Intelligent die cutting path control optimization method, system and equipment for heterogeneous packaging boxes and medium

Abstract

The invention discloses an intelligent die cutting path control optimization method, system and equipment for heterogeneous packaging boxes and a medium, and relates to the technical field of die cutting path control, and the method comprises the steps: recognizing a high-curvature die cutting path through the frequency domain distribution of curvature density, determining an initial preposition proportion according to the length ratio of the high-curvature die cutting path, and generating a plurality of candidate preposition proportions in a continuous range; according to curvature density integral values, sorting and forward moving corresponding high-curvature paths, and generating different die cutting path sequences; performing virtual die cutting in combination with digital twinning and extracting process features, and obtaining an evaluation value of each preposition proportion by utilizing reinforcement learning; and finally, the stable interval is analyzed through the continuous response curve, the optimal high-curvature die cutting path preposition proportion is determined and used for actual die cutting, and the problems that in the die cutting process of the heterogeneous packaging box, due to the unreasonable die cutting sequence, the decorative edge deforms, contour traction is conducted, and the edge precision is unstable are solved.

Description

Technical Field

[0001] This invention relates to the field of die-cutting path control technology, and more specifically, to an intelligent die-cutting path control optimization method, system, device, and medium for heterogeneous packaging boxes. Background Technology

[0002] As the packaging industry continues to demand higher standards for appearance consistency, structural precision, and production efficiency, the die-cutting of heterogeneous packaging boxes is gradually shifting from being driven by manual experience to intelligent die-cutting control technology based on digital path planning. Current technologies typically plan straight and curved die-cutting paths uniformly based on the unfolded geometry of the packaging box, and drive the die-cutting equipment to complete the cutting operation according to a predetermined path sequence. Some solutions introduce simulation or path optimization algorithms to reduce unnecessary strokes or improve processing efficiency, but limited attention is paid to the structural impact of different types of die-cutting paths on their execution sequence.

[0003] In the intelligent die-cutting process of heterogeneous packaging boxes, high-curvature die-cutting paths typically correspond to decorative edges, folded corners, or complex contour areas of the packaging box. Their execution order directly affects the stress release method and structural constraint state of the material during the die-cutting process. The pre-execution ratio of high-curvature die-cutting paths reflects the degree to which such paths are executed in advance in the overall die-cutting sequence. Different values ​​of this ratio will change the timing of decorative edge formation, thereby affecting the material's shape stability and contour retention ability in subsequent die-cutting processes.

[0004] When the proportion of high-curvature die-cutting paths is too high, the decorative edge is cut before the overall structure has formed effective constraints. Although the target outline can be formed in advance, the material is still in a free state and lacks sufficient support and connection. As subsequent straight die-cutting paths are executed, the overall stress state of the material changes, which can easily pull on the formed decorative edge, causing the outline to shift or be carried away, thus affecting the final appearance accuracy.

[0005] Conversely, when the proportion of high-curvature die-cutting paths at the beginning is low, decorative edges are arranged to be executed later in the die-cutting sequence. At this time, the material has accumulated a certain amount of residual cutting stress after a large number of straight-line die-cuts. During high-curvature die-cutting, these unreleased stresses will be concentrated and released during the curve cutting stage. Combined with the springback effect of the material itself, this can easily cause the edge position to undergo multiple springback changes after cutting, thereby reducing the forming accuracy of the curve edge.

[0006] Whether the front-end ratio is too high or too low, it will negatively impact the intelligent die-cutting path of heterogeneous packaging boxes. An excessively high front-end ratio mainly causes the decorative edge to become unstable in subsequent processing, while an excessively low front-end ratio will lead to springback overlap and reduced edge precision during the curve die-cutting stage. Both of these problems are difficult to avoid by a single fixed path sequence, and their performance varies significantly under different packaging structures and material conditions.

[0007] However, existing technologies mostly rely on empirical settings or simple rules to control the front ratio of high curvature die-cutting paths. They lack systematic analysis methods to understand the impact of changes in the front ratio on structural constraints, stress release, and morphological stability. This makes it impossible to adaptively determine a reasonable range of the front ratio in different heterogeneous packaging box scenarios. Consequently, the die-cutting path sequence cannot simultaneously take into account the precision of the decorative edge and the overall processing stability, thus hindering further improvement in the level of intelligent die-cutting path control.

[0008] To address the above problems, this invention proposes a solution. Summary of the Invention

[0009] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide an intelligent die-cutting path control optimization method, system, device, and medium for heterogeneous packaging boxes. By adaptively adjusting the pre-proportion of the high curvature die-cutting path based on curvature distribution characteristics, the present invention solves the problems of decorative edge deformation, contour pulling, and unstable edge accuracy caused by unreasonable die-cutting sequence during the die-cutting process of heterogeneous packaging boxes.

[0010] To achieve the above objectives, the present invention provides the following technical solution:

[0011] A method for intelligent die-cutting path control and optimization of heterogeneous packaging boxes includes the following steps: identifying high-curvature die-cutting paths based on the frequency domain distribution of curvature density, and using the length ratio of high-curvature die-cutting paths in all die-cutting paths as the initial high-curvature die-cutting path front ratio; generating several second high-curvature die-cutting path front ratios within a continuously varying range, centered on the initial high-curvature die-cutting path front ratio; sorting the high-curvature die-cutting path set according to the curvature density integral value, selecting high-curvature die-cutting paths with corresponding ratios according to the high-curvature die-cutting path front ratios and moving them forward as a whole, while maintaining the remaining... The relative order of the paths remains unchanged, generating a die-cutting path order; the preceding proportions of each second high curvature die-cutting path are used to generate the die-cutting path order, and virtual die-cutting is performed based on digital twins; features are extracted from the virtual die-cutting process, and a proportional scoring model is trained based on reinforcement learning to generate proportional evaluation values ​​corresponding to the preceding proportions of each second high curvature die-cutting path; a continuous response curve is constructed by combining the preceding proportions of each second high curvature die-cutting path with the corresponding proportional evaluation values, and the optimal preceding proportions of the high curvature die-cutting path are selected by analyzing the stable interval of the response change rate, and a die-cutting path order is generated for die-cutting.

[0012] A smart die-cutting path control and optimization system for heterogeneous packaging boxes includes a frequency domain distribution recognition module, a second high-curvature die-cutting path pre-proportion module, a die-cutting path sequence module, a virtual die-cutting module, a proportion evaluation module, and a die-cutting module. The frequency domain distribution recognition module identifies high-curvature die-cutting paths based on the frequency domain distribution of curvature density, and uses the length ratio of high-curvature die-cutting paths in all die-cutting paths as the initial high-curvature die-cutting path pre-proportion. The second high-curvature die-cutting path pre-proportion module generates several second high-curvature die-cutting path pre-proportions within a continuously varying range, centered on the initial high-curvature die-cutting path pre-proportion. The die-cutting path sequence module sorts the high-curvature die-cutting path set according to the curvature density integral value, and determines the pre-proportion based on the high curvature... The die-cutting path sequence is generated by selecting the corresponding proportion of high-curvature die-cutting paths in the lead-in ratio and moving them forward as a whole, while keeping the relative order of the remaining paths unchanged. The virtual die-cutting module generates the die-cutting path sequence based on the lead-in ratio of each second high-curvature die-cutting path and performs virtual die-cutting based on digital twins. The ratio evaluation module extracts features from the virtual die-cutting process, trains a ratio scoring model based on reinforcement learning, and generates ratio evaluation values ​​corresponding to the lead-in ratio of each second high-curvature die-cutting path. The die-cutting module constructs a continuous response curve with the lead-in ratio of each second high-curvature die-cutting path and the corresponding ratio evaluation value. By analyzing the stable range of the response change rate, the optimal lead-in ratio of high-curvature die-cutting paths is selected and the die-cutting path sequence is generated for die-cutting.

[0013] An electronic device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to execute an intelligent die-cutting path control optimization method for heterogeneous packaging boxes.

[0014] A computer-readable storage medium storing a computer program that, when executed by a processor, implements an intelligent die-cutting path control optimization method for heterogeneous packaging boxes.

