A method and apparatus for horizontal calibration of Micro LED chips

Abstract

The application provides a Micro LED chip horizontal correction method and device, and belongs to the technical field of optical detection of micro display devices. The method comprises the following steps: acquiring inclination angle deviation data of a Micro LED chip surface relative to a reference plane; converting the inclination angle deviation data into a physical position compensation amount of a multi-dimensional adjustment mechanism based on a pre-calibrated inclination angle-position mapping model; driving the multi-dimensional adjustment mechanism to perform horizontal adjustment based on the physical position compensation amount, and applying attenuation control to the physical position compensation amount in an iterative adjustment process, and stopping the iterative adjustment when the current inclination angle deviation data and / or the current iteration number meets a preset exit condition. The Micro LED chip horizontal correction method and device provided by the application effectively suppresses system overshoot oscillation and realizes high-precision automatic horizontal correction.

Description

Technical Field

[0001] This invention relates to the field of optical inspection technology for microdisplay devices, and in particular to a method and apparatus for horizontal calibration of Micro LED chips. Background Technology

[0002] In the optical inspection of micro-display devices (such as Micro LED chip modules), the levelness of the device surface directly affects the depth of field and inspection accuracy of the imaging system. As chip feature sizes evolve towards the sub-micron level, the usable depth of field of the optical system narrows dramatically, placing stringent requirements on the leveling accuracy of the stage. In existing technologies, some inspection equipment relies on manual intervention or simple open-loop adjustments to achieve stage leveling, which is difficult to accommodate material fluctuations and placement deviations. While automatic closed-loop adjustment schemes often ignore the cross-coupling and nonlinear distortion in the mechanical transmission chain, resulting in residual errors in the conversion of angular deviations to physical positions. Furthermore, in the closed-loop iterative adjustment process of high-inertia stages, directly applying theoretical compensation can easily trigger system overshoot and oscillations, leading to slow convergence or even failure of the correction process, failing to meet the requirements of high-precision, high-efficiency automated inspection. Summary of the Invention

[0003] This invention provides a method and apparatus for horizontal calibration of Micro LED chips, which solves the problems of low horizontal calibration accuracy and poor closed-loop adjustment stability in the detection of micro display devices in the prior art.

[0004] In a first aspect, the present invention provides a method for horizontal calibration of a Micro LED chip, comprising: acquiring tilt angle deviation data of the Micro LED chip surface relative to a reference plane; converting the tilt angle deviation data into a physical position compensation amount of a multi-dimensional adjustment mechanism based on a pre-calibrated tilt angle-position mapping model; driving the multi-dimensional adjustment mechanism to perform horizontal adjustment based on the physical position compensation amount, and applying attenuation control to the physical position compensation amount during the iterative adjustment process, and stopping the iterative adjustment when the current tilt angle deviation data and / or the current iteration number meet a preset exit condition.

[0005] According to the Micro LED chip horizontal calibration method provided by the present invention, the method for obtaining tilt angle deviation data of the Micro LED chip surface relative to a reference plane includes: arranging multiple measuring points on the Micro LED chip surface, measuring the height data of each measuring point to obtain spatial surface data; filtering the height data; calculating the tilt angle of the Micro LED chip surface relative to the reference plane based on the filtered height data using least squares plane fitting to obtain the tilt angle deviation data; or, directly measuring the absolute tilt angle of the Micro LED chip surface relative to the reference plane to obtain the tilt angle deviation data.

[0006] According to the Micro LED chip horizontal calibration method provided by the present invention, the tilt angle-position mapping model includes a linear mapping model, which includes a gain matrix and a bias vector; the gain matrix includes a main gain coefficient and a cross-coupling coefficient; the multi-dimensional adjustment mechanism includes at least two rotation axes, which include an Rx axis and a Ry axis; based on the pre-calibrated tilt angle-position mapping model, the tilt angle deviation data is converted into the physical position compensation amount of the multi-dimensional adjustment mechanism, including: when the ratio of the cross-coupling coefficient to the main gain coefficient is less than or equal to a preset ratio, ignoring the cross-coupling coefficient, and calculating the physical position compensation amount corresponding to each rotation axis based on the main gain coefficient and the bias vector; when the ratio of the cross-coupling coefficient to the main gain coefficient is greater than the preset ratio, decoupling calculation is performed using the inverse matrix of the complete gain matrix to obtain the physical position compensation amount corresponding to each rotation axis.

[0007] According to the Micro LED chip horizontal calibration method provided by the present invention, the calibration method of the linear mapping model includes: dividing the working range of each rotation axis of the multi-dimensional adjustment mechanism into a grid to form multiple grid nodes; sequentially driving the multi-dimensional adjustment mechanism to the physical position corresponding to each grid node, and collecting data pairs of physical position and measurement angle at each grid node; and solving the gain matrix and the bias vector by least squares method based on all the data pairs.

[0008] According to the Micro LED chip horizontal correction method provided by the present invention, the tilt angle-position mapping model includes a nonlinear mapping model, the nonlinear mapping model includes a pre-constructed inverse mapping table, the inverse mapping table characterizes the mapping relationship from the measured angle to the physical position; Based on a pre-calibrated tilt angle-position mapping model, the tilt angle deviation data is converted into a physical position compensation amount for a multi-dimensional adjustment mechanism. This includes: determining multiple grid nodes corresponding to the grid cell in the reverse mapping table based on the tilt angle deviation data; calculating the actual physical position corresponding to the tilt angle deviation data through interpolation based on the physical positions corresponding to the multiple grid nodes; and determining the physical position compensation amount based on the difference between the Micro LED chip's physical position corresponding to zero tilt angle and the actual physical position.

