Agricultural product processing content generation method based on parameter stress accumulation transmission

Abstract

The application discloses a kind of agricultural product processing content generation methods based on parameter stress accumulation transmission, it is related to agricultural product processing digital twin field.The method includes: constructing the processing link topological model including processing link node and state transmission relationship;Obtain the processing parameter of each processing link, calculate parameter deviation, parameter stress index and stress grade;Based on state transmission function, generate comprehensive state data including presequence stress accumulation influence;According to the macroscopic appearance content of comprehensive state data, microstructure content and abnormal superposition content corresponding to the processing stage are generated;According to the processing link time sequence, construct continuous evolution sequence and output final processing content;Based on stress transmission link, carry out defect tracing and parameter adjustment interval recommendation.The present application can express the cumulative transmission process of processing parameter anomaly in multiple links, synchronously generate internal organization and external forming content, improve the digital simulation, abnormal tracing and process optimization capability of agricultural product processing.

Description

Technical Field

[0001] This invention relates to the field of digital twin technology for agricultural product processing, specifically to a method for generating agricultural product processing content based on parameter stress accumulation and transmission. Background Technology

[0002] In agricultural product processing, processing parameters such as time, temperature, pressure, and concentration in different processing stages, such as soaking, pressing, braising, and drying, continuously affect the material's texture and final product. With the development of digital technology, some agricultural product processing processes are beginning to be visualized using 3D displays and virtual simulations.

[0003] However, most existing technologies use preset model switching or fixed animation playback methods for display, which cannot dynamically generate processing content for the corresponding stage based on real-time processing parameters. At the same time, existing technologies lack state transfer and anomaly accumulation mechanisms between upstream and downstream process steps, making it difficult to reflect the continuous impact of parameter anomalies on subsequent processing results. In addition, existing technologies can usually only perform simple rendering of the external morphology, failing to simultaneously express the correlation between internal structural changes and external forming anomalies, and also failing to perform process tracing and parameter optimization based on the final anomaly results.

[0004] Therefore, there is an urgent need for a method to generate agricultural product processing content that can dynamically generate processing content based on processing parameters and realize state transmission, anomaly evolution, and process traceability. Summary of the Invention

[0005] The purpose of this invention is to provide a method for generating agricultural product processing content based on the cumulative transmission of parameter stress, in order to solve the problems in the prior art that cannot dynamically generate agricultural product processing content according to processing parameters, cannot express the cross-stage cumulative transmission process of parameter anomalies, cannot simultaneously present internal organizational changes and external forming changes, and lack the ability to trace the process source and optimize parameters based on the final forming result.

[0006] To achieve the above objectives, this invention provides a method for generating agricultural product processing content based on the cumulative transmission of parameter stress, comprising the following steps: S1. Construct a topology model of agricultural product processing links, wherein the topology model includes multiple processing link nodes and the state transmission relationship defined between each processing link node; S2. Obtain the processing parameters for each processing stage, calculate the parameter deviation based on the preset parameter range, generate state characteristic data and corresponding parameter stress data, and determine the corresponding stress level. S3. Based on the state transfer relationship, the state feature data and parameter stress data of the previous processing stage are transferred to the current processing stage to generate comprehensive state data containing the stress accumulation effect. S4. Based on the comprehensive status data, generate agricultural product processing content corresponding to the processing stage according to the preset mapping rules. The processing content includes macroscopic appearance content, microscopic organization content, and abnormal superposition content corresponding to the stress level. S5. According to the temporal relationship of each processing stage, combine the generated processing content of each stage into a continuous evolution sequence and output the final processing content.

[0007] As a further technical solution of the present invention, the topological model of the agricultural product processing links is a directed acyclic graph; the processing links nodes include one or more of the following: raw material processing links, forming control links, seasoning processing links, and post-processing links; the state transit relationship is defined on the edges of the directed acyclic graph and is used to characterize the state association between the preceding and following processing links.

[0008] As a further technical solution of the present invention, the method for determining the stress level in step S2 includes: Stress evaluation function Calculate the stress index of the i-th processing parameter, where p i μ is the real-time input value of the i-th processing parameter. i σ is the median of the standard interval. i Let ω be the radius of the tolerance interval. i The parameter sensitivity weights are α > 1; The stress index is obtained by selecting the maximum value or weighting and fusing the stress indices of each processing parameter. The stress level is then mapped to at least one of no stress, mild stress, moderate stress, and severe stress based on the comprehensive stress index.

