Intelligent meat processing production line scheduling control method and system

Abstract

The invention provides an intelligent meat processing production line scheduling control method and system, and the method comprises the steps: collecting the image information of each piece of meat raw material of a processing production line, and extracting a structure complexity index and a deterioration risk score; generating a comprehensive score through weighted fusion according to the structural complexity index and the deterioration risk score, mapping the comprehensive score into a task priority, and generating a scheduling level; collecting equipment parameters, constructing a scheduling cost function in combination with the task priority and the scheduling level so as to calculate a scheduling cost value, allocating the task to the processing station with the lowest scheduling cost value, and generating a scheduling control instruction; and calculating the actual scheduling cost after the scheduling control instruction is executed, and calculating the deviation between the comprehensive score and the actual scheduling cost as a score error measure so as to periodically update the weight parameter of the comprehensive score and optimize the subsequent scheduling decision. The production line synergy efficiency and the raw material processing utilization rate can be remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of production line scheduling, and particularly relates to an intelligent meat processing production line scheduling and control method and system. Background Technology

[0002] As the meat processing industry increasingly moves towards large-scale, standardized, and intelligent operations, traditional production scheduling methods relying on manual experience and fixed rules are no longer sufficient to meet the higher demands for food safety, production efficiency, and flexible management. Particularly in cold chain meat processing, companies face complex challenges due to individual differences in raw materials from the upstream supply chain, such as inconsistent muscle texture, varying bone embedding patterns, and fluctuating fat layer thickness. Simultaneously, they must address the diverse needs of downstream customers regarding product stability, delivery cycles, processing uniformity, and shelf-life management. This results in highly heterogeneous processing tasks in terms of physical complexity and time sensitivity. Traditional scheduling methods based on flowchart templates or process libraries cannot effectively perceive the individual state of meat products, leading to unscientific task sequencing, low efficiency in production line resource utilization, uneven product quality, and even waste of high-value-added raw materials.

[0003] Furthermore, current scheduling systems generally only optimize specific processes at a single point, lacking a closed-loop mechanism that spans "state modeling—scoring and ranking—task assignment—model calibration." They also lack the ability to address the real-time cycle time, task fluctuations, and equipment heterogeneity inherent in the specific scenario of meat processing, making it difficult to form an intelligent control system with sustainable optimization capabilities. Therefore, how to directly transform the structural information and deterioration risk status of meat products into executable task priorities for the scheduling system, and dynamically allocate equipment resources and optimize processing paths accordingly, to build a scheduling control system for the meat processing industry with information perception capabilities and closed-loop self-learning capabilities, is one of the key challenges in the field of intelligent manufacturing. Summary of the Invention

[0004] The purpose of this invention is to propose an intelligent scheduling and control method and system for meat processing production lines to solve the above-mentioned technical problems.

[0005] To achieve the above objectives, a first aspect of the present invention provides an intelligent meat processing production line scheduling and control method, the method comprising the following steps: Image information of each piece of meat raw material in the processing production line is collected, and structural complexity index and deterioration risk score are extracted; A comprehensive score is generated by weighted fusion of the structural complexity index and the degradation risk score, and then mapped to task priority to generate a scheduling level, wherein raw materials with high degradation risk are scheduled first. Collect equipment parameters, combine them with the task priority and scheduling level, construct a scheduling cost function to calculate the scheduling cost value, assign the task to the processing station with the lowest scheduling cost value, and generate scheduling control instructions; The actual scheduling cost after the scheduling control is executed is calculated, and the deviation between the comprehensive score and the actual scheduling cost is calculated as a score error measure. The weight parameters of the comprehensive score are periodically adjusted to optimize subsequent scheduling decisions.

