detection head of the YOLO-V8 network module for feature processing to obtain predicted food ingredient types and predicted food ingredient positions; training the detection network model according to the standard food ingredient types, standard food ingredient positions, predicted food ingredient types and predicted food ingredient positions of the food ingredient sample images until the model training end condition is met to obtain the target food ingredient detection model.

[0053] The model parameters required for the model training process can be preset according to actual needs. For example, batch is set to 16, workers (the number of parallel processes for data loading) is set to 8, optimizer (stochastic gradient descent optimizer) selects SGD, the picture size of the food ingredient sample images is 640 pixels × 640 pixels, the learning rate is set to 0.01, and the maximum number of iterations is set to 1000 steps.

[0054] Input the food ingredient sample images into the backbone network of the YOLO-V8 network module, and the backbone network extracts features from low-level (edges / textures) to high-level features (semantics) to obtain multi-scale features. Input the multi-scale feature maps into the convolutional block attention module (CBAM) for feature map weighting processing to obtain weighted feature maps. Specifically, the weighted feature maps are processed through the channel attention module (CAM) in CBAM, and the feature maps of each channel are weighted to obtain channel attention weights, and the channel attention weights are multiplied by the weighted feature maps channel by channel to obtain the feature maps adjusted by channel attention. Input the feature maps adjusted by channel attention into the spatial attention module (SAM), and the adjusted feature maps are processed through the spatial attention module (SAM) in CBAM, and the feature maps of each spatial position are weighted to obtain spatial attention weights, and the spatial attention weights are multiplied by the adjusted feature maps channel by channel to obtain weighted feature maps. Input the weighted feature maps into the neck network of the YOLO-V8 network module for feature map fusion processing, fuse high-level semantic features and transmit low-level positioning information to obtain the fused feature maps after fusion. Input the fused feature maps into the detection head of the YOLO-V8 network module for food ingredient type and position prediction to obtain predicted food ingredient types and predicted food ingredient positions.

[0055] Based on the standard ingredient types, standard ingredient positions, predicted ingredient types, and predicted ingredient positions in the ingredient sample image, determine the loss value based on a preset loss function; perform model training on the detection network model according to the loss value until the model training end condition is satisfied, and obtain the trained target ingredient detection model. Among them, the model training end condition can be that the loss value tends to be stable, or the loss value reaches a set loss threshold, or the number of iterations reaches a set iteration threshold, etc., and this embodiment does not limit this.

[0056] In the above technical solution, the model integrating the YOLO network model and the CBAM mechanism is used as the detection network model, and the YOLO-V8 network model can achieve ideal results in the target detection task. However, in the process of identifying and detecting food sample images, due to the influence of light in the cooking equipment and the mutual superposition of some foods, the recognition results have errors, and the recognition results will also seriously affect the final cooking parameters, resulting in some foods not achieving the best steaming and roasting results. Therefore, the CBAM attention mechanism is added on the basis of the original YOLO-V8 network model, so that the model pays more attention to the foreground and target areas of the image during the training process, while SAM pays attention to the positions rich in context information during the network training process. Adding the CBAM attention mechanism can increase the weight of the food feature area and reduce the weight of the background feature when the network is learning features, making the model pay more attention to the target to be detected, improving the situation of missed detection and false detection in the image, improving the accuracy of food detection, further realizing the effective and accurate recognition of ingredients, improving the accurate determination of cooking parameters, and thus improving the cooking quality.

[0057] Embodiment 2

[0058] Figure 2 It is a flowchart of a method for controlling ingredient cooking provided by Embodiment 2 of the present invention. On the basis of the above technical solutions, this embodiment is optimized and improved.

[0059] Further, before the step of "performing image detection on the cooking ingredient image using the pre-trained lightweight ingredient detection model to obtain the ingredient types and ingredient positions of at least one target ingredient output by the model", after the step of "acquiring the cooking ingredient image", add the step of "monitoring the network status of the server to obtain the current network status information; according to the current network status information, determining whether to perform image detection on the cooking ingredient image using the pre-trained lightweight ingredient detection model; if not, sending the cooking ingredient image to the server for the server to perform image detection on the cooking ingredient image using the pre-trained target ingredient detection model locally deployed on itself to obtain the ingredient types and ingredient positions of at least one target ingredient output by the model, and feeding back the ingredient types and ingredient positions of the target ingredient to the cooking device; if so, performing the step of performing image detection on the cooking ingredient image using the pre-trained lightweight ingredient detection model to obtain the ingredient types and ingredient positions of at least one target ingredient output by the model." to improve the execution method of the ingredient image recognition task.

[0060] It should be noted that for the parts not detailed in the embodiments of the present invention, reference can be made to the descriptions of other embodiments. As Figure 2 shown, the method includes the following specific steps:

[0061] S210. Acquire a cooking ingredient image.