[0009] As a further technical solution of the present invention, in step S3, comprehensive state data including the cumulative effect of stress is generated using the following state transfer function: S n =f(P n ,S n-1 E n-1 )=A·P n +B·S n-1 +C·(E n-1 S n-1 ); Among them, S n P represents the comprehensive status data of the current processing stage n. n S is the input parameter for the current processing stage. n-1 E is the state characteristic data passed from the previous processing stage. n-1 The parameters and stress data are passed from the previous processing stage, and A, B, and C are the stage characteristic coefficient matrices, C·(E n-1 Sn-1 The term is used to characterize the nonlinear coupling effect of preceding stress on the current state.

[0010] As a further technical solution of the present invention, the step S4 of generating processing content according to preset mapping rules includes at least one of the following mapping sub-functions: Macroscopic appearance mapping: The morphological parameters in the comprehensive state data drive the offset of the three-dimensional mesh vertices, and the surface material is dynamically calculated based on the color and texture parameters; Microstructure mapping: Based on the moisture content or compactness parameters in the comprehensive state data, a microstructure distribution map of the three-dimensional cross-section is generated; Superimposed stress abnormalities: When the stress level is mild or above, deformation, color abnormalities or texture damage are superimposed at the macro level, and tissue abnormalities are superimposed at the micro level.

[0011] As a further technical solution of the present invention, the macroscopic appearance mapping includes: calling the geometric deformation basis function to drive the offset of the three-dimensional mesh vertices according to the morphological parameters in the comprehensive state data to generate a stage morphological mesh; and calling the material interpolation function based on physical rendering to dynamically calculate the base color, roughness and normal map according to the color parameters and texture parameters in the comprehensive state data.

[0012] As a further technical solution of the present invention, the microstructure mapping includes: generating a microstructure distribution map of a three-dimensional cross-section based on volume rendering technology, according to the moisture content parameter or compactness parameter in the comprehensive state data, wherein the microstructure distribution map includes pore structure and sap distribution.

[0013] As a further technical solution of the present invention, the step S5 of assembling a continuous evolution sequence specifically includes: aligning the stage processing content generated by each stage according to the temporal relationship of the processing stage topology model, and performing interpolation smoothing on the state changes between adjacent stages to form a multi-stage continuous evolution content sequence.

[0014] As a further technical solution of the present invention, it also includes: S6, calculating the contribution weight of the parameters of each processing stage to the defects of the final processed content based on the transmission link of the parameter stress data of each processing stage along the topology model; and deriving the recommended optimal parameter adjustment range in reverse with the goal of the final processed content approaching the standard state.

[0015] As a further technical solution of the present invention, it also includes: S7, outputting the continuous evolution sequence, the final processing content, and the traceability analysis data generated based on the parameter stress data transmission link as visualization content data and parameter analysis data; wherein, the visualization content data is used to display the stage appearance, abnormal evolution process and final product of the target agricultural product under different parameter conditions, and the parameter analysis data includes the processing parameter values, status results, stress level, defect contribution weight and parameter adjustment suggestions for each link.

[0016] Compared with the prior art, the present invention has the following significant advantages: 1. This invention generates agricultural product processing content based on processing parameters and state transmission relationships, which can realize the continuous dynamic evolution of agricultural product processing. Compared with the traditional static model switching method, it improves the authenticity and continuity of processing content generation.

[0017] 2. This invention establishes a parameter stress quantification and anomaly generation mechanism, which can dynamically generate corresponding abnormal processing content according to the degree of deviation of processing parameters, and realize the visual expression of abnormal parameter states.

[0018] 3. This invention establishes a cross-process stress accumulation and transmission mechanism, which can characterize the continuous impact of preceding processing anomalies on subsequent processing stages, thereby realizing the continuous transmission and evolution of processing anomalies.

