Baking surface partitioning precision temperature control method and system combined with visual semantic segmentation

Abstract

The application provides a baking surface partition accurate temperature control method and system combined with visual semantic segmentation, a plurality of frames of thermal radiation distribution maps under a continuous time sequence of a baking surface area are acquired by a thermal imaging acquisition unit, a spatial semantic segmentation processing based on the baking surface area is performed on the plurality of frames of thermal radiation distribution maps, and partition thermal radiation field data corresponding to each baking sub-partition is generated. According to the partition thermal radiation field data and the working state parameters of the heating execution unit preset in each baking sub-partition, a partition heating deviation analysis is performed to generate a compensation heating demand instruction of each baking sub-partition. A collaborative control processing of the heating execution unit is performed based on the compensation heating demand instructions of all baking sub-partitions to obtain a partition power adjustment signal for each baking sub-partition, and the partition power adjustment signal is sent to the corresponding heating execution unit to drive each baking sub-partition to perform a differential heating operation. The application can further improve the accuracy and stability of the baking process.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence, and more specifically, to a method and system for precise temperature control of baking dough by combining visual semantic segmentation. Background Technology

[0002] In the field of baked food processing, precise temperature control is crucial for ensuring the taste, color, and nutritional content of food. Traditional baking equipment typically uses a uniform heating method, controlling the baking process through preset, fixed heating temperatures and times. However, this method has significant limitations. Due to differences in heat conduction characteristics, food thickness, and initial temperature across different areas of the baking surface, uniform heating struggles to achieve even heating across all areas, easily leading to some areas being burnt while others remain uncooked. To address these issues, existing technologies have developed zoned heating baking equipment. By dividing the baking surface into zones and controlling the heating power of each zone separately, the uniformity of baking is improved to some extent. However, most of these devices rely on fixed heating strategies and lack real-time monitoring and dynamic adjustment of the actual temperature of each zone during the baking process. While some devices incorporate temperature sensors, the number of sensors is limited and their distribution is uneven, making it difficult to comprehensively and accurately reflect the temperature conditions of each area of ​​the baking surface, thus failing to achieve truly precise temperature control. Furthermore, existing technologies lack effective analysis methods when processing baking temperature data, failing to fully extract the information behind the temperature data and making it difficult to perform intelligent heating control based on the actual conditions of the baking process. Summary of the Invention

[0003] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide a method for precise temperature control of baking dough zones based on visual semantic segmentation, the method comprising: Obtain multi-frame thermal radiation distribution maps of the baking surface area in a continuous time series, acquired by the thermal imaging acquisition unit. Perform spatial semantic segmentation processing based on the baking surface region on the multi-frame thermal radiation distribution map to generate partition thermal radiation field data corresponding to each baking sub-partition. Based on the partitioned thermal radiation field data and the working status parameters of the heating execution unit preset in each baking sub-partition, partitioned heating deviation analysis is performed to generate compensation heating demand instructions for each baking sub-partition. Based on the compensation heating demand commands of all baking sub-zones, the heating execution unit is coordinated and controlled to obtain the partition power adjustment signal for each baking sub-zone. The partition power adjustment signal is sent to the corresponding heating execution unit to drive each baking sub-partition to perform differentiated heating operations.

[0004] In another aspect, embodiments of the present invention also provide a precise temperature control system for baking dough zones that combines visual semantic segmentation, including a processor and a machine-readable storage medium. The machine-readable storage medium is connected to the processor, the machine-readable storage medium is used to store programs, instructions or code, and the processor is used to execute the programs, instructions or code in the machine-readable storage medium to implement the above-described method.

[0005] Based on the above, this embodiment of the invention acquires multi-frame thermal radiation distribution maps of the baking surface area over a continuous time series using a thermal imaging acquisition unit. This captures temperature information from each area of ​​the baking surface, avoiding the limitations of traditional temperature sensors, which are limited in number and unevenly distributed. Spatial semantic segmentation processing based on the baking surface area is performed on the multi-frame thermal radiation distribution maps to generate thermal radiation field data for each baking sub-zone. This achieves fine division of the baking surface and accurate extraction of temperature data, accurately reflecting the temperature changes in each sub-zone. Based on the sub-zone thermal radiation field data and the pre-set operating state parameters of the heating execution units in each baking sub-zone, sub-zone heating deviation analysis is performed to generate compensation heating demand commands for each baking sub-zone. Based on these compensation heating demand commands, coordinated control processing of the heating execution units is performed to obtain a sub-zone power adjustment signal for each baking sub-zone. This achieves dynamic and precise adjustment of the heating power of each sub-zone, effectively solving the problem of uneven heating in different areas of the baking surface and improving the quality and consistency of baked goods. Meanwhile, the method also has the ability to monitor and handle abnormal situations, such as response lag partitioning, abnormal temperature rise partitioning, and excessive thermal radiation fluctuation partitioning, which can adjust the heating strategy in a timely manner to ensure the stability and reliability of the baking process. Attached Figure Description

[0006] Figure 1 This is a schematic diagram of the execution flow of the precise temperature control method for baking dough partitioning combined with visual semantic segmentation provided in an embodiment of the present invention.

[0007] Figure 2 This is a schematic diagram of exemplary hardware and software components of the precise temperature control system for baking dough zones that combines visual semantic segmentation, provided in an embodiment of the present invention. Detailed Implementation

[0008] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1 This is a flowchart illustrating a method for precise temperature control of baking dough zones based on visual semantic segmentation, provided in one embodiment of the present invention. The following is a detailed description of this method.

[0009] Step S110: Obtain multi-frame thermal radiation distribution maps of the baking surface area under continuous time series collected by the thermal imaging acquisition unit.

[0010] In this embodiment, the thermal imaging acquisition unit employs an infrared thermal imager, which continuously captures images of the baking surface area at a preset acquisition frequency, thereby obtaining multiple frames of thermal radiation distribution maps in a continuous time series. Each frame of the thermal radiation distribution map contains thermal radiation information for various locations within the baking surface area.

[0011] Step S111: Receive the initial thermal radiation distribution map sequence acquired by the thermal imaging acquisition unit in a continuous time series, extract the acquisition time point corresponding to each frame of the initial thermal radiation distribution map in the initial thermal radiation distribution map sequence, and identify the missing frame positions where the acquisition time point interval does not meet the preset time interval requirement according to the preset acquisition frequency requirement.

[0012] During actual data acquisition, due to equipment malfunctions, environmental interference, and other reasons, the initial thermal radiation distribution map sequence may contain missing frames. To ensure data continuity and integrity, the initial thermal radiation distribution map sequence needs to be processed. First, the acquisition time point corresponding to each frame of the initial thermal radiation distribution map is extracted from the sequence; for example, if the acquisition time point of one frame is T1, the next frame is T2. Then, the interval between these acquisition time points is compared with a preset acquisition time interval. If the interval between two adjacent acquisition time points is greater than the preset time interval, it indicates that there are missing frames between these two time points.

