Joint calibration method and system in optical communication chip package

Abstract

This invention relates to the field of optical communication technology, addressing the problem of low joint calibration accuracy between a camera and a rotating platform in existing optical communication chip packaging technologies. It provides a joint calibration method and system for optical communication chip packaging. The method includes controlling a rotating platform to rotate at multiple preset rotation angles and controlling a motion platform to move along a first preset direction and a second preset direction, respectively, to acquire angle data, encoder pulse count, and pixel position change data; calculating joint calibration parameters using a calibration module to obtain initial joint calibration parameters; obtaining verification results based on the initial joint calibration parameters; and adjusting the initial joint calibration parameters based on the verification results to obtain target joint calibration parameters that meet preset calibration accuracy conditions. This invention improves the calibration accuracy between the camera and the rotating platform.

Description

Technical Field

[0001] This invention relates to the field of optical communication technology, and in particular to a joint calibration method and system for optical communication chip packaging. Background Technology

[0002] In optical communication chip packaging, industrial cameras are used to acquire images of the packaged devices, and rotating and linear motion platforms are used to perform multi-angle, multi-position inspection operations. The mapping relationship between the camera imaging coordinate system and the equipment motion coordinate system, as well as the rotation center position of the rotating platform, directly affects the positioning accuracy and repeatability of the inspection area. If there is a deviation in the geometric relationship between the camera and the rotating platform, it will cause inconsistent pixel coordinates for the same packaged device at different rotation angles or displacement positions, thus affecting defect location, dimensional measurement, and the stability of subsequent inspection algorithms. Therefore, accurate calibration of the spatial relationship between the camera and the rotating platform is necessary in die bonding equipment.

[0003] In existing die bonding equipment, calibration methods typically employ single-object calibration or step-by-step calibration. For example, only the visual equivalent between the camera and the linear motion platform is calibrated, or only the rotation center of the rotating platform is measured independently. These methods often require separate acquisition of image and motion data, followed by manual or offline calculations to obtain various calibration parameters. Because the camera visual equivalent, coordinate system deflection angle, and rotation center parameters are coupled, separate calibration methods struggle to reflect the combined effects of these parameters, easily leading to accumulated errors under different motion directions or rotation angles. Furthermore, existing calibration methods largely rely on single calculation results, lacking mechanisms for motion verification and parameter correction of the calibration results, resulting in poor calibration accuracy stability during long-term operation or repeated calibration.

[0004] Chinese patent CN106228534B discloses a calibration method for the relationship between a rotating axis and a camera based on constrained global optimization. The calibration system includes a large-scene camera, a camera to be calibrated, a turntable calibration plate, a calibration plate, and a rotating platform. During the 360-degree rotation of the rotating platform's axis, the large-scene camera captures a series of images. The three-dimensional coordinates of the corner points on the calibration plate are extracted based on the rotation and translation matrices. The center of the corner point trajectory is calculated, and PCA is used to fit the straight line containing the center to determine the position of the rotating axis, thus obtaining the angle between the rotating axis and the camera's optical axis. This patent solution focuses on solving the geometric relationship between the rotating axis and the camera's optical axis and relies on an additional large-scene camera and dual calibration plates for collaborative observation. For die-bonding equipment, in addition to the parameters related to the rotating platform, it is also necessary to simultaneously calculate multiple parameters such as the camera's visual equivalent, deflection angle, and rotation center under the condition of coordinated rotation and linear motion of the equipment itself, combined with angle data, encoder pulse count, and pixel position change data. These parameters are then corrected through motion verification to meet the calibration accuracy requirements. Existing calibration schemes for the relationship between the rotating axis and the camera typically do not cover the visual equivalent calibration related to the pulse displacement of the linear motion platform, and also lack a joint verification and parameter closed-loop adjustment mechanism for rotation and displacement in online use scenarios. This results in errors being easily coupled and accumulated under multi-angle and multi-directional displacement, and limits the usability and stability of calibration results.

[0005] Therefore, how to achieve joint high-precision calibration of the camera and the rotating platform in the packaging of optical communication chips is a technical problem that urgently needs to be solved. Summary of the Invention

[0006] In view of this, the present invention provides a joint calibration method and system for optical communication chip packaging, in order to solve the problem of low joint calibration accuracy between camera and rotating platform in existing optical communication chip packaging technology.

[0007] The technical solution adopted in this invention is: In a first aspect, the present invention provides a joint calibration method for optical communication chip packaging, applied to a die-bonding device. The device includes: a composite calibration board, a rotating platform, a motion platform, a motion control module, and a calibration module. The composite calibration board is disposed on the rotating platform or the motion platform, and the motion control module is used to perform coordinated control of the rotating platform and the motion platform. The method includes: The motion control module controls the rotating platform to rotate to multiple preset rotation angles, and controls the motion platform to move along a first preset direction and a second preset direction at each preset rotation angle, respectively, to acquire angle data, encoder pulse count and pixel position change data before and after movement during each rotation and movement process, wherein the first preset direction and the second preset direction are perpendicular to each other; The calibration module performs joint calibration parameter calculation on the angle data, the encoder pulse count, and the pixel position change data to obtain initial joint calibration parameters, wherein the initial joint calibration parameters include camera visual equivalent, target deflection angle between camera coordinate system and device coordinate system, and rotation center parameters of rotating platform; Based on the initial joint calibration parameters, the rotating platform and the motion platform are controlled to perform preset motion verification to obtain the verification results; Based on the verification results, the initial joint calibration parameters are adjusted to obtain target joint calibration parameters that meet the preset calibration accuracy conditions; During the optical communication chip packaging process, the die bonding device is controlled to complete the packaging of the optical communication chip according to the target joint calibration parameters.

[0008] In an optional embodiment, the surface of the composite calibration plate is provided with composite calibration features, which include: alignment line segments for adjusting the alignment relationship of camera coordinate axes, a square array for visual equivalent calibration, a dot array for calibrating the rotation center of the rotating platform, and positioning marks for positioning the composite calibration plate.

[0009] In an optional embodiment, the die bonding device further includes an image acquisition module and an image processing module. The step of controlling the rotating platform to rotate to multiple preset rotation angles via the motion control module, and controlling the motion platform to move along a first preset direction and a second preset direction at each preset rotation angle, and acquiring angle data, encoder pulse count, and pixel position change data before and after movement during each rotation and movement includes: The image acquisition module acquires the initial calibration plate image at the preset position and the target calibration plate image after rotation and movement. The image processing module extracts pixel positions from the initial calibration board image and the target calibration board image to obtain the initial pixel position and the target pixel position. The difference between the initial pixel position and the target pixel position is calculated to obtain the pixel position change data; The motion control module acquires the corresponding angle data and encoder pulse count.

[0010] In an optional embodiment, the step of extracting pixel positions from the initial calibration board image and the target calibration board image using the image processing module to obtain the initial pixel positions and target pixel positions includes: The initial calibration board image and the target calibration board image are respectively input into a pre-trained calibration pattern recognition model to obtain a first calibration pattern and a second calibration pattern. Contour features are extracted from the first calibration pattern and the second calibration pattern respectively to obtain the corresponding contour position information; Based on the contour position information, the geometric center positions of the first calibration pattern and the second calibration pattern are calculated to obtain the first pixel coordinate position and the second pixel coordinate position; The first pixel coordinate position is used as the initial pixel position, and the second pixel coordinate position is used as the target pixel position.

[0011] In an optional embodiment, the step of calculating the initial joint calibration parameters by the calibration module on the angle data, the encoder pulse count, and the pixel position change data includes: Based on the pixel position change data, obtain the pixel displacement vector under different rotation angles and different movement directions; Calculate the target deflection angle between the camera coordinate system and the device coordinate system based on the pixel displacement vectors described above; The camera visual equivalent is calculated by least squares fitting based on the pixel position change data and the encoder pulse count. Based on the coordinates of the trajectory points corresponding to the pixel position change data at different rotation angles, the rotation center parameters of the rotating platform are calculated using a circular trajectory fitting method.

[0012] In an optional embodiment, calculating the target deflection angle between the camera coordinate system and the device coordinate system based on each of the pixel displacement vectors includes: Based on the direction information of the pixel displacement vector and the preset direction-to-deflection angle mapping relationship, each initial deflection angle is calculated; The average value of each initial deflection angle is calculated, and the calculated average deflection angle is taken as the target deflection angle.

