Self-adaptive laser cutting method and system based on characteristic parameters of sheet metal part

Abstract

The invention relates to the technical field of self-adaptive control laser cutting, in particular to a self-adaptive laser cutting method and system based on characteristic parameters of sheet metal parts. Firstly, by establishing a space coordinate system and obtaining a top view image of a cutting platform, contour deviation is calculated, a correction initial position is generated, and then a cutting path is optimized. And the laser cutting head moves to the corrected initial position according to the path and performs a cutting task. In addition, the method further comprises the steps that the to-be-processed sheet metal part image is segmented into a plurality of sub-images, and a sub-image cross list is established to describe the connection relation of the sub-images. And extracting a subimage feature vector through a feature learning model, and adjusting a control parameter by using cutting simulation information to compensate an error. And for each sub-image, based on the cutting result and the connection relationship, adaptive adjustment of control parameters is carried out, and the precision of the subsequent cutting step is ensured. And finally, a complete cutting task is realized by updating the cutting mark and merging the sub-images.

Description

Technical Field

[0001] The present invention relates to the technical field of adaptively controlled laser cutting, and in particular to an adaptive laser cutting method and system based on characteristic parameters of sheet metal parts. Background Art

[0002] As an important manufacturing workpiece, sheet metal parts are widely used in various industries, especially in the fields of automobiles, aviation, home appliances and construction. Sheet metal parts generally refer to parts made of thin sheet metal (such as steel, aluminum, stainless steel and copper) through a series of production processes (such as stamping, bending, cutting, welding, stretching, etc.).

[0003] In the production process of sheet metal parts, cutting is a very critical step, which is used to cut large pieces of metal sheets into the required shape and size according to design requirements. The cutting process not only affects the precision and appearance of the product, but is also closely related to production efficiency, material utilization, etc. In modern sheet metal processing, laser cutting, as an advanced cutting technology, has gradually become the mainstream cutting method.

[0004] Laser cutting is a high-precision, non-contact material cutting process that focuses a high-intensity laser beam on the surface of a sheet metal part, causing the local area to quickly heat up and melt or evaporate, thereby achieving cutting. During the laser cutting process, multiple laser parameters need to be strictly controlled to ensure cutting accuracy and quality.

[0005] At present, the existing laser cutting methods mainly rely on set fixed parameters, which are difficult to adapt to sheet metal parts of different thicknesses, materials and surface conditions. Based on the above, how to efficiently and accurately cut sheet metal parts of various specifications is still a difficult problem in production.

[0006] To this end, the present invention proposes an adaptive laser cutting method and system based on characteristic parameters of sheet metal parts. Summary of the invention

[0007] The purpose of the present invention is to provide an adaptive laser cutting method and system based on the characteristic parameters of sheet metal parts. The method and system are used for adaptively adjusting preset control parameters of sheet metal parts during laser cutting, specifically including: first, using the laser cutting head to accurately locate and correct the starting position, and generating the moving path of the laser cutting head according to the contour of the sheet metal part and the finished product features. By processing the images of the cutting platform and the sheet metal part to be processed, the contour deviation is further optimized and the correction starting position is calculated, and the contour deviation vector is extracted at the same time. Then, by dividing the sheet metal part image and establishing a cross-linked list, the features of each sub-image are obtained in turn and the cutting simulation is performed, and the error between the simulated cutting information and the expected cutting information is calculated, so as to dynamically adjust the control parameters.

[0008] In addition, the method automatically adjusts the control parameters in the subsequent processing steps by providing real-time feedback on the cutting status of each sub-image to be processed, so as to ensure the accuracy and efficiency of the cutting process. In particular, when there is a connection relationship between multiple sub-images, the system can expand according to the information of the cut sub-image, merge the current sub-image to be processed and the cut sub-image, and further perform standardization and influence analysis. Finally, the system optimizes the adjustment process of the control parameters by combining the sub-image features with the cut influence vector, so that the entire cutting process is more in line with the predetermined processing requirements.

[0009] To achieve the above object, the present invention provides the following technical solutions: An adaptive laser cutting method based on characteristic parameters of a sheet metal part, comprising: Using a control panel to move the laser cutting head to a corrected starting position; wherein moving the laser cutting head to the corrected starting position comprises: Establishing a spatial coordinate system with the current position of the laser cutting head as the origin; Further, obtaining a current top view image of the cutting platform; Further, denoising and contrast enhancement are performed on the current overhead view image to obtain a first overhead view image; Further, extracting the contour of the cutting platform and the contour of the sheet metal part to be processed in the first top view image to obtain a cutting platform contour image and a sheet metal part contour image to be processed; Further, calculating the contour deviation between the cutting platform contour image and the contour image of the sheet metal part to be processed, and constructing a contour deviation vector; Further, the expected cutting information of the sheet metal part to be processed is obtained, and characterized to obtain an expected feature vector; Further, generating a moving path of the laser cutting head and the corrected starting position according to the contour deviation vector and the expected feature vector; The generation process of the moving path and the corrected starting position includes: The correction starting position on the contour image of the sheet metal part to be processed is calculated according to the contour deviation vector, and the calculation formula is: ;in, is the correction starting position; is the preset starting position; Calculate the function for correcting the starting position; To obtain the centroid coordinate point through the contour image of the sheet metal part to be processed; is the expected feature vector; is a set of boundary points of the contour image of the sheet metal part to be processed; is the correction parameter of the correction starting position; Further, obtaining a set of candidate movable paths from the origin to the corrected starting position; Furthermore, the candidate movable path set is selected according to the contour deviation vector to obtain the moving path.