[0009] According to the Micro LED chip horizontal calibration method provided by the present invention, the calibration method of the nonlinear mapping model includes: the multidimensional adjustment mechanism provides adjustment of at least two rotation axes; dividing the working range of each rotation axis of the multidimensional adjustment mechanism into a grid to form multiple grid nodes; sequentially driving the multidimensional adjustment mechanism to the physical position corresponding to each grid node; collecting data pairs of physical position and measurement angle at each grid node to construct a forward mapping matrix; and converting the forward mapping matrix into the inverse mapping table characterizing the mapping relationship from the measurement angle to the physical position.

[0010] According to the Micro LED chip horizontal correction method provided by the present invention, during the iterative adjustment process, attenuation control is applied to the physical position compensation amount, and the iterative adjustment is stopped when the current tilt angle deviation data and / or the current iteration number meet a preset exit condition. The method includes: multiplying the physical position compensation amount by a preset relaxation factor to obtain the actual physical compensation amount, and driving the multi-dimensional adjustment mechanism to perform horizontal adjustment based on the actual physical compensation amount; after each iterative adjustment, determining whether the current tilt angle deviation data is less than or equal to a preset deviation threshold, or whether the current iteration number has reached a preset maximum iteration number; if either of the above conditions is met, the iterative adjustment is stopped.

[0011] Secondly, the present invention also provides a Micro LED chip horizontal correction device, comprising: A surface profile measuring instrument is used to measure the surface profile data or tilt angle of a Micro LED chip. A multi-dimensional adjustment mechanism for supporting the Micro LED chip and providing adjustment along at least two rotation axes; A controller is configured to perform the Micro LED chip leveling correction method as described in any one of the embodiments, the controller being communicatively connected to both the surface profile measuring instrument and the multidimensional adjustment mechanism.

[0012] Thirdly, this application also provides an optical inspection device, comprising: a surface profile measuring instrument for measuring surface profile data or tilt angle of a MicroLED chip surface; a multi-dimensional adjustment mechanism for supporting the MicroLED chip and providing adjustment of at least two rotation axes; and a controller for executing the MicroLED chip horizontal correction method as described above, wherein the controller is communicatively connected to the surface profile measuring instrument and the multi-dimensional adjustment mechanism respectively.

[0013] Fourthly, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the Micro LED chip horizontal calibration methods described above.

[0014] The Micro LED chip horizontal calibration method and apparatus provided by this invention have the following advantages compared with the prior art: (1) The present invention solves the nonlinear mapping problem between the sensor and the actuator by accurately converting the angle deviation into the physical compensation amount through a pre-calibrated mapping model; at the same time, by applying attenuation control during the iteration process and combining multiple exit conditions, the overshoot oscillation of the large inertia system is effectively suppressed, ensuring the rapid and stable convergence of the correction process.

[0015] (2) This invention achieves intelligent switching between single-axis independent compensation and dual-axis coupled compensation by introducing an adaptive judgment mechanism for the cross-coupling coefficient. When the inter-axis interference is small, the calculation model is simplified to improve efficiency; when the inter-axis interference is significant, the complete inverse matrix is ​​used for decoupling, eliminating the systematic residual error caused by ignoring the coupling term, and ensuring the correction accuracy of the linear model under different mechanical characteristics.

[0016] (3) This invention addresses the strong nonlinear characteristics present in mechanical and optical systems by using a reverse lookup table combined with interpolation to replace complex real-time inverse calculations. By pre-constructing a reverse mapping table of "angle → position", the target position can be quickly indexed and interpolated directly from the measured angle during online correction, avoiding the computational overhead of iteratively solving the Jacobian matrix. This significantly improves the real-time correction speed while ensuring the accuracy of nonlinear compensation. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1This is a flowchart illustrating the Micro LED chip horizontal calibration method provided by the present invention; Figure 2 This is a schematic diagram of the Micro LED chip horizontal correction device provided by the present invention; Figure 3 This is a schematic diagram of the structure of the optical detection device provided by the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0020] It should be noted that, in the description of the embodiments of the present invention, 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. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. Those skilled in the art can understand the specific meaning of the above terms in the present invention according to the specific circumstances.

[0021] Figure 1 This is a flowchart illustrating the Micro LED chip horizontal calibration method provided by the present invention, as shown below. Figure 1 As shown, this application provides a method for horizontal calibration of a Micro LED chip. This method can be applied to an optical inspection system, which includes a multi-dimensional adjustment mechanism for supporting the Micro LED chip, a surface profile measurement device for measuring the surface state of the Micro LED chip, and a controller communicatively connected to the surface profile measurement device and the multi-dimensional adjustment mechanism, thereby forming a complete closed-loop feedback control system. The method flow of this embodiment mainly includes the following steps.

[0022] It should be noted that the multi-dimensional adjustment mechanism (multi-dimensional adjustment platform) refers to a mechanical device capable of adjusting the orientation of a Micro LED chip in multiple dimensions. It includes at least two rotational axes (e.g., Rx and Ry axes) to drive the Micro LED chip to rotate around the X and Y axes, respectively. In practical applications, the multi-dimensional adjustment mechanism is typically driven by a motor and equipped with an encoder for real-time feedback of the current position of each rotational axis. By controlling the rotation angle of each axis, the multi-dimensional adjustment mechanism can precisely adjust the tilt orientation of the Micro LED chip placed on it relative to a reference plane, thereby achieving horizontal correction. It should be understood that the specific structural form of the multi-dimensional adjustment mechanism is not limited to this; any mechanism capable of providing at least two rotational degrees of freedom for orientation adjustment is acceptable.