[0062] S220. Monitor the network status of the server to obtain the current network status information.

[0063] S230. According to the current network status information, determine whether to perform image detection on the cooking ingredient image using the pre-trained lightweight ingredient detection model; if so, execute S240A; if not, execute S240B.

[0064] S240A. Perform image detection on the cooking ingredient image using the pre-trained lightweight ingredient detection model to obtain the ingredient types and ingredient positions of at least one target ingredient output by the model.

[0065] S240B. Send the cooking ingredient image to the server for the server to perform image detection on the cooking ingredient image using the pre-trained target ingredient detection model locally deployed on itself to obtain the ingredient types and ingredient positions of at least one target ingredient output by the model, and feed back the ingredient types and ingredient positions of the target ingredient to the cooking device.

[0066] S250. Determine cooking parameters according to the ingredient types and ingredient positions of each target ingredient.

[0067] S260. Perform corresponding device control actions based on the cooking parameters to obtain a cooking result.

[0068] Among them, the current network status information may include parameter information such as network speed, network latency, network jitter, and packet loss rate. The network status information can comprehensively evaluate the network condition or network quality of a device or server. Among them, the server may be a cloud server.

[0069] Based on the current network status information of the server, it can be determined whether the subject for image detection is the server or the cooking device (mobile device). Specifically, a pre-trained network quality detection model can be used to detect the current network status information, so as to obtain the network quality output by the model, which can specifically be a quality category, including categories such as high network quality, relatively high quality, medium quality, relatively low quality, and low quality; or, it can also be a specific quality score, and the range can be set from 0 to 100. The higher the score, the higher the network quality; the lower the score, the lower the network quality.

[0070] The model training method of the network quality detection model with the quality category as the output result can be: obtaining the historical network status in the historical time period as sample training data, and determining the corresponding standard classification results of each historical network status as the sample labels of the corresponding sample training data. Using the sample training data and its corresponding standard classification results to train the pre-constructed classification network model to obtain the trained network quality detection model. Among them, the classification network model can be a machine learning model or a neural network model. For example, the classification network model can be a decision tree, a random forest, or a convolutional neural network, etc.

[0071] It should be noted that a target ingredient detection model is deployed in the server. Compared with the lightweight ingredient detection model deployed on the cooking device (mobile device), its performance is higher and the accuracy of ingredient recognition is higher. However, compared with the mobile device, the model in the server is more affected by network latency. Therefore, when it is detected that the network quality of the server is high, the server performs image detection on the cooking ingredient image based on the target ingredient detection model with higher performance deployed by itself, and feeds back the ingredient type and ingredient position of the detected target ingredient to the cooking device. When it is detected that the network quality of the server is low, the cooking device (mobile device) performs image detection on the cooking ingredient image based on the lightweight ingredient detection model deployed by itself, and obtains the ingredient type and ingredient position of at least one target ingredient output by the model.

[0072] Furthermore, when the network status quality is poor, clearly prompt the cooking user to use the lightweight ingredient detection model deployed on the mobile device for recognition, and prompt the cooking user that the model recognition accuracy is relatively low. At the same time, provide the cooking user with a choice, allowing the cooking user to manually switch the recognition execution subject according to actual needs and network status.

[0073] Further, if the recognition of cooking ingredient images is based on a mobile device, it can be prompted to check after the network is restored and selectively re-upload the cooking ingredient images to the server for verification to provide a more accurate recognition result.

[0074] Further, this embodiment also provides a method for selecting an image recognition subject. Obtain the historical network status information and its corresponding user behavior pattern in the historical period. Among them, the user behavior pattern is the recognition subject (server or mobile device) selected under the historical network status information. Input the historical network status information and the user behavior pattern into a preset classification model to obtain the predicted behavior pattern output by the model. Train the classification model according to the user behavior pattern and the predicted behavior model to obtain a trained behavior pattern prediction model. Input the current network status information of the mobile device and / or the current network status information of the server in the current time period obtained into the behavior pattern prediction model to obtain the target behavior pattern output by the model, including the server or the mobile device.

[0075] The technical solution of this embodiment monitors the network status of the server to obtain the current network status information; according to the current network status information, determines the execution subject for performing image detection on the cooking ingredient images, realizes the dynamic determination of the execution subject, takes into account multiple factors such as network status and model performance, improves the flexibility of image detection, and better meets the requirements of actual image detection scenarios.

[0076] Embodiment III

[0077] Figure 3 It is a flowchart of a method for controlling ingredient cooking provided by Embodiment III of the present invention. This embodiment is optimized and improved on the basis of the above technical solutions.

[0078] Further, the step of "determining cooking parameters according to the ingredient types and ingredient positions of each target ingredient" is refined into "respectively determining the ingredient attributes of each target ingredient according to the ingredient types of each target ingredient; determining cooking parameters according to the ingredient attributes and ingredient positions of each target ingredient." to improve the method for determining cooking parameters.

