A laser cutting device for intelligently identifying irregular-shaped luggage and a control method thereof

Abstract

The application relates to the technical field of luggage special-shaped parts, in particular to a laser cutting device for intelligently identifying luggage special-shaped parts and a control method thereof. The laser cutting device comprises a cutting planning model, the cutting planning model comprising a reference cutting model and a record sub-library; image data and basic feature parameters of a to-be-processed special-shaped part are acquired, and a three-dimensional contour model of the to-be-processed special-shaped part is constructed according to the image data; the cutting planning model sets working parameters of a laser cutting unit according to the basic feature parameters and the three-dimensional contour model, the working parameters comprising execution laser parameters and an expected cutting path; cutting record data of the to-be-processed special-shaped part is generated, and the record sub-library is updated according to the cutting record data. Through the parallel double architecture of the reference cutting model and the record sub-library, the efficiency of identifying new special-shaped parts is ensured, and the risk of extreme parameter misjudgment caused by model overfitting is eliminated. Therefore, the precision, efficiency and flexible manufacturing capacity of the luggage special-shaped part laser cutting are improved.

Description

Technical Field

[0001] This application relates to the field of irregularly shaped bag parts technology, and in particular to a laser cutting device and its control method for intelligently identifying irregularly shaped bag parts. Background Technology

[0002] In the luggage manufacturing industry, laser cutting of irregularly shaped parts (such as curved splicing pieces, irregular reinforcing linings, and decorative contour parts) is a key process affecting product quality and production efficiency. Compared with traditional punching or mechanical milling, laser cutting has advantages such as non-contact operation, small heat-affected zone, and narrow kerf. However, irregularly shaped luggage parts often have complex three-dimensional curved surfaces, varied edge curvatures, and a wide variety of materials (such as microfiber, polyurethane, and composite fabrics), which poses a serious challenge to automatic cutting control.

[0003] Currently, most mainstream laser cutting systems use offline programming or manual teaching methods to set cutting parameters. Operators generate planar cutting paths based on two-dimensional unfolded diagrams and then set laser power, pulse frequency, auxiliary gas pressure, and feed speed based on experience. However, in actual production, planar path planning cannot accurately reflect the actual focal length changes and heat accumulation effects on the three-dimensional contour surface. Especially in areas with abrupt curvature changes such as corners and narrow ribs, fixed power and speed can easily lead to overheating or incomplete cutting, requiring operators to repeatedly test cuts and adjust, significantly lengthening the production preparation cycle. Bag styles are frequently iterated, and the same type of irregularly shaped parts may only have local size or curvature differences, but existing systems cannot recognize such similarities, requiring the process to be reset each time, resulting in a large amount of idle and repetitive work based on historical processing experience. Summary of the Invention

[0004] The purpose of this application is to solve the above-mentioned technical problems by providing an intelligent laser cutting device and control method for identifying irregularly shaped parts of bags, which aims to improve the accuracy, efficiency and flexible manufacturing capability of laser cutting of irregularly shaped parts of bags.

[0005] In some embodiments of this application, a dual architecture of a benchmark cutting model and a record sub-library is constructed in parallel. When a new type of irregularly shaped part appears, the cutting process can be dynamically adapted through the benchmark cutting model. When an already processed irregularly shaped part appears, relevant matching strategies can be called from the record sub-library. This ensures both the efficiency of identifying new irregularly shaped parts and eliminates the risk of misjudgment of extreme parameters due to model overfitting. This improves the accuracy, efficiency, and flexible manufacturing capabilities of laser cutting for irregularly shaped parts in bags and luggage.

[0006] In some embodiments of this application, a control method for a laser cutting device for intelligently identifying irregularly shaped parts of bags is provided, including: Construct a decision planning model, which includes: a baseline decision model and a record sub-library; Obtain image data and basic feature parameters of the irregular part to be processed, and construct a three-dimensional contour model of the irregular part to be processed based on the image data; The cutting planning model sets the working parameters of the laser cutting unit based on the basic feature parameters and the three-dimensional contour model. The working parameters include: laser execution parameters and expected cutting path. Generate cutting record data for the irregularly shaped part to be processed, and update the record sub-library based on the cutting record data.

[0007] In some embodiments of this application, the construction of the trimming planning model includes: Retrieve historical adjudication records; Multiple record sub-components are generated based on the historical adjudication record data, wherein each record sub-component includes: task feature parameters and adjudication control parameters; Aggregate all record sub-components; Multiple record sets are generated based on the aggregation results, where each record set includes multiple record sub-components; A record sub-library is constructed based on the entire record set, and a baseline irregular component is generated for each record set; A basic identification model is established based on all the reference irregular parts, and the cutting sub-models of each reference irregular part are generated in sequence. A baseline judgment model is established, which includes: a basic identification model and all judgment sub-models; A decision planning model is generated based on the baseline decision model and the record library.

