A laser heating system

Abstract

This invention relates to the field of laser processing and automated control technology, and particularly to a laser heating system, including an equipment device and a control device. The equipment device includes a support platform, a transmission mechanism, a laser, and vision components. The control device includes image acquisition, contour analysis, error compensation, coordinate optimization, and pulse sending modules. The image acquisition module drives the vision component to acquire material images and transmits them to the contour analysis module; the contour analysis module calculates deformation parameters and sends them to the error compensation module; the error compensation module combines pre-stored mechanical errors to generate a compensation matrix and sends it to the coordinate optimization module; the coordinate optimization module calculates and outputs the target coordinates to the pulse sending module; the pulse sending module drives the laser component to move and commands it to emit light. This invention solves the positioning deviation caused by hardware errors and changes in material size, improving control accuracy.

Description

Technical Field

[0001] The present invention relates to the technical fields of laser processing and automatic control, and particularly to a laser heating system. Background Art

[0002] The laser heating system mainly includes a beam emitting unit and a mechanical execution component. The laser module emits a beam with a high energy density. After the beam irradiates the surface of the material, it is quickly converted into heat energy, thereby completing the heating of a local area. The visual recognition technology uses an industrial camera to take digital images of the material. The upper computer processes the pixel distribution characteristics of the image and extracts the external contour of the material and the coordinates of the target processing point. Subsequently, the motion control module receives the coordinate command data, and the servo motor drives the transmission shaft to rotate, moving the laser module to a predetermined spatial position. The acquisition of visual data and the execution of mechanical displacement are closely coordinated to form an automatic processing closed-loop operation logic. According to the theoretical design expectation, the visual positioning and motion control can cooperate to complete the standardized production operation. However, in the actual industrial manufacturing process, the conventional operation mode relying on the mechanical structure often faces limitations and leads to positioning deviations.

[0003] The visual positioning and automatic precision processing technologies of existing laser heating equipment have the following technical pain points. Specifically, it is to solve the technical pain points that due to the hardware errors in the positioning method of existing laser heating equipment and the inability to adapt to the change of material size, the positioning of the processing starting point is inaccurate and the automation control precision of the laser heating position movement is low. Existing equipment usually relies on mechanical fixtures to fix the material and set the position reference. The long-term operation of mechanical components will cause wear and lead to physical position deviations. In addition, the fixtures with fixed structures can only match materials of a single specification. Once the outer contour of the input material fluctuates, the fixed fixture cannot perform self-adaptive adjustment of the size. The mechanical position deviation and the material size fluctuation directly cause the control system to be unable to extract the true processing origin coordinates. The control system executes the preset motion program according to the incorrect starting coordinate command, which will inevitably cause the linkage deviation of all subsequent heating points. For example, in the laser heating process of a circuit board, if there are manufacturing tolerances in the edge cutting of the fed circuit board, the mechanical baffle will push the circuit board to an incorrect coordinate deviating from the reference. When the laser module moves along the original trajectory set by the control system and executes the heating command, the laser beam will deviate from the specified heating target point. The deviation of the heating center not only causes the heating failure of the target area, but also causes the abnormal heating and damage of the surrounding sensitive electronic components. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides a laser heating system, which solves the technical problems that due to the hardware errors in the positioning method of existing laser heating equipment and the inability to adapt to the change of material size, the positioning of the processing starting point is inaccurate and the automation control precision of the laser heating position movement is low.

[0005] To solve the above-mentioned technical problems, the specific contents of the present invention are as follows: The present invention provides a laser heating system, comprising an equipment device and a control device, wherein the control device establishes a communication connection with the equipment device and is used to control the equipment device; the equipment device includes a support platform, a transmission mechanism, a laser component, and a vision component; the vision component moves synchronously with the laser component; characterized in that the control device includes an image acquisition module, a contour analysis module, an error compensation module, a coordinate optimization module, and a pulse transmission module; The image acquisition module drives the vision component to acquire the image matrix of the material to be processed, and transmits the image matrix to the contour analysis module; The contour analysis module receives the image matrix, analyzes the image matrix to calculate the deformation parameters, and transmits the deformation parameters to the error compensation module; The error compensation module receives the deformation parameters, extracts the pre-stored mechanical error data inside the control device, calculates the deformation parameters and the mechanical error data to generate a compensation matrix, and transmits the compensation matrix to the coordinate optimization module. The coordinate optimization module receives the compensation matrix, extracts the pre-stored processing point coordinates inside the control device, substitutes the processing point coordinates into the compensation matrix for calculation, outputs the target coordinates, and transmits the target coordinates to the pulse transmission module. The pulse transmission module receives the target coordinates, sends a drive signal to the transmission mechanism, drives the laser component to move to the physical position corresponding to the target coordinates, and controls the laser component to emit a laser beam after reaching the physical position.

[0006] Furthermore, in the laser heating system of the present invention, the device further includes a sliding base and a drive motor, and the transmission mechanism is configured to perform physical displacement processing: The drive motor and the sliding seat are connected to form the transmission mechanism, the laser component is mounted on the surface of the sliding seat, and the vision component is fixed to the side of the laser component; The contour parsing module is configured to perform proportional extraction processing: the image matrix obtained by the contour parsing module is a global pixel matrix, and the deformation parameter calculated by the parsing module is the space scaling ratio parameter of the material to be processed; The error compensation module is configured to perform tensor fusion processing: the error compensation module expands the space scaling ratio parameter of the material to be processed with the mechanical error data into a tensor operation to generate a multidimensional error compensation matrix as the compensation matrix; The coordinate optimization module is configured to perform algebraic mapping processing: the coordinate optimization module substitutes the processing point coordinates into the multidimensional error compensation matrix to perform algebraic operations, and outputs the absolute target coordinates as the target coordinates; The pulse sending module is configured to perform low-level drive processing: the pulse sending module converts the absolute target coordinates into motor drive pulse signals, and sends the motor drive pulse signals as drive signals to the drive motor inside the transmission mechanism; the equipment further includes a low-level base, a gantry frame, and a reference calibration plate; the support platform and the gantry frame are both installed above the low-level base, the gantry frame includes a main beam, the sliding seat is slidably connected to the surface of the main beam, and the reference calibration plate is fixed to the surface of the support platform; The image acquisition module is configured to perform reference data acquisition processing: before acquiring the global pixel matrix, the image acquisition module drives the vision component to acquire the reference calibration image matrix of the reference calibration board, and transmits the reference calibration image matrix to the error compensation module.

[0007] Furthermore, in the laser heating system of the present invention, the error compensation module is configured to perform spatial grid error extraction processing: The error compensation module receives the reference calibration image matrix, performs a corner feature extraction operator on the reference calibration image matrix, and obtains the pixel coordinates of multiple grid intersection points. The error compensation module extracts the preset ideal zero-point topological coordinates inside the control device, performs coordinate subtraction operation between the pixel coordinates of the multiple grid intersection points and the ideal zero-point topological coordinates, calculates the reference mechanical error vector as the mechanical error data, and stores the reference mechanical error vector inside the control device.

