High-precision optical module position correction method and device

Abstract

The application discloses a high-precision optical module position correction method and device, and relates to the technical field of optical measurement, which comprises the following steps: recognizing double-dial needle contacts through a vision system and establishing an initial mapping of images and mechanical coordinates; determining a safe operation height interval by using laser non-contact distance measurement; collecting images of multiple calibration points in the field of view to correct the mapping relationship and obtain a global accurate coordinate conversion matrix. The image of a chip is collected to calculate a tilt angle and a plane offset, when the tilt angle exceeds a threshold value, the double needle is controlled to be diagonally stepped and slightly pushed at a standard height by using an angle priority strategy, and after iteration, the double needle enters plane correction; in the plane correction, the double needle is synchronously and uniformly pushed at a low speed, and the angle is monitored in real time, and if the angle exceeds the threshold value, the angle correction is returned. Through closed-loop rechecking iteration compensation, the application finally realizes full-automatic high-precision angle correction.

Description

Technical Field

[0001] This invention relates to the field of optical measurement technology, and in particular to a high-precision optical module position correction method and device. Background Technology

[0002] After the optical module chip completes automatic bonding and pre-curing adhesive dispensing, it is affected by multiple factors, including mechanical assembly gaps in the front-end bonding equipment, positioning repeatability deviations, environmental micro-vibrations, and temperature drift. This results in a stable, albeit small, proportion of chip micro-skew and minute planar offset. These deviations are extremely small and represent inherent systemic biases of the bonding equipment. Under current technological conditions, the front-end bonding process cannot identify and effectively intercept or compensate for them in real time.

[0003] Although the aforementioned micro-positional deviations are small in magnitude, they directly affect the coaxial coupling accuracy of the optical path, leading to excessive optical insertion loss and a significant decrease in optical path coupling yield. They must be eliminated through post-processing and fine-tuning.

[0004] Currently, the industry relies entirely on manual operation for correcting chip misalignment and offset, primarily employing two modes: first, operators manually manipulate a single probe for minute adjustments by rotating precision mechanical knobs under a high-magnification microscope; second, operators perform purely manual adjustments using probes or tweezers under high-definition magnification. These delicate micro-operations heavily depend on the operator's feel and experience, resulting in lengthy training periods for newcomers, strong reliance on experienced workers, high labor costs, and limited production capacity. More importantly, manual operation cannot consistently achieve the ultra-high process requirements of angle accuracy below 0.1° and planar correction accuracy no greater than 3μm. Furthermore, uneven force during operation can easily lead to secondary damage such as chip chipping, coating scratches, or tearing of the pre-cured adhesive layer. Summary of the Invention

[0005] The main objective of this invention is to provide a high-precision optical module position correction method and device. This invention aims to solve the problem that after the optical module chip is bonded and pre-cured, systematic deviations in the front-end equipment can cause micron-level skewness and planar offset of the chip, which can damage the coaxial coupling accuracy of the optical path, resulting in excessive optical insertion loss and a significant decrease in coupling yield. Existing manual rework operations rely heavily on the operator's feel and experience, and cannot consistently achieve the ultra-high process requirements of angle accuracy below 0.1° and planar correction accuracy ≤3μm. Furthermore, these methods suffer from low efficiency, limited production capacity, long training cycles, and susceptibility to secondary damage.

[0006] To achieve the above objectives, this invention proposes a high-precision optical module position correction method for correcting the pose of a pre-cured optical module chip. The method is based on a vision system and a dual-pin mechanism, and includes the following steps: The system acquires images of the dual dial pins, identifies the contact points of the dial pins, and establishes an initial mapping relationship between the image coordinate system and the mechanical motion coordinate system; wherein multiple calibration points are set within the field of view of the vision system. Measure the vertical distance between the tip of the dial pin and the PCB board surface, and determine the safe operating height range of the dial pin in the vertical direction based on the vertical distance; The two pins are moved sequentially to each calibration point, and images of the two pins at each calibration point are acquired. The initial mapping relationship is corrected based on the coordinates of the images of the two pins at each calibration point and the corresponding mechanical motion coordinates to obtain a coordinate transformation matrix. Acquire an image of the chip to be corrected, and use the coordinate transformation matrix to determine the current tilt angle and planar offset of the chip; When the tilt angle exceeds the preset angle accuracy threshold, the required angle correction amount is converted into the target motion amount of the dual pins based on the coordinate transformation matrix, and the dual pins are controlled to move in coordination according to the target motion amount at the standard working height to correct the angle of the chip until the tilt angle meets the angle accuracy threshold. Based on the coordinate transformation matrix, the planar offset is converted into the target motion of the dual pins, and the dual pins are controlled to move at the standard working height to perform planar position correction; and the tilt angle of the chip is monitored in real time. If the tilt angle exceeds the angle accuracy threshold, the planar position correction is paused, and the process returns to the step of performing angle correction on the chip.

