Control management system and method for automatic stir-frying mechanical equipment of Dongpo braised pork

Abstract

The invention relates to the technical field of stir-frying mechanical equipment control, in particular to a control management system and method for automatic stir-frying mechanical equipment of Dongpo braised pork. Comprising an image acquisition unit, a stir-frying parameter matching unit, a stir-frying prediction unit, a stir-frying optimization unit and a stir-frying extension unit, the image acquisition unit is used for establishing communication connection with stir-frying mechanical equipment so as to acquire preset stir-frying parameters and real-time images of the Dongpo braised pork in a pot in the stir-frying mechanical equipment and also acquire historical stir-frying parameters and historical images; the stir-frying duration is divided into 10 isochronous time periods through the stir-frying parameter matching unit, a parameter time period association rule is established in combination with historical optimal data, scientificity and reasonability of initial stir-frying parameters are ensured, stir-frying imbalance caused by parameter mutation is avoided through parameter change rate verification of adjacent time periods, and the stir-frying efficiency is improved. The defects that in the prior art, parameter setting is extensive, and segmented regulation and control logic is lacked are overcome.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of stir-frying mechanical equipment control, in particular to a control management system and method of an automatic stir-frying mechanical equipment for Dongpo pork. BACKGROUND

[0002] As a traditional classic dish, the stir-frying process of Dongpo pork has very high requirements for process standardization and taste consistency. With the development of the scale of the catering chain and the promotion of the industrialization of food processing, the automatic stir-frying mechanical equipment for Dongpo pork gradually replaces manual stir-frying and becomes an important technology application in the field of food processing.

[0003] The existing technology has fixed preset stir-frying parameters, which can only set a unified standard based on a single dish specification and cannot dynamically adjust according to the real-time state of the materials in the pot (such as material distribution, stacking height, and stir-frying progress difference), resulting in uneven stir-frying of Dongpo pork in different areas of the pot, excessive material stacking and insufficient stir-frying in some areas, and excessive material stir-frying and poor taste in some areas, which directly affects the taste consistency and appearance uniformity of the dish. In addition, there is a lack of precise monitoring and prediction mechanism for the stir-frying process, which relies only on preset time and fixed parameters to promote stir-frying, cannot identify stir-frying deviations in advance, and often finds that the dish is unqualified after stir-frying is completed, causing waste of raw materials and reduction of production efficiency. Therefore, a control management system and method of an automatic stir-frying mechanical equipment for Dongpo pork are proposed. SUMMARY

[0004] The present application aims to provide a control management system and method of an automatic stir-frying mechanical equipment for Dongpo pork to solve the problems raised in the background.

[0005] To achieve the above technical problems, one of the purposes of the present application is to provide a control management system of an automatic stir-frying mechanical equipment for Dongpo pork, which includes an image acquisition unit, a stir-frying parameter matching unit, a stir-frying prediction unit, a stir-frying optimization unit, and a stir-frying extension unit. The image acquisition unit is used to establish a communication connection with the stir-frying mechanical equipment to acquire the preset stir-frying parameters and real-time images of Dongpo pork in the pot of the stir-frying mechanical equipment, as well as historical stir-frying parameters and historical images. The stir-frying parameter matching unit is used to set the stir-frying time and divide the stir-frying time into multiple stir-frying periods, and match the stir-frying periods with the preset stir-frying parameters to obtain the preset stir-frying parameters corresponding to each stir-frying period. The stir-frying prediction unit is used to screen a standard image set of Dongpo pork from the historical images, calculate an image difference value by combining the standard image set with the real-time image, and perform stir-frying completion prediction by combining the image difference value with the historical stir-frying process, the remaining stir-frying period, and the preset stir-frying parameters, and trigger the stir-frying optimization unit when the prediction shows that the stir-frying cannot be completed. The frying optimization unit is configured to determine a region to be optimized and an influence region according to the image difference value, perform directional frying optimization on the preset frying parameters according to the region to be optimized and the influence region, and substitute the optimized preset frying parameters into the corresponding frying period; The frying extension unit is configured to combine the real-time image corresponding to each frying period after optimization with the image difference value to perform frying completion prediction twice, and dynamically extend the frying time length when the optimization still fails to complete the frying.

[0006] As a further improvement of the technical solution, the image acquisition unit is configured to establish a communication connection with the frying mechanical equipment, so as to acquire the preset frying parameters and real-time images and historical frying parameters and historical images of the Dongpo pork in the pot in the frying mechanical equipment; The frying mechanical equipment is provided with a camera device and a frying device, and the camera device is configured to acquire the images of the Dongpo pork in the pot; The frying parameters include a frying frequency and a frying direction.