[0079] It should be noted that for the parts not detailed in the embodiments of the present invention, reference can be made to the descriptions of other embodiments. As Figure 3 shown, the method includes the following specific steps:

[0080] S310. Obtain a cooking ingredient image.

[0081] S320. Use a pre-trained lightweight ingredient detection model to perform image detection on the cooking ingredient image to obtain the ingredient types and ingredient positions of at least one target ingredient output by the model.

[0082] S330. Determine the ingredient attributes of each target ingredient according to the ingredient types of each target ingredient.

[0083] S340. Determine the cooking parameters according to the ingredient attributes and ingredient positions of each target ingredient.

[0084] S350. Perform corresponding equipment control actions based on the cooking parameters to obtain a cooking result.

[0085] Among them, the ingredient attributes may include ingredient texture, nutritional value, cooking difficulty, etc. The ingredient attributes corresponding to different types of ingredients are different. The mapping relationship between the ingredient attributes and the ingredient types can be established and stored in advance by relevant technical personnel.

[0086] Among them, the association relationship between the ingredient attributes and ingredient positions and the cooking parameters can also be established and stored in advance by relevant technical personnel.

[0087] To further improve the accuracy of determining the cooking parameters, in an optional embodiment, determining the cooking parameters according to the ingredient attributes and ingredient positions of each target ingredient includes: determining a target cooking temperature and a target cooking time according to the ingredient attributes of each target ingredient; determining the heating weight parameters of each heating tube deployed in the cooking equipment according to the ingredient positions of each target ingredient; generating cooking parameters including the target cooking temperature, the target cooking time, and the heating weight parameters of the heating tubes.

[0088] It should be noted that different ingredient attributes can correspond to different cooking temperatures and cooking times. Specifically, a first mapping relationship table between the ingredient attributes and the cooking temperature is established and stored in advance, and a second mapping relationship table between the ingredient attributes and the cooking time is established and stored in advance. After determining the ingredient attributes of the target ingredient, the target cooking temperature corresponding to the target ingredient can be determined based on the first mapping relationship table; the target cooking time corresponding to the target ingredient can be determined based on the second mapping relationship table.

[0089] Since there are multiple heating tubes in the cavity of the cooking equipment, for any heating tube, when the ingredient is close to the position of the heating tube, the heating power of the heating tube should be set lower; when the ingredient is far from the position of the heating tube, the heating power of the heating tube should be set higher. Therefore, the setting of the heating power of different heating tubes is related to the ingredient position in the cavity.

[0090] Among them, the heating weight parameter of the heating tube can be the power weight corresponding to the heating power of the heating tube. When the position distance between the ingredient and the heating tube is close, the corresponding power weight is lower; when the position distance between the ingredient and the heating tube is far, the corresponding power weight is higher.

[0091] Exemplarily, if there are heating tubes A, B, and C, according to the position of the food ingredients, it is known that the distance between the target food ingredients and heating tube A is a, the distance between the target food ingredients and heating tube B is b, and the distance between the target food ingredients and heating tube C is c, and the distance a is greater than the distance b which is greater than the distance c. Correspondingly, the heating weight parameter of heating tube A is set to be greater than the heating weight parameter of heating tube B which is greater than the heating weight parameter of heating tube C.

[0092] It can be understood that the number of food ingredients of the identified target food ingredients may be one or more. When the number of target food ingredients is multiple and the cooking parameters corresponding to each target food ingredient are different, then the target food ingredient with a larger cooking parameter setting is selected as the final cooking parameter. Exemplarily, if the cooking temperature of target food ingredient A is Ca, the cooking time is Ta, and the heating weight parameter of the heating tube is Wa; the cooking temperature of target food ingredient B is Cb, the cooking time is Tb, and the heating weight parameter of the heating tube is Wb. Then the larger value between cooking temperature Ca and cooking temperature Cb is used as the final target cooking temperature. The larger value between cooking time Ta and cooking time Tb is used as the final target cooking time. The maximum value among the heating weight parameters corresponding to each heating tube between heating weight parameter Wa and heating weight parameter Wb of the heating tube is compared and taken out to generate a new heating weight parameter Wc of the heating tube as the final heating weight parameter of the heating tube.

[0093] In the above embodiment, by considering the food ingredient attributes and food ingredient positions during the process of determining the cooking parameters, and determining the heating weight parameter of the heating tube based on the positional relationship between the food ingredient position and the heating tube, the accuracy of determining the cooking parameters is improved.

[0094] It should be noted that due to the different altitudes, air pressures, and boiling points in each regional area, there are also certain differences in the temperature and time required for the food ingredients to be fully cooked. Therefore, this embodiment also provides a temperature and time automatic compensation method to improve the accuracy of determining the cooking temperature and cooking time of the food ingredients.