[0023] It should also be noted that the aforementioned surface profile measurement device refers to a measuring device capable of measuring the tilt state of the Micro LED chip surface relative to a reference plane. In specific implementation, the surface profile measurement device can employ two different technical approaches: the first is to use a multi-point ranging sensor (such as a laser displacement sensor array), which arranges multiple measuring points on the Micro LED chip surface and measures the height data of each point, then calculates the tilt angle of the Micro LED chip surface through plane fitting; the second is to use precision angle measuring instruments such as a laser collimator to directly measure the absolute tilt angle of the Micro LED chip surface relative to the reference plane. Regardless of the technical approach used, the output of the surface profile measurement device is used to characterize the deviation between the Micro LED chip surface and the ideal horizontal state, providing an accurate feedback benchmark for subsequent closed-loop calibration. It should be understood that the specific type and measurement principle of the surface profile measurement device are not limited to the above examples, as long as they can obtain parameters reflecting the tilt state of the Micro LED chip surface.

[0024] Step S100: Obtain the tilt angle deviation data of the Micro LED chip surface relative to the reference plane.

[0025] Specifically, this step aims to quantify the difference between the current orientation of the Micro LED chip and its ideal horizontal state. In practice, a surface profile measurement device probes the surface of the Micro LED chip, which is placed on a multi-dimensional adjustment mechanism, and collects raw signals reflecting the surface topography or tilt angle. These raw signals are then processed and converted into numerical deviation data characterizing the degree of tilt.

[0026] It should be understood that there are various ways to obtain the tilt angle deviation data. It can be calculated by plane fitting after multi-point height measurement, or the absolute tilt angle can be directly read by a precision collimator. This embodiment does not specifically limit this, as long as the parameters reflecting the tilt state of the Micro LED chip surface relative to the reference plane can be obtained. This step provides an accurate feedback benchmark for subsequent closed-loop calibration.

[0027] Step S200: Based on the pre-calibrated tilt angle-position mapping model, convert the tilt angle deviation data into the physical position compensation amount of the multi-dimensional adjustment mechanism.

[0028] Specifically, due to complex factors such as installation errors, mechanical transmission clearances, and inter-axis cross-coupling, the measurement coordinate system of the sensor and the motion coordinate system of the multi-dimensional adjustment mechanism are not simply linearly related. Therefore, this application introduces a pre-calibrated tilt angle-position mapping model. This model is pre-stored in the form of function expressions, matrix parameters, or lookup tables. Essentially, it is a mathematical mapping relationship verified offline, accurately describing the correspondence between the "measured angle deviation" and the "mechanical shaft displacement required to eliminate the deviation." During correction, the tilt angle deviation data obtained in step S100 is input into this model, and the model outputs the corresponding physical position compensation. This conversion mechanism based on the pre-calibrated model effectively shields the nonlinear characteristics of the underlying mechanical system, allowing the upper-level control logic to focus on deviation elimination itself without resolving complex inverse mechanical kinematics equations during each correction, thus ensuring the real-time performance and accuracy of the correction.

[0029] Step S300: Drive the multi-dimensional adjustment mechanism to perform horizontal adjustment based on the physical position compensation amount, and apply attenuation control to the physical position compensation amount during the iterative adjustment process. Stop the iterative adjustment when the current tilt angle deviation data and / or the current iteration number meet the preset exit conditions.

[0030] Specifically, the calculated physical position compensation amount is sent to the multi-dimensional adjustment mechanism, driving each rotational degree of freedom to perform corresponding movements to counteract the current tilt deviation. It should be noted that the iterative adjustment process refers to a cyclical process of repeatedly executing the series of steps: "acquiring tilt angle deviation data → calculating physical position compensation amount → driving the multi-dimensional adjustment mechanism to perform horizontal adjustment." After each iteration, the tilt state of the Micro LED chip is remeasured, and adjustments are continued based on the new deviation until the preset exit condition is met.

[0031] However, in practical engineering applications, multidimensional adjustment mechanisms often have a certain moment of inertia, and there is a lag between sensor measurements and mechanical responses. If the theoretically calculated full compensation value is directly applied to the system all at once, it is very easy to cause overshoot or even continuous oscillation, resulting in the inability to converge the correction process. To address this, this embodiment introduces an attenuation control strategy in the iterative adjustment loop. That is, in each iteration, the theoretical compensation amount is not fully executed, but multiplied by a relaxation factor less than or equal to 1, so that the actual compensation amount exhibits a gradual approximation characteristic.

[0032] Meanwhile, to prevent the calibration process from getting stuck in a loop due to environmental interference or hardware malfunctions, this embodiment also includes a dual exit mechanism: on the one hand, it monitors whether the current tilt angle deviation data has fallen within the allowable accuracy threshold range; on the other hand, it monitors whether the number of iterations has reached the preset upper limit. As long as either condition is met, the system determines that the calibration process has ended and locks the current position. This design ensures that the expected leveling accuracy can be achieved under normal operating conditions, while also ensuring the safety of the equipment and the certainty of the production cycle under extreme abnormal conditions.

[0033] Through the coordinated operation of steps S100 to S300, this embodiment constructs a universal leveling correction framework that combines high precision and high stability. This framework decouples specific measurement methods from compensation algorithms, enabling the same set of control logic to adapt to different types of sensors and motion platforms. Furthermore, by employing attenuation control and multiple exit mechanisms, it solves common oscillation and divergence problems in precision leveling, providing reliable technical support for high-precision optical inspection scenarios such as micro-display devices.