[0008] In some embodiments of this application, the generation of cut sub-models for each reference irregular part includes: Select the target reference part sequentially from all the reference irregular parts, and select the mapping record set of the target reference part; The basic trimming strategy for the target reference component is set according to the mapping record set; The basic decision-making strategy includes: expected laser parameters and path planning strategy; Multiple working condition scenarios are constructed based on the mapping record set, and the target working condition scenario is selected sequentially from all working condition scenarios. Set the path compensation strategy for the target reference component in the target working condition scenario; The path compensation strategy for the target reference component is set sequentially for each working scenario; Establish a cutting sub-model for the target reference part; The judgment sub-model includes: a basic judgment strategy and a full path compensation strategy.

[0009] In some embodiments of this application, setting the operating parameters of the laser cutting unit includes: Generate structural feature packages and operating condition feature packages based on the aforementioned basic feature parameters; Input the structural feature package into the basic recognition model; The basic recognition model generates similarity values ​​between the irregular part to be processed and each reference irregular part, and selects the maximum value among all similar values ​​as the first-level similarity value; If the first-level similarity value is greater than the preset similarity threshold, a first-level matching instruction is generated; If the first-level similarity value is less than the preset similarity threshold, a first-level calling instruction is generated.

[0010] In some embodiments of this application, the first-level matching instruction includes: Select the subset to be matched based on the first-level similarity value; The matching values ​​of the irregular part to be processed and each recorded sub-part in the matching subset are generated based on the structural feature package and the working condition feature package. The recording sub-component corresponding to the maximum value among all matching values ​​is selected as the mapping recording sub-component, and the working parameters of the laser cutting unit are set according to the cutting control parameters of the mapping recording sub-component.

[0011] In some embodiments of this application, the first-level invocation instruction includes: The reference irregular part corresponding to the first-level similarity value is set as the mapping reference part; The cut sub-model of the mapping reference component is defined as the mapping cut model; The working condition feature package is input into the mapping and trimming model, and the execution compensation strategy is output. Set the base trimming strategy of the mapping reference component to the execution trimming strategy; The execution laser parameters and expected cutting path of the irregular part to be processed are set according to the execution cutting strategy and the execution compensation strategy.

[0012] In some embodiments of this application, updating the record sub-database based on the cut record data includes: Generate a cutoff deviation value based on the cutoff record data; If the cut deviation value is greater than the preset cut deviation value threshold, an optimization instruction is generated, and a new record sub-component is generated according to the optimization instruction; If the cutting deviation value is less than the preset cutting deviation value threshold, a new record sub-component is generated based on the cutting record data; The newly added record sub-components include: task feature parameters and truncation control parameters; The update instruction is set according to the first-level similarity value; If the first-level similarity value is greater than the preset similarity threshold, a first-level update instruction is generated, and the newly added record sub-component is added to the record sub-library according to the first-level update instruction; If the first-level similarity value is less than the preset similarity threshold, a second-level update instruction is generated, and the newly added record sub-component is added to the record sub-library according to the second-level update instruction.

[0013] In some embodiments of this application, a laser cutting device for intelligently identifying irregularly shaped parts of bags is provided, comprising: A cutting platform, on which the irregularly shaped part to be processed is fixed during the cutting process; The acquisition unit includes a vision submodule and an acquisition submodule; The vision submodule is used to acquire image data of the irregularly shaped parts to be processed. The acquisition submodule is used to acquire the basic feature parameters of the irregularly shaped part to be processed; A laser cutting unit is used to cut the irregularly shaped part to be processed; The central control unit is used to construct the decision planning model, which includes: a baseline decision model and a record sub-library; The central control unit includes: The analysis module is used to construct a three-dimensional contour model of the irregular part to be processed based on the image data; The control module is used to set the working parameters of the laser cutting unit according to the basic feature parameters and the three-dimensional contour model. The working parameters include: laser execution parameters and expected cutting path. The recording module is used to generate cutting record data for the irregularly shaped part to be processed, and to update the recording sub-library based on the cutting record data.

[0014] In some embodiments of this application, the central control unit further includes: The execution module is used to obtain historical judgment record data; Multiple record sub-components are generated based on the historical adjudication record data, wherein each record sub-component includes: task feature parameters and adjudication control parameters; Aggregate all record sub-components; Multiple record sets are generated based on the aggregation results, where each record set includes multiple record sub-components; A record sub-library is constructed based on the entire record set, and a baseline irregular component is generated for each record set; A basic identification model is established based on all the reference irregular parts, and the cutting sub-models of each reference irregular part are generated in sequence. A baseline judgment model is established, which includes: a basic identification model and all judgment sub-models; Generate a trimming planning model based on the baseline trimming model and the record library; The process of generating cut sub-models for each baseline irregular part includes: Select the target reference part sequentially from all the reference irregular parts, and select the mapping record set of the target reference part; The basic trimming strategy for the target reference component is set according to the mapping record set; The basic decision-making strategy includes: expected laser parameters and path planning strategy; Multiple working condition scenarios are constructed based on the mapping record set, and the target working condition scenario is selected sequentially from all working condition scenarios. Set the path compensation strategy for the target reference component in the target working condition scenario; The path compensation strategy for the target reference component is set sequentially for each working scenario; Establish a cutting sub-model for the target reference part; The judgment sub-model includes: a basic judgment strategy and a full path compensation strategy.