[0019] 4. This invention can simultaneously generate the internal structure and external forming content of agricultural products, and perform process traceability and parameter adjustment analysis based on the processing anomaly results, thereby improving the digital analysis and process optimization capabilities of agricultural product processing. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating the overall process of the method of the present invention. Figure 2 Flowchart for parameter stress quantization; Figure 3 A flowchart for cross-stage status transfer and content generation; Figure 4 Output a flowchart for 3D visualization and interaction. Detailed Implementation

[0021] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] like Figure 1-4As shown, this invention provides a method for generating agricultural product processing content based on the cumulative transmission of parameter stress, comprising the following steps: S1. Constructing a topology model for agricultural product processing. Based on the standard processing technology of the target agricultural product, a topological model is established, consisting of multiple processing stage nodes and directed edges. The processing stage topological model includes at least: stage identification information, stage execution sequence relationships, input parameter sets for each stage, state feature vectors for each stage, 3D content generation rules for each stage, output result sets for each stage, and state propagation relationships defined on the directed edges. Preferably, the processing stage topological model is a directed acyclic graph (DAG). The processing stage nodes include one or more of the following: raw material processing, shaping control, seasoning processing, and post-processing. The state propagation relationships are defined on the edges of the DAG and are used to characterize the state associations between preceding and subsequent processing stages.

[0023] S2. Obtain parameters and calculate parameter stress level. The processing parameters for each processing stage are obtained, the parameter deviation is calculated based on the preset parameter range, state characteristic data and corresponding parameter stress data are generated, and the corresponding stress level is determined. The specific steps are as follows: For each processing stage, a multi-dimensional parameter space is established, and standard process ranges, effective tolerance ranges, and failure ranges are set for each parameter. Real-time processing parameter input values ​​for each processing stage are obtained, and state characteristic data are calculated using state determination rules. Furthermore, a stress evaluation function is used to calculate the stress index of the i-th processing parameter. ; Where, p i μ is the real-time input value of the i-th processing parameter. i σ is the median of the standard interval. i Let ω be the radius of the tolerance interval. i α is the parameter sensitivity weight, and α is a nonlinear exponent with α > 1, used to characterize the nonlinear growth characteristics of the stress effect as the degree of parameter deviation increases; A comprehensive stress index is calculated, which can be determined by the maximum value or weighted fusion value of the stress indices of each parameter. The stress level is then mapped to at least one of the following categories based on the comprehensive stress index: no stress, mild stress, moderate stress, and severe stress. For example, a comprehensive stress index less than 0.3 indicates no stress, 0.3–0.6 indicates mild stress, 0.6–0.9 indicates moderate stress, and greater than 0.9 indicates severe stress.

[0024] S3, Generate the cumulative stress state Based on the aforementioned state transfer relationship, the state characteristic data and parameter stress data from the previous processing stage are transferred to the current processing stage to generate comprehensive state data that includes the cumulative effect of stress. The specific steps are as follows: Using the output of the previous stage as the initial input of the next stage, a state transfer function incorporating the cumulative effect of stress is established: S n =f(P n ,S n-1 E n-1 )=A·P n +B·S n-1 +C·(E n-1 S n-1 ); Among them, S n P represents the comprehensive status data of the current processing stage n. n S is the input parameter for the current processing stage. n-1 E is the state characteristic data passed from the previous processing stage. n-1 The parameters and stress data are passed from the previous processing stage, and A, B, and C are the stage characteristic coefficient matrices, C·(E n-1 S n-1 The term is used to characterize the nonlinear coupling effect of preceding stress on the current state.

[0025] S4, Macro- and Micro-level Content Mapping and Anomaly Overlay Based on the comprehensive status data, agricultural product processing content corresponding to the processing stage is generated according to the preset mapping rules. The processing content includes macroscopic appearance content, microscopic organization content, and abnormal superposition content corresponding to the stress level. For each processing stage, a mapping relationship is established between comprehensive status data and 3D visualization features, specifically including at least one of the following mapping sub-functions: Macroscopic appearance mapping: Based on the morphological parameters in the comprehensive state data, the 3D mesh vertex offset is driven, and the surface material is dynamically calculated based on the color and texture parameters; further, the geometric deformation basis function is called to drive the mesh vertex offset based on the morphological parameters (such as expansion rate and compression ratio) to generate a staged morphological mesh; the physically based rendering (PBR) material interpolation function is called to dynamically calculate the base color, roughness and normal map based on the color and texture parameters; Microstructure mapping: Based on the moisture content or compactness parameters in the comprehensive state data, a microstructure distribution map of the three-dimensional cross-section is generated. Further, based on volume rendering technology, a microstructure distribution map of the three-dimensional cross-section is generated according to internal parameters such as moisture content and compactness. This microstructure distribution map includes pore structure and sap distribution. Stress Anomaly Overlay: When the stress level is mild or above, the anomaly operator is triggered, which overlays deformation distortion (such as topological breakage and collapse), color anomaly (such as scorched spots and color difference noise) and texture damage (such as normal perturbation simulating cracks) at the macro level; and overlays tissue anomaly (such as excessively large pores and structural fractures) at the micro level.