[0007] Furthermore, the step of measuring the vertical distance between the tip of the dial pin and the PCB board surface specifically includes: The vertical distance between the tip of the dial pin and the PCB board surface is measured using a laser non-contact ranging method.

[0008] Furthermore, the lower limit safety height is calibrated by the reference plane of the PCB board surface, which is used to limit the maximum downward travel of the Z-axis, and through a dual interlock of algorithm and hardware, it prevents the pin from scratching the PCB base plate and the circuit. The standard operating height is configured to ensure that the pins are properly aligned with the chip edge, avoiding excessive height that could lead to push failure. During angle correction and planar position correction, the Z-axis movement of the dial pin is locked within the range formed by the lower safety height and the standard working height.

[0009] Furthermore, the multiple calibration points are nine points that uniformly cover the complete field of view of the vision system, including the center of the field of view, the midpoints of the four sides, and the four corners; the correction of the initial mapping relationship includes: sequentially controlling the pin to move to each calibration point, acquiring images and obtaining the visual pixel coordinates of each point, and synchronously recording the corresponding motor motion mechanical coordinates; calculating the deviation distance between the visual pixel coordinates after transformation by the current mapping relationship and the mechanical coordinates point by point; when the deviation distance of any calibration point exceeds 1μm, the affine transformation parameters are refitted and corrected based on the current correspondence of all nine sets of coordinates and the coordinate transformation matrix is ​​updated, and all calibration points are re-traversed for deviation verification until all nine calibration points meet the deviation distance ≤ 1μm, thus completing the full-domain accuracy calibration.

[0010] Furthermore, the preset angle accuracy threshold is 0.1°.

[0011] Furthermore, the step of converting the required angle correction amount into the target motion amount of the dual pins based on the coordinate transformation matrix, and controlling the dual pins to move collaboratively according to the target motion amount at the standard working height includes: controlling the dual pins to push diagonally stepwise with the chip center as the rotation reference, and the single angle correction step distance being less than or equal to 0.02°.

[0012] Furthermore, the method also includes: after completing the planar position correction, re-checking the tilt angle and planar offset of the chip. If the tilt angle is not less than 0.1°, the X-axis offset is ≤3μm and the Y-axis offset is not ≤3μm at the same time, then the angle correction and planar position correction are re-executed in sequence until all meet the standards.

[0013] Furthermore, during the planar position correction process, the left and right pins are controlled to move synchronously and in the same direction, and pushed at a preset low speed and small step distance. The Z-axis height is maintained at the standard working height throughout the process and is limited within the safe working height range to ensure uniform force on the chip and avoid chip edge breakage, coating damage or adhesive layer tearing.

[0014] Furthermore, determining the current tilt angle and planar offset of the chip specifically includes: performing ROI region masking, image enhancement, and adaptive threshold segmentation preprocessing on the acquired chip image to suppress interference from glue overflow, glue fog, and strong reflections from the chip mirror surface; A dual-feature fusion detection algorithm combining contour and baseline is adopted, and the coordinate transformation matrix is ​​used to calculate the chip's real-time tilt angle Δθ, lateral offset ΔX, and longitudinal offset ΔY, so as to achieve the recognition of tilt angle changes as low as 0.01°.

[0015] The present invention also proposes a high-precision optical module position correction device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The device is characterized in that, when the processor executes the computer program, it implements the steps of the high-precision optical module position correction method as described in any one of claims 1 to 9.