[0015] In some embodiments of this application, the control module is further configured to: Generate structural feature packages and operating condition feature packages based on the aforementioned basic feature parameters; Input the structural feature package into the basic recognition model; The basic recognition model generates similarity values ​​between the irregular part to be processed and each reference irregular part, and selects the maximum value among all similar values ​​as the first-level similarity value; If the first-level similarity value is greater than the preset similarity threshold, a first-level matching instruction is generated; If the first-level similarity value is less than a preset similarity threshold, a first-level invocation instruction is generated; The first-level matching instructions include: Select the subset to be matched based on the first-level similarity value; The matching values ​​of the irregular part to be processed and each recorded sub-part in the matching subset are generated based on the structural feature package and the working condition feature package. Select the recording sub-component corresponding to the maximum value among all matching values ​​as the mapping recording sub-component, and set the working parameters of the laser cutting unit according to the cutting control parameters of the mapping recording sub-component; The first-level invocation instruction includes: The reference irregular part corresponding to the first-level similarity value is set as the mapping reference part; The cut sub-model of the mapping reference component is defined as the mapping cut model; The working condition feature package is input into the mapping and trimming model, and the execution compensation strategy is output. Set the base trimming strategy of the mapping reference component to the execution trimming strategy; The execution laser parameters and expected cutting path of the irregular part to be processed are set according to the execution cutting strategy and the execution compensation strategy.

[0016] Compared with the prior art, the laser cutting device and its control method for intelligently identifying irregularly shaped parts of bags and luggage, as described in this application, have the following advantages: By constructing a dual architecture that combines a baseline cutting model and a record sub-library, when new types of irregularly shaped parts appear, the cutting process can be dynamically adapted through the baseline cutting model. When already processed irregularly shaped parts appear, relevant matching strategies can be called from the record sub-library. This ensures both the efficiency of identifying new irregularly shaped parts and eliminates the risk of misjudgment of extreme parameters due to model overfitting. This improves the accuracy, efficiency, and flexible manufacturing capabilities of laser cutting for irregularly shaped bag parts. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the control method of a laser cutting device for intelligently identifying irregularly shaped parts of bags in a preferred embodiment of this application. Detailed Implementation

[0018] The specific embodiments of this application will be described in further detail below with reference to the accompanying drawings and examples. The following examples are used to illustrate this application, but are not intended to limit the scope of this application.

[0019] In the description of this application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0020] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "multiple" means each or more.

[0021] In the description of this application, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal communication of various components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0022] like Figure 1As shown in the preferred embodiment of this application, a control method for a laser cutting device for intelligently identifying irregularly shaped parts of bags includes: S101: Construct a decision planning model, which includes: a baseline decision model and a record sub-library; S102: Obtain image data and basic feature parameters of the irregular part to be processed, and construct a three-dimensional contour model of the irregular part to be processed based on the image data; S103: The cutting planning model sets the working parameters of the laser cutting unit based on the basic feature parameters and the three-dimensional contour model. The working parameters include: the laser execution parameters and the expected cutting path. S104: Generate cutting record data for the irregularly shaped parts to be processed, and update the record sub-library based on the cutting record data.

[0023] Specifically, the control acquisition unit acquires image data (i.e., after the irregularly shaped part to be processed is fixed on the cutting platform, the vision submodule performs global image acquisition), uses a Gaussian filter for filtering and noise reduction to remove noise interference introduced during image acquisition and transmission, and employs a histogram equalization method for contrast enhancement, making the contour edges of the irregularly shaped part to be processed clearer and more distinguishable. The contour recognition algorithm module uses the Canny edge detection operator to perform edge detection on the preprocessed image, generating contour coordinate data of the irregularly shaped part to be processed. For cases where the contour is incomplete or has partial occlusion, the missing parts are predicted and repaired based on the symmetry and continuity features of the contour. A three-dimensional contour model of the irregularly shaped part to be processed is constructed based on its contour coordinate data, providing data support for the subsequent laser cutting unit to perform laser cutting.

[0024] Specifically, the acquisition unit can also acquire the basic feature parameters of the irregular part to be processed. These basic feature parameters include, but are not limited to, the structural parameters of the irregular part that affect the cutting effect, such as material, thickness, and contour structure.