[0010] Furthermore, the laser cutting head moves to the corrected starting position according to the moving path.

[0011] Further, obtaining an image of the sheet metal part to be processed; Further, the image of the sheet metal part to be processed is divided according to the cutting content to obtain N sub-images; Furthermore, a sub-image cross-link list is established for the sub-image according to the sheet metal part image to be processed; wherein the sub-image cross-link list includes: processing mark, sub-image sequence number and sub-image connection relationship; Further, obtaining a starting sub-image to be processed; Further, the initial sub-image to be processed is preprocessed to obtain a standard first sub-image; Further, the standard first sub-image is input into a sheet metal feature learning model to obtain a sub-image feature vector; Further, obtaining pixel coordinates corresponding to the sub-image feature vector; Further, a sub-image feature-position set is established according to the pixel coordinates and the sub-image feature vector; Further, using preset control parameters, a cutting simulation is performed on the sub-graph feature-position set to obtain simulated cutting information; Further, calculating the error between the simulated cutting information and the expected cutting information to obtain a cutting error vector; Further, the compensation control parameter is calculated using a control parameter compensation function; wherein the control parameter compensation function is expressed as: ;in, is the control parameter compensation function; is the pixel coordinate Compensation control parameters at ; is the pixel coordinate The cutting error vector at ; Simulates the function for cutting; is the pixel coordinate Preset control parameters at is the pixel coordinate The sub-image feature vector at ; is the pixel coordinate location information.

[0012] Further, the laser cutting head is controlled to perform cutting according to the compensation control parameter, and the cutting mark of the starting sub-image to be processed is updated to 1; Furthermore, the preset control parameters in the subsequent processing steps are adaptively adjusted according to the processing result of the initial sub-image to be processed, and cutting is performed using the laser cutting head; The process of adaptively adjusting the preset control parameters in the subsequent processing steps includes: Further, according to the sub-image sequence number, the corresponding next sub-image to be processed is obtained and recorded as the current sub-image to be processed; Further, a sub-image having the sub-image connection relationship with the current sub-image to be processed and the cutting mark being 1 is obtained, and recorded as a cut sub-image; wherein the sub-image connection relationship includes: upper connection, lower connection, left connection and right connection; Further, the current sub-image to be processed and the cut sub-image are merged to obtain a current extended sub-image to be processed; Further, the preprocessing is performed on the current extended sub-image to be processed to obtain a standard current extended sub-image to be processed; Further, an influence analysis is performed on the cut sub-image in the standard current extended sub-image to be processed to obtain a cut influence vector; Further, the current sub-image to be processed in the standard current extended sub-image to be processed is input into the sheet metal feature learning model to obtain a feature vector of the current sub-image; Further, adjusting the current preset control parameters according to the cut influence vector and the current sub-image feature vector; When all the cutting marks in the sub-image cross list are 1, the cutting task is completed.

[0013] An adaptive laser cutting system based on characteristic parameters of sheet metal parts, comprising: Intelligent control unit, which controls the laser cutting head through the control panel; A monitoring unit, used to monitor the sheet metal cutting process; A collection unit, used to collect images during the monitoring process of the monitoring unit; a preprocessing unit, used to preprocess the collected images; The starting movement calculation unit is used to calculate the moving path of the laser cutting head and correct the starting position; wherein the starting movement calculation unit includes: Establish a spatial coordinate system with the current position of the laser cutting head as the origin; Further, obtaining a current top view image of the cutting platform; Further, denoising and contrast enhancement are performed on the current overhead view image to obtain a first overhead view image; Further, extracting the contour of the cutting platform and the contour of the sheet metal part to be processed in the first top view image to obtain a cutting platform contour image and a sheet metal part contour image to be processed; Further, calculating the contour deviation between the cutting platform contour image and the contour image of the sheet metal part to be processed, and constructing a contour deviation vector; Further, the expected cutting information of the sheet metal to be processed is obtained and characterized to obtain an expected feature vector; Further, generating a moving path of the laser cutting head and the corrected starting position according to the contour deviation vector and the expected feature vector; The correction starting position on the contour image of the sheet metal part to be processed is calculated according to the contour deviation vector, and the calculation formula is: ;in, is the correction starting position; is the preset starting position; Calculate the function for correcting the starting position; To obtain the centroid coordinate point through the contour image of the sheet metal part to be processed; is the expected feature vector; is a set of boundary points of the contour image of the sheet metal part to be processed; is the correction parameter of the correction starting position; Further, obtaining a set of candidate movable paths from the origin to the corrected starting position; Furthermore, the candidate movable path set is selected according to the contour deviation vector to obtain the moving path.