[0034] Based on the above embodiments, as an optional embodiment, the Micro LED chip horizontal correction method provided by the present invention obtains tilt angle deviation data of the Micro LED chip surface relative to a reference plane, including: arranging multiple measuring points on the Micro LED chip surface, measuring the height data of each measuring point to obtain spatial surface data; filtering the height data; calculating the tilt angle of the Micro LED chip surface relative to the reference plane based on the filtered height data using least squares plane fitting to obtain the tilt angle deviation data; or, directly measuring the absolute tilt angle of the Micro LED chip surface relative to the reference plane to obtain the tilt angle deviation data.

[0035] Specifically, this application provides two parallel data acquisition paths to adapt to different hardware configurations and detection accuracy requirements. The first path is an indirect calculation method based on multi-point height measurement. For example, a surface shape measurement device (such as a laser triangulation rangefinder or a spectral confocal rangefinder) can be used to arrange multiple measurement points on the surface of the Micro LED chip according to a preset grid, such as a 3×3 or 5×5 matrix distribution, and collect the height value Z_uv point by point, thereby constructing spatial surface shape data reflecting the micro-morphology of the surface. Since the original measurement data inevitably contains environmental vibration noise or sensor inherent noise, this embodiment first filters the height data, for example, by using moving average filtering or median filtering, to obtain a smoother and more stable height value. Furthermore, in order to enhance the robustness of the algorithm to local surface anomalies (such as dust, scratches, or dirt), an outlier removal mechanism can be introduced: a removal quantity parameter N_out is set, the distance from each measurement point to the initial fitting plane is calculated, and the N_out points with the largest deviation are marked as outliers and removed, and then the remaining valid measurement points are used for subsequent calculations. When N_out is 0, no removal is performed. Based on the height data after filtering and elimination, the least squares method is used for plane fitting, and its mathematical model is as follows:

[0036] Where Z is the height of the measuring point, and x and y are the planar coordinates of the measuring point on the surface of the Micro LED chip. a , b These are the slope coefficients of the fitted plane in the X and Y directions, respectively. c Let be the intercept constant. Based on the slope coefficient obtained from the fitting, the tilt angle of the Micro LED chip surface relative to the reference plane can be calculated using the following formula:

[0037] in, Let be the tilt angle about the X-axis. Let be the tilt angle around the Y-axis. Under the condition of a small angle approximation (typically, the tilt deviation in micro-display detection is much less than 1°), the above formula can be simplified to: ≈ b, ≈ - a (Unit: radians).

[0038] As an alternative implementation, the second approach is to directly measure the absolute tilt angle of the Micro LED chip surface relative to a reference plane. In this case, a laser collimator is used for surface measurement. The laser collimator utilizes the principle of parallel light reflection to emit a collimated beam towards the Micro LED chip surface and receives the positional changes of the reflected light spot. Through internal photoelectric conversion and calculation circuitry, it directly outputs the absolute tilt angle of the surface within the beam coverage area in the Rx and Ry directions. and Compared to multi-point fitting paths, direct angle measurement paths eliminate the need for complex scanning motions and post-processing data. They offer extremely short single-measurement response times (e.g., milliseconds) and the measurement results characterize the overall average tilt angle within the light spot area. Naturally immune to interference from localized minor defects, they exhibit higher repeatability and interference resistance. Therefore, this approach is particularly suitable for high-throughput precision inspection scenarios with extremely high detection cycle requirements and where product surfaces may have localized unevenness.

[0039] It should be understood that the two data acquisition paths described above are functionally equivalent and interchangeable. In actual engineering implementation, technicians can flexibly choose one path based on specific production line speed targets, accuracy budgets, and hardware cost constraints, or integrate two types of sensors in the same device to achieve multi-mode switching, thereby maximizing the overall efficiency of the equipment while ensuring calibration accuracy.

[0040] Based on the above embodiments, as an optional embodiment, the tilt angle-position mapping model includes a linear mapping model, which includes a gain matrix and a bias vector; the gain matrix includes a main gain coefficient and a cross-coupling coefficient; the rotation axis of the multi-dimensional adjustment mechanism includes an Rx axis and a Ry axis.

[0041] Specifically, to accurately describe the correspondence between the physical motion of the multi-dimensional adjustment mechanism and the readings of the surface measurement device, this application defines a physical position vector P and a measurement angle vector Q. The physical position vector P represents the encoder feedback value or command value of each rotating axis of the multi-dimensional adjustment mechanism, and the measurement angle vector Q represents the tilt state of the Micro LED chip surface relative to the reference plane as measured by the surface measurement device. Their mathematical definitions are as follows:

[0042] In the formula, The physical position (RX_POS) of the stage as it rotates around the X-axis. The physical position (RY_POS) of the stage rotated about the Y-axis. The measured tilt angle around the X-axis (RX_AGL). The measured tilt angle around the Y-axis (RY_AGL).

[0043] Based on the above vector definition, the linear mapping model establishes a positive linear relationship from physical location to measurement angle, and its expression is:

[0044] In the formula, Here is the gain matrix. This is the bias vector. The gain matrix. and bias vector The expansion is as follows:

[0045] in, and These are the main gain coefficients, which respectively characterize the rate of change of the same-direction angle caused by a unit physical displacement of the Rx-axis and Ry-axis; and The cross-coupling coefficient characterizes the degree to which the motion of one axis interferes with the angle readings of another axis; and The bias value reflects the zero-position error of the system. The model parameters are obtained through pre-calibration, for example, by dividing the working range of each rotation axis of the multi-dimensional adjustment mechanism into multiple grid nodes, sequentially driving the mechanism to each grid node to collect data pairs of physical position and measurement angle, and solving them using the least squares method.