[0026] S5. Construct a continuous evolution sequence and output the final content. According to the temporal relationship of each processing stage, the stage processing content generated by each stage is aligned on the timeline, and the state changes between adjacent stages are smoothed by interpolation to form a multi-stage continuous evolution content sequence; after the final stage is completed, the final formed content result is rendered and output.

[0027] S6. Defect Origin Contribution Calculation and Reverse Optimization Based on the parameter stress data of each processing stage along the transmission link of the topological model, the contribution weight of the parameters of each processing stage to the defects in the final processed content is calculated. Specifically, methods based on Shapley values ​​or partial derivative sensitivity analysis can be used to quantify the marginal contribution of each stage parameter to the final appearance defects. Furthermore, a reverse optimization objective function is constructed, aiming to make the final appearance state approach the standard state. Based on gradient descent or heuristic algorithms, the recommended optimal parameter adjustment range is derived in reverse.

[0028] S7. Output of Visualized Content Data and Parameter Analysis Data The continuous evolution sequence, final processing content, and traceability analysis data generated based on the parameter stress data transmission link are output as visualized content data and parameter analysis data. The visualized content data can be used to display the stage appearance, abnormal evolution process, and final product of the target agricultural product under different parameter conditions; the parameter analysis data includes processing parameter values, status results, stress levels, defect contribution weights, and parameter adjustment suggestions for each stage.

[0029] Example 1: System Architecture and Execution Flow for Generating Molded Content Based on Multi-Stage Parameters This embodiment uses a general-purpose digital simulation system for agricultural product processing as an example to explain in detail the core execution logic of the present invention. The system runs on a computer device with three-dimensional graphics processing capabilities, and each step is implemented through software programs.

[0030] 1. Establish a topology model (corresponding to step S1) During system initialization, a directed acyclic graph (DAG) topological model is first constructed based on the standard processing flow of the target agricultural product. Taking plant protein recombinant foods (such as vegetarian meat) as an example, its typical processing flow includes four main stages: raw material soaking, molding and pressing, seasoning and braising, and drying and post-processing. The system establishes four nodes accordingly: Node A: Raw material processing stage (soaking, softening); Node B: Molding control process (pressing, shaping); Node C: Seasoning process (braising, coloring); Node D: Post-processing stage (drying, packaging).

[0031] The directed edges between nodes define the state propagation relationship: Node A → Node B → Node C → Node D. Each node records the following information: stage identifier (ID), stage execution sequence (sequential execution), set of input parameters (such as time, temperature, pressure, concentration, etc.), state feature vector (such as moisture content, compactness, color, etc.), 3D content generation rules (geometric deformation function, material interpolation function), and set of output results.

[0032] 2. Coercion Quantification and Transmission (corresponding steps S2 to S4) In node A (raw material processing), the key parameter is set as soaking time. The standard process range is [8h, 10h], with a median of μ=9h and a tolerance radius of σ=2h (i.e., the effective range is 6h-12h; exceeding this range results in failure). The parameter sensitivity weight is ω=0.8, and the nonlinearity exponent is α=1.5 (characterizing the accelerated increase in destructive force after the parameter exceeds its limit). When the system obtains the actual input soaking time T soak When the time is 15h, first calculate the deviation: |(15-9) / 2|=3. Substitute this into the stress assessment function: SI = ω × [|(p-μ) / σ|] α =0.8×3 1.5 ≈4.156; If the value is greater than 0.9, the system classifies it as "severe stress." Simultaneously, based on preset threshold judgment rules, a state feature vector is generated: due to excessive soaking time and over-absorption of water, the state feature is "loose organization," with specific parameters of moisture content 0.85 (normal is 0.70) and compactness 0.3 (normal is 0.65). Furthermore, a stress flag E is generated. A =Severe.