[0006] Furthermore, the image information of each piece of meat raw material in the processing line is acquired using an industrial near-infrared camera, an ultrasonic array probe, an electrochemical electrode module, and an edge AI controller; among which, The image information of each piece of meat raw material collected from the processing production line is used to extract the structural complexity index and deterioration risk score, specifically as follows: Image information of each piece of meat raw material in the processing production line is collected. The images are input into the tissue structure map model in grayscale three-channel format, and the output is a map containing the fat layer, bone region and muscle texture of each piece of meat. Three indicators were extracted from the atlas: fat-bone edge variability, muscle texture direction change rate, and outer contour geometric deviation. The structural complexity index is obtained by weighted summation of the fat-bone edge line variability, the rate of change of muscle texture direction, and the geometric deviation of the outer contour. The industrial near-infrared camera acquires the oxidation-reduction current response curve of the raw material surface, inputs it into a one-dimensional convolutional neural network, and outputs a feature map; the feature map is then connected to a fully connected layer and mapped to a risk score.

[0007] Furthermore, the organizational structure map model has four stages, each stage containing a PatchPartition layer, a ShiftedWindowAttention mechanism, and a feedforward channel integration module; the one-dimensional convolutional neural network has a three-layer Conv1D, ReLU, and BN network structure, and outputs a feature map.

[0008] Furthermore, the comprehensive score is generated based on the structural complexity index, the degradation risk score, and the penalty-incentive function; wherein, the penalty-incentive function is used to strengthen the score distribution pull under extreme conditions, and the penalty-incentive function is obtained based on the degradation risk upper limit, the structural complexity lower limit, the structural complexity index, and the degradation risk score; The upper limit of the risk of deterioration indicates that the meat product is nearing its maximum acceptable shelf life; the lower limit of structural complexity indicates that the raw material structure is simple and the processing is fast.

[0009] Furthermore, if a piece of meat has a complex structural complexity index but an extremely high degradation risk score, a positive incentive term will be introduced into the comprehensive score to improve its scheduling ranking; if a piece of meat has a simple structural complexity index and a low degradation risk score, a negative penalty will be imposed in the comprehensive score to prevent it from seizing resources.

[0010] Furthermore, the task priority is generated by a discrete mapping of the comprehensive score, with a value range of 1 to 5. When the comprehensive score is 0.86 and the number of levels is set to 5, the corresponding priority level is 5.

[0011] Furthermore, the equipment parameters include load factor and processing adaptability; The parameters of the acquired equipment, combined with the task priority and scheduling level, are used to construct a scheduling cost function to calculate the scheduling cost value. The task is then assigned to the processing station with the lowest scheduling cost value, and scheduling control instructions are generated. Specifically, this includes: The current set of devices is constructed by combining the device parameters, task priorities, and scheduling levels; The scheduling cost of assigning tasks to equipment is calculated based on the load coefficient and processing adaptability; the smaller the value, the better. For all tasks to be assigned in each scheduling cycle, calculate the minimum cost device from the current device set to obtain the assigned device number: The scheduling control instruction is generated by combining the allocated device number, the current timestamp of the allocated device number as the expected task start time, the task number of the allocated device number, and the task level of the allocated device number.

[0012] Furthermore, the scheduling control commands are sent to the PLC or edge controller through the scheduling central control service to control specific equipment to complete the execution actions.

[0013] Furthermore, the actual scheduling cost after the execution of the scheduling control instruction is calculated, and the deviation between the comprehensive score and the actual scheduling cost is calculated as a scoring error measure. The subsequent scheduling decisions are then optimized by periodically adjusting the weight parameters of the comprehensive score. Specifically: Obtain the comprehensive score; The weighting coefficients of the comprehensive score are optimized based on the comprehensive score and actual scheduling cost of the current period to obtain the squared residual between the score and the actual scheduling cost. In each scoring optimization cycle, the squared residual is minimized once. The feasible weight coefficient space is traversed using a simple grid search method, and the weight combination corresponding to the least squared residual is retained.