[0013] In an optional embodiment, the step of calculating the camera visual equivalent by least squares fitting based on the pixel position change data and the encoder pulse count includes: Based on the pixel position change data, obtain the displacement of each first pixel corresponding to the first preset direction and the displacement of each second pixel corresponding to the second preset direction; Based on the encoder pulse count, obtain the first pulse displacement amount corresponding to the first preset direction and the second pulse displacement amount corresponding to the second preset direction; The least squares method is used to linearly fit each pixel displacement to the corresponding pulse displacement, and the slope of the fitted line is used as the visual equivalent in the corresponding direction.

[0014] In an optional embodiment, controlling the rotating platform and the motion platform to perform a preset motion verification based on the initial joint calibration parameters, and obtaining the verification result, includes: Based on the camera visual equivalent and rotation center parameters, the preset motion amount is converted into the corresponding theoretical pixel displacement amount; Control the rotating platform and the motion platform to perform corresponding rotation and movement operations to acquire verification images; Pixel displacement is calculated on the verification image to obtain the actual pixel displacement amount; The difference between the actual pixel displacement and the theoretical pixel displacement is calculated to obtain the verification result.

[0015] In an optional embodiment, adjusting the initial joint calibration parameters based on the verification results to obtain target joint calibration parameters that meet preset calibration accuracy conditions includes: Obtain the error threshold based on the preset calibration accuracy conditions; The verification result is compared with the error threshold. If the verification result is less than or equal to the error threshold, the initial joint calibration parameter is used as the target joint calibration parameter. If the verification result is greater than the error threshold, the initial joint calibration parameters are corrected based on the verification result to obtain the target joint calibration parameters.

[0016] In a second aspect, the present invention provides a joint calibration system in an optical communication chip package, comprising: at least one processor, at least one memory, and computer program instructions stored in the memory, wherein when the computer program instructions are executed by the processor, the method of the first aspect described above is implemented.

[0017] In summary, the beneficial effects of the present invention are as follows: The joint calibration method and system for optical communication chip packaging provided by this invention collects angle data, encoder pulse count, and pixel position change data under the linkage of a rotating platform and a motion platform using a composite calibration board. Based on this multi-source data, joint calibration parameters are calculated, thereby establishing a holistic correspondence between image coordinates and device motion coordinates simultaneously during the same calibration process. This more realistically reflects the coupling influence between parameters under the combined action of rotational and linear motion, avoiding the error accumulation problem caused by fragmented parameter solving in existing step-by-step calibration methods. Furthermore, instead of relying on a single calculation to obtain initial parameters, this invention further controls the rotating platform and motion platform to perform preset motion verification under the actual motion conditions corresponding to the optical communication chip packaging process, using the initial joint calibration parameters. The verification results are then used to adjust the initial joint calibration parameters in reverse, allowing the calibration results to be corrected and converged based on the actual operating state. Therefore, this invention improves the calibration accuracy between the camera and the rotating platform. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments of the present invention will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort, and these are all within the protection scope of the present invention.

[0019] Figure 1 This is a schematic diagram illustrating the overall operation of the joint calibration method in the optical communication chip packaging in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the composite calibration plate in Embodiment 1 of the present invention; Figure 3 This is a flowchart illustrating the process of calculating the initial joint calibration parameters by means of the calibration module in Embodiment 1 of the present invention, based on the angle data, the encoder pulse count, and the pixel position change data. Figure 4 This is a schematic diagram of the joint calibration system in the optical communication chip packaging in Embodiment 2 of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. In the description of the present invention, it should be understood that the terms "center," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, the element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. Where there is no conflict, embodiments of the present invention and the various features thereof can be combined with each other, all of which are within the scope of protection of the present invention.

[0021] Example 1 Please see Figure 1 Embodiment 1 of this invention discloses a joint calibration method in optical communication chip packaging, applied to a die-bonding device. The device includes: a composite calibration board, a rotating platform, a motion platform, a motion control module, and a calibration module. The composite calibration board is disposed on the rotating platform or the motion platform, and the motion control module is used to perform linkage control on the rotating platform and the motion platform. Specifically, please see Figure 2The composite calibration board is designed as a 100mm×100mm plate structure, integrating three types of calibration patterns: line segments, square arrays, and dot arrays. It can simultaneously meet the requirements of preliminary camera coordinate system calibration, visual equivalent calibration, and rotation center and transmission ratio calibration on a single calibration board. The rotation platform provides controllable angular attitude changes, the motion platform provides controllable linear displacement sampling, and the motion control module handles the timing control of multi-axis linkage and pause sampling, ensuring that the corresponding encoder position or encoder pulse count is obtained at each preset angle and displacement position, triggering camera image acquisition. The encoder pulse count refers to the pulse count output by the motion platform encoder, which can be converted into actual displacement through pulse equivalent. The calibration module is used to uniformly organize and calculate the sub-pixel coordinates of feature points in the image and the displacement and angle data of the motion side. In practical implementation, the composite calibration board can be installed on the rotating platform or the motion platform using a quick-change fixing and positioning structure. This allows the calibration process to cover commonly used workstation postures of the equipment. The motion control module drives the rotating platform to rotate according to a set angle sequence and stop at each angle, while simultaneously driving the motion platform to move along a planned trajectory. This allows the camera's field of view to traverse the feature point array on the calibration board. The calibration module synchronously records the motor position and platform angle corresponding to each frame of image on the host computer, forming a set of multi-pose datasets that can be used for joint solution. By incorporating the calibration reference, rotational motion, and linear motion into the same device and control link, subsequent calculations of the mapping relationship between the camera and the platform can be based on the same physical reference and the same data sampling sequence. This reduces the risk of inconsistencies in the reference and error propagation during step-by-step calibration, improving the repeatability and overall reliability of the calibration results.

[0022] The method includes: The motion control module controls the rotating platform to rotate to multiple preset rotation angles, and controls the motion platform to move along a first preset direction and a second preset direction at each preset rotation angle, respectively, to acquire angle data, encoder pulse count and pixel position change data before and after movement during each rotation and movement process, wherein the first preset direction and the second preset direction are perpendicular to each other; Specifically, by setting the composite calibration plate at a preset position on the rotating or moving platform, the composite calibration plate is aligned with the actual rotating and linear motion structures in the die-bonding equipment under the same calibration reference. This facilitates the subsequent data acquisition in accurately reflecting the spatial correspondence between the camera, the rotating platform, and the moving platform. The high-precision calibration pattern integrated on the composite calibration plate forms extractable calibration feature points in the camera's field of view. The motion control module controls the rotating platform to rotate sequentially to multiple preset rotation angles according to the preset calibration process. At each preset rotation angle, the motion platform is further controlled to move along a first preset direction and a second preset direction, respectively. The first and second preset directions are perpendicular to each other, enabling the camera to acquire imaging changes of the calibration features under different rotational attitudes and different orthogonal displacements. During each rotation and movement, the angle data of the rotating platform, the encoder pulse count of the moving platform, and the pixel position change data of the calibration features in the image before and after movement are simultaneously acquired, thereby forming joint calibration data that simultaneously includes angle attitude, linear displacement, and pixel response relationships. By using the above-mentioned linkage acquisition method, we can obtain the constraints of the feature point position changes on the rotation center parameters and coordinate system deflection relationship under different rotation angles, and we can also obtain the correspondence between pixel displacement and actual displacement when moving along two mutually perpendicular directions.

[0023] The calibration module performs joint calibration parameter calculation on the angle data, the encoder pulse count, and the pixel position change data to obtain initial joint calibration parameters, wherein the initial joint calibration parameters include camera visual equivalent, target deflection angle between camera coordinate system and device coordinate system, and rotation center parameters of rotating platform; Specifically, the calibration module performs unified modeling and joint calculation on the angle data, encoder pulse count, and pixel position change data acquired above. The camera visual equivalent describes the pixel change ratio corresponding to a unit mechanical displacement in the image; it is the scale factor between pixel coordinates and device displacement. The target deflection angle between the camera coordinate system and the device coordinate system reflects the angular deviation between the camera's mounting posture and the device's motion axis. The rotation center parameter of the rotating platform is used to characterize the position of the platform's rotation axis in the camera coordinate system. During the calculation process, the calibration module correlates the pixel change trajectories acquired at different rotation angles with the corresponding angle data to fit the rotation center position. Simultaneously, it performs corresponding calculations on the pixel change and pulse displacement amounts under different linear movement directions to obtain the visual equivalent, and combines this with the relative relationship between the two displacement vectors to determine the coordinate system deflection angle. By simultaneously solving these three types of parameters within the same computational framework, the originally interrelated scale, angle, and center position relationships can be mutually constrained, avoiding parameter coupling errors generated during separate calibration and improving the overall consistency of the initial calibration results from the source.