[0008] Furthermore, in the laser heating system of the present invention, the contour analysis module is configured to perform image denoising and continuous contour point set extraction processing: After receiving the global pixel matrix, the contour parsing module uses a Gaussian smoothing filter algorithm to remove electromagnetic noise from the global pixel matrix and outputs a noise-reduced pixel matrix. The contour parsing module uses a sub-pixel polynomial fitting operator to identify the gray-level step change region within the denoised pixel matrix and extracts a continuous set of physical contour points.

[0009] Furthermore, in the laser heating system of the present invention, the contour analysis module is configured to calculate the spatial scaling ratio parameter of the material to be processed based on the continuous physical contour point set: The contour parsing module performs a geometric center determination operation based on the continuous physical contour point set, calculates the position coordinates of the physical centroid of the material to be processed, and measures the radial distance values ​​from the position coordinates of the physical centroid to multiple pixels in the continuous physical contour point set. The contour analysis module extracts the standard material radial dimension values ​​pre-stored inside the control device, divides the multiple radial distance values ​​by the standard material radial dimension values, and outputs the space scaling ratio parameter of the material to be processed.

[0010] Furthermore, in the laser heating system of the present invention, the error compensation module is configured to perform cross-dimensional data underlying tensor fusion processing: After the error compensation module extracts the reference mechanical error vector stored inside the control device, it performs Kronecker product tensor operation on the reference mechanical error vector and the space scaling ratio parameter of the material to be processed. The error compensation module outputs the multidimensional error compensation matrix through the Kronecker product tensor operation.

[0011] Furthermore, in the laser heating system of the present invention, the coordinate optimization module is configured to perform coordinate nonlinear mapping processing: After receiving the multidimensional error compensation matrix, the coordinate optimization module substitutes the processing point coordinates into the multidimensional error compensation matrix to perform a nonlinear topological mapping dot product operation and calculates the preliminary compensation coordinates. The coordinate optimization module extracts a preset horizontal constant, adds the preset horizontal constant to the value of the preliminary compensation coordinates, and outputs the absolute target coordinates.

[0012] Furthermore, in the laser heating system of the present invention, the coordinate optimization module is configured to perform physical hardware spacing fusion processing: When the coordinate optimization module adds the preset horizontal constant to the value of the initial compensation coordinate, the preset horizontal constant is a fixed horizontal physical distance value between the central axis of the vision component and the central axis of the laser component. The coordinate optimization module performs a vector addition operation between the horizontal physical distance value and the preliminary compensation coordinates, and outputs the absolute target coordinates.

[0013] Furthermore, in the laser heating system of the present invention, the pulse sending module is configured to perform anti-resonance trajectory conversion processing: The pulse sending module receives the absolute target coordinates, obtains the current mechanical coordinates of the transmission mechanism, calls the inverse kinematics model to perform a difference calculation between the absolute target coordinates and the current mechanical coordinates, and calculates the physical step distance required for the drive motor to travel. The pulse sending module extracts a preset flexible acceleration / deceleration curve, modulates the motor drive pulse signal according to the physical step distance and the flexible acceleration / deceleration curve, and sends the motor drive pulse signal to the drive motor of the device.

[0014] Furthermore, in the laser heating system of the present invention, the pulse sending module is configured to perform optomechanical low-level linkage processing: The pulse sending module sends the motor driving pulse signal to the drive motor, driving the sliding seat to move the laser component to the physical position corresponding to the absolute target coordinates; After the sliding seat reaches the physical position corresponding to the absolute target coordinates, the pulse sending module sends a high-level enable signal to the laser component to control the laser component to emit a laser beam.

[0015] Beneficial effects of this invention: This invention provides a laser heating system that constructs a control architecture that deeply collaborates with digital signal processing algorithms on the basis of physical hardware spatial execution. An image acquisition module drives a vision component to acquire the pixel feature distribution state of the material surface. This, combined with a contour analysis module, outputs a scaling ratio parameter of the material's spatial dimensions, reflecting the material's cutting tolerances and thermal expansion / contraction. An error compensation module performs a Kronecker product tensor operation on the scaling ratio parameter of the material's spatial dimensions and a reference mechanical error vector, including the physical displacement component of the gantry's long-term wear, to generate a multidimensional error compensation matrix with nonlinear spatial compensation features. This tensor operation mechanism overcomes the physical limitations of existing single rigid mechanical fixtures with fixed positioning, integrating the underlying spatial mechanical deformation deviation factor into the surface material's asymmetric distortion coefficient, constructing a dynamic correction model with superimposed dual physical attributes. A coordinate optimization module substitutes the processing point coordinates from the theoretical design drawings into the multidimensional error compensation matrix for algebraic operations. Combined with the fixed horizontal physical distance between the vision component and the laser component, it performs vector addition operations, directly offsetting hardware assembly offsets and misalignments at the digital logic calculation level, outputting absolute target coordinates reflecting the absolute physical landing point coordinates. The pulse transmission module calls the inverse kinematics model to convert the absolute target coordinates into physical step distances, and introduces flexible acceleration and deceleration curve modulation to generate a nonlinear frequency-modulated motor drive pulse signal. The anti-resonance trajectory conversion processing logic smooths the acceleration and deceleration speed fluctuations during the high-frequency physical displacement of the sliding seat, eliminating the metal rigidity resonance caused by the long-span movement of the transmission mechanism. The closed-loop flow of multi-dimensional module data within the control device effectively solves the technical pain points of existing laser heating equipment that cannot adapt to changes in material size and is compounded by underlying mechanical hardware physical errors, improving the positioning accuracy of the laser module to the processing starting point and the automation control precision throughout its entire life cycle. Attached Figure Description

[0016] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on the accompanying drawings without creative effort.

[0017] Figure 1 This is a system architecture diagram of a laser heating system according to the present invention. Detailed Implementation

[0018] To make the technical solution of the present invention clearer, the present invention will be clearly and completely described below with reference to specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. The various embodiments of the present invention will be described in detail below with reference to the accompanying drawings. To better understand the purpose of the present invention, the present invention will be described in further detail below.

[0019] Please see Figure 1 This invention provides a laser heating system that constructs a collaborative architecture at the hardware level, integrating a physical space execution foundation and a digital signal processing hub. The equipment is built on a base, which provides a stable physical bearing reference surface for the entire system. A gantry and a support platform are fixed parallel to each other above the base, with the support platform providing a placement plane for the materials to be processed. The gantry has a main beam, and a sliding seat is connected to the surface of the main beam via a sliding guide connection. A drive motor and a transmission connection between the drive motor and the sliding seat form a transmission mechanism. This mechanism receives electrical commands and converts them into mechanical kinetic energy, thereby driving the sliding seat to physically displace along the direction of the main beam. A laser component is mounted on the surface of the sliding seat, and a vision component is rigidly fixed to the side of the laser component, ensuring that the vision component and the laser component maintain a fixed relative position in space and achieve synchronous displacement. Additionally, a reference calibration plate is fixedly placed on a predetermined edge area of ​​the support platform surface. The surface of the reference calibration plate is covered with a pre-defined pattern, providing an absolute physical space reference anchor point to the control device.