[0014] A control parameter optimization unit is used to optimize the control parameters of sheet metal cutting; wherein the control parameter optimization unit includes: Acquire the image of the sheet metal part to be processed; Further, the image of the sheet metal part to be processed is divided according to the cutting content to obtain N sub-images; Furthermore, a sub-image cross-link list is established for the sub-image according to the sheet metal part image to be processed; wherein the sub-image cross-link list includes: processing mark, sub-image sequence number and sub-image connection relationship; Further, obtaining a starting sub-image to be processed; Further, the initial sub-image to be processed is preprocessed to obtain a standard first sub-image; Further, the standard first sub-image is input into a sheet metal feature learning model to obtain a sub-image feature vector; Further, obtaining pixel coordinates corresponding to the sub-image feature vector; Further, a sub-image feature-position set is established according to the pixel coordinates and the sub-image feature vector; Further, using preset control parameters, a cutting simulation is performed on the sub-graph feature-position set to obtain simulated cutting information; Further, calculating the error between the simulated cutting information and the expected cutting information to obtain a cutting error vector; Further, the compensation control parameter is calculated using a control parameter compensation function; wherein the control parameter compensation function is expressed as: ;in, is the control parameter compensation function; is the pixel coordinate Compensation control parameters at ; is the pixel coordinate The cutting error vector at ; Simulates the function for cutting; is the pixel coordinate Preset control parameters at is the pixel coordinate The sub-image feature vector at ; is the pixel coordinate location information; Further, the laser cutting head is controlled to perform cutting according to the compensation control parameter, and the cutting mark of the starting sub-image to be processed is updated to 1; Further, the preset control parameters in the subsequent processing steps are adaptively adjusted according to the processing result of the initial sub-image to be processed, and the corresponding next sub-image to be processed is obtained according to the sub-image sequence number and recorded as the current sub-image to be processed; Further, a sub-image having the sub-image connection relationship with the current sub-image to be processed and the cutting mark being 1 is obtained, and recorded as a cut sub-image; wherein the sub-image connection relationship includes: upper connection, lower connection, left connection and right connection; Further, the current sub-image to be processed and the cut sub-image are merged to obtain a current extended sub-image to be processed; Further, the preprocessing is performed on the current extended sub-image to be processed to obtain a standard current extended sub-image to be processed; Further, an influence analysis is performed on the cut sub-image in the standard current extended sub-image to be processed to obtain a cut influence vector; Further, the current sub-image to be processed in the standard current extended sub-image to be processed is input into the sheet metal feature learning model to obtain a feature vector of the current sub-image; Further, adjusting the current preset control parameters according to the cut influence vector and the current sub-image feature vector; The display unit is used to display the cutting status and control parameters of the sheet metal parts.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention proposes a method for calculating the corrected starting position of a sheet metal part; this method can effectively optimize the moving path of the laser cutting head, reduce errors in the cutting process, and improve cutting accuracy. By establishing a spatial coordinate system, obtaining and processing the top view of the cutting platform, extracting the contours of the cutting platform and the sheet metal part to be processed, and calculating the contour deviation vector, the moving path of the cutting head and the corrected starting position are generated in combination with the expected feature vector. The method of the present invention can not only accurately determine the starting position of laser cutting, but also dynamically adjust the path during the cutting process to adapt to the shape of the sheet metal part and the requirements of the finished product, thereby improving cutting efficiency and accuracy.

[0016] 2. The present invention proposes a method for adaptively adjusting control parameters; this method can accurately calculate and compensate for processing errors through step-by-step processing and cutting simulation of sheet metal images, thereby dynamically optimizing control parameters in the laser cutting process. By dividing and extracting features from the images of sheet metal to be processed, combined with cross-link list management of sub-images, control parameters can be monitored and adjusted in real time during the cutting process to ensure that the cutting accuracy of each sub-image meets expectations. This method can effectively reduce processing errors, thereby significantly improving the cutting quality and processing efficiency of sheet metal parts.

[0017] 3. The present invention proposes a control parameter optimization method for subsequent processing steps; the method tracks the cutting progress of the sub-images and obtains the current sub-image to be processed and the cut sub-image in real time according to the connection relationship and cutting mark status of the sub-images. Combined with the influence analysis and sheet metal feature learning model, the method can accurately calculate the feature vector of the current sub-image and adjust the preset control parameters according to the cut influence vector to ensure the cutting quality and efficiency of the subsequent processing steps. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic diagram of a scene for sheet metal cutting provided by an embodiment of the present invention; Figure 2 A flowchart of an adaptive laser cutting method based on characteristic parameters of a sheet metal part provided by an embodiment of the present invention; Figure 3 A structural diagram of an adaptive laser cutting system based on sheet metal feature parameters provided by an embodiment of the present invention; Figure 4 A schematic diagram of the image division of a sheet metal part to be processed provided by an embodiment of the present invention; Figure 5 A schematic diagram of sub-image merging provided by an embodiment of the present invention.

[0019] In the figure: 1. Control panel; 2. Cutting platform; 3. Laser cutting head; 4. Monitoring equipment; 5. Sheet metal parts to be processed. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] As an important manufacturing workpiece, sheet metal parts are widely used in various industries, especially in the fields of automobiles, aviation, home appliances and construction. Sheet metal parts generally refer to parts made of thin sheet metal (such as steel, aluminum, stainless steel and copper) through a series of production processes (such as stamping, bending, cutting, welding and stretching).

[0022] Among the many production processes of sheet metal parts, cutting is one of the most common and important processes. This process is mainly used to cut metal sheets according to the shape and size required by the design. Common cutting methods include laser cutting, plasma cutting, flame cutting and water jet cutting; among them, laser cutting is a high-precision non-contact material cutting process, which mainly focuses on the surface of sheet metal parts through a high-intensity laser beam, so that the local area quickly heats up and melts or evaporates, thereby realizing the cutting process. In the laser cutting process, it is indeed necessary to strictly control multiple laser parameters to ensure the accuracy and quality of cutting.

[0023] In the existing laser cutting method, it mainly relies on set fixed parameters, which is difficult to adapt to sheet metal parts of different sizes, thicknesses, materials and surface conditions; especially in mass production, how to efficiently and accurately cut sheet metal parts of various specifications is still a difficult problem in production. To this end, the present invention proposes an adaptive laser cutting method and system based on the characteristic parameters of sheet metal parts. It will be described in detail from the following two embodiments.