[0015] The technical effects and advantages of the intelligent die-cutting path control optimization method, system, equipment and medium for heterogeneous packaging boxes of this invention are as follows:

[0016] 1. This invention introduces a high-curvature die-cutting path identification method based on the frequency domain distribution of curvature density, and uses the length ratio of the high-curvature die-cutting path in all die-cutting paths as the core parameter for pre-proportion adjustment. Combined with a continuous proportion generation and overall path forward reorganization mechanism, it achieves fine-grained control of the execution timing of high-curvature die-cutting paths. This method can adaptively generate multiple sets of candidate die-cutting path sequences for different heterogeneous packaging box structures while maintaining the relative order stability of the original die-cutting paths. This avoids the problems of premature or delayed decorative edge formation caused by traditional experience-based fixed path sequences, fundamentally reducing the risks of insufficient structural constraints, contour pulling, and edge deformation caused by unreasonable die-cutting sequences. It also improves the adaptability of die-cutting path planning to complex geometries and multiple material properties.

[0017] 2. This invention combines digital twin virtual die-cutting with reinforcement learning to systematically quantify the structural retention, contour traction effect, stress accumulation, and edge springback characteristics under different pre-cutting path ratios for high curvature die-cutting paths. It also constructs a ratio evaluation model and continuous response curves, thereby selecting the optimal pre-cutting ratio from multiple ratio schemes where the evaluation response tends to be stable. This technical approach avoids the limitations of judging the merits of pre-cutting ratios solely based on a single indicator or extreme value. It ensures that the final determined die-cutting path sequence achieves a balance between decorative edge accuracy, material morphology stability, and overall die-cutting consistency, effectively improving the finished product quality stability and process reliability of heterogeneous packaging box die-cutting processing, and has significant engineering application value. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating an intelligent die-cutting path control optimization method for heterogeneous packaging boxes according to the present invention.

[0019] Figure 2 This is a schematic diagram of the intelligent die-cutting path control and optimization system for heterogeneous packaging boxes according to the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0021] Example 1, Figure 1 This invention presents an intelligent die-cutting path control optimization method for heterogeneous packaging boxes, comprising the following steps:

[0022] S1, based on the frequency domain distribution of curvature density, identify high curvature die-cutting paths, and use the length ratio of high curvature die-cutting paths in all die-cutting paths as the initial high curvature die-cutting path front ratio;

[0023] In this embodiment, the high-curvature die-cutting path is identified based on the frequency domain distribution of curvature density, and the proportion of the length of the high-curvature die-cutting path in all die-cutting paths is used as the initial high-curvature die-cutting path pre-proportion, specifically:

[0024] Discretely sample the die-cutting paths of heterogeneous packaging boxes and represent each die-cutting path as a coordinate sequence arranged in path order;

[0025] Based on the coordinate sequence, the curvature change of adjacent sampling points is calculated, and a curvature density sequence arranged sequentially along the die-cutting path is constructed.

[0026] The curvature density sequence is subjected to frequency domain transformation to obtain the spectral distribution of curvature density in the path order dimension;

[0027] Based on the spectrum distribution, the path intervals corresponding to the high-frequency components are extracted, and the path intervals are mapped back to the original die-cutting paths and marked as a set of high-curvature die-cutting paths.

[0028] The cumulative path length of the high curvature die-cutting path set is calculated and normalized with the total length of all die-cutting paths to obtain the initial high curvature die-cutting path front ratio.

[0029] It should be noted that the following is a feasible example of calculating the curvature change between adjacent sampling points to construct a curvature density sequence:

[0030] Calculate the discrete curvature of three consecutive points on the path:

[0031] ;

[0032] In the formula, Let be the curvature intensity at the k-th sampling position. is the planar coordinate of the k-th sampling point; i is the die-cutting path number.

[0033] In this embodiment, a heterogeneous packaging box refers to a packaging box that contains different material types, different thickness areas, or different structural connection forms in the same packaging structure. During the die-cutting process, the inconsistent stress and deformation responses of each area of ​​such a packaging box will affect the die-cutting path sequence and cutting stability.

[0034] In this embodiment, the die-cutting path refers to the cutting trajectory formed by the continuous movement of the cutting tool in a plane according to a preset execution sequence during the process of cutting the packaging box. Each die-cutting path corresponds to an actual cutting boundary on the packaging box.

[0035] Discrete sampling in this embodiment refers to dividing a continuous die-cutting path into several continuous sampling points according to a unified path advancement rule, so that the die-cutting path can be described in the form of a discrete point sequence for subsequent curvature change analysis and path feature extraction.

[0036] In this embodiment, the coordinate sequence refers to the set of sampling points arranged in the order of die-cutting execution. Each sampling point contains its spatial position in the die-cutting plane, which is used to completely describe the movement trajectory of the tool along the die-cutting path.

[0037] In this embodiment, the curvature change refers to the degree of directional deflection of the die-cutting path between adjacent sampling points. It is used to reflect the bending situation when the path transitions from one direction to the next. This change is used to characterize the geometric complexity of the die-cutting path.

[0038] In this embodiment, curvature density refers to the intensity of directional change accumulated within a unit path advancement length during the die-cutting process. This description takes into account the relationship between the path bending amplitude and the path advancement distance, and is used to distinguish between gentle turning paths and sharp bending paths.

[0039] In this embodiment, the curvature density sequence refers to the set of curvature densities arranged in the order of execution along the die-cutting path. This sequence fully reflects the change distribution of bending intensity from the beginning to the end of the die-cutting path and is the basic input for frequency domain analysis.

[0040] In this embodiment, frequency domain transformation refers to converting the curvature density sequence from the path sequence dimension to the change frequency dimension, which is used to analyze the change rhythm of curvature during the path advancement process, thereby revealing the distribution characteristics of the degree of bend concentration as the path advances.

[0041] In this embodiment, the spectral distribution refers to the distribution of curvature density at different frequencies of change, where the part with higher frequency of change corresponds to the region in the path where the direction changes multiple times within a short distance.

[0042] In this embodiment, the high-frequency component refers to the part of the spectrum distribution that corresponds to rapidly changing components. This part reflects the path segment with dense bends and frequent turns in the die-cutting path, and is an important basis for identifying high curvature die-cutting paths.

[0043] In this embodiment, the path interval refers to a continuous range of paths in the die-cutting path sequence, which is used to accurately map the high-frequency change region identified in the frequency domain back to the specific position in the original die-cutting path.

[0044] In this embodiment, the high curvature die-cutting path set refers to the set of die-cutting paths that have significant bending characteristics during the die-cutting process, obtained by mapping path intervals. The paths in this set are more likely to cause material stress concentration and structural stability changes during the cutting process.

[0045] In this embodiment, the cumulative path length refers to the sum of the actual cutting lengths of all paths in the high curvature die-cutting path set, which is used to reflect the proportion of high curvature paths in the overall die-cutting task.

[0046] In this embodiment, the total length of all die-cutting paths refers to the sum of the lengths of all die-cutting paths required to complete the die-cutting of the entire packaging box, and is used as a unified length reference basis.

[0047] In this embodiment, the initial high curvature die-cutting path forward ratio refers to the proportion of the cumulative path length of the high curvature die-cutting path set to the total length of all die-cutting paths. This ratio is used to characterize the natural distribution of high curvature paths in the overall die-cutting process before path order adjustment, and serves as a benchmark for the subsequent forward ratio construction.

[0048] It should be noted that the above-mentioned high curvature die-cutting path recognition method based on curvature density frequency domain distribution is not only used to classify the geometric shape of the path, but also to provide a unified proportional expression basis for subsequent die-cutting path sequence adjustment, so that the degree of forward movement of the high curvature path can be continuously adjusted by the path length ratio.

[0049] It should be noted that by introducing frequency domain analysis in the path sequence dimension, the interference of local path jitter or single-point anomalies on the high curvature recognition results can be effectively suppressed, making the obtained high curvature die-cutting path set more concentrated in the bending area that has a real impact on material stress and structural stability, thus providing a stable and consistent input basis for subsequent scale generation, virtual die-cutting and scoring model training.

[0050] S2, with the initial high curvature die-cutting path front ratio as the center, generates several second high curvature die-cutting path front ratios within a continuously varying range;

[0051] In this embodiment, the step of generating several second high-curvature die-cutting path front ratios within a continuously varying range, centered on the initial high-curvature die-cutting path front ratio, specifically involves:

[0052] The cumulative length of the high-curvature die-cutting path corresponding to the initial high-curvature die-cutting path front ratio is used as the reference forward shift length;

[0053] Based on the curvature density integral value of each path in the high curvature die-cutting path set, the high curvature die-cutting paths are sorted in descending order to form a path sorting sequence from high to low curvature.