[0095] In an alternative embodiment, determining the target cooking temperature and target cooking time according to the food ingredient attributes of each target food ingredient includes: determining the initial cooking temperature and initial cooking time according to the food ingredient attributes of each target food ingredient; determining the geographical location parameter information according to the communication address information of its own device; determining the temperature compensation parameter and time compensation parameter according to the geographical location parameter information; determining the target cooking temperature according to the temperature compensation parameter and the initial cooking temperature; and determining the target cooking time according to the time compensation parameter and the initial cooking time.

[0096] Among them, the geographical location parameter information may include the region to which the device belongs, as well as the altitude, air pressure, boiling point, etc. corresponding to this region. Different geographical location parameter information corresponds to different temperature compensation parameters and time compensation parameters, which can be specifically determined based on a preset mapping relationship table. This mapping relationship table can be pre-generated and stored locally. Alternatively, it can also be obtained by predicting based on a pre-trained parameter compensation model. Specifically, historical geographical location parameter information can be used to train a pre-constructed network model to obtain a prediction network model for predicting temperature parameter compensation and time parameter compensation.

[0097] It should be noted that, to further improve the determination efficiency of the compensation parameters, the execution entity for determining the time compensation parameter and the temperature compensation parameter described above can be executed by the server (cloud). Specifically, the cooking device (mobile terminal) uploads the communication address information of its own device to the server according to its own networking module. The server converts the communication address information of the mobile terminal into a regional address to obtain the region to which the corresponding mobile terminal belongs. According to the region to which the mobile terminal belongs, query the altitude parameter mapping table to obtain the temperature compensation parameter and the time compensation parameter of the region to which the mobile terminal belongs. Among them, the altitude parameter mapping table is pre-constructed by the cloud and stored locally, which records the altitude, boiling point, air pressure, temperature compensation parameter, time compensation parameter, etc. of different regions. The mobile terminal feeds back the determined temperature compensation parameter and time compensation parameter to the cooking device (mobile terminal), and the cooking device determines the target cooking temperature according to the temperature compensation parameter and the initial cooking temperature; determines the target cooking time according to the time compensation parameter and the initial cooking time.

[0098] The above technical solution determines the temperature compensation parameter and the time compensation parameter based on the geographical location information of the cooking device, reduces the determination error of the target cooking time and the target cooking temperature, improves the determination accuracy of the target cooking time and the target cooking temperature, and thus ensures that the ingredients can be fully cooked.

[0099] Embodiment 4

[0100] Figure 4 This is a flowchart of a method for controlling the cooking of ingredients provided in Embodiment 4 of the present invention. Based on the above embodiments, this embodiment provides a preferred example. This embodiment uses the cloud as the server and the cooking device as the mobile terminal. As Figure 4 shown, the method includes the following specific steps:

[0101] This embodiment uses the Pytoch deep learning framework and takes a model that combines the YOLO-V8 model structure and the CBAM (Convolutional Block Attention Module) structure as the training base model, that is, the detection network model. The specific structure of the detection network model includes the YOLO-V8 network module and CBAM. Among them, the YOLO-V8 network module includes a backbone network, a neck network, and a detection head; among them, the input end of CBAM is connected to the output end of the backbone network; the output end of CBAM is connected to the input end of the neck network. Specifically, the model training method for the target ingredient detection model used for ingredient image target recognition and category detection is as follows:

[0102] Obtain ingredient sample images of different ingredients in different cooking devices, and label the ingredient types and ingredient positions in the ingredient sample images to obtain ingredient sample images with labels of standard ingredient types and standard ingredient positions. Input the ingredient sample images into the backbone network of the YOLO-V8 network module for feature extraction to obtain multi-scale feature maps; input the multi-scale feature maps into the convolutional block attention module for feature map weighting processing to obtain weighted feature maps; input the weighted feature maps into the neck network of the YOLO-V8 network module for feature map fusion processing to obtain fused feature maps; input the fused feature maps into the detection head of the YOLO-V8 network module for feature processing to obtain predicted ingredient types and predicted ingredient positions; train the detection network model according to the standard ingredient types, standard ingredient positions, predicted ingredient types, and predicted ingredient positions of the ingredient sample images until the model training end condition is met to obtain the target ingredient detection model.

[0103] Deploy the target ingredient detection model on the server side, perform lightweight processing on the target ingredient detection model to obtain a lightweight ingredient detection model, and deploy the lightweight ingredient detection model on the mobile side. The server side can make full use of the high performance of the model for ingredient recognition, but this model is vulnerable to network latency and other influences; while the model detection accuracy of the lightweight model on the mobile side is relatively low, but it is not easily affected by network latency.