[0014] A second aspect of the present invention provides an intelligent meat processing production line scheduling and control system, the system comprising: The state perception module is used to collect image information of each piece of meat raw material in the processing production line and extract the structural complexity index and deterioration risk score; The scoring generation module is used to generate a comprehensive score by weighted fusion based on the structural complexity index and the degradation risk score, and map it to the task priority to generate a scheduling level, wherein raw materials with high degradation risk are scheduled first. The scheduling execution module is used to collect equipment parameters, combine the task priority and scheduling level, construct a scheduling cost function to calculate the scheduling cost value, allocate the task to the processing station with the lowest scheduling cost value, and generate scheduling control instructions. The optimization module is used to calculate the actual scheduling cost after the execution of the scheduling control command, calculate the deviation between the comprehensive score and the actual scheduling cost as a score error metric, and periodically optimize subsequent scheduling decisions by adjusting the weight parameters of the comprehensive score.

[0015] The beneficial technical effects of the present invention are at least as follows: This invention provides an intelligent scheduling and control method and system for addressing issues such as high heterogeneity of scheduling tasks, imbalanced resource allocation, and uncoordinated system rhythm in meat processing production lines. It constructs a state-awareness system based on the structural complexity of individual meat products and the urgency of processing. By deploying graph modeling and electrochemical acquisition modules in the raw material feeding stage to extract structural indicators and risk scores, and introducing a scoring function with a safety incentive and constraint mechanism to prioritize tasks, it guides the optimal allocation of tasks among processing stations. In the scheduling execution phase, the system incorporates task-equipment compatibility and equipment load information to form a scheduling cost function. Scoring drives task ranking and resource matching. Simultaneously, the system constructs a scoring deviation measurement mechanism by collecting scheduling feedback data and uses historical execution data to back-optimize the scoring function parameters, forming a closed-loop control chain of state modeling, scoring generation, task scheduling, and feedback optimization. Compared to traditional rule-based production scheduling schemes, this invention establishes for the first time a scheduling and control mechanism in the meat processing process that uses state perception as the entry point, a scoring function as the scheduling center, and task execution deviation as the model learning entry point. It features strong real-time performance, high equipment adaptability, and continuous optimization, which can significantly improve production line collaboration efficiency and raw material processing utilization rate. It has good practical deployment value and promotion potential. Attached Figure Description

[0016] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0017] Figure 1 This is a flowchart of the intelligent meat processing production line scheduling and control method disclosed in an embodiment of the present invention. Detailed Implementation Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0018] Example 1 like Figure 1 As shown in the embodiment of the present invention, the intelligent meat processing production line scheduling and control method includes: S1. Collect image information of each piece of meat raw material in the processing production line, and extract the structural complexity index and deterioration risk score.

[0019] Specifically, this step is used to perform state-aware modeling on each piece of meat entering the processing production line, and extract two core indicators for subsequent scheduling and control: structural complexity index. Risk score of deterioration To address the current challenges of diverse meat varieties, significant morphological differences, and uncontrollable individual quality, this step establishes a joint modeling scheme using two sensing methods: visual mapping and electrochemical response. This allows the system to stably acquire quantifiable raw material state characteristics without disrupting the production line rhythm.

[0020] All input data comes from sensor devices at the front end of the production line, deployed in the raw material buffer section after the feeding conveyor belt. The data acquisition process is completed after automatic weighing and cleaning and before cutting, specifically including: Industrial near-infrared camera: Model is global shutter type, resolution 1280×1024, frame rate 30fps, used to acquire images of surface and shallow tissue structures; Ultrasonic array probe: frequency 15MHz, array arrangement spans the width of the belt, used to identify internal bone distribution and tissue interface; Electrochemical electrode module: Three sets of platinum-containing working electrode, reference electrode and auxiliary electrode, which automatically touch the surface of the raw material to perform voltammetric scanning and record the current-voltage curve; Edge AI Controller: Integrates NVIDIA Jetson Orin module to perform graph modeling and signal feature extraction, with a processing speed of no less than 50 pieces of meat per unit time.