[0024] Based on the initial joint calibration parameters, the rotating platform and the motion platform are controlled to perform preset motion verification to obtain the verification results; Specifically, during the optical communication chip packaging process, using the obtained initial joint calibration parameters, a rotating platform and a motion platform are driven to perform combined motion according to a preset rotation angle and displacement. The corresponding theoretical pixel positions are calculated based on the initial joint calibration parameters. Simultaneously, the pixel positions of the calibration features in the actual image are acquired through image acquisition. The actual pixel positions are compared and analyzed with the theoretical pixel positions to obtain verification results reflecting the calibration error. This verification process directly assesses the ability of the initial joint calibration parameters to describe the actual motion behavior of the device. When there is a deviation in the rotation center parameter, the pixel trajectory of the same physical feature point at different rotation angles is difficult to maintain consistency. When there is an error in the visual equivalent or deflection angle, a systematic deviation will occur between the theoretical pixel displacement and the actual pixel displacement. By quantifying and analyzing these deviations, the error of the calibration parameters is transformed into observable verification results.

[0025] Based on the verification results, the initial joint calibration parameters are adjusted to obtain target joint calibration parameters that meet the preset calibration accuracy conditions; Specifically, the initial joint calibration parameters are adjusted based on the verification results until the preset calibration accuracy conditions are met. The verification results reflect the pixel error level of the current parameters under real motion conditions. When this error exceeds the allowable range, it indicates that the existing parameters still cannot accurately describe the mapping relationship between the camera and the rotating platform. At this time, the calibration module corrects the visual equivalent, deflection angle, and rotation center parameters based on the verification error, so that the model output gradually approximates the real motion behavior. This adjustment process can form a closed loop through multiple verifications and corrections, allowing the parameters to continuously converge towards a smaller error until the verification results meet the preset accuracy requirements. By introducing this motion verification-based parameter adjustment mechanism, not only can residual errors that may exist in the initial calculation stage be eliminated, but the effects of mechanical assembly errors or installation deviations can also be compensated. This ensures that the final target joint calibration parameters have high consistency under different angles and directions of movement, thereby significantly improving the overall accuracy and long-term reliability of the mapping relationship between the camera and the rotating platform.

[0026] During the optical communication chip packaging process, the die bonding device is controlled to complete the packaging of the optical communication chip according to the target joint calibration parameters.

[0027] Specifically, in the chip picking and pre-alignment stage, the die bonding device converts the pixel coordinates of the chip image acquired by the camera into physical coordinates of the device motion coordinate system based on the camera visual equivalent and the target deflection angle between the camera coordinate system and the device coordinate system in the target joint calibration parameters. This enables the device to accurately control the nozzle to move above the chip and complete the picking, while simultaneously calculating the initial position and angle deviation of the chip through visual calculation.

[0028] Subsequently, in the visual positioning and angle rotation correction stage, the die bonding equipment controls the rotating platform to rotate the chip (or substrate) to the target angle. Since the calibration parameters already include precise rotation center parameters, the system can accurately predict and compensate for the positional offset of chip feature points caused by platform rotation. By combining the movement of the motion platform in the first and second preset directions, key features (such as waveguide coupling areas or electrode pads) on the chip or substrate can be quickly and repeatedly located using a camera at different rotation angles, and the required position and angle compensation amounts can be calculated.

[0029] Finally, in the final mounting stage, the motion control module drives the motion platform to precisely move the chip above the corresponding position on the substrate based on the above visual positioning results and compensation amount. Through coordinate transformation and motion compensation, the mounting head is finally controlled to accurately place the chip on the substrate with preset bonding pressure and posture, completing optical alignment and mechanical fixation, thereby meeting the stringent requirements of high coupling efficiency and high reliability for optical communication chip packaging.

[0030] In an optional embodiment, controlling the die bonding device to complete the packaging of the optical communication chip according to the target joint calibration parameters during the optical communication chip packaging process includes: Obtain the reference ambient temperature data corresponding to the target joint calibration parameters and the real-time ambient temperature data during the optical communication chip packaging process; The temperature offset is determined based on the difference between the real-time ambient temperature data and the reference ambient temperature data; Based on the temperature offset and the preset temperature compensation relationship, determine the parameter compensation amount corresponding to the target joint calibration parameter; Based on the parameter compensation amount, the target joint calibration parameters are compensated to obtain the compensated joint calibration parameters; Based on the compensated joint calibration parameters, the die bonding device is controlled to complete the packaging of the optical communication chip.

[0031] Specifically, when obtaining the reference ambient temperature data corresponding to the target joint calibration parameters, the equipment temperature state when the joint calibration is completed and the target joint calibration parameters are obtained is used as the reference. The reference ambient temperature data may include the temperature values ​​of the camera mounting area, the rotating platform mounting area, the motion platform mounting area, the composite calibration plate mounting area, or the internal working space of the die bonding equipment. During the packaging process of optical communication chips, a temperature detection module continuously or at preset intervals collects real-time ambient temperature data to reflect temperature changes caused by factors such as heat generation from the light source, motor operation, platform friction, and heat conduction from the mounting structure after prolonged operation. The difference between the real-time ambient temperature data and the reference ambient temperature data is used to form a temperature offset. This temperature offset is not simply environmental data, but rather a basis for determining whether thermal drift has occurred in the current packaging state relative to the calibration state.

[0032] Since the target joint calibration parameters include the camera visual equivalent, the target deflection angle between the camera coordinate system and the equipment coordinate system, and the rotation center parameters of the rotating platform, the effects of temperature changes on different parameters are not entirely the same. The preset temperature compensation relationship can be established based on the parameter change patterns obtained from repeated calibration under different temperature conditions during the equipment debugging phase, or it can be set based on the thermal expansion characteristics of the composite calibration plate, camera mounting part, rotating platform, and motion platform. After obtaining the temperature offset, the parameter compensation amount corresponding to each parameter in the target joint calibration parameters is determined according to the temperature compensation relationship. The parameter compensation amount can be expressed as a visual equivalent correction value, a deflection angle correction value, or a rotation center coordinate correction value.

[0033] The parameter compensation is then superimposed or mapped to the original target joint calibration parameters to obtain the compensated joint calibration parameters adapted to the current temperature state. During the packaging of the optical communication chip, the camera recognition position, platform movement position, and rotation correction position are calculated using these compensated joint calibration parameters, so that the die bonding equipment can maintain the coordinate consistency between camera positioning, rotation positioning, and linear motion even after the actual operating temperature changes.

[0034] In an optional embodiment, determining the parameter compensation amount corresponding to the target joint calibration parameter based on the temperature offset and a preset temperature compensation relationship includes: Based on the temperature offset, the temperature offset range corresponding to the current optical communication chip packaging process is determined, wherein the temperature offset range includes at least a low temperature offset range, a normal offset range, and a high temperature offset range; According to the temperature offset range, the corresponding parameter compensation coefficient is called from the preset temperature compensation relationship. The parameter compensation coefficient includes a first compensation coefficient corresponding to the camera visual equivalent, a second compensation coefficient corresponding to the target deflection angle, and a third compensation coefficient corresponding to the rotation center parameter. Based on the temperature offset and the first compensation coefficient, determine the visual equivalent compensation amount corresponding to the camera visual equivalent; Based on the temperature offset and the second compensation coefficient, determine the deflection angle compensation amount corresponding to the target deflection angle; Based on the temperature offset and the third compensation coefficient, determine the rotation center compensation amount corresponding to the rotation center parameters; At least one of the visual equivalent compensation amount, the deflection angle compensation amount, and the rotation center compensation amount is used as the parameter compensation amount corresponding to the target joint calibration parameter.

[0035] Specifically, after obtaining the temperature offset, the impact of temperature changes on different joint calibration parameters is further processed separately. The temperature offset refers to the change in real-time ambient temperature relative to the reference ambient temperature during joint calibration. The low-temperature offset range, normal offset range, and high-temperature offset range can be pre-divided according to the actual operating temperature range of the die bonding equipment. For example, small changes close to the calibration temperature are classified into the normal offset range, cases significantly below the reference temperature are classified into the low-temperature offset range, and cases where the equipment experiences a significant temperature rise due to heat generated by the light source, motor, rotating platform, and moving platform after long-term operation are classified into the high-temperature offset range.