[0025] Specifically, constructing a decision planning model includes: Retrieve historical adjudication records; Multiple record sub-components are generated based on historical adjudication record data. Each record sub-component includes: task feature parameters and adjudication control parameters. Aggregate all record sub-components; Multiple record sets are generated based on the aggregation results, where each record set includes multiple record sub-components; A record sub-library is constructed based on the entire record set, and a baseline irregular component is generated for each record set; A basic identification model is established based on all the reference irregular parts, and the cutting sub-models of each reference irregular part are generated in sequence. Establish a baseline judgment model, which includes: a basic identification model and all judgment sub-models; A decision planning model is generated based on the baseline decision model and the record library.

[0026] Specifically, the historical cutting record data includes the structural parameters (material, thickness, and contour structure, etc.) of the irregularly shaped parts that have already been cut, the corresponding working parameters of the laser cutting unit (i.e., laser power, cutting speed, pulse frequency, auxiliary air pressure, and corresponding cutting path, etc.), and the working condition parameters during the cutting process (temperature curve and aging status of the laser cutting device).

[0027] Specifically, by filtering historical cutting records, multiple record sub-components are generated. Each record sub-component includes the relevant processing procedures for a non-standard part, and each record sub-component specifically represents a different non-standard part. The number of record sub-components is equal to the number of non-standard parts that have been processed.

[0028] Specifically, a storage duration is set for each record sub-component (the storage start time is based on the completion time of the processed irregular-shaped part corresponding to the current record sub-component, and the current record sub-component is removed after the storage duration is completed). The storage duration of each record sub-component is set according to the number of record sub-components in its corresponding record set; the more record sub-components, the shorter the corresponding storage duration. By setting the storage duration, historical record sub-components are periodically cleared from the record sub-database, thereby reducing storage pressure.

[0029] Specifically, the task characteristic parameters in the recording sub-component include the structural parameters (material, thickness, and contour structure, etc.) of the irregular part of the current recording sub-component and the corresponding working condition parameters (temperature profile and aging state of the laser cutting unit). The cutting control parameters are the cutting sub-strategies corresponding to this recording sub-component (the working parameters corresponding to the current recording sub-component; if the irregular part corresponding to the current recording sub-component exhibits a large deviation during the cutting process, the working parameters are optimized, and the optimization result is set as the corresponding cutting sub-strategy). These include control parameters for the laser cutting unit, such as cutting path, laser power, cutting speed, pulse frequency, and auxiliary air pressure.

[0030] Specifically, based on the structural parameters in the task feature parameters corresponding to each recording sub-component, the recording sub-components are aggregated to generate multiple record sets. Among them, the irregular parts corresponding to all recording sub-components in a single record set are irregular parts with the same material, similar contour structure, and similar thickness.

[0031] Specifically, the irregular parts corresponding to all record sub-parts in the current record set are compared with each other, and a reference value for each irregular part is generated in turn. The irregular part with the maximum value among the reference values ​​is selected as the reference irregular part of the current record set.

[0032] Specifically, similarity values ​​are generated between the current irregular part and each irregular part in the record set (i.e., by comparing contour similarity and thickness similarity), and the variance and average of all similar values ​​are generated. A first reference value is set based on the variance; the smaller the variance, the larger the first reference value. A second reference value is generated based on the average value; the larger the average value, the larger the corresponding second reference value. The sum of the first and second reference values ​​is set as the reference value of the current irregular part.

[0033] Specifically, generating cut sub-models for each baseline irregular part includes: Select the target reference part sequentially from all the reference irregular parts, and select the mapping record set of the target reference part; The basic trimming strategy for the target reference component is set based on the mapping record set; The basic decision-making strategies include: expected laser parameters and path planning strategies; Multiple working condition scenarios are constructed based on the mapping record set, and the target working condition scenario is selected sequentially from all working condition scenarios. Set the path compensation strategy for the target reference component in the target working condition scenario; The path compensation strategy for the target reference component is set sequentially for each working scenario; Establish a cutting sub-model for the target reference part; The judgment sub-model includes: basic judgment strategy and full path compensation strategy.

[0034] Specifically, the historical adjudication record data of the record set corresponding to the target reference component is selected as the mapping record set of the target reference component.

[0035] Specifically, based on the mapping record set, the truncation control parameters in the record sub-component corresponding to the target reference component are set to the expected laser parameters (laser power, cutting speed, pulse frequency, auxiliary air pressure) and path planning strategy (i.e., how to identify the waste area and the safe distance to be ensured during cutting. The cutting start point and cutting end point are set in the waste area, and the corresponding safe distance is ensured with the effective area of ​​the irregular part when planning the cutting path).

[0036] Specifically, operating condition indicators are generated, including ambient temperature and equipment aging status. By quantifying each operating condition indicator, the reference values ​​of each operating condition indicator are made to be within the same range, and multiple value intervals for each operating condition indicator are generated. Multiple operating condition scenarios are generated based on the random combination of all value intervals.