[0021] The raw materials are simultaneously collected by the above-mentioned equipment while being transported on the belt. The images are input into the model in grayscale three-channel format, the electrochemical curve sampling frequency is 100Hz, and the total acquisition time is about 1 second. The state modeling can be completed without affecting the overall production rhythm.

[0022] The acquired image information is first input into a tissue mapping model based on SwinTransformer. This model has four stages, each containing a PatchPartition layer, a ShiftedWindowAttention mechanism, and a feedforward channel integration module, which can automatically perform spatial segmentation and local structure enhancement on the input image. The model output is a map. It includes the segmentation boundaries and corresponding attribute tags of each piece of meat, such as the fat layer, bone area, and muscle texture.

[0023] Three metrics for calculating structural complexity are extracted from the graph: First, the undulation of the fat-bone boundary. The system extracts the boundary between fat and bone, constructs a cubic polynomial fitting function based on sampling points, uses a sliding window width of 5 pixels, and calculates the standard deviation of the derivative segment by segment. This standard deviation reflects the local undulation of the structural boundary; the more uneven the edge, the higher the value. For example, in the pig scapula, due to the large bone embedding angle, the derivative changes drastically, resulting in a value close to 1.

[0024] Secondly, the rate of change of muscle texture direction The method involves calculating the gradient orientation field within the muscle region, extracting the principal orientation angle of each 16×16 pixel grid, and then calculating the standard deviation of the principal orientation distribution. A greater change in orientation indicates a more disordered arrangement of muscle fibers, making path planning more difficult. This method is suitable for recognizing textured areas such as the inner and outer sides of a chicken leg.

[0025] Third, geometric deviation of the outer contour The system extracts the outer boundary of the raw material and performs Bézier curve fitting. It then calculates the rate of change of curvature for each segment of equal length and establishes a deviation reference table with a standard meat block (such as trimmed pork belly). The greater the actual deviation, the worse the feasibility of the standardized processing path.

[0026] The structural complexity score is obtained by linearly combining the above three indicators: in: This represents the structural complexity index and is one of the core output variables in this step. It is the local derivative fluctuation of the fat-bone interface line; It is the rate of change of muscle texture direction; It is the curvature deviation of the outer contour; The normalized weighting coefficients are manually set, and their values ​​are not dynamically adjusted based on real-time input.

[0027] In terms of electrochemical sensing, the system acquires the redox current response curves on the raw material surface and inputs them into a one-dimensional convolutional neural network. The network structure is a three-layer Conv1D (kernel width 5) + ReLU + BN, and the output is a feature map. Feature maps are mapped to risk scores when connected to a fully connected layer. Rating range The scoring is determined based on the distribution pattern of meat oxidation response under conditions such as high-temperature treatment, room-temperature exposure, and cold chain disruption in the training samples.

[0028] The oxidation-reduction current response curve of the raw material surface is obtained by an electrochemical electrode module. The core is to use the current change generated by the oxidation-reduction reaction between the meat raw material surface and the electrode to reflect the freshness of the raw material (such as the degree of oxidation and deterioration). This curve is the core input for the deterioration risk score: Fresh meat exhibits distinct characteristic peaks (such as the reduction peak of myoglobin) and stable current intensity; Oxidized and deteriorated meat has increased peroxides, resulting in a decrease in the intensity of characteristic peaks or a shift in peak direction, and a reduction in curve smoothness. After processing by a one-dimensional convolutional neural network, the degradation risk score can be quantified as 0~1, supporting subsequent scheduling priority decisions (high-risk raw materials are processed first).

[0029] Output : Structural complexity index, representing the difficulty of structural processing of the meat piece in the three dimensions of image, geometry, and texture; The risk score for deterioration indicates the urgency of processing corresponding to the current oxidation state of the meat.