[0013] Step S112: Based on the position index of the missing frame in the initial thermal radiation distribution map sequence, obtain the initial thermal radiation distribution map of the frame before and the frame after the missing frame, extract the thermal radiation intensity value of each pixel unit in the initial thermal radiation distribution map of the frame before as a first intensity value set, and extract the thermal radiation intensity value of each pixel unit in the initial thermal radiation distribution map of the frame after as a second intensity value set.

[0014] Once the location index of the missing frame is determined, the initial thermal radiation distribution maps of the preceding and following frames can be found. For example, if the location index of the missing frame is k, then the preceding frame is k-1 and the following frame is k+1. Next, the thermal radiation intensity value of each pixel unit in the preceding and following frames is extracted, forming a first intensity value set and a second intensity value set. Each intensity value set is a two-dimensional array, where each element represents the thermal radiation intensity of the corresponding pixel unit.

[0015] Step S113: Calculate the rate of change of thermal radiation intensity of each pixel unit based on the difference between the thermal radiation intensity values ​​of the corresponding pixel units in the first set of intensity values ​​and the second set of intensity values, and the time interval between the acquisition time of the initial thermal radiation distribution map of the previous frame and the acquisition time of the initial thermal radiation distribution map of the next frame.

[0016] For each pixel, the difference in thermal radiation intensity is obtained by subtracting the corresponding pixel's thermal radiation intensity value from the first intensity value set. Then, this difference is divided by the time interval between the acquisition points of the previous and next frames to obtain the rate of change of thermal radiation intensity for each pixel. For example, if a pixel has value A in the first intensity value set and value B in the second intensity value set, with a time interval of Δt, then the rate of change of thermal radiation intensity for that pixel is (B - A) / Δt.

[0017] Step S114: Based on the time difference between the missing acquisition time point corresponding to the missing frame position and the acquisition time point of the initial thermal radiation distribution map of the previous frame, and the rate of change of thermal radiation intensity of each pixel unit, calculate the interpolated thermal radiation intensity value of each pixel unit at the missing acquisition time point, and generate an interpolated thermal radiation distribution map as a supplementary frame based on the interpolated thermal radiation intensity value.

[0018] Given that the time difference between the missing acquisition time point and the previous frame acquisition time point is Δt', for each pixel unit, the interpolated thermal radiation intensity value at the missing acquisition time point is obtained by multiplying the rate of change of thermal radiation intensity of that pixel unit by Δt' and adding it to the thermal radiation intensity value of that pixel unit in the first intensity value set. Combining the interpolated thermal radiation intensity values ​​of all pixel units generates an interpolated thermal radiation distribution map, which is then inserted as a supplementary frame into the missing frame position of the initial thermal radiation distribution map sequence.

[0019] Step S115: Insert the interpolated thermal radiation distribution map into the missing frame position of the initial thermal radiation distribution map sequence to generate a continuous multi-frame thermal radiation distribution map with uniform time intervals.

[0020] After obtaining the supplementary frames through the above interpolation process, they are inserted into the corresponding missing frame positions in the initial thermal radiation distribution map sequence, so that the thermal radiation distribution maps in the entire sequence have uniform intervals in time.

[0021] Step S120: Perform spatial semantic segmentation processing based on the baking surface region on the multi-frame thermal radiation distribution map to generate partition thermal radiation field data corresponding to each baking sub-partition.

[0022] After obtaining continuous multi-frame thermal radiation distribution maps, spatial semantic segmentation is required to divide the baking surface area into different baking sub-zones and acquire the thermal radiation field data of each sub-zone. This step is crucial for achieving precise temperature control of the zones, as semantic segmentation can accurately identify the location and extent of different sub-zones.

[0023] Step S121: Obtain the thermal radiation intensity distribution matrix of each frame of the multi-frame thermal radiation distribution map, and delineate the boundary contour line of the baking surface area according to the intensity value distribution of the pixel units in the thermal radiation intensity distribution matrix, and determine the area inside the boundary contour line as the baking surface area range.

[0024] Each frame of the thermal radiation distribution map can be represented as a thermal radiation intensity distribution matrix, where each element corresponds to the thermal radiation intensity value of a pixel unit. By analyzing the intensity value distribution of the pixel units in this matrix, the boundary of the baked surface region can be found. For example, an intensity threshold can be set, and pixel units with intensity values ​​higher than the threshold can be considered as part of the baked surface region. Then, an edge detection algorithm can be used to extract the boundary contour line of the region formed by these pixel units, and the area inside the boundary contour line is the range of the baked surface region.

[0025] Step S122: Perform superpixel segmentation processing on the baked surface area based on spatial continuity and consistency of thermal radiation intensity to obtain multiple initial segmentation units. Each initial segmentation unit contains a set of connected pixels with similar thermal radiation intensity values.

[0026] Superpixel segmentation is the process of dividing an image into pixel blocks with similar characteristics. In this step, superpixel segmentation is performed based on the spatial continuity and thermal radiation intensity consistency of pixel units within the baked surface region. Specifically, the spatial distance and thermal radiation intensity difference between each pixel unit and its surrounding pixel units are first calculated. Then, pixel units that are spatially close and have small thermal radiation intensity differences are grouped together to form multiple initial segmentation units. Each initial segmentation unit is a connected set of pixels, in which the pixel units have tending-to-consistent thermal radiation intensity values.

[0027] Step S123: Input the multiple initial segmentation units into the semantic segmentation network for category classification prediction. The semantic segmentation network outputs the category label of the baked sub-region to which each initial segmentation unit belongs, and merges the initial segmentation units belonging to the same baked sub-region category label to generate the corresponding baked sub-region partition mask image.

[0028] Semantic segmentation networks employ deep learning-based architectures, such as the U-Net network. The network takes multiple initial segmentation units as input, processing each unit as a whole. By extracting and classifying features from the initial segmentation units, the network outputs a category label for the baked sub-region to which each initial segmentation unit belongs. Then, initial segmentation units with the same category label are merged to form the corresponding baked sub-region, and a sub-region mask image is used to represent the location and extent of this region. In the sub-region mask image, pixels belonging to the baked sub-region are labeled 1, and other pixels are labeled 0.

[0029] Step S124: Based on the spatial coordinates of each pixel unit in the partitioned mask image, extract the thermal radiation intensity value of the pixel unit located inside the partitioned mask image from the thermal radiation distribution map of the corresponding frame, and generate a set of thermal radiation intensity values ​​for each baking sub-partition in the corresponding frame.

[0030] The partitioned mask image provides the location information of each baking sub-partition in the thermal radiation distribution map. Based on the spatial coordinates of the pixel units marked as 1 in the mask image, these pixel units are located in the thermal radiation distribution map of the corresponding frame, and their thermal radiation intensity values ​​are extracted. These values ​​form the set of thermal radiation intensity values ​​for each baking sub-partition in the corresponding frame.

[0031] Step S125: Perform time-series correlation modeling on the set of multi-frame thermal radiation intensity values ​​of each baking sub-partition in a continuous time series to obtain partition thermal radiation field data describing the change law of thermal radiation intensity of each baking sub-partition in the time dimension.