[0036] Since the camera visual equivalent mainly reflects the proportional relationship between pixel displacement and actual device displacement, the target deflection angle reflects the angular correspondence between the camera coordinate system and the device coordinate system, and the rotation center parameter reflects the position of the physical rotation center of the rotating platform in the calibration coordinate system, the different mechanisms by which these parameters are affected by temperature changes are different. Therefore, a single compensation coefficient is not used for unified correction. In actual implementation, during the device debugging or factory calibration phase, joint calibration is repeatedly performed under different temperature conditions. The changes in camera visual equivalent, target deflection angle, and rotation center parameter with temperature offset are statistically analyzed, and preset temperature compensation relationships are established for different temperature offset intervals. When the current temperature offset is detected during the packaging process, the temperature offset interval to which it belongs is first determined. Then, the first compensation coefficient, the second compensation coefficient, and the third compensation coefficient are called from the compensation relationship corresponding to that interval and calculated with the temperature offset to obtain the visual equivalent compensation amount, deflection angle compensation amount, and rotation center compensation amount.

[0037] Subsequently, one or more compensation values ​​can be selected as the parameter compensation values ​​for the target joint calibration parameters according to actual needs, and used to correct subsequent camera positioning, platform movement conversion, and rotation angle correction processes. Through this interval-based and parameter-based compensation method, it is possible to avoid over-compensation or under-compensation of certain parameters due to overly coarse temperature compensation, making the joint calibration results more suitable for the actual motion relationship of the die bonding equipment under different thermal states.

[0038] In an optional embodiment, the surface of the composite calibration plate is provided with composite calibration features, which include: alignment line segments for adjusting the alignment relationship of camera coordinate axes, a square array for visual equivalent calibration, a dot array for calibrating the rotation center of the rotating platform, and positioning marks for positioning the composite calibration plate.

[0039] Specifically, by integrating multiple calibration features with different functions on the same composite calibration plate surface, the calibration process can simultaneously serve the solution of multiple types of parameters under a unified physical reference. Alignment segments refer to high-contrast straight line patterns arranged in a known direction on the calibration plate. They form stable straight line edges in the image and can be used to calculate the angular relationship between the camera imaging coordinate axis and the device motion coordinate axis, thereby identifying and correcting installation deviations of the camera coordinate axis. Square arrays consist of regularly distributed square or intersection structures. The actual distance between each feature point is known. When the motion platform moves in a straight line, by comparing the pixel displacement of the feature points in the image with the corresponding encoder pulse number, the proportional relationship between the pixel change and the physical displacement can be established, thus obtaining the visual equivalent parameters. Dot arrays are usually composed of several eccentrically distributed circular features. When the rotating platform rotates, the position of the same dot in the image will change along an approximately circular trajectory. By fitting this trajectory, the actual rotation center position of the rotating platform can be deduced, so that the rotation axis parameters can be accurately extracted from the imaging data. Positioning marks serve as the installation reference for the composite calibration plate on the rotating or motion platform. They are used to cooperate with the positioning elements on the fixed structure to ensure that each installation is in a consistent reference position, avoiding new system deviations introduced by repeated clamping errors. By integrating the aforementioned alignment segments, square arrays, dot arrays, and positioning marks onto the same calibration plate, the solutions for different types of parameters share the same imaging conditions and the same mechanical reference. This not only reduces the cumulative error caused by switching between multiple calibration plates, but also enables the three key parameters—axis deflection, scale ratio, and rotation center—to mutually constrain each other within the same data system, thereby improving the stability and overall accuracy of the joint calibration results.

[0040] In an optional embodiment, the die bonding device further includes an image acquisition module and an image processing module. The step of controlling the rotating platform to rotate to multiple preset rotation angles via the motion control module, and controlling the motion platform to move along a first preset direction and a second preset direction at each preset rotation angle, and acquiring angle data, encoder pulse count, and pixel position change data before and after movement during each rotation and movement includes: The image acquisition module acquires the initial calibration plate image at the preset position and the target calibration plate image after rotation and movement. Specifically, the image acquisition module acquires an initial image of the composite calibration board when it is in a preset installation position, and acquires corresponding images of the target calibration board after the rotating platform rotates and the motion platform moves. This allows the same calibration board to form paired image data under different postures and displacements. The image acquisition module typically consists of an industrial camera and its triggering circuit, which establishes a synchronous relationship with the motion control module. Image acquisition is triggered after the rotating platform completes an angle positioning or the motion platform completes a displacement, ensuring that each frame corresponds to a specific angle and displacement state. By acquiring the initial calibration board image as a reference, and then acquiring the target calibration board images after rotation and movement, the imaging state before and after equipment movement can be completely recorded. This provides a stable comparison for subsequent analysis of pixel changes, making the calibration data unaffected by random sampling and human operation, and facilitating the formation of a repeatable and comparable data foundation.

[0041] The image processing module extracts pixel positions from the initial calibration board image and the target calibration board image to obtain the initial pixel position and the target pixel position. Specifically, the image processing module extracts pixel positions from both the initial and target calibration board images, locating the coordinates of feature points corresponding to the composite calibration features. Pixel position extraction involves determining the center position of features such as the intersection of aligned line segments, corner points of square arrays, or centers of dot arrays in the image through grayscale analysis, edge detection, or contour fitting, and then converting these positions into pixel coordinates for output. To improve positioning accuracy, the image processing module can perform magnification and sub-pixel interpolation calculations on local areas, allowing feature point positions to be expressed in floating-point form rather than integer pixels, thus more accurately reflecting the actual imaging position. By performing the same feature extraction process on both the initial and target images, a unified judgment standard is ensured, ensuring that subsequent pixel differences accurately reflect imaging changes caused by device movement, rather than errors due to inconsistent algorithms.

[0042] The difference between the initial pixel position and the target pixel position is calculated to obtain the pixel position change data; Specifically, the difference between the initial pixel position and the target pixel position is calculated to obtain pixel position change data. This data directly describes the amount and direction of movement of the calibration feature in the image under a certain rotation angle or displacement condition. The pixel position change data includes both horizontal and vertical components, and also implicitly reflects the combined effect of the rotation of the rotating platform and the movement of the motion platform on the imaging plane. In implementation, the coordinates of the same feature point in the two images can be mapped one-to-one, and the coordinate difference can be calculated to form a set of pixel displacement vectors. These vectors can be used to analyze the pixel change patterns corresponding to different directions of movement, and can also serve as the basic input for subsequent joint calibration parameter calculations. By introducing this intermediate quantity of pixel position change data, a direct link is established between the changes in the image domain and the angle data and encoder pulse count on the device side, allowing the relationship between physical motion and imaging response to be quantitatively expressed. This provides reliable data support for subsequent solutions to visual equivalent, deflection angle, and rotation center parameters.

[0043] The motion control module acquires the corresponding angle data and encoder pulse count.

[0044] Specifically, the rotation angle data originates from the angle feedback unit of the rotating platform, reflecting the actual rotation angle position of the calibration plate under different postures. The encoder pulse count is output by the drive motor or encoder of the motion platform, characterizing the actual movement of the motion platform in a first or second preset direction. In practice, the motion control module simultaneously records the motor pulse count or encoder reading while controlling rotation or linear movement, establishing a one-to-one correspondence between this data and the currently acquired calibration image, ensuring that each frame is associated with a unique angle and displacement state. In this way, pixel position change data is no longer an isolated image result but is bound to a specific mechanical motion quantity, allowing for a direct establishment of the mapping relationship between pixel changes and physical motion in subsequent calculations, making the calibration process traceable and consistent.

[0045] In an optional embodiment, the step of extracting pixel positions from the initial calibration board image and the target calibration board image using the image processing module to obtain the initial pixel positions and target pixel positions includes: The initial calibration board image and the target calibration board image are respectively input into a pre-trained calibration pattern recognition model to obtain a first calibration pattern and a second calibration pattern. Specifically, the initial calibration board image and the target calibration board image are input into the pre-trained calibration pattern recognition model to quickly and stably locate the effective region containing the composite calibration features in the entire image, avoiding the misinterpretation of background textures, reflections, noise, or device edges as calibration patterns in the calculation. The pre-trained calibration pattern recognition model is a recognizer built based on the appearance features of the calibration board pattern. It is trained on a convolutional neural network, and the training data covers calibration board images under different lighting and poses. It supports online updates to adapt to new environments. During the training phase, it learns the performance of typical shapes such as square arrays, line segments, or dots under different lighting and imaging conditions, thereby outputting the region results of the first and second calibration patterns in the actual acquired initial image and the image after motion, respectively. In implementation, the input image is first normalized in grayscale, denoised, and scaled. Then, the model outputs the bounding box or segmentation mask of the pattern region. Subsequent processing is only performed within this region, which can significantly reduce the false detection rate and false negative rate, make pixel position extraction more stable, and thus improve the data consistency of subsequent pixel difference and joint calibration calculation.