[0030] S2. A comprehensive score is generated by weighted fusion based on the structural complexity index and the degradation risk score, and then mapped to the task priority to generate a scheduling level, wherein raw materials with high degradation risk are scheduled first.

[0031] Specifically, the purpose of this step is to extract the structural complexity index from step one. and deterioration risk score Transform it into a unified score for task scheduling and ranking. And further generate task priority levels. This serves as the basis for decision-making in task scheduling and control. In real-world meat processing scenarios, production lines often need to simultaneously handle high-quality raw materials with extremely complex structures but still within their shelf life, as well as raw materials with relatively uniform appearance but at risk of spoilage due to cold chain interruptions or prolonged shelf time. Scheduling based solely on a single indicator can easily lead to wasted equipment resources or depreciation of high-value-added components. Therefore, the scoring model constructed in this step must be able to comprehensively evaluate processing costs and processing time, and possess adjustability and on-site deployability.

[0032] The core design idea of ​​the scoring function is: in most cases, linear fusion by weight. and However, when approaching certain risk ranges (such as a degradation score higher than 0.8 or a structural complexity exceeding 0.9), safety scheduling incentives or penalties are added to provide support through human control strategies.

[0033] The basic scoring model is as follows: in: The scoring result indicates the overall scheduling priority of this raw material; , : Static weights, satisfying ; The weight of the scoring adjustment item typically ranges from 0.05 to 0.2. The penalty-incentive function is designed to reinforce the score distribution in extreme cases, and its specific form is shown below. The design is as follows: in: This indicates an indicator function that activates the corresponding item only when the condition is met. This represents the upper limit of the risk of deterioration (usually set to 0.8), when... This indicates that the meat product is nearing its maximum acceptable shelf life and needs to be processed as a priority. This represents the lower bound of structural complexity (usually set to 0.3), when... This indicates that the raw materials have a simple structure and can be processed quickly, allowing them to be inserted into low-load workstations to fill production capacity. and The excitation intensity is controlled separately, with typical values ​​as follows: , This definition allows the scoring model to respond more accurately in the following two key scenarios: If a piece of meat has a complex structure but carries extremely high risk (such as...) , A positive incentive will be introduced into the scoring to improve its scheduling ranking; If a certain piece of meat has a simple structure and low risk (such as...) , The system will impose a negative penalty to prevent it from hogging resources, and the processing can be postponed.

[0034] This mechanism is particularly suitable for priority adjustment in industrial production lines under the scenario of "peak capacity limited". Combined with the scheduling system, it supports the control strategy of "high risk first, then high load", which can effectively improve resource utilization efficiency.

[0035] To facilitate sorting and task queuing, the system will continuously score. Discrete mapping to task priority levels : in: This is the maximum number of priority levels supported by the scheduling system, typically set to 5. This represents the scheduling priority level for the task, with a value range of [value range missing]. (Minimum) to (highest); if and ,but This indicates that it should be given priority.

[0036] This scoring mechanism is compatible with multiple scheduling systems, including pipelined models based on task fetching and parallel distribution models based on task pools. Parameters and The system is configured by the enterprise during the deployment phase based on the types of raw materials and the performance of the equipment, and is not dynamically adjusted during production line operation to ensure system stability.

[0037] Output : Comprehensive score, used for global ranking control; Task priority is used for task queue control, production line division of labor, or time window control.

[0038] S3. Collect equipment parameters, combine them with the task priority and scheduling level, construct a scheduling cost function to calculate the scheduling cost value, assign the task to the processing station with the lowest scheduling cost value, and generate scheduling control instructions.

[0039] Specifically, this step involves the scoring results given in step two. and ranking results Based on this, the core control operation is to rationally allocate meat processing tasks to processing stations. Unlike general manufacturing, the processing of meat raw materials is affected by the individual structural differences and quality timeliness, resulting in highly heterogeneous tasks and performance differences between equipment.