[0032] For each baking sub-region, its multi-frame thermal radiation intensity values ​​in a continuous time series are arranged chronologically. Then, temporal correlation modeling methods, such as time series analysis and Markov chains, are used to analyze the variation of thermal radiation intensity values ​​over time. For example, the mean, variance, and trend of thermal radiation intensity values ​​at different time points can be calculated, as well as the correlation between thermal radiation intensity values ​​at adjacent time points, thereby obtaining the sub-region's thermal radiation field data. This data can comprehensively reflect the thermal radiation characteristics of the baking sub-region in the time dimension.

[0033] Step S126: Obtain the acquisition timestamp corresponding to each frame of the multi-frame thermal radiation distribution map, and construct a thermal radiation distribution map sequence according to the order of acquisition timestamps.

[0034] Each frame of the thermal radiation distribution map is accompanied by a corresponding acquisition timestamp. These thermal radiation distribution maps are arranged in chronological order according to their acquisition timestamps to form an ordered sequence of thermal radiation distribution maps.

[0035] Step S127: Extract two frames of thermal radiation distribution maps with adjacent timestamps from the thermal radiation distribution map sequence, calculate the difference in thermal radiation intensity values ​​of corresponding pixel units in the two frames of thermal radiation distribution maps, and generate a difference thermal radiation distribution map.

[0036] Select two adjacent frames of thermal radiation distribution maps from the sequence, for example, frame i and frame i+1. For each corresponding pixel unit in these two frames, calculate the difference in their thermal radiation intensity values, which is the intensity value of the pixel unit in frame i+1 minus the intensity value of the corresponding pixel unit in frame i. Combining the differences of all pixel units generates the difference thermal radiation distribution map.

[0037] Step S128: Perform connected component analysis on the difference thermal radiation distribution map to identify connected regions where the difference in thermal radiation intensity values ​​exceeds a preset difference threshold as regions with significant changes in thermal radiation.

[0038] Connected component analysis is a method for identifying connected regions with similar characteristics in an image. In a differential thermal radiation distribution map, a preset difference threshold is set, and pixel units whose thermal radiation intensity differences exceed this threshold are considered pixel units with significant thermal radiation changes. Then, a connected component analysis algorithm is used to connect these pixel units into connected regions, which are the regions with significant thermal radiation changes.

[0039] Step S129: Extract the spatial coordinates of the region with significant thermal radiation changes, and locate the baking sub-region that overlaps with the region with significant thermal radiation changes in the thermal radiation distribution map of the corresponding frame according to the spatial coordinates, and mark the baking sub-region as a thermal radiation fluctuation region.

[0040] Obtain the spatial coordinates of all pixel units in the region of significant thermal radiation change, and then find the baked sub-regions that overlap with these coordinates in the thermal radiation distribution map of the corresponding frame. Since the baked sub-regions are determined by the partition mask image, coordinate matching can be used to determine which baked sub-regions overlap with the region of significant thermal radiation change, and these baked sub-regions are marked as thermal radiation fluctuation regions.

[0041] Step S1210: When generating the partition thermal radiation field data corresponding to each baking sub-partition, add a fluctuation marker to the partition thermal radiation field data of the thermal radiation fluctuation partition. The fluctuation marker includes the fluctuation occurrence timestamp and fluctuation intensity characterization parameters.

[0042] During the generation of partitioned thermal radiation field data, for the baking sub-partition marked as a thermal radiation fluctuation partition, fluctuation markers are added to its corresponding partitioned thermal radiation field data. The timestamp of the fluctuation occurrence is the acquisition timestamp corresponding to the area with significant thermal radiation change. The fluctuation intensity characterization parameter can be determined based on the magnitude of the difference in thermal radiation intensity values, for example, using the mean or maximum value of the difference as the fluctuation intensity characterization parameter. By adding fluctuation markers, thermal radiation fluctuation partitions can be specially processed in subsequent heating deviation analysis.

[0043] Step S130: Analyze the heating deviation of each baking sub-zone based on the partitioned thermal radiation field data and the working status parameters of the heating execution unit preset in each baking sub-zone, and generate a compensation heating requirement command for each baking sub-zone.

[0044] The zoned thermal radiation field data reflects the actual thermal radiation of each baking sub-zone, while the operating status parameters of the heating execution unit reflect the current heating status. By comparing and analyzing the two, it is possible to determine whether there is a heating deviation in each baking sub-zone and generate corresponding compensating heating demand commands to adjust the operating status of the heating execution unit so that the temperature of the baking sub-zone reaches the preset target.

[0045] Step S131: Obtain the set of thermal radiation intensity values ​​at the current moment in the thermal radiation field data corresponding to each baking sub-zone, and calculate the statistical average of all thermal radiation intensity values ​​in the set of thermal radiation intensity values ​​at the current moment as the current partition characterization temperature parameter of each baking sub-zone.

[0046] The set of thermal radiation intensity values ​​for each baking sub-region at the current moment is extracted from the regional thermal radiation field data, and then the statistical average of all thermal radiation intensity values ​​in the set is calculated. This average value comprehensively reflects the overall temperature level of the baking sub-region at the current moment and is used as the temperature parameter characterizing the current sub-region.

[0047] Step S132: Obtain the preset target temperature range corresponding to each baking sub-zone, and compare the current zone's characteristic temperature parameter with the upper and lower limits of the preset target temperature range. When the current zone's characteristic temperature parameter is lower than the lower limit of the preset target temperature range, determine that each baking sub-zone is in an under-temperature state and record the under-temperature difference value. When the current zone's characteristic temperature parameter is higher than the upper limit of the preset target temperature range, determine that each baking sub-zone is in an over-temperature state and record the over-temperature difference value.

[0048] Each baking sub-zone has a preset target temperature range, which is determined based on factors such as the type of food being baked and the baking stage. The current sub-zone's characteristic temperature parameter is compared with the upper and lower limits of the target temperature range. If the current temperature parameter is lower than the lower limit, the sub-zone is under-temperatured, and the under-temperature difference is the lower limit minus the current temperature parameter. If the current temperature parameter is higher than the upper limit, the sub-zone is over-temperatured, and the over-temperature difference is the current temperature parameter minus the upper limit.

[0049] Step S133: Obtain the current operating status parameters of the heating execution unit corresponding to each baking sub-zone. The current operating status parameters include the current output power value and the current duty cycle adjustment value. Calculate the required power adjustment amount for each baking sub-zone based on the under-temperature difference value or over-temperature difference value and the current operating status parameters.

[0050] The current operating status parameters of the heating actuator directly affect its heating effect. The current output power value represents the energy output by the heating actuator per unit time, while the current duty cycle adjustment value reflects the proportion of the heating actuator's working time within one cycle. Based on the magnitude of the under-temperature or over-temperature difference and the current operating status parameters, the required power adjustment is calculated using a specific algorithm. For example, in an under-temperature state, the required power adjustment is positive to increase the output power; in an over-temperature state, the required power adjustment is negative to decrease the output power.