[0046] Furthermore, based on the pre-calibrated tilt angle-position mapping model, the tilt angle deviation data is converted into the physical position compensation amount of the multi-dimensional adjustment mechanism, including: when the ratio of the cross-coupling coefficient to the main gain coefficient is less than or equal to a preset ratio, the cross-coupling coefficient is ignored, and the physical position compensation amount corresponding to each rotation axis is calculated based on the main gain coefficient and the offset vector; when the ratio of the cross-coupling coefficient to the main gain coefficient is greater than the preset ratio, the inverse matrix of the complete gain matrix is ​​used for decoupling calculation to obtain the physical position compensation amount corresponding to each rotation axis.

[0047] Specifically, in practical systems, due to machining errors or assembly stress, there is often varying degrees of motion coupling between the Rx and Ry axes. If a uniform dual-axis decoupling or single-axis independent model is used without differentiation, it may lead to computational redundancy or compensation residuals. Therefore, this embodiment introduces coupling strength evaluation logic. First, the ratio of the cross-coupling coefficient to the main gain coefficient is calculated:

[0048] In the formula, Characterizes the relative coupling strength between Ry-axis motion and Rx-axis readings. This characterizes the relative coupling strength between the Rx-axis motion and the Ry-axis reading. The calculated ratio is then compared to a preset threshold, typically set at 5%, but can be adjusted to other values ​​depending on system accuracy requirements. and When both are less than 5%, the inter-axis coupling is determined to be weak and negligible. In this case, the model degenerates into a single-axis independent calibration model.

[0049] In this mode, the calculation of physical position compensation requires only simple division, eliminating the need for matrix inversion, significantly reducing the computational load and avoiding the risk of amplifying sensor noise due to excessively large inverse matrix elements. Conversely, if any ratio is greater than or equal to 5%, significant coupling is determined, and the complete gain matrix inverse must be used. Decoupling calculations are performed to eliminate systematic errors caused by interaxial interference.

[0050] Specifically, when the coupling is determined to be weak, the formula for calculating the physical position compensation of each rotation axis is as follows: ; in, This represents the physical compensation amount in the Rx-axis direction. This represents the physical compensation amount along the Ry axis.

[0051] Conversely, if any ratio is greater than or equal to 5%, significant coupling is determined to exist, and the complete inverse gain matrix must be used. Decoupling calculations are performed to eliminate systematic errors caused by inter-axis interference. The formula for calculating the physical position compensation is as follows: .

[0052] Based on the above embodiments, as an optional embodiment, in this embodiment, the tilt angle-position mapping model includes a nonlinear mapping model, which contains a pre-built inverse mapping table that represents the mapping relationship from the measured angle to the physical position.

[0053] Specifically, when the mechanical transmission chain of the multidimensional adjustment mechanism exhibits significant hysteresis, backlash, or geometric nonlinearity, or when the surface measurement equipment itself suffers from optical distortion in a large field of view, the global linear model described in the above embodiments may fail to control the correction residuals within an acceptable range (e.g., the residuals exceed twice the accuracy of the collimator). In this case, the nonlinear mapping model provided in this embodiment, through discretized grid interpolation, can accurately fit complex continuous mapping surfaces, thereby achieving high-precision compensation for strongly nonlinear systems. Unlike the linear model, which relies on analytical expressions, the core of this model lies in a pre-stored inverse mapping table, Map_Q2P, which directly establishes a lookup relationship between "measured angle → physical position," eliminating the need for complex real-time inverse calculations during the online correction process and significantly improving computational efficiency and response speed.

[0054] Furthermore, based on a pre-calibrated tilt angle-position mapping model, the tilt angle deviation data is converted into a physical position compensation amount for the multi-dimensional adjustment mechanism, including: determining multiple grid nodes of the corresponding grid cell in the reverse mapping table according to the tilt angle deviation data; calculating the actual physical position corresponding to the tilt angle deviation data by interpolation based on the physical positions corresponding to the multiple grid nodes; and determining the physical position compensation amount based on the difference between the target physical position corresponding to zero tilt angle and the actual physical position.

[0055] Specifically, in ( The angular space range is evenly divided. Each grid node (grid step size can be set according to system accuracy) is used to establish the measurement angle via Map_Q2P. With physical location The mapping relationship.

[0056] During online calibration, the acquired tilt angle deviation data will be used. ,in, and The tilt angle measurements are taken in two orthogonal directions, and these are used as angular coordinates. The grid cell containing the angular coordinate is then searched in the inverse mapping table Map_Q2P, which means finding the four adjacent grid nodes surrounding the angular coordinate.

[0057] Given that the measured angles and physical locations of these four grid nodes are known, the tilt angle deviation data is used. As the interpolation target point, the actual physical location corresponding to the tilt angle deviation data can be calculated using bilinear interpolation. .

[0058] Because the reverse mapping table has a zero tilt angle Corresponding target physical location Given a quantity, the difference between the target's physical location and its actual physical location is the required physical location compensation amount, calculated using the following formula:

[0059] In the formula, This represents the physical compensation amount in the Rx-axis direction. This represents the physical compensation amount along the Ry axis.