[0007] As a further improvement of the technical solution, the frying parameter matching unit includes a frying time length setting module and a parameter matching module; The frying time length setting module is configured to set a frying time length range of the Dongpo pork frying process, and select the lowest time length in the frying time length range to set the frying time length of the Dongpo pork this time; The parameter matching module is configured to divide the frying time length into multiple frying periods, and the number of frying periods can be initially set to ten, and the time lengths of the frying periods are the same; The frying periods are matched with the preset frying parameters, so that the preset frying parameters are segmented and divided into the frying periods, and the preset frying parameters corresponding to each frying period are acquired.

[0008] As a further improvement of the technical solution, the frying prediction unit includes a standard image screening module and an image difference calculation module, and a frying prediction module; The standard image screening module is configured to screen the images of dishes that have completed frying in the historical images, and then aggregate the images of dishes that have completed frying to generate a standard image set; The image difference calculation module is configured to calculate the image difference value by combining the standard image set with the real-time image after completing a frying period, and select the most similar standard image and the real-time image to calculate, so as to acquire the image difference value from the completed frying; The frying prediction module is configured to establish a frying prediction model by combining the completed frying process of this frying and the historical frying process by using an LSTM neural network, and then input the remaining frying period and the preset frying parameters into the frying prediction model to perform frying completion prediction; When the prediction indicates that stir-frying cannot be completed, the stir-frying optimization unit is triggered. Continue monitoring if the forecast indicates that the cooking process can be completed.

[0009] As a further improvement to this technical solution, in the stir-frying prediction module, the completed stir-frying process in this stir-frying is the completed stir-frying period, and its corresponding preset stir-frying parameters and real-time images; The historical frying process consists of historical frying parameters and historical images.

[0010] As a further improvement to this technical solution, the stir-frying optimization unit includes a region determination module and a parameter optimization module; The region determination module is used to receive the trigger signal from the stir-fry prediction module, extract the image difference value corresponding to the real-time image data of the latest time node, divide the dish area in the wok of the stir-frying machine, obtain multiple dish areas, and then determine the difference value of each dish area in the stir-frying machine based on the image difference value. Based on the completed cooking time, a deviation threshold is set, and then the difference values ​​of each dish area are compared with the deviation threshold. If the difference value of a dish area exceeds the deviation threshold, the area is determined to be an area to be optimized. If the difference value of a dish area is less than the deviation threshold, the area is determined to be the affected area. The parameter optimization module is used to perform targeted stir-fry optimization on the preset stir-fry parameters according to the area to be optimized and the area affected, summarize the difference values ​​between the area to be optimized and the area affected to set an average difference value, and then analyze the dish position adjustment parameters according to the difference between the area affected and the area to be optimized and the average deviation value to obtain the dish position adjustment parameters. Then, the preset stir-fry parameters are optimized according to the dish position adjustment parameters to obtain the optimized preset stir-fry parameters. The optimized preset stir-frying parameters are used to stir-fry the dishes in the area to be optimized towards the affected area, so that the difference values ​​of each dish area conform to the average difference value.

[0011] As a further improvement to this technical solution, the region determination module simultaneously sets a stacking height threshold and analyzes the stacking height of each dish region by combining real-time images with the cooking machinery to obtain the stacking height of each dish region. For dish regions whose stacking height exceeds the stacking height threshold, they are determined to be regions to be optimized. Meanwhile, a height average stir-fry parameter was added to the parameter optimization module to limit the stacking height of each dish area to below the stacking height threshold.

[0012] As a further improvement to this technical solution, in the extended frying unit, the real-time images corresponding to each frying period after optimization are combined with the image difference values ​​to predict the completion of frying for the second time. After the optimized frying period is completed, the real-time images and optimized frying parameters are input into the frying prediction model to predict the completion of frying. If the prediction results show that the optimized frying parameters are still insufficient to complete the frying process, the frying time will be dynamically extended based on the image difference value; however, the extension of the frying time will not exceed the frying time range. If the prediction results show that the optimized frying parameters can complete the frying process, continue monitoring.

