Electric baking pan based on food material identification and dynamic temperature control and control method

Abstract

The invention relates to the technical field of intelligent kitchen appliances, and discloses an electric baking pan based on food material recognition and dynamic temperature control and a control method.The electric baking pan comprises a base and an upper cover rotationally connected with the base through a hinge, and the upper cover and the base jointly define a cooking cavity used for containing a cooking object in a closed state; the electric baking pan further comprises a first lock catch and a second lock catch, wherein the first lock catch is used for connecting and fixing the upper cover body in a closed state and the base, the second lock catch is used for unlocking the first lock catch, and heating assemblies are arranged in the upper cover body and the base respectively. The electric baking pan further comprises an image collecting module, and the collecting view field of the image collecting module covers the cooking cavity. Through the image acquisition module and the auxiliary light source, the food material type can be identified, individualized physical characteristics such as the thickness and the size of the food material can be accurately measured, and it is ensured that the stable and high-quality cooking effect can be achieved no matter what food materials are used for several times and what food materials are cooked.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent kitchen appliances, in particular to an electric baking pan based on food material recognition and dynamic temperature control and a control method. BACKGROUND

[0002] As a popular household kitchen appliance, the electric baking pan can be used for grilling, baking and frying, and is particularly suitable for making pancakes, barbecued meat and fried dumplings.

[0003] The control mode of the currently marketed electric baking pan is relatively simple. Most products rely on a simple timer and a few fixed power levels (such as low, medium and high heat), and the user needs to manually set the heating time and heat according to personal experience. Even some advanced models equipped with preset menus for specific foods (such as steak mode and pancake mode) are essentially still executing a fixed, open-loop heating program.

[0004] The common defect of these prior art solutions is the lack of perception and adaptability to the physical state of the cooking object. Whether it is manual setting or preset menu, the heating strategy is fixed and completely ignores the specific differences in size, thickness, initial temperature, etc. of the food material in each cooking. For example, a fixed steak mode cannot distinguish between a 1 cm thick sirloin and a 3 cm thick fillet, resulting in unstable cooking results and often resulting in a burnt exterior and raw interior or a dry and overcooked whole. Therefore, the user often needs to repeatedly open the lid and subjectively judge the doneness by visual observation and manual pressing, which not only complicates the operation, disrupts the stable heating environment and prolongs the cooking time, but also greatly depends on the user's cooking skills and experience, and cannot guarantee the stability and consistency of the cooking quality, far from achieving true automation and intelligence.

[0005] In view of this, the present application aims to provide an electric baking pan based on food material recognition and dynamic temperature control and a control method to solve the problems existing in the prior art. SUMMARY

[0006] In view of the deficiencies of the prior art, the present application provides an electric baking pan based on food material recognition and dynamic temperature control and a control method, which solves the problem that the existing electric baking pan cannot adaptively adjust according to the specific physical characteristics such as the thickness and size of the food material due to the use of a fixed heating program, resulting in unstable cooking results and high dependence on user experience.

[0007] To achieve the above object, the present application is implemented by the following technical solutions: An electric baking pan based on food material recognition and dynamic temperature control, comprising a base and an upper cover body rotatably connected to the base through a hinge, the upper cover body and the base jointly enclose a cooking cavity for placing a cooking object in a closed state; The electric baking pan further comprises a first lock for connecting and fixing the upper cover body and the base in the closed state, and a second lock for unlocking the first lock, and the upper cover body and the base are respectively provided with a heating assembly; The electric baking pan further comprises: An image acquisition module for acquiring a field of view covering the cooking cavity; A temperature sensor for monitoring the temperature in the cooking cavity; A controller electrically connected to the heating assembly, the image acquisition module and the temperature sensor.

[0008] Preferably, the electric baking pan further comprises an auxiliary light source arranged in the cooking cavity for projecting a shadow of the cooking object in the cooking cavity, and the controller is used to determine the equivalent thickness of the cooking object.

[0009] A control method of an electric baking pan based on food material recognition and dynamic temperature control, comprising the following steps: a) After the upper cover body and the base are closed through the hinge to form a cooking cavity, initial image data and initial temperature of a cooking object in the cooking cavity are acquired by an image acquisition module and a temperature sensor arranged inside the cooking cavity; b) The controller determines at least one physical feature of the cooking object based on the initial image data using an image recognition algorithm, the physical feature including size and equivalent thickness; c) The controller reversely generates an initial heating power curve based on the physical feature and the initial temperature by solving a predictive process model with a target cooking state as a terminal constraint; d) The controller controls the heating assembly arranged in the upper cover body and the base respectively to perform heating according to the initial heating power curve, and the image acquisition module continuously acquires real-time image data of the cooking object during the cooking process; e) The controller analyzes the real-time image data in real time to monitor the surface color change of the cooking object, and performs control strategy re-planning on the heating power curve when the actual progress of the surface color change deviates from the progress predicted by the predictive process model.