[0037] Specifically, by analyzing the mapping record set, a path compensation strategy for the working conditions is generated (i.e., the increase in safe distance when in the target working condition). By adding the path compensation strategy, the problem of the laser cutting device cutting into the effective area of ​​irregularly shaped parts due to cutting deviations caused by external working condition interference is avoided. This improves the cutting efficiency of irregularly shaped parts.

[0038] In a preferred embodiment of this application, the operating parameters of the laser cutting unit are set, including: Generate structural feature packages and operating condition feature packages based on basic feature parameters; Input the structural feature package into the basic recognition model; The basic identification model generates similarity values ​​between the irregular part to be processed and each reference irregular part, and selects the maximum value among all similar values ​​as the first-level similarity value; If the first-level similarity value is greater than the preset similarity threshold, a first-level matching instruction is generated. If the first-level similarity value is less than the preset similarity threshold, a first-level calling instruction is generated.

[0039] Specifically, the structural feature package contains the structural parameters (material, thickness, and contour structure) of the irregular part to be processed, while the working condition feature package contains the current ambient temperature and the actual aging state of the laser cutting unit.

[0040] Specifically, by comparing the degree of difference between the structural feature package and the structural parameters corresponding to each benchmark irregular part, the smaller the degree of difference, the greater the corresponding similarity value.

[0041] Specifically, the similarity threshold can be set based on historical parameters. When the first-level similarity value is greater than the preset similarity threshold, it indicates that the part to be processed is not a newly added part, and relevant cutting record data is stored in the record sub-library. At this time, a first-level matching instruction is generated, and the optimal working parameters are matched by filtering and calling from the record sub-library. If the first-level similarity value is less than the preset similarity threshold, it indicates that the part to be processed is a newly added part, and there is no relevant cutting record data in the record sub-library. At this time, a first-level calling instruction is generated, and the expected optimal working parameters are quickly planned through real-time optimization, thereby improving the cutting and processing efficiency of the part to be processed.

[0042] Specifically, the first-level matching instructions include: Select the subset to be matched based on the first-level similarity value; Generate matching values ​​for each recorded sub-part in the non-standard part to be processed and the matching subset based on the structural feature package and the working condition feature package; Select the recording sub-component corresponding to the maximum value among all matching values ​​as the mapping recording sub-component, and set the working parameters of the laser cutting unit according to the cutting control parameters of the mapping recording sub-component.

[0043] Specifically, the subset of records corresponding to the first-level similarity value is set as the subset to be matched.

[0044] Specifically, a matching value is set by comparing the degree of difference between the structural parameters of the structural feature package and each record sub-component in the subset to be matched. The smaller the degree of difference, the larger the corresponding matching value.

[0045] Specifically, the cutting control parameters within the mapping record sub-component are set to the operating parameters of the laser cutting unit.

[0046] Specifically, first-level invocation instructions include: The reference irregular part corresponding to the first-level similarity value is set as the mapping reference part; Set the cut sub-model of the mapping reference component as the mapping cut model; Input the working condition feature package into the mapping and trimming model, and output the compensation strategy to be executed. Set the base trimming strategy of the mapping reference component to the execution trimming strategy; The laser parameters and expected cutting path for the irregularly shaped parts to be processed are set according to the execution cutting strategy and the execution compensation strategy.

[0047] Specifically, by implementing a compensation strategy and a mapping cut-off model, the three-dimensional contour model of the irregular part to be processed is analyzed to generate the optimal cut-off path (i.e., the predicted cut-off path), and the laser parameters are set according to the expected laser parameters of the mapping reference part.

[0048] It is understandable that, in the above embodiments, by constructing a dual architecture of a benchmark cutting model and a record sub-library in parallel, when a new type of irregularly shaped part appears, the cutting process can be dynamically adapted through the benchmark cutting model. When an irregularly shaped part that has already been processed appears, the relevant matching strategy can be called from the record sub-library. This ensures both the efficiency of identifying new irregularly shaped parts and eliminates the risk of misjudgment of extreme parameters that may be caused by model overfitting. This improves the accuracy, efficiency, and flexible manufacturing capabilities of laser cutting of irregularly shaped parts for bags.

[0049] In a preferred embodiment of this application, updating the record sub-database based on the cut record data includes: Generate a cutoff deviation value based on the cutoff record data; If the cut deviation value is greater than the preset cut deviation value threshold, an optimization instruction is generated, and a new record sub-component is generated based on the optimization instruction; If the cutting deviation value is less than the preset cutting deviation value threshold, a new record sub-component is generated based on the cutting record data; The newly added record sub-components include: task characteristic parameters and trimming control parameters; Set update instructions based on first-level similarity values; If the first-level similarity value is greater than the preset similarity threshold, a first-level update instruction is generated, and the newly added record sub-component is added to the record sub-database according to the first-level update instruction; If the first-level similarity value is less than the preset similarity threshold, a second-level update instruction is generated, and the newly added record sub-component is added to the record sub-database according to the second-level update instruction.