[0054] Using the cumulative path length change points of adjacent paths in the path sorting sequence as the forward adjustment boundary, the boundary of the high curvature die-cut path subset that can participate in the forward adjustment is determined.

[0055] Within the path length range defined by the boundary of the high curvature die-cutting path subset, the reference forward shift length is gradually increased or decreased using the actual path length of a single high curvature die-cutting path as the adjustment unit, forming several candidate forward shift lengths;

[0056] Each candidate forward shift length is normalized to the total path length of all die-cutting paths to obtain the corresponding second high curvature die-cutting path forward shift ratio;

[0057] Apply adjacent difference constraints to the preceding proportions of the second high curvature die-cutting path, so that the sets of forward paths corresponding to adjacent proportions differ by at least one complete high curvature die-cutting path.

[0058] In this embodiment, the initial high-curvature die-cutting path front ratio refers to the natural proportion of the high-curvature die-cutting path in the total die-cutting path length before any path order adjustment is made. This ratio is derived from the correspondence between the cumulative length of the high-curvature die-cutting path set and the total length of the overall die-cutting path, and is used as the central reference for subsequent forward adjustment.

[0059] In this embodiment, the cumulative length of the high curvature die-cutting path refers to the sum of the actual cutting lengths of each high curvature die-cutting path in the set of high curvature die-cutting paths corresponding to the initial high curvature die-cutting path front ratio. This length is directly used to represent the size of the path that needs to be moved forward first in the die-cutting path sequence.

[0060] In this embodiment, the reference forward shift length refers to the cumulative length of the high curvature die-cutting path determined by the initial high curvature die-cutting path forward ratio. This length serves as the starting scale for forward shift adjustment and is used to limit the expansion or contraction of subsequent candidate forward shift lengths around this scale.

[0061] In this embodiment, the curvature density integral value refers to the overall bending strength characterization obtained by continuously accumulating the curvature density at each position of the path during the advancement along a single high-curvature die-cutting path. It is used to comprehensively reflect the degree of influence of the path on material bending and stress concentration throughout the entire cutting process.

[0062] In this embodiment, the path sorting sequence refers to the ordered sequence formed by arranging the high curvature die-cutting paths from largest to smallest according to the curvature density integral value corresponding to each high curvature die-cutting path. This sequence is used to identify which paths have a higher priority in forward adjustment in terms of geometric complexity and potential risks.

[0063] In this embodiment, the degree of bending refers to the overall turning strength of the path as reflected by the integral value of curvature density. This description takes into account both the frequency and duration of bending in the path, and is used to distinguish between slightly bent paths and severely bent paths.

[0064] In this embodiment, the forward adjustment boundary refers to the cumulative length change node formed as the actual length of each path is accumulated in the path sorting sequence. This node is used to clarify the position where the set of high curvature die-cut paths participating in the forward movement switches when the cumulative forward length changes.

[0065] In this embodiment, the high curvature die-cutting path subset boundary refers to the path segmentation position determined by the forward adjustment boundary. This boundary is used to limit which complete high curvature die-cutting paths can participate in the forward movement as a whole under a certain forward movement length condition, so as to avoid the path being split or truncated.

[0066] In this embodiment, the adjustment unit refers to the actual path length of a single complete high-curvature die-cut path as the minimum scale for the change in forward length. By introducing or removing each path, the change in forward length always corresponds to the change in the complete path set.

[0067] In this embodiment, the candidate forward shift length refers to a set of discrete forward shift length results formed by gradually increasing or decreasing the baseline forward shift length within the range defined by the boundary of the high curvature die-cutting path subset. Each candidate forward shift length corresponds to a specific high curvature die-cutting path subset.

[0068] In this embodiment, the total path length of all die-cutting paths refers to the sum of the lengths of all die-cutting paths required to complete the entire heterogeneous packaging box die-cutting task, and is used as a unified proportional conversion benchmark.

[0069] In this embodiment, the second high curvature die-cutting path front ratio refers to the ratio obtained after establishing a correspondence between each candidate forward length and the total path length of all die-cutting paths. This ratio is used to quantitatively represent the degree of fronting of the high curvature die-cutting path in the overall die-cutting sequence under different forward schemes.

[0070] In this embodiment, the adjacent difference constraint means that when generating the first ratio of the second high curvature die-cutting path, there must be at least one difference in the set of forward paths corresponding to two adjacent ratios, so as to ensure that the ratio change can truly reflect the structural change of the die-cutting path sequence.

[0071] In this embodiment, the forward path set refers to the set of high curvature die-cutting paths that are executed earlier in the die-cutting path sequence under a certain second high curvature die-cutting path forward ratio.

[0072] It should be noted that by limiting the change in the forward length to be adjusted in units of complete high curvature die-cutting paths, problems such as path splitting, partial truncation, or discontinuous execution order can be avoided during the path sequence generation process. This ensures that the forward ratio of each second high curvature die-cutting path can be directly mapped to an executable die-cutting path sequence.

[0073] It should be noted that the aforementioned continuous variation range does not rely on a fixed interval set by humans, but is naturally formed by the path length distribution and bending intensity sorting of the high curvature die-cutting path set itself. This ensures that the generated second high curvature die-cutting path front ratio is structurally consistent with the actual die-cutting path composition, providing a stable and interpretable ratio input for subsequent virtual die-cutting and scoring analysis.

[0074] S3, sort the set of high curvature die-cutting paths according to the curvature density integral value, select the corresponding proportion of high curvature die-cutting paths according to the front ratio of high curvature die-cutting paths and move them forward as a whole, while keeping the relative order of the remaining paths unchanged, and generate the die-cutting path order.

[0075] In this embodiment, the process of sorting the high-curvature die-cutting path set according to the curvature density integral value, selecting a corresponding proportion of high-curvature die-cutting paths based on the preceding proportion of high-curvature die-cutting paths and shifting them forward as a whole, while keeping the relative order of the remaining paths unchanged, to generate the die-cutting path order, is as follows:

[0076] For each second high curvature die-cutting path front ratio, based on its corresponding forward shift length, in the high curvature die-cutting path sorting sequence arranged in descending order of curvature density integral value, the path length is accumulated sequentially from the beginning of the sequence until the accumulated path length reaches the forward shift length, thereby determining the high curvature die-cutting path subset participating in the forward shift.

[0077] The high curvature die-cutting path subset is extracted as a whole from the original die-cutting path sequence and rearranged into preceding path segments according to their relative order in the sorting sequence.

[0078] The preceding path segment is inserted into the starting position of the die-cutting path sequence to form the adjusted preceding die-cutting path sequence;

[0079] For the remaining die-cutting paths that are not selected into the high curvature die-cutting path subset, their relative order in the original die-cutting path sequence remains unchanged, and they are sequentially spliced ​​to the preceding die-cutting path sequence.

[0080] The completed splicing path sequence is determined as the die-cutting path order corresponding to the preceding ratio of the second high curvature die-cutting path, and is used for subsequent virtual die-cutting processing.

[0081] It should be noted that the following are feasible examples for calculating the integral value of curvature density:

[0082] In the m-th die-cutting path, two adjacent sampling points form a straight line, and its direction angle is defined as:

[0083] ;

[0084] In the formula, , These represent the x and y coordinates of sampling point m in the die-cutting plane coordinate system, respectively. This represents the motion direction angle of the m-th die-cutting path on the k-th path segment.

[0085] The change in direction between two adjacent path segments is defined as:

[0086] ;

[0087] In the formula, To represent the magnitude of the directional change of the m-th die-cutting path at the k-th sampling position, it is used to characterize the bending intensity of the path at that position.

[0088] The path length of the k-th path segment is defined as follows:

[0089] ;

[0090] In the formula, This represents the actual path length of the m-th die-cutting path on the k-th sampling segment.

[0091] Curvature density is defined as:

[0092] ;

[0093] In the formula, This represents the curvature density value.

[0094] It should be noted that the following are feasible examples for calculating the integral value of curvature density:

[0095] ;

[0096] In the formula, The integral value of curvature density, The number of discrete sampling points for the m-th die-cutting path.