[0040] enter: Task scoring measures the combined result of the urgency and complexity of a task. Task priority: This indicates the priority number of the task in the scheduling system. : No. The load factor of the equipment is calculated as the ratio of the number of processing tasks in the last 5 minutes to the maximum cycle time capacity of the equipment. :Task With equipment The processing adaptability is derived from the average processing error and operational stability statistics of historical tasks on this equipment.

[0041] Constructing a device collection It is dynamically maintained by the system configuration and scheduling cycle.

[0042] To simultaneously meet the response requirements of high-scoring tasks and the adaptability of equipment capabilities, this step uses the following scheduling cost function for the task. and equipment The combination of scores is used for evaluation: in: Indicates task Assigned to device The smaller the scheduling cost, the better; Indicates equipment For the task The degree of compatibility, the average accuracy score obtained by the device in handling this type of task during historical operation, and the numerical range. ; Indicates the current load of the device, with a value of It is generated by the MES system based on real-time beat data; To determine the scheduling weight and control the sensitivity to device load, a value of 0.5 to 0.7 is recommended.

[0043] The core idea behind this function design is to prioritize assigning high-scoring tasks to low-load, historically reliable (i.e., highly adaptable) equipment, thereby ensuring optimal utilization of critical production line resources. It is particularly suitable for critical tasks in meat processing, such as structurally complex meat pieces nearing their expiration date; the system will prioritize assigning them to workstations with higher stability and greater idle time to complete the processing.

[0044] The scheduling system processes all pending tasks in each scheduling cycle (e.g., every 5 seconds). In the current set of devices Calculate the minimum cost equipment: Then, a scheduling control instruction is generated: Task Number ; Task Level Assign equipment number The system's current timestamp is used as the estimated task start time. Allocation costs As an indicator for performance evaluation, the scheduling results are sent to the PLC or edge controller through the central scheduling control service to control specific equipment to complete automatic feeding, cutting, conveying and other actions.

[0045] Output Dispatch and control commands; : Scheduling cost value, which can be used for subsequent model optimization and quality traceability analysis.

[0046] S4. Calculate the actual scheduling cost after the scheduling control is executed, calculate the deviation between the comprehensive score and the actual scheduling cost as a score error measure, and periodically optimize subsequent scheduling decisions using the weight parameters of the comprehensive score.

[0047] Specifically, this step is used to achieve closed-loop optimization of the scoring function parameters. The goal is to adjust the weight parameters in the scoring function through feedback analysis of the task scheduling execution results. and This allows the scoring model to better reflect the actual consumption and urgency of scheduled resources in future cycles, thereby improving the scheduling accuracy and overall stability of the system.

[0048] The core idea of ​​the steps is: if the score value High, but scheduling costs are high. However, the score is very low, indicating that the current scoring model overestimates the urgency of the task; conversely, the score is high. Therefore, the deviation between the two is used as a measure of scoring error, and the weights in the scoring function are adjusted accordingly.

[0049] The scoring function is in the form of (derived from step two): In this step, we keep the scoring structure unchanged and only target... and Optimization is performed. To this end, the fitting error for the scoring cost is defined as follows: in: The squared residual between the score and the actual scheduling cost; Indicates restarting with the current cycle. and candidates Recalculate the score; The actual cost of the tasks executed by the scheduling system; Let be the normalization coefficient, so that and With uniform scale (take all) (reciprocal of the maximum value) For example: If the original score of a task However, it was assigned to an idle device, resulting in Therefore, the rating system overemphasizes risk and should be adjusted accordingly in subsequent iterations. The weight. Conversely, if However, the final equipment load is high (e.g.) This indicates that the task underestimated the processing burden, and improvements should be made in the future. Weights.

[0050] In each scoring optimization cycle (e.g., after processing 200 batches), the system performs objective function minimization once, which can be done using a simple grid search method. Space, retain minimum Corresponding weight combinations: Search step size: 0.05; Search scope: , ; The results are confirmed by system maintenance personnel and written into the scoring function for the next cycle.