[0016] This invention establishes a hand-eye coordinated benchmark through visual calibration of dual pin contacts, 9-point global coordinate calibration, and laser non-contact automatic height detection. It adopts logic control with angle priority and real-time interlocking of 0.1° threshold to ensure the coordinated movement of dual pins within the safe operating height range. First, it uses the chip center as the rotation benchmark to achieve precise convergence of the tilt angle to within 0.1° by diagonally pushing step by step. Then, it uses synchronous and unidirectional low-speed small-step push to complete the plane offset correction to within 3μm. The entire process is closed-loop re-inspection and iterative compensation, realizing fully automatic, high-precision, and non-destructive chip pose correction, which significantly improves the yield of optical path coupling and mass production consistency. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the high-precision optical module position correction method of the present invention. Figure 2 This is a logic diagram of the standardized closed-loop correction process of the present invention; Figure 3 This is a schematic diagram showing the distribution of calibration points across the entire field of view in this invention.

[0020] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of the present invention and are not intended to limit the present invention.

[0022] To better understand the technical solution of the present invention, a detailed description will be provided below in conjunction with the accompanying drawings and specific embodiments.

[0023] After the optical module chips complete automatic bonding and pre-curing adhesive dispensing, a certain proportion of chip micro-skew and minute planar offset will stably occur due to the coupling effects of factors such as the mechanical assembly gaps of the front-end bonding equipment, positioning repeatability deviations, environmental micro-vibrations, and temperature drift. These deviations are extremely small and are inherent systemic biases of the bonding equipment, which the front-end bonding process cannot identify and effectively intercept or compensate for in real time. These micro-positional deviations will directly damage the coaxial coupling accuracy of the optical path, leading to excessive optical insertion loss and a significant decrease in optical path coupling yield, and must be eliminated through post-processing precision correction.

[0024] Currently, the industry relies entirely on manual operation for correcting chip misalignment and offset, primarily employing two work modes: First, under a high-magnification microscope, operators manually manipulate a single probe to slightly adjust the chip by rotating a precision mechanical knob; second, under high-definition magnification, operators directly hold probes or tweezers for purely manual adjustments. These delicate micro-operations heavily depend on the operator's feel and experience, resulting in long training periods for new employees, strong reliance on experienced workers, high labor costs, and limited production capacity, making it difficult to meet the demands of large-scale mass production.

[0025] To address the above problems, the present invention proposes the following solution: like Figure 1 As shown, Figure 1 This is a flowchart illustrating the high-precision optical module position correction method of the present invention.

[0026] This invention proposes a high-precision optical module position correction method for correcting the pose of a pre-cured optical module chip. The method is based on a vision system and a dual-pin mechanism, and includes the following steps: S10, acquire an image of the dual dial pins, identify the contact point position of the dial pins, and establish an initial mapping relationship between the image coordinate system and the mechanical motion coordinate system; wherein multiple calibration points are set within the field of view of the vision system; S20, Measure the vertical distance between the tip of the dial pin and the PCB board surface, and determine the safe operating height range of the dial pin in the vertical direction based on the vertical distance; S30, move the two pins to each calibration point in sequence, and obtain images of the two pins at each calibration point. Then, correct the initial mapping relationship based on the coordinates of the images of the two pins at each calibration point and the corresponding mechanical motion coordinates to obtain a coordinate transformation matrix. S40, acquire the chip image to be corrected, and use the coordinate transformation matrix to determine the current tilt angle and plane offset of the chip; S50, when the tilt angle exceeds the preset angle accuracy threshold, the required angle correction amount is converted into the target motion amount of the dual pins based on the coordinate transformation matrix, and the dual pins are controlled to move in coordination according to the target motion amount at the standard working height to correct the angle of the chip until the tilt angle meets the angle accuracy threshold. S60, based on the coordinate transformation matrix, the planar offset is converted into the target motion amount of the dual pins, and the dual pins are controlled to move at the standard working height to perform planar position correction; and, the tilt angle of the chip is monitored in real time. If the tilt angle is detected to exceed the angle accuracy threshold, the planar position correction is paused, and the process returns to the step of performing angle correction on the chip.