[0013] The second objective of this invention is to provide a control and management method for an automated stir-frying machine for Dongpo braised pork, based on any one of the above-mentioned control and management systems for the automated stir-frying machine for Dongpo braised pork, comprising the following steps: S1. Establish a communication connection with the stir-frying machinery and equipment to obtain the preset stir-frying parameters and real-time images of Dongpo braised pork in the pot of the stir-frying machinery and equipment, and at the same time obtain the historical stir-frying parameters and historical images. S2. Set the stir-frying time and divide the stir-frying time into multiple stir-frying periods. At the same time, match the stir-frying periods with preset stir-frying parameters to obtain the preset stir-frying parameters corresponding to each stir-frying period. S3. Select a standard image set of Dongpo braised pork from historical images, and combine the standard image set with real-time images to calculate the image difference value. At the same time, combine the image difference value with the historical frying process, the remaining frying time and preset frying parameters to predict the completion of frying. When the prediction shows that frying cannot be completed, trigger S4. S4. Determine the region to be optimized and the region affected based on the image difference value. Optimize the preset frying parameters in a targeted manner based on the region to be optimized and the region affected. Substitute the optimized preset frying parameters into the corresponding frying time period. S5. Combine the optimized real-time images corresponding to each frying period with the image difference values ​​to predict the completion of frying. If the frying cannot be completed after optimization, the frying time will be dynamically extended.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. A control management system and method for an automated stir-frying machine for Dongpo braised pork belly, which divides the stir-frying time into 10 equal time periods through a stir-frying parameter matching unit, establishes parameter time period association rules based on historical best data to ensure the scientificity and rationality of the initial stir-frying parameters, and avoids stir-frying imbalance caused by parameter mutations by verifying the parameter change rate of adjacent time periods. This solves the defects of the existing technology, which has coarse parameter setting and lacks segmented control logic. Secondly, an LSTM neural network is introduced to construct a stir-frying prediction model, which integrates multi-dimensional features such as the completed stir-frying process, historical data, and parameters of the remaining time period to predict the stir-frying results in advance, realizing early warning of stir-frying deviations. This effectively avoids the generation of unqualified products due to lack of prediction in the existing technology, and reduces raw material waste and production losses.

[0015] 2. A control management system and method for an automated stir-frying machine for Dongpo braised pork belly, which determines the area to be optimized and the area affected by the machine by using a dual threshold of difference value and stacking height, and calculates and adjusts the stir-frying parameters in a targeted manner by combining the average difference value, thereby achieving precise transfer and uniform distribution of materials. This fundamentally solves the core problems of uneven stir-frying and material stacking in the existing technology. At the same time, by limiting the stacking height of each area by using height average stir-frying parameters, the physical form and stir-frying uniformity of the dish are further guaranteed, ensuring that the stir-frying conditions of each piece of Dongpo braised pork belly are consistent, and improving the uniformity of the dish's taste and appearance.

[0016] 3. A control management system and method for an automated stir-frying machine for Dongpo braised pork belly. When the stir-frying extension unit still cannot meet the stir-frying requirements after parameter optimization, it dynamically calculates the extension time based on the current difference value and the remaining time, and strictly controls it within the time range allowed by the process. This ensures the quality of the dish and avoids the decline in taste caused by excessive extension, achieving a balance between efficiency and quality. Through a closed-loop control logic of data acquisition, parameter matching, prediction, early warning, targeted optimization, and fallback extension, combined with historical data self-learning and real-time dynamic adjustment, the standardization level and quality stability of automated stir-frying of Dongpo braised pork belly are significantly improved, while also taking into account production efficiency. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall structure of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] like Figure 1As shown, one of the objectives of this invention is to provide a control and management system for an automated stir-frying machine for Dongpo braised pork, including an image acquisition unit, a stir-frying parameter matching unit, a stir-frying prediction unit, a stir-frying optimization unit, and a stir-frying extension unit. The image acquisition unit is used to establish a communication connection with the stir-frying machinery and equipment, thereby acquiring the preset stir-frying parameters and real-time images of Dongpo braised pork in the pot of the stir-frying machinery and equipment, and at the same time acquiring historical stir-frying parameters and historical images. The image acquisition unit is used to establish a communication connection with the stir-frying machinery and equipment, thereby acquiring the preset stir-frying parameters and real-time images of Dongpo braised pork in the wok of the stir-frying machinery and equipment, as well as the historical stir-frying parameters and historical images. The communication interface of the frying machinery is standardized and adapted, and the built-in Ethernet interface (wired communication) of the equipment is enabled first. At the same time, the built-in 5G communication module (wireless redundancy) of the equipment is activated to ensure seamless switching between the two communication methods. Then, the image acquisition unit sends a connection request to the frying machinery. The request information includes the equipment's unique identifier (ID) and the communication protocol version (using Modbus TCP protocol). After receiving the request, the machinery verifies the identity. If the verification is successful, it returns a connection success response, and the communication link is established. The stir-frying equipment is equipped with a camera and a stir-frying device, which captures images of the Dongpo braised pork inside the pot through the camera. The stir-frying parameters include the frequency and direction of stirring.

[0020] Simultaneously acquire the preset stir-frying parameters (stir-frying frequency, stir-frying direction) and real-time images of Dongpo braised pork in the pot; among them, the image data is acquired by a high-definition camera device equipped with mechanical equipment (installed directly above the pot opening, with a field of view covering the entire pot), and then transmitted to the image acquisition unit after compression encoding (H.265 encoding).

