Method, system and device for reducing food vacuum cooling loss by moisture intelligent compensation

Abstract

The application provides a moisture intelligent compensation method, system and device for reducing loss of food vacuum cooling, and relates to the technical field of intelligent food processing. The method of the application is applied to a vacuum cooling system (an internal vacuum chamber for placing food to be cooled is arranged). The method of the application comprises the following steps: obtaining a temperature field, a three-dimensional profile, a mass, humidity in the vacuum chamber and pressure in the vacuum chamber of the food; inputting the temperature field, the three-dimensional profile, the mass, the humidity and the pressure into a pre-trained water content prediction model to obtain water contents of multiple different regions of the food; determining target regions to be watered and water supplement amounts corresponding to each target region according to the water contents of each region and target water contents corresponding to each region; and watering the target regions according to coordinates corresponding to each target region and the water supplement amounts corresponding to each target region. The application can realize accurate and quantitative compensation of moisture in local regions of food and reduce loss in the vacuum cooling process of food.

Description

Technical Field

[0001] This application relates to the field of intelligent food processing technology, and in particular to an intelligent method, system and device for reducing moisture loss during vacuum cooling of food. Background Technology

[0002] Vacuum cooling rapidly lowers the boiling point of water by creating a vacuum. The evaporation of this water carries away heat from the surface and interior of the food, achieving rapid cooling and inhibiting bacterial growth. This method is widely used in food processing. However, current vacuum cooling technologies often rely on rapid vacuuming to reduce pressure and cause rapid evaporation of moisture. During this process, the drastic pressure drop leads to significant water evaporation, resulting in a substantial reduction in food weight and severely lowering the economic value of food sold by weight.

[0003] Currently, some companies and research institutions use deionized water (such as sodium hypochlorite) to soak some foods (e.g., meat products) to reduce quality loss during cooling, but this poses a risk of cross-contamination. Other institutions use ultrasonic atomization to replenish water to food within a specific temperature range, but this method is imprecise, has large errors, and can cause excessive moisture content in some areas, thus affecting food quality. Summary of the Invention

[0004] This application provides a method, system, and device for intelligent moisture compensation to reduce vacuum cooling loss in food. It can achieve precise quantitative compensation of moisture in local areas of food, solving the defects of quality loss in deionized food treatment and the defects of inaccuracy, large error, and excessive moisture content in some areas in ultrasonic atomization methods.

[0005] This application provides a smart moisture compensation method for reducing moisture loss during vacuum cooling of food, applied to a vacuum cooling system. The vacuum cooling system includes a vacuum chamber for holding food to be cooled. The method includes: The temperature field information, three-dimensional contour information, mass information, humidity information and pressure information of the vacuum chamber are obtained for the food. The temperature field information, the three-dimensional contour information, the mass information, the humidity information, and the pressure information are input into a pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food. Based on the water content of each region and the target water content corresponding to each region, determine the target region to be replenished and the water replenishment amount corresponding to each region; Water is replenished to the target areas based on their coordinates and the amount of water replenishment required.

[0006] According to the intelligent moisture compensation method provided in this application, the moisture content prediction model is trained using a multimodal fusion network. The multimodal fusion network includes at least a feature extraction layer, a feature fusion layer, and an output layer. The feature extraction layer includes three distinct feature extraction networks: a first feature extraction network, a second feature extraction network, and a third feature extraction network. The step of inputting the temperature field information, the three-dimensional contour information, the mass information, the humidity information, and the pressure information into the pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food includes: The temperature field information is input into the first feature extraction network to obtain a first feature vector; the three-dimensional contour information is input into the second feature extraction network to obtain a second feature vector; and the mass information, humidity information, and pressure information are input into the third feature extraction network to obtain a third feature vector. The first feature vector, the second feature vector, and the third feature vector are input into the feature fusion layer to obtain the fused features; The fusion features are input into the output layer to obtain the water content of each region.

[0007] According to the intelligent moisture compensation method provided in this application, the output layer is provided with a deconvolution layer, and the step of replenishing water to the target regions according to the coordinates corresponding to each target region and the water replenishment amount corresponding to each target region includes: The fused features are input into the deconvolution layer to obtain the moisture content distribution information of the food. Determine at least one coordinate corresponding to each of the target areas based on the water content distribution information; Water is replenished to the target areas based on at least one coordinate corresponding to each target area and the amount of water replenishment corresponding to each target area.

[0008] According to the intelligent moisture compensation method provided in this application, the vacuum cooling system includes a water replenishment module, the nozzle of which is located within the vacuum chamber; the step of replenishing water to the target areas according to the coordinates of each target area and the corresponding water replenishment amount includes: Based on the coordinates corresponding to each target area and the water replenishment amount corresponding to each coordinate in each target area, and with the goal of minimizing the total movement path through all coordinates in descending order of water replenishment amount, the optimal movement path corresponding to the nozzle is determined by a preset path planning algorithm. The nozzle is controlled to move according to the optimal movement path in order to replenish water to the target area.

[0009] According to the intelligent moisture compensation method provided in this application, the step of replenishing water to the target areas based on the coordinates corresponding to each target area and the water replenishment amount corresponding to each target area includes: The target flow rate of the nozzle is determined based on the moisture loss rate of the target area over a preset historical period. According to the target flow rate, water is replenished to the target areas based on the coordinates of each target area and the water replenishment amount corresponding to each target area.

[0010] According to the intelligent moisture compensation method provided in this application, based on the moisture content of each region and the target moisture content corresponding to each region, the target region to be replenished and the corresponding amount of water replenishment for each target region are determined, including: Determine the target moisture content and the difference between the moisture content and the target moisture content for each of the aforementioned regions; Regions with differences greater than a preset threshold are identified as target regions. The difference corresponding to the target area is determined as the water replenishment amount corresponding to the target area.

