A collaborative control method, system, terminal, and medium for laser flexible processing based on image analysis and intelligent scheduling

Abstract

This invention relates to the field of flexible laser processing, specifically providing a collaborative control method, system, terminal, and medium for flexible laser processing based on image analysis and intelligent scheduling. The method includes: receiving a conventional task instruction containing known parameters and an image file to be processed; performing image processing and path planning on the image file to extract the process attribute parameters of the image processing task; merging the parameters of the two types of tasks to construct a multi-objective model with the optimization objectives of minimizing production cycle, equipment load imbalance, total energy consumption, and total tool changes; solving the model using an improved genetic algorithm to output the optimal task allocation and sorting scheme; and controlling the laser equipment to execute tasks according to the scheduling scheme, supporting automatic tool changing and real-time monitoring. This invention improves the intelligence level, production efficiency, and resource utilization of flexible laser processing systems.

Description

Technical Field

[0001] This invention relates to the field of laser flexible processing, specifically to a collaborative control method, system, terminal, and medium for laser flexible processing based on image analysis and intelligent scheduling. Background Technology

[0002] In the current laser processing manufacturing industry, flexible production has become an important trend. This is especially true for customized orders involving mixed materials (metal and non-metal) and multiple processes (engraving, cutting, welding, cleaning, etc.) in fields such as automotive parts and cultural and creative products. Typical production scenarios involve diverse task scales, including both standardized batch production tasks (routine tasks) from Enterprise Resource Planning (ERP) systems and personalized image processing tasks such as hand-drawn patterns or photos from end users. The need for mixed scheduling and collaborative execution of these two types of tasks on multiple integrated laser machines is increasingly urgent. However, existing technologies face the following prominent problems in this scenario:

[0003] First, in image processing tasks, traditional methods heavily rely on human experience. Operators need to use specialized computer-aided design (CAD) software to manually outline, extract contours, and plan paths for user-provided images before converting them into numerical control (G-code) instructions. This process is not only time-consuming but also demands high skill levels, making it difficult to adapt to the flexible production rhythm of small batches and rapid responses, thus becoming a bottleneck restricting production efficiency and personalized service capabilities.

[0004] Secondly, regarding multi-device collaborative scheduling, existing Manufacturing Execution Systems (MES) or scheduling methods typically optimize routine tasks with known and defined parameters, focusing primarily on traditional indicators such as completion time and equipment utilization. When image processing tasks with unknown parameters and complex processing are introduced into the scheduling system, existing solutions exhibit significant drawbacks: First, they cannot achieve unified parameterized description and integrated optimization for both types of tasks, often treating image tasks as a "black box" or subjecting them to simple estimations. This leads to large deviations between the scheduling scheme and actual execution, resulting in frequent equipment idleness or overload. Second, they fail to fully consider the significant time costs and efficiency losses caused by the frequent laser head replacements required by integrated laser equipment to adapt to mixed material processing. The equipment downtime caused by tool changing operations is not accurately modeled and considered a core optimization objective, potentially leading to unnecessary frequent tool changes in actual production and severely impacting the overall production cycle time.

[0005] While some existing laser processing methods combine image processing, they typically focus on path generation for a single device and fail to deeply integrate with intelligent scheduling of multiple devices at the workshop level. Some research employs metaheuristic algorithms such as genetic algorithms to solve workshop scheduling problems, but their decision variables and optimization models are usually based on simplified task assumptions, failing to cover the entire closed-loop process from image to path to schedulable parameters. Furthermore, they lack systematic modeling and targeted solution strategies for the crucial optimization dimension of tool change count in flexible laser processing. Summary of the Invention

[0006] To address the aforementioned issues, this invention provides a collaborative control method, system, terminal, and medium for flexible laser processing based on image analysis and intelligent scheduling, which improves the intelligence level, production efficiency, and resource utilization of flexible laser processing systems.

[0007] In a first aspect, the technical solution of the present invention provides a collaborative control method for laser flexible processing based on image analysis and intelligent scheduling, comprising the following steps:

[0008] S1, receive a task set containing routine task instructions and image files; wherein, the routine task instructions contain basic process requirements for determining the final process attribute parameters of the routine task, and the image files contain the pattern to be processed;

[0009] S2, perform image processing and path planning on the pattern to be processed in the image file to extract the process attribute parameters of the image processing task; the process attribute parameters include: estimated processing time, energy consumption per unit time, and required laser type;

[0010] S3, construct and solve the multi-device task scheduling optimization model;

[0011] S3.1: Merge the process attribute parameters of regular tasks with the process attribute parameters generated by image processing tasks into a unified task process parameter list;

[0012] S3.2: Based on the task process parameter list, the current availability status of each laser device and the laser head loading type, construct a multi-objective optimization model with the optimization objectives of minimizing production cycle, equipment load imbalance, total energy consumption and total number of tool changes, and with equipment capacity and task priority as constraints;

[0013] S3.3: Use a genetic algorithm to solve the multi-objective optimization model and output the optimal task-equipment allocation scheme and the task processing sequence on each device;

[0014] S4. Based on the scheduling scheme output in step S3, control the corresponding laser devices to execute their assigned task sequences in sequence.

[0015] Secondly, the technical solution of the present invention provides a laser flexible processing collaborative control system based on image analysis and intelligent scheduling, comprising:

[0016] The task set receiving module is used to receive a task set containing regular task instructions and image files; wherein, the regular task instructions contain basic process requirements for determining the final process attribute parameters of the regular task, and the image files contain the pattern to be processed;

[0017] The image task parameter extraction module is used to perform image processing and path planning on the pattern to be processed in the image file to extract the process attribute parameters of the image processing task; the process attribute parameters include: estimated processing time, energy consumption per unit time, and required laser type;

[0018] The task scheduling optimization model construction and solution module is used to construct and solve multi-device task scheduling optimization models.

[0019] a) Merge the process attribute parameters of regular tasks with the process attribute parameters generated by image processing tasks into a unified task process parameter list;

[0020] b) Based on the task process parameter list, the current availability status of each laser device and the laser head loading type, construct a multi-objective optimization model with the optimization objectives of minimizing production cycle, equipment load imbalance, total energy consumption and total number of tool changes, and constraints of equipment capacity and task priority;

[0021] c) Use a genetic algorithm to solve the multi-objective optimization model and output the optimal task-equipment allocation scheme and the task processing sequence on each equipment;

[0022] The task allocation and execution module is used to construct the scheduling scheme output by the solution module based on the task scheduling optimization model, and control the corresponding laser equipment to execute its assigned task sequence in sequence.

[0023] Thirdly, the technical solution of the present invention provides a terminal, comprising:

[0024] The memory is used to store the collaborative control program for laser flexible processing based on image analysis and intelligent scheduling;

[0025] A processor is configured to implement the steps of the laser flexible processing collaborative control method based on image analysis and intelligent scheduling as described above when executing the laser flexible processing collaborative control program based on image analysis and intelligent scheduling.