[0024] Embodiment 1: In the embodiment of the present application, by combining the method and system of the present invention, the adjustment of the adaptive control parameters in the laser cutting process of the sheet metal to be processed is realized; wherein the sheet metal 5 to be processed is cut by a laser cutting device; in the example of the present application, the laser cutting device includes: a control panel 1, a cutting platform 2, a laser cutting head 3 and a monitoring device 4, and the specific structure is shown in FIG. Figure 1 .exist Figure 1 In the embodiment, a sheet metal part 5 is placed on a cutting platform 2, and a laser cutting head 3 is controlled by a control panel 1 to cut the sheet metal part 5. A monitoring device 4 is installed above the cutting platform 2 to monitor the cutting process in real time.

[0025] for Figure 1 The implementation process, through Figure 2 and Figure 3 content realization; Figure 2 The method flow of the present invention includes: S10. placing the sheet metal to be processed on the cutting platform; S20. using the control panel to move the laser cutting head to the corrected starting position; S30. obtaining the image of the sheet metal to be processed; S40. dividing the image of the sheet metal to be processed according to the cutting content to obtain N sub-images; S50. establishing a sub-image cross-link list; S60. obtaining the starting sub-image to be processed; S70. calculating the starting compensation control parameters according to the starting sub-image to be processed; S80. adaptively adjusting the preset control parameters in the subsequent processing steps according to the processing result of the starting sub-image to be processed, and performing cutting. Figure 3 This is a system structure diagram of the present invention, which includes: an intelligent control unit, a monitoring unit, a collection unit, a preprocessing unit, an initial movement calculation unit, a control parameter optimization unit and a display unit.

[0026] According to the above content, the cutting process of the sheet metal part 5 to be processed is described, which specifically includes the following contents; According to the above step S10, the sheet metal part 5 to be processed is placed on the cutting platform 2; Further, the corrected starting position and moving path are obtained by the starting movement calculation unit of the system, and the intelligent control unit of the system is used to control the laser cutting head 3 to move to the corrected starting position, and the above step S10 is used; wherein the control panel 1 integrates all the functions realized by the intelligent control unit of the system; The specific movement process of the laser cutting head 3 includes: Taking the current position of the laser cutting head 3 as the origin, a spatial coordinate system is established; Further, the acquisition unit of the system is used to obtain the current top view image of the cutting platform 2 taken by the monitoring device 4; Further, the current overhead image is denoised and the contrast is enhanced to obtain a first overhead image; Further, the contour images of the cutting platform 2 and the sheet metal part to be processed 5 in the first top view image are extracted to obtain the contour image of the cutting platform and the contour image of the sheet metal part to be processed; Further, the contour deviation between the contour image of the cutting platform and the contour image of the sheet metal to be processed is calculated, and a contour deviation vector is constructed; wherein the contour deviation vector includes: scale deviation, relative position deviation and angle deviation; The calculation formula of scale deviation is: ; If the sizes are consistent, the scale deviation is 0; The calculation formula for relative position deviation is: ;in, and are the centroids of the sheet metal to be processed and the cutting platform respectively; if the centroids of the two coincide, the position deviation is 0; The calculation formula for the angle deviation is: ;in, and are the main direction vectors of the contours of the cutting platform and the sheet metal part to be processed respectively; Further, the expected cutting information of the sheet metal part 5 to be processed is obtained; wherein the expected cutting information includes: cutting shape, cutting path and cutting shape size; Further, the expected cutting information is characterized to obtain an expected feature vector; Furthermore, the corrected starting position of the sheet metal part 5 to be processed is calculated; wherein the calculation formula of the corrected starting position is: ; in, To correct the starting position; is the preset starting position; Calculate the function for correcting the starting position; To obtain the centroid coordinate point through the contour image of the sheet metal part to be processed; is the expected feature vector; is the boundary point set of the contour image of the sheet metal part to be processed; Correction parameters for correcting the starting position; For the correction starting position calculation function It can be defined as: ; in, is the number of boundary point sets; For the boundary point The distance from the centroid coordinate point; For the boundary point , ; Expected Features , ; Expected features The weight of is the expected number of features; by Figure 1 Take the sheet metal part 5 to be processed as an example; wherein, , that is, there are 3 points in the boundary point set; the distance between each point in the boundary point set and the centroid coordinate point is calculated as: , and ; Among them, the geometric influence part is: ; Calculate the impact part of the expected cutting information; where the expected feature vector is: ; The expected feature weights are: , and ;therefore, ; Set the correction parameter to 0.05; the preset starting position is known is (101.5,125); According to the above calculation results, the correction starting position is: ; Further, a candidate movable path set from the origin to the correction start is obtained; wherein the candidate movable path set includes multiple movable paths, and the available space The algorithm generates the corresponding path; Furthermore, a candidate movable path set is selected according to the contour deviation vector to obtain a movable path; wherein the movable path generation step includes: Step 1: Evaluate the adaptability of each candidate path based on the calculated contour deviation vector; Step 2: According to the adaptability score of each candidate path, the path with the smallest contour deviation and the most consistent with the expected requirements is selected as the final moving path; Step 3: Smooth the selected optimal path to reduce unnecessary sharp turns and complex bends in the path, thereby improving the stability and execution efficiency of the path; Step 4: The optimized and smoothed path is used as the final moving path and is ready for execution by the laser cutting head 3.