[0104] Therefore, this embodiment also provides an intelligent decision-making method for the execution location of image recognition of cooking ingredient images, so as to realize the dynamic recognition of the execution location of the ingredient image detection task. Specifically, monitor the network status of the server side and the network status of the mobile side respectively to obtain the current network status of the server side and the current network status of the mobile side. According to the current network status of the server side and the current network status of the mobile side, determine the network quality of the server side and the network quality of the mobile side respectively; according to the network quality of the server side and the network quality of the mobile side, determine the execution location for performing the ingredient image detection task.

[0105] Specifically, based on machine learning or deep learning techniques, by learning the user behavior patterns (selection results of performing food ingredient image detection tasks on the server side or the mobile side) in historical time periods, as well as the network status under such user behavior patterns, a classification model can be trained to obtain a trained behavior pattern network model for determining the execution location on the mobile side or the server side.

[0106] When the network condition is poor, it can clearly prompt the user that the lightweight food ingredient detection model on the mobile side will be used for recognition and inform that the recognition accuracy may decrease. At the same time, provide the user with the option to manually switch the recognition mode according to actual needs and network conditions. If the recognition on the mobile side is inaccurate, it can prompt to check the network. After the network recovers, selectively re-upload the cooking detection picture to the server for verification to provide more accurate results. For misrecognized pictures, provide a correction mechanism to allow the user to manually correct or re-recognize.

[0107] If it is executed on the mobile side, a pre-trained lightweight food ingredient detection model is used to perform image detection on the cooking food ingredient image to obtain the food ingredient types and food ingredient positions of at least one target food ingredient output by the model. If it is executed on the server side, a pre-trained target food ingredient detection model is used to perform image detection on the cooking food ingredient image to obtain the food ingredient types and food ingredient positions of at least one target food ingredient output by the model, and feedback the food ingredient types and food ingredient positions to the mobile side.

[0108] The mobile side determines the food ingredient attributes of each target food ingredient respectively according to the food ingredient types of each target food ingredient. Among them, the food attributes can include food texture, nutritional value, cooking difficulty, etc. The mapping relationship between the food ingredient types and the corresponding food ingredient attributes can be pre-generated by relevant technical personnel and stored locally in the form of a mapping relationship database table.

[0109] According to the food ingredient attributes of each target food ingredient, the cooking temperature and cooking time are determined. Among them, different food ingredient attributes correspond to the same or different cooking temperatures and cooking times. According to the food ingredient positions of each target food ingredient, the heating weight parameters of each heating tube deployed in the cooking device are determined. Since there are multiple heating tubes in the cavity of the cooking device, for any heating tube, when the food ingredient is closer to the position of the heating tube, the heating tube should be set to a lower heating power; when the food ingredient is farther from the position of the heating tube, the heating tube should be set to a higher heating power. Therefore, the setting of the heating power of different heating tubes is related to the food ingredient position in the cavity. Among them, the heating weight parameter of the heating tube can be the power weight corresponding to the heating power of the heating tube. When the position distance between the food ingredient and the heating tube is closer, the corresponding power weight is lower; when the position distance between the food ingredient and the heating tube is farther, the corresponding power weight is higher.

[0110] Due to the different altitudes, air pressures, and boiling points across the country, the temperatures and times required for food to be fully cooked vary. Therefore, the server-based automatic compensation algorithm compensates the parameters of the cooking temperature and cooking time, and sends the compensated parameters to the mobile terminal. Specifically, the mobile terminal sends the communication address information of its own device to the server, and the server converts the communication address information of the mobile terminal into regional address information, and determines the geographical location parameter information of the region to which the regional address information belongs in a table lookup manner; according to the geographical location parameter information, the temperature compensation parameter and the time compensation parameter are determined, which can be specifically implemented by table lookup. The cooking temperature is compensated according to the temperature compensation parameter to obtain the target cooking temperature; and, the cooking time is compensated according to the time compensation parameter to obtain the target cooking time. Cooking parameters including the target cooking temperature, the target cooking time, and the heating weight parameter of the heating tube are generated.

[0111] Embodiment 5

[0112] Figure 5 The following is a schematic structural diagram of a food cooking control device provided in Embodiment 5 of the present invention. The food cooking control device provided in the embodiment of the present invention can be applied to the situation of accurately determining the cooking parameters of a cooking device to control food cooking based on the cooking parameters. The food cooking control device can be implemented in the form of hardware and / or software, such as Figure 5 As shown, the device specifically includes: a food ingredient image acquisition module 501, an image detection module 502, a cooking parameter determination module 503, and a control execution module 504. Among them,

[0113] The food ingredient image acquisition module 501 is used to acquire a cooking food ingredient image;

[0114] The image detection module 502 is used to perform image detection on the cooking food ingredient image by using a pre-trained lightweight food ingredient detection model to obtain the food ingredient type and food ingredient position of at least one target food ingredient output by the model;

[0115] The cooking parameter determination module 503 is used to determine cooking parameters according to the food ingredient types and food ingredient positions of the target food ingredients;

[0116] The control execution module 504 is used to perform corresponding device control actions based on the cooking parameters to obtain a cooking result.