[0026] Fourthly, the present invention provides a computer-readable storage medium storing a laser flexible processing collaborative control program based on image analysis and intelligent scheduling. When the laser flexible processing collaborative control program based on image analysis and intelligent scheduling is executed by a processor, it implements the steps of the laser flexible processing collaborative control method based on image analysis and intelligent scheduling as described in any of the above claims.

[0027] As can be seen from the above technical solutions, this application has the following advantages:

[0028] A complete image processing and path planning workflow was designed. This workflow employs an adaptive sampling algorithm based on curvature variation for contour feature point extraction, and combines this with the kinematic characteristics of the CNC platform for path smoothing and processing time estimation. Finally, it automatically converts unstructured image files into structured process attribute parameters, including estimated processing time, energy consumption per unit time, and required laser type. This solves the problems of low efficiency and large errors associated with traditional manual conversion, allowing image processing tasks to be input into the scheduling system in a quantitative form, just like regular tasks, thereby improving the quality of global optimization.

[0029] The constructed optimization model not only includes classic objectives such as minimizing production cycle time, balancing equipment load, and total energy consumption, but also incorporates minimizing the total number of tool changes as an independent and important optimization objective. This model fully considers the physical tool changing actions and their time costs caused by switching processing materials and functions of laser equipment, solving the problem of the disconnect between existing scheduling models and the actual operation of the equipment. This allows the generated scheduling scheme to proactively avoid unnecessary laser head changes, thereby significantly reducing equipment downtime and directly improving equipment utilization and overall production efficiency.

[0030] A dual-layer encoded chromosome structure was further designed, and heuristic rules prioritizing clustering of similar tasks and constraint-aware mutation acceptance criteria were introduced into the crossover and mutation operators. This addresses the issues of low search efficiency and poor solution quality in general optimization algorithms when solving scheduling problems with special process constraints. The algorithm can effectively explore the solution space and tend towards high-quality scheduling schemes with fewer tool changes and balanced load, ensuring high-performance and executable scheduling results within a limited computation time. Attached Figure Description

[0031] To more clearly illustrate the technical solution of this application, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1This is a schematic diagram of a collaborative control method for laser flexible processing based on image analysis and intelligent scheduling, provided in an embodiment of the present invention.

[0033] Figure 2 A flowchart illustrating the process of extracting technological attribute parameters for image processing tasks.

[0034] Figure 3 This is a schematic block diagram of a collaborative control system for laser flexible processing based on image analysis and intelligent scheduling, provided as an embodiment of the present invention.

[0035] Figure 4 This is a schematic diagram of the structure of a terminal provided in an embodiment of the present invention. Detailed Implementation

[0036] To make the purpose, features, and advantages of this application more apparent and understandable, specific embodiments and accompanying drawings will be used to clearly and completely describe the technical solution protected by this application. Obviously, the embodiments described below are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0037] Unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this application and in the specification of this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0038] Figure 1 This is a schematic flowchart illustrating a collaborative control method for laser flexible processing based on image analysis and intelligent scheduling, provided in an embodiment of the present invention. Figure 1 The executing entity can be a collaborative control system for flexible laser processing based on image analysis and intelligent scheduling. The collaborative control method for flexible laser processing based on image analysis and intelligent scheduling provided in this embodiment of the invention is executed by a computer device; correspondingly, the collaborative control system for flexible laser processing based on image analysis and intelligent scheduling runs on the computer device. Depending on different requirements, the order of the steps in this flowchart can be changed, and some steps can be omitted.

[0039] like Figure 1 As shown, the method includes the following steps.

[0040] S1, receive a task set containing routine task instructions and image files; wherein, the routine task instructions contain basic process requirements for determining the final process attribute parameters of the routine task, and the image files contain the pattern to be processed.

[0041] The "task set" received in step S1 is a complex consisting of two types of task instructions with different properties: regular task instructions and image files.

[0042] Regular task instructions originate from the upper-level production management system. Their structured data contains the basic information needed to generate the final scheduling parameters, collectively referred to as basic process requirements. In one embodiment, these may include:

[0043] Basic identification information: Task ID, Order Number;

[0044] Processing target definition: processing material type, processing function type;

[0045] Geometric and dimensional constraints: the geometric dimensions of the workpiece, the target machining accuracy, and the surface quality requirements;

[0046] Initial process reference: Recommended initial parameter ranges such as laser power and scanning speed based on the enterprise's standard process library;

[0047] Management attributes: task priority, customer delivery date.

[0048] Using the basic process requirements contained in the routine task instructions as input, determine the process attribute parameters of the routine task, including: estimated processing time, energy consumption per unit time, and required laser type.

[0049] The required laser type can be obtained directly by querying the preset "Material-Processing Function-Laser Mapping Table", which is a rule base predefined by the enterprise based on the equipment configuration.

[0050] The processing time consists of effective processing time and auxiliary / idle travel time. Effective processing time is calculated based on the total length of the processing path and the effective average processing speed. Total path length estimation: For contour cutting, the total path length can be approximated as the perimeter of the processed contour; for rectangles, the total path length is the perimeter of the rectangle; for complex graphics, it can be directly read from the task-related CAD file or estimated using simple geometric formulas. For area engraving / marking, the total path length can be estimated based on fill density and processing area, i.e., it is equal to the product of processing area, fill density, and a coefficient related to the fill mode. Effective average speed determination: This is adjusted based on the processing function, material thickness, and precision requirements. For example, for thick plate cutting, the speed needs to be reduced according to the thickness; for high-precision engraving, the speed also needs to be reduced accordingly. The correction coefficient is obtained by querying the process library; i.e., the effective speed equals the recommended speed multiplied by the correction coefficient. Therefore, the effective processing time equals the total path length divided by the effective speed. Auxiliary time estimation: This includes idle travel positioning time, laser start / stop time, platform lifting time, etc. This part of the time can be set to an empirical value based on the equipment model and historical data. Ultimately, the total estimated time equals the sum of the effective processing time and the auxiliary time.

[0051] Energy consumption per unit time is the sum of average processing power and auxiliary system power consumption. Considering that the laser does not always operate at full recommended power during processing, but rather varies depending on the processing function (duty cycle or power modulation), the average processing power is the recommended power at full load multiplied by the power efficiency factor. Auxiliary system power consumption includes the basic operating power of the cooling system, exhaust system, CNC system, etc., and can be set as a device constant.

[0052] Based on the above calculations, each regular task object is assigned three core process attribute parameters that are isomorphic to the image task object: estimated processing time, energy consumption per unit time, and required laser type.

[0053] Image files originate from end users or designers and are uploaded via dedicated client software, web applications, or mobile applications (APPs). These tasks cater to personalized, small-batch customization needs. Files are common image formats such as JPG, PNG, or BMP containing the images to be processed.

[0054] S2, perform image processing and path planning on the pattern to be processed in the image file to extract the process attribute parameters of the image processing task; the process attribute parameters include: estimated processing time, energy consumption per unit time, and required laser type.