[0011] According to the intelligent moisture compensation method provided in this application, determining the target area to be replenished and the corresponding replenishment amount for each target area includes: Determine the target areas to be replenished with water, and the historical correction information corresponding to each target area to be replenished with water, wherein the historical correction information includes the correction value of the water replenishment amount corresponding to the target area; The water replenishment amount corresponding to each of the target areas is adjusted according to the adjustment value; The step of replenishing water to the target areas based on the coordinates of each target area and the corresponding water replenishment amount includes: Water is replenished to the target areas based on the coordinates of each target area and the corrected water replenishment amount for each target area.

[0012] This application also provides a vacuum cooling system, including: The cooling module has an internal vacuum chamber for holding food to be cooled. A three-dimensional laser scanning module is used to acquire the three-dimensional contour information of the food. Temperature field measurement module, used to acquire temperature field information of the food; A weight sensing module is used to acquire the quality information of the food. A pressure sensing module is used to acquire pressure information within the vacuum chamber; A humidity sensing module is used to acquire humidity information within the vacuum chamber; The lower-level machine is communicatively connected to the three-dimensional laser scanning module, the temperature field measurement module, the weight sensing module, the pressure sensing module, and the humidity sensing module, respectively. The lower-level machine is used to acquire the temperature field information, the three-dimensional contour information, the mass information, the humidity information, and the pressure information. The host computer is communicatively connected to the slave computer. The host computer is used to input the temperature field information, the three-dimensional contour information, the mass information, the humidity information, and the pressure information into a pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food; and to determine the target region to be replenished and the amount of water replenishment corresponding to each target region based on the moisture content of each region and the target moisture content corresponding to each region. The water replenishment module is used to replenish water to the target areas according to the coordinates of each target area and the water replenishment amount corresponding to each target area.

[0013] This application also provides a food moisture compensation device, including the following modules: The acquisition module is used to acquire temperature field information, three-dimensional contour information, mass information, humidity information and pressure information of the food located in the vacuum chamber. The input module is used to input the temperature field information, the three-dimensional contour information, the mass information, the humidity information, and the pressure information into a pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food. The determining module is used to determine the target area to be replenished and the amount of water replenishment corresponding to each target area based on the water content of each area and the target water content corresponding to each area; The compensation module is used to replenish water to the target areas according to the coordinates of each target area and the water replenishment amount of each target area.

[0014] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements a method for intelligently compensating for moisture loss in food vacuum cooling as described above.

[0015] This application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for intelligent moisture compensation to reduce vacuum cooling loss in food as described above.

[0016] This application also provides a computer program product, including a computer program that, when executed by a processor, implements a method for intelligently compensating for moisture loss in food vacuum cooling as described above.

[0017] The intelligent moisture compensation method for reducing food vacuum cooling losses according to this application first acquires the food's temperature field information, three-dimensional contour information, mass information, humidity information within the vacuum chamber, and pressure information within the vacuum chamber. Next, the temperature field information, three-dimensional contour information, mass information, humidity information, and pressure information are input into a pre-trained moisture content prediction model to obtain the moisture content of multiple different regions within the food. Then, based on the moisture content of each region and its corresponding target moisture content, the target regions to be replenished and the corresponding replenishment amounts for each target region are determined. Finally, based on the coordinates of each target region and the corresponding replenishment amounts, water is replenished to the target regions. This method enables precise quantitative compensation of moisture in localized areas of food, thereby effectively reducing losses during the food vacuum cooling process. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the structure of a vacuum cooling system provided in this application; Figure 2 This is a flowchart illustrating an embodiment of the present application of an intelligent moisture compensation method for reducing moisture loss during vacuum cooling of food; Figure 3 This is a schematic diagram of the structure of a cooling module according to an embodiment of this application; Figure 4 This is a schematic diagram of the water replenishment area of ​​a food product according to an embodiment of this application; Figure 5 This is a schematic diagram illustrating a water replenishment point according to an embodiment of this application; Figure 6 This is a schematic diagram of a water replenishment movement path shown in one embodiment of this application; Figure 7 This is a structural block diagram of an embodiment of the present application illustrating an intelligent moisture compensation device for reducing food vacuum cooling loss; Figure 8 This is a schematic diagram of the physical structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0021] The food moisture compensation method provided in this application is specifically applied to a vacuum cooling system. Figure 1 This is a structural schematic diagram of a vacuum cooling system provided in this application. (Refer to...) Figure 1 The vacuum cooling system of this application includes a cooling module, a water replenishment module, a three-dimensional laser scanning module, a temperature field measurement module, a weight sensing module, a pressure sensing module, a humidity sensing module, a host computer, and a slave computer.

[0022] The cooling module includes a vacuum chamber for holding food items to be cooled.

[0023] The water replenishment module is a Delta-3D compensation actuator used to receive and execute control commands sent by the lower-level machine. The module includes a nozzle located within the vacuum chamber, which can move within the chamber and spray water onto the food.

[0024] The lower-level machine can be a programmable logic controller (PLC). On the one hand, the lower-level machine communicates with the 3D laser scanning module, temperature field measurement module, weight sensing module, pressure sensing module, and humidity sensing module through input interfaces, and on the other hand, it communicates with the cooling module and water replenishment module through output interfaces.

[0025] The host computer can be an industrial computer, which communicates with the slave computer through a serial communication interface and is used to send control commands to the slave computer.

[0026] Figure 2 This is a flowchart illustrating an embodiment of the present application of a method for intelligent moisture compensation to reduce moisture loss during vacuum cooling of food. (Refer to...) Figure 2 The food moisture compensation method of this application may include the following steps: Step 101: Obtain the temperature field information, three-dimensional contour information, mass information, humidity information and pressure information inside the vacuum chamber of the food.