[0027] S10, acquire images of the dual dial pins, identify the contact points of the dial pins, and establish an initial mapping relationship between the image coordinate system and the mechanical motion coordinate system; wherein multiple calibration points are set within the field of view of the vision system.

[0028] During this step, the high-magnification microscopic vision system automatically focuses and images the left and right pins, acquiring high-resolution images containing the complete outline of the pins. Considering the micron-level wear and plastic deformation of the pin tips during continuous operation, and the unavoidable installation position deviations introduced during assembly, directly using the geometric center or theoretical endpoint of the pins as the working reference would introduce systematic errors. Therefore, it is necessary to accurately locate the actual working contact point through image recognition. The system calls a sub-pixel edge extraction algorithm to calculate the gradient magnitude and fit the edge chain code to the pin image, obtaining a sub-pixel-level accurate description of the pin outline. Then, it locks the coordinates of the working contact point where the pin makes mechanical contact with the chip edge along the outline features. This contact point constitutes the spatial reference origin for all subsequent pushing actions, and its positioning accuracy directly determines the reference accuracy level of the correction system.

[0029] After contact point calibration, the system controls the XY motion module to traverse a set of preset standard points, synchronously acquiring the mechanical coordinates fed back from the motor encoder at each point, and extracting the contact point pixel coordinates in the corresponding visual frame. Based on the corresponding datasets of multiple sets of pixel coordinates and mechanical coordinates, the least squares method is used to fit and solve the affine transformation matrix, establishing an initial mapping relationship between the image pixel coordinate system and the mechanical motion coordinate system. Thus, the micron-level pose deviation detected by the vision system can be accurately converted into motor stepping commands executable by the motion system, providing a unified coordinate transformation benchmark for subsequent correction stages and eliminating calibration errors between the visual perception end and the motion execution end at the root.

[0030] S20, Measure the vertical distance between the tip of the dial pin and the PCB board surface, and determine the safe operating height range of the dial pin in the vertical direction based on the vertical distance.

[0031] During this step, the system invokes a laser ranging module mounted on the motion platform in parallel with the vision system. This module emits a focused laser beam to illuminate the PCB surface in a non-contact manner, and calculates the absolute vertical distance between the needle tip and the PCB surface in real time by receiving the time difference or phase difference of the reflected light. Before each batch of work, the equipment performs an automatic height detection process: it controls the needle to move to a preset coarse positioning area above the PCB surface, activates laser ranging for continuous sampling, and averages the distance data from multiple measurement points to eliminate random errors caused by local unevenness of the board surface, thus obtaining a uniform reference plane height for that batch of work.

[0032] Using this reference plane as the zero point, the system automatically calculates and locks two key height thresholds. The lower safety height limits the maximum downward travel of the Z-axis. Its value comprehensively considers factors such as PCB board thickness, copper layer thickness, allowable board surface deformation tolerance, and chip adhesive layer thickness to ensure that even under the worst-case tolerance conditions, the tip of the pin still has sufficient safety margin from the PCB board surface, eliminating the risk of the pin scratching the PCB base and circuitry. The standard operating height is set according to the thickness and edge height of the chip to be corrected, ensuring that the pin does not press down excessively when contacting the chip edge, causing chip damage, nor does it dangle above the chip edge, causing the pushing force to fail due to ineffective transmission. Throughout the correction process, the Z-axis movement is locked within this safe operating height range. The soft limit of the control algorithm and the hardware limit of the driver form a double interlock protection, fundamentally solving the common problems of scraping scrap and pin dangling failure in traditional manual operation.

[0033] After completing the pre-calibration and safety height setting, the standardized closed-loop correction logic of the entire correction process is as follows: Figure 2 As shown, Figure 2 A logic diagram for a standardized closed-loop correction process.