[0055] Figure 2 A flowchart illustrating the process of extracting technological attribute parameters for image processing tasks includes the following steps.

[0056] S2.1: Perform noise filtering, contrast enhancement, edge sharpening, and background redundancy removal on the pattern to be processed to generate an enhanced image.

[0057] Noise filtering: The noise filtering algorithm built into the Convolutional Neural Network (CNN) is used to accurately filter out brush stroke shakiness in hand-drawn patterns and shooting noise in album images. The image pixels are smoothed by 3×3 or 5×5 convolution kernels, which remove invalid noise points while preserving the image edge features. The noise filtering accuracy is ≥95%.

[0058] Contrast Enhancement: Automatically detects the grayscale value distribution range of the image, and uses an adaptive histogram equalization algorithm to adjust the pixel grayscale values ​​of images with insufficient contrast, thereby expanding the grayscale difference between the effective feature area and the background.

[0059] Edge sharpening: The Laplacian operator or Sobel operator is used to enhance the edges of the image. By calculating the gray-level gradient between a pixel and its surrounding pixels, the edge contours of the effective processing area in the image are strengthened.

[0060] Background Redundancy Removal: Based on image semantic segmentation technology, it automatically identifies and separates the effective processing area and background area in the image. For hand-drawn patterns, it automatically removes the canvas background. For album images, it separates the foreground processing object and the background environment through threshold segmentation, retaining only the core image content that needs to be processed.

[0061] S2.2: Binarize the enhanced image to obtain a binary image that distinguishes the processing area from the background; use an edge detection algorithm to traverse the binary image, extract the edge pixel set of the processing area, and record the coordinate information of each pixel; connect and regularize the extracted discrete edge feature points, and generate smooth and continuous closed or non-closed contours through line fitting or arc fitting algorithms; convert the regularized contours into vector path data.

[0062] Specifically, firstly, based on the preprocessed enhanced image, the image is converted into a black and white binary image through binarization to clearly distinguish between processed and unprocessed objects. Then, the Canny edge detection algorithm is used to traverse the binary image, identify the edge pixels of the effective processing area, and record the coordinate information of each pixel (X / Y axis coordinates, with an accuracy of ±0.001mm). For curved contours, the feature point sampling density is automatically increased, and for straight contours, the feature point distribution is optimized. Next, the detected discrete edge feature points are connected to form continuous contour segments. Through algorithms such as straight line fitting and circular arc fitting, irregular contour details are regularized, and contour burrs or breaks caused by residual image noise are removed to generate smooth and continuous closed or non-closed contours. Finally, the regularized continuous contours are converted into vector path data, and complex contours are represented by Bézier curves or spline curves to ensure the smoothness and continuity of the path.

[0063] In one specific embodiment, an edge detection algorithm is used to extract the initial set of contour pixels from the binarized contour image. For the initial contour pixel set Perform adaptive sampling based on curvature variation to generate an optimized subset of feature points. .

[0064] Among them, for point Its approximate curvature Estimate using the following formula:

[0065]

[0066] And based on this curvature With the preset basic sampling interval and curvature sensitivity coefficient Dynamically determine the sampling interval , is represented as:

[0067]

[0068] in accordance with from Selecting a subset of feature points .

[0069] An iterative endpoint fitting algorithm is used to fit the feature point subset. Segmentation recognition is performed, and a continuous vector processing path is generated using a straight line or circular arc fitting algorithm. Smoothing is then applied at the connection points of the path to ensure first-order geometric continuity.

[0070] S2.3: Based on the combined objective of minimizing total processing time and motion stability, the vector path is optimized.

[0071] Specifically, this includes: adapting the path analysis to the motion characteristics of the CNC platform, planning the movement path with minimal idle travel time, smoothing geometric abrupt changes in the path, and removing redundant path points that do not affect contour accuracy. Specifically, this includes: motion characteristic adaptation analysis, idle travel path planning, path smoothing, and removal of redundant path points.

[0072] Motion characteristic adaptation analysis: Obtain the motion parameter constraints of the CNC platform, including the maximum traverse speed, acceleration, and motion smoothness requirements of the X / Y / Z axes (S-curve acceleration and deceleration control), and determine the boundary conditions for path optimization in combination with the process requirements of laser processing.

[0073] Idle travel path planning: The shortest path Dijkstra algorithm is used to analyze the location distribution of each processing area in the vector path, plan the processing start point, processing sequence and movement path between areas, avoid non-processing areas, and minimize idle travel time.

[0074] Path smoothing: For abrupt changes such as corners and polylines in the vector path, circular arc transitions or spline curve interpolation are used to make the path conform to the S-curve acceleration and deceleration control requirements of the CNC platform.

[0075] Redundant point removal in the path: Optimize and filter overly dense feature points in the vector path, remove redundant points without affecting the contour accuracy, and simplify the amount of path data.

[0076] Specifically, it includes the following steps.

[0077] S2.31: Obtain the kinematic constraint parameters of the target CNC platform, including the maximum speed, maximum acceleration, S-curve acceleration / deceleration time constant, and the maximum centripetal acceleration allowed by the laser head for each axis.

[0078] Obtain the key kinematic parameters of the target CNC platform and construct hard constraints for path optimization:

[0079] Maximum speed of each axis: ;

[0080] Maximum acceleration of each axis: ;

[0081] S-curve acceleration / deceleration time constant: ;

[0082] Maximum permissible centripetal acceleration of the laser head (to avoid jitter): .

[0083] S2.32: Construct a comprehensive cost function to evaluate the merits of each path segment; define a comprehensive cost for each path segment, which is a weighted sum of the estimated travel time cost and the processing stability risk cost; wherein: the estimated travel time cost includes the constant speed time based on the path segment length and the time loss due to acceleration and deceleration control; the processing stability risk cost is calculated based on the motion jitter risk caused by the coupling of local curvature of the path and feed speed, and the processing thermal shock risk caused by the coupling of excessively rapid changes in laser power per unit path length and feed speed.

[0084] Define a comprehensive cost for each path segment (such as a straight line or arc) and its connecting points. For each point... Time The path segment, its cost It consists of two weighted parts:

[0085]

[0086] in, The estimated travel time cost includes acceleration losses, based on segment length. and preset feed rate constant time It also includes the acceleration / deceleration time loss caused by speed planning (S-curve). Estimated using the following model:

[0087]

[0088] in, .

[0089] The model shows that shorter line segments are less efficient because they cannot accelerate to full speed.

[0090] To mitigate the risks associated with processing stability, local trajectory curvature was introduced. and laser power The coupling effect is calculated using the following formula:

[0091]

[0092] In the formula, For the local curvature of the path segment, To preset the feed rate, To preset the laser power, The path segment length, These are adjustable weighting coefficients that include the corresponding dimension conversion factors.

[0093] First item (centripetal force risk): High curvature With high speed This will result in a large centripetal acceleration, causing the laser head to vibrate.

[0094] Second item (thermal shock risk): Rate of change of laser power per unit length Too large, at high speed This can lead to uneven heating of the material, affecting processing quality (such as rough cut surfaces and inconsistent carving depth).