[0021] The stir-frying parameter matching unit is used to set the stir-frying time and divide the stir-frying time into multiple stir-frying periods. At the same time, it matches the stir-frying periods with preset stir-frying parameters to obtain the preset stir-frying parameters corresponding to each stir-frying period. The stir-frying parameter matching unit includes a stir-frying time setting module and a parameter matching module; The stir-frying time setting module is used to set the stir-frying time range for Dongpo braised pork, and select the minimum time within the stir-frying time range to set the stir-frying time for this Dongpo braised pork. Retrieve qualified cooking records of Dongpo braised pork of the same specifications (consistent number of pieces, weight per piece, and lean-fat ratio) from the historical database, extract the actual cooking time of all records, determine the lower limit (shortest qualified time) and upper limit (longest qualified time) of the cooking time, and form the cooking time range for this cooking. The lower limit of the time range was selected as the target cooking time for this Dongpo braised pork, so as to balance cooking efficiency and product quality stability. The parameter matching module is used to divide the frying time into frying periods, dividing the frying time into multiple frying periods (10 equal-length frying periods). The number of frying periods can be initially set to ten, and each frying period has the same length of time, which facilitates subsequent segmentation parameter adjustment. The stir-frying time period is matched with the preset stir-frying parameters, thereby dividing the preset stir-frying parameters into segments according to the stir-frying time period, and obtaining the preset stir-frying parameters corresponding to each stir-frying time period.

[0022] Extract the preset stir-frying parameters (stir-frying frequency, stir-frying direction) corresponding to the best stir-frying record with the duration limit in the historical database, and establish the parameter time period association rules according to the temporal characteristics of the stir-frying process stages (initial ingredient distribution → uniform stir-frying → enhanced flavor absorption); Based on the established association rules, the preset stir-frying parameters are matched one by one to the 10 stir-frying time periods, completing the segmentation of the preset parameters and forming a precise set of stir-frying parameters corresponding to each time period. Then, the continuity and rationality of the parameters in each time period are verified (the rate of change of parameters between adjacent time periods does not exceed the preset threshold). If the verification passes, the final segmented parameters are determined; if the verification fails, the parameters from the second-best historical records are retrieved again for matching until the requirements are met. The formula is as follows: ; in, The preset parameter value (stirring frequency) for the i-th time period. This is the preset parameter value for the (i+1)th time period. The maximum allowable rate of change is preset (valued at 15% to ensure smooth parameter transition). This represents the actual rate of change.

[0023] The stir-frying prediction unit is used to filter a standard image set of Dongpo braised pork from historical images, and calculate the image difference value by combining the standard image set with real-time images. At the same time, the image difference value is combined with the historical stir-frying process, the remaining stir-frying time and preset stir-frying parameters to predict the completion of stir-frying. When the prediction shows that stir-frying cannot be completed, the stir-frying optimization unit is triggered. The stir-fry prediction unit includes a standard image screening module, an image difference calculation module, and a stir-fry prediction module; The standard image filtering module is used to filter images of dishes that have been cooked in the historical image library, and then summarize the images of cooked dishes to generate a standard image set. The selection criteria for images of stir-fried dishes are as follows: the image features meet the standards (color RGB values ​​are within the preset acceptable range, texture uniformity is ≥90%, and material stacking height is ≤2cm) and the corresponding stir-frying parameters are complete and compliant (no abnormal parameter records). Then, all images of Dongpo braised pork marked as stir-fried are retrieved from the historical database, and each image is verified according to the above selection criteria. Images with abnormal image features or incomplete parameter records are removed, and qualified images that meet the conditions are retained. These images are then compiled into a standardized set of standard images and stored in the system's local cache for easy retrieval later. The image difference calculation module is used to calculate the image difference value by combining the standard image set with the real-time image after a stir-frying period is completed. It selects the most similar standard image and the real-time image for calculation to obtain the image difference value from the time when the stir-frying is completed. After a stir-frying period is completed, real-time images of the Dongpo braised pork belly in the pot are captured by a camera device. The same preprocessing process as the standard image set is used (size normalized to 1920×1080 pixels, Gaussian filtering for noise reduction) to ensure the comparability of the real-time images with the standard images. At the same time, the core features (color mean, texture variance, material distribution entropy) of all images in the standard image set are extracted. Meanwhile, the corresponding features of the preprocessed real-time images are extracted. By calculating the feature similarity between the real-time images and each standard image, the standard image with the highest similarity is selected as the reference image. Based on the selected most similar standard image and the real-time image, the structural similarity (SSIM) algorithm is used to calculate the difference between the two. This value directly reflects the gap between the real-time image and the completed cooking state, that is, the difference value from the image of completed cooking. The formula is as follows: ; in, This represents the feature similarity (values ​​range from 0 to 1, with higher similarity values ​​closer to 1). , , , The RGB mean and texture variance of the real-time image. , , , These are the corresponding features of the standard image. = = = =0.25 (feature weights are equal to ensure comprehensive evaluation), selected The standard image corresponding to the maximum value is used as the most similar reference image; ;

[0024] ; in, For real-time images, For the most similar standard image, , Let x and y be the mean gray values, respectively. , These are the grayscale variances, For grayscale covariance, For structural similarity (the closer to 1, the more similar). The image difference value (ranging from 0 to 1, with values ​​closer to 0 indicating closer to completion of the stir-frying process). (Preset constant) =0.01, =0.03, =255 represents the grayscale range of the image); The stir-frying prediction module uses an LSTM neural network to build a stir-frying prediction model by combining the completed stir-frying process and the historical stir-frying process. Then, the remaining stir-frying time and preset stir-frying parameters are input into the stir-frying prediction model to predict the completion of stir-frying. When the prediction indicates that stir-frying cannot be completed, the stir-frying optimization unit is triggered. Continue monitoring if the forecast indicates that the cooking process can be completed.