[0020] At the initial stage of the automated machining task, the image acquisition module embedded in the control device outputs a trigger level signal. The vision component receives the trigger level signal and activates the optical sensor to perform optical scanning sampling on the reference calibration plate area on the surface of the support platform. The optical sensor converts the acquired light signal into a digital grayscale array, generating a reference calibration image matrix. The reference calibration image matrix includes the actual pixel distribution characteristics of the geometric pattern of the reference calibration plate in the current mechanical coordinate system. Subsequently, the image acquisition module transmits the reference calibration image matrix to the error compensation module for low-level mechanical feature alignment. The error compensation module extracts the coordinates of the geometric edge intersection points in the reference calibration image matrix and performs difference calculation by comparing the extracted actual intersection point coordinates with the internally pre-stored ideal calibration coordinate data. Through coordinate difference calculation, the error compensation module calculates the physical space offset caused by long-term wear of the transmission mechanism and deformation of the gantry, generating mechanical error data representing the low-level physical state of the mechanical system and pre-stored it in the memory.

[0021] After completing the baseline alignment phase, the image acquisition module outputs a displacement command to the transmission mechanism, driving the vision component to move directly above the support platform and acquire the image matrix of the material to be processed. To meet the overall material size assessment requirements, the image matrix acquired by the image acquisition module is represented as a global pixel matrix covering the complete outline of the material. The image acquisition module transmits this global pixel matrix to the contour analysis module via an internal data bus. The contour analysis module applies an image thresholding segmentation algorithm to the global pixel matrix, separating the material entity pixel set from the support platform background pixel set, thereby extracting the outer contour boundary of the material entity. Along the outer contour boundary, the contour analysis module calculates the number of span pixels of the material in different spatial dimensions, and divides the current span pixel count by the internally stored standard material design drawing size parameters. Through this division, the contour analysis module outputs deformation parameters reflecting the actual cutting tolerance or thermal expansion and contraction state of the material. These deformation parameters are specifically represented as the spatial scaling ratio parameters of the material to be processed, which the contour analysis module transmits to the error compensation module.

[0022] In actual production applications of laser heating processes for circuit boards, the circuit boards undergo irregular stretching of their physical contours after the initial lamination and cutting processes. Direct positioning with a fixed fixture can cause surrounding sensitive electronic components to deviate from their design coordinates. The error compensation module receives the spatial scaling parameters of the material to be processed and extracts pre-stored mechanical error data. To eliminate the combined effects of circuit board contour stretching and underlying mechanical wear, the error compensation module constructs a two-dimensional spatial feature tensor. The module performs tensor operations on the vector features of the spatial scaling parameters of the material to be processed and the deviation vector of the mechanical error data, projecting the material's dimensional change rate in the length and width directions onto a grid coordinate system with a mechanical deformation factor. After the tensor operation is completed, the error compensation module generates a compensation matrix including nonlinear spatial compensation features; this compensation matrix is ​​expressed as a multidimensional error compensation matrix, which is then transmitted to the coordinate optimization module.

[0023] The coordinate optimization module receives the multidimensional error compensation matrix and extracts the pre-stored processing point coordinates from the control device's storage unit. These processing point coordinates represent the laser heating center anchor point specified on the theoretical design drawings. The module substitutes the horizontal and vertical values ​​of the processing point coordinates into the multidimensional error compensation matrix for algebraic operations, reconstructing the theoretical heating center anchor point according to the transformation coefficients within the matrix. The algebraic operation process uses matrix multiplication to convert theoretical coordinates to actual physical coordinates, outputting target coordinates that eliminate material deformation and mechanical wear errors. These target coordinates are specifically absolute target coordinates, representing the final landing point determined in physical space. Subsequently, the coordinate optimization module transmits the absolute target coordinates to the pulse transmission module. The pulse transmission module receives the absolute target coordinates, calls the inverse kinematics algorithm function, and converts the distance difference between the absolute target coordinates into a motor drive pulse signal recognizable by the servo controller. The pulse transmission module uses this motor drive pulse signal as a drive signal and sends it to the drive motor inside the transmission mechanism. The drive motor generates rotational torque based on the drive signal, driving the sliding seat and the laser component to move to the physical position corresponding to the absolute target coordinates. The transmission mechanism provides feedback on the electrical level of the physical position. After receiving the feedback level, the pulse transmission module controls the laser component to excite and emit a laser beam.

[0024] The error compensation module receives the reference calibration image matrix, loads the input data in the form of a two-dimensional pixel grayscale array into the memory processing stream, and performs corner feature extraction on the reference calibration image matrix. The corner feature extraction operator traverses the two-dimensional pixel grayscale array, searches for local maxima in the grayscale gradient covariance matrix by calculating the pixel grayscale gradient change rate in the two orthogonal directions (horizontal and vertical), and then outputs the pixel coordinates of multiple grid intersection points representing the local maxima. The extracted pixel coordinates of multiple grid intersection points represent the true mapping coordinates of the intersection points of the physical lines on the reference calibration plate on the current optical imaging physical surface. Subsequently, the error compensation module calls the internal communication bus to extract the preset ideal zero-point topological coordinates. The ideal zero-point topological coordinates are the coordinate array data of absolutely orthogonal grid nodes under theoretical conditions without any external stress interference. The error compensation module performs node-by-node subtraction operations on the pixel coordinates of multiple grid intersection points and the ideal zero-point topological coordinates, outputting a spatial displacement difference network with magnitude and direction attributes. The spatial displacement difference network constitutes the reference mechanical error vector as mechanical error data. The reference mechanical error vector is essentially a two-dimensional spatial vector field covering the entire processing area of ​​the bearing platform. In the application scenario of laser heating of large-format circuit boards, the gantry is affected by gravity and the transmission mechanism is worn by long-term operation. The reference mechanical error vector stored in the control device directly maps the spatial physical deformation state of the physical equipment at each grid node.

[0025] To obtain the true physical state of the material and eliminate environmental interference, the contour analysis module receives a global pixel matrix containing background and material entity information. It then performs a two-dimensional convolution operation on all pixels in the global pixel matrix using a Gaussian smoothing filter algorithm. The Gaussian smoothing filter algorithm weights the gray values ​​of the center pixel with those of the surrounding pixels according to spatial distance weights, filtering out electromagnetic noise signals mixed in by the high-frequency operation of motors in industrial production sites, thus outputting a smoothly transitioning, denoised pixel matrix. Based on the generated denoised pixel matrix, the contour analysis module applies a sub-pixel polynomial fitting operator to identify gray-level step change regions within the denoised pixel matrix. The sub-pixel polynomial fitting operator breaks the integer coordinate limitation of the physical pixel grid, constructing a continuous gray-level distribution polynomial function along the direction of maximum gray-level gradient. By using the differentiation rules of calculus to calculate the zero-crossing points of the second derivative of the polynomial function, the contour analysis module locates the true continuous physical boundary of the gray-level step change region and extracts a continuous physical contour point set with floating-point high-precision numerical characteristics. When using the Gaussian smoothing filter algorithm, the contour analysis module calls a preset five-row, five-column two-dimensional discrete Gaussian kernel matrix. The Gaussian smoothing filter algorithm extracts local pixel blocks within the global pixel matrix and performs spatial convolution on these local pixel blocks with a two-dimensional discrete Gaussian kernel matrix. The weight distribution of the convolution operation follows a two-dimensional normal probability density function, and weights are assigned based on the Euclidean distance from the center pixel to its neighboring pixels, outputting a denoised pixel matrix including smoothed gradient information. When applying the sub-pixel polynomial fitting operator, it extracts the integer pixel edge coordinates in the denoised pixel matrix where the gray-level gradient magnitude exceeds a preset threshold. A one-dimensional quadratic polynomial fitting curve equation is constructed centered on these integer pixel edge coordinates, following the direction of the maximum gradient. The quadratic polynomial fitting curve equation receives the gray-level values ​​of three adjacent pixels as input parameters and uses the least squares method to calculate the algebraic coefficients of each term. After calculating the algebraic coefficients, the sub-pixel polynomial fitting operator performs a first-order derivative operation on the quadratic polynomial fitting curve equation and sets the derivative value to zero, calculating the sub-pixel floating-point coordinates corresponding to the extreme points. Subpixel floating-point coordinates break through the integer spacing limitation of the physical grid of the photosensitive element. By using subpixel floating-point coordinates as node elements in a continuous set of physical contour points, a transformation mapping relationship between optical input signals and high-precision physical geometric coordinates is established.