[0095] S2.33: Based on the comprehensive cost function, a heuristic search algorithm is used to plan the empty travel path with the minimum total cost between processing regions.

[0096] The movement (empty journey) between processing areas is modeled as a graph theory problem. The nodes of the graph are the entry / exit points of each processing area, and the weights of the edges are the comprehensive costs mentioned above. The A* search algorithm is used, whose heuristic function h(n) combines Euclidean distance and directional consistency towards the destination to plan the empty journey path with the minimum total cost globally, rather than just the shortest path.

[0097] S2.34: Adaptive smoothing is performed on the corners in the path. The smoothing radius is dynamically determined based on the preset speed and maximum centripetal acceleration at the corner. Redundant points are removed from the smoothed path based on dynamic tolerance, which is related to the local curvature.

[0098] For all corners in the optimized path, calculate the theoretical centripetal acceleration of the corners. .like If this occurs, smoothing is triggered. The smoothing radius R is dynamically calculated based on the current velocity V.

[0099]

[0100] The corner is smoothed using a tangent circular arc of radius R, and resampling is performed to ensure that the smoothed path points meet the machining accuracy requirements. Based on the smoothing, the path point sequence is compressed using the Douglas-Peucker algorithm, with a dynamic tolerance inversely proportional to the local curvature set. :

[0101]

[0102] That is, retain more points in high curvature areas to ensure accuracy, and remove more points in low curvature areas to improve efficiency.

[0103] S2.4, based on the material type and processing function of the processing task, call the corresponding processing strategy package from the preset path planning strategy library; based on the parameter rules in the called processing strategy package, parameterize the optimized vector path to generate the final processing path and associated process parameter set.

[0104] The path planning strategy library has at least two processing strategy packages pre-built: speed-first strategy package and accuracy-first strategy package.

[0105] The speed-priority strategy package is associated with a set of parameter rules that favor processing efficiency, including: a larger path interpolation tolerance (e.g., 0.05-0.1mm), an aggressive idle stroke merging threshold (e.g., ≤5mm), coarse corner smoothing (transition radius ≥0.5mm), and a higher range of process speed and power parameters.

[0106] The precision-first strategy package is associated with a set of parameter rules that favor machining quality, including: smaller path interpolation tolerance (e.g., 0.001-0.01mm), conservative idle stroke merging threshold (e.g., ≤1mm), fine corner smoothing (transition radius ≤0.1mm), and lower, more stable range of process speed and power parameters.

[0107] For metal cutting or metal welding tasks: the speed-first strategy package is invoked by default. Its specific process strategies aim to improve processing efficiency; for example, it optimizes the cutting sequence through the "inner and outer contour layer cutting" strategy and uses the high-speed parameters in this strategy package.

[0108] For metal engraving, metal cleaning, non-metal engraving, or non-metal marking tasks: the precision-priority strategy package is invoked by default. Its specific process strategies aim to ensure processing quality; for example, it controls the engraving depth through a "surface micro-etching" strategy and employs high-precision interpolation and smoothing parameters from this strategy package.

[0109] For non-metallic cutting tasks: The selection is adaptive based on material thickness. When the thickness is ≥ a preset threshold (e.g., 10mm), the speed-priority strategy package is invoked; when the thickness is < the threshold, the accuracy-priority strategy package is invoked. The specific process strategy needs to control the heat-affected zone.

[0110] After selecting the basic strategy package and specific process strategy, the system performs parameter adjustments according to the following steps:

[0111] Step 1: Apply the parameters of the basic strategy package. Apply the rules defined in the selected strategy package, such as path interpolation tolerance, empty path merging threshold, and corner handling parameters, to the vector path to complete the geometric optimization of the path;

[0112] Step 2: Bind the process parameter range. Use the process speed and power parameter ranges defined in the selected strategy package as the constraint boundaries for subsequent process parameter settings;

[0113] Step 3: Fine-tune the specific process strategy. Within the parameter framework established in Steps 1 and 2, the path parameters are fine-tuned based on the specific material and processing function to instantiate the abstract process strategy. For example:

[0114] For metal engraving, within the precision-priority parameter framework, specific values ​​for "line width step size ≤ 0.05mm" and "depth ≤ 0.5mm" are further set according to the material hardness to achieve the "surface micro-etching" strategy.

[0115] For metal cutting, within the speed priority parameter framework, a specific value is selected based on the thickness within the speed range provided in the package (e.g., 0.1-5 m / min), and the processing sequence of inner and outer contour layers is configured to achieve a layered cutting strategy.

[0116] For non-metallic cutting, within the selected strategy package parameter framework, the power is fine-tuned according to the material characteristics to ensure that the heat-affected zone is ≤0.1mm.

[0117] S2.5, the final processing path and associated process parameter set are converted into G-code executable by the target laser equipment; based on the kinematic parameters of the CNC platform corresponding to the target laser equipment, motion simulation is performed on the G-code, the total processing time is extracted from the motion simulation results as the estimated processing time, and the average power is extracted from the power change curve during the simulation as the basis for the energy consumption per unit time. The required laser type is determined according to the material type and processing function.

[0118] The optimized and process-adapted paths are automatically converted into equipment-recognizable G-code, supporting G-code syntax adaptation for different brands of laser equipment. Key parameters for scheduling are extracted from the generated G-code, including:

[0119] Time-related parameters: estimated total machining time (including effective machining time, idle travel time, auxiliary action time, and tool change time), and time consumption of critical actions;

[0120] Process and equipment load parameters: material identification, processing function type, core process parameter range, and equipment load parameters;

[0121] Resource usage parameters: laser head type requirements, auxiliary resource requirements;

[0122] Priority-related parameters: machining accuracy level, urgency level indicator.

[0123] S3, construct and solve the multi-device task scheduling optimization model.

[0124] S3.1: Merge the process attribute parameters of regular tasks with the process attribute parameters generated by image processing tasks into a unified task process parameter list.

[0125] Processing tasks from different sources are standardized and integrated. Specifically, the process attribute parameters obtained through rapid calculation for routine tasks are merged with similar parameters extracted from image processing tasks via S2.6 high-fidelity simulation to form a unified, structured list of task process parameters. Each entry in this list represents a schedulable atomic task and contains all the quantization information required by the scheduling algorithm.

[0126] S3.2: Based on the task process parameter list, the current availability status of each laser device and the laser head loading type, construct a multi-objective optimization model with the optimization objectives of minimizing production cycle, equipment load imbalance, total energy consumption and total tool change times, and with equipment capacity and task priority as constraints.

[0127] The decision variables in this embodiment include the task allocation matrix. and task sorting vector .

[0128] Task allocation matrix It is The binary (0-1) matrix is ​​represented as:

[0129]

[0130] in, The total number of devices, numbered ; Number the total number of tasks ; Indicates task Assigned to device implement; Indicates task Unassigned to device .

[0131] The constraint is that each task must be assigned to one and only one device, i.e. .