[0051] Step S134: Determine the adjustment direction of the heating execution unit according to the positive or negative direction of the demand power adjustment amount. When the demand power adjustment amount is positive, it is determined to be the power increase direction. When the demand power adjustment amount is negative, it is determined to be the power decrease direction. Combine the absolute value of the demand power adjustment amount and the adjustment direction to generate a compensation heating demand instruction for each baking sub-zone. The compensation heating demand instruction includes the demand power increase value or demand power decrease value and the corresponding adjustment duration parameter.

[0052] The direction of the power demand adjustment determines whether the heating actuator needs to increase or decrease power. A positive value indicates an increase in power, and the compensation heating demand command includes the increased power demand. A negative value indicates a decrease in power, and the command includes the decreased power demand. Simultaneously, the adjustment duration parameter needs to be determined. This parameter is based on factors such as the magnitude of the power demand adjustment and the response characteristics of the heating actuator to ensure that the expected temperature adjustment effect is achieved within this duration.

[0053] Step S135: Associate and store the compensation heating demand instruction for each baking sub-zone with the corresponding baking sub-zone identifier to generate a set of compensation heating demand instructions indexed by the baking sub-zone identifier.

[0054] To facilitate subsequent coordinated control and processing, the compensation heating demand instructions for each baking sub-zone are associated with their identifiers, for example, by storing them in a dictionary or database to form a set of compensation heating demand instructions. This way, when a compensation instruction for a specific baking sub-zone needs to be invoked, it can be quickly found using its identifier.

[0055] Step S136: Obtain the change curve of the thermal radiation intensity value in the thermal radiation field data of each baking sub-zone under continuous time series, extract the slope change feature of the change curve within a preset time window, compare the slope change feature with a preset heating rate abnormality threshold, and mark the corresponding baking sub-zone as a heating abnormality zone when the slope change feature exceeds the heating rate abnormality threshold.

[0056] For each baking sub-zone, the variation curve of thermal radiation intensity values ​​over a continuous time series is extracted from the sub-zone's thermal radiation field data. Then, within a preset time window, the slope variation characteristics of this curve are calculated, such as the maximum value, minimum value, and rate of change of the slope. These slope variation characteristics are compared with a preset abnormal heating rate threshold. If the threshold is exceeded, it indicates that the heating rate of the baking sub-zone is abnormal, and it is marked as a heating abnormality sub-zone.

[0057] Step S137: Obtain the current output power value in the working status parameters of the heating execution unit corresponding to the abnormal heating zone, and calculate the current heating rate value based on the difference between adjacent time points in the thermal radiation intensity value in the partition thermal radiation field data of the abnormal heating zone.

[0058] For zones with abnormal temperature rise, the current output power value of the corresponding heating execution unit is obtained. Simultaneously, the current heating rate value is calculated by dividing the difference in thermal radiation intensity values ​​of that zone at adjacent time points by the time interval. For example, if the thermal radiation intensity values ​​at time points t1 and t2 (t2>t1) are I1 and I2 respectively, then the current heating rate value is (I2 - I1) / (t2 - t1).

[0059] Step S138: Input the current output power value and the current heating rate value into the power-temperature rise relationship model. The power-temperature rise relationship model outputs a predicted temperature rise rate value. Compare the current heating rate value with the predicted temperature rise rate value. When the current heating rate value is greater than the predicted temperature rise rate value, determine that the abnormal heating zone is in an overshoot heating state and generate a power reduction compensation requirement.

[0060] The power-temperature rise relationship model is trained using a large amount of experimental data. This model can predict the corresponding temperature rise rate based on the current output power value. Inputting the current output power value and the current temperature rise rate value into the model yields the predicted temperature rise rate value. If the current temperature rise rate value is greater than the predicted temperature rise rate value, it indicates that the actual temperature rise rate exceeds expectations, and this abnormal temperature rise zone is in an overshoot state, requiring a power reduction. Therefore, a power reduction compensation requirement is generated.

[0061] Step S139: When the current temperature rise rate is less than the predicted temperature rise rate, determine that the abnormal temperature rise zone is in an insufficient response state and generate a power increase compensation demand. The power increase compensation demand includes a power increase range calculated based on the difference between the predicted temperature rise rate and the current temperature rise rate.

[0062] If the current temperature rise rate is less than the predicted temperature rise rate, it indicates that the actual temperature rise rate is lower than expected, and the zone with abnormal temperature rise is in an under-response state, requiring an increase in power. The power increase is calculated based on the difference between the predicted temperature rise rate and the current temperature rise rate; for example, the larger the difference, the larger the power increase.

[0063] Step S1310: Generate a compensation heating requirement instruction for the abnormal temperature zone based on the power reduction compensation requirement or the power increase compensation requirement.

[0064] Based on the power reduction compensation demand or the power increase compensation demand, and combined with the method of determining the adjustment direction, the required power adjustment amount, and the adjustment duration parameters in the previous steps, the compensation heating demand command for the abnormal temperature zone is generated.

[0065] Step S140: Based on the compensation heating demand commands of all baking sub-zones, perform coordinated control processing of the heating execution unit to obtain the partition power adjustment signal for each baking sub-zone.

[0066] Since the heating execution units of each baking sub-zone may share the same power supply or have mutual influence, it is necessary to coordinate and regulate the compensation heating demand commands of all baking sub-zones to ensure the stable operation of the overall heating system and the reasonable distribution of energy, and finally obtain the partition power adjustment signal of each baking sub-zone.

[0067] Step S141: Analyze the demand power increase or decrease value and adjustment duration parameter in the compensation heating demand instruction of each baking sub-zone, and combine the rated power upper limit value and rated power lower limit value of the heating execution unit corresponding to each baking sub-zone to determine whether the heating execution unit is within the rated power range after executing the demand power increase or decrease value within the adjustment duration parameter.

[0068] First, the increased or decreased power demand value and the adjustment duration parameter are parsed from the compensated heating demand command of each baking sub-zone. Then, the upper and lower limits of the rated power are obtained for each heating execution unit. The increased power demand value (or decreased power demand value) is added to the current output power value to obtain the adjusted power value. It is then determined whether the adjusted power value is between the upper and lower limits of the rated power, and whether it will remain within this range throughout the adjustment duration parameter.

[0069] Step S142: When the determination result is that the heating execution unit exceeds the rated power limit after executing the demand power increase value, a limited power increase value is generated based on the difference between the rated power limit value and the current output power value of the heating execution unit, and the limited power increase value replaces the original demand power increase value to regenerate the corrected compensated heating demand command.

[0070] If the increased power demand would exceed the rated power limit, the original demand cannot be adjusted. In this case, the limited power increase is the rated power limit minus the current output power. This limited power increase replaces the original power demand increase, and a revised compensated heating demand command is generated to ensure that the power of the heating execution unit does not exceed the rated limit.

[0071] Step S143: When the determination result is that the heating execution unit is lower than the rated power lower limit after executing the demand power reduction value, a restricted power reduction value is generated based on the difference between the current output power value of the heating execution unit and the rated power lower limit value, and the restricted power reduction value replaces the original demand power reduction value to regenerate the corrected compensation heating demand command.