[0026] After extracting the continuous physical contour point set, the contour analysis module performs a geometric center calculation based on the continuous physical contour point set. It then performs an arithmetic average of the horizontal and vertical coordinates of all boundary pixels within the continuous physical contour point set, outputting the position coordinates of the physical centroid of the material to be processed. Using the position coordinates of the physical centroid as the origin of the polar coordinate system, the contour analysis module scans outwards along a specified angular step parameter, calculating multiple radial distance values ​​from the position coordinates of the physical centroid to the corresponding angular pixels within the continuous physical contour point set. These multiple radial distance values ​​construct a polar coordinate distance array reflecting the actual outer contour shape of the material to be processed. Next, the contour analysis module extracts the standard material radial dimension values ​​pre-stored within the control device. These standard material radial dimension values ​​are derived from the theoretical reference distance array on the design drawings at the same step angle. The contour analysis module performs a division operation between the multiple radial distance values ​​and the standard material radial dimension values ​​according to the corresponding index dimension, outputting the spatial scaling parameter of the material to be processed. The scaling parameter of the material to be processed is represented as a floating-point array structure that includes stretching or shrinkage coefficients in multiple angular directions. In the laser hot pressing process of flexible circuit boards, the flexible circuit board undergoes asymmetric thermal expansion and contraction deformation due to the influence of the previous pressing process. The resulting scaling parameter of the material to be processed independently characterizes the degree of local physical stretching and shrinkage of the flexible circuit board in all angular dimensions.

[0027] The error compensation module extracts the reference mechanical error vector stored internally in the control device. This reference mechanical error vector includes two-dimensional physical deformation mapping network data covering the processing area of ​​the underlying base. After extracting the reference mechanical error vector, the error compensation module receives the scaling parameters of the material space to be processed output from the pre-calculation stage. These scaling parameters influence the local stretching and shrinkage property coefficients of the material in multiple specific angular dimensions. In large-format flexible circuit board laser thermoforming applications, the flexible circuit board is prone to asymmetric physical stretching deformation after high-temperature pressing and lamination. Simultaneously, the gantry frame experiences physical sagging in the middle region due to the gravity of its metal components. To integrate the physical sagging phenomenon with the asymmetric physical stretching deformation in the underlying algorithm architecture, the error compensation module performs a Kronecker product tensor operation between the scaling parameters of the material space to be processed and the reference mechanical error vector. The Kronecker product multiplication tensor operation breaks the dimensionality limitation of linear addition, expanding the two-dimensional physical deformation mapping network data into a higher-order tensor space based on the stretching and contraction attribute coefficients at specific angular dimensions. This expansion calculation creates a deep algebraic coupling between the local deformation properties of the flexible circuit board and the physical offset of the underlying gantry mesh. The error compensation module outputs a multidimensional error compensation matrix through the Kronecker product multiplication tensor operation. This matrix establishes a full-space nonlinear coordinate transformation data set, including the underlying mechanical physical loss components and the surface material size distortion components. When the error compensation module executes the Kronecker product multiplication tensor operation, the input reference mechanical error vector is represented as a two-dimensional matrix structure with the number of rows and columns corresponding to the number of physical mesh divisions of the bearing platform. Each array element within the reference mechanical error vector includes both horizontal and vertical physical offset values. The input scaling parameter of the material to be processed is represented as a one-dimensional floating-point array structure, which includes local deformation scalar coefficients at multiple angular step dimensions. The Kronecker product multiplication tensor operation expands the dimension of each array element within a two-dimensional matrix structure by performing an outer product operation with the entire one-dimensional floating-point array structure, outputting a structured, multi-dimensional error compensation matrix. This multi-dimensional error compensation matrix synchronously includes the two-dimensional spatial position index information of the reference mechanical error vector and the one-dimensional angular direction variable information of the scaling parameter of the material to be processed. Through the dimensional expansion calculation of the input and output matrices, the Kronecker product multiplication tensor operation mathematically achieves data-level coupling between the displacement deviation data of the underlying gantry physical mesh and the scalar data of the local deformation of the flexible circuit board, providing the coordinate optimization module with the underlying calculation basis including all-dimensional physical compensation parameters.

[0028] The coordinate optimization module receives the multidimensional error compensation matrix and uses the internal communication bus to extract the processing point coordinates stored within the control device. These processing point coordinates represent the theoretical design position data under ideal conditions, signifying the original absolute anchor point for laser action planned in the product design drawings. The coordinate optimization module substitutes the horizontal and vertical values ​​of the processing point coordinates into the multidimensional error compensation matrix and performs a nonlinear topological mapping dot product operation on the substituted values. This nonlinear topological mapping dot product operation utilizes the algebraic coupling relationship within the multidimensional error compensation matrix to perform spatial dimension distortion and coordinate reconstruction on the theoretical design position data. The coordinate reconstruction mechanism corrects the original absolute anchor point for laser action based on physical sag phenomena and asymmetric physical stretching deformation, calculating preliminary compensation coordinates that perfectly match the actual physical undulations of the flexible circuit board. These preliminary compensation coordinates represent intermediate transformation state data from a single sensor perspective, reflecting the physical three-dimensional coordinates required for the optical lens center of the vision component to accurately align with the target processing position. The coordinate optimization module extracts a preset horizontal constant stored within the control device and adds this constant to the values ​​of the preliminary compensation coordinates, advancing the intermediate transformation state data to the final drive control dimension.

[0029] When the coordinate optimization module adds a preset horizontal constant to the initial compensation coordinates, this preset horizontal constant originates from the underlying hardware assembly physical constraints of the equipment. In the physical spatial layout of the transmission mechanism, the vision component and the laser component are fixedly mounted side-by-side on the sliding seat surface. There is an objectively fixed horizontal physical distance between the central axis of the vision component and the central axis of the laser component. This fixed horizontal physical distance constitutes the core physical meaning of the preset horizontal constant. Since the initial compensation coordinates represent the command data when the optical field center of the vision component aligns with the actual processing position, if the control device directly follows the generator drive pulse signal under the initial compensation coordinates, the beam emitted by the laser component will deviate from the actual processing anchor point. The coordinate optimization module extracts the fixed horizontal physical distance value and performs a vector addition operation with the initial compensation coordinates. The vector addition operation performs a one-dimensional vector translation adjustment calculation on the command data along the straight line direction of the fixed assembly offset between the vision component and the laser component. This one-dimensional vector translation adjustment calculation mathematically cancels out the physical installation distance misalignment between the vision component and the laser component, directly outputting the absolute target coordinates of the final beam's actual landing point spatial information to the pulse sending module.