[0050] Specifically, a cutting deviation value is generated by analyzing the cutting results (i.e., the degree of difference between the actual cutting area and the expected cutting area, and the degree of difference between the actual state and the expected state of the cutting surface). The larger the cutting deviation value, the worse the cutting effect of the irregular part to be processed.

[0051] Specifically, when a secondary new instruction is generated, the new record component is set as a new record set, and this record set is added to the record sub-library.

[0052] It is understandable that, in the above embodiments, by dynamically updating the record sub-library, the adaptation and cutting efficiency of different irregular parts can be improved, thereby enhancing the accuracy, efficiency, and flexible manufacturing capabilities of laser cutting of irregular parts for bags.

[0053] Another preferred embodiment of the control method for a laser cutting device for intelligently identifying irregularly shaped parts of bags, based on any of the above preferred embodiments, provides a laser cutting device for intelligently identifying irregularly shaped parts of bags, comprising: The cutting platform is where the irregularly shaped parts to be processed are fixed during the cutting process; The acquisition unit includes a vision submodule and an acquisition submodule; The vision submodule is used to acquire image data of the irregularly shaped parts to be processed; The acquisition submodule is used to acquire the basic feature parameters of the irregularly shaped parts to be processed; Laser cutting unit, used to cut irregularly shaped parts to be processed; The central control unit is used to construct the decision planning model, which includes: a baseline decision model and a record sub-library. The central control unit includes: The analysis module is used to construct a three-dimensional contour model of the irregular part to be processed based on the image data; The control module is used to set the working parameters of the laser cutting unit based on the basic feature parameters and the three-dimensional contour model. The working parameters include: the laser execution parameters and the expected cutting path. The recording module is used to generate cutting record data for irregularly shaped parts to be processed, and to update the record sub-library based on the cutting record data.

[0054] Specifically, the laser cutting unit includes a laser, a laser cutting head, and an auxiliary air blowing device. The laser is a carbon dioxide laser, and the auxiliary air blowing device is used to blow away molten material and fumes from the cutting kerf during the cutting process.

[0055] Specifically, the vision submodule is preferably an industrial area array camera, which can acquire global images of the irregularly shaped parts to be processed.

[0056] In a preferred embodiment of this application, the central control unit further includes: The execution module is used to obtain historical judgment record data; Multiple record sub-components are generated based on historical adjudication record data. Each record sub-component includes: task feature parameters and adjudication control parameters. Aggregate all record sub-components; Multiple record sets are generated based on the aggregation results, where each record set includes multiple record sub-components; A record sub-library is constructed based on the entire record set, and a baseline irregular component is generated for each record set; A basic identification model is established based on all the reference irregular parts, and the cutting sub-models of each reference irregular part are generated in sequence. Establish a baseline judgment model, which includes: a basic identification model and all judgment sub-models; Generate a trimming planning model based on the baseline trimming model and the record library; The process of generating cut sub-models for each baseline irregular part includes: Select the target reference part sequentially from all the reference irregular parts, and select the mapping record set of the target reference part; The basic trimming strategy for the target reference component is set based on the mapping record set; The basic decision-making strategies include: expected laser parameters and path planning strategies; Multiple working condition scenarios are constructed based on the mapping record set, and the target working condition scenario is selected sequentially from all the working condition scenarios; Set the path compensation strategy for the target reference component in the target working condition scenario; The path compensation strategy for the target reference component is set sequentially for each working scenario; Establish a cutting sub-model for the target reference part; The judgment sub-model includes: basic judgment strategy and full path compensation strategy.

[0057] In a preferred embodiment of this application, the control module is further configured to: Generate structural feature packages and operating condition feature packages based on basic feature parameters; Input the structural feature package into the basic recognition model; The basic identification model generates similarity values ​​between the irregular part to be processed and each reference irregular part, and selects the maximum value among all similar values ​​as the first-level similarity value; If the first-level similarity value is greater than the preset similarity threshold, a first-level matching instruction is generated. If the first-level similarity value is less than the preset similarity threshold, a first-level invocation instruction is generated. The first-level matching instructions include: Select the subset to be matched based on the first-level similarity value; Generate matching values ​​for each recorded sub-part in the non-standard part to be processed and the matching subset based on the structural feature package and the working condition feature package; Select the recording sub-component corresponding to the maximum value among all matching values ​​as the mapping recording sub-component, and set the working parameters of the laser cutting unit according to the cutting control parameters of the mapping recording sub-component; Level 1 invocation instructions include: The reference irregular part corresponding to the first-level similarity value is set as the mapping reference part; Set the cut sub-model of the mapping reference component as the mapping cut model; Input the working condition feature package into the mapping and trimming model, and output the compensation strategy to be executed. Set the base trimming strategy of the mapping reference component to the execution trimming strategy; The laser parameters and expected cutting path for the irregularly shaped parts to be processed are set according to the execution cutting strategy and the execution compensation strategy.