[0034] refer to Figure 2The closed-loop control logic of the entire correction process is as follows: After correction starts, system initialization is performed first. If initialization fails, the calibration process on the right is entered. The automatic height detection of the dual pin Z-axis and PCB reference zero point calibration are completed in sequence, the automatic calibration of the camera's 9-point full-domain field of view and the coordinate transformation matrix interval update are completed, and the visual recognition and contact positioning calibration of the dual pins are completed to establish a unified mapping of image / mechanical coordinates. After calibration, the process returns to the initialization judgment node. After initialization is successful, the chip tilt angle and XY plane deviation are detected by high-magnification visual acquisition. Following the principle of "angle priority", the tilt angle is first determined to meet the accuracy requirements. If the deviation exceeds the limit, the angle correction of the dual pins is triggered. After the angle meets the standard, the X-axis offset and Y-axis offset are determined to meet the thresholds in sequence, and the X-axis and Y-axis position correction of the dual pins is triggered accordingly. After all three indicators are qualified, the correction is determined to be completed. The three judgment nodes correspond to the accuracy thresholds of the angle, X-axis and Y-axis, respectively. If any dimension exceeds the limit, the corresponding correction process is entered. After correction, the process returns to the detection stage for judgment, forming a closed-loop iterative mechanism of "detection-correction-re-inspection".

[0035] S30, move the two pins to each calibration point in sequence, and obtain images of the two pins at each calibration point. Then, correct the initial mapping relationship based on the coordinates of the images of the two pins at each calibration point and the corresponding mechanical motion coordinates to obtain a coordinate transformation matrix.

[0036] In this embodiment, nine uniformly distributed calibration points are pre-planned within the complete imaging field of view of the vision system, and the point layout is as follows: Figure 3 As shown, Figure 3 This is a schematic diagram showing the distribution of calibration points across the entire field of view.

[0037] refer to Figure 3 The nine points completely cover the central area of ​​the field of view, the midpoints of the four sides, and the four corners, arranged symmetrically in a nine-square grid to ensure no precision blind spots across the entire field of view. The system controls the XY module to move the pins precisely to each preset calibration point in sequence. After reaching each point, the motion platform pauses briefly to eliminate residual vibrations, and then triggers the camera to capture a stable image. For the image captured at each point, a sub-pixel feature localization algorithm is used to extract the precise visual pixel coordinates of the pin contacts. Simultaneously, the actual mechanical coordinates of the current module's motor motion are read and recorded in a one-to-one correspondence.

[0038] After acquiring data from all nine points, the system calculates the mapping deviation for each point, which is the Euclidean distance deviation between the predicted mechanical coordinates obtained by transforming the visual pixel coordinates of that point using the current mapping matrix and the actually recorded mechanical coordinates. If the deviation distance of any calibration point exceeds a preset threshold of 1μm, it indicates that there is local accuracy decay in the current coordinate transformation matrix. The system then refits and corrects the affine transformation parameters and updates the global coordinate transformation matrix based on the corresponding data of all nine sets of pixel coordinates and mechanical coordinates. After the update, the system automatically re-traverses all nine calibration points to verify the deviation, forming an automatic iterative closed loop of "acquisition, comparison, correction, and verification". This iterative process is completed fully automatically by the system and usually converges after 2 to 3 rounds of iteration, ultimately achieving a mapping deviation of no more than 1μm for all nine points, completing the global accuracy calibration. After this global calibration, regardless of where the chip appears in the field of view, its visual detection deviation can be accurately converted into mechanical motion coordinates with a uniform mapping accuracy, completely eliminating the hidden danger of decreased edge area correction accuracy caused by lens distortion and temperature drift.

[0039] S40: Acquire an image of the chip to be corrected, and use the coordinate transformation matrix to determine the current tilt angle and plane offset of the chip.

[0040] After the optical module chip is pre-cured, varying degrees of adhesive overflow and adhesive fog often exist around the chip. Furthermore, the chip's surface typically has a mirror coating, which easily generates strong reflections under high-magnification microscopic imaging conditions. These three types of interference factors severely reduce the edge extraction accuracy of traditional image processing algorithms. To address this complex situation, this step first performs ROI (Region of Interest) masking on the acquired raw chip image. This involves defining the region of interest in the image based on the chip's theoretical installation location, masking all irrelevant background and interference information outside the region, reducing computational load and eliminating non-target feature interference. Subsequently, image enhancement processing is performed on the image within the ROI region. Histogram equalization and contrast stretching algorithms are used to improve the grayscale gradient between the chip edge and the background. Simultaneously, an adaptive threshold segmentation algorithm dynamically adjusts the segmentation threshold in different local areas of the image, effectively suppressing light spots caused by strong reflections and edge blurring caused by adhesive overflow, thus completely separating the chip's true outline from the complex background.