[0033] The output of node A (state feature vector S) A and the coercion marker E A The state is passed to node B through a state transfer function, the expression of which is: S B =A·P B +B·S A +C·(E A S A ); Among them, P BFor node B, the input parameters are (e.g., pressing pressure, pressing time). A, B, and C are the process characteristic coefficient matrices (obtained by fitting historical process data). The third term is C·(E A S A ) represents the nonlinear coupling effect term, where ⊗ denotes the vector outer product. Since E A Under severe stress, the nonlinear coupling effect term is significantly amplified, which amplifies the instability of the pressing process during the calculation of node B, simulating the physical causal relationship that "excessive soaking leads to easy breakage of materials during pressing".

[0034] 3. Macro- and Micro-level Generation and Origin Tracing (corresponding to steps S5-S7): In step S4, the system generates processing content based on the current comprehensive state of node B (including the cumulative effect of stress) according to preset mapping rules. Specifically: the macroscopic appearance mapping first detects that the stress level is severe, triggering the anomaly operator. When generating the compressed mesh, the geometric deformation basis function superimposes random vertex offsets (amplitude 0.15mm) in the surface edge region to simulate micro-fractures. The microstructure mapping is based on volume rendering technology, generating an internal three-dimensional cross-section based on a moisture content of 0.85% and a compactness of 0.3, showing a porosity as high as 45% (normal is 15%), with irregularly connected pores.

[0035] The system continues to generate nodes C and D. Finally, in step S6, the source analysis module calculates the contribution weight of each parameter along the transmission link to the final product defect. Through partial derivative sensitivity analysis, it is found that: abnormal molding pressure contributes 40%, and the cumulative contribution of severe stress from raw material immersion is 60%. The system generates a source analysis report: "Final product edge damage (contribution: abnormal molding pressure accounts for 40%, and the cumulative contribution of severe stress from raw material immersion is 60%)."

[0036] In step S7, the graphical user interface displays a real-time animation of the continuous evolution of the plant protein recombinant food from the initial raw materials to the final product. Users can drag the timeline to view the 3D model and internal cross-section of each stage. The dashboard on the right side of the interface displays the stress level of each stage (node ​​A displays a red "severe" label).

[0037] Example 2: Stage Content Generation and Stress Transmission Driven by Suppression Parameters (Taking Dried Tofu Suppression as an Example) This embodiment takes the pressing and shaping process in tofu processing as an example, focusing on how parameters in the shaping control process drive three-dimensional morphological changes, and how stress states are quantified at the macroscopic and microscopic levels. In this embodiment, the system has pre-built a topology model containing three nodes: soaking, pressing, and braising.

[0038] Setting parameters: The main input parameters for the pressing stage include pressing pressure P (unit: MPa) and pressing time T. press (unit: min), and the initial water content W received from the previous stage (immersion). init (Dimensionless, 0–1). Standard process range: P ∈ [0.4, 0.6], T press ∈[15,25], W init The normal value is 0.70±0.05.

[0039] 1. Normal state calculation When the system obtains P=0.5MPa, T press =20min, and the W passed in the previous stage init When the value is 0.71 (within the normal range), the stress assessment result is no stress. The content of step S4 is generated as follows: Macroscopic appearance mapping: The geometric deformation basis function (using a mesh deformer based on radial basis functions) is invoked to drive the vertices of the 3D mesh to uniformly shift along the Z-axis according to the compression ratio (target thickness / original thickness = 0.6) in the shape parameters. The original tofu blank thickness is 2.5cm; after calculation, the thickness becomes 1.5cm after vertex offset, while maintaining a regular right-angled shape at the edges. Simultaneously, the PBR material interpolation function is invoked to generate a base color RGB=(0.82,0.65,0.42) based on the color parameters (normal tofu color), with a roughness set to 0.4, and the normal map remains unchanged.

[0040] Microstructure mapping: Based on volume rendering technology, a three-dimensional cross-sectional image was generated according to a compactness parameter of 0.8 (normal range 0.75–0.85) and a water content of 0.70. The image shows that the pores are evenly distributed, with the pore size concentrated in the range of 50–80 μm. The pores are connected by a continuous protein network, and the juice (water) is evenly distributed on the pore walls.

[0041] 2. Stress Anomaly Calculation Input parameters: P = 0.2 MPa, T press= 20min, W init =0.71.

[0042] Calculate the deviation: (0.2-0.5) / 0.1=-3.0 (with the standard interval half-width of 0.1 as a benchmark, the deviation is -300%).

[0043] Taking ω=0.6 (pressure sensitivity) and α=1.2, calculate the stress index: SI = 0.6 × | -3.0 | 1.2 ≈0.6×3.737=2.242, triggering "moderate stress".