[0097] In this embodiment, the high-curvature die-cutting path set refers to the set of paths selected from the complete die-cutting path after segmental analysis of path direction changes. These paths exhibit significant directional changes, frequent turning points, and concentrated tool posture adjustment requirements within a unit travel distance. Such paths are more prone to causing transient tool deviation, uneven material stretching, or edge micro-tears during actual die-cutting processes. Therefore, they are identified separately and prioritized for optimization, rather than being treated uniformly across all die-cutting paths.

[0098] In this embodiment, the curvature density integral value is a comprehensive measure reflecting the cumulative level of overall turning strength of a single die-cutting path. It is not based on a single point or corner, but rather on the continuous accumulation of the degree of curvature along the direction of travel, reflecting the path's influence on tool stability and material stress continuity throughout the cutting process. The more concentrated the curvature changes and the more frequent the turning, the higher the corresponding integral value, thus being prioritized as a path with a greater impact on die-cutting stability during sorting.

[0099] In this embodiment, the curvature density integral value sorting refers to uniformly arranging all identified high-curvature die-cutting paths based on the aforementioned comprehensive measurement results. Paths with a greater impact on cutting stability are placed at the front, while paths with a relatively smaller impact are placed at the back. This sorting does not change the geometry of any path; it is only used as a selection basis for subsequent path order adjustments to ensure that the adjustment process has clear physical direction and interpretability.

[0100] In this embodiment, the high-curvature die-cutting path pre-positioning ratio is a pre-set path adjustment control variable used to limit the execution range of high-curvature paths that need to be executed in advance within the overall set of high-curvature paths. This ratio does not directly affect the number of paths, but rather guides the subsequent path selection process constrained by the forward shift length, thereby avoiding the sudden change in execution load caused by simply truncating based on quantity or position.

[0101] In this embodiment, the forward movement length is a path travel scale calculated based on the forward ratio of the high-curvature die-cutting path. It is used to characterize the coverage of the high-curvature path that needs to be moved forward as a whole in terms of cumulative travel distance. This length is used to perform segment-by-segment accumulation judgment in the sorted high-curvature die-cutting path sequence to ensure that the forward-moved path subset has continuity and integrity in the actual cutting journey, rather than being scattered.

[0102] In this embodiment, the high-curvature die-cutting path subset refers to the set of paths determined by accumulating the path travel length one by one from the beginning of the high-curvature path sequence arranged in descending order of curvature density integral value until the forward movement length constraint is satisfied. The paths in this subset have strong similarities in physical characteristics and have a concentrated impact on the dynamic response of the equipment and the stability of the materials during execution. Therefore, they are extracted as a whole and their execution positions are uniformly adjusted.

[0103] In this embodiment, the original die-cutting path sequence refers to the path execution order directly generated based on the pattern outline, machining logic, or design input before any sequence optimization. This order is usually based on geometric proximity or design generation order and is not specifically optimized for tool continuity stability or material stress evolution.

[0104] In this embodiment, the preceding path segment refers to a continuous execution segment formed by extracting the high-curvature die-cutting path subset from the original die-cutting path sequence as a whole, maintaining its relative order in the sorting sequence, and then recombining it. This path segment structurally maintains the inherent correlation between the high-curvature paths, avoiding unnecessary cutting impacts caused by isolated execution of individual paths due to rearrangement.

[0105] In this embodiment, the adjusted pre-die-cutting path sequence refers to the starting part of a new path execution sequence formed after inserting the pre-cutting path segment into the starting position of the die-cutting path sequence. By concentrating the high-curvature path in the early stage of die-cutting, the tool and material complete the difficult path before experiencing multiple superpositions of cutting stress, thereby reducing the risk of edge drift or cumulative shape errors in subsequent cutting stages.

[0106] In this embodiment, the remaining die-cutting paths refer to the path portions that were not selected into the high-curvature die-cutting path subset. These paths typically have a gentler change in direction and a relatively uniform cutting load. Therefore, in adjusting the path sequence, it is only required to maintain their relative order in the original die-cutting path sequence and sequentially splice them after the preceding path segment to maintain the coherence of the overall processing logic.

[0107] In this embodiment, the die-cutting path sequence refers to the final complete path execution sequence after the high-curvature path is moved forward and the remaining path sequences are maintained. This sequence not only reflects the geometric relationship of the path, but also implies a comprehensive consideration of cutting stability, material stress release rhythm, and tool dynamic response, and serves as the direct input basis for subsequent virtual die-cutting processing.

[0108] It should be noted that in this embodiment, when performing the high curvature die-cutting path forward processing, no changes are made to the geometry of the path, the cutting direction, or the cutting method itself. All adjustments are limited to the path execution order, thereby ensuring that the method is directly compatible with existing die-cutting equipment and existing pattern data, without introducing additional processing risks.

[0109] It should be noted that by concentrating the high-curvature die-cutting path in the early stage of die-cutting, the influence trend of the high-curvature path on the edge morphology evolution can be observed more clearly in the virtual die-cutting stage. This helps to identify potential unstable edge areas in advance and provides a more targeted reference basis for subsequent path fine-tuning, tool status verification, or material adaptation strategies, thereby improving the consistency and controllability of the overall die-cutting quality.

[0110] S4, generate the die-cutting path sequence by pre-proportioning each second high curvature die-cutting path, and perform virtual die-cutting based on digital twin;

[0111] In this embodiment, the step of generating the die-cutting path sequence by prioritizing the proportions of each second high curvature die-cutting path and performing virtual die-cutting based on digital twins specifically involves:

[0112] For each of the preceding proportions of the second high curvature die-cutting path, read the corresponding generated die-cutting path sequence and use the die-cutting path sequence as the execution input for virtual die-cutting;

[0113] Based on the geometric structure data, material hierarchy information, and motion constraints of the die-cutting equipment of the heterogeneous packaging boxes, a digital twin die-cutting model consistent with the actual die-cutting environment is constructed.

[0114] The die-cutting paths are sequentially mapped into the digital twin die-cutting model, and the virtual cutter is driven to move along the corresponding die-cutting paths in sequence according to the path order.

[0115] During the virtual die-cutting process, the cutting status, local connection relationships, and distribution of remaining support structures of the packaging box material are updated synchronously, forming a virtual die-cutting state sequence that changes as the path progresses;

[0116] Record the execution timing of each die-cutting path in the virtual die-cutting state sequence and the corresponding material constraint change process to generate virtual die-cutting process data that corresponds one-to-one with the preceding ratio of the second high curvature die-cutting path.

[0117] In this embodiment, the second high curvature die-cutting path pre-position ratio is a set of preset control quantities determined in the preceding steps to control the degree of advancement of the high curvature die-cutting path in the overall execution order. Its essential function is to limit the range of differences between different path order schemes, so as to generate multiple comparable die-cutting execution orders, so as to systematically evaluate the impact of different order strategies in the virtual die-cutting stage.

[0118] In this embodiment, generating the die-cutting path order by the second high curvature die-cutting path front ratio means that, based on the path forward shifting rules established in the previous embodiment, the forward shifting result of each high curvature path subset corresponding to the front ratio is completely mapped into a die-cutting path arrangement order that can be directly used for execution. This process only involves the reconstruction of the path execution order and does not involve changes to the path shape, cutting method, or processing object.

[0119] In this embodiment, the die-cutting path sequence refers to the complete path execution sequence obtained for a specific second high curvature die-cutting path pre-proportion. It clearly defines the execution order of each die-cutting path in the overall processing and serves as the direct input basis for the subsequent virtual die-cutting process, so that the path sequence corresponding to different pre-proportion ratios can be evaluated and compared one by one.

[0120] In this embodiment, virtual die-cutting refers to simulating and reproducing the execution process of the die-cutting path sequence in a virtual environment without performing actual physical cutting, using digital modeling and simulation methods, in order to observe the influence of different path sequences on material state evolution, structural support changes, and cutting continuity. This process is a common way of applying digital twins to processing simulation in the prior art, so its basic principles will not be elaborated here.

[0121] The digital twin die-cutting model in this embodiment is a virtual mapping model built based on the actual die-cutting scenario. It comprehensively incorporates the geometric structural features of heterogeneous packaging boxes, the superposition relationship of multi-layer materials, and the motion restrictions of die-cutting equipment, so that the virtual environment is consistent with the real die-cutting process at both the structural and behavioral levels, thereby ensuring that the virtual die-cutting results have reference value and engineering significance.