[0030] The pulse sending module receives the absolute target coordinates, which are the final three-dimensional spatial position command data that integrates the underlying hardware offset vector and the surface material size distortion coefficient. Before executing the actual physical drive command, the pulse sending module obtains the current mechanical coordinates of the transmission mechanism, which represent the original physical stopping position of the sliding seat on the gantry main beam. For spatial movement requirements, the pulse sending module calls the inverse kinematics model to perform a difference operation between the absolute target coordinates and the current mechanical coordinates. The difference operation extracts the three-dimensional coordinate values ​​of the target endpoint and the original position point and executes the corresponding subtraction operation logic to calculate the data array representing the spatial linear offset vector. The inverse kinematics model, combined with the gear transmission ratio parameters inside the equipment, performs a scalar transformation on the spatial linear offset vector, thereby calculating the physical step distance required for the drive motor to travel. The physical step distance is reflected in the number of rotations and angle values ​​that the drive motor rotor needs to complete.

[0031] The pulse sending module extracts a preset flexible acceleration / deceleration curve, which includes a set of nonlinear discrete data relating time and velocity variables. In high-speed electronic manufacturing applications, a gantry crane drives a wide sliding block in high-frequency start-stop physical movements. Simple linear acceleration triggers high-frequency rigid mechanical resonance in the mechanical structure. Based on the physical step distance and the flexible acceleration / deceleration curve, the pulse sending module divides the total travel distance into three continuous control stages: acceleration, constant speed, and deceleration. It dynamically changes the frequency change rate of the output pulse sequence based on the nonlinear discrete data set. After dynamic frequency adjustment calculations, the pulse sending module modulates a motor drive pulse signal, which is a high-low level pulse train with a constant duty cycle but a nonlinearly changing trigger frequency. After generating the pulse train, the pulse sending module sends the motor drive pulse signal to the drive motor of the equipment.

[0032] The pulse transmission module sends motor drive pulse signals to the drive motor. The internal coil of the drive motor converts the received pulse train into a fixed rotating magnetic field torque. This rotating magnetic field torque is converted into a linear physical thrust through a transmission mechanism, driving the sliding block to move the laser component to the physical position corresponding to the absolute target coordinates. The control device continuously monitors the real-time feedback level signal uploaded by the underlying position encoder. After the sliding block reaches the physical position corresponding to the absolute target coordinates, the pulse transmission module triggers the optomechanical linkage electrical execution logic. The pulse transmission module sends a high-level enable signal to the laser component. This high-level enable signal directly connects the internal power supply circuit of the semiconductor light-emitting array through the drive power interface. The semiconductor light-emitting array receives high-level excitation and activates its circuit, instructing the laser component to emit a laser beam.

[0033] The error compensation module receives the reference calibration image matrix from the image acquisition module and performs spatial grid error extraction data stream processing. The error compensation module performs corner feature extraction on the reference calibration image matrix to extract the pixel coordinates of multiple grid intersection points representing the mapping positions of the physical scribe line intersections. Simultaneously, the error compensation module extracts the preset ideal zero-point topological coordinates within the control device, and performs a coordinate subtraction operation between the multiple grid intersection pixel coordinates and the ideal zero-point topological coordinates to calculate the reference mechanical error vector. The specific mathematical formula for the coordinate subtraction operation is as follows:

[0034] In the formula, Represents the reference mechanical error vector. Represents the pixel coordinates of the intersection of multiple grid points. Represents the topological coordinates of the ideal zero point. Represents the horizontal index number of a two-dimensional spatial grid. This represents the vertical index of the two-dimensional spatial grid. The generated baseline mechanical error vector, obtained through matrix interpolation, is stored within the control device as the basic mechanical error data input source for subsequent tensor fusion algorithms. The contour analysis module receives the global pixel matrix from the image acquisition module, removes electromagnetic noise from the industrial environment using a Gaussian smoothing filter, and extracts a continuous physical contour point set from regions of abrupt grayscale changes. The contour analysis module then performs a geometric center calculation based on this continuous physical contour point set, determining the coordinates of the physical centroid of the material to be processed. The formula for the geometric center calculation is as follows:

[0035] In the formula, The coordinates representing the physical centroid. Represents the total number of boundary pixels within a continuous set of physical contour points. Represents the first point within a continuous set of physical contour points. The coordinates of each pixel This represents the summation of the coordinates of all boundary pixels within a continuous set of physical contour points. This represents the sequence number of the cumulative operation.

[0036] After calculating the coordinates of the physical centroid, the contour analysis module measures the radial distances from the centroid to multiple pixels within a continuous set of physical contour points. The calculation process uses the Euclidean distance equation, and the specific formula is as follows:

[0037] In the formula, Represents the first of multiple radial distance values A radial distance value, Represents the first in the continuous physical contour point set The horizontal coordinates of each pixel Represents the first in the continuous physical contour point set The vertical coordinates of each pixel A horizontal scalar representing the position coordinates of the physical centroid. The vertical scalar representing the position coordinates of the physical centroid. This represents the pixel sequence number. After obtaining the actual measured distance data, the contour analysis module extracts the standard material radial dimension values ​​pre-stored within the control device. The contour analysis module divides multiple radial distance values ​​by the standard material radial dimension values ​​and outputs the space scaling ratio parameter for the material to be processed. The scaling ratio extraction and division formula is as follows:

[0038] In the formula, The first parameter in the sequence representing the scaling ratio of the space of the material to be processed One proportional parameter, Represents the first of multiple radial distance values A radial distance value, The first of the standard material radial dimension values A theoretical design dimension value, The symbol represents the feature sequence number, and the fraction bar represents the division operator for proportional calculation. The generated scaling parameters of the material to be processed are then transmitted to the error compensation module to perform underlying data fusion.

[0039] The error compensation module extracts the reference mechanical error vector stored internally in the control device. It then performs a Kronecker product tensor operation on the scaling parameters of the material to be processed against the reference mechanical error vector, outputting a multidimensional error compensation matrix containing both physical mechanical deformation components and material dimensional deformation components. The Kronecker product tensor operation formula is as follows:

[0040] In the formula, Represents a multidimensional error compensation matrix. Represents the reference mechanical error vector. A vector parameter array representing the scaling parameters of the space of the material to be processed. This represents the Kronecker product tensor multiplication operator. After the multidimensional error compensation matrix integrates the underlying error characteristics, it is transmitted to the coordinate optimization module via the system communication bus. The coordinate optimization module receives the multidimensional error compensation matrix and extracts the pre-stored machining point coordinates within the control device. The coordinate optimization module substitutes the machining point coordinates into the multidimensional error compensation matrix to perform a nonlinear topological mapping dot product operation, thereby calculating the preliminary compensation coordinates reflecting the actual deformation position. The formula for the nonlinear topological mapping dot product operation is as follows:

[0041] In the formula, Represents the initial compensation coordinates. This represents the coordinates of the machining points specified in the design drawings. This represents a multi-dimensional error compensation matrix. The coordinate optimization module extracts a preset horizontal constant, adds this preset horizontal constant to the initial compensation coordinate values, and performs physical hardware spacing fusion processing to output the absolute target coordinates. The vector addition formula is expressed as:

[0042] In the formula, Represents the absolute target coordinates. Represents the initial compensation coordinates. This represents the preset horizontal constant, which is a fixed horizontal physical distance between the central axis of the vision component and the central axis of the laser component.