[0117] The technical solution of the present invention comprehensively considers the types and positions of ingredients during the process of determining cooking parameters. At the same time, based on a pre-trained lightweight ingredient detection model, image detection is performed on the cooking ingredient images, realizing effective and accurate identification of ingredients, rather than completely controlling cooking based on the cooking parameters corresponding to pre-selected prefabricated recipes, improving the flexibility of the cooking process and the accurate determination of cooking parameters, thereby improving the cooking quality.

[0118] Optionally, the device further includes a model training module; the model training module includes:

[0119] An ingredient sample image acquisition unit, configured to acquire ingredient sample images of different ingredients in different cooking devices, and label the ingredient types and ingredient positions in the ingredient sample images to obtain ingredient sample images with labels of standard ingredient types and standard ingredient positions;

[0120] A model training unit, configured to use the ingredient sample images to perform model training on a pre-constructed detection network model to obtain a trained target ingredient detection model;

[0121] A model quantization processing unit, configured to perform model quantization processing on the target ingredient detection model to obtain a lightweight ingredient detection model.

[0122] Optionally, the model structure of the detection network model includes a YOLO-V8 network module and a convolutional block attention module; the YOLO-V8 network module includes a backbone network, a neck network, and a detection head; wherein, the input end of the convolutional block attention module is connected to the output end of the backbone network; the output end of the convolutional block attention module is connected to the input end of the neck network; correspondingly, the model training unit is specifically configured to:

[0123] Input the ingredient sample images into the backbone network of the YOLO-V8 network module for feature extraction to obtain multi-scale feature maps;

[0124] Input the multi-scale feature maps into the convolutional block attention module for feature map weighting processing to obtain weighted feature maps;

[0125] Input the weighted feature maps into the neck network of the YOLO-V8 network module for feature map fusion processing to obtain fused feature maps;

[0126] Input the fused feature maps into the detection head of the YOLO-V8 network module for feature processing to obtain predicted ingredient types and predicted ingredient positions;

[0127] The detection network model is trained according to the standard ingredient types, standard ingredient positions, predicted ingredient types, and predicted ingredient positions of the ingredient sample images until the model training end condition is met, and a target ingredient detection model is obtained.

[0128] Optionally, the device further includes:

[0129] A network status detection module, configured to monitor the network status of the server to obtain the current network status information before performing image detection on the cooking ingredient image using the pre-trained lightweight ingredient detection model to obtain the ingredient types and ingredient positions of at least one target ingredient output by the model;

[0130] An image detection judgment module, configured to judge whether to perform image detection on the cooking ingredient image using the pre-trained lightweight ingredient detection model according to the current network status information;

[0131] A server detection module, configured to, if the pre-trained lightweight ingredient detection model is not used to perform image detection on the cooking ingredient image, send the cooking ingredient image to the server for the server to perform image detection on the cooking ingredient image based on the pre-trained target ingredient detection model locally deployed on the server to obtain the ingredient types and ingredient positions of at least one target ingredient output by the model, and feedback the ingredient types and ingredient positions of the target ingredient to the cooking device;

[0132] A mobile device detection module, configured to, if the pre-trained lightweight ingredient detection model is used to perform image detection on the cooking ingredient image, perform the image detection on the cooking ingredient image using the pre-trained lightweight ingredient detection model to obtain the ingredient types and ingredient positions of at least one target ingredient output by the model.

[0133] Optionally, the cooking parameter determination module 503 includes:

[0134] An ingredient attribute determination unit, configured to respectively determine the ingredient attributes of each target ingredient according to the ingredient types of each target ingredient;

[0135] A cooking parameter determination unit, configured to determine cooking parameters according to the ingredient attributes and ingredient positions of each target ingredient.

[0136] Optionally, the cooking parameter determination unit includes:

[0137] A target parameter determination subunit, configured to determine a target cooking temperature and a target cooking time according to the ingredient attributes of each target ingredient;

[0138] A heating weight determination subunit, configured to determine heating tube heating weight parameters of each heating tube deployed in the cooking device according to the food ingredient positions of the target food ingredients;

[0139] A cooking parameter determination subunit, configured to generate cooking parameters including a target cooking temperature, a target cooking time, and heating tube heating weight parameters.