[0122] In this embodiment, the heterogeneous packaging box geometric structure data refers to the structural description information established for packaging boxes with different specifications, different folding methods and different contour shapes. It is used to clarify the positional relationship of each die-cutting path in space and the mutual influence between different areas during the cutting process. This data provides the basic geometric constraints for the digital twin model.

[0123] In this embodiment, the material layer information refers to the structural description of the materials used in the packaging box in terms of layering relationships, connection methods, and local support states. It is used to determine the impact of different cutting paths on the remaining connection strength and overall stability of adjacent areas during the virtual die-cutting process. This information changes gradually as the path is executed.

[0124] In this embodiment, the motion constraints of the die-cutting equipment refer to the behavioral restrictions on the virtual tool in terms of movement speed, turning mode and movement continuity. These constraints are used to ensure that the movement process of the virtual tool conforms to the actual working characteristics of the real equipment. This part belongs to the conventional constraint settings in digital twin modeling and is existing technology, so it will not be elaborated further.

[0125] In this embodiment, mapping the die-cutting path sequence to the digital twin die-cutting model means inputting the generated path execution sequence one by one into the virtual die-cutting environment, and driving the virtual cutter to move along the corresponding path in sequence according to the predetermined order, so that the difference in path sequence can be truly reflected in the virtual cutting process.

[0126] In this embodiment, the virtual cutting tool is the execution subject used in the digital twin die-cutting model to simulate the cutting behavior of real die-cutting equipment. Its motion trajectory strictly follows the input die-cutting path sequence and is used to trigger material state updates and structural change judgments, thereby realizing dynamic simulation of the die-cutting process.

[0127] In this embodiment, the virtual die-cutting state sequence refers to a series of continuous state records formed in the virtual environment as the die-cutting path is executed one by one. Each state corresponds to the completion status of the packaging box material after the current cutting path has been completed, the distribution of the remaining connection area, and the overall support structure shape, which is used to reflect the evolution trajectory of the die-cutting process.

[0128] In this embodiment, the material cutting state refers to the distribution of the cut and uncut areas of the packaging box material during the virtual die-cutting process. This state is continuously updated as the die-cutting path progresses and is an important basis for judging whether subsequent path execution may cause structural instability.

[0129] In this embodiment, the local connection relationship refers to the relationship between different material regions that remain connected or have been severed during the virtual die-cutting process. The changes in this relationship directly reflect the rhythm of the disruption of material continuity under different path sequences, which is of key significance for evaluating the rationality of the path sequence.

[0130] In this embodiment, the distribution of the remaining support structure refers to the spatial distribution of material areas that still support the overall structure of the packaging box in any virtual die-cutting state. This distribution is used to determine whether there are potential risks such as local suspension or premature loss of support during the cutting process of the packaging box.

[0131] In this embodiment, the virtual die-cutting process data refers to the set of path execution timing and corresponding material constraint change information recorded during the virtual die-cutting process for each second high curvature die-cutting path pre-proportion. This data is used to provide a data basis for comparing and optimizing the effects of different path sequence schemes in the future.

[0132] It should be noted that this embodiment introduces multiple sets of pre-proportion ratios for the second high curvature die-cutting path and generates corresponding virtual die-cutting process data, so that the path order optimization no longer relies on a single experience judgment, but has a repeatable and comparable virtual verification basis, thereby providing more evidence-based support for subsequent path order selection.

[0133] It should be noted that the digital twin virtual die-cutting process is only an intermediate step in the evaluation and screening of path order. Its purpose is to identify the influence trend of different path orders on the material state evolution in advance, and it does not replace the actual die-cutting process itself, thereby ensuring that this embodiment has good feasibility and compatibility in engineering applications.

[0134] S5, extract features from the virtual die-cutting process, train the ratio scoring model based on reinforcement learning, and generate the ratio evaluation value corresponding to the preceding ratio of each second high curvature die-cutting path.

[0135] In this embodiment, the feature extraction of the virtual die-cutting process, and the generation of proportional evaluation values ​​corresponding to the preceding proportions of each second high curvature die-cutting path based on a reinforcement learning-trained proportional scoring model, specifically involves:

[0136] Based on the virtual die-cutting process data, the distribution of the remaining connection regions of the material at each execution time along the die-cutting path is extracted sequentially. The number of support regions that still maintain continuous connection during the high curvature die-cutting path execution stage is counted to form a structural retention feature group used to characterize the insufficient overall structural constraint caused by the high pre-proportion ratio.

[0137] Based on the material morphology change sequence during the execution stage of the straight die-cutting path in the virtual die-cutting process, the relative displacement distribution characteristics of the decorative edge contour after straight die-cutting are extracted to form a contour traction feature group for characterizing the subsequent straight die-cutting traction deformation caused by the high proportion of the preceding step.

[0138] Based on the virtual material state sequence before the high curvature die-cutting path is executed, the spatial cumulative distribution characteristics of the unreleased cutting stress in the material are extracted to form a stress accumulation feature group to characterize the residual stress accumulation caused by the low pre-cutting ratio.

[0139] Based on the virtual tool motion and material springback response sequence during the high curvature die-cutting path execution process, the springback superposition change characteristics of the edge position after curve die-cutting are extracted to form an edge springback feature group to characterize the curve springback superposition caused by the low pre-proportion.

[0140] The structure retention feature group, contour traction feature group, stress accumulation feature group and edge springback feature group are combined in a fixed order to form a die-cutting process feature vector that corresponds one-to-one with the preceding proportion of each second high curvature die-cutting path.

[0141] Using the pre-proportion of the second high-curvature die-cutting path as the action input in the reinforcement learning environment, the feature vector of the die-cutting process is used as the environmental state representation, and the spatial offset distribution of the actual cutting path of the decorative edge after virtual die-cutting relative to the design path, the number of continuous support areas that still maintain uncut connections in the material at the die-cutting completion stage and their distribution position in the path sequence, and the springback displacement change process of the edge position after the high-curvature die-cutting path ends as the path advances are used as feedback sources to construct a proportional scoring reinforcement learning training environment.

[0142] The reinforcement learning model is trained based on a multi-round virtual die-cutting interaction process, so that the comprehensive influence relationship of various defect features under different pre-ratios is learned, and a ratio scoring model is obtained for outputting ratio evaluation values.

[0143] The feature vector of the die-cutting process corresponding to the pre-cutting ratio of each second high curvature die-cutting path is input into the ratio scoring model, and the corresponding ratio evaluation value is output for subsequent optimal pre-cutting ratio selection.

[0144] In this embodiment, the virtual die-cutting process refers to the complete path execution simulation process completed in the digital twin die-cutting model for different pre-proportions of the second high curvature die-cutting path. Its essence is to make the influence of different path sequence strategies on the material state evolution explicit in the form of time series, providing a unified data source for subsequent defect feature extraction and proportion evaluation.

[0145] In this embodiment, feature extraction refers to extracting distinguishable characterization quantities from the multidimensional state change information generated during the virtual die-cutting process, focusing on the impact of the die-cutting sequence on structural stability, contour accuracy, and stress release effect, to characterize the differences in the strength of potential defect risks under different pre-cutting ratios.

[0146] In this embodiment, the structural preservation feature group refers to a set of structural description information formed after statistical analysis and organization of the support areas that still maintain a continuous connection state during the high curvature die-cutting path execution stage. Its inherent meaning is to reflect whether the overall support structure is prematurely destroyed when the high curvature path is executed too early, thus causing the subsequent die-cutting process to lack sufficient structural constraints.

[0147] In this embodiment, the distribution of the remaining connection area of ​​the material refers to the distribution state of the material area that has not been cut off and still plays a supporting or constraining role in space during the virtual die-cutting process. This distribution directly affects the overall stability of the packaging box during the die-cutting process and is an important basis for evaluating whether the path front ratio is too high.

[0148] In this embodiment, the number of supporting regions refers to the number of regions that remain continuously connected and support the structure at a specific path execution time. This number is not used in isolation, but is combined with its changing trend as the path progresses to determine whether the structural constraints are weakened prematurely.

[0149] In this embodiment, the contour traction feature group refers to a set of features formed during the straight die-cutting path execution stage. By analyzing the spatial offset of the decorative edge contour after cutting, it is used to characterize whether subsequent straight die-cutting will be pulled and deformed due to too many preceding high curvature paths. It reflects the influence of path sequence on morphological stability.