[0041] In the feature detection stage, the system employs a dual-feature fusion detection algorithm combining contour and reference features: on one hand, it extracts the outer contour edges of the chip for straight-line fitting and geometric center calculation; on the other hand, it identifies preset reference markers on the chip as angle orientation references. After fusing the two feature information streams, and combining them with the calibrated and updated precise coordinate transformation matrix, the pixel-level deviations detected in the image are converted into actual geometric deviations in physical space, accurately calculating the chip's real-time tilt angle Δθ, lateral offset ΔX, and longitudinal offset ΔY. This algorithm achieves a tilt angle recognition resolution as low as 0.01°, capturing minute angular deviations imperceptible to the human eye, providing a reliable quantitative basis for subsequent high-precision correction.

[0042] S50, when the tilt angle exceeds the preset angle accuracy threshold, the required angle correction amount is converted into the target motion amount of the dual pins based on the coordinate transformation matrix, and the dual pins are controlled to move in coordination according to the target motion amount at the standard working height to correct the angle of the chip until the tilt angle meets the angle accuracy threshold.

[0043] This step embodies the core "angle-first" process logic of this invention. The preset angle accuracy threshold is 0.1°, determined based on engineering experience with coaxial coupling of the optical module's optical path. If the detected chip tilt angle Δθ is not less than 0.1°, the system immediately executes the angle correction process, temporarily suspending planar position adjustment regardless of the magnitude of the planar offset. At the start of angle correction, the system inputs the visually detected tilt angle deviation Δθ into the control algorithm. Through a calibrated coordinate transformation matrix, the angle correction amount is precisely decomposed into the micro-displacement components (i.e., target motion) of the left and right pins in the X and Y axes, respectively. Driven by their respective Z-axis motors, the two pins descend to the calibrated standard operating height, ensuring effective contact between the pins and the chip edge.

[0044] During the correction process, the two adjustment pins on the left and right coordinate from opposite corners of the chip, using the chip's center point as the rotation reference: one pin pushes forward slightly, while the other pushes back slightly, creating a pure torque effect to rotate the chip around its center. To avoid excessive stress between the chip and the adhesive layer, which could tear the pre-cured adhesive layer, or cause edge damage to the brittle optical chip due to a single large-amplitude correction, the system adopts a small-step, step-by-step pushing strategy, with the single angle correction step strictly controlled within 0.02°. After each small-amplitude angle push, the vision system re-acquires the chip image for retesting. If the tilt angle still does not meet the standard, the iteration continues until the tilt angle stabilizes and converges to within 0.1°. This flexible, small-step iterative correction method ensures, from a mechanical perspective, that the chip experiences uniform and controllable force during the correction process, maximizing the protection of the chip and adhesive layer's integrity.

[0045] S60, based on the coordinate transformation matrix, the planar offset is converted into the target motion amount of the dual pins, and the dual pins are controlled to move at the standard working height to perform planar position correction; and, the tilt angle of the chip is monitored in real time. If the tilt angle is detected to exceed the angle accuracy threshold, the planar position correction is paused, and the process returns to the step of performing angle correction on the chip.

[0046] In this embodiment, after the angle correction is qualified, the system enters the planar position correction stage; at this time, the chip tilt angle has stabilized within 0.1°, and only the lateral offset ΔX and longitudinal offset ΔY need to be corrected. The system uses a coordinate transformation matrix to convert the planar offsets in the two directions into synchronous movement step distances of the left and right dial pins on the X and Y axes. The two dial pins maintain synchronous and unidirectional movement at the standard working height, using a preset low-speed, small-step pushing mode to translate the entire chip to the target position. The low-speed, small-step movement mode ensures that the pushing force is applied smoothly, avoiding impact damage to the chip due to excessive speed or excessive step distance.