[0027] In this application, food may include meat, meat products, pre-cooked dishes, etc., and the specific type of food can be set according to actual needs.

[0028] In this application, temperature field information of the food surface can be collected by a temperature field measurement module, three-dimensional contour information of the food can be collected by a three-dimensional laser scanning module, mass information of the food can be collected by a weight sensing module, pressure information inside the vacuum chamber can be collected by a pressure sensing module, and humidity information inside the vacuum chamber can be collected by a humidity sensing module.

[0029] After receiving temperature field information, three-dimensional contour information, mass information, humidity information, and pressure information from each sensor module, the lower-level computer uploads this information to the upper-level computer.

[0030] Step 102: Input the temperature field information, three-dimensional contour information, mass information, humidity information and pressure information into the pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food.

[0031] In this application, the host computer analyzes temperature field information, three-dimensional contour information, mass information, humidity information and pressure information through a pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food.

[0032] The division of different regions in the food can be determined based on the structure of the vacuum chamber, and this application does not impose any restrictions on this.

[0033] In this application, multi-dimensional information such as temperature field, three-dimensional contour, mass, humidity, pressure, moisture content distribution, and moisture content of different regions within the food can be pre-collected during the vacuum cooling process. This information is then used as sample data to train a moisture content prediction model. During this process, the moisture content distribution and moisture content of different regions within the food can be acquired using insertion sensors.

[0034] Next, based on the collected information, mapping relationships are established between the food's temperature field, three-dimensional contour, mass, humidity within the vacuum chamber, and moisture content distribution under different vacuum pressures. Additionally, mapping relationships are established between these relationships and the moisture content of each region under different vacuum pressures. Thus, during subsequent vacuum cooling of the food, the host computer can directly input the collected temperature field, three-dimensional contour, mass, humidity, and pressure information into the moisture content prediction model. Based on the input information and the previously established mapping relationships, the moisture content prediction model analyzes and obtains the current moisture content distribution information of the food and the moisture content of each region. This eliminates the need for insertable sensors to obtain the moisture content distribution information and the moisture content of each region, avoiding food contamination.

[0035] Step 103: Based on the water content of each region and the target water content corresponding to each region, determine the target areas to be replenished and the corresponding water replenishment amount for each target area.

[0036] The water content of each region refers to the current water content Q of each region obtained through the water content prediction model. 当前 Target moisture content Q 目标 This refers to the expected moisture content for each region. Target moisture content Q 目标 It can be set according to actual needs.

[0037] Specifically, step 103 may include: Determine the target moisture content and the difference between the target moisture content and the actual moisture content for each region; Regions with differences greater than a preset threshold are identified as target regions. The difference corresponding to the target area is determined as the water replenishment amount for the target area.

[0038] In this application, for each region, Q is obtained. 目标 With Q 当前 The difference (i.e. Q) 目标 -Q 当前 If the difference is greater than a preset threshold (set according to actual needs, for example, it can be set to 0), it indicates that the area has lost a lot of water and needs to be replenished. Therefore, this area can be considered a target area. After determining all target areas that need water replenishment based on the relationship between the difference and the preset threshold, the difference corresponding to each target area can be used as the water replenishment amount for each target area, i.e., Q. 补 =Q 目标 -Q 当前 .

[0039] Step 104: Replenish water to the target areas according to the coordinates of each target area and the water replenishment amount for each target area.

[0040] In this application, the locations of the nozzles that replenish water to each target area can be determined first, and then the three-dimensional spatial coordinates of these locations can be obtained and used as the coordinates of the target area.

[0041] A target area can correspond to at least one point (which can be determined based on the nozzle diameter), meaning a target area can correspond to at least one coordinate. This allows the nozzle to spray water from multiple points when replenishing a large area, ensuring sufficient hydration. Therefore, after determining the water replenishment amount for each target area, we can further determine at least one point corresponding to that target area and the water replenishment amount at each point. Finally, based on each point (coordinate) and the corresponding water replenishment amount, we can replenish water to all target areas.

[0042] In this application, after determining the coordinates and water replenishment volume corresponding to each target area, the host computer can generate control commands and send them to the slave computer. The slave computer executes the control commands to control the movement of the water replenishment module, so that the nozzles spray water sequentially at each point according to the corresponding water replenishment volume, thereby replenishing the corresponding target area until the Q corresponding to each target area is reached. 目标 With moisture content Q 当前 The difference is less than or equal to the preset threshold.

[0043] Figure 3 This is a schematic diagram illustrating the structure of a cooling module according to an embodiment of this application. Figure 3 In the diagram, the nine boxes correspond to nine different areas of the food. If all nine areas are target areas, then the specific areas for hydration within those target areas can be defined as follows: Figure 4 The slashed part is shown in the image. Figure 4 This is a schematic diagram of the water replenishment area of ​​a food product according to an embodiment of this application.

[0044] Steps 101-104 described above can be applied both before and during vacuum cooling. Before vacuum cooling, i.e., before starting the cooling module, steps 101-104 can be used to determine the target areas requiring water replenishment and the corresponding water replenishment amount for each target area, and pre-replenish water to each target area according to the replenishment amount. Then, after starting the cooling module, at preset intervals (set according to actual needs), a water replenishment strategy (i.e., determining the target areas requiring water replenishment and the corresponding water replenishment amount for each target area) can be obtained through steps 101-104, and the water replenishment strategy (i.e., replenishing water to each target area according to the replenishment amount) can be executed until all target areas Q 目标 With Q 当前 The difference is not greater than a preset threshold. In other words, during the entire vacuum cooling process, the vacuum cooling system of this application can replenish water to the areas that need to be replenished, maintaining Q in all areas. 目标 With Q 当前 The difference is never greater than the preset threshold.