[0150] In this embodiment, the straight-line die-cutting path execution stage refers to the stage in which the die-cutting path is cut in a straight line or near-straight line form. This stage is more sensitive to the overall stiffness of the material and the edge shape. Therefore, its deformation performance can be used to reverse evaluate the rationality of the preceding high-curvature path arrangement.

[0151] In this embodiment, the relative displacement distribution of the decorative edge contour refers to the spatial offset of the actual contour formed by the decorative edge relative to the design contour after the virtual die-cutting is completed. This distribution is used to measure whether the path sequence introduces unacceptable morphological errors.

[0152] In this embodiment, the stress accumulation feature set refers to the feature set formed by extracting the spatial accumulation state of the internal stress in the material that has not yet been released by cutting before the high curvature die-cutting path is executed. Its inherent meaning is to reveal whether the low proportion of the high curvature path leads to stress retention for a long time and concentration in a local area.

[0153] In this embodiment, the unreleased cutting stress refers to the stress state that still exists inside the material before it is cut, which is introduced by forming, folding or previous cutting. If this stress is not released in time, it may cause sudden deformation or abnormal springback in subsequent die cutting.

[0154] In this embodiment, the stress spatial cumulative distribution refers to the distribution state of unreleased stress in a material plane or local area, and its changing trend reflects the influence of the path sequence on the release rhythm of internal forces in the material.

[0155] In this embodiment, the edge springback feature group refers to a set of features formed by analyzing the correspondence between the virtual tool movement and the material springback response during the execution of the high curvature die-cutting path, and extracting the springback superposition of the edge position after curve die-cutting. It is used to characterize the springback superposition effect caused by the continuous execution of multiple high curvature paths when the pre-proportion is low.

[0156] In this embodiment, the material springback response refers to the instantaneous or continuous deformation behavior of the material after cutting due to the release of internal stress. This response is more pronounced when high curvature die-cutting paths are concentrated.

[0157] In this embodiment, the feature vector of the die-cutting process refers to the unified feature expression formed by combining the structure retention feature group, the contour traction feature group, the stress accumulation feature group and the edge springback feature group in a preset order. Its purpose is to represent the virtual die-cutting results under different pre-scale ratios with fixed dimensions and consistent semantics.

[0158] In this embodiment, the ratio scoring model refers to the evaluation model obtained through reinforcement learning training. Its input is the feature vector of the die-cutting process, and its output is the comprehensive evaluation result of the corresponding preceding ratio, which is used to reflect the overall performance of the ratio in terms of structural stability, morphological accuracy and stress control.

[0159] In this embodiment, the reinforcement learning training environment refers to an interactive training framework that uses the pre-process ratio of the second high curvature die-cutting path as the action input, the feature vector of the die-cutting process as the environmental state representation, and the information reflecting the degree of defects in the virtual die-cutting result as the feedback source. This framework is used to guide the model to learn the comprehensive influence relationship of different pre-process ratios on defect features.

[0160] In this embodiment, the ratio evaluation value refers to the comprehensive evaluation result given by the ratio scoring model for a certain second high curvature die-cutting path front ratio. This result is used for the subsequent screening and determination of the optimal front ratio.

[0161] It should be noted that this embodiment does not rely solely on a single defect index to judge the path pre-position ratio. Instead, it uses joint modeling of multiple types of defect features to enable the ratio evaluation results to simultaneously reflect the effects of insufficient structural constraints, morphological tensile deformation, and abnormal stress and springback, thereby improving the robustness of path sequence optimization.

[0162] It should be noted that the training method of the reinforcement learning model and the basic construction method of state-action-feedback are existing technologies. The innovation of this embodiment does not lie in the reinforcement learning algorithm itself, but in introducing the process control quantity of the pre-cutting path ratio into the reinforcement learning environment, and realizing the automatic evaluation of the ratio quality through virtual die-cutting feature construction. Therefore, the general principles of reinforcement learning will not be elaborated further.

[0163] S6. Construct a continuous response curve by combining the preceding proportions of each second high curvature die-cutting path with the corresponding proportion evaluation values. By analyzing the stable range of the response change rate, select the optimal preceding proportion of the high curvature die-cutting path and generate the die-cutting path sequence for die-cutting.

[0164] In this embodiment, the intelligent die-cutting path control optimization method for heterogeneous packaging boxes is characterized by the following steps: First, a continuous response curve is constructed by combining the preceding proportions of each second high-curvature die-cutting path with the corresponding proportion evaluation values. Then, by analyzing the stable interval of the response change rate, the optimal high-curvature die-cutting path preceding proportion is selected, and a die-cutting path sequence is generated for die-cutting. Specifically:

[0165] Sort the preceding proportions of each second high curvature die-cutting path according to their numerical values, and map the corresponding proportion evaluation values ​​one by one according to the sorting order to form a proportion-evaluation value correspondence sequence.

[0166] Based on the ratio-evaluation value correspondence sequence, a continuous response curve of the preceding ratio as the evaluation value changes is constructed using a piecewise continuous interpolation method;

[0167] Along the preceding proportional direction of the continuous response curve, extract the difference sequence of the evaluation value change at adjacent proportional positions to form a rate of change sequence reflecting the trend of the evaluation response change;

[0168] In the rate of change sequence, the proportional segments where the direction of change is consistent and the magnitude of change is convergent at multiple consecutive proportional positions are identified as the stable intervals for evaluating the response change.

[0169] Within the stable interval, the preceding proportion where the corresponding evaluation value converges at an extreme value within the interval is selected and determined as the optimal preceding proportion for the high curvature die-cutting path.

[0170] Based on the optimal high curvature die-cutting path front ratio, a corresponding high curvature die-cutting path front shift set is regenerated, and the die-cutting path order is organized accordingly for actual die-cutting.

[0171] In this embodiment, the second high curvature die-cutting path advance ratio refers to multiple discrete ratio values ​​generated around the initial advance ratio in step S2. Each ratio corresponds to a specific high curvature die-cutting path advance strategy and a complete die-cutting path sequence. Essentially, it is a different control scheme for the degree of advancement of the high curvature path in the overall die-cutting sequence.

[0172] In this embodiment, the ratio evaluation value refers to the comprehensive evaluation result given by the ratio scoring model in S5 for the virtual die-cutting result corresponding to each second high curvature die-cutting path front ratio. This evaluation value inherently reflects the overall quality level of the front ratio in terms of structural preservation, contour stability, stress release and springback control.

[0173] In this embodiment, the scale-evaluation value correspondence sequence refers to the ordered sequence formed by sorting the preceding scales of the second high curvature die-cutting path according to their numerical values ​​and pairing them one by one with their corresponding scale evaluation values. This sequence is used to eliminate the influence of the scale generation order on subsequent analysis and to keep the evaluation relationship monotonically ordered on the scale axis.

[0174] In this embodiment, the continuous response curve refers to the expansion of the originally discrete proportional evaluation relationship into a continuously changing response relationship on the proportional axis by using piecewise continuous interpolation based on the proportional-evaluation value correspondence sequence. Its inherent purpose is to reveal the overall trend of the impact of proportional changes on the evaluation results, rather than relying on single-point results to make judgments.

[0175] In this embodiment, the segmented continuous interpolation method refers to the method of constructing a local continuous transition relationship between adjacent preceding proportional values ​​based on the trend of their evaluation value changes. This method ensures that the curve maintains the continuity of response within each proportional segment, while avoiding the introduction of drastic fluctuations that do not conform to the trend of the original evaluation data.

[0176] In this embodiment, the rate of change sequence refers to the descriptive sequence formed by extracting the trend of evaluation value change at adjacent ratio positions segment by segment along the preceding ratio direction of the continuous response curve. This sequence is used to reflect the sensitivity and direction of change of the evaluation result with the ratio.

[0177] In this embodiment, the direction of change remains consistent, meaning that in multiple consecutive proportional positions, the evaluation value exhibits the same trend of change as the preceding proportional increases or decreases. This consistency is used to determine whether the response curve is in a monotonic change phase within that proportional range.

[0178] In this embodiment, the change range is in a convergent state, which means that the change range of the evaluation value gradually decreases in the continuous proportional position, indicating that the impact of the adjustment of the preceding proportion on the evaluation result tends to be stable. This state is used to distinguish between sensitive sections and stable sections.