[0044] Macroscopic appearance mapping: The system detected a moderate stress level. The anomaly operator adjusted the geometric deformation basis function: reducing Z-axis compression (compression ratio decreased from 0.6 to 0.88), while increasing random vertex offsets on the X / Y axes, using Gaussian noise (standard deviation 0.1mm) to simulate edge irregularities. The final mesh thickness was 2.2cm, with jagged edges and localized depressions. In PBR material interpolation, insufficient pressure resulted in residual surface moisture, causing a slightly lighter base color (RGB=(0.78,0.61,0.38)). The roughness decreased to 0.35 (the surface became more moist).

[0045] Microstructure mapping: The compactness parameter dropped from the normal 0.8 to 0.4. The system used volume rendering to generate a cross-sectional view, showing that a large area of ​​the pores were not closed, the pore diameter increased to 120-200μm, the pore walls were fractured, and the juice seeped out to form irregular liquid pools.

[0046] Scenario B: Overload Input parameters: P = 0.9 MPa, T press =20 min, W init =0.71.

[0047] Deviation: (0.9-0.5) / 0.1=4.0, SI=0.6×4.0 1.2 ≈3.167, triggering "severe stress".

[0048] Macroscopic appearance mapping: The compression ratio is set to 0.32, and the thickness after strong Z-axis compression is 0.8 cm. Anomaly operators are applied to the surface with a normal map, and a crack texture (frequency 0.2, amplitude 0.3 mm) is generated using a Perlin noise function to simulate surface cracking due to excessive compression. The base color darkens due to compression and density, RGB=(0.55,0.42,0.30), and the roughness increases to 0.7 due to increased surface friction.

[0049] Microstructure mapping: The compactness is abnormally increased to 1.2 (exceeding the normal upper limit of 1.0). Volume rendering shows that the protein fibers are broken, forming fractured structures, the pores are almost completely compressed and closed, and water is squeezed out to form local water accumulation areas.

[0050] 3. Transmission of coercion across stages In scenario A (insufficient stress, loose organization), the state feature vector S output by node B is... B =[Moisture content 0.78, Compactness 0.4] and stress indicator E B =Medium, passed to node C (seasoning processing stage) through the state transfer function. The nonlinear coupling term C·(E) in the transfer function... B ⊗S BThe system is activated. When calculating the brine penetration rate at node C, the base value of the penetration coefficient is amplified from 0.2 to 0.3. This causes the brine to penetrate the tofu too quickly and excessively during the subsequent braising process, simulating a chain reaction of "over-infusion and darkening of color." Users can observe in the visualization output that at the beginning of the braising process, the surface of the tofu darkens rapidly, reaching the normal color depth required for 10 minutes within 3 minutes, resulting in a final product with a dark brown appearance. This embodiment reproduces the chain amplification effect where preceding stress is amplified by subsequent stages in a physical causal relationship.

[0051] Example 3: Stage content generation driven by halogenation parameters (taking color and texture rendering as an example) This embodiment takes the seasoning (braising) process in dried bean curd processing as an example, and focuses on illustrating the material dynamic rendering technology based on physical rendering (PBR), as well as the reverse optimization mechanism unique to this invention.

[0052] 1. Dynamic rendering of PBR materials The system employs a physically based rendering model in the seasoning processing stage. Key parameters affecting the appearance of materials include: braising temperature T. boil (Unit: °C), brine concentration C (unit: %), referring to the mass fraction of salt and pigment components, and braising time t (unit: min). The system establishes the following mapping functions: Basic color RGB: RGB = RGB base +k1·C-k2·(T boil -T ref ) Among them, RGB base =(0.90,0.85,0.80) represents the original color before braising, T ref =80°C (standard starting temperature), k1=0.03, k2=0.005.

[0053] Roughness: Roughness = Rough base +k3·max(0,T boil -T ref ); Among them, Rough base =0.3, k3=0.02.

[0054] Normal mapping: Remains unchanged under normal conditions; when a stress occurs, a perturbation is superimposed.

[0055] Under normal operating conditions, set T boil=85°C (standard range [80,95]), C=12% (standard range [10%,15%]), braising time t=30min (standard range [25,35]). The system calculates: RGB=(1.235,1.185,1.135); After clamping and texture sampling mapping, the actual rendered color is caramel (RGB=(0.65,0.32,0.12)).