[0043] The pulse transmission module receives the absolute target coordinates, obtains the current mechanical coordinates of the transmission mechanism, and calls the inverse kinematics model to perform a difference calculation between the absolute target coordinates and the current mechanical coordinates to calculate the physical step distance required for the drive motor to travel. The formula for the inverse kinematics model difference calculation is as follows:

[0044] In the formula, Represents the physical step distance. Represents the absolute target coordinates. Represents the current mechanical coordinates of the transmission mechanism. This represents the matrix subtraction operator. The pulse sending module extracts the preset flexible acceleration / deceleration curve, modulates the motor drive pulse signal according to the physical step distance and the flexible acceleration / deceleration curve, and sends the motor drive pulse signal to the drive motor of the device.

[0045] The actual data operation implementation in specific production application scenarios provides a benchmark for verifying the underlying data processing link. The image acquisition module drives the vision component to capture images of the benchmark calibration board to obtain the benchmark calibration image matrix. The error compensation module extracts multiple grid intersection pixel coordinates as two-dimensional feature points composed of a horizontal coordinate of 102 mm and a vertical coordinate of 101 mm. The error compensation module extracts the ideal zero-point topological coordinates as two-dimensional target points composed of a horizontal coordinate of 100 mm and a vertical coordinate of 100 mm. The error compensation module subtracts the multiple grid intersection pixel coordinates from the ideal zero-point topological coordinates, outputting a benchmark mechanical error vector as a spatial vector composed of a horizontal deviation of +2 mm and a vertical deviation of +1 mm. The contour analysis module performs geometric center calculation on a continuous physical contour point set to obtain the position coordinates of the physical centroid. Based on the calculation equation, it obtains multiple radial distance values ​​of 51 mm. The contour analysis module extracts the standard material radial dimension value of 50 mm, divides the multiple radial distance values ​​by the standard material radial dimension value, and outputs a space scaling parameter of 1.02 for the material to be processed. The error compensation module performs a Kronecker product tensor operation on the material to be processed space scaling parameter of 1.02 and the reference mechanical error vector, outputting a multidimensional error compensation matrix with a horizontal coordinate component of 2.04 mm and a vertical coordinate component of 1.02 mm. The coordinate optimization module extracts the processing point coordinates as a horizontal coordinate of 200 mm and a vertical coordinate of 200 mm, and performs a nonlinear topological mapping dot product operation on the processing point coordinates with the multidimensional error compensation matrix to obtain preliminary compensation coordinates of a horizontal coordinate of 202.04 mm and a vertical coordinate of 201.02 mm. The coordinate optimization module extracts the preset horizontal constant as the horizontal physical distance value of a horizontal coordinate of 60 mm and a vertical coordinate of 0 mm, and adds the preset horizontal constant to the preliminary compensation coordinates to calculate the absolute target coordinates of a horizontal coordinate of 262.04 mm and a vertical coordinate of 201.02 mm. The pulse sending module obtains the current mechanical coordinates of the transmission mechanism as 0 mm in the horizontal coordinate and 0 mm in the vertical coordinate. It performs a difference calculation between the absolute target coordinates and the current mechanical coordinates, and outputs the physical step distance required for the drive motor to travel as 262.04 mm in the horizontal axis and 201.02 mm in the vertical axis. The pulse sending module modulates the motor drive pulse signal according to the physical step distance and sends it to the drive motor, which drives the sliding seat to move the laser component precisely to the corresponding physical position.

[0046] The construction process of the inverse kinematics model includes mechanical transmission ratio parameter mapping and matrix transformation steps. The control device pre-extracts the rated pulse equivalent parameters of the drive motor inside the transmission mechanism and the physical parameters of the transmission screw lead, storing them in memory. The inverse kinematics model receives the physical step distance input from the pulse sending module and inputs it into the internal scalar transformation matrix. The scalar transformation matrix divides the linear offset distance in three-dimensional space by the physical parameters of the transmission screw lead to calculate the absolute number of revolutions required by the drive motor. The inverse kinematics model multiplies the absolute number of revolutions by the rated pulse equivalent parameters to calculate the total number of pulses required for the motor drive pulse signal. The construction process of the flexible acceleration / deceleration curve adopts a seven-segment nonlinear Bézier curve algorithm. The control device presets the maximum operating speed parameters, maximum acceleration parameters, and abrupt change limit parameters. The control device inputs the physical step distance of the total stroke into the seven-segment nonlinear Bézier curve algorithm. The seven-segment nonlinear Bézier curve algorithm calculates the smooth transition frequency conversion time node sequence based on the speed limit parameter, modulates the frequency change rate of the motor drive pulse signal, and outputs a high and low level pulse train with a constant duty cycle but a nonlinear change in trigger frequency.

[0047] When the sub-pixel polynomial fitting operator extracts integer pixel edge coordinates in the denoised pixel matrix where the gray-level gradient magnitude exceeds a preset threshold, the specific value of the preset threshold is set to an integer pixel gray-level difference between 45 and 60. The control device obtains the global average gray-level parameter of the global pixel matrix and dynamically adjusts the accurate value of the preset threshold based on the global average gray-level parameter. When the global average gray-level parameter is greater than 128, the control device sets the preset threshold to 60; when the global average gray-level parameter is less than or equal to 128, the control device sets the preset threshold to 45. By setting the preset threshold between 45 and 60, the sub-pixel polynomial fitting operator can effectively eliminate false edge pixels caused by local environmental reflections, accurately locate the gray-level step change region where the real physical boundary is located, and establish a conversion mapping relationship between the optical input signal and high-precision physical geometric coordinates.