[0027] In the embodiment of the present application, the starting position of the sheet metal to be processed is corrected and the moving path of the laser cutting head to the starting position is obtained to improve the cutting accuracy and reduce the error. Specifically, by analyzing the contour deviation vector of the sheet metal to be processed, the starting position can be corrected to ensure that the laser cutting head cuts at the correct starting position, thereby minimizing the cutting error caused by the starting point deviation.

[0028] Furthermore, the control panel 1 controls the laser cutting head 3 to move to the corrected starting position according to the moving path.

[0029] In an embodiment of the present application, through precise image processing and path optimization technology, the laser cutting head can accurately move to the corrected starting position. First, a spatial coordinate system is established based on the current position of the laser cutting head, and a real-time top view image of the cutting platform is obtained. After denoising and contrast enhancement, the contour images of the cutting platform and the sheet metal are extracted, the contour deviation is calculated, and the deviation vector is constructed. By characterizing the expected cutting information and combining the deviation vector, the correction path of the laser cutting head is generated, and finally the cutting head is accurately guided to move to the corrected starting position. This process effectively solves the cutting deviation caused by position error or deformation, and ensures the high precision and high stability of the laser cutting process.

[0030] Further, according to the content of step S30, the image of the sheet metal part 5 to be processed is acquired by using the acquisition unit; Furthermore, the control parameter optimization unit is used to optimize the control parameters of the laser cutting head 3 during the cutting process of the sheet metal part 5 to be processed; the specific process includes: The image of the sheet metal part 5 to be processed is divided into a plurality of sub-images according to the cutting content, corresponding to the content of step S40; wherein the image of the sheet metal part 5 to be processed is divided into 9 sub-images in total; see Table 1; Further, according to the content of step S50, a sub-image cross linked list is established for the sub-image; wherein the sub-image cross linked list includes: processing mark, sub-image sequence number and sub-image connection relationship; wherein the processing mark includes: 0 and 1; "0" indicates uncut; "1" indicates cut; the sub-image sequence number corresponds to the cutting step; the sub-image connection relationship includes: upper connection, lower connection, left connection and right connection between sub-images; refer to the content of Table 1, which gives the initial information in the sub-image cross linked list; "NULL" in Table 1 indicates empty; Table 1 Initial information in the sub-image cross list

[0031] Further, according to the content of step S60, a starting sub-image to be processed is obtained; wherein the starting sub-image to be processed is a sub-image with a sub-image sequence number of 1; Further, the initial sub-image to be processed is pre-processed by a pre-processing unit to obtain a standard first sub-image; wherein the pre-processing of the pre-processing unit includes: image denoising, image enhancement and geometric correction; Furthermore, the standard first sub-image is input into a sheet metal feature learning model to obtain a sub-image feature vector; wherein the sheet metal feature learning model includes: A cutting state recognition layer, used to recognize the cutting state of the input image; The sub-item feature learning layer is used to learn the sub-item features of the input image according to the cutting recognition results, including: a cut feature learning module and an uncut feature learning module; wherein the cut feature learning module uses a convolutional neural network to learn the local features of the cutting area, such as the smoothness, precision, burrs, burns and coherence of the cutting edge; The uncut feature learning module uses a deep learning model to obtain information such as geometric shape, boundary features, size, and texture in the uncut image; According to the recognition result of the cutting state recognition layer, it can be known that there is a cut part in the standard first sub-image; therefore, only the uncut feature learning module is used for feature learning; Furthermore, the features learned by the uncut feature learning module are input into the attention enhancement layer to focus on the relevant defective features; the focused features are output from the output layer to obtain the sub-image feature vector; Further, obtaining pixel coordinates in the sub-image corresponding to the sub-image feature vector; Further, a sub-image feature-position set is established based on the pixel coordinates and the sub-image feature vector; Further, using preset control parameters, a cutting simulation is performed on the sub-graph feature-position set to obtain simulated cutting information; Furthermore, the error between the simulated cutting information and the expected cutting information is calculated to obtain a cutting error vector; wherein the cutting error vector includes: size error, path error, geometric error and surface damage error; see Table 2, in which the following are given: Figure 4 Error information between simulated cutting and expected cutting at different pixel points of neutron image 1; Table 2 Error information between simulated cutting and expected cutting

[0032] Furthermore, the compensation control parameter is calculated using a control parameter compensation function; wherein the control parameter compensation function is expressed as: ;in, is the control parameter compensation function; is the pixel coordinate Compensation control parameters at ; is the pixel coordinate The cutting error vector at ; Simulates the function for cutting; is the pixel coordinate Preset control parameters at is the pixel coordinate The sub-image feature vector at ; is the pixel coordinate location information; In the embodiment of the present application, the compensation control parameters are calculated using the control parameter compensation function, so as to accurately adjust and compensate the key cutting parameters such as the cutting angle, speed, power, temperature and position of the laser cutting head. Refer to Table 3 and Table 4, which respectively give the changes of the control parameters before and after compensation for the pixel coordinate numbers L1 and L2; Table 3 Changes of control parameters before and after compensation of pixel coordinate number L1

[0033] Table 4 Changes of control parameters before and after compensation of pixel coordinate number L2

[0034] In the embodiment of the present application, a control parameter compensation function is used to adjust the preset control parameters. The function accurately calculates the error of each pixel coordinate and uses cutting simulation and feature vectors to dynamically adjust the preset control parameters to improve cutting accuracy and quality.