[0072] Similarly, if the power demand reduction would result in a power lower than the rated power lower limit, the restricted power reduction value is the current output power value minus the rated power lower limit. This restricted power reduction value replaces the original power demand reduction value, generating a corrected compensation heating demand command to ensure that the power does not fall below the rated lower limit.

[0073] Step S144: Obtain the corrected compensated heating demand instructions corresponding to all baking sub-zones, and identify the baking sub-zones containing increased demand power values ​​as heating control zones, and identify the baking sub-zones containing decreased demand power values ​​as cooling control zones.

[0074] After receiving all the corrected compensation heating demand instructions, the baking sub-zone is divided into a heating control zone and a cooling control zone, depending on whether the instruction contains an increase or decrease in demand power, so that they can be processed separately.

[0075] Step S145: Based on the difference between the sum of the power demand increases of the heating control zones and the sum of the power demand decreases of the cooling control zones, the power demand increases of the heating control zones are proportionally scaled and adjusted so that the difference between the adjusted sum of the power demand increases of the heating control zones and the sum of the power demand decreases of the cooling control zones is within a preset power balance threshold range, thus obtaining the final power demand adjustment amount for each baking sub-zone.

[0076] Calculate the difference between the sum of the power demand increases in the heating control zones and the sum of the power demand decreases in the cooling control zones. If this difference exceeds a preset power balance threshold, the power demand increases in the heating control zones need to be scaled proportionally. For example, multiply the power demand increase in each heating control zone by a scaling factor so that the difference between the adjusted sum and the sum in the cooling control zones is within the threshold range, thus obtaining the final power demand adjustment for each baking sub-zone.

[0077] Step S146: Generate a partition power adjustment signal for each baking sub-partition based on the final required power adjustment amount and the corresponding adjustment duration parameter for each baking sub-partition. The partition power adjustment signal includes the output power target value and duty cycle adjustment waveform parameter.

[0078] The final required power adjustment is added to the current output power value to obtain the target output power value. Combined with the adjustment duration parameter and duty cycle adjustment waveform parameter (such as square wave, sine wave, etc.), a zoned power adjustment signal is generated. The duty cycle adjustment waveform parameter determines the proportion of the working time of the heating execution unit within one cycle. By adjusting the duty cycle, precise control of the output power can be achieved.

[0079] Step S147: Obtain the increase or decrease in demand power from the compensation heating demand instructions corresponding to all baking sub-zones, and calculate the absolute value of the power difference between the sum of all demand power increases and the sum of all demand power decreases.

[0080] In the process of coordinated regulation, in addition to the aforementioned proportional scaling adjustment, it is also necessary to consider the overall power balance. The absolute value between the sum of all increases in demand power and the sum of all decreases in demand power is calculated; this absolute value reflects the degree of overall power imbalance.

[0081] Step S148: When the absolute value of the power difference exceeds the preset total power balance threshold, obtain the thermal response time constant of the heating execution unit corresponding to each baking sub-zone, and divide each baking sub-zone into a fast response zone and a slow response zone according to the thermal response time constant.

[0082] The thermal response time constant is a parameter that measures the speed at which the heating actuator responds to power changes. The smaller the time constant, the faster the response. When the absolute value of the power difference exceeds the total power balance threshold, the baking sub-zone is divided into a fast response zone (smaller time constant) and a slow response zone (larger time constant) based on the thermal response time constant.

[0083] Step S149: The absolute value of the power difference is allocated to the fast response partition and the slow response partition according to a preset power allocation ratio, wherein the power adjustment amount allocated to the fast response partition accounts for a first allocation ratio of the absolute value of the power difference, and the power adjustment amount allocated to the slow response partition accounts for a second allocation ratio of the absolute value of the power difference, and the first allocation ratio is greater than the second allocation ratio.

[0084] To quickly balance overall power, the absolute value of the power difference is allocated to fast-response and slow-response zones according to a preset ratio. Since the fast-response zone can respond to power adjustments more quickly, the first allocation ratio is greater than the second allocation ratio to achieve power balance more rapidly.

[0085] Step S1410: For a fast response zone that has been allocated a power adjustment amount, the allocated power adjustment amount is superimposed on the original increase or decrease in demand power to generate the superimposed fast response zone compensation demand.

[0086] The power adjustment amount allocated to the fast response zone is superimposed on its original increase or decrease in demand power to obtain the superimposed compensation demand, thereby achieving rapid adjustment of the overall power.

[0087] Step S1411: For slow response zones that have been allocated power adjustment amounts, the allocated power adjustment amounts are superimposed on the original increase or decrease in demand power to generate the superimposed slow response zone compensation demand.

[0088] Similarly, the power adjustment allocated to the slow-response partition is superimposed on its original demand to generate the superimposed compensation demand.

[0089] Step S1412: Generate a partition power adjustment signal for each baking sub-partition based on the superimposed fast response partition compensation requirements and the superimposed slow response partition compensation requirements.

[0090] Taking into account both the fast response partition compensation requirements and the slow response partition compensation requirements after superposition, and following the method for generating partition power adjustment signals mentioned earlier, the partition power adjustment signal for each baking sub-partition is finally determined.

[0091] Step S150: Send the partition power adjustment signal to the corresponding heating execution unit to drive each baking sub-partition to perform differentiated heating operations.

[0092] The generated partition power adjustment signal is sent to the corresponding heating execution unit through the communication interface. The heating execution unit adjusts its own output power and working status according to the signal, thereby realizing differentiated heating of each baking sub-partition and making the temperature of each sub-partition reach the preset target.

[0093] Step S151: Within a preset feedback period after sending the partition power adjustment signal, reacquire the multi-frame thermal radiation distribution map within the feedback period acquired by the thermal imaging acquisition unit for the baked surface area.

[0094] After sending the zone power adjustment signal, feedback monitoring of the heating effect is required. Within a preset feedback period, the thermal imaging acquisition unit re-acquires multiple frames of thermal radiation distribution maps of the baking surface area to evaluate the response of the heating execution unit to the adjustment signal.

[0095] Step S152: Perform spatial semantic segmentation processing based on the baking surface region on the multi-frame thermal radiation distribution map within the feedback period to generate partition thermal radiation field data within the feedback period corresponding to each baking sub-partition.

[0096] Using the same spatial semantic segmentation method as in step S120, the multi-frame thermal radiation distribution map within the feedback period is processed to obtain the thermal radiation field data of each baking sub-partition within the feedback period.

[0097] Step S153: Calculate the temperature change rate of each baking sub-zone within the feedback period based on the partitioned thermal radiation field data within the feedback period, and compare the temperature change rate with a preset temperature change rate threshold. When the temperature change rate is lower than the preset temperature change rate threshold, mark the corresponding baking sub-zone as a response hysteresis zone.

[0098] The changes in thermal radiation intensity are extracted from the thermal radiation field data of the zones within the feedback period, and the rate of temperature change is calculated. This rate is compared with a preset temperature change rate threshold. If it is lower than the threshold, it indicates that the baking sub-zone responds slowly to the power adjustment signal and is marked as a lag zone.

[0099] Step S154: Extract the partition identifier of the response lag partition and the historical power adjustment signal record of the heating execution unit corresponding to the response lag partition. Input the partition identifier and the historical power adjustment signal record into the control parameter optimization model for control parameter correction processing to obtain the optimized partition power adjustment signal for the response lag partition.