[0048] Embodiment 1 of the present invention: In a laser heating production scenario for a rigid printed circuit board with dimensions of 500 mm by 500 mm, the equipment and control device are in an online communication state. The image acquisition module sends a trigger command to the vision component, which optically scans the reference calibration plate fixed on the surface of the support platform to generate a reference calibration image matrix. The error compensation module performs corner feature extraction operations on the reference calibration image matrix to obtain the pixel coordinates of multiple grid intersection points with a grid size of 10 mm by 10 mm. The control device internally presets an ideal zero-point topological coordinate system with absolute orthogonality. The error compensation module performs coordinate subtraction operations between the pixel coordinates of multiple grid intersection points and the ideal zero-point topological coordinate system to calculate the reference mechanical error vector, including the small physical deformation of the gantry, and stores the reference mechanical error vector internally in the control device. After completing the underlying physical reference alignment operation, the image acquisition module drives the vision component to move directly above the rigid printed circuit board to obtain the global pixel matrix and transmit it to the contour analysis module. The contour analysis module uses a Gaussian smoothing filter algorithm to remove industrial electromagnetic noise from the global pixel matrix, outputting a denoised pixel matrix. By applying a sub-pixel polynomial fitting operator, the contour analysis module locates regions of gray-level step changes within the denoised pixel matrix, extracting a continuous physical contour point set. Based on this continuous physical contour point set, the module calculates the position coordinates of the physical centroid and measures multiple radial distances from these coordinates to multiple pixels within the continuous physical contour point set. The control device pre-stores standard material radial dimensions from the design drawings; the module divides these radial distances by the standard material radial dimensions to output the scaling parameters for the material to be processed. The error compensation module extracts a baseline mechanical error vector and performs a Kronecker product tensor operation on the scaling parameters for the material to be processed and the baseline mechanical error vector to generate a multidimensional error compensation matrix. The coordinate optimization module substitutes the processing point coordinates into the multidimensional error compensation matrix and performs a nonlinear topological mapping dot product operation to calculate the preliminary compensation coordinates. There is a 60mm horizontal physical distance between the central axis of the vision component and the central axis of the laser component. The coordinate optimization module performs vector addition on the horizontal physical distance and the preliminary compensation coordinates to output the absolute target coordinates. The pulse sending module calls the inverse kinematics model to perform difference calculation between the absolute target coordinates and the current mechanical coordinates, calculates the physical step distance required for the drive motor to travel, and modulates the motor drive pulse signal. The pulse sending module sends the motor drive pulse signal to the drive motor, and the sliding seat moves the laser component to the physical position corresponding to the absolute target coordinates. After the sliding seat reaches the physical position corresponding to the absolute target coordinates, the pulse sending module sends a high-level enable signal to the laser component to command the laser beam to be emitted.

[0049] Embodiment 2 of the present invention: In the laser heating scenario of flexible circuit boards made of polyimide, the flexible circuit boards are prone to asymmetric physical stretching and shrinkage distortion after the initial high-temperature pressing process. The image acquisition module drives the vision component to acquire a global pixel matrix including the physical region of the flexible circuit board; the contour analysis module uses a Gaussian smoothing filter algorithm to process the global pixel matrix and output a denoised pixel matrix, and applies a sub-pixel polynomial fitting operator to identify the gray-level step change region within the denoised pixel matrix to extract a continuous physical contour point set. The contour analysis module performs a geometric center calculation operation on the continuous physical contour point set to obtain the position coordinates of the physical centroid, and calculates 180 radial distance values ​​from the position coordinates of the physical centroid to the corresponding pixel point within the continuous physical contour point set in a 360-degree step angle. The control device pre-stores standard material radial dimension values, and the contour analysis module divides the 180 radial distance values ​​one by one with the standard material radial dimension values, outputting a space scaling ratio parameter of the material to be processed, including multi-directional deformation coefficients. The error compensation module extracts the reference mechanical error vector stored internally in the control device. This reference mechanical error vector records the inherent mechanical wear deviation of the gantry and sliding seat across their length and width spans. Considering that the asymmetric stretching of the flexible circuit board and the underlying mechanical wear are deformation factors in different physical dimensions, the error compensation module performs a Kronecker product tensor operation on the reference mechanical error vector and the scaling parameter of the material to be processed. The tensor operation incorporates the two-dimensional spatial mechanical deformation factor into the stretching coefficient of the angular dimension, generating a multidimensional error compensation matrix that includes cross-integrated physical quantity features. The coordinate optimization module receives the multidimensional error compensation matrix and substitutes the processing point coordinates defined in the original product drawings into the multidimensional error compensation matrix to perform a nonlinear topological mapping dot product operation to calculate the preliminary compensation coordinates. The coordinate optimization module extracts a preset horizontal constant, which corresponds to the mechanically fixed distance between the vision component and the laser component on the main beam. The coordinate optimization module adds the preset horizontal constant to the value of the preliminary compensation coordinates and outputs the absolute target coordinates. The pulse sending module obtains the current mechanical coordinates based on the absolute target coordinates and calculates the physical step distance required for the drive motor to travel. The pulse sending module modulates the motor drive pulse signal according to the physical step distance and sends it to the drive motor of the device. The drive motor drives the sliding seat to move the laser component to the physical position corresponding to the absolute target coordinates. After reaching the physical position, the laser component is instructed to release the beam energy.

[0050] Embodiment 3 of the present invention: In the application scenario of laser heating curing of edge sealant for large-size LCD panels, the span of the support platform reaches 1500 mm, and the transmission mechanism is prone to rigid mechanical resonance during long-distance physical displacement. The error compensation module inside the control device pre-extracts the pixel coordinates of multiple grid intersection points of the benchmark calibration board, subtracts these coordinates from the ideal zero-point topological coordinates to obtain the benchmark mechanical error vector, and stores it inside the control device. The image acquisition module drives the vision component to acquire the global pixel matrix of the large-size LCD panel. The contour analysis module applies a sub-pixel polynomial fitting operator to extract a continuous set of physical contour points, calculates the position coordinates of the physical centroid and multiple radial distance values, and outputs the scaling ratio parameter of the material to be processed. The error compensation module performs a Kronecker product tensor operation on the scaling ratio parameter of the material to be processed and the benchmark mechanical error vector to output a multidimensional error compensation matrix. The coordinate optimization module performs a nonlinear topological mapping dot product operation on the multidimensional error compensation matrix and the processing point coordinates to obtain preliminary compensation coordinates. It then performs a vector addition operation with the preliminary compensation coordinates, combining the fixed horizontal physical distance between the central axis of the vision component and the central axis of the laser component, to output the absolute target coordinates. The pulse sending module acquires the current mechanical coordinates of the transmission mechanism and uses the inverse kinematics model to perform a difference operation between the absolute target coordinates and the current mechanical coordinates, converting it into the physical step distance required for the drive motor to travel. To suppress metal structure vibration caused by long-span movement, a smooth physical speed transition mechanism is needed. The pulse sending module extracts a preset flexible acceleration / deceleration curve from the control device and modulates the motor drive pulse signal according to the physical step distance and the flexible acceleration / deceleration curve. The frequency change rate of the motor drive pulse signal exhibits a smooth nonlinear state during the start-up and stop phases. The pulse sending module sends a motor drive pulse signal to the drive motor. The rotational torque generated by the drive motor drives the sliding seat, which in turn moves the laser component smoothly across the long distance to the physical position corresponding to the absolute target coordinates. After detecting that the physical displacement of the sliding seat has stopped, the pulse sending module sends a high-level enable signal to the laser component to activate the internal power supply circuit and emit a high-energy beam.

[0051] To verify the positioning accuracy and automated control precision of the laser heating system, this invention conducted a long-term comparative experiment. The comparative experiment used a traditional mechanical fixture positioning system as a parallel comparison with the laser heating system provided by this invention. The test conditions were set to run continuously for 1000 hours, processing 10,000 rigid printed circuit boards with dimensions of 500 mm x 500 mm. Test data shows that after 500 hours of continuous operation, the existing mechanical fixture positioning system experienced a processing point offset of 0.08 mm due to objective wear in the transmission mechanism. The laser heating system provided by this invention, using Kronecker product tensor operations and an inverse kinematics model for dynamic error compensation, maintained a final positioning offset of the absolute target coordinates within 0.01 mm after 1000 hours of continuous operation. The experimental data fully demonstrates that the laser heating system effectively solves the positioning deviation problem caused by the superposition of underlying mechanical hardware physical errors and material size fluctuations, substantially improving the product yield in automated processing.