[0051] Example 2 This invention also provides an intelligent meat processing production line scheduling and control system, the system comprising: The state perception module 101 is used to collect image information of each piece of meat raw material in the processing production line and extract the structural complexity index and deterioration risk score; The scoring generation module 102 is used to generate a comprehensive score by weighted fusion based on the structural complexity index and the degradation risk score, and map it to the task priority to generate a scheduling level, wherein raw materials with high degradation risk are scheduled first. The scheduling execution module 103 is used to collect equipment parameters, combine the task priority and scheduling level, construct a scheduling cost function to calculate the scheduling cost value, allocate the task to the processing station with the lowest scheduling cost value, and generate scheduling control instructions. The optimization module 104 is used to calculate the actual scheduling cost after the scheduling control instruction is executed, calculate the deviation between the comprehensive score and the actual scheduling cost as a score error measure, and periodically optimize subsequent scheduling decisions by adjusting the weight parameters of the comprehensive score.

[0052] The foregoing has described specific embodiments of this specification; other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims may be performed in a different order than those shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily have to follow the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0053] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0054] For ease of description, the above devices are described in terms of function, divided into various units. Of course, in implementing this specification, the functions of each unit can be implemented in one or more software and / or hardware components.

[0055] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the embodiments of this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0056] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0057] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0058] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0059] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0060] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0061] Computer-readable media include both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0062] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0063] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0064] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0065] Finally, it should be noted that the intelligent meat processing production line scheduling and control platform disclosed in the embodiments of the present invention is only a preferred embodiment of the present invention and is only used to illustrate the technical solutions of the present invention, not to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for scheduling and controlling an intelligent meat processing production line, characterized in that, The method includes the following steps: Image information of each piece of meat raw material in the processing production line is collected, and structural complexity index and deterioration risk score are extracted; A comprehensive score is generated by weighted fusion of the structural complexity index and the degradation risk score, and then mapped to task priority to generate a scheduling level, wherein raw materials with high degradation risk are scheduled first. Collect equipment parameters, combine them with the task priority and scheduling level, construct a scheduling cost function to calculate the scheduling cost value, assign the task to the processing station with the lowest scheduling cost value, and generate scheduling control instructions; The actual scheduling cost after the scheduling control is executed is calculated, and the deviation between the comprehensive score and the actual scheduling cost is calculated as a score error measure. The weight parameters of the comprehensive score are periodically adjusted to optimize subsequent scheduling decisions.

2. The intelligent meat processing production line scheduling and control method according to claim 1, characterized in that, The image information of each piece of meat raw material in the processing line is acquired using an industrial near-infrared camera, an ultrasonic array probe, an electrochemical electrode module, and an edge AI controller; among which... The image information of each piece of meat raw material collected from the processing production line is used to extract the structural complexity index and deterioration risk score, specifically as follows: Image information of each piece of meat raw material in the processing production line is collected. The images are input into the tissue structure map model in grayscale three-channel format, and the output is a map containing the fat layer, bone region and muscle texture of each piece of meat. Three indicators were extracted from the atlas: fat-bone edge variability, muscle texture direction change rate, and outer contour geometric deviation. The structural complexity index is obtained by weighted summation of the fat-bone edge line variability, the rate of change of muscle texture direction, and the geometric deviation of the outer contour. The industrial near-infrared camera acquires the oxidation-reduction current response curve of the raw material surface, inputs it into a one-dimensional convolutional neural network, and outputs a feature map; the feature map is then connected to a fully connected layer and mapped to a risk score.

3. The intelligent meat processing production line scheduling and control method according to claim 2, characterized in that, The organizational structure map model has four stages, each stage containing a PatchPartition layer, a ShiftedWindowAttention mechanism, and a feedforward channel integration module; the one-dimensional convolutional neural network has a three-layer Conv1D, ReLU, and BN network structure, and outputs a feature map.