[0045] The food moisture compensation method of this application first acquires the food's temperature field information, three-dimensional contour information, mass information, humidity information within the vacuum chamber, and pressure information within the vacuum chamber. Next, the temperature field information, three-dimensional contour information, mass information, humidity information, and pressure information are input into a pre-trained moisture content prediction model to obtain the moisture content of multiple different regions within the food. Then, based on the moisture content of each region and its corresponding target moisture content, the target regions to be replenished and the corresponding replenishment amounts for each target region are determined. Finally, based on the coordinates of each target region and the corresponding replenishment amounts, moisture is replenished to the target regions. This method enables precise quantitative compensation of moisture in localized areas of food, thereby effectively reducing losses during the vacuum cooling process.

[0046] In conjunction with the above embodiments, in one implementation, the water content prediction model is trained using a multimodal fusion network (Hybrid Neural Network). The multimodal fusion network includes at least a feature extraction layer, a feature fusion layer, and an output layer. The feature extraction layer includes a first feature extraction network, a second feature extraction network, and a third feature extraction network that are different from each other.

[0047] Accordingly, temperature field information, three-dimensional contour information, mass information, humidity information, and pressure information are input into a pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food, which may include: Temperature field information is input into the first feature extraction network to obtain the first feature vector; three-dimensional contour information is input into the second feature extraction network to obtain the second feature vector; and mass information, humidity information and pressure information are input into the third feature extraction network to obtain the third feature vector. The first feature vector, the second feature vector, and the third feature vector are input into the feature fusion layer to obtain the fused features; The fused features are input to the output layer to obtain the water content of each region.

[0048] In this application, the first feature extraction network can be a convolutional neural network (CNN). The temperature field information of the food collected by the temperature field measurement module is image data. The image data representing the temperature field information is input into the first feature extraction network, and the spatial features of the temperature field can be extracted through the first feature extraction network to obtain the first feature vector.

[0049] The 3D contour information of the food acquired by the 3D laser scanning module is point cloud data. The second feature extraction network can use PointNet or VoxelNet to process the 3D contour information and obtain a second feature vector representing the 3D contour of the food.

[0050] The third feature extraction network can use fully connected layers to extract features from scalar information such as mass information, humidity information, and pressure information, and obtain the third feature vector.

[0051] The feature fusion layer can use a fully connected layer to fuse the first feature vector, the second feature vector, and the third feature vector to obtain fused features.

[0052] The output layer can use a fully connected layer to output data representing the water content of each region, thereby obtaining the water content of each region.

[0053] In one implementation, a deconvolution layer is provided in the output layer. Accordingly, water replenishment is performed on the target regions based on the coordinates of each target region and the corresponding water replenishment amount, which may include: The fused features are input into the deconvolution layer to obtain the moisture content distribution information of the food. Determine at least one coordinate corresponding to each target area based on the water content distribution information; Water is replenished to the target areas based on at least one coordinate corresponding to each target area and the corresponding water replenishment amount for each target area.

[0054] In this application, the moisture content prediction model, based on input temperature field information, three-dimensional contour information, mass information, humidity information, and pressure information, can obtain both the moisture content distribution information of the food and the moisture content of each region. The moisture content of each region differs from the overall moisture content distribution information of the food. While the moisture content of each region indicates the amount of water contained within it, it doesn't reveal how the water is distributed. The moisture content distribution information, being image information, clearly defines the distribution of moisture within each region. Figure 4 As shown.

[0055] This application determines all target areas requiring water replenishment based on the relationship between the difference between the target moisture content and the current moisture content of each region of the food and a preset threshold. Then, based on the target areas and the moisture content distribution information of the food, it determines the specific water replenishment areas within the target areas (e.g., ...). Figure 4 (as shown in the water replenishment area), and further determine at least one coordinate corresponding to each water replenishment area and the water replenishment amount corresponding to each coordinate, such as Figure 5 As shown. At least one coordinate corresponding to a target area is also at least one coordinate corresponding to a specific water replenishment area within that target area. Figure 5 In the middle, the water replenishment area is relatively large, so two points (point 1 and point 2) are set with corresponding coordinates. When the nozzle sprays water at these two coordinates, the water in the area can be fully replenished. Figure 5This is a schematic diagram illustrating a water replenishment point according to an embodiment of this application. Finally, based on at least one coordinate corresponding to the target area and the water replenishment amount corresponding to each coordinate, water replenishment can be achieved for all target areas.

[0056] In this application, the moisture content prediction model establishes a mapping relationship during the training phase through the following process: First, a large amount of training sample data is collected. Each set of sample data should include two parts: one part is the input information, namely, the food temperature field information, three-dimensional contour information, mass information, humidity information, and pressure information collected in real time by the temperature field measurement module, three-dimensional laser scanning module, weight sensing module, humidity sensing module, and pressure sensing module under specific vacuum pressure and cooling conditions; the other part is the corresponding real label information, namely, the actual moisture content distribution information of the food under this condition and the moisture content of each region. This real label information can be obtained through methods such as insertable sensors during the training data acquisition phase. During the training process, the multimodal fusion network receives this input information. The first, second, and third feature extraction networks in its feature extraction layer, which are different from each other, process the data of different modalities such as temperature field (image), three-dimensional contour (point cloud), and mass, humidity, and pressure (scalar), respectively, to obtain their respective first, second, and third feature vectors. Then, the feature fusion layer fuses these feature vectors to obtain a unified fused feature. Finally, the output layer decodes based on the fused features, outputting the predicted moisture content and moisture distribution information for each region. The model compares its predicted output with the real label information obtained through the inserted sensors and calculates the loss, then continuously adjusts the network parameters through optimization algorithms such as backpropagation. By repeating this process on a large amount of sample data, the multimodal fusion network can autonomously learn and establish a complex mapping relationship between multi-dimensional, multimodal sensor inputs and the moisture content and moisture distribution inside food.