[0060] To support the generation of the aforementioned inverse mapping table, this application also provides an efficient calibration strategy. The calibration method for the nonlinear mapping model includes: a multidimensional adjustment mechanism providing adjustment for at least two rotation axes; dividing the working range of each rotation axis of the multidimensional adjustment mechanism into a grid to form multiple grid nodes; sequentially driving the multidimensional adjustment mechanism to the physical position corresponding to each grid node; collecting data pairs of physical position and measurement angle at each grid node to construct a forward mapping matrix; and converting the forward mapping matrix into an inverse mapping table representing the mapping relationship from the measurement angle to the physical position.

[0061] Specifically, the working range of the Rx and Ry axes is first evenly divided. Each grid node is sequentially instructed to move the platform to its known physical position, and the measured angle at that position is recorded to construct a forward mapping matrix of "physical position → measured angle". Subsequently, the forward data is subjected to coordinate exchange processing, that is, with the measured angle as the independent variable and the physical position as the dependent variable, a uniform grid is re-divided in the angle space, and the irregularly distributed forward sampling points are resampled into a regularly arranged inverse mapping table Map_Q2P using a scatter interpolation algorithm (such as bilinear interpolation or radial basis function interpolation).

[0062] It should be understood that although this embodiment uses bilinear interpolation as an example for explanation, in other embodiments, more advanced algorithms such as cubic spline interpolation or kriging interpolation can be used according to the accuracy requirements, as long as they can achieve angle-position mapping reconstruction based on discrete nodes.

[0063] Based on the above embodiments, as an optional embodiment, in this embodiment, attenuation control is applied to the physical position compensation amount during the iterative adjustment process, and the iterative adjustment is stopped when the current tilt angle deviation data and / or the current iteration number meet the preset exit conditions.

[0064] Specifically, the aforementioned embodiments mainly address the problem of accurately calculating the theoretical physical position compensation amount from a measurement perspective. However, in practical engineering applications, the multidimensional adjustment mechanism, as a mechanical system with mass and rotational inertia, does not exhibit ideal rigidity in its dynamic response characteristics. If the theoretically calculated full compensation value is directly applied to the drive unit at once, overshoot can easily occur due to factors such as mechanical transmission backlash, motor response lag, or sensor noise, potentially causing the system to oscillate continuously near the target position and fail to converge. Therefore, this embodiment introduces a relaxation factor-based attenuation control strategy in the closed-loop iterative circuit, coupled with a dual exit mechanism, to ensure both stability and safety during the correction process.

[0065] Regarding the specific implementation of attenuation control, the system multiplies the calculated physical position compensation amount by a preset relaxation factor to obtain the actual physical compensation amount, and drives the multi-dimensional adjustment mechanism to perform horizontal adjustment based on the actual physical compensation amount.

[0066] In this embodiment, the calculated physical location compensation amount (the calculation method can be any of those mentioned in the above embodiments) is set as follows: The final actual compensation amount applied to the stage must be adjusted by the relaxation factor. The weighted processing is calculated using the following formula:

[0067] In the formula, and These are the actual physical compensation amounts applied to the Rx and Ry axes of the stage, respectively. Let be the relaxation factor, and its value range is . In practice, The preferred value range is 0.8 to 0.9. When When the system degenerates to full compensation mode, and when When the value is between 0.8 and 0.9, it is equivalent to performing only 80% to 90% of the theoretical compensation in each iteration. The reserved margin can effectively absorb the dynamic overshoot energy of the mechanical system, so that the stage attitude can smoothly approach the target horizontal state in a progressive manner.

[0068] Regarding the dual exit mechanism, after each iteration adjustment, the system determines whether the current tilt angle deviation data is less than or equal to the preset deviation threshold, or whether the current iteration number has reached the preset maximum iteration number; if either of the above conditions is met, the iteration adjustment stops.

[0069] Specifically, the preset deviation threshold is an accuracy index set according to the depth-of-field tolerance of the optical detection system. For example, for a high-precision Micro LED chip inspection scenario, the threshold can be set to 0.01° or the corresponding Z-direction height difference of 5μm@2mm. Once the measured deviation falls within this range, the correction is determined to be qualified, and the loop is immediately terminated to save cycle time. On the other hand, the preset maximum number of iterations (for example, the default value is 3 times) is the system's safety fallback strategy. In the actual production line environment, the deviation may always fail to converge within the threshold due to reasons such as severe contamination on the product surface, abnormal loosening of the fixture, or instantaneous sensor failure. Without the iteration number limit, the system will fall into an infinite loop, causing equipment downtime or capacity loss. By setting , even in the extreme case where the correction is not fully up to standard, the system can forcibly exit after a limited number of attempts, and feed back and report the abnormal state, which facilitates subsequent manual intervention or re-judgment processing.

[0070] Figure 2 is a schematic structural diagram of the Micro LED chip horizontal correction device provided by the present invention, as shown in Figure 2 , including: A data acquisition module 210, configured to acquire tilt angle deviation data of the Micro LED chip surface relative to a reference plane; A compensation conversion module 220, configured to convert the tilt angle deviation data into a physical position compensation amount of the multi-dimensional adjustment mechanism based on a pre-calibrated tilt angle-position mapping model; A closed-loop control module 230, configured to drive the multi-dimensional adjustment mechanism to perform horizontal adjustment based on the physical position compensation amount, apply attenuation control to the physical position compensation amount during the iterative adjustment process, and stop the iterative adjustment when the current tilt angle deviation data and / or the current number of iterations satisfy a preset exit condition.

[0071] It should be noted that the Micro LED chip horizontal correction device provided by the embodiment of the present invention can execute the Micro LED chip horizontal correction method described in any of the above embodiments during specific operation, which will not be repeated in this embodiment.