[0046] Contour features are extracted from the first calibration pattern and the second calibration pattern respectively to obtain the corresponding contour position information; Specifically, contour feature extraction is performed on the first and second calibration patterns respectively. The aim is to convert the pattern region from a grayscale image into a boundary description that can be used for geometric calculations, thereby providing a more reliable geometric basis for high-precision positioning. Contour feature extraction typically revolves around high-contrast edges. Threshold segmentation can be used to obtain a binary image, then edge operators can be used to extract the set of boundary points, and closed curves can be obtained through contour tracking. For dotted patterns, further edge point filtering and ellipse fitting can be performed to remove discrete points caused by burrs and gaps. For square or line segment patterns, straight line segments can be extracted and corner point constraints can be applied to make the contour information have stronger structural consistency. By obtaining the corresponding contour position information, the pattern position no longer depends on the brightness extreme value of a single pixel, but is determined by a set of boundary points as a whole. This makes it less sensitive to noise, local stains, and slight blurring, and can improve the repeatability of feature extraction results between different images.

[0047] Based on the contour position information, the geometric center positions of the first calibration pattern and the second calibration pattern are calculated to obtain the first pixel coordinate position and the second pixel coordinate position; Specifically, calculating the geometric center positions of the first and second calibration patterns based on the contour position information involves further transforming the contour description into pixel coordinate representations that can be directly used for displacement calculation. The position of the geometric center can be obtained by taking the centroid of the contour point set and considering area weights, or, in the case of circular dots, by directly using the fitted center of a circle or ellipse as the center coordinates. In the case of square arrays or regular patterns, robustness can be enhanced by using the intersection point or average center of multiple corner points. To improve accuracy, the center coordinates are often output in sub-pixel form, that is, floating-point coordinates are obtained by interpolation or fitting based on an integer pixel grid, so that the center position can reflect minute displacement changes. The first and second pixel coordinate positions obtained in this way can stably represent the imaging position of the same physical calibration pattern in two states before and after motion, providing high-resolution input for subsequent pixel position change data, thereby improving the upper limit of the accuracy of the fitting of visual equivalent, deflection angle, and rotation center.

[0048] The first pixel coordinate position is used as the initial pixel position, and the second pixel coordinate position is used as the target pixel position.

[0049] Specifically, using the first pixel coordinate position as the initial pixel position and the second pixel coordinate position as the target pixel position is to clearly define the one-to-one correspondence before and after motion in the data structure, providing a unified reference for subsequent difference calculations and parameter solutions. The initial pixel position serves as the benchmark, while the target pixel position corresponds to the response results after rotation and movement. Storing these two in pairs allows for direct calculation of the pixel displacement vector, which is then correlated with the angle data and encoder pulse count at the same time. In implementation, each identified calibration pattern or feature point is typically assigned a unique identifier to ensure that the initial frame and target frame match the same physical feature, avoiding abnormal displacement caused by mismatches. For scenarios with multiple feature points, a feature point set can be formed and its consistency verified according to the pattern's topological relationship. Through this clear benchmark and target definition, the directionality and magnitude of pixel changes can be stably extracted, making subsequent joint calibration calculations easier to converge and verify, ultimately improving the reliability and repeatability of the entire calibration process under different angles and movement directions.

[0050] In an alternative embodiment, please refer to Figure 3 The initial joint calibration parameters obtained by calculating the joint calibration parameters of the angle data, the encoder pulse count, and the pixel position change data through the calibration module include: Based on the pixel position change data, obtain the pixel displacement vector under different rotation angles and different movement directions; Specifically, obtaining pixel displacement vectors under different rotation angles and movement directions based on pixel position change data represents the positional changes of feature points in the image before and after motion as a two-dimensional quantity with direction and magnitude. This facilitates the subsequent description of the combined influence of linear displacement and rotational attitude on imaging using the same set of mathematical relationships. Pixel displacement vectors are typically represented as two-dimensional coordinate differences, containing horizontal and vertical components. They reflect the dominant displacement direction in the image when the motion platform moves along a first or second preset direction, while also incorporating perturbations caused by camera mounting angle, lens distortion residuals, and minor attitude changes of the rotating platform. In implementation, an index is created for each feature point, and the corresponding initial pixel coordinates are paired with the target pixel coordinates one by one to calculate the set of pixel displacement vectors. The vectors are then grouped and statistically analyzed according to rotation angle and movement direction. For example, two sets of orthogonal displacement vectors are formed at the same rotation angle for subsequent calculation of deflection angle and visual equivalent, while cross-angle data are retained for rotation center fitting. By first forming an intermediate expression of pixel displacement vectors, the original discrete coordinate data can be organized into structured samples, reducing the repeated dependence on the original coordinate matching in subsequent calculations, and making the parameter solution more stable and traceable.

[0051] Calculate the target deflection angle between the camera coordinate system and the device coordinate system based on the pixel displacement vectors described above; Specifically, the target deflection angle between the camera coordinate system and the device coordinate system is calculated based on the displacement vectors of each pixel. The core principle is to use the known motion relationship of the motion platform in two mutually perpendicular directions to deduce the overall rotational deviation of the camera's imaging coordinate axis relative to the device's motion axis. The target deflection angle reflects the angle between the camera's mounting posture or imaging coordinate definition and the device's X and Y motion directions. Without compensation, the device's purely linear motion along the first preset direction will appear as an oblique displacement with both horizontal and vertical components in the image, and the same cross component will appear in the second preset direction. In implementation, pixel displacement vectors corresponding to the first and second preset directions can be selected respectively, and their principal direction angles can be calculated. For example, by calculating the statistical center value of the direction angles of multiple sets of vectors, or by performing direction consistency fitting on the vector set, the influence of noise points and local errors can be offset. Combined with the orthogonal constraints between the two sets of principal directions, a more reliable deflection angle estimate is obtained. The target deflection angle obtained in this way can be used to subsequently decompose the pixel displacements onto the device coordinate axis directions, so that the visual equivalent calculation is no longer affected by the inter-axis coupling caused by the camera's tilted mounting, thereby improving the direction consistency of the mapping relationship and the reliability of subsequent motion reproduction.

[0052] The camera visual equivalent is calculated by least squares fitting based on the pixel position change data and the encoder pulse count. Specifically, the camera visual equivalent is calculated by fitting pixel position change data and encoder pulse count using the least squares method. This involves using a statistically optimal proportional relationship to correlate physical displacement with pixel displacement, based on a large number of samples. The camera visual equivalent can be understood as the pixel change ratio corresponding to a unit physical displacement. In engineering, it is usually necessary to obtain visual equivalent parameters in two orthogonal directions to match the displacement control of the motion platform in both directions. In implementation, the number of encoder pulses for each movement is converted into the corresponding actual displacement, or the pulse count is used directly as the independent variable, and the pixel displacement component after deflection angle compensation is used as the dependent variable. An approximate linear relationship is established, and the slope is obtained by least squares fitting multiple sets of samples as the visual equivalent. The advantage of least squares fitting is that it can average random errors using multiple samples, avoiding proportional coefficient drift caused by a single measurement, even in the presence of measurement noise, image extraction errors, and small motion backlash errors. Once the visual equivalent is determined, any given pulse displacement or target displacement can be quickly converted into the theoretical pixel displacement that should appear in the image domain, which can be used for motion verification, path planning or defect localization and correction, significantly improving the scale accuracy and stability of pixel-displacement mapping.

[0053] Based on the coordinates of the trajectory points corresponding to the pixel position change data at different rotation angles, the rotation center parameters of the rotating platform are calculated using a circular trajectory fitting method.

[0054] Specifically, a least-squares circle fitting method is used to fit the pixel coordinates of the same feature point at multiple rotation angles, obtaining the rotation center coordinates as the rotation center parameter of the rotating platform. The rotation center parameter describes the center coordinates of the physical axis of the rotating platform projected onto the camera coordinate system. If this parameter has a large deviation, the pixel trajectory of the same feature point at different angles cannot be consistently interpreted, directly affecting rotation compensation, angle positioning, and coordinate alignment during multi-angle detection. In implementation, a dot array or other suitable feature points for rotation analysis are selected, and their pixel coordinates are extracted at multiple preset rotation angles. The pixel position of the same physical point changes with the angle to form a trajectory point set. Ideally, this trajectory approximates an arc or circle centered at the rotation center. By performing circle fitting or least-squares circle center estimation on the trajectory point set, the position of the rotation center in the pixel coordinate system can be obtained. If multiple point trajectories are used simultaneously, consistency constraints or joint fitting can be applied to multiple fitting results to reduce the impact of single-point extraction errors. By fitting based on cross-angle distribution relationships, the rotation center does not depend on a single angle or a single measurement, but is determined by the statistical properties of the complete angle sequence. This can significantly reduce the sensitivity of the results to angle repeatability errors and feature extraction noise, and enable rotation compensation and coordinate alignment to maintain higher consistency and reliability across the entire angle range.