[0010] Preferably, in step b), the method for determining the equivalent thickness of the cooking object comprises: The auxiliary light source is arranged in the cooking cavity to project a shadow of the cooking object on the heating plate of the base, thereby constructing a triangular geometric model defined by an equivalent thickness of the cooking object, a length of the shadow, and a light path of the auxiliary light source; The controller analyzes the initial image data, identifies and measures a characteristic length of the shadow, thereby obtaining a known side of the triangular geometric model; Based on the obtained known side length and a preset light incidence angle of the auxiliary light source as a known angle in the triangular geometric model, the equivalent thickness of the cooking object is calculated according to an inherent trigonometric function relationship of the triangular geometric model.

[0011] Preferably, in the step c), the predictive process model comprises a coupled dynamic equation for describing changes of a core temperature and a surface color of the cooking object with heating power and time; The process of reversely generating the initial heating power curve is an optimal control problem of solving a target function aiming to minimize a cooking time, and a terminal state constraint is a target core temperature and a target surface color of the cooking object.

[0012] Preferably, the controller further defines a cooking progress index for quantifying a cooking completion degree of the cooking object in the cooking cavity; In the step e), the actual progress of the surface color change is monitored, specifically, the cooking progress index is continuously updated and corrected by fusing a prediction of the predictive process model and an actual surface color of the cooking object in the cooking cavity analyzed based on the real-time image data.

[0013] Preferably, the step of updating and correcting the cooking progress index comprises: According to a state at a previous time, a predicted cooking progress index increment is calculated by the predictive process model; An observation residual error between an actual surface color of the cooking object in the cooking cavity analyzed based on the real-time image data and a surface color predicted by the predictive process model is calculated; The predicted cooking progress index increment and the observation residual error are fused by weighting to obtain a corrected cooking progress index at a current time.

[0014] Preferably, in the step e), when a deviation between the corrected cooking progress index and a planned progress trajectory corresponding to the initial heating power curve exceeds a preset threshold, a control strategy re-planning for the upper cover body and the heating assembly in the base is triggered.

[0015] Preferably, when the cooking process index reaches a preset completion threshold, and the surface color of the cooking object in the cooking cavity monitored by the real-time image data enters the target range and stabilizes for a preset duration, the heating of the upper cover body and the heating assembly in the base is terminated.

[0016] Preferably, in the step b), the controller further constructs the type ID, size, equivalent thickness and initial temperature of the cooking object in the cooking cavity into an initial state vector, and takes the initial state vector as the input of the predictive process model.

[0017] The application provides an electric baking pan based on food material recognition and dynamic temperature control and a control method. 1、The application can not only identify the type of food material, but also accurately measure the individual physical characteristics such as thickness and size through the image acquisition module and auxiliary light source. The controller customizes the exclusive optimal heating curve for each food material based on the real-time accurate data, instead of using a fixed preset program. This targeted control scheme fundamentally solves the problem of uneven cooking or different cooking time caused by the individual differences of food materials (such as the thickness of beef and the size of dumplings) in traditional kitchen utensils, ensuring stable and high-quality cooking results regardless of the number of uses or the size of food materials.

[0018] 2、The heating curve generation module of the application constructs the cooking process as an optimal control problem with the shortest time as the objective function. Through inverse solving based on the predictive process model, the controller can calculate a heating power path that meets all process constraints (such as core thorough cooking and proper surface charring) and has the shortest time. This method avoids the time redundancy and heat waste caused by traditional segmented control or constant power heating, realizes the most efficient energy transfer from the initial state to the target state, and thus maximizes the saving of user time and electrical energy consumption while ensuring cooking quality.

[0019] 3、In the application, the user only needs to put the food material into the electric baking pan and close the upper cover, without any selection or setting. The device automatically completes the whole process of food material recognition, state perception, strategy planning, process monitoring and dynamic adjustment. In particular, the dual termination condition judgment mechanism based on the cooking process index and the actual surface color can intelligently judge whether the cooking is truly completed. This highly automated put-and-go reduces the use threshold and brings users an unprecedented convenient and easy cooking experience. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 is the overall structural diagram of the application; Figure 2A control method flowchart of the present application; Figure 3 An initial state acquisition flowchart of the present application; Figure 4 A physical feature determination flowchart of the present application; Figure 5 A foodstuff equivalent thickness calculation principle diagram of the present application.