[0035] Further, the control panel 1 controls the laser cutting head 3 to perform cutting according to the compensation control parameter, and updates the cutting mark of the starting sub-image to be processed to 1; Furthermore, the preset control parameters of the uncut sub-image are adaptively adjusted according to the cut sub-image, corresponding to step S80; the specific process includes: Further, the next sub-image to be processed is obtained according to the sub-image sequence number and recorded as the current sub-image to be processed; Figure 4 From the content, we can know that the sub-image corresponding to sub-image number 2 is the sub-image to be processed; Further, see Figure 5 , merge the current sub-image 2 to be processed and the cut sub-image 1 to obtain the current extended sub-image to be processed; Further, the current extended processed sub-image is preprocessed by a preprocessing unit to obtain a standard current extended sub-image to be processed; Further, the cut sub-image in the standard current extended sub-image to be processed is subjected to influence analysis to obtain a cut influence vector; The process of obtaining the cut influence vector includes: The cut sub-image is identified as cut by the cutting state identification layer; Further, the cut sub-image is input into a cut feature learning module to obtain a cut feature vector; Further, the cut feature vector is outputted through the output layer; Furthermore, according to the sub-image connection relationship between the current sub-image to be processed and the cut image in the sub-image cross list, combined with the cut feature vector, the cut influence vector is obtained; it is known that the sub-image connection relationship between sub-image number 2 and sub-image number 1 is "left connection"; wherein, the calculation method of the cut influence vector is: ;in, is the cut impact vector; To build an analysis model through big data training of the previous and next steps in the sheet metal cutting process; is the cut feature vector; The weight analysis function of the influence of different connected cut areas on the current area to be cut; is the sub-image to be processed; is the cut sub-image; Further, the current sub-image to be processed in the standard current extended sub-image to be processed is input into the sheet metal feature learning model to obtain a feature vector of the current sub-image; Furthermore, the cut influence vector and the current sub-image feature vector are used to adjust the current preset control parameters; wherein, the adjustment formula for the preset control parameters of the existing cut area to the current area to be cut is: ; in, The pixel point of the current sub-image to be processed Compensation control parameters at ; is the number of cut influence vectors The weight value of the influencing factors after nonlinear changes; The pixel point of the current sub-image to be processed Preset control parameters at The pixel point of the current sub-image to be processed The sub-image feature vector of The pixel point of the current sub-image to be processed location information; is the number of influencing factors in the cut influence vector.

[0036] Referring to Table 5, according to the above calculation process, it can be obtained that sub-image 2 at pixel point Changes of control parameters before and after adjustment; Table 5 Pixels Changes of control parameters before and after compensation

[0037] In the embodiment of the present application, by analyzing the influence of the cut sub-image and the adaptive adjustment mechanism based on the feature learning model, the preset control parameters can be adjusted more accurately during the cutting process of the sheet metal. This method can effectively optimize the cutting quality, improve the processing efficiency and reduce the error.

[0038] Further, according to the adjusted control parameters, the laser cutting head 3 is controlled by the control panel 1 to Figure 5 The sub-image 2 is cut corresponding to the sheet metal part 5 to be processed; and the cutting state of the sub-image 2 in the sub-image cross list is recorded as 1; Further, according to the above content, the uncut areas of the sheet metal part 5 to be processed corresponding to sub-image 3, sub-image 4, sub-image 5, sub-image 6, sub-image 6, sub-image 7, sub-image 8 and sub-image 9 can be cut in sequence.

[0039] Furthermore, during the cutting process of the sheet metal part 5 to be processed, the display unit displays the cutting status and control parameters in real time.

[0040] In the embodiment of the present application, the management and control method based on the sub-image cross-link list can effectively improve the accuracy and efficiency of the laser cutting process. Specifically, by dividing the image of the sheet metal to be processed and combining the sheet metal feature learning model, the cutting parameters can be adjusted and optimized in real time, thereby reducing errors and improving the final cutting quality. During the cutting process, the laser cutting head simulates cutting according to the preset control parameters and the sub-image feature-position set, and dynamically adjusts the control parameters by comparing the errors between the simulation results and the expected results. This process further eliminates the influence caused by the processing error by calculating the compensation control parameters, thereby ensuring that the cutting accuracy of each sub-image meets the requirements. With the feedback of each step of the processing results, the control parameters will be gradually adjusted adaptively to adapt to the processing characteristics of different sub-images. In addition, as the entire cutting task proceeds, when the cutting marks of all sub-images in the sub-image cross-link list are updated to 1, it means that all the scheduled cutting tasks have been completed and the cutting operation is successfully completed. At this point, through the complete adaptive control process, the execution efficiency and cutting accuracy of the entire cutting task have been significantly improved.

[0041] Embodiment 2: In implementing a Figure 1 , Figure 2 and Figure 3 The content realizes the process control of adaptive laser cutting of sheet metal parts; in order to further reflect the applicability of the method of the present invention to the cutting of different sheet metal parts, the following is further explained through the implementation of this application, and the specific process is as follows: Use the control panel to move the laser cutting head to the corrected starting position of the sheet metal part to be processed; Further, acquiring an image of the sheet metal part to be processed; Further, the image of the sheet metal part to be processed is divided according to the cutting content to obtain N sub-images; Further, a sub-image cross-link list is established for the sub-image according to the image of the sheet metal part to be processed; Further, obtaining a starting sub-image to be processed; Further, the initial sub-image to be processed is preprocessed to obtain a standard first sub-image; Further, the standard first sub-image is input into the sheet metal feature learning model to obtain a sub-image feature vector; Further, obtaining pixel coordinates corresponding to the sub-image feature vector; Further, a sub-image feature-position set is established based on the pixel coordinates and the sub-image feature vector; Further, using preset control parameters, a cutting simulation is performed on the sub-graph feature-position set to obtain simulated cutting information; Further, the error between the simulated cutting information and the expected cutting information is calculated to obtain a cutting error vector; Further, the compensation control parameter is calculated using a control parameter compensation function; wherein the control parameter compensation function is expressed as: ;in, is the control parameter compensation function; is the pixel coordinate Compensation control parameters at ; is the pixel coordinate The cutting error vector at ; Simulates the function for cutting; is the pixel coordinate Preset control parameters at is the pixel coordinate The sub-image feature vector at ; is the pixel coordinate location information.