[0025] In the stir-frying prediction module, the completed stir-frying process for this stir-frying is the completed stir-frying period, along with its corresponding preset stir-frying parameters and real-time images; The historical trading process consists of historical trading parameters and historical images, and the steps are as follows: Two types of core data are summarized: completed data of this stir-frying (completed stir-frying time period number, preset stir-frying parameters for the corresponding time period, and real-time image features) and historical stir-frying data (all-time parameters and image features of historical qualified / unqualified stir-frying records). At the same time, high-dimensional features (such as color histograms and texture feature vectors) are extracted from all images (real-time images and historical images) using convolutional neural networks (CNNs), transforming 2D images into 1D feature vectors to facilitate time series model processing. The stir-frying parameters (frequency, direction) and image feature vectors are normalized by minmax to eliminate the difference in dimensions and uniformly map them to the [0,1] interval. Then, taking the stir-frying period as the time step, the completed data and historical data are organized into time series samples according to the time period order. Each sample contains the parameter + image feature input sequence of the first N time periods, as well as a label indicating whether the stir-frying is qualified (1 = can be completed, 0 = cannot be completed). The input layer has a dimension of time sequence length (i.e., number of completed time periods) × feature dimension (stirring parameter dimension + image feature vector dimension). The hidden layer consists of two LSTM layers (to enhance temporal feature extraction capabilities), with 64 neurons in the first layer and 32 neurons in the second layer. Both layers use the ReLU activation function to alleviate the gradient vanishing problem. The fully connected layer maps the temporal features output by the LSTM layer into a 1-dimensional vector and connects it to a fully connected layer with 16 neurons. The output layer uses the Sigmoid activation function to output a single probability value (0~1), which corresponds to the confidence level of being able to complete the stir-frying process. Then, model training is performed. First, the dataset is divided into a training set (for model training) and a validation set (for performance evaluation) in a 7:3 ratio. Then, the Adam optimizer is used with the cross-entropy loss function as the optimization objective, the learning rate is set to 0.001, the batch size is 32, and the number of iterations is 50. An early stopping mechanism is introduced: training is stopped when the validation set loss does not decrease for 5 consecutive rounds to avoid overfitting. Finally, the model accuracy is calculated using the validation set. The accuracy must be ≥95%. Otherwise, the number of hidden layer neurons is adjusted and retraining is performed until the target is met.

[0026] The completed frying time sequence, the number of remaining frying time periods, and the preset frying parameters for the remaining time periods are combined into a prediction input sequence. After preprocessing, the sequence is input into a trained LSTM model, and the model outputs a confidence probability value that the frying can be completed. The confidence threshold is set to 90%. If the confidence level is ≥90%, it is determined that the stir-frying can be completed and monitoring continues. If the confidence level is <10% (i.e., it cannot be completed with a confidence level ≥90%), it is determined that the stir-frying cannot be completed and the stir-frying optimization unit is triggered. In this process, after each new cooking period is completed, the parameters and image features of that period are added to the time series, re-inputted into the model for real-time prediction, and the prediction results are dynamically updated to ensure the continuity of monitoring.