[0021] Wherein, 1, upper cover body; 2, base; 3, hinge; 4, cooking cavity; 5, second lock catch; 6, first lock catch. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the specification of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0023] Referring to the drawings Figure 1 The embodiment of the present application provides an electric baking pan based on foodstuff recognition and dynamic temperature control, which comprises a base 2 and an upper cover body 1 rotatably connected with the base 2 through a hinge 3. The upper cover body 1 and the base 2 jointly enclose a cooking cavity 4 for placing a cooking object in the closed state. The electric baking pan further comprises a first lock catch 6 for connecting and fixing the upper cover body 1 and the base 2 in the closed state, and a second lock catch 5 for unlocking the first lock catch 6. The upper cover body 1 and the base 2 are respectively provided with a heating assembly. The electric baking pan further comprises: an image acquisition module for acquiring a field of view covering the cooking cavity; a temperature sensor for monitoring the temperature in the cooking cavity; a controller electrically connected with the heating assembly, the image acquisition module and the temperature sensor; The electric baking pan further comprises an auxiliary light source arranged in the cooking cavity 4 for projecting a shadow of the cooking object in the cooking cavity 4, and the controller is used for determining the equivalent thickness of the cooking object. Specifically, the electric baking pan provided by the present application comprises a base 2 and an upper cover body 1. The upper cover body 1 is rotatably connected with the base 2 through a hinge 3 arranged on the back side of the upper cover body 1, so as to allow the upper cover body 1 to open and close relative to the base 2. When the upper cover body 1 is in the closed position, it jointly encloses a cooking cavity 4 with the base 2 for accommodating and heating a cooking object. A lower heating disc is arranged on the upper surface of the base 2, and the cooking object is directly placed on the lower heating disc.

[0024] To reliably secure the upper cover body 1 and the base 2 in the closed state during cooking, the electric baking pan is further provided with a set of locking mechanisms. The mechanisms include a first lock 6 provided on the upper cover body 1 and a card slot provided on the base 2 and cooperating with the first lock 6. A second lock 5 for unlocking the first lock 6 is provided as a pressable button, which is mechanically linked to the first lock 6. By pressing the second lock 5, the user can make the first lock 6 disengage from the card slot, thereby opening the upper cover body 1.

[0025] The heating function of the electric baking pan is realized by heating assemblies provided inside the upper cover body 1 and the base 2, respectively. Specifically, an upper heating disc is provided inside the upper cover body 1, and a lower heating disc is provided inside the base 2. The two heating discs work together to heat the cooking object in the cooking cavity 4 from both the top and bottom.

[0026] To realize intelligent perception and control of the cooking process, the electric baking pan further integrates an electronic system. The system includes an image acquisition module, a temperature sensor, and a controller 7. The image acquisition module, for example, a digital camera with a built-in wide-angle lens, is installed on the inner surface of the upper cover body 1, and its position and field of view are designed to ensure that its acquisition field of view can completely cover the area of the lower heating disc on the base 2. The temperature sensor, for example, an NTC thermistor or an infrared temperature sensor, is provided in the cooking cavity 4 to monitor the temperature inside the cavity in real time. In an embodiment, the electric baking pan further includes an auxiliary light source, for example, an LED lamp bead, which is also provided on the inner surface of the upper cover body 1, and its position has a fixed geometric angle relative to the image acquisition module. The role of the auxiliary light source is to illuminate under the instruction of the controller 7, so as to project a clear shadow on the cooking object. The shadow image information is used by the controller 7 for subsequent thickness calculation.

[0027] The controller 7 is the core of the entire electronic system, and its physical entity is a printed circuit board assembly PCBA with a microprocessor MCU. The printed circuit board assembly PCBA also integrates a storage unit, such as a non-volatile flash memory and a random access memory RAM. The non-volatile flash memory is used to permanently store the control program of the present application, the preset predictive process model, the food material feature database, various parameter thresholds, etc.; the random access memory is used as a cache area for program runtime, temporarily storing the collected image data, intermediate calculation results, and generated heating power curve. The electric baking pan can also include a user interface module electrically connected to the controller 7 for receiving user input cooking preferences such as desired doneness or displaying cooking status to the user. The module can include a display screen or status indicator light.

[0028] The controller 7 is connected with various functional components through different electrical interfaces: it is connected with the upper and lower heating components through a driving circuit to control the output of their heating power; it is connected with the image acquisition module through a data bus, such as a MIPI or USB interface, to receive the digital image data acquired by the module; it is connected with the temperature sensor through an analog-digital conversion interface to read the digitized temperature value; it is also connected with the auxiliary light source through an I / O port to control its turning on and off. The controller 7 coordinates the work of all these components by executing the internal fixed program algorithm, thereby realizing the complete automated cooking method of the present application.

[0029] With reference to the accompanying drawings Figure 2 , the present application provides a control method for an electric baking pan based on food material recognition and dynamic temperature control, which is implemented in an electric baking pan comprising a controller 7. The controller 7 can internally include, but is not limited to, an initial state acquisition module, a physical feature determination module, a heating curve generation module, a closed-loop control module, and a cooking termination module.