[0057] In this application, a moisture content prediction model is trained through a multimodal fusion network. Based on the current input temperature field information, three-dimensional contour information, mass information, humidity information, and pressure information, the model can accurately obtain the moisture content distribution information of food and the moisture content of each region. This helps to accurately obtain the water replenishment points and the amount of water replenishment at each point, thus achieving precise quantitative water replenishment to local areas of the food.

[0058] In conjunction with the above embodiments, in one implementation, replenishing water to the target areas based on the coordinates of each target area and the corresponding water replenishment amount for each target area may include: Based on the coordinates of each target area and the water replenishment volume corresponding to each coordinate in each target area, the optimal movement path for the nozzle is determined by using a preset path planning algorithm, with the goal of minimizing the total movement path through all coordinates and ranking the water replenishment volume corresponding to each coordinate from largest to smallest. Control the nozzle movement according to the optimal movement path to replenish water to the target area.

[0059] The preset path planning algorithm can be either Dijkstra's algorithm or A* algorithm.

[0060] In this application, after determining the water replenishment amount corresponding to a target region, a pre-trained image recognition model can be used to determine at least one coordinate, the position of each coordinate, and the water replenishment amount corresponding to each coordinate for each target region. Specifically, for any target region X, the image recognition model first performs image segmentation on the specific water replenishment area within it, obtaining multiple sub-regions, the center position of each sub-region, and the area proportion of each sub-region within the specific water replenishment area of ​​target region X. Next, the number of sub-regions corresponding to target region X is determined as the number of coordinates corresponding to target region X (one sub-region corresponds to one coordinate), the center position of each sub-region corresponding to target region X is determined as the position of each coordinate corresponding to target region X, and the product of the water replenishment amount corresponding to target region X and the area proportion of each sub-region is taken as the water replenishment amount corresponding to each coordinate. When using the image recognition model, constraints can be set for the area of ​​each sub-region in the image recognition model to avoid obtaining multiple unnecessary coordinates. This application does not restrict the structure and training method of the image recognition model and can set it according to actual needs.

[0061] In this application, after obtaining the coordinates of each water replenishment point and the corresponding water replenishment volume, higher priority can be assigned to coordinates with larger water replenishment volumes and higher priority to shorter straight-line paths between coordinates. The optimal movement path for the nozzle is then determined using the Dijkstra algorithm. Figure 6 As shown. In Figure 6 In the diagram, the arrows indicate the direction of nozzle movement. The coordinates corresponding to the water replenishment points are (x1, y1, z1), (x2, y2, z2), (x3, y3, z3), (x4, y4, z4), (x5, y5, z5), and (x6, y6, z6), with the corresponding water replenishment volume being Q. 补1 Q 补2 Q 补3 Q 补4 Q 补5 Q 补6 . Figure 6 This is a schematic diagram of a water replenishment movement path shown in one embodiment of this application.

[0062] In one implementation, when obtaining the optimal path, the straight-line distances between all coordinates are first calculated. Then, a path weight is applied that considers both the amount of water replenishment and the length of the straight-line path. The path weight is calculated as path length / water replenishment at a point. This path weight indicates that the shorter the path length and the greater the water replenishment at a point, the lower the path weight (and the higher the priority). All points are then converted into a weighted graph, and the optimal movement path is obtained using Dijkstra's algorithm.

[0063] In another implementation, when using the A* algorithm to obtain the optimal movement path, each water replenishment point is treated as a node, containing the coordinates of that point and the corresponding water replenishment amount. Next, connecting edges are established between all nodes, and the weight of each edge can be calculated based on the Euclidean distance (i.e., the length of the straight path) between the nodes. To prioritize coordinates with larger water replenishment amounts, a heuristic function can be defined. h ( n ),For example h ( n )=- α × Water replenishment n ( α It is a positive weighting coefficient used to adjust the impact of water replenishment on path planning; the larger the water replenishment, the greater the impact. h ( n The smaller the value of ), the higher the priority it will be in the cost function. To consider paths with shorter straight-line distances, we can define a path cost function g(n), using the distance between nodes as . g ( n The value of ), that is g ( n () equals from the starting node to the node n The sum of the weights of the edges traversed. Finally, the total cost function is f(n) = g(n) + h(n). The A* algorithm selects the nodes to expand based on the value of f(n), thus obtaining the optimal movement path.

[0064] In conjunction with the above embodiments, in one implementation, replenishing water to the target areas based on the coordinates of each target area and the corresponding water replenishment amount for each target area may include: The target flow rate of the nozzle is determined based on the moisture loss rate of the target area over a preset historical period. Based on the target flow rate, water is replenished to the target areas according to their coordinates and the corresponding water replenishment volume.

[0065] In this application, the nozzle flow rate refers to the amount of water output from the nozzle per unit time. When replenishing water to a target area, the nozzle flow rate should be greater than or equal to the moisture loss rate of that target area to ensure that the amount of water received by the food is greater than the evaporation rate under vacuum cooling.

[0066] The preset historical duration can be set according to actual needs.

[0067] In conjunction with the above embodiments, in one implementation, step 103, determining the target area to be replenished with water and the corresponding water replenishment amount for each target area, includes: Determine the target areas to be replenished with water, and the historical correction information corresponding to each target area to be replenished with water, wherein the historical correction information includes the correction value of the water replenishment amount corresponding to the target area; The water replenishment amount corresponding to each of the target areas is adjusted according to the correction value.