[0100] The system acquires the identifiers of the lag zones and their corresponding historical power regulation signal records for the heating actuators. These records include past power regulation signals and their corresponding temperature changes. This data is then input into a control parameter optimization model. This model analyzes the historical data to identify the causes of the lag and corrects the control parameters, generating optimized zone power regulation signals.

[0101] Step S155: Resend the optimized partition power adjustment signal to the heating execution unit corresponding to the response lag partition to replace the previously sent partition power adjustment signal.

[0102] The optimized partition power adjustment signal is sent to the heating execution unit corresponding to the partition with lag response, replacing the previous adjustment signal, so as to improve the heating response speed and effect of that partition.

[0103] For example, in step S156: within a preset monitoring period after sending the partition power adjustment signal, multiple frames of thermal radiation distribution maps within the monitoring period are continuously acquired by the thermal imaging acquisition unit, and spatial semantic segmentation processing based on the baking surface area is performed on the multiple frames of thermal radiation distribution maps within the monitoring period to generate partition thermal radiation field data within the monitoring period corresponding to each baking sub-partition.

[0104] In addition to monitoring the feedback cycle, it is also necessary to continuously monitor the thermal radiation of the baking sub-zone within a preset monitoring cycle. Multiple frames of thermal radiation distribution maps are acquired through a thermal imaging acquisition unit within the monitoring cycle, and spatial semantic segmentation is performed to obtain the zoned thermal radiation field data within the monitoring cycle.

[0105] Step S157: Extract the fluctuation amplitude parameter of the thermal radiation intensity value in the monitoring period of the monitoring period corresponding to each baking sub-zone, and compare the fluctuation amplitude parameter with the preset fluctuation amplitude threshold. When the fluctuation amplitude parameter exceeds the fluctuation amplitude threshold, mark the corresponding baking sub-zone as a thermal radiation fluctuation exceeding the limit zone.

[0106] The fluctuation of thermal radiation intensity values ​​is extracted from the regional thermal radiation field data within the monitoring period, and fluctuation amplitude parameters, such as variance and maximum difference, are calculated. This parameter is compared with a preset fluctuation amplitude threshold. If it exceeds the threshold, it indicates that the thermal radiation fluctuation of the baking sub-zone is too large, and it is marked as a zone with excessive thermal radiation fluctuation.

[0107] Step S158: Obtain the output power target value and duty cycle adjustment waveform parameters in the partition power adjustment signal corresponding to the partition with excessive thermal radiation fluctuation. Based on the fluctuation frequency parameters of the thermal radiation intensity value in the partition thermal radiation field data of the partition with excessive thermal radiation fluctuation within the monitoring period, adjust the frequency of the duty cycle adjustment waveform parameters to generate a partition power adjustment signal with adjusted frequency.

[0108] The fluctuation frequency parameter of the zone where thermal radiation fluctuations exceed limits is analyzed, which is the period of fluctuation in thermal radiation intensity value. Then, the frequency of the duty cycle adjustment waveform parameter is adjusted according to this fluctuation frequency to match the fluctuation frequency, thereby suppressing thermal radiation fluctuations. For example, if the fluctuation frequency is high, the duty cycle adjustment frequency is appropriately increased to make heating more uniform.

[0109] Step S159: Send the frequency-adjusted partition power adjustment signal to the heating execution unit corresponding to the thermal radiation fluctuation exceeding limit partition to replace the previously sent partition power adjustment signal, and re-execute the monitoring operation within the preset monitoring cycle after sending, until the fluctuation amplitude parameter of the thermal radiation fluctuation exceeding limit partition is lower than the fluctuation amplitude threshold, and then stop the adjustment operation.

[0110] The frequency-adjusted partition power regulation signal is sent to the heating execution unit corresponding to the partition with excessive thermal radiation fluctuation, and monitoring continues. If the fluctuation amplitude parameter still exceeds the threshold, the above adjustment process is repeated until the fluctuation amplitude parameter is lower than the threshold to ensure the temperature stability of the baking sub-partition.

[0111] Figure 2 The illustration shows exemplary hardware and software components of a precise temperature control system 100 for baking areas incorporating visual semantic segmentation, which can implement the ideas of this application, according to some embodiments of this application. For example, a processor 120 can be used in the precise temperature control system 100 for baking areas incorporating visual semantic segmentation and to perform the functions in this application.

[0112] The precise temperature control system 100 for baking dough partitions, which combines visual semantic segmentation, can be a general-purpose server or a special-purpose server. Both can be used to implement the precise temperature control method for baking dough partitions based on visual semantic segmentation of this application. Although only one server is shown in this application, for convenience, the functions described in this application can be implemented in a distributed manner on multiple similar platforms to balance the load.

[0113] For example, the toaster zone precise temperature control system 100 incorporating visual semantic segmentation may include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and various forms of storage media 140, such as a disk, ROM, or RAM, or any combination thereof. Exemplarily, the toaster zone precise temperature control system 100 incorporating visual semantic segmentation may also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The methods of this application can be implemented according to these program instructions. The toaster zone precise temperature control system 100 incorporating visual semantic segmentation also includes an I / O interface 150 between the computer and other input / output devices.

[0114] For ease of explanation, only one processor is described in the toaster zone precise temperature control system 100 incorporating visual semantic segmentation. However, it should be noted that the toaster zone precise temperature control system 100 incorporating visual semantic segmentation may also include multiple processors. Therefore, the steps executed by one processor as described in this application may also be executed jointly or individually by multiple processors. For example, if the processor of the toaster zone precise temperature control system 100 incorporating visual semantic segmentation executes steps A and B, it should be understood that steps A and B may also be executed jointly by two different processors or individually by one processor. For example, the first processor executes step A, the second processor executes step B, or the first processor and the second processor jointly execute steps A and B.

[0115] Furthermore, this embodiment of the invention also provides a readable storage medium, wherein computer-executable instructions are preset in the readable storage medium, and when the processor executes the computer-executable instructions, the above-mentioned method for precise temperature control of baking dough partitions combined with visual semantic segmentation is realized.

[0116] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. A method for precise temperature control of baking dough by combining visual semantic segmentation, characterized in that, The method includes: Obtain multi-frame thermal radiation distribution maps of the baking surface area in a continuous time series, acquired by the thermal imaging acquisition unit. Perform spatial semantic segmentation processing based on the baking surface region on the multi-frame thermal radiation distribution map to generate partition thermal radiation field data corresponding to each baking sub-partition. Based on the partitioned thermal radiation field data and the working status parameters of the heating execution unit preset in each baking sub-partition, partitioned heating deviation analysis is performed to generate compensation heating demand instructions for each baking sub-partition. Based on the compensation heating demand commands of all baking sub-zones, the heating execution unit is coordinated and controlled to obtain the partition power adjustment signal for each baking sub-zone. The partition power adjustment signal is sent to the corresponding heating execution unit to drive each baking sub-partition to perform differentiated heating operations.