[0030] The specific flow of the method is as follows: After the user places the cooking object on the heating disc of the base 2 and closes the upper cover body 1, the initial state acquisition module in the controller 7 is activated. This module drives the image acquisition module arranged in the cooking cavity 4 to capture an initial image of the cooking object and simultaneously reads the reading of the temperature sensor to obtain the initial temperature in the cooking cavity 4. The initial image data and the initial temperature are transmitted to the physical feature determination module. After receiving the initial image data, the physical feature determination module executes an image recognition algorithm. This module first calculates the size of the cooking object in the two-dimensional plane, such as length, width, or area, through edge detection and contour analysis. Then, the module determines the equivalent thickness of the cooking object by using the shadow formed on the cooking object by the auxiliary light source. Specifically, the module identifies and measures the characteristic length of the shadow from the initial image data , and reads the preset light incidence angle of the auxiliary light source from the memory The equivalent thickness is calculated by the following trigonometric function relationship : ; wherein, is the equivalent thickness of the cooking object; is the measured characteristic length of the shadow; is the preset light incidence angle of the auxiliary light source.

[0031] After determining the physical characteristics, the module also integrates the recognized cooking object type ID, size, equivalent thickness H, and the obtained initial temperature into an initial state vector. Subsequently, the initial state vector is passed to the heating curve generation module. This module takes this vector as the initial condition of the predictive process model and takes the preset cooking target, such as the target core temperature and the target surface color, as the terminal state constraint of the model, and solves an optimal control problem with the objective function of minimizing the cooking time. The output of this solving process is an initial heating power curve, which defines the power value that the heating assembly should output at each time from the current time to the completion of the cooking. The initial heating power curve is sent to the closed-loop control module. According to the curve, the closed-loop control module controls the heating assembly arranged in the upper cover body 1 and the base 2 to start heating. During the heating process, the closed-loop control module continuously obtains real-time cooking object image data from the image acquisition module and monitors the cooking process in real time based on the real-time image data. To this end, the closed-loop control module maintains a cooking process index , and continuously updates it using a prediction correction mechanism. This updating process can be represented by the following formula: ; Wherein, is the corrected cooking process index at the current time k; is the cooking process index at the previous time , and is the process index increment predicted by the predictive process model based on the state at the previous time; is the weighting factor at the current time; is the quantitative value of the actual surface color analyzed by the real-time image data; and is the quantitative value of the surface color predicted by the model.

[0032] In one specific embodiment, may be a constant that is set empirically and remains unchanged throughout the entire cooking process. In another preferred embodiment, may be a dynamic variable, and the value thereof is the Kalman gain calculated according to the Kalman filtering theory. The size of the gain depends on the uncertainty of the model prediction and the uncertainty of the actual observation, so as to achieve the optimal fusion of the model prediction and the actual observation.

[0033] The closed-loop control module also continuously compares the corrected cooking process index with the theoretical process trajectory planned by the initial heating power curve. When the deviation between the two exceeds a preset threshold, it indicates that the actual cooking process deviates significantly from the planning, and at this time, the control strategy re-planning is triggered. The re-planning process is to update the current corrected cooking state, including , the current temperature, etc. as new initial conditions, re-call the heating curve generation module, calculate a new optimal heating power curve from the current state to the final target state, to replace the remaining part of the original curve.

[0034] Finally, the cooking termination module monitors two termination conditions in parallel throughout the entire cooking process: First, the corrected cooking progress index whether it reaches the preset completion threshold; Second, whether the color of the cooking object surface monitored by real-time image data enters the target range and stabilizes within the target range for a preset duration. When and only when both conditions are met, the cooking termination module will send a stop signal to the driving circuit of the heating assembly, terminate the heating of the heating assembly in the upper cover body 1 and the base 2, and thus the entire cooking process ends.

[0035] Referring to the accompanying Figure 3 , the execution of the control method of the present application begins with a clear physical trigger event. When the user closes the upper cover body 1 with the base 2 and locks it by the first lock 6, a position sensor mechanically linked with the first lock 6, such as a microswitch or a Hall effect sensor, is triggered. The sensor sends a level signal to the controller 7 indicating that the cooking cavity is closed and locked. After receiving this signal, the controller 7 starts its internal initial state acquisition module and begins the initial state acquisition process.

[0036] The stable closure of the upper cover body 1 and the base 2 is ensured by the lock mechanism, which aims to create a repeatable, closed physical environment that is not disturbed by external light and air flow for subsequent image acquisition and temperature measurement, which is a prerequisite for ensuring the accuracy of initial state measurement.