[0042] Further, the control panel controls the laser cutting head to perform cutting according to the compensation control parameter, and updates the cutting mark of the starting sub-image to be processed to 1; Further, according to the processing result of the initial sub-image to be processed, the preset control parameters in the subsequent processing steps are adaptively adjusted, and cutting is performed using a laser cutting head; The process of adaptively adjusting the preset control parameters in the subsequent processing steps includes: Further, the next sub-image to be processed is obtained according to the sub-image sequence number and recorded as the current sub-image to be processed; Further, a sub-image having the sub-image connection relationship with the current sub-image to be processed and having the cutting mark 1 is obtained, and recorded as a cut sub-image; Further, the current sub-image to be processed and the cut sub-image are merged to obtain the current extended sub-image to be processed; Further, the preprocessing is performed on the current extended sub-image to be processed to obtain a standard current extended sub-image to be processed; Further, an influence analysis is performed on the cut sub-image in the standard current extended sub-image to be processed to obtain a cut influence vector; Further, the current sub-image to be processed in the standard current extended sub-image to be processed is input into the sheet metal feature learning model to obtain a feature vector of the current sub-image; Further, adjusting the current preset control parameters according to the cut influence vector and the current sub-image feature vector; When the cutting marks in the sub-image cross list are all 1, the cutting task of the sheet metal part to be processed is completed.

[0043] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An adaptive laser cutting method based on characteristic parameters of sheet metal, characterized in that: include: Use the control panel to move the laser cutting head to the correction starting position; Acquire the image of the sheet metal part to be processed; Dividing the image of the sheet metal part to be processed according to the cutting content to obtain N sub-images; Establishing a sub-image cross-link list for the sub-image according to the sheet metal part image to be processed; wherein the sub-image cross-link list includes: processing marks, sub-image serial numbers and sub-image connection relationships; Obtaining a starting sub-image to be processed; Preprocessing the initial sub-image to be processed to obtain a standard first sub-image; Inputting the standard first sub-image into a sheet metal feature learning model to obtain a sub-image feature vector; Obtaining pixel coordinates corresponding to the sub-image feature vector; Establishing a sub-image feature-position set according to the pixel coordinates and the sub-image feature vector; Using preset control parameters to simulate cutting of the sub-graph feature-position set to obtain simulated cutting information; Calculating the error between the simulated cutting information and the expected cutting information to obtain a cutting error vector; Compensating the control parameter using a control parameter compensation function; Control the laser cutting head to perform cutting according to the compensation control parameter, and update the cutting mark of the starting sub-image to be processed to 1; The preset control parameters in the subsequent processing steps are adaptively adjusted according to the processing result of the initial sub-image to be processed, and cutting is performed; when the cutting marks in the cross linked list of the sub-image are all 1, the cutting task is completed.

2. The adaptive laser cutting method based on sheet metal feature parameters according to claim 1, characterized in that: Moving the laser cutting head to the corrected starting position includes: establishing a spatial coordinate system with the current position of the laser cutting head as the origin; Get the current top view image of the cutting platform; De-noising and contrast enhancing the current overhead image to obtain a first overhead image; Extracting the contour of the cutting platform and the contour of the sheet metal part to be processed in the first top view image to obtain a contour image of the cutting platform and a contour image of the sheet metal part to be processed; Calculating the contour deviation between the cutting platform contour image and the to-be-processed sheet metal part contour image, and constructing a contour deviation vector; Acquiring the expected cutting information of the sheet metal part to be processed, and characterizing it to obtain an expected feature vector; The moving path of the laser cutting head and the corrected starting position are generated according to the contour deviation vector and the expected feature vector; the laser cutting head moves to the corrected starting position according to the moving path.

3. The adaptive laser cutting method based on sheet metal feature parameters according to claim 2, characterized in that: The generation process of the moving path and the corrected starting position includes: The correction starting position on the contour image of the sheet metal part to be processed is calculated according to the contour deviation vector, and the calculation formula is: ;in, is the correction starting position; is the preset starting position; Calculate the function for correcting the starting position; To obtain the centroid coordinate point through the contour image of the sheet metal part to be processed; is the expected feature vector; is a set of boundary points of the contour image of the sheet metal part to be processed; is the correction parameter of the correction starting position; Acquire a set of candidate movable paths from the origin to the corrected starting position; The candidate movable path set is selected according to the contour deviation vector to obtain the moving path.

4. The adaptive laser cutting method based on sheet metal feature parameters according to claim 1, characterized in that: The control parameter compensation function is expressed as: ;in, is the control parameter compensation function; is the pixel coordinate Compensation control parameters at ; is the pixel coordinate The cutting error vector at ; Simulates the function for cutting; is the pixel coordinate Preset control parameters at is the pixel coordinate The sub-image feature vector at ; is the pixel coordinate location information.