[0140] Optionally, the target parameter determination subunit is specifically configured to:

[0141] Determine an initial cooking temperature and an initial cooking time according to the food ingredient attributes of the target food ingredients;

[0142] Determine geographical location parameter information according to the communication address information of its own device;

[0143] Determine a temperature compensation parameter and a time compensation parameter according to the geographical location parameter information;

[0144] Determine a target cooking temperature according to the temperature compensation parameter and the initial cooking temperature; and,

[0145] Determine a target cooking time according to the time compensation parameter and the initial cooking time.

[0146] The food ingredient cooking control device provided by the embodiments of the present invention can execute the food ingredient cooking control method provided by any embodiment of the present invention, and has corresponding functional modules and beneficial effects for executing the method.

[0147] Embodiment Six

[0148] Figure 6 FIG. shows a schematic structural diagram of a cooking device 60 that can be used to implement the embodiments of the present invention. The cooking device is intended to represent various forms of digital computers, such as, a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The cooking device can also represent various forms of mobile devices, such as, a personal digital processor, a cellular phone, a smart phone, a wearable device (such as a helmet, glasses, a watch, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are only examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0149] As Figure 6As shown, the cooking device 60 includes at least one processor 61 and a memory communicatively connected to the at least one processor 61, such as a read-only memory (ROM) 62, a random access memory (RAM) 63, etc. The memory stores a computer program executable by the at least one processor. The processor 61 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 62 or the computer program loaded from the storage unit 68 into the random access memory (RAM) 63. In the RAM 63, various programs and data required for the operation of the cooking device 60 can also be stored. The processor 61, the ROM 62, and the RAM 63 are connected to each other via a bus 64. An input / output (I / O) interface 65 is also connected to the bus 64.

[0150] Multiple components in the cooking device 60 are connected to the I / O interface 65, including: an input unit 66, such as a keyboard, a mouse, etc.; an output unit 67, such as various types of displays, speakers, etc.; a storage unit 68, such as a magnetic disk, an optical disc, etc.; and a communication unit 69, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 69 allows the cooking device 60 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0151] The processor 61 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 61 include but are not limited to a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 61 executes the various methods and processes described above, such as the food cooking control method.

[0152] In some embodiments, the food cooking control method can be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as the storage unit 68. In some embodiments, part or all of the computer program can be loaded and / or installed onto the cooking device 60 via the ROM 62 and / or the communication unit 69. When the computer program is loaded into the RAM 63 and executed by the processor 61, one or more steps of the food cooking control method described above can be executed. Alternatively, in other embodiments, the processor 61 can be configured to execute the food cooking control method in any other appropriate manner (e.g., by means of firmware).

[0153] The various embodiments of the systems and techniques described above in this specification can be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems-on-chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which may be a special-purpose or general-purpose programmable processor that receives data and instructions from, and transmits data and instructions to, a storage system, at least one input device, and at least one output device.

[0154] The computer programs for implementing the methods of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions / operations specified in the flowchart and / or block diagram to be implemented. The computer programs can be executed entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.

[0155] In the context of the present invention, a computer-readable storage medium can be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, the computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0156] To provide interaction with a user, the systems and techniques described herein can be implemented on a cooking device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) through which the user can provide input to the cooking device. Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0157] The systems and techniques described herein can be implemented in a computing system including backend components (e.g., as a data server), or a computing system including middleware components (e.g., an application server), or a computing system including frontend components (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system including any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected to each other by digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include: local area network (LAN), wide area network (WAN), blockchain network, and the Internet.

[0158] The computing system can include a client and a server. The client and the server are generally remote from each other and typically interact through a communication network. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system and solves the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.

[0159] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps recited in the present invention can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved, and no limitation is made herein.

[0160] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for controlling the cooking of food ingredients, characterized in that, Including: Obtain a cooking ingredient image; Monitor the network status of the server to obtain the current network status information; Based on the current network status information, determine whether to perform image detection on the cooking ingredient image using a pre-trained lightweight ingredient detection model; If not, send the cooking ingredient image to the server for the server to perform image detection on the cooking ingredient image based on a pre-trained target ingredient detection model locally deployed on itself, obtain the ingredient types and ingredient positions of at least one target ingredient output by the model, and feedback the ingredient types and ingredient positions of the target ingredient to the cooking device; If so, perform image detection on the cooking ingredient image using a pre-trained lightweight ingredient detection model to obtain the ingredient types and ingredient positions of at least one target ingredient output by the model; the model performance and ingredient recognition accuracy of the target ingredient detection model are higher than those of the lightweight ingredient detection model; Determine cooking parameters based on the ingredient types and ingredient positions of each target ingredient; Execute corresponding device control actions based on the cooking parameters to obtain a cooking result.

2. The food ingredient cooking control method according to claim 1, wherein The model training method of the lightweight ingredient detection model is as follows: Obtain ingredient sample images of different ingredients in different cooking devices, and label the ingredient types and ingredient positions in the ingredient sample images to obtain ingredient sample images with labels of standard ingredient types and standard ingredient positions; Use the ingredient sample images to perform model training on a pre-constructed detection network model to obtain a trained target ingredient detection model; Perform model quantization processing on the target ingredient detection model to obtain a lightweight ingredient detection model.