[0055] In an optional embodiment, calculating the target deflection angle between the camera coordinate system and the device coordinate system based on each of the pixel displacement vectors includes: Based on the direction information of the pixel displacement vector and the preset direction-to-deflection angle mapping relationship, each initial deflection angle is calculated; Specifically, when calculating the initial deflection angles based on the direction information of the pixel displacement vector and the preset direction-to-deflection angle mapping relationship, the direction information of the pixel displacement vector can be obtained from the ratio of its components in the image coordinate system, usually expressed as the direction angle of the vector relative to the horizontal or vertical axis of the image. Since the movement directions of the motion platform in the first and second preset directions are known and perpendicular to each other in the device coordinate system, the preset direction-to-deflection angle mapping relationship is used to establish a correspondence between the known movement directions on the device side and the vector directions observed on the image side. For example, the angle between the main direction of the displacement vector corresponding to the first preset direction and the camera coordinate axis is used as a candidate value for the deflection angle, and the angle between the main direction of the displacement vector corresponding to the second preset direction and another coordinate axis is used as another candidate value. The direction angle is then unified by combining the orthogonal constraint of the two directions. In the specific implementation, the direction angle of the pixel displacement vector obtained for each movement can be calculated first, and then the corresponding mapping rule can be called according to its movement direction category to complete the angle conversion. At the same time, quadrant correction and sign consistency processing are performed on the direction angle to avoid angle jumps caused by vector direction reversal or differences in image coordinate system definition. The initial deflection angles obtained in this way can reflect the local estimates of the camera coordinate system relative to the device coordinate system under different sampling conditions, providing a basis for subsequent statistical fusion.

[0056] The average value of each initial deflection angle is calculated, and the calculated average deflection angle is taken as the target deflection angle.

[0057] Specifically, calculating the mean of each initial deflection angle and using the average deflection angle as the target deflection angle is a method that utilizes multi-sample statistical fusion to improve the stability of deflection angle estimation. Since the direction of the pixel displacement vector is affected by factors such as feature point extraction noise, lens distortion residue, motion platform backlash, and minor vibrations, the initial deflection angles obtained under different sampling conditions for the same true deflection angle will exhibit discrete fluctuations. Directly using a single result can easily introduce random errors. By calculating the mean of the initial deflection angles obtained from multiple acquisitions, the impact of random errors can be weakened overall, making the result closer to the true installation deflection angle. In implementation, the initial deflection angles are usually first screened for consistency, for example, by removing outliers that significantly deviate from the main distribution or by periodically expanding the angles to avoid mean shifts caused by angles crossing boundaries. Then, the average deflection angle is calculated as the target deflection angle. The obtained target deflection angle can be used to subsequently rotate the pixel displacement components to the device coordinate axis, so that visual equivalent fitting and rotation center fitting are performed under a more consistent coordinate reference, thereby reducing inter-axis coupling errors and improving the overall reliability of joint calibration parameters under different motion directions and angles.

[0058] In an optional embodiment, the step of calculating the camera visual equivalent by least squares fitting based on the pixel position change data and the encoder pulse count includes: Based on the pixel position change data, obtain the displacement of each first pixel corresponding to the first preset direction and the displacement of each second pixel corresponding to the second preset direction; Specifically, when obtaining the first pixel displacement corresponding to the first preset direction and the second pixel displacement corresponding to the second preset direction based on pixel position change data, the pixel displacement is the displacement amplitude of a specified feature in the image from its initial position to its target position, usually represented by the component of the pixel displacement vector in the specified direction or its projection. Since the motion platform moves linearly in two mutually perpendicular directions, the pixel position change data needs to be categorized according to the direction of movement to ensure that the first pixel displacement comes from the image change corresponding to the movement of the motion platform along the first preset direction, and the second pixel displacement comes from the image change corresponding to the movement along the second preset direction. In implementation, the direction of each set of data can be determined by the motion command of the motion control module or the recorded displacement axis identifier, and multiple movement data in the same direction can be processed uniformly, such as averaging the pixel changes of the same feature point in multiple displacements, or performing combined statistics on the pixel changes of multiple feature points to reduce the impact of single-point extraction errors. The two sets of pixel displacements obtained in this way have clear physical meaning in direction, providing clean data input for subsequently establishing proportional relationships on two orthogonal axes.

[0059] Based on the encoder pulse count, obtain the first pulse displacement amount corresponding to the first preset direction and the second pulse displacement amount corresponding to the second preset direction; Specifically, when obtaining the first pulse displacement corresponding to the first preset direction and the second pulse displacement corresponding to the second preset direction based on the encoder pulse count, the pulse displacement is a discrete measurement result of the displacement by the motion platform drive system, derived from the motor pulse count or encoder reading, and can objectively reflect the actual displacement of the platform in each movement. Consistent with the grouping of pixel displacements, this step also requires splitting the encoder pulse count according to the movement direction, ensuring a one-to-one correspondence between the first pulse displacement and the first pixel displacement, and between the second pulse displacement and the second pixel displacement, to avoid cross-axis mismatches that could distort the fitting relationship. In practice, the initial pulse count or encoder position is typically recorded before each movement begins, and the termination count is recorded after the movement ends. The difference is then used to obtain the pulse displacement for that movement. Simultaneously, the platform's motion mode can be combined to perform a consistency check on the data. For example, the pulse displacement of multiple movements in the same direction should be consistent with the set displacement step size. Abnormal fluctuations can be used as indicators of mechanical return or slide jamming and should be removed before fitting. By obtaining a stable set of pulse displacements, the subsequent calculation of the visual equivalent has a reliable physical scale benchmark.

[0060] The least squares method is used to linearly fit each pixel displacement to the corresponding pulse displacement, and the slope of the fitted line is used as the visual equivalent in the corresponding direction.

[0061] Specifically, when linearly fitting pixel displacement to the corresponding pulse displacement using the least squares method and taking the slope as the visual equivalent in the corresponding direction, the visual equivalent is essentially the proportionality coefficient between pixel change and mechanical displacement. This can be understood as the pixel change corresponding to each unit pulse displacement in the image, or further converted to how many pixels per millimeter. Due to noise in image feature extraction and small errors in platform motion, the proportionality coefficient obtained from a single measurement is prone to fluctuation. The least squares method can utilize multiple sets of data samples to find the best-fitting straight line that minimizes the sum of squared residuals, thus obtaining a more stable proportional estimate. In practice, linear models are established for the data pairs in the first and second directions, with pixel displacement as the dependent variable and pulse displacement as the independent variable. The slope of the fitted line is used as visual equivalent parameters such as Kx and Ky. If necessary, the intercept can also be fitted simultaneously to compensate for zero-point deviation or the intercept can be fixed at zero to conform to the assumption of a pure proportional relationship. The obtained visual equivalents in two directions can be used to convert any pulse displacement into theoretical pixel displacement, or to inversely calculate pixel displacement into platform displacement, thereby achieving scale unification between the pixel domain and the device motion domain. This allows subsequent rotation compensation, position reproduction, and motion verification to be based on a more accurate scale mapping, thus improving the overall accuracy and stability of the joint calibration.

[0062] In an optional embodiment, controlling the rotating platform and the motion platform to perform a preset motion verification based on the initial joint calibration parameters, and obtaining the verification result, includes: Based on the camera visual equivalent and rotation center parameters, the preset motion amount is converted into the corresponding theoretical pixel displacement amount; Specifically, when converting the preset motion amount into the theoretical pixel displacement amount based on the camera visual equivalent and rotation center parameters, the preset motion amount can be the displacement step size of the motion platform in the first or second preset direction, or a preset combination of rotation angles of the rotation platform, used to manually set a set of repeatable standard actions during the verification phase. The camera visual equivalent is a proportionality coefficient between the pixel scale and the mechanical scale, which can convert pulse displacement or actual displacement into pixel displacement components in the image coordinate system; the rotation center parameters are used to describe the pixel displacement law caused by rotation. When the calibration plate or the observed feature point changes angle relative to the rotation center, its pixel position will change along an approximately circular trajectory. The theoretical pixel displacement needs to be calculated by combining the initial pixel radius from the rotation center to the feature point and the rotation angle. In implementation, the feature point or feature point set used for verification can be selected first, and the preset linear displacement and rotation angle can be converted into the expected displacement vector in the pixel domain based on the initial joint calibration parameters. Then, the linear displacement and rotation displacement are superimposed in the same image coordinate system to obtain the theoretical pixel displacement amount. Through this conversion, the expected response of the verification action in the pixel domain is clearly quantified, so that subsequent comparisons do not rely on empirical judgment but have a calculable benchmark.