4. The intelligent meat processing production line scheduling and control method according to claim 1, characterized in that, The comprehensive score is generated based on the structural complexity index, the degradation risk score, and the penalty-incentive function; wherein, the penalty-incentive function is used to strengthen the score distribution pull under extreme conditions, and the penalty-incentive function is obtained based on the degradation risk upper limit, the structural complexity lower limit, the structural complexity index, and the degradation risk score; The upper limit of the risk of deterioration indicates that the meat product is nearing its maximum acceptable shelf life; the lower limit of structural complexity indicates that the raw material structure is simple and the processing is fast.

5. The intelligent meat processing production line scheduling and control method according to claim 4, characterized in that, If a piece of meat has a complex structural complexity index but an extremely high degradation risk score, a positive incentive will be introduced into the comprehensive score to improve its scheduling ranking; if a piece of meat has a simple structural complexity index and a low degradation risk score, a negative penalty will be imposed in the comprehensive score to prevent it from seizing resources.

6. The intelligent meat processing production line scheduling and control method according to claim 1, characterized in that, The task priority is generated by a discrete mapping of the comprehensive score, with a value range of 1 to 5. When the comprehensive score is 0.86 and the number of levels is set to 5, the corresponding priority level is 5.

7. The intelligent meat processing production line scheduling and control method according to claim 1, characterized in that, The equipment parameters include load factor and processing adaptability; The parameters of the acquired equipment, combined with the task priority and scheduling level, are used to construct a scheduling cost function to calculate the scheduling cost value. The task is then assigned to the processing station with the lowest scheduling cost value, and scheduling control instructions are generated. Specifically, this includes: The current set of devices is constructed by combining the device parameters, task priorities, and scheduling levels. The scheduling cost of assigning tasks to equipment is calculated based on the load coefficient and processing adaptability; the smaller the value, the better. For all tasks to be assigned in each scheduling cycle, calculate the minimum cost device from the current device set to obtain the assigned device number: The scheduling control instruction is generated by combining the allocated device number, the current timestamp of the allocated device number as the expected task start time, the task number of the allocated device number, and the task level of the allocated device number.

8. The intelligent meat processing production line scheduling and control method according to claim 7, characterized in that, The scheduling and control commands are sent to the PLC or edge controller through the central scheduling and control service to control specific equipment to perform actions.

9. The intelligent meat processing production line scheduling and control method according to claim 1, characterized in that, The actual scheduling cost after the execution of the scheduling control command is calculated, and the deviation between the comprehensive score and the actual scheduling cost is calculated as a scoring error measure. Subsequent scheduling decisions are then optimized by periodically adjusting the weight parameters of the comprehensive score. Specifically: Obtain the comprehensive score; The weighting coefficients of the comprehensive score are optimized based on the comprehensive score and actual scheduling cost of the current period to obtain the squared residual between the score and the actual scheduling cost. In each scoring optimization cycle, the squared residual is minimized once. The feasible weight coefficient space is traversed using a simple grid search method, and the weight combination corresponding to the least squared residual is retained.

10. An intelligent scheduling and control system for a meat processing production line, characterized in that, The system includes: The state perception module is used to collect image information of each piece of meat raw material in the processing production line and extract the structural complexity index and deterioration risk score; The scoring generation module is used to generate a comprehensive score by weighted fusion based on the structural complexity index and the degradation risk score, and map it to the task priority to generate a scheduling level, wherein raw materials with high degradation risk are scheduled first. The scheduling execution module is used to collect equipment parameters, combine the task priority and scheduling level, construct a scheduling cost function to calculate the scheduling cost value, allocate the task to the processing station with the lowest scheduling cost value, and generate scheduling control instructions. The optimization module is used to calculate the actual scheduling cost after the execution of the scheduling control command, calculate the deviation between the comprehensive score and the actual scheduling cost as a score error metric, and periodically optimize subsequent scheduling decisions by adjusting the weight parameters of the comprehensive score.

Images