[0058] Based on the first concept of this application, by constructing a dual architecture of a benchmark cutting model and a record sub-library in parallel, when a new type of irregularly shaped part appears, the cutting process can be dynamically adapted through the benchmark cutting model. When an irregularly shaped part that has already been processed appears, relevant matching strategies can be called from the record sub-library. This ensures both the efficiency of identifying new irregularly shaped parts and eliminates the risk of misjudgment of extreme parameters that may be caused by model overfitting. This improves the accuracy, efficiency, and flexible manufacturing capabilities of laser cutting for irregularly shaped parts in bags and luggage.

[0059] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and substitutions can be made without departing from the technical principles of this application, and these improvements and substitutions should also be considered within the scope of protection of this application.

Claims

1. A control method for a laser cutting device for intelligently identifying irregularly shaped parts of bags, characterized in that, include: Construct a decision planning model, which includes: a baseline decision model and a record sub-library; Obtain image data and basic feature parameters of the irregular part to be processed, and construct a three-dimensional contour model of the irregular part to be processed based on the image data; The cutting planning model sets the working parameters of the laser cutting unit based on the basic feature parameters and the three-dimensional contour model. The working parameters include: laser execution parameters and expected cutting path. Generate cutting record data for the irregularly shaped part to be processed, and update the record sub-library based on the cutting record data.

2. The control method of the laser cutting device for intelligent identification of irregularly shaped parts of bags as described in claim 1, characterized in that, The construction of the cutoff planning model includes: Retrieve historical adjudication records; Multiple record sub-components are generated based on the historical adjudication record data, wherein each record sub-component includes: task feature parameters and adjudication control parameters; Aggregate all record sub-components; Multiple record sets are generated based on the aggregation results, where each record set includes multiple record sub-components; A record sub-library is constructed based on the entire record set, and a baseline irregular component is generated for each record set; A basic identification model is established based on all the reference irregular parts, and the cutting sub-models of each reference irregular part are generated in sequence. A baseline judgment model is established, which includes: a basic identification model and all judgment sub-models; A decision planning model is generated based on the baseline decision model and the record library.

3. The control method of the laser cutting device for intelligent identification of irregularly shaped parts of bags as described in claim 2, characterized in that, Generate cut sub-models for each baseline irregular part, including: Select the target reference part sequentially from all the reference irregular parts, and select the mapping record set of the target reference part; The basic trimming strategy for the target reference component is set according to the mapping record set; The basic decision-making strategy includes: expected laser parameters and path planning strategy; Multiple working condition scenarios are constructed based on the mapping record set, and the target working condition scenario is selected sequentially from all working condition scenarios. Set the path compensation strategy for the target reference component in the target working condition scenario; The path compensation strategy for the target reference component is set sequentially for each working scenario; Establish a cutting sub-model for the target reference part; The judgment sub-model includes: a basic judgment strategy and a full path compensation strategy.

4. The control method of the laser cutting device for intelligent identification of irregularly shaped parts of bags as described in claim 3, characterized in that, The operating parameters of the laser cutting unit include: Generate structural feature packages and operating condition feature packages based on the aforementioned basic feature parameters; Input the structural feature package into the basic recognition model; The basic recognition model generates similarity values ​​between the irregular part to be processed and each reference irregular part, and selects the maximum value among all similar values ​​as the first-level similarity value; If the first-level similarity value is greater than the preset similarity threshold, a first-level matching instruction is generated; If the first-level similarity value is less than the preset similarity threshold, a first-level calling instruction is generated.

5. The control method of the laser cutting device for intelligent identification of irregularly shaped parts of bags as described in claim 4, characterized in that, The first-level matching instructions include: Select the subset to be matched based on the first-level similarity value; The matching values ​​of the irregular part to be processed and each recorded sub-part in the matching subset are generated based on the structural feature package and the working condition feature package. The recording sub-component corresponding to the maximum value among all matching values ​​is selected as the mapping recording sub-component, and the working parameters of the laser cutting unit are set according to the cutting control parameters of the mapping recording sub-component.

6. The control method of the laser cutting device for intelligent identification of irregularly shaped parts of bags as described in claim 4, characterized in that, The first-level invocation instruction includes: The reference irregular part corresponding to the first-level similarity value is set as the mapping reference part; The cut sub-model of the mapping reference component is defined as the mapping cut model; The working condition feature package is input into the mapping and trimming model, and the execution compensation strategy is output. Set the base trimming strategy of the mapping reference component to the execution trimming strategy; The execution laser parameters and expected cutting path of the irregular part to be processed are set according to the execution cutting strategy and the execution compensation strategy.