[0063] Control the rotating platform and the motion platform to perform corresponding rotation and movement operations to acquire verification images; Specifically, when controlling the rotating platform and the motion platform to perform corresponding rotation and movement operations and acquire verification images, the core is to reproduce a set of standard actions under the same motion link as the actual detection, in order to verify the interpretability of the initial joint calibration parameters for real motion. The motion control module drives the rotating platform to rotate according to preset angle commands and stop at the target angle, while simultaneously driving the motion platform to move to the target position along the specified direction according to preset displacement commands. Throughout the process, it is necessary to ensure that the image acquisition time is consistent with the motion state. Generally, the camera is triggered to acquire verification images after the platform is in place and vibration decay has been completed. To improve the credibility of the verification, multiple sets of verification actions can be repeatedly executed at different angles, directions, and displacements, and corresponding verification images can be acquired for each set of actions. The verification images obtained in this way come from the actual hardware execution results and can include the comprehensive influence of actual factors such as assembly deviations, return gaps, and rigid deformation, making the verification conclusions closer to the working state of the equipment and helping to discover systematic errors that cannot be exposed by calculation alone.

[0064] Pixel displacement is calculated on the verification image to obtain the actual pixel displacement amount; Specifically, when calculating the actual pixel displacement of the verification image, it is necessary to extract the pixel position of the same calibration feature point corresponding to the theoretical calculation from the verification image and compare it with the reference pixel position before verification to obtain the true pixel displacement response. The actual pixel displacement is usually represented in the form of a two-dimensional displacement vector, which includes both displacement magnitude and displacement direction, and can directly reflect the composite effect of linear movement and rotation in the image. In implementation, the same feature extraction method as in the calibration stage can be used to locate the corner points of the square array or the center of the circle array in the verification image. If necessary, sub-pixel fitting can be used to improve the displacement calculation resolution, and the matching with the same physical point in the reference frame can be ensured by feature point identification or topological relationship. If multiple feature points are used, the displacement results of each feature point can be statistically fused to obtain a more stable actual pixel displacement, reduce the fluctuation caused by single-point extraction error or local occlusion, and make the subsequent difference evaluation more reflective of the true accuracy level of the calibration parameters.

[0065] The difference between the actual pixel displacement and the theoretical pixel displacement is calculated to obtain the verification result.

[0066] Specifically, when calculating the difference between the actual pixel displacement and the theoretical pixel displacement to obtain the verification result, the difference reflects the magnitude of the prediction error of the initial joint calibration parameters on the actual movement of the device, and is a direct indicator of whether the calibration accuracy meets the requirements. The difference calculation can be performed at the two-dimensional vector level, comparing the deviations of the horizontal and vertical components separately, or calculating the vector magnitude error and direction error to distinguish different error sources such as scale error, deflection angle error, and rotation center error. In implementation, the corresponding error vector can be calculated for each set of verification actions, and the root mean square error or maximum error can be further calculated as the verification result output, which is convenient for comparison with the preset calibration accuracy conditions. By subtracting the theoretical prediction from the actual observation in the same pixel coordinate system, the verification result can objectively reveal the applicability of the calibration parameters under different angles and different movement directions. If the error shows directional shift or amplifies with changes in angle, it can also provide a clear basis for subsequent parameter adjustment, thus forming a closed-loop correction basis, and ultimately improving the overall reliability and repeatability of the camera-platform mapping relationship.

[0067] In an optional embodiment, adjusting the initial joint calibration parameters based on the verification results to obtain target joint calibration parameters that meet preset calibration accuracy conditions includes: Obtain the error threshold based on the preset calibration accuracy conditions; Specifically, when obtaining the error threshold based on preset calibration accuracy conditions, the preset calibration accuracy conditions are used to clarify the acceptable error range that the joint calibration results need to achieve in the equipment application scenario. The error threshold is the quantified boundary that places this requirement on calculable indicators, facilitating subsequent automatic judgment. The error threshold can be set for the pixel domain or the machine domain, for example, represented by the root mean square value of pixel displacement error or reprojection error, or by the maximum deviation of the rotation center closure error. The specific value is matched with the equipment's requirements for positioning accuracy and angular accuracy. In implementation, the calibration module can pre-configure this threshold in the software parameters, or it can be converted based on factors such as equipment model, lens magnification, pixel resolution, and minimum step distance of the motion platform to ensure that the threshold is consistent with the actual application accuracy. By first solidifying the accuracy requirements into error thresholds, subsequent verification conclusions can be judged according to a unified standard, avoiding inconsistencies caused by relying on human experience.

[0068] The verification result is compared with the error threshold. If the verification result is less than or equal to the error threshold, the initial joint calibration parameter is used as the target joint calibration parameter. Specifically, when comparing the verification results with the error threshold, the verification results are typically error indices derived from the difference between the theoretical and actual pixel displacements, including error vectors from a single verification or statistical indices from multiple verifications. The comparison process doesn't simply consider a single numerical value; it ensures that the object of comparison aligns with the threshold definition. For example, if the threshold is defined as root mean square error, the errors from multiple frames and points need to be aggregated before comparison. If the threshold is defined as the maximum error, the maximum value in the error set needs to be extracted for comparison. In practice, errors under different rotation angles and movement directions can be calculated separately, and the largest value is taken as the conservative verification result, ensuring that the calibration parameters meet requirements across the entire working range. Through this comparison step, the calibration process establishes clear pass / fail judgment conditions, transforming joint calibration from a single calculation into a verifiable and definable engineering process.

[0069] When the verification result is less than or equal to the error threshold, the initial joint calibration parameters are used as the target joint calibration parameters, which confirms and solidifies the validity of the current parameters. The initial joint calibration parameters are obtained in the preceding steps from angle data, encoder pulse count, and pixel position change data, and have already undergone pre-defined motion verification. If the error meets the threshold, it indicates that the camera visual equivalent, target deflection angle, and rotation center parameters can accurately describe the mapping relationship between the pixel domain and the machine domain under the current equipment operating conditions. In practice, this set of parameters is usually written into the equipment configuration file or calibration database, and the corresponding calibration time, calibration board number, verification error index, and other information are recorded for easy traceability and recalibration comparison. By directly outputting the target parameters when the threshold is met, unnecessary iterative calculations can be reduced, calibration efficiency can be improved, and the parameters used for production testing can be ensured to be in a verified and reliable state.

[0070] If the verification result is greater than the error threshold, the initial joint calibration parameters are corrected based on the verification result to obtain the target joint calibration parameters.

[0071] Specifically, the verification results are used to drive parameter convergence for closed-loop correction. The verification results not only indicate that the parameters do not meet accuracy requirements, but also often reveal the characteristics of error sources. For example, when the error is proportionally amplified in two orthogonal directions, it is more likely to correspond to visual equivalent deviation; when the error changes regularly with the rotation angle, it is more likely to be related to the rotation center parameter or deflection angle. The calibration module can reintroduce the verification results into the parameter solving process, incrementally adjusting the visual equivalent, deflection angle, and rotation center parameters, or re-perform joint solving with the current parameters as initial values, making the direction of error minimization clearer. Multiple sets of verification data can also be used to expand the sample set, reducing the impact of random errors. After correction, a preset motion verification is usually performed again until the error drops below the threshold. This ensures that the target joint calibration parameters not only originate from computational solutions but also undergo verification feedback correction, effectively suppressing coupling errors and cumulative deviations that are difficult to eliminate in individual calibrations, and improving the stability and reliability of the calibration results under multi-angle and multi-directional displacement conditions.

[0072] After solving the joint calibration parameters of the target, the calibration parameters such as camera visual equivalent, coordinate system deflection angle, and rotation center are uniformly stored in a structured data format. The data format includes calibration time, equipment identifier, original acquisition data index, and corresponding calibration result parameters. Based on this, each calibration result is version-managed, with a unique version identifier assigned to different calibration batches, and historical version data is retained during parameter updates, thus forming a continuous record of calibration parameter evolution. At the same time, a historical record query mechanism is established based on the calibration time and equipment identifier information, which can retrieve and call any historical calibration data to support quality traceability, anomaly analysis, and parameter comparison and inheritance during recalibration in the production process. This ensures that the calibration results can not only be used for current equipment operation but also maintain good traceability and consistency during long-term equipment use and multiple calibration processes.