[0027] The stir-frying optimization unit is used to determine the region to be optimized and the affected region based on the image difference value, to perform targeted stir-frying optimization on the preset stir-frying parameters based on the region to be optimized and the affected region, and to substitute the optimized preset stir-frying parameters into the corresponding stir-frying time period; The stir-frying optimization unit includes a region determination module and a parameter optimization module; The region determination module is used to receive the trigger signal from the stir-fry prediction module, extract the image difference value corresponding to the real-time image data of the latest time node, divide the dish area in the wok of the stir-frying machine, obtain multiple dish areas, and then determine the difference value of each dish area in the stir-frying machine based on the image difference value. The system receives the trigger signal indicating that the stir-frying cannot be completed from the stir-frying prediction module, and simultaneously extracts the real-time image data of the latest time node and the corresponding global image difference value (i.e., the difference value between the distance to the completion of stir-frying calculated by the image difference calculation module). Then, it divides the area inside the pot in the real-time image into an equal-area grid (6×6 grid, a total of 36 dish areas) to ensure that each area has a uniform coverage and that no position inside the pot is missed, which facilitates the accurate location of the imbalance area. Based on the global image difference value calculation logic, the image features (color, texture, material distribution) of each dish area are extracted and compared with the corresponding area features of the most similar image in the standard image set. The SSIM algorithm is used to calculate the local difference value of each dish area. The threshold for deviation is set based on the completed cooking period. The fewer the completed cooking period, the more lenient the threshold (allowing for larger deviations in the early stages). The more the completed cooking period, the stricter the threshold (precise control is required near the end point) to ensure that the threshold is adapted to the characteristics of the cooking stage. Then, the differences between the various dish regions are compared with the deviation threshold. When the difference value of a dish area exceeds the deviation threshold, the area is determined to be an area to be optimized (uneven stir-frying, large gap from the finished state). If the difference value of a dish area is less than the deviation threshold, the area is determined to be an affected area (the stir-frying condition is qualified and the dish can be accepted). The parameter optimization module is used to perform targeted stir-fry optimization on the preset stir-fry parameters based on the area to be optimized and the area affected. It summarizes the difference values ​​between the area to be optimized and the area affected and sets an average difference value. Then, it analyzes the dish position adjustment parameters based on the difference between the area affected and the area to be optimized and the average deviation value, obtains the dish position adjustment parameters, and then optimizes the preset stir-fry parameters based on the dish position adjustment parameters to obtain the optimized preset stir-fry parameters. The optimized preset stir-frying parameters are used to stir-fry the dishes in the area to be optimized towards the affected area, so that the difference values ​​of each dish area conform to the average difference value.

[0028] Summarize the local difference values ​​of all areas to be optimized and affected areas, calculate the average difference value as the target convergence standard for the difference values ​​of each area, and then calculate the deviation of each area to be optimized from the average difference value and the deviation of each affected area from the average difference value. Combined with the location coordinates of the area in the pot, analyze the material transfer requirements and determine the position adjustment parameters (including the adjustment amount of the stirring frequency and the offset amount of the stirring direction). By adjusting the stirring direction (pointing to the adjacent affected area) and stirring intensity of the corresponding area, it is ensured that the accumulation height gradually decreases to below the accumulation height threshold when the material is transferred to the affected area; at the same time, for other areas to be optimized, the stirring frequency and direction are optimized according to the location, so as to promote the convergence of their difference value to the average difference value, as shown in the following formula; ; in, This represents the adjustment amount of the stir-frying frequency corresponding to region i to be optimized (a positive value indicates an increase in frequency). The original preset stirring frequency for this area, This is the weighting coefficient for stacking height (optimal value is 1.2, prioritizing stacking height optimization). The difference in the region to be optimized. For the deviation in stacking height, For the number of regions; ; in, Adjust the angle of the stirring direction for region i to be optimized. Let i be the center coordinates of the region to be optimized. The coordinates are the center coordinates of the nearest influence region j to this region; ; in, The frequency of stirring after achieving a high degree of uniformity. The optimized stir-frying frequency is calculated based on the difference value. The stacking height threshold, This represents the actual stacking height.

[0029] In the area determination module, a stacking height threshold is set (the optimal value is 3cm, which can be increased as the amount of vegetables increases). The stacking height of each vegetable area is analyzed by combining real-time images with the cooking machinery to obtain the stacking height of each vegetable area. Vegetable areas whose stacking height exceeds the stacking height threshold are identified as areas to be optimized. By analyzing the pixel depth of real-time images (combined with the known dimensions of the pot for calibration), the actual stacking height of each food area is calculated. When the actual stacking height exceeds the stacking height threshold, regardless of whether the difference value of the area meets the standard, it is further identified as an area to be optimized (prioritizing the resolution of physical stacking imbalance). Meanwhile, a height average stir-fry parameter was added to the parameter optimization module to limit the stacking height of each dish area to below the stacking height threshold.

[0030] The stir-frying extension unit is used to combine the real-time images corresponding to each optimized stir-frying period with the image difference value to predict the completion of stir-frying. If the completion is still not possible after optimization, the stir-frying time is dynamically extended.

[0031] In the extended frying unit, the real-time images corresponding to each frying period after optimization are combined with the image difference values ​​to predict the completion of frying for the second time. After the optimized frying period is completed, the real-time images and optimized frying parameters are input into the frying prediction model to predict the completion of frying. If the prediction results show that the optimized cooking parameters are still insufficient to complete the cooking process, the cooking time is dynamically extended based on the image difference value. The extension time does not exceed the normal cooking time range. The extended time is then evenly distributed across the remaining unexecuted cooking periods (keeping the duration consistent across periods), and the cooking parameters for each extended period are adapted according to the original parameter matching rules, as shown in the following formula: ; in, The duration is to be extended. The latest image difference value. The target difference value (takes a value of 0.1). This is the duration adjustment factor (calibrated based on historical extension effects, default 1.0). The remaining cooking time (total remaining time before extension); The larger the difference and the less time remaining, the more time is extended to ensure targeted reduction of the gap; If the prediction results show that the optimized frying parameters can complete the frying process, continue monitoring.