[0056] Roughness: Roughness = 0.3 + 0.02 × (85 - 80) = 0.3 + 0.1 = 0.4; The normal map was not disturbed, and the final rendered tofu surface was smooth, slightly glossy, and presented an attractive dark red color.

[0057] 2. High Temperature / Timeout Stress Calculation Set abnormal input: T boil =110°C (exceeding the upper limit of 100°C), braising time extended to 45 min (standard upper limit 35 min), C=12% remains unchanged. System calculation deviation: temperature deviation (110-95) / 5=3 (based on half-width 5), time deviation (45-35) / 5=2. Comprehensive stress index: SI≈0.5×4.655+0.5×2.462=3.5585 triggers "severe heat stress".

[0058] System execution of superimposed coercive anomalies: The base color RGB has been recalculated: RGB = (1.11, 1.06, 1.01); After tone mapping, it shifts towards dark brown and is finally rendered as (0.25, 0.12, 0.05), simulating over-charring.

[0059] Roughness: Roughness = 0.3 + 0.02 × max(0, 110 - 80) = 0.3 + 0.6 = 0.9. The surface becomes dull and rough due to water loss.

[0060] A normal map is overlaid with a local depression displacement map (amplitude 0.2mm, frequency 0.5) to simulate surface wrinkling and shrinkage deformation.

[0061] Microstructure mapping: Based on volume rendering, it shows that the surface proteins denature and coagulate to form a hard shell, while the internal moisture decreases sharply and the pores shrink.

[0062] 3. Reverse optimization mechanism When a user observes an abnormal "severely charred" appearance in the final product through visualization (e.g., RGB values ​​significantly deviating from the target value and excessive roughness), the reverse optimization function can be activated. The system will perform the following operations: First, the source analysis module pinpointed the anomaly to the brining temperature T during the seasoning process. boil And brining time t; targeting the standard appearance color feature vector: target RGB target =(0.65, 0.32, 0.12), target Roughness target =0.4, target normal perturbation amplitude=0, construct the objective function: J=w1·‖RGB-RGB target || 2 +w2·(Roughness-0.4)²+regular term; The system employs a gradient descent algorithm (learning rate 0.1, 50 iterations) to optimize within the parameter space. The initial parameters are the current outlier parameters (T=110°C, t=45min), adjusted in the opposite direction of the gradient. After calculation, the system outputs recommended parameter adjustments: "It is recommended to reduce the braising temperature by 15°C to 95°C, or shorten the braising time by 10 minutes to 35 minutes." Users can substitute these parameters into the simulation again, and the system generates the optimized appearance of the dried bean curd in real time (restoring normal caramel color and gloss). Furthermore, the system supports batch optimization: if a user adjusts multiple parameters simultaneously, the system can output a Pareto front for process engineers to select the optimal operating point.

[0063] The above three embodiments fully disclose the present invention from three dimensions: overall system architecture, deformation transfer driven by pressing parameters, and material rendering and reverse optimization driven by halogenation parameters, combined with specific numerical examples and algorithm details. Those skilled in the art should understand that the parameter values ​​(such as pressure 0.5 MPa, temperature 85°C, etc.) and parameter sensitivity weights ω in the above embodiments are... i The specific values ​​of model parameters such as the nonlinear exponent α and the characteristic coefficient matrices A, B, and C are for illustrative purposes only. In practical applications, the above model parameters can be determined through regression analysis of historical process data or conventional numerical optimization methods. Those skilled in the art can obtain values ​​suitable for specific products based on a limited number of experiments. The processing parameter range can be adjusted accordingly based on the processing requirements of different agricultural products, and all such adjustments do not depart from the protection scope of this invention.