3. The food ingredient cooking control method according to claim 2, characterized in that, The model structure of the detection network model includes a YOLO-V8 network module and a convolutional block attention module; the YOLO-V8 network module includes a backbone network, a neck network, and a detection head; wherein, the input end of the convolutional block attention module is connected to the output end of the backbone network; the output end of the convolutional block attention module is connected to the input end of the neck network; Correspondingly, the step of using the ingredient sample images to perform model training on a pre-constructed detection network model to obtain a trained target ingredient detection model includes: Input the ingredient sample images into the backbone network of the YOLO-V8 network module for feature extraction to obtain multi-scale feature maps; Input the multi-scale feature maps into the convolutional block attention module for feature map weighting processing to obtain weighted feature maps; Input the weighted feature maps into the neck network of the YOLO-V8 network module for feature map fusion processing to obtain fused feature maps; Input the fused feature maps into the detection head of the YOLO-V8 network module for feature processing to obtain predicted ingredient types and predicted ingredient positions; Perform model training on the detection network model according to the standard ingredient types, standard ingredient positions, predicted ingredient types, and predicted ingredient positions of the ingredient sample images until the model training end condition is met to obtain a target ingredient detection model.

4. The food ingredient cooking control method according to claim 1, wherein Determining cooking parameters according to the food types and food positions of each target food ingredient, including: Respectively determining the food ingredient attributes of each target food ingredient according to the food types of each target food ingredient; Determining cooking parameters according to the food ingredient attributes and food positions of each target food ingredient.

5. The method for controlling ingredient cooking according to claim 4, wherein, The determining of cooking parameters according to the food ingredient attributes and food positions of each target food ingredient includes: Determining a target cooking temperature and a target cooking time according to the food ingredient attributes of each target food ingredient; Determining the heating weight parameters of each heating tube deployed in the cooking device according to the food positions of each target food ingredient; Generating cooking parameters including the target cooking temperature, the target cooking time, and the heating weight parameters of the heating tube.

6. The method for controlling food cooking according to claim 5, wherein The determining of the target cooking temperature and the target cooking time according to the food ingredient attributes of each target food ingredient includes: Determining an initial cooking temperature and an initial cooking time according to the food ingredient attributes of each target food ingredient; Determining geographical location parameter information according to the communication address information of its own device; Determining a temperature compensation parameter and a time compensation parameter according to the geographical location parameter information; Determining the target cooking temperature according to the temperature compensation parameter and the initial cooking temperature; and, Determining the target cooking time according to the time compensation parameter and the initial cooking time.

7. An ingredient cooking control device, characterized in that, Including: A food ingredient image acquisition module, configured to acquire a cooking food ingredient image; An image detection module, configured to perform image detection on the cooking food ingredient image by using a pre-trained lightweight food ingredient detection model, and obtain the food types and food positions of at least one target food ingredient output by the model; A cooking parameter determination module, configured to determine cooking parameters according to the food types and food positions of each target food ingredient; A control execution module, configured to perform corresponding device control actions based on the cooking parameters to obtain a cooking result; The device further includes: A network status detection module, configured to monitor the network status of the server before performing image detection on the cooking food ingredient image by using the pre-trained lightweight food ingredient detection model to obtain at least one target food ingredient's food type and food position output by the model, and obtain the current network status information; An image detection judgment module, configured to judge whether to perform image detection on the cooking food ingredient image by using the pre-trained lightweight food ingredient detection model according to the current network status information; A server detection module, configured to, if not performing image detection on the cooking food ingredient image by using the pre-trained lightweight food ingredient detection model, send the cooking food ingredient image to the server for the server to perform image detection on the cooking food ingredient image by using the pre-trained target food ingredient detection model deployed locally thereon, obtain the food types and food positions of at least one target food ingredient output by the model, and feedback the food types and food positions of the target food ingredient to the cooking device; The mobile detection module is configured to, if the lightweight food ingredient detection model obtained by pre-training is used to perform image detection on the cooking food ingredient image, execute the image detection on the cooking food ingredient image using the lightweight food ingredient detection model obtained by pre-training, and obtain the food ingredient types and food ingredient positions of at least one target food ingredient output by the model; the model performance and food ingredient recognition accuracy of the target food ingredient detection model are higher than those of the lightweight food ingredient detection model.

8. A cooking device, characterized in that, The cooking device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores a computer program executable by the at least one processor, and when the computer program is executed by the at least one processor, the at least one processor is enabled to execute the food ingredient cooking control method according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a processor, the food ingredient cooking control method according to any one of claims 1-6 is implemented.

Images