[0032] The second objective of this invention is to provide a control and management method for an automated stir-frying machine for Dongpo braised pork. The control and management system for the automated stir-frying machine for Dongpo braised pork, based on any one of the above-mentioned methods, includes the following steps: S1. Establish a communication connection with the stir-frying machinery and equipment to obtain the preset stir-frying parameters and real-time images of Dongpo braised pork in the pot of the stir-frying machinery and equipment, and at the same time obtain the historical stir-frying parameters and historical images. S2. Set the stir-frying time and divide the stir-frying time into multiple stir-frying periods. At the same time, match the stir-frying periods with preset stir-frying parameters to obtain the preset stir-frying parameters corresponding to each stir-frying period. S3. Select a standard image set of Dongpo braised pork from historical images, and combine the standard image set with real-time images to calculate the image difference value. At the same time, combine the image difference value with the historical frying process, the remaining frying time and preset frying parameters to predict the completion of frying. When the prediction shows that frying cannot be completed, trigger S4. S4. Determine the region to be optimized and the region affected based on the image difference value. Optimize the preset frying parameters in a targeted manner based on the region to be optimized and the region affected. Substitute the optimized preset frying parameters into the corresponding frying time period. S5. Combine the optimized real-time images corresponding to each frying period with the image difference values ​​to predict the completion of frying. If the frying cannot be completed after optimization, the frying time will be dynamically extended.

[0033] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A control and management system for an automated stir-frying machine for Dongpo braised pork, characterized in that: It includes an image acquisition unit, a cooking parameter matching unit, a cooking prediction unit, a cooking optimization unit, and a cooking extension unit; The image acquisition unit is used to establish a communication connection with the stir-frying machinery and equipment, thereby acquiring the preset stir-frying parameters and real-time images of Dongpo braised pork in the pot of the stir-frying machinery and equipment, and at the same time acquiring historical stir-frying parameters and historical images. The stir-frying parameter matching unit is used to set the stir-frying time and divide the stir-frying time into multiple stir-frying periods. At the same time, it matches the stir-frying periods with preset stir-frying parameters to obtain the preset stir-frying parameters corresponding to each stir-frying period. The stir-frying prediction unit is used to filter a standard image set of Dongpo braised pork from historical images, and calculate the image difference value by combining the standard image set with real-time images. At the same time, the image difference value is combined with the historical stir-frying process, the remaining stir-frying time and preset stir-frying parameters to predict the completion of stir-frying. When the prediction shows that stir-frying cannot be completed, the stir-frying optimization unit is triggered. The stir-frying optimization unit is used to determine the region to be optimized and the region affected based on the image difference value, to perform targeted stir-frying optimization on the preset stir-frying parameters based on the region to be optimized and the region affected, and to substitute the optimized preset stir-frying parameters into the corresponding stir-frying time period; The stir-frying extension unit is used to combine the real-time images corresponding to each optimized stir-frying period with the image difference value to predict the completion of stir-frying for the second time. If the stir-frying cannot be completed after optimization, the stir-frying time is dynamically extended.

2. The control and management system for the automated stir-frying machinery for Dongpo braised pork according to claim 1, characterized in that: The image acquisition unit is used to establish a communication connection with the stir-frying machinery and equipment, thereby acquiring the preset stir-frying parameters and real-time images of Dongpo braised pork in the pot of the stir-frying machinery and equipment, as well as the historical stir-frying parameters and historical images. The stir-frying equipment is equipped with a camera and a stir-frying device, which captures images of the Dongpo braised pork inside the pot through the camera. The stir-frying parameters include the frequency and direction of stirring.

3. The control and management system for the automated stir-frying machinery for Dongpo braised pork according to claim 1, characterized in that: The stir-frying parameter matching unit includes a stir-frying time setting module and a parameter matching module; The stir-frying time setting module is used to set the stir-frying time range of Dongpo braised pork, and select the lowest time within the stir-frying time range to set the stir-frying time of this Dongpo braised pork. The parameter matching module is used to divide the frying time into frying periods, and divide the frying time into multiple frying periods. The number of frying periods can be initially set to ten, and the duration of each frying period is the same. The stir-frying time period is matched with the preset stir-frying parameters, thereby dividing the preset stir-frying parameters into segments according to the stir-frying time period, and obtaining the preset stir-frying parameters corresponding to each stir-frying time period.