[0068] Accordingly, in step 104, replenishing the target areas with water according to the coordinates of each target area and the water replenishment amount corresponding to each target area includes: Water is replenished to the target areas based on the coordinates of each target area and the corrected water replenishment amount for each target area.

[0069] In this application, the vacuum cooling process may involve multiple stages of water replenishment (e.g., water replenishment at preset intervals). Each stage of water replenishment may involve multiple water replenishments, and steps 101-104 above illustrate the specific principle of performing a single water replenishment operation.

[0070] In a single water replenishment phase, after each replenishment operation, a correction operation is performed. This correction operation involves collecting temperature field information, three-dimensional contour information, mass information, humidity information within the vacuum chamber, and pressure information within the vacuum chamber. This information is then input again into a pre-trained moisture content prediction model to obtain the moisture content of each region and the moisture content distribution information of the food. Based on this information, the following decisions are made: (1) For a certain region A, if Q 第一次补后 >Q 目标 This indicates that area A has received excessive water replenishment, and no further water replenishment is needed for the time being. Q 第一次补后 This is the water content after the first water replenishment in the current stage. The host computer will record this information (including areas A and Q). 第一次补后 Q 目标(etc.). If region A experiences water shortage again during subsequent vacuum cooling, the host computer will make another decision to replenish water to region A. When replenishing water to region A for the second time in a later stage, the host computer will use the Q value from the first water replenishment. 第一次补后 -Q 目标 The difference (i.e., the correction value) reduces the amount of water to be replenished the second time, meaning the actual amount of water to be replenished the second time is Q. 第二次补实际 =Q 补 -(Q 第一次补后 -Q 目标 ), where Q 补 The theoretical water replenishment amount Q calculated according to steps 101-104 above. 补 In actual hydration, Q will be used. 第二次补实际 Replace Q 补 The correction process occurs throughout the entire cooling process, effectively conserving water resources. The information recorded by the host computer is the historical correction information. After determining the target areas to be replenished with water, each target area needs to be corrected according to the method of area A.

[0071] (2) For a certain region A, if Q 第一次补后 目标 This indicates that area A is under-watered and requires additional water replenishment during this phase until area A reaches Q. 第一次补后 ≥Q 目标 Only then does it indicate that the water replenishment for area A in this phase has ended. If Q 第一次补后 >Q 目标 Execute the strategy in step (1) above to reduce Q. 第一次补后 With Q 目标 The difference between them.

[0072] This application also provides a vacuum cooling system, including: The cooling module has an internal vacuum chamber for holding food to be cooled. A three-dimensional laser scanning module is used to acquire the three-dimensional contour information of the food. Temperature field measurement module, used to acquire temperature field information of the food; A weight sensing module is used to acquire the quality information of the food. A pressure sensing module is used to acquire pressure information within the vacuum chamber; A humidity sensing module is used to acquire humidity information within the vacuum chamber; ​The lower-level machine is communicatively connected to the three-dimensional laser scanning module, the temperature field measurement module, the weight sensing module, the pressure sensing module, and the humidity sensing module, respectively. The lower-level machine is used to acquire the temperature field information, the three-dimensional contour information, the mass information, the humidity information, and the pressure information. The host computer is communicatively connected to the slave computer. The host computer is used to input the temperature field information, the three-dimensional contour information, the mass information, the humidity information, and the pressure information into a pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food; and to determine the target region to be replenished and the amount of water replenishment corresponding to each target region based on the moisture content of each region and the target moisture content corresponding to each region. The water replenishment module is used to replenish water to the target areas according to the coordinates of each target area and the amount of water to be replenished for each target area.

[0073] The structure of the vacuum cooling system can be as follows: Figure 1 As shown, the functions of each part in the vacuum cooling system can be referred to the above description, and will not be repeated here.

[0074] The following describes an intelligent moisture compensation device for reducing food vacuum cooling loss. The intelligent moisture compensation device for reducing food vacuum cooling loss described below can be referred to in conjunction with the intelligent moisture compensation method for reducing food vacuum cooling loss described above.

[0075] Figure 7 This is a structural block diagram of an intelligent moisture compensation device for reducing moisture loss during vacuum cooling of food, as illustrated in one embodiment of this application. (Refer to...) Figure 7 The intelligent moisture compensation device 700 for reducing food vacuum cooling loss in this application may include: The acquisition module 701 is used to acquire temperature field information, three-dimensional contour information, mass information, humidity information and pressure information of the food located in the vacuum chamber. The input module 702 is used to input the temperature field information, the three-dimensional contour information, the mass information, the humidity information, and the pressure information into a pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food. The determining module 703 is used to determine the target area to be replenished and the amount of water replenishment corresponding to each target area based on the water content of each area and the target water content corresponding to each area; The compensation module 704 is used to replenish water to the target area according to the coordinates of each target area and the water replenishment amount of each target area.

[0076] According to the intelligent moisture compensation device 700 for reducing vacuum cooling loss in food provided in this application, the moisture content prediction model is trained using a multimodal fusion network. The multimodal fusion network includes at least a feature extraction layer, a feature fusion layer, and an output layer. The feature extraction layer includes three distinct feature extraction networks: a first feature extraction network, a second feature extraction network, and a third feature extraction network. The input module 702 includes: The first input submodule is used to input the temperature field information into the first feature extraction network to obtain a first feature vector, input the three-dimensional contour information into the second feature extraction network to obtain a second feature vector, and input the mass information, the humidity information, and the pressure information into the third feature extraction network to obtain a third feature vector. The second input submodule is used to input the first feature vector, the second feature vector and the third feature vector into the feature fusion layer to obtain fused features; The third input submodule is used to input the fusion features into the output layer to obtain the water content of each region.