[0132] Task sorting vector A set of matrices The associated integer sequence.

[0133] For equipment Define its task sorting vector as . To be assigned to the device The total number of tasks, It is an integer representing the task. In the equipment The processing sequence number on it.

[0134] The constraint requires that the sequence numbers of tasks on the same device must be unique and form a sequence from 1 to... The arrangement.

[0135] The multi-objective optimization function of the multi-objective optimization model is expressed as:

[0136]

[0137] in, These are weighting coefficients that are dynamically adjusted based on the production scenario.

[0138] Total production cycle This refers to the longest total time required for all devices to complete their assigned tasks. The goal is to minimize the overall completion time, expressed as:

[0139]

[0140] In the formula, The total number of devices. Total number of tasks; For binary decision variables, when the task Assigned to device The value is 1 if it is true, and 0 otherwise. For the task Estimated processing time; For equipment Total number of tool changes; This is the standard time for a single tool change.

[0141] In a parallel production system, the final completion time does not depend on the sum of the working times of all equipment, but rather on the slowest equipment. In this embodiment, the total production cycle time is... Based on the actual wall clock time from the start of the first task to the end of the last task, the maximum bottleneck is shortened by balancing the load of each device, thereby achieving true optimization of the overall cycle. Furthermore, tool change time is included in the load calculation, allowing the optimization scheme to proactively avoid situations where frequent tool changes cause a particular device to become a bottleneck.

[0142] Equipment load imbalance This is used to measure the degree of difference in the total processing time of each piece of equipment. It is measured by the coefficient of variation of the total processing time of each piece of equipment. The goal is to make the workload of each piece of equipment as balanced as possible, which is expressed as:

[0143]

[0144] in, Indicates equipment Total processing time This indicates the average processing time for all equipment.

[0145] Equipment load imbalance in this embodiment This ensures that regardless of the average load of each device, the load imbalance of the equipment is reduced. All can fairly measure their relative equilibrium and minimize It can effectively reduce the idleness of individual equipment, thereby improving the overall asset utilization rate and return on investment.

[0146] Total energy cost The total energy cost consumed by all equipment in performing machining and tool changing processes, with the goal of reducing energy consumption while ensuring production, is expressed as:

[0147]

[0148] in, For the task Energy consumption per unit time; This represents the energy consumption per unit time during the tool changing process.

[0149] This embodiment treats energy consumption as an independent optimization objective, enabling the system to automatically allocate high-energy-consuming tasks to high-efficiency equipment or reduce high-power standby time by adjusting the task sequence, thereby achieving energy conservation and emission reduction, while meeting delivery deadlines. Furthermore, its coupling with time and tool-changing objectives allows the algorithm to perceive the inherent relationships between objectives during optimization.

[0150] Total number of tool changes The calculation method is to iterate through each device. The assigned task sequence is used to count the number of tool changes caused by different laser head types required for adjacent tasks, and these counts are accumulated. The goal is to reduce non-value-added tool change downtime, expressed as:

[0151]

[0152] in, To be assigned to the device The number of tasks; Indicates equipment Upper The "material-processing function" combination identifier for each task. ; is the discriminant function, when The value is 1 if it is active, and 0 otherwise.

[0153] The value depends entirely on the order in which the tasks are arranged on each device. Minimize This powerfully drives the algorithm to cluster and sequentially arrange tasks for similar laser heads, thereby significantly reducing non-productive downtime. By prioritizing tool change frequency as a primary optimization objective alongside time and energy consumption, the integration of process planning and production scheduling is achieved, enhancing the practical feasibility and efficiency of the solution.

[0154] The constraints of the multi-objective optimization model include: task assignment uniqueness constraint: each task must be assigned to one and only one device; device capability adaptability constraint: the laser head type assigned to the task must match the current processing capacity of the device or the processing capacity that can be obtained through tool changing; device load rate constraint: the total processing time of each device must not exceed its available production time; task priority constraint: on the same device, the processing order of high-priority tasks must precede that of low-priority tasks; tool changing process constraint: the cumulative tool changing time of a single device must not exceed the target proportion of its available production time.

[0155] S3.3: Use a genetic algorithm to solve the multi-objective optimization model and output the optimal task-equipment allocation scheme and the task processing sequence on each device.

[0156] A complete scheduling scheme consists of Common definition. In improved genetic algorithms: each "individual" or "chromosome" represents a complete... A combination is a possible scheduling scheme. Multiple such individuals are randomly generated to form the initial population. Substitute into the optimization objective function In the function The specific calculations all depend on the current and .For example, The calculation of (number of tool changes) requires iterating through each machine and pressing... After sorting the task sequences, check whether the laser head type of adjacent tasks has changed. Genetic operations such as selection, crossover, and mutation alter individual chromosomes, thus changing... and / or The algorithm uses a structure to explore new scheduling schemes. The final output of the algorithm is the optimal individual, i.e., the optimal [schedule / scheme]. This is the optimal task allocation and sorting scheme.

[0157] In this embodiment, the chromosomes of the genetic algorithm use a two-layer coding structure to represent a complete scheduling scheme.

[0158] The first layer (allocation layer) employs task-based segmented coding. A chromosome is a sequence of elements equal to the total number of tasks. sequence Among them, gene loci The value represents the task. The assigned device number This encodes the assignment matrix in the decision variables. .

[0159] Second layer (sorting layer): After the allocation layer is determined, for each device Generate a daughter chromosome It is a list that stores all task IDs assigned to the device in sequence. The order of this sub-chromosome represents the processing order of tasks on the device, thus encoding the order vector in the decision variables. .

[0160] Encoding Example: Suppose there are 3 devices and 5 tasks. A chromosome might be represented as: (Task 1 to Device 2, Task 2 to Device 1, Task 3 to Device 2...). From this, we can deduce:

[0161] Task list for device 1 ( ):[2, 5]→ Processing order: Task 2 first, then Task 5;

[0162] Task list for device 2 ( ):[1, 3]→ Processing order: Task 1 first, then Task 3;

[0163] Device 3's task list ( ):[4]→ Processing sequence: Task 4.

[0164] The fitness value is taken as the reciprocal of the multi-objective optimization function to reflect the minimization objective.

[0165] In this embodiment, the genetic manipulation is designed to actively guide the reduction of tool changes during the global search. A tournament selection method is used to randomly select from the population. Each individual is selected, and the one with the highest fitness is chosen to enter the next generation.

[0166] The crossover operation in a genetic algorithm includes: assigning encodings to two parent individuals. Perform two-point crossover; after the crossover is complete, construct the sorting code for each device for the offspring individuals. At that time, heuristic rules are used to prioritize arranging tasks with the same type of laser required adjacent to each other in the sequence.

[0167] Specifically, for the two parent individuals A two-point crossover approach is used. Two crossover points are randomly selected, and the device allocation information in the middle segment is exchanged. After the allocation layer crossover, a device in a child instance may receive tasks from both parents. At this point, a new task is generated for that device. At the same time, a heuristic rule of prioritizing clustering of similar tasks is adopted: tasks requiring the same type of laser are prioritized to be grouped together in the sequence. By introducing local optimization during the crossover process, the chromosome is actively guided to evolve in a direction that reduces the number of laser replacements.