[0072] Figure 3 is a schematic structural diagram of the optical detection apparatus provided by the present invention, as shown in Figure 3 , the optical detection apparatus includes a surface profile measurement device, a multi-dimensional adjustment mechanism, and a controller.

[0073] Specifically, the surface profile measuring device is used to measure the surface profile data or tilt angle of the Micro LED chip. It is mounted above or to the side of the multi-dimensional adjustment mechanism, with the optical path axis aligned with the Micro LED chip placed on the stage. In practical deployments, the surface profile measuring device can be a laser triangulation rangefinder, a spectral confocal sensor, or a laser collimator, etc., with the specific selection depending on the detection accuracy and cycle time requirements. The multi-dimensional adjustment mechanism is used to support the Micro LED chip and provide adjustment for at least two rotation axes. As the execution terminal, it directly determines the physical limit accuracy of the horizontal correction. The controller is communicatively connected to both the surface profile measuring device and the multi-dimensional adjustment mechanism to execute the Micro LED chip horizontal correction method as described in any of the foregoing embodiments.

[0074] It should be noted that the controller can be physically represented as a combination of a host computer and a measurement controller: the host computer is responsible for running the core correction algorithm, model solving and human-computer interaction, while the measurement controller is responsible for parsing the underlying motion control commands and servo driving; the two communicate with each other in real time via a high-speed bus or Ethernet.

[0075] Regarding the specific mechanical configuration of the multi-dimensional adjustment mechanism, this application provides several optional implementation methods to adapt to different application scenarios. In one implementation, the multi-dimensional adjustment mechanism adopts a serial motion platform configuration, where the Rx axis, Ry axis, and possibly X / Y / Z axes are sequentially superimposed and connected. This configuration is compact, low-cost, and easy to integrate into existing production lines, making it suitable for medium-to-high precision detection scenarios where space is limited or cost is sensitive. In another implementation, the multi-dimensional adjustment mechanism adopts a parallel motion platform configuration, such as a 6SPS Stewart platform, which connects the upper and lower flanges through six linear actuation legs, utilizing multi-axis linkage to achieve adjustment of Rx / Ry and other degrees of freedom. Compared to the serial configuration, the parallel platform has higher structural rigidity, lower moment of inertia, and better dynamic response characteristics, effectively suppressing vibration residue during high-frequency adjustment, making it particularly suitable for sub-micron ultra-high precision detection or high-acceleration high-throughput detection scenarios. It should be noted that the innovation of this invention is not that regardless of the configuration used, as long as the multi-dimensional adjustment machine meets appropriate angular resolution and repeatability accuracy, it is sufficient to ensure that the theoretical accuracy of the aforementioned correction algorithm can be reproduced by the physical system.

[0076] Through the collaborative operation of the aforementioned hardware architecture, the optical inspection equipment provided in this embodiment achieves a complete closed loop from deviation perception and algorithm calculation to precise execution. The surface measurement device provides a high-precision feedback reference, the multi-dimensional adjustment mechanism provides reliable physical correction capabilities, and the controller acts as the "brain" to organically combine the two. By executing the mapping model transformation, attenuation control, and multiple exit logic in the aforementioned embodiment, it eliminates the influence of mechanical nonlinearity and inter-axis coupling, and suppresses overshoot oscillations during dynamic adjustment. This deep integration of hardware and software design enables the equipment to stably achieve Z-axis deviation correction accuracy within 0.5μm@2mm in actual mass production environments, and compresses the single-chip correction time to less than 1 second, fully meeting the high-standard inspection requirements of micro-display devices such as Micro LED chips.

[0077] Furthermore, this application also provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, this computer program implements the Micro LED chip horizontal calibration method as described in any of the foregoing embodiments. Specifically, the storage medium can be any form of non-volatile or volatile data storage medium, such as read-only memory (ROM), random access memory (RAM), flash memory, disk, optical disk, or cloud storage server. By embedding the calibration method into a software program and storing it in the medium, this application not only protects specific hardware devices but also covers independent distribution and licensing scenarios for software products. For example, device manufacturers can upgrade the program version in the storage medium to enable older devices to obtain the latest nonlinear compensation algorithms or adaptive coupling / decoupling functions without replacing the hardware, thereby significantly extending the device's lifespan and improving the user experience. This software-level protection strategy, together with the foregoing hardware device embodiments, constitutes comprehensive and multi-dimensional protection for the technical solution of this application.

[0078] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the Micro LED chip horizontal correction method provided in the above embodiments. The method includes: acquiring tilt angle deviation data of the Micro LED chip surface relative to a reference plane; converting the tilt angle deviation data into a physical position compensation amount of a multi-dimensional adjustment mechanism based on a pre-calibrated tilt angle-position mapping model; driving the multi-dimensional adjustment mechanism to perform horizontal adjustment based on the physical position compensation amount, applying attenuation control to the physical position compensation amount during the iterative adjustment process, and stopping the iterative adjustment when the current tilt angle deviation data and / or the current iteration number meet a preset exit condition.

[0079] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for horizontal calibration of a Micro LED chip, characterized in that, include: Obtain the tilt angle deviation data of the Micro LED chip surface relative to the reference plane; Based on a pre-calibrated tilt angle-position mapping model, the tilt angle deviation data is converted into the physical position compensation amount of the multi-dimensional adjustment mechanism; The multi-dimensional adjustment mechanism is driven to perform horizontal adjustment based on the physical position compensation amount, and attenuation control is applied to the physical position compensation amount during the iterative adjustment process. The iterative adjustment is stopped when the current tilt angle deviation data and / or the current iteration number meet the preset exit conditions.