[0064] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for generating agricultural product processing content based on parameter stress accumulation and propagation, characterized in that, Includes the following steps: S1. Construct a topology model of agricultural product processing links, wherein the topology model includes multiple processing link nodes and the state transmission relationship defined between each processing link node; The topological model of the agricultural product processing links is a directed acyclic graph; the processing links nodes include one or more of the following: raw material processing links, forming control links, seasoning processing links, and post-processing links; the state transit relationships are defined on the edges of the directed acyclic graph and are used to characterize the state associations between the preceding and following processing links. S2. Obtain the processing parameters for each processing stage, calculate the parameter deviation based on a preset parameter range, generate state characteristic data and corresponding parameter stress data, and determine the corresponding stress level. The method for determining the stress level includes: Stress evaluation function Calculate the stress index of the i-th processing parameter, where p i μ is the real-time input value of the i-th processing parameter. i σ is the median of the standard interval. i Let ω be the radius of the tolerance interval. i The parameter sensitivity weights are α > 1; The stress index is obtained by selecting the maximum value or weighting and fusing the stress indices of each processing parameter. The stress level is then mapped to at least one of no stress, mild stress, moderate stress, and severe stress based on the comprehensive stress index. S3. Based on the state transfer relationship, the state characteristic data and parameter stress data of the previous processing stage are transferred to the current processing stage to generate comprehensive state data containing the cumulative stress effect; the generation of comprehensive state data containing the cumulative stress effect adopts the following state transfer function: S n =f(P n ,S n-1 ,E n-1 )=A·P n +B·S n-1 +C·(E n-1 S n-1 ); Among them, S n P represents the comprehensive status data of the current processing stage n. n S is the input parameter for the current processing stage. n-1 E is the state characteristic data passed from the previous processing stage. n-1 The parameters and stress data are passed from the previous processing stage, and A, B, and C are the stage characteristic coefficient matrices, C·(E n-1 S n-1 The term is used to characterize the nonlinear coupling effect of preceding stress on the current state; S4. Based on the comprehensive status data, generate agricultural product processing content corresponding to the processing stage according to the preset mapping rules. The processing content includes macroscopic appearance content, microscopic organization content, and abnormal superposition content corresponding to the stress level. S5. According to the temporal relationship of each processing stage, combine the generated processing content of each stage into a continuous evolution sequence and output the final processing content.

2. The method for generating agricultural product processing content based on parameter stress accumulation and transmission according to claim 1, characterized in that, The step S4, which generates processing content according to preset mapping rules, includes at least one of the following mapping sub-functions: Macroscopic appearance mapping: The morphological parameters in the comprehensive state data drive the offset of the three-dimensional mesh vertices, and the surface material is dynamically calculated based on the color and texture parameters; Microstructure mapping: Based on the moisture content or compactness parameters in the comprehensive state data, a microstructure distribution map of the three-dimensional cross-section is generated; Superimposed stress abnormalities: When the stress level is mild or above, deformation, color abnormalities or texture damage are superimposed at the macro level, and tissue abnormalities are superimposed at the micro level.

3. The method for generating agricultural product processing content based on parameter stress accumulation and transmission according to claim 2, characterized in that, The macroscopic appearance mapping includes: calling the geometric deformation basis function to drive the offset of the three-dimensional mesh vertices according to the morphological parameters in the comprehensive state data to generate a stage morphological mesh; and calling the material interpolation function based on physical rendering to dynamically calculate the base color, roughness and normal map according to the color parameters and texture parameters in the comprehensive state data.

4. The method for generating agricultural product processing content based on parameter stress accumulation and transmission according to claim 2, characterized in that, The microstructure mapping includes: based on volume rendering technology, generating a three-dimensional cross-sectional microstructure distribution map according to the moisture content parameter or compactness parameter in the comprehensive state data, the microstructure distribution map including pore structure and sap distribution.

5. The method for generating agricultural product processing content based on parameter stress accumulation and transmission according to claim 1, characterized in that, The step S5, which involves assembling a continuous evolution sequence, specifically includes: aligning the stage processing content generated by each stage according to the temporal relationship of the processing stage topology model, and performing interpolation smoothing on the state changes between adjacent stages to form a multi-stage continuous evolution content sequence.

6. The method for generating agricultural product processing content based on parameter stress accumulation and transmission according to claim 1, characterized in that, Also includes: S6. Based on the parameter stress data of each processing stage along the transmission link of the topology model, calculate the contribution weight of the parameters of each processing stage to the defects of the final processed content; and with the goal of the final processed content approaching the standard state, deduce the recommended optimal parameter adjustment range in reverse.

7. The method for generating agricultural product processing content based on parameter stress accumulation and transmission according to claim 1, characterized in that, Also includes: S7. Output the continuous evolution sequence, final processing content, and traceability analysis data generated based on the parameter stress data transmission link as visualization content data and parameter analysis data. The visualization content data is used to display the stage appearance, abnormal evolution process, and final product of the target agricultural product under different parameter conditions. The parameter analysis data includes the processing parameter values, status results, stress level, defect contribution weight, and parameter adjustment suggestions for each stage.

Images