[0037] The initial state acquisition module first activates the auxiliary light source set in the cooking cavity 4 to a preset, constant luminous intensity through a control signal. This step aims to provide a stable and external environment light-affected lighting condition for image acquisition. After the brightness of the auxiliary light source is stable, the module immediately instructs the image acquisition module to capture a frame of static digital image covering the heating plate area in the cooking cavity 4. The static digital image, for example, in the form of a two-dimensional pixel matrix, exists as initial image data, which is transmitted by the image acquisition module to the internal storage of the controller 7 through the data interface.

[0038] At the same time or immediately after the image acquisition instruction is issued, the initial state acquisition module sends a data read request to a temperature sensor disposed within the cooking cavity 4. The temperature sensor is disposed at a location that can reflect the initial ambient temperature inside the cooking cavity 4, such as a cavity side wall, and its measured value can be used as a reference for the initial temperature of the cooking object. The temperature sensor converts the measured analog signal into a digital temperature value via an analog-to-digital conversion circuit ADC and transmits it to the controller 7.

[0039] Finally, the initial state acquisition module stores the received initial image data and digitized initial temperature value as a unified data set in a designated data buffer area within the controller 7. The successful generation and storage of the data set mark the completion of the initial state acquisition step and serve as direct input for the subsequent physical feature determination module to process.

[0040] Referring to the accompanying drawings Figure 4 The flow is executed by the physical feature determination module within the controller 7, which receives the initial image data and initial temperature provided by the initial state acquisition module as input.

[0041] The physical feature determination module first pre-processes the received initial image data. The pre-processing step includes: converting the multi-channel color image into a single-channel grayscale image to simplify subsequent calculations; using a denoising algorithm such as Gaussian filtering to eliminate random noise generated during image acquisition; and enhancing image contrast through histogram equalization and other methods to highlight the outline of the cooking object and its shadow.

[0042] After pre-processing, the physical feature determination module begins to determine the two-dimensional size of the cooking object. This process is achieved by performing a binaryzation operation on the pre-processed grayscale image, such as using Otsu's method to automatically determine a threshold value and divide the image into pixel regions representing the cooking object and pixel regions representing the background. Subsequently, the physical feature determination module executes a contour detection algorithm on the binaryzation image to extract a closed contour representing the outer boundary of the cooking object. Based on this contour, a series of two-dimensional geometric parameters can be calculated, such as the total number of pixels enclosed by the contour, the minimum circumscribed rectangle of the contour, and the centroid of the contour.

[0043] Next, the physical feature determination module performs equivalent thickness calculation. Based on the fixed position of the auxiliary light source, a shadow will appear on the other side of the cooking object relative to the light source. The physical feature determination module performs pixel intensity analysis in the expected shadow area based on the determined cooking object contour in the pre-processed image, and divides out the shadow area by setting a darker grayscale threshold value. The physical feature determination module then calculates the characteristic length of the shadow , the calculation can be measuring the distance from the cooking object boundary to the outer boundary of the shadow on multiple preset axes perpendicular to the light direction, and taking the average. The physical feature determination module reads a preset constant value from its non-volatile memory, which represents the angle between the light of the auxiliary light source and the heating plate surface of the base 2, i.e. the light incidence angle . Then the equivalent thickness of the cooking object is calculated : After obtaining the size and thickness information, the physical feature determination module can also perform type recognition of the cooking object. This process combines multiple features calculated in the previous steps, such as area, aspect ratio, circularity of the contour, and initial average chroma extracted from the initial color image, into a multi-dimensional feature vector. The physical feature determination module inputs this feature vector into a pre-trained and stored classification model in the controller 7, such as a support vector machine or a lightweight convolutional neural network. The classification model analyzes the input feature vector and outputs a type ID corresponding to a certain category in the predefined food library, such as dumplings, steak, pancakes, etc.

[0044] In another embodiment, type recognition can also not use machine learning models, but be completed by a rule-based decision tree system. For example, the controller 7 can determine the type ID of the cooking object according to a set of preset logical rules based on the calculated area, thickness and aspect ratio, etc. If the thickness < 3mm and the circularity > 0.9, it is determined as a pancake.

[0045] At the end of the process, the physical feature determination module organizes and constructs all the key parameters obtained by this calculation and recognition, including the determined type ID, the calculated two-dimensional size, the equivalent thickness and the initial temperature passed in by the initial state acquisition module, into a structured initial state vector. This vector is a complete digital description of the physical state of the cooking object before cooking starts, and it will be the final output of the physical feature determination module, passed to the subsequent heating curve generation module.

[0046] Referring to the accompanying Figure 5 , this process is performed by the heating curve generation module inside the controller 7, which receives the initial state vector output by the physical feature determination module as input.