2. The Micro LED chip horizontal calibration method according to claim 1, characterized in that, Obtain the tilt angle deviation data of the Micro LED chip surface relative to the reference plane, including: Multiple measuring points are arranged on the surface of the Micro LED chip, and the height data of each measuring point is measured to obtain spatial surface shape data; the height data is filtered; based on the filtered height data, the tilt angle of the Micro LED chip surface relative to the reference plane is calculated using least squares plane fitting to obtain the tilt angle deviation data; or... The tilt angle deviation data is obtained by directly measuring the absolute tilt angle of the Micro LED chip surface relative to the reference plane.

3. The Micro LED chip horizontal calibration method according to claim 1, characterized in that, The tilt angle-position mapping model includes a linear mapping model, which contains a gain matrix and a bias vector; the gain matrix contains a main gain coefficient and a cross-coupling coefficient; the multidimensional adjustment mechanism includes at least two rotation axes, which include an Rx axis and a Ry axis. Based on a pre-calibrated tilt angle-position mapping model, the tilt angle deviation data is converted into physical position compensation amounts for the multi-dimensional adjustment mechanism, including: If the ratio of the cross-coupling coefficient to the main gain coefficient is less than or equal to a preset ratio, the cross-coupling coefficient is ignored, and the physical position compensation amount corresponding to each rotation axis is calculated based on the main gain coefficient and the offset vector. If the ratio of the cross-coupling coefficient to the main gain coefficient is greater than the preset ratio, decoupling calculation is performed using the inverse of the complete gain matrix to obtain the physical position compensation amount corresponding to each rotation axis.

4. The Micro LED chip horizontal calibration method according to claim 3, characterized in that, The calibration method for the linear mapping model includes: The working range of each rotation axis of the multidimensional adjustment mechanism is divided into a grid to form multiple grid nodes; The multi-dimensional adjustment mechanism is driven sequentially to the physical positions corresponding to each of the grid nodes, and data pairs of physical positions and measurement angles at each grid node are collected. Based on all the data pairs, the gain matrix and the bias vector are solved using the least squares method.

5. The Micro LED chip horizontal calibration method according to claim 1, characterized in that, The tilt angle-position mapping model includes a nonlinear mapping model, which contains a pre-built inverse mapping table that represents the mapping relationship from the measured angle to the physical position. Based on a pre-calibrated tilt angle-position mapping model, the tilt angle deviation data is converted into physical position compensation amounts for the multi-dimensional adjustment mechanism, including: In the reverse mapping table, multiple grid nodes corresponding to the grid cell are determined based on the tilt angle deviation data; Based on the physical locations corresponding to multiple grid nodes, the actual physical locations corresponding to the tilt angle deviation data are obtained by interpolation calculation. The physical position compensation amount is determined based on the difference between the theoretical position of the Micro LED chip corresponding to zero tilt angle and the actual physical position.

6. The Micro LED chip horizontal calibration method according to claim 5, characterized in that, The calibration method for the nonlinear mapping model includes: The multidimensional adjustment mechanism provides adjustment for at least two rotation axes, and divides the working range of each rotation axis of the multidimensional adjustment mechanism into a grid to form multiple grid nodes; The multi-dimensional adjustment mechanism is driven sequentially to the physical positions corresponding to each of the grid nodes, and data pairs of physical positions and measurement angles at each grid node are collected to construct a positive mapping matrix; The forward mapping matrix is ​​converted into the reverse mapping table, which represents the mapping relationship from the measurement angle to the physical location.

7. The Micro LED chip horizontal calibration method according to claim 5, characterized in that, During the iterative adjustment process, attenuation control is applied to the physical position compensation amount, and the iterative adjustment is stopped when the current tilt angle deviation data and / or the current iteration number meet the preset exit condition, including: The physical position compensation amount is multiplied by a preset relaxation factor to obtain the actual physical compensation amount, and the multi-dimensional adjustment mechanism is driven to perform horizontal adjustment based on the actual physical compensation amount. After each iteration adjustment, determine whether the current tilt angle deviation data is less than or equal to the preset deviation threshold, or whether the current iteration number has reached the preset maximum iteration number; If any of the above conditions are met, the iterative adjustment shall be stopped.

8. A Micro LED chip horizontal calibration device, characterized in that, include: The data acquisition module is used to acquire the tilt angle deviation data of the Micro LED chip surface relative to the reference plane; The compensation conversion module is used to convert the tilt angle deviation data into the physical position compensation amount of the multi-dimensional adjustment mechanism based on a pre-calibrated tilt angle-position mapping model. The closed-loop control module is used to drive the multi-dimensional adjustment mechanism to perform horizontal adjustment based on the physical position compensation amount, and to apply attenuation control to the physical position compensation amount during the iterative adjustment process, and to stop the iterative adjustment when the current tilt angle deviation data and / or the current iteration number meet the preset exit conditions.

9. An optical inspection device, characterized in that, include: Surface profile measuring equipment is used to measure the surface profile data or tilt angle of Micro LED chips; A multi-dimensional adjustment mechanism for supporting Micro LED chips, the multi-dimensional adjustment mechanism including at least two rotation axes; A controller is configured to perform the Micro LED chip horizontal calibration method as described in any one of claims 1 to 7, wherein the controller is communicatively connected to the surface measurement device and the multidimensional adjustment mechanism, respectively.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the Micro LED chip horizontal correction method as described in any one of claims 1 to 7.

Images