[0168] In this embodiment, the mutation operation of the genetic algorithm includes sorting layer mutation. After the sorting layer mutation is executed, the number of tool changes in the task sequence of the device before and after the mutation is calculated and compared, and whether the number of tool changes increases is used as one of the evaluation factors for whether to accept the mutation result.

[0169] Specifically, changing randomly with a small probability The value of a specific gene locus, i.e., changing the allocation device for that task. This involves a certain probability on a specific device. Perform local search mutations, such as: a) swap mutation: randomly swap the positions of two tasks in the sequence; b) reverse mutation: randomly select a segment of the sequence and reverse its order; c) insertion mutation: randomly select a task and insert it into another random position in the sequence. After performing any sorting-level mutation, calculate the tool change count of the task sequence before and after the mutation. A mutation is accepted only if the tool change count of the sequence after mutation does not increase, or if the increase is acceptable to the global optimization. This constraint-aware mutation strategy effectively avoids disrupting existing low-tool-change task groups.

[0170] The algorithm terminates when the number of iterations reaches a preset maximum, or when the improvement in the optimal fitness value over multiple consecutive generations is less than a preset threshold. The algorithm outputs the chromosome with the highest fitness in the current generation of the population and decodes it into the final task assignment matrix. and task sorting vector set That is, the optimal scheduling scheme.

[0171] S4. Based on the scheduling scheme output in step S3, control the corresponding laser devices to execute their assigned task sequences in sequence.

[0172] MES distributes the optimal scheduling scheme (task allocation and sequence) to the corresponding devices. The devices execute tasks in sequence, automatically calling the preset programs for regular tasks or the G-code generated by S2 for image tasks.

[0173] Automatic tool change and process switching: When adjacent tasks in a task sequence require different laser head types, the equipment automatically executes the standard tool change procedure: safe positioning → automatic laser head replacement → in-situ detection and calibration. After completion, it automatically continues to the next task.

[0174] Real-time monitoring and dynamic adjustment: The system monitors the power, position, progress, and other statuses of the equipment in real time. When an anomaly occurs (such as tool change failure or process parameter deviation), the system triggers a tiered alarm and can automatically pause the task, adjust parameters, or initiate task rescheduling for the faulty equipment to ensure production continuity.

[0175] A specific example: Includes multiple processing functions + AI image processing tasks.

[0176] For automotive parts processing orders and cultural and creative product customization orders, there are 60 regular processing tasks (covering the five major metal processing functions and the three major non-metal processing functions) and 20 AI image processing tasks:

[0177] I. Routine Tasks

[0178] 1. 30 metal processing tasks: 5 metal engravings (stainless steel panel patterns), 10 metal markings (parts traceability codes), 5 metal cuttings (carbon steel parts), 5 metal weldings (aluminum alloy frames), and 5 metal cleanings (mold oil stain cleaning).

[0179] 2. 30 non-metallic processing tasks: 10 non-metallic carvings (wooden ornaments), 10 non-metallic markings (plastic shell markings), and 10 non-metallic cuttings (acrylic decorative parts).

[0180] 3. Task priority distribution: Level 5 (urgent) 10 tasks, Level 4 15 tasks, Level 3 20 tasks, Level 2 10 tasks, Level 1 5 tasks.

[0181] 4. AI Image Processing Tasks: 20 customized cultural and creative product tasks (15 for non-metallic materials: wood carving, acrylic decoration cutting; 5 for metallic materials: metal pendant marking, personalized welding trajectory). Users can upload hand-drawn patterns or album images (such as logos, cartoon characters) through a mobile APP.

[0182] II. Order Import and Task Analysis

[0183] 1. The MES system receives regular order data from the ERP system via the OPC UA protocol. The task management module automatically parses and generates a processing task list, specifying the material type, processing function, priority, and estimated processing time for each task (automatic tool change time 2-3 minutes, AI processing time ≤ 3 seconds).

[0184] 2. Twenty AI image processing tasks were uploaded by users via a mobile app: 12 were hand-drawn cartoon patterns (brush thickness 3-5 pixels), and 8 were logo images imported from the album (resolution 500dpi). The app automatically completed the cropping and editing before uploading to the MES system.

[0185] 3. The MES system calls the AI ​​image analysis engine (to perform image preprocessing, contour extraction, path optimization, process adaptation, and G-code simulation) to process the image, completing the processing within 3 seconds and extracting high-precision estimated processing time, energy consumption, and required laser type, among other process attribute parameters. Simultaneously, it performs rapid parameterization calculations on routine tasks from the ERP system. Finally, the process attribute parameters of all tasks are merged to generate a unified list of processing task parameters.

[0186] III. Intelligent Task Allocation

[0187] The intelligent scheduling algorithm module, based on the real-time status of 10 devices (initial state: 5 devices with metal laser heads and 5 devices with non-metal laser heads, all in standby mode), and a unified list of task process parameters and device status, constructs a multi-objective model aimed at minimizing production cycle time, load imbalance, energy consumption cost, and tool change frequency. An improved genetic algorithm is used for task sorting and allocation. Algorithm parameters are set as follows: population size 80, crossover probability 0.7, mutation probability 0.03, 80 generations, and weighting coefficients α=0.4, β=0.3, γ=0.1, δ=0.2 (balancing production cycle time and tool change frequency). The final optimal task allocation scheme is:

[0188] 1. 30 routine metal tasks (including five major processing functions) + 5 metal image tasks are assigned to 5 metal laser head devices. Tasks with the same processing function are assigned consecutively (e.g., welding tasks are assigned to 2 devices).

[0189] 2. Non-metallic routine tasks (30, including three major processing functions) + 15 non-metallic image tasks are allocated to 5 non-metallic laser head devices. The engraving tasks and cutting tasks are allocated separately to improve efficiency.

[0190] 3. Since the scheduling model takes the total number of tool changes as an independent optimization objective, and the genetic algorithm adopts a task coding and operation strategy aimed at reducing tool changes, the final generated scheme realizes the clustering and arrangement of similar machining tasks. Therefore, only 3 tool changes are required, which reduces the number of tool changes by 7 compared with the traditional scheduling method that takes completion time as the single objective, and significantly reduces non-productive downtime.

[0191] IV. Processing Execution and Terminal Control

[0192] 1. Operators log in to the operating software via computer terminal, view the task allocation plan, and issue processing instructions after confirmation. The computer and equipment establish a binding connection, and control permissions for other terminals are locked.

[0193] 2. After receiving the instruction, the equipment control unit determines whether the current laser head type matches the task. For 3 devices that need tool replacement, it automatically performs a safe shutdown, moves to the safe tool replacement position, and triggers an audible and visual prompt and a tool replacement notification from the MES system.