[0179] In this embodiment, the stable interval refers to the proportion segment in the rate of change sequence that simultaneously satisfies the requirement that the direction of change remains consistent and the magnitude of change gradually converges. The proportion change within this interval will not cause drastic fluctuations in the evaluation results, representing a region where a relative balance is achieved between the degree of pre-positioning of the high curvature die-cutting path and the die-cutting quality.

[0180] In this embodiment, the extreme convergence position refers to the position where the ratio evaluation value reaches a local optimum within the stable interval and is no longer sensitive to small changes in the ratio. This position reflects the preliminary ratio point where the overall defect risk is the lowest or the overall quality performance is the best within the stable interval.

[0181] In this embodiment, the optimal high curvature die-cutting path front ratio refers to the front ratio corresponding to the extreme convergence position selected from the stable interval. This ratio ensures the stability of die-cutting quality while taking into account the overall optimal balance of structural constraints, morphological accuracy and stress release.

[0182] In this embodiment, the high curvature die-cutting path forward shift set refers to the path subset that is reselected from the high curvature die-cutting path sorting sequence and shifted forward as a whole based on the optimal high curvature die-cutting path forward ratio. This set directly determines the front-end structure of the final die-cutting path order.

[0183] In this embodiment, the actual die-cutting path sequence refers to the complete die-cutting path execution sequence formed after reorganization based on the optimal high curvature die-cutting path pre-proportion. This sequence serves as the final output result and is used to drive the actual die-cutting equipment to perform the die-cutting operation.

[0184] It should be noted that this embodiment avoids making optimal judgments based solely on a single ratio evaluation value by analyzing continuous response curves and stable intervals, thereby reducing the impact of virtual evaluation noise or local anomalies on the final path sequence decision and improving the stability and repeatability of the die-cutting path optimization results in actual production.

[0185] It should be noted that this step does not change the path rearrangement rules and digital twin simulation methods already determined in S3 and S4. Instead, it introduces a continuous analysis mechanism of the scale-response relationship based on the results, which is used to select the final execution scheme that is most robust to the die-cutting quality from multiple candidate pre-scales.

[0186] Example 2, Figure 2 This invention presents an intelligent die-cutting path control and optimization system for heterogeneous packaging boxes, comprising a frequency domain distribution identification module, a second high-curvature die-cutting path pre-proportion module, a die-cutting path sequence module, a virtual die-cutting module, a proportion evaluation module, and a die-cutting module. The frequency domain distribution identification module identifies high-curvature die-cutting paths based on the frequency domain distribution of curvature density, and uses the length ratio of high-curvature die-cutting paths in all die-cutting paths as the initial high-curvature die-cutting path pre-proportion. The second high-curvature die-cutting path pre-proportion module generates several second high-curvature die-cutting path pre-proportions within a continuously varying range, centered on the initial high-curvature die-cutting path pre-proportion. The die-cutting path sequence module sorts the high-curvature die-cutting path set according to the curvature density integral value, based on... The high-curvature die-cutting path is selected by prior to the high-curvature die-cutting path and moved forward as a whole, while keeping the relative order of the remaining paths unchanged, thus generating the die-cutting path order. The virtual die-cutting module is used to generate the die-cutting path order from the prior proportions of each second high-curvature die-cutting path and perform virtual die-cutting based on digital twins. The proportion evaluation module is used to extract features from the virtual die-cutting process, train the proportion scoring model based on reinforcement learning, and generate the proportion evaluation value corresponding to the prior proportion of each second high-curvature die-cutting path. The die-cutting module is used to construct a continuous response curve with the prior proportions of each second high-curvature die-cutting path and the corresponding proportion evaluation value. By analyzing the stable range of the response change rate, the optimal high-curvature die-cutting path prior proportion is selected and the die-cutting path order is generated for die-cutting.

[0187] The present invention also includes an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to execute an intelligent die-cutting path control optimization method for heterogeneous packaging boxes.

[0188] The present invention also includes a computer-readable storage medium storing a computer program that, when executed by a processor, implements an intelligent die-cutting path control optimization method for heterogeneous packaging boxes.

[0189] In the embodiments provided by this invention, it should be understood that the disclosed system can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0190] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0191] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the invention. No appended diagram markings in the claims should be construed as limiting the scope of the claims.

[0192] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0193] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0194] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

[0195] In the embodiments provided in this disclosure, it should be understood that the disclosed systems and methods can also be implemented in other ways. The system embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0196] It should be noted that, in this disclosure, 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 limited by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0197] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended 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, and should all be included within the protection scope of the present invention.

Claims

1. A method for intelligent die-cutting path control and optimization of heterogeneous packaging boxes, characterized in that, Includes the following steps: High curvature die-cutting paths are identified based on the frequency domain distribution of curvature density, and the length ratio of high curvature die-cutting paths in all die-cutting paths is used as the initial high curvature die-cutting path front ratio. Using the initial high-curvature die-cutting path front ratio as the center, several second high-curvature die-cutting path front ratios are generated within a continuously varying range. The set of high curvature die-cutting paths is sorted according to the curvature density integral value. Based on the front ratio of high curvature die-cutting paths, the corresponding proportion of high curvature die-cutting paths is selected and moved forward as a whole, while keeping the relative order of the remaining paths unchanged, thus generating the die-cutting path order. The die-cutting path order is generated by the preceding proportion of each second high curvature die-cutting path, and virtual die-cutting is performed based on digital twin; Feature extraction is performed on the virtual die-cutting process, and a proportional evaluation value corresponding to the preceding proportionality of each second high curvature die-cutting path is generated based on reinforcement learning training proportional scoring model. A continuous response curve is constructed by comparing the preceding proportion of each second high curvature die-cutting path with the corresponding proportion evaluation value. By analyzing the stable range of the response change rate, the optimal preceding proportion of the high curvature die-cutting path is selected and the die-cutting path sequence is generated for die-cutting.

2. The intelligent die-cutting path control optimization method for heterogeneous packaging boxes according to claim 1, characterized in that, The frequency domain distribution based on curvature density identifies high-curvature die-cutting paths, and the proportion of the length of high-curvature die-cutting paths in all die-cutting paths is used as the initial high-curvature die-cutting path prefix ratio, specifically: Discretely sample the die-cutting paths of heterogeneous packaging boxes and represent each die-cutting path as a coordinate sequence arranged in path order; Based on the coordinate sequence, the curvature change of adjacent sampling points is calculated, and a curvature density sequence arranged sequentially along the die-cutting path is constructed. The curvature density sequence is subjected to frequency domain transformation to obtain the spectral distribution of curvature density in the path order dimension; Based on the spectrum distribution, the path intervals corresponding to the high-frequency components are extracted, and the path intervals are mapped back to the original die-cutting paths and marked as a set of high-curvature die-cutting paths. The cumulative path length of the high curvature die-cutting path set is calculated and normalized with the total length of all die-cutting paths to obtain the initial high curvature die-cutting path front ratio.

3. The intelligent die-cutting path control optimization method for heterogeneous packaging boxes according to claim 2, characterized in that, The process of generating several second high-curvature die-cutting path front ratios within a continuously varying range, centered on the initial high-curvature die-cutting path front ratio, specifically involves: The cumulative length of the high-curvature die-cutting path corresponding to the initial high-curvature die-cutting path front ratio is used as the reference forward shift length; Based on the curvature density integral value of each path in the high curvature die-cutting path set, the high curvature die-cutting paths are sorted in descending order to form a path sorting sequence from high to low curvature. Using the cumulative path length change points of adjacent paths in the path sorting sequence as the forward adjustment boundary, the boundary of the high curvature die-cut path subset that can participate in the forward adjustment is determined. Within the path length range defined by the boundary of the high curvature die-cutting path subset, the reference forward shift length is gradually increased or decreased using the actual path length of a single high curvature die-cutting path as the adjustment unit, forming several candidate forward shift lengths; Each candidate forward shift length is normalized to the total path length of all die-cutting paths to obtain the corresponding second high curvature die-cutting path forward shift ratio; Apply adjacent difference constraints to the preceding proportions of the second high curvature die-cutting path, so that the sets of forward paths corresponding to adjacent proportions differ by at least one complete high curvature die-cutting path.