7. The control method of the laser cutting device for intelligent identification of irregularly shaped parts of bags as described in claim 4, characterized in that, Updating the record sub-database based on the cut record data includes: Generate a cutoff deviation value based on the cutoff record data; If the cut deviation value is greater than the preset cut deviation value threshold, an optimization instruction is generated, and a new record sub-component is generated according to the optimization instruction; If the cutting deviation value is less than the preset cutting deviation value threshold, a new record sub-component is generated based on the cutting record data; The newly added record sub-components include: task feature parameters and truncation control parameters; The update instruction is set according to the first-level similarity value; If the first-level similarity value is greater than the preset similarity threshold, a first-level update instruction is generated, and the newly added record sub-component is added to the record sub-library according to the first-level update instruction; If the first-level similarity value is less than the preset similarity threshold, a second-level update instruction is generated, and the newly added record sub-component is added to the record sub-library according to the second-level update instruction.

8. A laser cutting device for intelligently identifying irregularly shaped parts of bags, employing the control method of any one of claims 1-7, characterized in that, include: A cutting platform, on which the irregularly shaped part to be processed is fixed during the cutting process; The acquisition unit includes a vision submodule and an acquisition submodule; The vision submodule is used to acquire image data of the irregularly shaped parts to be processed. The acquisition submodule is used to acquire the basic feature parameters of the irregularly shaped part to be processed; A laser cutting unit is used to cut the irregularly shaped part to be processed; The central control unit is used to construct the decision planning model, which includes: a baseline decision model and a record sub-library; The central control unit includes: The analysis module is used to construct a three-dimensional contour model of the irregular part to be processed based on the image data; The control module is used to set the working parameters of the laser cutting unit according to the basic feature parameters and the three-dimensional contour model. The working parameters include: laser execution parameters and expected cutting path. The recording module is used to generate cutting record data for the irregularly shaped part to be processed, and to update the recording sub-library based on the cutting record data.

9. The laser cutting device for intelligent identification of irregularly shaped parts of bags as described in claim 8, characterized in that, The central control unit also includes: The execution module is used to obtain historical judgment record data; Multiple record sub-components are generated based on the historical adjudication record data, wherein each record sub-component includes: task feature parameters and adjudication control parameters; Aggregate all record sub-components; Multiple record sets are generated based on the aggregation results, where each record set includes multiple record sub-components; A record sub-library is constructed based on the entire record set, and a baseline irregular component is generated for each record set; A basic identification model is established based on all the reference irregular parts, and the cutting sub-models of each reference irregular part are generated in sequence. A baseline judgment model is established, which includes: a basic identification model and all judgment sub-models; Generate a trimming planning model based on the baseline trimming model and the record library; The process of generating cut sub-models for each baseline irregular part includes: Select the target reference part sequentially from all the reference irregular parts, and select the mapping record set of the target reference part; The basic trimming strategy for the target reference component is set according to the mapping record set; The basic decision-making strategy includes: expected laser parameters and path planning strategy; Multiple working condition scenarios are constructed based on the mapping record set, and the target working condition scenario is selected sequentially from all the working condition scenarios. Set the path compensation strategy for the target reference component in the target working condition scenario; The path compensation strategy for the target reference component is set sequentially for each working scenario; Establish a cutting sub-model for the target reference part; The judgment sub-model includes: a basic judgment strategy and a full path compensation strategy.

10. The laser cutting device for intelligent identification of irregularly shaped parts of bags as described in claim 9, characterized in that, The control module is also used for: Generate structural feature packages and operating condition feature packages based on the aforementioned basic feature parameters; Input the structural feature package into the basic recognition model; The basic recognition model generates similarity values ​​between the irregular part to be processed and each reference irregular part, and selects the maximum value among all similar values ​​as the first-level similarity value; If the first-level similarity value is greater than the preset similarity threshold, a first-level matching instruction is generated; If the first-level similarity value is less than a preset similarity threshold, a first-level invocation instruction is generated; The first-level matching instructions include: Select the subset to be matched based on the first-level similarity value; The matching values ​​of the irregular part to be processed and each recorded sub-part in the matching subset are generated based on the structural feature package and the working condition feature package. Select the recording sub-component corresponding to the maximum value among all matching values ​​as the mapping recording sub-component, and set the working parameters of the laser cutting unit according to the cutting control parameters of the mapping recording sub-component; The first-level invocation instruction includes: The reference irregular part corresponding to the first-level similarity value is set as the mapping reference part; The cut sub-model of the mapping reference component is defined as the mapping cut model; The working condition feature package is input into the mapping and trimming model, and the execution compensation strategy is output. Set the base trimming strategy of the mapping reference component to the execution trimming strategy; The execution laser parameters and expected cutting path of the irregular part to be processed are set according to the execution cutting strategy and the execution compensation strategy.

Images