[0047] Furthermore, throughout the entire planar position correction process, the vision system continuously acquires images of the chip and monitors its tilt angle in real time, maintaining a 0.1° angle interlock threshold. This real-time monitoring mechanism prevents secondary chip tilting during planar translation due to uneven force or localized adhesion of the adhesive layer. If the system detects that the chip tilt angle exceeds 0.1° again, it immediately pauses the current planar translation, locks the current position, and automatically returns to the angle correction process to re-correct and lock the angle. Planar position adjustment can only continue after the angle is again within acceptable limits. This "angle interlock" mechanism ensures, from the process control logic level, that the correction processes in both angle and planar dimensions are strictly decoupled and executed sequentially, preventing the problem of tilting while moving or worsening with each correction, and guaranteeing the stability and consistency of the final correction result.

[0048] The aforementioned correction method also includes a closed-loop re-inspection step: after completing the planar position correction, a full visual re-inspection is performed in situ, the chip image is re-acquired, and the pose detection process is fully executed to obtain the corrected tilt angle, X-axis offset, and Y-axis offset, which are then compared one by one with the final accuracy standard. The final accuracy standard includes three indicators that must be met simultaneously: tilt angle Δθ less than 0.1°, X-axis planar offset not greater than 3μm, and Y-axis planar offset not greater than 3μm. If any parameter exceeds the tolerance, the system automatically triggers a cyclic iterative correction process, re-executing angle correction, planar position correction, and visual re-inspection sequentially until all three accuracy indicators simultaneously meet the standard. Once this is achieved, the chip is deemed to have passed correction, and a pass signal is sent to the production line to release the workstation. This closed-loop iterative mechanism ensures that the correction results of each chip are rigorously verified, guaranteeing consistency and traceability in mass production from a quality control perspective. Meanwhile, all process data generated during the correction process, such as tilt angle data, plane offset data, Z-axis height parameters, calibration deviation records, and correction motion trajectory, are completely recorded and stored. This data can be used for process traceability of the chip and can also be fed back to the front-end bonding equipment for process parameter optimization, forming a continuous improvement closed loop for the entire packaging production line.

[0049] Furthermore, during the planar position correction process, the left and right pins ensure that their start-stop time and displacement speed are strictly consistent through the synchronous interpolation algorithm of the motion controller, achieving synchronous and unidirectional advancement. The advancement speed is set to a low speed setting, and the step size is set to a micrometer-level small step size, ensuring that the mechanical impulse of each push is sufficiently small to avoid impact damage to the chip and adhesive layer. The independent Z-axis motors of the two pins precisely maintain the pin tip height at the standard working height, and compensate for possible minute positional drifts in real time through servo closed-loop control. The upper and lower limits of the safe working height range are written into the motion planner as boundary constraints. Motion commands exceeding the range will be directly truncated and rejected, forming soft limit protection; at the same time, the driver and limit sensor constitute hardware limit protection, and the two are superimposed to form a double interlock. This symmetrical and coordinated force-bearing control method with small step size and flexible push-correction ensures that the thrust is evenly distributed on the contact points of the pins on both sides during the chip's planar movement, avoiding torque deflection and local stress concentration caused by unilateral force, preventing edge chipping of brittle optical chips, coating scratches, and tearing of the pre-cured adhesive layer, thus maximizing the protection of product integrity while ensuring correction accuracy.

[0050] This invention also proposes a high-precision optical module position correction device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the aforementioned high-precision optical module position correction method. This device, as the control core of the correction system, is responsible for coordinating and managing the execution of all algorithm tasks, including visual image acquisition and processing, laser ranging data processing, coordinate transformation matrix operations, deviation conversion, motion trajectory planning, multi-axis motor synchronous control, and closed-loop re-inspection judgment. The memory pre-stores key process parameters such as coordinate transformation matrix parameters, standard coordinates of nine calibration points, safe operating height range thresholds, angle accuracy thresholds, and plane offset accuracy indicators. The processor sequentially calls each functional module to execute the corresponding steps according to the instructions of the computer program, realizing fully automatic high-precision closed-loop correction of the optical module chip.

[0051] The above description is only a part of the embodiments of the present invention and does not limit the patent scope of the present invention. All equivalent structural transformations made under the technical concept of the present invention using the contents of the present invention specification and drawings, or direct / indirect applications in other related technical fields, are included within the patent protection scope of the present invention.