4. The intelligent die-cutting path control optimization method for heterogeneous packaging boxes according to claim 3, characterized in that, The process involves sorting the high-curvature die-cutting path set according to its curvature density integral value, selecting a corresponding proportion of high-curvature die-cutting paths based on their preceding proportion, and shifting them forward as a whole, while maintaining the relative order of the remaining paths. This generates the die-cutting path order as follows: For each second high curvature die-cutting path front ratio, based on its corresponding forward shift length, in the high curvature die-cutting path sorting sequence arranged in descending order of curvature density integral value, the path length is accumulated sequentially from the beginning of the sequence until the accumulated path length reaches the forward shift length, thereby determining the high curvature die-cutting path subset participating in the forward shift. The high curvature die-cutting path subset is extracted as a whole from the original die-cutting path sequence and rearranged into preceding path segments according to their relative order in the sorting sequence. The preceding path segment is inserted into the starting position of the die-cutting path sequence to form the adjusted preceding die-cutting path sequence; For the remaining die-cutting paths that are not selected into the high curvature die-cutting path subset, their relative order in the original die-cutting path sequence remains unchanged, and they are sequentially spliced ​​to the preceding die-cutting path sequence. The completed splicing path sequence is determined as the die-cutting path order corresponding to the preceding ratio of the second high curvature die-cutting path, and is used for subsequent virtual die-cutting processing.

5. The intelligent die-cutting path control optimization method for heterogeneous packaging boxes according to claim 4, characterized in that, The step of generating the die-cutting path sequence by prioritizing the proportions of each second high curvature die-cutting path and performing virtual die-cutting based on digital twins specifically involves: For each of the preceding proportions of the second high curvature die-cutting path, read the corresponding generated die-cutting path sequence and use the die-cutting path sequence as the execution input for virtual die-cutting; Based on the geometric structure data, material hierarchy information, and motion constraints of the die-cutting equipment of the heterogeneous packaging boxes, a digital twin die-cutting model consistent with the actual die-cutting environment is constructed. The die-cutting paths are sequentially mapped into the digital twin die-cutting model, and the virtual cutter is driven to move along the corresponding die-cutting paths in sequence according to the path order. During the virtual die-cutting process, the cutting status, local connection relationships, and distribution of remaining support structures of the packaging box material are updated synchronously, forming a virtual die-cutting state sequence that changes as the path progresses; Record the execution timing of each die-cutting path in the virtual die-cutting state sequence and the corresponding material constraint change process to generate virtual die-cutting process data that corresponds one-to-one with the preceding ratio of the second high curvature die-cutting path.

6. The intelligent die-cutting path control optimization method for heterogeneous packaging boxes according to claim 5, characterized in that, The process of extracting features from the virtual die-cutting process and generating proportional evaluation values ​​corresponding to the preceding proportions of each second high curvature die-cutting path based on a reinforcement learning-trained proportional scoring model is as follows: Based on the virtual die-cutting process data, the distribution of the remaining connection regions of the material at each execution time along the die-cutting path is extracted sequentially. The number of support regions that still maintain continuous connection during the high curvature die-cutting path execution stage is counted to form a structural retention feature group used to characterize the insufficient overall structural constraint caused by the high pre-proportion ratio. Based on the material morphology change sequence during the execution stage of the straight die-cutting path in the virtual die-cutting process, the relative displacement distribution characteristics of the decorative edge contour after straight die-cutting are extracted to form a contour traction feature group for characterizing the subsequent straight die-cutting traction deformation caused by the high proportion of the preceding step. Based on the virtual material state sequence before the high curvature die-cutting path is executed, the spatial cumulative distribution characteristics of the unreleased cutting stress in the material are extracted to form a stress accumulation feature group to characterize the residual stress accumulation caused by the low pre-cutting ratio. Based on the virtual tool motion and material springback response sequence during the high curvature die-cutting path execution process, the springback superposition change characteristics of the edge position after curve die-cutting are extracted to form an edge springback feature group to characterize the curve springback superposition caused by the low pre-proportion. The structure retention feature group, contour traction feature group, stress accumulation feature group and edge springback feature group are combined in a fixed order to form a die-cutting process feature vector that corresponds one-to-one with the preceding proportion of each second high curvature die-cutting path. Using the pre-proportion of the second high-curvature die-cutting path as the action input in the reinforcement learning environment, the feature vector of the die-cutting process is used as the environmental state representation, and the spatial offset distribution of the actual cutting path of the decorative edge after virtual die-cutting relative to the design path, the number of continuous support areas that still maintain uncut connections in the material at the die-cutting completion stage and their distribution position in the path sequence, and the springback displacement change process of the edge position after the high-curvature die-cutting path ends as the path advances are used as feedback sources to construct a proportional scoring reinforcement learning training environment. The reinforcement learning model is trained based on a multi-round virtual die-cutting interaction process, so that the comprehensive influence relationship of various defect features under different pre-ratios is learned, and a ratio scoring model is obtained for outputting ratio evaluation values. The feature vector of the die-cutting process corresponding to the pre-cutting ratio of each second high curvature die-cutting path is input into the ratio scoring model, and the corresponding ratio evaluation value is output for subsequent optimal pre-cutting ratio selection.

7. The intelligent die-cutting path control optimization method for heterogeneous packaging boxes according to claim 6, characterized in that, The process involves constructing a continuous response curve by comparing the preceding proportions of each second high-curvature die-cutting path with their corresponding evaluation values. By analyzing the stable range of the response change rate, the optimal preceding proportions of the high-curvature die-cutting path are selected, and a die-cutting path sequence is generated for die-cutting. Specifically: Sort the preceding proportions of each second high curvature die-cutting path according to their numerical values, and map the corresponding proportion evaluation values ​​one by one according to the sorting order to form a proportion-evaluation value correspondence sequence. Based on the ratio-evaluation value correspondence sequence, a continuous response curve of the preceding ratio as the evaluation value changes is constructed using a piecewise continuous interpolation method; Along the preceding proportional direction of the continuous response curve, extract the difference sequence of the evaluation value change at adjacent proportional positions to form a rate of change sequence reflecting the trend of the evaluation response change; In the rate of change sequence, the proportional segments where the direction of change is consistent and the magnitude of change is convergent at multiple consecutive proportional positions are identified as the stable intervals for evaluating the response change. Within the stable interval, the preceding proportion where the corresponding evaluation value converges at an extreme value within the interval is selected and determined as the optimal preceding proportion for the high curvature die-cutting path. Based on the optimal high curvature die-cutting path front ratio, a corresponding high curvature die-cutting path front shift set is regenerated, and the die-cutting path order is organized accordingly for actual die-cutting.

8. A system using the intelligent die-cutting path control optimization method for heterogeneous packaging boxes as described in any one of claims 1-7, characterized in that, It includes a frequency domain distribution identification module, a second high curvature die-cutting path pre-proportioning module, a die-cutting path sequence module, a virtual die-cutting module, a proportioning evaluation module, and a die-cutting module; The frequency domain distribution recognition module is used to identify high curvature die-cutting paths based on the frequency domain distribution of curvature density, and uses the length ratio of high curvature die-cutting paths in all die-cutting paths as the initial high curvature die-cutting path front ratio. The second high curvature die-cutting path front ratio module is used to generate several second high curvature die-cutting path front ratios within a continuously varying range, centered on the initial high curvature die-cutting path front ratio. The die-cutting path sequence module is used to sort the set of high curvature die-cutting paths according to the curvature density integral value, select the corresponding proportion of high curvature die-cutting paths according to the preceding proportion of high curvature die-cutting paths and move them forward as a whole, while keeping the relative order of the remaining paths unchanged, and generate the die-cutting path sequence. The virtual die-cutting module is used to generate the die-cutting path sequence by the preceding proportion of each second high curvature die-cutting path, and to perform virtual die-cutting based on digital twin; The proportion evaluation module is used to extract features from the virtual die-cutting process, train the proportion scoring model based on reinforcement learning, and generate the proportion evaluation value corresponding to the preceding proportion of each second high curvature die-cutting path. The die-cutting module is used to construct a continuous response curve by comparing the preceding proportions of each second high curvature die-cutting path with the corresponding proportion evaluation values. By analyzing the stable range of the response change rate, the optimal high curvature die-cutting path preceding proportions are selected and the die-cutting path sequence is generated for die-cutting.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform the intelligent die-cutting path control optimization method for heterogeneous packaging boxes as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the intelligent die-cutting path control optimization method for heterogeneous packaging boxes as described in any one of claims 1 to 7.

Images