[0047] The core of the heating profile generation module is to solve an optimal control problem based on a predictive process model. The predictive process model is in the form of a set of coupled ordinary differential equations, which describes the state evolution of the cooking object during the heating process. The model contains at least two parts: a heat conduction sub-model and a surface color change sub-model. The heat conduction sub-model is based on the heat conduction law, which describes the change of the internal and surface temperature of the cooking object over time through one or more state variables, whose rate of change is a function of the applied heating power and the current temperature distribution. In particular, the specific parameters of the heat conduction sub-model are determined according to the geometric parameters determined by the physical feature determination module, such as area and equivalent thickness The equivalent thickness directly determines the time required for heat to be transferred from the surface to the core, while the area affects the distribution of the total heating power. By substituting these real-time measured geometric parameters into the model, the model can more accurately reflect the physical reality of the current specific cooking object. The surface color change sub-model is based on chemical reaction kinetics, which describes the change of a quantitative surface color index, whose rate of change is a function of the surface temperature calculated by the heat conduction sub-model. Therefore, the two sub-models are coupled through the surface temperature.

[0048] The state of the entire system can be represented by a state vector , which contains at least the core temperature and the surface color index . The heating profile generation module converts the initial state vector received from the physical feature determination module into the initial condition of the model, while reading the target state vector corresponding to the type of cooking object from its internal memory, which defines the target core temperature and the target surface color index at the end of cooking. The goal of the heating profile generation module is to calculate a heating power profile , i.e. the control variable, which can drive the state of the cooking object from to in the shortest time . This problem is constructed as an optimal control problem with the objective function , and the constraint conditions include: System dynamic equation: ; where represents the aforementioned coupled predictive process model.

[0049] Initial state constraint: is equal to the obtained initial state vector.

[0050] Terminal state constraint: .

[0051] Control variable constraints: where is the rated maximum power of the heating assembly.

[0052] Since the goal of this problem is to solve for the required control input given the known terminal state, this process is referred to as inverse generation. The heating profile generation module employs a numerical optimization algorithm to solve this optimal control problem. Specifically, this module discretizes the continuous-time problem, converting it into a large-scale nonlinear programming (NLP) problem. This conversion process parameterizes the total time and the continuous control profile into a series of discrete decision variables. Subsequently, a processor within the controller 7 executes an iterative optimization solver to solve this NLP problem. The result of solving this NLP problem is a discrete time series containing the optimal total cooking duration and a series of heating power setpoints ordered in time over this duration. This sequence of power values constitutes the final generated initial heating profile. This profile is then output by the heating profile generation module to the closed-loop control module as a command sequence, to be used to guide the subsequent heating execution and monitoring process.

[0053] This process is executed by the closed-loop control module within the controller 7, which receives as input the initial heating profile generated by the heating profile generation module.

[0054] Upon receiving the initial heating profile, the closed-loop control module begins to apply control to the heating assembly disposed within the lid body 1 and base 2 by driving the circuitry according to the power values and time sequence defined by the profile. At the same time, the closed-loop control module initiates a real-time monitoring and adjustment cycle at a fixed time period.

[0055] At the beginning of each monitoring period, the closed-loop control module first instructs the image acquisition module to capture a frame of the current cooking object image. The closed-loop control module processes this real-time image to obtain a quantified actual surface color value . This process involves: locating one or more pre-determined surface regions as regions of interest (ROIs) in the image according to the previously determined cooking object location; computing the average RGB color value of all pixels within these ROIs; and finally, converting this average RGB value to a single scalar value representing the actual surface color at the current time instant via a pre-determined mapping function.

[0056] At the same time, the closed loop control module performs a prediction step. It calls the same predictive process model as used by the heating profile generation module, inputs the state at the previous time step and the heating power that has been applied , and computes the predicted state at the current time step . From this predicted state , the closed loop control module extracts the predicted surface color quantification value and a predicted cooking progress index increment .

[0057] Next, the closed loop control module performs a correction step to fuse the model prediction with the actual observation. It first computes the observation residual between the actual observation and the model prediction, i.e. . Then, the closed loop control module fuses the predicted cooking progress index with the observation residual, weighted by a correction factor, to compute the corrected, more accurate cooking progress index at the current time step .

[0058] After updating the cooking progress index , the closed loop control module compares it with the value of the theoretical planned progress trajectory corresponding to the initial heating power profile at the current time step , and computes the absolute deviation between the two . When this deviation exceeds a pre-set deviation threshold stored in the controller 7 memory, the control strategy is triggered for re-planning.

[0059] Upon triggering re-planning, the closed loop control module passes the complete state vector at the current time step, including the actual temperature and other state variables corresponding to , as the new initial condition to the heating profile generation module. The heating profile generation module, based on this new starting point, performs the complete optimal control problem solving process again to generate a new optimal heating power profile from the current state to the final target state. Upon receiving this new profile, the closed loop control module discards the old profile and starts to follow the new profile for subsequent heating control. If the deviation does not exceed the threshold, the current profile is continued to be followed for heating. This loop continues until the cooking termination module issues a termination instruction.