Claims

1. A laser heating system, characterized in that, The system includes a device and a control device. The control device establishes a communication connection with the device and is used to control the device. The device includes a support platform, a transmission mechanism, a laser component, and a vision component. The vision component moves synchronously with the laser component. The control device includes an image acquisition module, a contour analysis module, an error compensation module, a coordinate optimization module, and a pulse transmission module. The image acquisition module drives the vision component to acquire the image matrix of the material to be processed, and transmits the image matrix to the contour analysis module; The contour analysis module receives the image matrix, analyzes the image matrix to calculate the deformation parameters, and transmits the deformation parameters to the error compensation module; The error compensation module receives the deformation parameters, extracts the pre-stored mechanical error data inside the control device, calculates the deformation parameters and the mechanical error data to generate a compensation matrix, and transmits the compensation matrix to the coordinate optimization module. The coordinate optimization module receives the compensation matrix, extracts the pre-stored processing point coordinates inside the control device, substitutes the processing point coordinates into the compensation matrix for calculation, outputs the target coordinates, and transmits the target coordinates to the pulse transmission module. The pulse transmission module receives the target coordinates, sends a drive signal to the transmission mechanism, drives the laser component to move to the physical position corresponding to the target coordinates, and controls the laser component to emit a laser beam after reaching the physical position.

2. The laser heating system according to claim 1, characterized in that, The device further includes a sliding base and a drive motor, and the transmission mechanism is configured to perform physical displacement processing: The drive motor and the sliding seat are connected to form the transmission mechanism, the laser component is mounted on the surface of the sliding seat, and the vision component is fixed to the side of the laser component; The contour parsing module is configured to perform proportional extraction processing: the image matrix obtained by the contour parsing module is a global pixel matrix, and the deformation parameter calculated by the parsing module is the space scaling ratio parameter of the material to be processed; The error compensation module is configured to perform tensor fusion processing: the error compensation module expands the space scaling ratio parameter of the material to be processed with the mechanical error data into a tensor operation to generate a multidimensional error compensation matrix as the compensation matrix; The coordinate optimization module is configured to perform algebraic mapping processing: the coordinate optimization module substitutes the processing point coordinates into the multidimensional error compensation matrix to perform algebraic operations, and outputs the absolute target coordinates as the target coordinates; The pulse sending module is configured to perform low-level drive processing: the pulse sending module converts the absolute target coordinates into motor drive pulse signals, and sends the motor drive pulse signals as drive signals to the drive motor inside the transmission mechanism; the equipment further includes a low-level base, a gantry frame, and a reference calibration plate; the support platform and the gantry frame are both installed above the low-level base, the gantry frame includes a main beam, the sliding seat is slidably connected to the surface of the main beam, and the reference calibration plate is fixed to the surface of the support platform; The image acquisition module is configured to perform reference data acquisition processing: before acquiring the global pixel matrix, the image acquisition module drives the vision component to acquire the reference calibration image matrix of the reference calibration board, and transmits the reference calibration image matrix to the error compensation module.

3. The laser heating system according to claim 2, characterized in that, The error compensation module is configured to perform spatial grid error extraction processing: The error compensation module receives the reference calibration image matrix, performs a corner feature extraction operator on the reference calibration image matrix, and obtains the pixel coordinates of multiple grid intersection points. The error compensation module extracts the preset ideal zero-point topological coordinates inside the control device, performs coordinate subtraction operation between the pixel coordinates of the multiple grid intersection points and the ideal zero-point topological coordinates, calculates the reference mechanical error vector as the mechanical error data, and stores the reference mechanical error vector inside the control device.

4. The laser heating system according to claim 3, characterized in that, The contour parsing module is configured to perform image denoising and continuous contour point set extraction processing: After receiving the global pixel matrix, the contour parsing module uses a Gaussian smoothing filter algorithm to remove electromagnetic noise from the global pixel matrix and outputs a noise-reduced pixel matrix. The contour parsing module uses a sub-pixel polynomial fitting operator to identify the gray-level step change region within the denoised pixel matrix and extracts a continuous set of physical contour points.

5. The laser heating system according to claim 4, characterized in that, The contour parsing module is configured to calculate the spatial scaling parameter of the material to be processed based on the continuous set of physical contour points. The contour parsing module performs a geometric center determination operation based on the continuous physical contour point set, calculates the position coordinates of the physical centroid of the material to be processed, and measures the radial distance values ​​from the position coordinates of the physical centroid to multiple pixels in the continuous physical contour point set. The contour analysis module extracts the standard material radial dimension values ​​pre-stored inside the control device, divides the multiple radial distance values ​​by the standard material radial dimension values, and outputs the space scaling ratio parameter of the material to be processed.

6. The laser heating system according to claim 5, characterized in that, The error compensation module is configured to perform cross-dimensional data underlying tensor fusion processing: After the error compensation module extracts the reference mechanical error vector stored inside the control device, it performs Kronecker product tensor operation on the reference mechanical error vector and the space scaling ratio parameter of the material to be processed. The error compensation module outputs the multidimensional error compensation matrix through the Kronecker product tensor operation.

7. The laser heating system according to claim 6, characterized in that, The coordinate optimization module is configured to perform nonlinear coordinate mapping processing: After receiving the multidimensional error compensation matrix, the coordinate optimization module substitutes the processing point coordinates into the multidimensional error compensation matrix to perform a nonlinear topological mapping dot product operation and calculates the preliminary compensation coordinates. The coordinate optimization module extracts a preset horizontal constant, adds the preset horizontal constant to the value of the preliminary compensation coordinates, and outputs the absolute target coordinates.

8. The laser heating system according to claim 7, characterized in that, The coordinate optimization module is configured to perform physical hardware spacing fusion processing: When the coordinate optimization module adds the preset horizontal constant to the value of the initial compensation coordinate, the preset horizontal constant is a fixed horizontal physical distance value between the central axis of the vision component and the central axis of the laser component. The coordinate optimization module performs a vector addition operation between the horizontal physical distance value and the preliminary compensation coordinates, and outputs the absolute target coordinates.

9. The laser heating system according to claim 8, characterized in that, The pulse transmission module is configured to perform anti-resonance trajectory conversion processing: The pulse sending module receives the absolute target coordinates, obtains the current mechanical coordinates of the transmission mechanism, calls the inverse kinematics model to perform a difference calculation between the absolute target coordinates and the current mechanical coordinates, and calculates the physical step distance required for the drive motor to travel. The pulse sending module extracts a preset flexible acceleration / deceleration curve, modulates the motor drive pulse signal according to the physical step distance and the flexible acceleration / deceleration curve, and sends the motor drive pulse signal to the drive motor of the device.

10. The laser heating system according to claim 9, characterized in that, The pulse transmission module is configured to perform optomechanical low-level linkage processing: The pulse sending module sends the motor driving pulse signal to the drive motor, driving the sliding seat to move the laser component to the physical position corresponding to the absolute target coordinates; After the sliding seat reaches the physical position corresponding to the absolute target coordinates, the pulse sending module sends a high-level enable signal to the laser component to control the laser component to emit a laser beam.

Images