5. The adaptive laser cutting method based on sheet metal feature parameters according to claim 1, characterized in that: The process of adaptively adjusting the preset control parameters in the subsequent processing steps includes: According to the sub-image sequence number, the next sub-image to be processed is obtained, which is recorded as the current sub-image to be processed; Obtain a sub-image that has the sub-image connection relationship with the current sub-image to be processed and the cutting mark is 1, and record it as a cut sub-image; wherein the sub-image connection relationship includes: upper connection, lower connection, left connection and right connection; Merging the current sub-image to be processed and the cut sub-image to obtain a current extended sub-image to be processed; Performing the preprocessing on the current extended sub-image to be processed to obtain a standard current extended sub-image to be processed; Performing an influence analysis on the cut sub-image in the standard current extended sub-image to be processed to obtain a cut influence vector; Inputting the current sub-image to be processed in the standard current extended sub-image to be processed into the sheet metal feature learning model to obtain a feature vector of the current sub-image; The current preset control parameter is adjusted according to the cut influence vector and the current sub-image feature vector.

6. An adaptive laser cutting system based on characteristic parameters of sheet metal parts, characterized in that: The system includes: an intelligent control unit, which controls the laser cutting head through a control panel; a monitoring unit, which is used to monitor the sheet metal cutting process; a collection unit, which is used to collect images during the monitoring process of the monitoring unit; a preprocessing unit, which is used to preprocess the collected images; a starting movement calculation unit, which is used to calculate the movement path of the laser cutting head and correct the starting position; a control parameter optimization unit, which is used to optimize the control parameters of the sheet metal cutting; and a display unit, which is used to display the cutting status and control parameters of the sheet metal.

7. The adaptive laser cutting system based on sheet metal feature parameters according to claim 6, characterized in that: The initial movement calculation unit comprises: Establish a spatial coordinate system with the current position of the laser cutting head as the origin; Get the current top view image of the cutting platform; De-noising and contrast enhancing the current overhead image to obtain a first overhead image; Extracting the contour of the cutting platform and the contour of the sheet metal part to be processed in the first top view image to obtain a contour image of the cutting platform and a contour image of the sheet metal part to be processed; Calculating the contour deviation between the cutting platform contour image and the to-be-processed sheet metal part contour image, and constructing a contour deviation vector; Obtain the expected cutting information of the sheet metal to be processed, and characterize it to obtain the expected feature vector; The moving path of the laser cutting head and the corrected starting position are generated according to the contour deviation vector and the expected feature vector.

8. The adaptive laser cutting system based on sheet metal feature parameters according to claim 6, characterized in that: The process of generating the moving path and correcting the starting position includes: The correction starting position on the contour image of the sheet metal part to be processed is calculated according to the contour deviation vector, and the calculation formula is: ;in, To correct the starting position; is the preset starting position; Calculate the function for correcting the starting position; To obtain the centroid coordinate point through the contour image of the sheet metal part to be processed; is the expected feature vector; is the boundary point set of the contour image of the sheet metal part to be processed; Correction parameters for correcting the starting position; Acquire a set of candidate movable paths from the origin to the corrected starting position; The candidate movable path set is selected according to the contour deviation vector to obtain the movable path.

9. The adaptive laser cutting system based on sheet metal feature parameters according to claim 6, characterized in that: The control parameter optimization unit comprises: Acquire the image of the sheet metal part to be processed; Dividing the image of the sheet metal part to be processed according to the cutting content to obtain N sub-images; Establishing a sub-image cross-link list for the sub-image according to the sheet metal part image to be processed; wherein the sub-image cross-link list includes: processing marks, sub-image serial numbers and sub-image connection relationships; Obtaining a starting sub-image to be processed; Preprocessing the initial sub-image to be processed to obtain a standard first sub-image; Inputting the standard first sub-image into a sheet metal feature learning model to obtain a sub-image feature vector; Obtaining pixel coordinates corresponding to the sub-image feature vector; Establishing a sub-image feature-position set according to the pixel coordinates and the sub-image feature vector; Using preset control parameters to simulate cutting of the sub-graph feature-position set to obtain simulated cutting information; Calculating the error between the simulated cutting information and the expected cutting information to obtain a cutting error vector; Calculating compensation control parameters using a control parameter compensation function; Control the laser cutting head to perform cutting according to the compensation control parameter, and update the cutting mark of the starting sub-image to be processed to 1; Adaptively adjusting the preset control parameters in the subsequent processing steps according to the processing result of the initial sub-image to be processed, and obtaining the corresponding next sub-image to be processed according to the sub-image sequence number, which is recorded as the current sub-image to be processed; Obtain a sub-image that has the sub-image connection relationship with the current sub-image to be processed and the cutting mark is 1, and record it as a cut sub-image; wherein the sub-image connection relationship includes: upper connection, lower connection, left connection and right connection; Merging the current sub-image to be processed and the cut sub-image to obtain a current extended sub-image to be processed; Performing the preprocessing on the current extended sub-image to be processed to obtain a standard current extended sub-image to be processed; Performing an influence analysis on the cut sub-image in the standard current extended sub-image to be processed to obtain a cut influence vector; Inputting the current sub-image to be processed in the standard current extended sub-image to be processed into the sheet metal feature learning model to obtain a feature vector of the current sub-image; The current preset control parameters are adjusted according to the cut influence vector and the current sub-image feature vector.

10. The adaptive laser cutting system based on sheet metal feature parameters according to claim 9, characterized in that: The control parameter compensation function is expressed as: ;in, is the control parameter compensation function; is the pixel coordinate Compensation control parameters at ; is the pixel coordinate The cutting error vector at ; Simulates the function for cutting; is the pixel coordinate Preset control parameters at is the pixel coordinate The sub-image feature vector at ; is the pixel coordinate location information.

Images