2. The precise temperature control method for baking dough by combining visual semantic segmentation according to claim 1, characterized in that, The step of performing spatial semantic segmentation processing based on the baked surface region on the multi-frame thermal radiation distribution map to generate partition thermal radiation field data corresponding to each baked sub-partition includes: Obtain the thermal radiation intensity distribution matrix of each frame of the multi-frame thermal radiation distribution map, and delineate the boundary contour line of the baking surface area according to the intensity value distribution of the pixel unit in the thermal radiation intensity distribution matrix, and determine the area inside the boundary contour line as the baking surface area range. The baked surface area is segmented using superpixel segmentation based on spatial continuity and consistency of thermal radiation intensity to obtain multiple initial segmentation units. Each initial segmentation unit contains a set of connected pixels with similar thermal radiation intensity values. The multiple initial segmentation units are input into the semantic segmentation network for category classification prediction. The semantic segmentation network outputs the category label of the baked sub-region to which each initial segmentation unit belongs, and merges the initial segmentation units belonging to the same baked sub-region category label to generate the corresponding baked sub-region partition mask image. Based on the spatial coordinates of each pixel unit in the partitioned mask image, the thermal radiation intensity value of the pixel unit located inside the partitioned mask image is extracted from the thermal radiation distribution map of the corresponding frame, and a set of thermal radiation intensity values ​​of each baking sub-partition in the corresponding frame is generated. Temporal correlation modeling is performed on the set of multi-frame thermal radiation intensity values ​​of each baking sub-region in a continuous time series to obtain the sub-region thermal radiation field data describing the variation law of thermal radiation intensity of each baking sub-region in the time dimension.

3. The precise temperature control method for baking dough partitioning based on visual semantic segmentation according to claim 2, characterized in that, The step of analyzing the heating deviation of each baking sub-zone based on the partitioned thermal radiation field data and the pre-set operating status parameters of the heating execution unit in each baking sub-zone, and generating compensation heating demand instructions for each baking sub-zone, includes: Obtain the set of thermal radiation intensity values ​​at the current moment from the thermal radiation field data corresponding to each baking sub-zone, and calculate the statistical average of all thermal radiation intensity values ​​in the set of thermal radiation intensity values ​​at the current moment as the current zone characterization temperature parameter for each baking sub-zone; Obtain the preset target temperature range corresponding to each baking sub-zone, and compare the current zone's characteristic temperature parameter with the upper and lower limits of the preset target temperature range. When the current zone's characteristic temperature parameter is lower than the lower limit of the preset target temperature range, determine that each baking sub-zone is in an under-temperature state and record the under-temperature difference value. When the current zone's characteristic temperature parameter is higher than the upper limit of the preset target temperature range, determine that each baking sub-zone is in an over-temperature state and record the over-temperature difference value. Obtain the current operating status parameters of the heating execution unit corresponding to each baking sub-zone. The current operating status parameters include the current output power value and the current duty cycle adjustment value. Calculate the required power adjustment amount for each baking sub-zone based on the under-temperature difference value or over-temperature difference value and the current operating status parameters. The adjustment direction of the heating execution unit is determined based on the positive or negative direction of the required power adjustment amount. When the required power adjustment amount is positive, it is determined to be the power increase direction, and when the required power adjustment amount is negative, it is determined to be the power decrease direction. The compensation heating demand instruction for each baking sub-zone is generated by combining the absolute value of the required power adjustment amount and the adjustment direction. The compensation heating demand instruction includes the required power increase value or the required power decrease value and the corresponding adjustment duration parameter. The compensation heating requirement instructions for each baking sub-zone are associated and stored with the corresponding baking sub-zone identifier, generating a set of compensation heating requirement instructions indexed by the baking sub-zone identifier.

4. The precise temperature control method for baking dough by combining visual semantic segmentation according to claim 3, characterized in that, The coordinated control processing of the heating execution unit based on the compensated heating demand commands of all baking sub-zones yields a partition power adjustment signal for each baking sub-zone, including: The system analyzes the increase or decrease in required power and the adjustment duration parameter in the compensation heating demand instruction of each baking sub-zone, and combines the upper limit and lower limit of the rated power of the heating execution unit corresponding to each baking sub-zone to determine whether the heating execution unit is within the rated power range after executing the increase or decrease in required power within the adjustment duration parameter. When the determination result is that the heating execution unit exceeds the rated power limit after executing the demand power increase value, a limited power increase value is generated based on the difference between the rated power limit value and the current output power value of the heating execution unit, and the limited power increase value replaces the original demand power increase value to regenerate the corrected compensation heating demand command; When the determination result is that the heating execution unit is lower than the rated power lower limit after executing the demand power reduction value, a limited power reduction value is generated based on the difference between the current output power value of the heating execution unit and the rated power lower limit value, and the limited power reduction value replaces the original demand power reduction value to regenerate the corrected compensation heating demand command. Obtain the corrected compensated heating demand instructions corresponding to all baking sub-zones, and identify the baking sub-zones containing increased demand power values ​​as heating control zones, and identify the baking sub-zones containing decreased demand power values ​​as cooling control zones. Based on the difference between the sum of the power demand increases of the heating control zones and the sum of the power demand decreases of the cooling control zones, the power demand increases of the heating control zones are proportionally scaled and adjusted so that the difference between the adjusted sum of the power demand increases of the heating control zones and the sum of the power demand decreases of the cooling control zones is within a preset power balance threshold range, thus obtaining the final power demand adjustment amount for each baking sub-zone. Based on the final required power adjustment amount and corresponding adjustment duration parameters of each baking sub-zone, a sub-zone power adjustment signal is generated for each baking sub-zone. The sub-zone power adjustment signal includes the output power target value and duty cycle adjustment waveform parameters.

5. The precise temperature control method for baking dough partitioning based on visual semantic segmentation according to claim 4, characterized in that, After sending the partition power adjustment signal to the corresponding heating execution unit to drive each baking sub-partition to perform differentiated heating operations, the method further includes: Within a preset feedback period after sending the partition power adjustment signal, the multi-frame thermal radiation distribution map of the baking surface area acquired by the thermal imaging acquisition unit within the feedback period is re-acquired. Perform spatial semantic segmentation processing based on the baking surface region on the multi-frame thermal radiation distribution map within the feedback period to generate partition thermal radiation field data within the feedback period corresponding to each baking sub-partition. The temperature change rate of each baking sub-zone is calculated based on the partition thermal radiation field data within the feedback period, and the temperature change rate is compared with a preset temperature change rate threshold. When the temperature change rate is lower than the preset temperature change rate threshold, the corresponding baking sub-zone is marked as a response hysteresis zone. Extract the partition identifier of the response lag partition and the historical power adjustment signal record of the heating execution unit corresponding to the response lag partition. Input the partition identifier and the historical power adjustment signal record into the control parameter optimization model for control parameter correction processing to obtain the optimized partition power adjustment signal for the response lag partition. The optimized partition power adjustment signal is resent to the heating execution unit corresponding to the lag partition to replace the previously sent partition power adjustment signal.