[0077] According to the intelligent moisture compensation device 700 for reducing vacuum cooling loss in food provided in this application, the output layer is provided with a deconvolution layer, and the compensation module 704 includes: The fourth input submodule is used to input the fused features into the deconvolution layer to obtain the moisture content distribution information of the food. The first determining submodule is used to determine at least one coordinate corresponding to each of the target areas based on the water content distribution information; The first compensation submodule is used to replenish water to the target area based on at least one coordinate corresponding to each target area and the water replenishment amount corresponding to each target area.

[0078] According to the intelligent moisture compensation device 700 for reducing vacuum cooling loss in food provided in this application, a nozzle is provided in the vacuum chamber; the compensation module 704 includes: The second determining submodule is used to determine the optimal movement path of the nozzle based on the coordinates corresponding to each of the target areas and the water replenishment volume corresponding to each coordinate in each of the target areas, with the goal of minimizing the total movement path through all coordinates in descending order of water replenishment volume. The control submodule is used to control the movement of the nozzle according to the optimal movement path in order to replenish water to the target area.

[0079] According to the intelligent moisture compensation device 700 for reducing vacuum cooling loss in food provided in this application, the compensation module 704 includes: The third determining submodule is used to determine the target flow rate of the nozzle based on the moisture loss rate of the target area within a preset historical time period; The second compensation submodule is used to replenish water to the target area according to the target flow rate, based on the coordinates of each target area and the water replenishment amount corresponding to each target area.

[0080] According to the intelligent moisture compensation device 700 for reducing vacuum cooling loss in food provided in this application, the determining module 703 includes: The fourth determining submodule is used to determine the difference between the target moisture content and the actual moisture content for each of the aforementioned regions; The fifth determination submodule is used to determine the target area as the region where the difference is greater than a preset threshold; The sixth determination submodule is used to determine the difference corresponding to the target area as the water replenishment amount corresponding to the target area.

[0081] According to the intelligent moisture compensation device 700 for reducing vacuum cooling loss in food provided in this application, the determining module 703 includes: The seventh determination submodule is used to determine the target area to be replenished with water and the historical correction information corresponding to each target area to be replenished with water, wherein the historical correction information includes the correction value of the water replenishment amount corresponding to the target area; The correction submodule is used to correct the water replenishment amount corresponding to each of the target areas according to the correction value; The compensation module 704 includes: The third compensation submodule is used to replenish water to the target areas based on the coordinates corresponding to each target area and the corrected water replenishment amount corresponding to each target area.

[0082] Figure 8 This is a schematic diagram of the physical structure of an electronic device according to an embodiment of this application, as shown below. Figure 8 As shown, the electronic device may include: a processor 810, a communications interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communications interface 820, and the memory 830 communicate with each other via the communication bus 840. The processor 810 can call logical instructions in the memory 830 to execute a food moisture compensation method, which includes: The temperature field information, three-dimensional contour information, mass information, humidity information and pressure information of the vacuum chamber are obtained for the food. The temperature field information, the three-dimensional contour information, the mass information, the humidity information, and the pressure information are input into a pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food. Based on the water content of each region and the target water content corresponding to each region, determine the target region to be replenished and the water replenishment amount corresponding to each region; Water is replenished to the target areas based on their coordinates and the amount of water replenishment required.

[0083] Furthermore, the logical instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0084] On the other hand, this application also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to perform a food moisture compensation method provided by the above methods, the method comprising: The temperature field information, three-dimensional contour information, mass information, humidity information and pressure information of the vacuum chamber are obtained for the food. The temperature field information, the three-dimensional contour information, the mass information, the humidity information, and the pressure information are input into a pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food. Based on the water content of each region and the target water content corresponding to each region, determine the target region to be replenished and the water replenishment amount corresponding to each region; Water is replenished to the target areas based on their coordinates and the amount of water replenishment required.

[0085] Furthermore, this application also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, is implemented to perform a food moisture compensation method provided by the methods described above, the method comprising: The temperature field information, three-dimensional contour information, mass information, humidity information and pressure information of the vacuum chamber are obtained for the food. The temperature field information, the three-dimensional contour information, the mass information, the humidity information, and the pressure information are input into a pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food. Based on the water content of each region and the target water content corresponding to each region, determine the target region to be replenished and the water replenishment amount corresponding to each region; Water is replenished to the target areas based on their coordinates and the amount of water replenishment required.

[0086] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0087] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for intelligently compensating for moisture loss during vacuum cooling of food, characterized in that, A method applicable to a vacuum cooling system, wherein the vacuum cooling system includes a vacuum chamber for holding food to be cooled, the method comprising: The temperature field information, three-dimensional contour information, mass information, humidity information and pressure information of the vacuum chamber are obtained for the food. The temperature field information, the three-dimensional contour information, the mass information, the humidity information, and the pressure information are input into a pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food. Based on the water content of each region and the target water content corresponding to each region, determine the target region to be replenished and the water replenishment amount corresponding to each region; Water is replenished to the target areas based on their coordinates and the amount of water replenishment required.

2. The intelligent moisture compensation method for reducing vacuum cooling loss in food according to claim 1, characterized in that, The moisture content prediction model is trained using a multimodal fusion network. This multimodal fusion network includes at least a feature extraction layer, a feature fusion layer, and an output layer. The feature extraction layer includes three distinct feature extraction networks: a first feature extraction network, a second feature extraction network, and a third feature extraction network. The process of inputting the temperature field information, the three-dimensional contour information, the mass information, the humidity information, and the pressure information into the pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food includes: The temperature field information is input into the first feature extraction network to obtain a first feature vector; the three-dimensional contour information is input into the second feature extraction network to obtain a second feature vector; and the mass information, humidity information, and pressure information are input into the third feature extraction network to obtain a third feature vector. The first feature vector, the second feature vector, and the third feature vector are input into the feature fusion layer to obtain the fused features; The fusion features are input into the output layer to obtain the water content of each region.