[0060] This judgment process is performed in parallel by the cooking termination module inside the controller 7, which continuously receives the corrected cooking progress index and the analyzed actual surface color quantification value from the closed loop control module.

[0061] The cooking termination module incorporates two independent condition monitors to simultaneously evaluate two termination conditions. The first condition monitor assesses the internal state, receiving real-time cooking progress indices. With a preset completion threshold stored in the non-volatile memory of controller 7 The comparison is performed. This threshold is a value close to 1.0, for example, 0.98. When When the condition is met, the first condition monitor sets one of its internal status flags to true.

[0062] The second condition monitor is responsible for evaluating the external appearance. It first quantizes the received actual surface color value. It is compared with a preset target color range. This target color range is read from a parameter library based on the type ID of the cooking object, and can be defined as a multi-dimensional color space, such as a region in the CIELAB color space. , , .when When the corresponding color coordinates fall within this area, the monitor starts an internal timer; if If the monitoring continues to fall within this area during subsequent monitoring cycles, the timer will continue counting. once Once removed from the area, the timer is immediately reset to zero. When the timer's count reaches or exceeds a preset stable duration, for example, 5 seconds, the second condition monitor sets one of its internal status flags to true. At the end of each control cycle, the cooking termination module performs a logical AND operation on the status flags of both monitors. If both status flags are false or only one is true, no operation is performed, and the heating process continues. The cooking termination module is triggered if and only if the status flags of both the first and second condition monitors are true simultaneously.

[0063] In addition, as a parallel safety measure, the cooking termination module continuously monitors the real-time temperature value fed back by the temperature sensor. Once this temperature value exceeds a preset safety upper limit threshold that is much higher than the normal cooking temperature, the module will be triggered immediately, regardless of whether the aforementioned two cooking completion conditions are met, to prevent overheating or dry burning due to unexpected circumstances.

[0064] After triggering, the cooking termination module immediately sends a clear stop instruction signal to the driving circuit of the heating assembly in the control upper cover body 1 and the base 2, for example, sets the control pin to low level. The instruction makes the driving circuit cut off the power supply to the heating assembly, thereby terminating the entire cooking process. At the same time, the cooking termination module can send a completion signal to the user interface module to drive the buzzer to sound or light up an indicator light to prompt the user that the cooking is complete.

[0065] Working principle: Based on food material recognition and dynamic temperature control, the electric baking pan is formed by closing the upper cover body 1 and the base 2 through the hinge 3 to form the cooking cavity 4, fixed by the first lock 6 and unlocked by the second lock 5, and the heating assembly is arranged in the upper cover body 1 and the base 2 respectively. At the same time, the image acquisition module covering the cooking cavity 4, the temperature sensor monitoring the temperature in the cavity, and the core controller are electrically connected with each component. The upper cover body 1 also has an auxiliary light source built-in; when in use, after closing the equipment, the image acquisition module first captures the initial image of the food material, and the temperature sensor synchronously obtains the initial temperature in the cavity. The controller uses image recognition algorithm, combines the food material shadow projected by the auxiliary light source to construct a triangular geometric model, calculates the equivalent thickness of the food material through the shadow length and the preset light incidence angle, determines the size, equivalent thickness and type ID of the food material and constructs the initial state vector, inputs the predictive process model containing the core temperature and surface color coupling dynamic equation, solves the optimal control problem with the minimum cooking time as the target and the target cooking state as the terminal constraint, reversely generates the initial heating power curve, and drives the upper and lower heating assemblies to heat according to the curve. During heating, the image acquisition module continuously obtains real-time images of the food material, the controller analyzes the image to monitor the surface color change, defines the cooking process index, fuses model prediction and actual color data to continuously correct the index, and when the corrected index deviates from the planned trajectory by more than the preset threshold, the heating power curve is immediately re-planned; until the index reaches the completion threshold, and the surface color of the food material enters the target range and stabilizes for a preset time length, the controller terminates heating, which not only solves the pain points of traditional electric baking pans relying on user experience, but also improves the cooking success rate and convenience, and realizes intelligent and precise cooking.

Claims

1. An electric griddle based on food ingredient recognition and dynamic temperature control, characterized in that, Includes a base (2) and an upper cover (1) rotatably connected to the base (2) via a hinge (3), wherein the upper cover (1) and the base (2) together form a cooking cavity (4) for placing cooking objects in the closed state. The electric griddle also includes a first latch (6) for connecting and fixing the upper cover (1) in the closed state to the base (2), and a second latch (5) for unlocking the first latch (6). Heating components are respectively provided inside the upper cover (1) and the base (2). The electric griddle also includes: The image acquisition module has a field of view that covers the cooking cavity; A temperature sensor is used to monitor the temperature inside the cooking cavity; The controller is electrically connected to the heating component, the image acquisition module, and the temperature sensor.