Claims

1. A high-precision optical module position correction method, used to correct the pose of a pre-cured optical module chip, characterized in that, include: Images of the left and right dial pins are acquired through a vision system, the contact positions of the dial pins are identified, and an initial mapping relationship is established between the image coordinate system and the mechanical motion coordinate system; wherein multiple calibration points are set within the field of view of the vision system. Measure the vertical distance between the tip of the dial pin and the PCB board surface, and determine the safe operating height range of the dial pin in the vertical direction based on the vertical distance. The safe operating height range includes the lower limit safe height and the standard operating height. After sequentially moving the dial pin to each calibration point, images are acquired, and the initial mapping relationship is corrected based on the image coordinates of each calibration point and the corresponding mechanical motion coordinates to obtain a coordinate transformation matrix. Acquire an image of the chip to be corrected, and use the coordinate transformation matrix to determine the current tilt angle and planar offset of the chip; When the tilt angle exceeds the preset angle accuracy threshold, the required angle correction amount is converted into the target motion amount of the dual pins based on the coordinate transformation matrix, and the dual pins are controlled to move in coordination according to the target motion amount at the standard working height to correct the angle of the chip until the tilt angle meets the angle accuracy threshold. After the tilt angle meets the angle accuracy threshold, the plane offset is converted into the target motion of the dual pins based on the coordinate transformation matrix, and the dual pins are controlled to move at the standard working height to perform plane position correction; during the plane position correction process, the tilt angle of the chip is monitored in real time. If the tilt angle exceeds the angle accuracy threshold, the plane position correction is paused and the angle correction step is returned.

2. The high-precision optical module position correction method according to claim 1, characterized in that, In the step of measuring the vertical distance between the tip of the dial pin and the PCB board surface, and determining the safe operating height range of the dial pin in the vertical direction based on the vertical distance, wherein the safe operating height range includes a lower limit safe height and a standard operating height, the vertical distance between the tip of the dial pin and the PCB board surface is measured using a laser non-contact ranging method.

3. The high-precision optical module position correction method according to claim 1, characterized in that, The lower safety height is configured to prevent the pin from scratching the PCB board surface and circuits, and the standard working height is configured to ensure that the pin effectively contacts the chip edge; in the angle correction step and the plane correction step, the Z-axis movement of the pin is restricted within the safe working height range.

4. The high-precision optical module position correction method according to claim 1, characterized in that, The multiple calibration points are nine points that uniformly cover the complete field of view of the vision system, including the center of the field of view, the midpoints of the four sides, and the four corners; the correction includes: comparing the image coordinates of each point with the corresponding mechanical motion coordinates, and when the single-point mapping error exceeds 1μm, iteratively performing calibration until the overall accuracy is qualified.

5. The high-precision optical module position correction method according to claim 1, characterized in that, The preset angle accuracy threshold is 0.1°.

6. The high-precision optical module position correction method according to claim 5, characterized in that, The step of converting the required angle correction amount into the target motion amount of the dual pins based on the coordinate transformation matrix, and controlling the dual pins to move in coordination according to the target motion amount at the standard working height includes: controlling the dual pins to push diagonally step by step with the chip center as the rotation reference, and the single angle correction step distance is less than or equal to 0.02°.

7. The high-precision optical module position correction method according to claim 1, characterized in that, The method further includes: after completing the planar position correction, re-checking the tilt angle and planar offset of the chip. If the tilt angle is not less than 0.1°, the X-axis offset is ≤3μm and the Y-axis offset is not ≤3μm at the same time, then the angle correction and planar position correction are re-executed in sequence until all meet the standards.

8. The high-precision optical module position correction method according to claim 1, characterized in that, During the planar position correction process, the dual pins are controlled to maintain synchronous and unidirectional movement, pushed at a preset low speed and small step distance, and the Z-axis height of the pins is maintained at the standard working height.

9. The high-precision optical module position correction method according to claim 1, characterized in that, The determination of the current tilt angle and planar offset of the chip specifically includes: performing ROI region masking and image enhancement processing on the acquired chip image, and using a contour and reference dual feature fusion algorithm to identify the chip pose, so as to achieve the recognition of tilt angle changes as low as 0.01°.

10. A high-precision optical module position correction device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, when the processor executes the computer program, it implements the steps of the high-precision optical module position correction method as described in any one of claims 1-9.

Images