[0073] Example 2 In addition, combined Figure 1 The joint calibration method in the optical communication chip package described in Embodiment 1 of the present invention can be implemented by the joint calibration system in the optical communication chip package. Figure 4 A schematic diagram of the hardware structure of the joint calibration system in the optical communication chip package provided in Embodiment 2 of the present invention is shown.

[0074] The joint calibration system in an optical communication chip package may include a processor and a memory storing computer program instructions.

[0075] Specifically, the processor may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement embodiments of the present invention.

[0076] The memory may include a large-capacity storage device for data or instructions. For example, and not limitingly, the memory may include a hard disk drive (HDD), a floppy disk drive, flash memory, an optical disk drive, a magneto-optical disk drive, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory may include removable or non-removable (or fixed) media. Where appropriate, the memory may be internal or external to a data processing device. In a particular embodiment, the memory is a non-volatile solid-state memory. In a particular embodiment, the memory includes a read-only memory (ROM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or flash memory, or a combination of two or more of these.

[0077] The processor reads and executes computer program instructions stored in the memory to implement the joint calibration method in any of the optical communication chip packages described in the above embodiments.

[0078] In one example, the joint calibration system in an optical communication chip package may also include a communication interface and a bus. For example, Figure 4 As shown, the processor, memory, and communication interface are connected via a bus and communicate with each other.

[0079] The communication interface is mainly used to enable communication between various modules, devices, units and / or equipment in the embodiments of the present invention.

[0080] A bus, including hardware, software, or both, couples components of the device together. For example, and not limitingly, a bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, a bus may include one or more buses. While specific buses are described and illustrated in embodiments of the invention, the invention contemplates any suitable bus or interconnect.

[0081] In summary, the embodiments of the present invention provide a joint calibration method and system for optical communication chip packaging.

[0082] It should be clarified that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of the present invention.

[0083] The functional blocks shown in the above-described structural diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this invention are programs or code segments used to perform the required tasks. The programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried in a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0084] The user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data shall comply with the relevant laws, regulations and standards of the relevant locality, and corresponding operation entry points shall be provided for the user to choose to authorize or refuse.

[0085] It should also be noted that the exemplary embodiments mentioned in this invention describe methods or systems based on a series of steps or apparatus. However, this invention is not limited to the order of the steps described above; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0086] The above description is merely a specific embodiment of the present invention. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the protection scope of the present invention.

Claims

1. A method for joint calibration in an optical communication chip package, the method comprising: This invention is applied to die bonding equipment, which includes: an image processing module, a composite calibration plate, a rotating platform, a motion platform, a motion control module, and a calibration module. The composite calibration plate is disposed on the rotating platform or the motion platform, and the motion control module is used to perform coordinated control of the rotating platform and the motion platform. The method includes: The motion control module controls the rotating platform to rotate to multiple preset rotation angles, and controls the motion platform to move along a first preset direction and a second preset direction at each preset rotation angle, respectively, to acquire angle data, encoder pulse count and pixel position change data before and after movement during each rotation and movement process, wherein the first preset direction and the second preset direction are perpendicular to each other; The calibration module performs joint calibration parameter calculation on the angle data, the encoder pulse count, and the pixel position change data to obtain initial joint calibration parameters, wherein the initial joint calibration parameters include camera visual equivalent, target deflection angle between camera coordinate system and device coordinate system, and rotation center parameters of rotating platform; Based on the initial joint calibration parameters, the rotating platform and the motion platform are controlled to perform preset motion verification to obtain the verification results; Based on the verification results, the initial joint calibration parameters are adjusted to obtain target joint calibration parameters that meet the preset calibration accuracy conditions; During the optical communication chip packaging process, the die bonding device is controlled to complete the packaging of the optical communication chip according to the target joint calibration parameters.

2. The joint calibration method in optical communication chip packaging according to claim 1, characterized in that, The composite calibration plate has composite calibration features on its surface, including: alignment segments for adjusting the alignment of camera coordinate axes, a square array for visual equivalent calibration, a dot array for calibrating the rotation center of the rotating platform, and positioning marks for positioning the composite calibration plate.

3. The joint calibration method in optical communication chip packaging according to claim 1, characterized in that, The die bonding device also includes an image acquisition module. The step of controlling the rotating platform to rotate to multiple preset rotation angles via the motion control module, and controlling the motion platform to move along a first preset direction and a second preset direction at each preset rotation angle, and acquiring angle data, encoder pulse count, and pixel position change data before and after movement during each rotation and movement includes: The image acquisition module acquires the initial calibration plate image at the preset position and the target calibration plate image after rotation and movement. The image processing module extracts pixel positions from the initial calibration board image and the target calibration board image to obtain the initial pixel position and the target pixel position. The difference between the initial pixel position and the target pixel position is calculated to obtain the pixel position change data; The motion control module acquires the corresponding angle data and encoder pulse count.

4. The joint calibration method in optical communication chip packaging according to claim 3, characterized in that, The step of extracting pixel positions from the initial calibration board image and the target calibration board image using the image processing module to obtain the initial pixel positions and target pixel positions includes: The initial calibration board image and the target calibration board image are respectively input into a pre-trained calibration pattern recognition model to obtain a first calibration pattern and a second calibration pattern. Contour features are extracted from the first calibration pattern and the second calibration pattern respectively to obtain the corresponding contour position information; Based on the contour position information, the geometric center positions of the first calibration pattern and the second calibration pattern are calculated to obtain the first pixel coordinate position and the second pixel coordinate position; The first pixel coordinate position is used as the initial pixel position, and the second pixel coordinate position is used as the target pixel position.

5. The joint calibration method in optical communication chip packaging according to claim 1, characterized in that, The initial joint calibration parameters obtained by calculating the joint calibration parameters of the angle data, the encoder pulse count, and the pixel position change data through the calibration module include: Based on the pixel position change data, obtain the pixel displacement vector under different rotation angles and different movement directions; Calculate the target deflection angle between the camera coordinate system and the device coordinate system based on the pixel displacement vectors described above; The camera visual equivalent is calculated by least squares fitting based on the pixel position change data and the encoder pulse count. Based on the coordinates of the trajectory points corresponding to the pixel position change data at different rotation angles, the rotation center parameters of the rotating platform are calculated using a circular trajectory fitting method.

6. The joint calibration method in optical communication chip packaging according to claim 5, characterized in that, The step of calculating the target deflection angle between the camera coordinate system and the device coordinate system based on each pixel displacement vector includes: Based on the direction information of the pixel displacement vector and the preset direction-to-deflection angle mapping relationship, each initial deflection angle is calculated; The average value of each initial deflection angle is calculated, and the calculated average deflection angle is taken as the target deflection angle.

7. The joint calibration method in optical communication chip packaging according to claim 5, characterized in that, The step of calculating the camera visual equivalent by least squares fitting based on the pixel position change data and the encoder pulse count includes: Based on the pixel position change data, obtain the displacement of each first pixel corresponding to the first preset direction and the displacement of each second pixel corresponding to the second preset direction; Based on the encoder pulse count, obtain the first pulse displacement amount corresponding to the first preset direction and the second pulse displacement amount corresponding to the second preset direction; The least squares method is used to linearly fit each pixel displacement to the corresponding pulse displacement, and the slope of the fitted line is used as the visual equivalent in the corresponding direction.

8. The joint calibration method in optical communication chip packaging according to claim 1, characterized in that, The step of controlling the rotating platform and the motion platform to perform preset motion verification based on the initial joint calibration parameters, and obtaining the verification results includes: Based on the camera visual equivalent and rotation center parameters, the preset motion amount is converted into the corresponding theoretical pixel displacement amount; Control the rotating platform and the motion platform to perform corresponding rotation and movement operations to acquire verification images; Pixel displacement is calculated on the verification image to obtain the actual pixel displacement amount; The difference between the actual pixel displacement and the theoretical pixel displacement is calculated to obtain the verification result.

9. The joint calibration method in optical communication chip packaging according to any one of claims 1-8, characterized in that, The step of adjusting the initial joint calibration parameters based on the verification results to obtain target joint calibration parameters that meet the preset calibration accuracy conditions includes: Obtain the error threshold based on the preset calibration accuracy conditions; The verification result is compared with the error threshold. If the verification result is less than or equal to the error threshold, the initial joint calibration parameter is used as the target joint calibration parameter. If the verification result is greater than the error threshold, the initial joint calibration parameters are corrected based on the verification result to obtain the target joint calibration parameters.

10. A joint calibration system in optical communication chip packaging, characterized in that, include: At least one processor, at least one memory, and computer program instructions stored in the memory, which, when executed by the processor, implement the method as described in any one of claims 1-9.

Images