6. The precise temperature control method for baking dough by combining visual semantic segmentation according to claim 3, characterized in that, The step of analyzing the heating deviation of each baking sub-zone based on the partitioned thermal radiation field data and the pre-set operating status parameters of the heating execution unit in each baking sub-zone, and generating a compensation heating demand instruction for each baking sub-zone, further includes: Obtain the change curve of the thermal radiation intensity value in the thermal radiation field data corresponding to each baking sub-zone under continuous time series, extract the slope change feature of the change curve within a preset time window, compare the slope change feature with a preset heating rate abnormal threshold, and mark the corresponding baking sub-zone as a heating abnormal zone when the slope change feature exceeds the heating rate abnormal threshold. Obtain the current output power value from the working status parameters of the heating execution unit corresponding to the abnormal temperature rise zone, and calculate the current heating rate value based on the difference between the thermal radiation intensity values ​​in the partition thermal radiation field data of the abnormal temperature rise zone and adjacent time points. The current output power value and the current heating rate value are input into the power-temperature rise relationship model. The power-temperature rise relationship model outputs the predicted temperature rise rate value. The current heating rate value is compared with the predicted temperature rise rate value. When the current heating rate value is greater than the predicted temperature rise rate value, it is determined that the heating abnormal zone is in an overshoot heating state and a power reduction compensation requirement is generated. When the current temperature rise rate is less than the predicted temperature rise rate, the abnormal temperature rise zone is determined to be in an insufficient response state and a power increase compensation demand is generated. The power increase compensation demand includes a power increase range calculated based on the difference between the predicted temperature rise rate and the current temperature rise rate. The compensation heating requirement instruction for the abnormal temperature zone is generated based on the power reduction compensation requirement or the power increase compensation requirement.

7. The precise temperature control method for baking dough partitioning based on visual semantic segmentation according to claim 2, characterized in that, The step of performing spatial semantic segmentation processing based on the baked surface region on the multi-frame thermal radiation distribution map to generate partition thermal radiation field data corresponding to each baked sub-partition also includes: Obtain the acquisition timestamp corresponding to each frame of the multi-frame thermal radiation distribution map, and construct a thermal radiation distribution map sequence according to the order of acquisition timestamps. Extract two frames of thermal radiation distribution maps with adjacent timestamps from the thermal radiation distribution map sequence, calculate the difference in thermal radiation intensity values ​​of corresponding pixel units in the two frames of thermal radiation distribution maps, and generate a difference thermal radiation distribution map. Perform connected component analysis on the difference thermal radiation distribution map to identify connected regions where the difference in thermal radiation intensity values ​​exceeds a preset difference threshold as regions with significant changes in thermal radiation. Extract the spatial coordinates of the region with significant thermal radiation changes, and locate the baking sub-region that overlaps with the region with significant thermal radiation changes in the thermal radiation distribution map of the corresponding frame based on the spatial coordinates. Mark the baking sub-region as a thermal radiation fluctuation region. When generating the thermal radiation field data corresponding to each baking sub-partition, a fluctuation marker is added to the thermal radiation field data of the thermal radiation fluctuation partition. The fluctuation marker includes the fluctuation occurrence timestamp and fluctuation intensity characterization parameters.

8. The precise temperature control method for baking dough partitioning based on visual semantic segmentation according to claim 4, characterized in that, The method of coordinating the control and regulation of the heating execution unit based on the compensated heating demand commands of all baking sub-zones to obtain a partition power adjustment signal for each baking sub-zone also includes: Obtain the increase or decrease in demand power from the compensation heating demand instructions corresponding to all baking sub-zones, and calculate the absolute value of the power difference between the sum of all demand power increases and the sum of all demand power decreases. When the absolute value of the power difference exceeds the preset total power balance threshold, the thermal response time constant of the heating execution unit corresponding to each baking sub-zone is obtained, and each baking sub-zone is divided into a fast response zone and a slow response zone according to the thermal response time constant. The absolute value of the power difference is allocated to the fast response partition and the slow response partition according to a preset power allocation ratio. The power adjustment amount allocated to the fast response partition accounts for a first allocation ratio of the absolute value of the power difference, and the power adjustment amount allocated to the slow response partition accounts for a second allocation ratio of the absolute value of the power difference. The first allocation ratio is greater than the second allocation ratio. For a fast response zone that has been allocated a power adjustment amount, the allocated power adjustment amount is superimposed on the original increase or decrease in demand power to generate the superimposed fast response zone compensation demand. For slow response zones that have been allocated power adjustment amounts, the allocated power adjustment amounts are superimposed on the original increase or decrease in demand power to generate the superimposed slow response zone compensation demand. Based on the superimposed fast response partition compensation requirements and the superimposed slow response partition compensation requirements, a partition power adjustment signal is generated for each baking sub-partition.

9. The precise temperature control method for baking dough partitioning based on visual semantic segmentation according to claim 1, characterized in that, The acquisition of multi-frame thermal radiation distribution maps of the baked surface area under continuous time series acquired by the thermal imaging acquisition unit includes: The system receives the initial thermal radiation distribution map sequence acquired by the thermal imaging acquisition unit in a continuous time series, extracts the acquisition time point corresponding to each frame of the initial thermal radiation distribution map in the initial thermal radiation distribution map sequence, and identifies missing frame positions where the acquisition time point interval does not meet the preset time interval requirement according to the preset acquisition frequency requirement. Based on the position index of the missing frame in the sequence of initial thermal radiation distribution maps, the initial thermal radiation distribution map of the frame before and the initial thermal radiation distribution map of the missing frame are obtained. The thermal radiation intensity value of each pixel unit in the initial thermal radiation distribution map of the frame before is extracted as a first set of intensity values, and the thermal radiation intensity value of each pixel unit in the initial thermal radiation distribution map of the frame after is extracted as a second set of intensity values. The rate of change of thermal radiation intensity of each pixel unit is calculated based on the difference between the thermal radiation intensity values ​​of the corresponding pixel units in the first set of intensity values ​​and the second set of intensity values, as well as the time interval between the acquisition time of the initial thermal radiation distribution map of the previous frame and the acquisition time of the initial thermal radiation distribution map of the next frame. Based on the time difference between the missing acquisition time point corresponding to the missing frame position and the acquisition time point of the initial thermal radiation distribution map of the previous frame, and the rate of change of thermal radiation intensity of each pixel unit, the interpolated thermal radiation intensity value of each pixel unit at the missing acquisition time point is calculated, and an interpolated thermal radiation distribution map is generated as a supplementary frame based on the interpolated thermal radiation intensity value. The interpolated thermal radiation distribution map is inserted into the missing frame positions of the initial thermal radiation distribution map sequence to generate a continuous multi-frame thermal radiation distribution map with uniform time intervals.

10. A precise temperature control system for baking dough with visual semantic segmentation, characterized in that, The precise temperature control system for baking zones combined with visual semantic segmentation includes a processor and a memory. The memory and the processor are connected. The memory is used to store programs, instructions, or code. The processor is used to execute the programs, instructions, or code in the memory to implement the precise temperature control method for baking zones combined with visual semantic segmentation as described in any one of claims 1-9.

Images