[0194] 3. The electric intelligent automatic tool changer performs tool changing actions according to the preset program. The in-situ detection sensor verifies the installation status (positioning accuracy ±0.015mm). After passing the verification, the equipment automatically calibrates the focus position and adapts the parameters according to the processing function (such as welding penetration depth 2mm, engraving depth 0.3mm).

[0195] V. Processing Execution Example

[0196] 1. Metal welding task: The equipment calls the weld trajectory G code generated by AI, the welding power is set to 1500W, the welding speed is 0.8m / min, the weld penetration is monitored in real time (sampling accuracy ±0.1mm), and the weld strength test after processing reaches 92% of the base material, with no porosity defects.

[0197] 2. Non-metallic engraving task: Users upload hand-drawn patterns via mobile APP. AI algorithm generates a 0.1-0.5mm gradient depth path based on grayscale values. The engraving power is 30W, the processing speed is 1m / min, and the engraved pattern has a detail resolution of 1200dpi with no carbonized burrs.

[0198] 3. Metal cleaning task: The equipment uses AI image recognition to identify oily areas on the mold and adopts a local cleaning mode. The cleaning power is 800W, the cleaning efficiency is 0.6m² / min, and the surface roughness Ra after cleaning is 0.6μm, which meets the requirements of subsequent processing.

[0199] The foregoing has described in detail an embodiment of a laser flexible processing collaborative control method based on image analysis and intelligent scheduling. Based on the laser flexible processing collaborative control method based on image analysis and intelligent scheduling described in the above embodiment, this invention also provides a laser flexible processing collaborative control system based on image analysis and intelligent scheduling corresponding to the method.

[0200] Figure 3 This is a schematic block diagram of a laser flexible processing collaborative control system based on image analysis and intelligent scheduling, provided as an embodiment of the present invention. In this embodiment, the laser flexible processing collaborative control system 300 based on image analysis and intelligent scheduling can be divided into multiple functional modules according to the functions it performs. A module, as referred to in this invention, is a series of computer program segments that can be executed by at least one processor and perform a fixed function, and is stored in memory.

[0201] The task set receiving module 310 is used to receive a task set containing regular task instructions and image files; wherein, the regular task instructions contain basic process requirements for determining the final process attribute parameters of the regular task, and the image files contain the pattern to be processed.

[0202] The image task parameter extraction module 320 is used to perform image processing and path planning on the pattern to be processed in the image file to extract the process attribute parameters of the image processing task; the process attribute parameters include: estimated processing time, energy consumption per unit time, and required laser type.

[0203] The task scheduling optimization model construction and solution module 330 is used to construct and solve the multi-device task scheduling optimization model.

[0204] a) Merge the process attribute parameters of regular tasks with the process attribute parameters generated by image processing tasks into a unified task process parameter list;

[0205] b) Based on the task process parameter list, the current availability status of each laser device and the laser head loading type, construct a multi-objective optimization model with the optimization objectives of minimizing production cycle, equipment load imbalance, total energy consumption and total number of tool changes, and constraints of equipment capacity and task priority;

[0206] c) Use a genetic algorithm to solve the multi-objective optimization model and output the optimal task-equipment allocation scheme and the task processing sequence on each equipment.

[0207] The task allocation and execution module 340 is used to construct the scheduling scheme output by the solution module based on the task scheduling optimization model, and control the corresponding laser equipment to execute its assigned task sequence in sequence.

[0208] The laser flexible processing collaborative control system based on image analysis and intelligent scheduling in this embodiment is used to implement the aforementioned laser flexible processing collaborative control method based on image analysis and intelligent scheduling. Therefore, the specific implementation of this system can be found in the embodiment section of the laser flexible processing collaborative control method based on image analysis and intelligent scheduling mentioned above. Thus, the specific implementation can be referred to the description of the corresponding embodiments, and will not be elaborated here.

[0209] Furthermore, since the laser flexible processing collaborative control system based on image analysis and intelligent scheduling in this embodiment is used to implement the aforementioned laser flexible processing collaborative control method based on image analysis and intelligent scheduling, its function corresponds to the function of the above method, and will not be repeated here.

[0210] Figure 4 This is a schematic diagram of a terminal 400 provided in an embodiment of the present invention, including: a processor 410, a memory 420, and a communication unit 430. The processor 410 is used to implement the steps of the above-described embodiment of the laser flexible processing collaborative control method based on image analysis and intelligent scheduling when implementing the laser flexible processing collaborative control program based on image analysis and intelligent scheduling stored in the memory 420.

[0211] This invention also provides a computer storage medium, which may be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc. The computer storage medium stores a laser flexible processing collaborative control program based on image analysis and intelligent scheduling. When executed by a processor, this program implements the steps of the aforementioned embodiment of the laser flexible processing collaborative control method based on image analysis and intelligent scheduling.

[0212] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A collaborative control method for laser flexible processing based on image analysis and intelligent scheduling, characterized in that, Includes the following steps: S1, receive a task set containing routine task instructions and image files; wherein, the routine task instructions contain basic process requirements for determining the final process attribute parameters of the routine task, and the image files contain the pattern to be processed; S2, perform image processing and path planning on the pattern to be processed in the image file to extract the process attribute parameters of the image processing task; the process attribute parameters include: estimated processing time, energy consumption per unit time, and required laser type; S3, construct and solve the multi-device task scheduling optimization model; S3.1: Merge the process attribute parameters of regular tasks with the process attribute parameters generated by image processing tasks into a unified task process parameter list; S3.2: Based on the task process parameter list, the current availability status of each laser device and the laser head loading type, construct a multi-objective optimization model with the optimization objectives of minimizing production cycle, equipment load imbalance, total energy consumption and total number of tool changes, and with equipment capacity and task priority as constraints; S3.3: Use a genetic algorithm to solve the multi-objective optimization model and output the optimal task-equipment allocation scheme and the task processing sequence on each device; S4. Based on the scheduling scheme output in step S3, control the corresponding laser devices to execute their assigned task sequences in sequence. The multi-objective optimization function of the multi-objective optimization model is expressed as: in, These are weighting coefficients that are dynamically adjusted based on the production scenario. Total production cycle The longest total time required for all devices to complete their assigned tasks is expressed as: In the formula, For the total number of devices, Total number of tasks; For binary decision variables, when the task Assigned to device The value is 1 if it is true, and 0 otherwise. For the task Estimated processing time; For equipment Total number of tool changes; The standard time for a single tool change; Equipment load imbalance The coefficient of variation of the total processing time of each piece of equipment is used to measure the degree of difference between the total processing times of each piece of equipment, and is expressed as: in, Indicates equipment Total processing time This represents the average processing time for all equipment. Total energy cost The total energy cost consumed for all equipment to perform machining and tool changing processes is expressed as: in, For the task Energy consumption per unit time; Energy consumption per unit time during the tool changing process; Total number of tool changes The calculation method is to iterate through each device. The assigned task sequence is used to count the number of tool changes caused by the different laser head types required for adjacent tasks, and these counts are accumulated and represented as follows: in, To be assigned to the device The number of tasks; Indicates equipment Upper The "material-processing function" combination identifier for each task; For the discriminant function, when The value is 1 if it is active, and 0 otherwise.