4. The control and management system for the automated stir-frying machinery for Dongpo braised pork according to claim 1, characterized in that: The stir-frying prediction unit includes a standard image screening module, an image difference calculation module, and a stir-frying prediction module; The standard image filtering module is used to filter images of dishes that have been cooked in the historical images, and then summarize the images of dishes that have been cooked in the historical images to generate a standard image set. The image difference calculation module is used to calculate the image difference value by combining the standard image set with the real-time image after a stir-frying period is completed. It selects the most similar standard image and the real-time image for calculation to obtain the image difference value at the time of completion of stir-frying. The stir-frying prediction module is used to establish a stir-frying prediction model by using an LSTM neural network, combining the completed stir-frying process and the historical stir-frying process, and then inputting the remaining stir-frying time and preset stir-frying parameters into the stir-frying prediction model to predict the completion of stir-frying. When the prediction indicates that stir-frying cannot be completed, the stir-frying optimization unit is triggered. Continue monitoring if the forecast indicates that the cooking process can be completed.

5. The control and management system for the automated stir-frying machinery for Dongpo braised pork according to claim 4, characterized in that: In the stir-frying prediction module, the completed stir-frying process is the completed stir-frying period, along with its corresponding preset stir-frying parameters and real-time images; The historical frying process consists of historical frying parameters and historical images.

6. The control and management system for the automated stir-frying machinery for Dongpo braised pork according to claim 4, characterized in that: The stir-frying optimization unit includes a region determination module and a parameter optimization module; The region determination module is used to receive the trigger signal from the stir-fry prediction module, extract the image difference value corresponding to the real-time image data of the latest time node, divide the dish area in the wok of the stir-frying machine, obtain multiple dish areas, and then determine the difference value of each dish area in the stir-frying machine based on the image difference value. Based on the completed cooking time, a deviation threshold is set, and then the difference values ​​of each dish area are compared with the deviation threshold. If the difference value of a dish area exceeds the deviation threshold, the area is determined to be an area to be optimized. If the difference value of a dish area is less than the deviation threshold, the area is determined to be the affected area. The parameter optimization module is used to perform targeted stir-fry optimization on the preset stir-fry parameters according to the area to be optimized and the area affected, summarize the difference values ​​between the area to be optimized and the area affected to set an average difference value, and then analyze the dish position adjustment parameters according to the difference between the area affected and the area to be optimized and the average deviation value to obtain the dish position adjustment parameters. Then, the preset stir-fry parameters are optimized according to the dish position adjustment parameters to obtain the optimized preset stir-fry parameters. The optimized preset stir-frying parameters are used to stir-fry the dishes in the area to be optimized towards the affected area, so that the difference values ​​of each dish area conform to the average difference value.

7. The control and management system for the automated stir-frying machinery for Dongpo braised pork according to claim 6, characterized in that: In the area determination module, a stacking height threshold is set, and the stacking height of each dish area is analyzed by combining real-time images with the cooking machinery to obtain the stacking height of each dish area. For dish areas whose stacking height exceeds the stacking height threshold, they are determined to be areas to be optimized. Meanwhile, a height average stir-fry parameter was added to the parameter optimization module to limit the stacking height of each dish area to below the stacking height threshold.

8. The control and management system for the automated stir-frying machinery for Dongpo braised pork according to claim 1, characterized in that: In the extended frying unit, the real-time images corresponding to each frying period after optimization are combined with the image difference values ​​to predict the completion of frying for the second time. After the optimized frying period is completed, the real-time images and optimized frying parameters are input into the frying prediction model to predict the completion of frying. If the prediction results show that the optimized frying parameters are still insufficient to complete the frying process, the frying time will be dynamically extended based on the image difference value; however, the extension of the frying time will not exceed the frying time range. If the prediction results show that the optimized frying parameters can complete the frying process, continue monitoring.

9. A control and management method for an automated stir-frying machine for Dongpo braised pork, based on the control and management system for the automated stir-frying machine for Dongpo braised pork as described in any one of claims 1-8, characterized in that: Includes the following steps: S1. Establish a communication connection with the stir-frying machinery and equipment to obtain the preset stir-frying parameters and real-time images of Dongpo braised pork in the pot of the stir-frying machinery and equipment, and at the same time obtain the historical stir-frying parameters and historical images. S2. Set the stir-frying time and divide the stir-frying time into multiple stir-frying periods. At the same time, match the stir-frying periods with preset stir-frying parameters to obtain the preset stir-frying parameters corresponding to each stir-frying period. S3. Select a standard image set of Dongpo braised pork from historical images, and combine the standard image set with real-time images to calculate the image difference value. At the same time, combine the image difference value with the historical frying process, the remaining frying time and preset frying parameters to predict the completion of frying. When the prediction shows that frying cannot be completed, trigger S4. S4. Determine the region to be optimized and the region affected based on the image difference value. Optimize the preset frying parameters in a targeted manner based on the region to be optimized and the region affected. Substitute the optimized preset frying parameters into the corresponding frying time period. S5. Combine the optimized real-time images corresponding to each frying period with the image difference values ​​to predict the completion of frying. If the frying cannot be completed after optimization, the frying time will be dynamically extended.

Images