3. The intelligent moisture compensation method for reducing vacuum cooling loss in food according to claim 2, characterized in that, The output layer includes a deconvolution layer. The step of replenishing water to the target regions based on their coordinates and the corresponding water replenishment amounts includes: The fused features are input into the deconvolution layer to obtain the moisture content distribution information of the food. Determine at least one coordinate corresponding to each of the target areas based on the water content distribution information; Water is replenished to the target areas based on at least one coordinate corresponding to each target area and the amount of water replenishment corresponding to each target area.

4. The intelligent moisture compensation method for reducing vacuum cooling loss in food according to claim 1, characterized in that, The vacuum cooling system includes a water replenishment module, the nozzle of which is located within the vacuum chamber; the step of replenishing water to the target areas according to the coordinates and water replenishment amounts corresponding to each target area includes: Based on the coordinates corresponding to each target area and the water replenishment amount corresponding to each coordinate in each target area, and with the goal of minimizing the total movement path through all coordinates in descending order of water replenishment amount, the optimal movement path corresponding to the nozzle is determined by a preset path planning algorithm. The nozzle is controlled to move according to the optimal movement path in order to replenish water to the target area.

5. The intelligent moisture compensation method for reducing vacuum cooling loss in food according to claim 4, characterized in that, The step of replenishing water to the target areas based on the coordinates of each target area and the corresponding water replenishment amount includes: The target flow rate of the nozzle is determined based on the moisture loss rate of the target area over a preset historical period. According to the target flow rate, water is replenished to the target areas based on the coordinates of each target area and the water replenishment amount corresponding to each target area.

6. The intelligent moisture compensation method for reducing vacuum cooling loss in food according to claim 1, characterized in that, Based on the moisture content of each region and the target moisture content corresponding to each region, the target regions to be replenished and the corresponding replenishment amounts for each target region are determined, including: Determine the difference between the target moisture content and the moisture content corresponding to each of the aforementioned regions; Regions with differences greater than a preset threshold are identified as target regions. The difference corresponding to the target area is determined as the water replenishment amount corresponding to the target area.

7. The intelligent moisture compensation method for reducing vacuum cooling loss in food according to claim 1, characterized in that, The determination of the target areas to be replenished and the corresponding water replenishment amounts for each target area includes: Determine the target areas to be replenished with water, and the historical correction information corresponding to each target area to be replenished with water, wherein the historical correction information includes the correction value of the water replenishment amount corresponding to the target area; The water replenishment amount corresponding to each of the target areas is adjusted according to the adjustment value; The step of replenishing water to the target areas based on the coordinates of each target area and the corresponding water replenishment amount includes: Water is replenished to the target areas based on the coordinates of each target area and the corrected water replenishment amount for each target area.

8. A vacuum cooling system, characterized in that, include: The cooling module has an internal vacuum chamber for holding food to be cooled. A three-dimensional laser scanning module is used to acquire the three-dimensional contour information of the food. Temperature field measurement module, used to acquire temperature field information of the food; A weight sensing module is used to acquire the quality information of the food. A pressure sensing module is used to acquire pressure information within the vacuum chamber; A humidity sensing module is used to acquire humidity information within the vacuum chamber; The lower-level machine is communicatively connected to the three-dimensional laser scanning module, the temperature field measurement module, the weight sensing module, the pressure sensing module, and the humidity sensing module, respectively. The lower-level machine is used to acquire the temperature field information, the three-dimensional contour information, the mass information, the humidity information, and the pressure information. The host computer is connected to the slave computer. The host computer is used to input the temperature field information, the three-dimensional contour information, the mass information, the humidity information and the pressure information into the pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food. And based on the water content of each region and the target water content corresponding to each region, determine the target region to be replenished and the water replenishment amount corresponding to each target region; The water replenishment module is used to replenish water to the target areas according to the coordinates of each target area and the amount of water to be replenished for each target area.

9. A smart moisture compensation device for reducing moisture loss during vacuum cooling of food, characterized in that, include: The acquisition module is used to acquire temperature field information, three-dimensional contour information, mass information, humidity information and pressure information of the food located in the vacuum chamber. The input module is used to input the temperature field information, the three-dimensional contour information, the mass information, the humidity information, and the pressure information into a pre-trained moisture content prediction model to obtain the moisture content of multiple different regions in the food. The determining module is used to determine the target area to be replenished and the amount of water replenishment corresponding to each target area based on the water content of each area and the target water content corresponding to each area; The compensation module is used to replenish water to the target areas according to the coordinates of each target area and the water replenishment amount of each target area.

10. A smart moisture compensation device for reducing vacuum cooling loss in food according to claim 9, characterized in that, The water content prediction model is trained using a multimodal fusion network, which includes at least a feature extraction layer, a feature fusion layer, and an output layer. The feature extraction layer includes a first feature extraction network, a second feature extraction network, and a third feature extraction network that are different from each other. The input module includes: The first input submodule is used to input the temperature field information into the first feature extraction network to obtain a first feature vector, input the three-dimensional contour information into the second feature extraction network to obtain a second feature vector, and input the mass information, the humidity information, and the pressure information into the third feature extraction network to obtain a third feature vector. The second input submodule is used to input the first feature vector, the second feature vector and the third feature vector into the feature fusion layer to obtain fused features; The third input submodule is used to input the fusion features into the output layer to obtain the water content of each region.

Images