2. The laser flexible processing collaborative control method based on image analysis and intelligent scheduling according to claim 1, characterized in that, Step S2 involves performing image processing and path planning on the pattern to be processed in the image file to extract the process attribute parameters of the image processing task. Specifically, this includes: S2.1: Perform noise filtering, contrast enhancement, edge sharpening, and background redundancy removal on the pattern to be processed to generate an enhanced image; S2.2: The enhanced image is binarized to obtain a binary image that distinguishes the processing area from the background; an edge detection algorithm is used to traverse the binary image to extract the edge pixel set of the processing area and record the coordinate information of each pixel; the extracted discrete edge feature points are connected and regularized, and a smooth and continuous closed or non-closed contour is generated by a straight line fitting or circular arc fitting algorithm; the regularized contour is converted into vector path data. S2.3: Based on the combined objective of minimizing total processing time and motion stability, the vector path is optimized; S2.4, Based on the material type and processing function of the processing task, call the corresponding processing strategy package from the preset path planning strategy library; according to the parameter rules in the called processing strategy package, perform parameterized adjustment on the optimized vector path to generate the final processing path and associated process parameter set; S2.5, the final processing path and associated process parameter set are converted into G-code executable by the target laser equipment; based on the kinematic parameters of the CNC platform corresponding to the target laser equipment, motion simulation is performed on the G-code, the total processing time is extracted from the motion simulation results as the estimated processing time, and the average power is extracted from the power change curve during the simulation as the basis for the energy consumption per unit time. The required laser type is determined according to the material type and processing function.

3. The laser flexible processing collaborative control method based on image analysis and intelligent scheduling according to claim 2, characterized in that, Step S2.3 specifically includes: Obtain the kinematic constraint parameters of the target CNC platform, including the maximum speed, maximum acceleration, S-curve acceleration / deceleration time constant, and the maximum centripetal acceleration allowed by the laser head for each axis; A comprehensive cost function is constructed to evaluate the merits of each path segment. A comprehensive cost is defined for each path segment, which is a weighted sum of the estimated travel time cost and the processing stability risk cost. The estimated travel time cost includes the constant speed time based on the path segment length and the time loss due to acceleration and deceleration control. The processing stability risk cost is calculated based on the motion jitter risk caused by the coupling of local curvature of the path and feed rate, and the processing thermal shock risk caused by the coupling of excessively rapid changes in laser power per unit path length and feed rate. Based on the comprehensive cost function, a heuristic search algorithm is used to plan the empty travel path with the minimum total cost between processing areas; The corners in the path are adaptively smoothed. The smoothing radius is dynamically determined based on the preset speed and maximum centripetal acceleration at the corner. Redundant points in the smoothed path are removed based on dynamic tolerance, which is related to the local curvature.

4. The laser flexible processing collaborative control method based on image analysis and intelligent scheduling according to claim 2, characterized in that, The path planning strategy library has at least two processing strategy packages pre-built: speed-first strategy package and accuracy-first strategy package.

5. The laser flexible processing collaborative control method based on image analysis and intelligent scheduling according to claim 1, characterized in that, The constraints of the multi-objective optimization model include: Task assignment uniqueness constraint: Each task must be assigned to one and only one device; Equipment capability adaptability constraint: The type of laser head assigned to the task must match the current processing capability of the equipment or the processing capability that can be obtained through tool changing; Equipment load factor constraint: The total processing time of each piece of equipment shall not exceed its available production time; Task priority constraint: On the same device, higher priority tasks must be processed before lower priority tasks. Tool changing process constraints: The cumulative tool changing time of a single machine must not exceed the target proportion of its available production time.

6. The laser flexible processing collaborative control method based on image analysis and intelligent scheduling according to claim 5, characterized in that, The genetic algorithm's chromosomes use a two-layer coding structure to represent a complete scheduling scheme: the first layer is the allocation code, which is a code with a length equal to the total number of tasks. sequence Among them, gene loci The value is the task Assigned equipment number This encoding defines the task assignment matrix. ; The second layer is a sorting code, for each device. Its sorting code It is a list of all task IDs assigned to the device, arranged in processing order, and this encoding defines the task sorting vector; The crossover operation in a genetic algorithm includes: assigning encodings to two parent individuals. Perform two-point crossover; after the crossover is complete, construct the sorting code for each device for the offspring individuals. At that time, heuristic rules are used to prioritize arranging tasks with the same required laser type adjacently in the sequence; The mutation operation of the genetic algorithm includes sorting layer mutation. After the sorting layer mutation is executed, the number of tool changes in the task sequence of the device before and after the mutation is calculated and compared, and whether the number of tool changes increases is used as one of the evaluation factors for whether to accept the mutation result.

7. A collaborative control system for laser flexible processing based on image analysis and intelligent scheduling, characterized in that, For implementing the collaborative control method for laser flexible processing as described in claim 1, the system comprises: The task set receiving module is used to receive a task set containing regular task instructions and image files; wherein, the regular task instructions contain basic process requirements for determining the final process attribute parameters of the regular task, and the image files contain the pattern to be processed; The image task parameter extraction module is used to perform image processing and path planning on the pattern to be processed in the image file to extract the process attribute parameters of the image processing task; the process attribute parameters include: estimated processing time, energy consumption per unit time, and required laser type; The task scheduling optimization model construction and solution module is used to construct and solve multi-device task scheduling optimization models. a) Merge the process attribute parameters of regular tasks with the process attribute parameters generated by image processing tasks into a unified task process parameter list; b) Based on the task process parameter list, the current availability status of each laser device and the laser head loading type, construct a multi-objective optimization model with the optimization objectives of minimizing production cycle, equipment load imbalance, total energy consumption and total number of tool changes, and constraints of equipment capacity and task priority; c) Use a genetic algorithm to solve the multi-objective optimization model and output the optimal task-equipment allocation scheme and the task processing sequence on each equipment; The task allocation and execution module is used to construct the scheduling scheme output by the solution module based on the task scheduling optimization model, and control the corresponding laser equipment to execute its assigned task sequence in sequence.

8. A terminal, characterized in that, include: The memory is used to store the collaborative control program for laser flexible processing based on image analysis and intelligent scheduling; A processor is configured to implement the steps of the laser flexible processing collaborative control method based on image analysis and intelligent scheduling as described in any one of claims 1 to 6 when executing the laser flexible processing collaborative control program based on image analysis and intelligent scheduling.

9. A computer-readable storage medium, characterized in that, The readable storage medium stores a laser flexible processing collaborative control program based on image analysis and intelligent scheduling. When the laser flexible processing collaborative control program based on image analysis and intelligent scheduling is executed by the processor, it implements the steps of the laser flexible processing collaborative control method based on image analysis and intelligent scheduling as described in any one of claims 1 to 6.

Images