2. The electric griddle based on food ingredient recognition and dynamic temperature control according to claim 1, characterized in that, The electric griddle also includes an auxiliary light source, which is disposed in the cooking cavity (4) for casting the shadow of the cooking object in the cooking cavity (4), and the controller is used to determine the equivalent thickness of the cooking object.

3. A control method for an electric griddle based on food ingredient recognition and dynamic temperature control, applied to an electric griddle based on food ingredient recognition and dynamic temperature control as described in claims 1-2, characterized in that, Includes the following steps: a) After the upper cover (1) and the base (2) are closed by the hinge (3) to form the cooking cavity (4), the initial image data and initial temperature of a cooking object in the cooking cavity (4) are obtained by the image acquisition module and temperature sensor set inside the cooking cavity (4); b) Based on the initial image data, the controller uses an image recognition algorithm to determine at least one physical feature of the cooking object, the physical feature including size and equivalent thickness; c) Based on the physical characteristics and the initial temperature, the controller generates an initial heating power curve in reverse by solving a predictive process model with the target cooking state as the terminal constraint; d) The controller controls the heating components respectively installed in the upper cover (1) and the base (2) to perform heating according to the initial heating power curve, and the image acquisition module continuously acquires real-time image data of the cooking object during the cooking process; e) The controller analyzes the real-time image data in real time to monitor the surface color change of the cooking object, and when the actual process of the surface color change deviates from the process predicted by the predictive process model, it performs control strategy replanning on the heating power curve.

4. The control method for an electric griddle based on food ingredient recognition and dynamic temperature control according to claim 3, characterized in that, In step b), the method for determining the equivalent thickness of the cooking object includes: Using an auxiliary light source located in the cooking cavity (4), the shadow of the cooking object is projected onto the heating plate of the base (2), thereby constructing a triangular geometric model defined by the equivalent thickness of the cooking object, the shadow length, and the lighting path; The controller analyzes the initial image data, identifies and measures the feature length of the shadow, thereby obtaining a known side of the triangular geometric model; Based on the known side lengths and the preset incident angle of the auxiliary light source, which is a known angle in the triangular geometric model, the equivalent thickness of the cooking object is calculated according to the inherent trigonometric function relationships of the triangular geometric model.

5. The control method for an electric griddle based on food ingredient recognition and dynamic temperature control according to claim 3, characterized in that, In step c), the predictive process model includes coupled dynamic equations for describing the changes in core temperature and surface color of the cooked object with heating power and time. The process of reversely generating the initial heating power curve is to solve an optimal control problem with an objective function aimed at minimizing cooking time, and with the terminal state constraints being the target core temperature and target surface color of the object being cooked.

6. The control method for an electric griddle based on food ingredient recognition and dynamic temperature control according to claim 3, characterized in that, The controller also defines a cooking process index for quantifying the degree of completion of cooking of the objects in the cooking cavity (4); In step e), monitoring the actual process of the surface color change specifically involves continuously updating and correcting the cooking process index by fusing the predictions of the predictive process model with the actual surface color of the cooking object in the cooking cavity (4) based on the analysis of the real-time image data.

7. The control method for an electric griddle based on food ingredient recognition and dynamic temperature control according to claim 6, characterized in that, The steps for updating and correcting the cooking process index include: Based on the state at the previous moment, a predicted cooking progress index increment is calculated by the predictive process model. And calculate the observation residual between the actual surface color of the cooking object in the cooking cavity (4) as analyzed from the real-time image data and the surface color predicted by the predictive process model; The predicted cooking progress index increment is then weighted and fused with the observed residual to obtain the corrected cooking progress index at the current time.

8. The control method for an electric griddle based on food ingredient recognition and dynamic temperature control according to claim 6, characterized in that, In step e), when the deviation between the corrected cooking process index and the planned process trajectory corresponding to the initial heating power curve exceeds a preset threshold, the control strategy replanning for the heating components in the upper cover (1) and the base (2) is triggered.

9. The control method for an electric griddle based on food ingredient recognition and dynamic temperature control according to claim 3, characterized in that, When the cooking process index reaches the preset completion threshold, and the surface color of the cooking object in the cooking cavity (4) enters the target range and remains stable for a preset duration as monitored by the real-time image data, the heating of the heating components in the upper cover (1) and the base (2) is terminated.

10. The control method for an electric griddle based on food ingredient recognition and dynamic temperature control according to claim 3, characterized in that, In step b), the controller also constructs an initial state vector from the type ID, size, equivalent thickness and initial temperature of the cooking object in the cooking cavity (4), and uses the initial state vector as the input of the predictive process model.

Images