A mosaic mirror boundary change detection apparatus and method

Abstract

This invention discloses a device and method for detecting boundary changes in a video wall splicing mirror, relating to the field of optical inspection technology. The detection device includes a contact sensing structure, a non-contact optical structure, and a processor arranged relative to the video wall splicing mirror body. The detection of boundary changes in the video wall splicing mirror is achieved based on the contact sensing structure and the non-contact optical structure. Furthermore, when using the contact sensing structure, the edge sensor readings are corrected by introducing a mounting arm length parameter, eliminating Abbe errors caused by the misalignment of the mounting reference plane and the optical neutral plane, thus improving the accuracy and reliability of the contact detection. An event sensor is introduced into the non-contact optical structure; this event sensor can typically capture transient events of edge changes with microsecond-level time resolution, making it suitable for real-time detection in dynamic environments. In addition, the contact sensing structure and the non-contact optical structure can serve as backups for each other, improving the reliability of the detection device.

Description

Technical Field

[0001] This invention relates to the field of optical inspection technology, and in particular to a device and method for detecting changes in the boundary of a splicing mirror. Background Technology

[0002] With the development of technologies such as deep space exploration and space optics, and the requirements of the Ruili criterion, optical systems now generally require larger telescope sizes to achieve stronger overall optical performance. However, due to limitations in manufacturing processes and material properties, directly manufacturing a single ultra-large aperture mirror surface faces extremely high technical difficulties and costs. Large-aperture spliced ​​plane mirrors and co-phase beam combining systems, by combining multiple sub-mirrors to obtain a larger effective light-gathering area, have become a key approach to solving this aperture challenge. To ensure that the optical performance of multiple spliced ​​sub-mirrors is completely equivalent to that of a single large-aperture mirror, strict co-phase control between the sub-mirrors must be achieved. This requires controlling the root mean square error of the surface splicing to the tens of nanometers or even sub-nanometer levels.

[0003] In actual telescope field operations, large-aperture mosaicking systems face complex environmental disturbances. These include, but are not limited to, changes in the system's working attitude, large-scale zenith angle and rotational movements, as well as changes in ambient temperature gradients due to day-night cycles and dynamic wind loads. All of these factors inevitably cause gravitational deformation and heat release in the overall support truss of the system. These macroscopic deformations, when transmitted to each sub-mirror unit, result in multi-degree-of-freedom relative displacements between adjacent sub-mirrors, primarily including relative translation (piston) and relative tilting (tip / tilt) around the X / Y axes. Therefore, it is necessary to monitor the boundary state changes of the mosaicking mirrors in real time to provide a basis for subsequent conformal mirror correction and maintaining overall imaging accuracy.

[0004] Therefore, how to detect changes in the boundary of splicing mirrors is a technical problem that urgently needs to be solved by those in this field. Summary of the Invention

[0005] The purpose of this invention is to provide a device and method for detecting changes in the boundary of a mosaic mirror, so as to solve the technical problem of detecting changes in the boundary of a mosaic mirror.

[0006] To solve the above-mentioned technical problems, the present invention provides a device for detecting changes in the boundary of a splicing mirror, comprising: a processor, and a contact sensing and detection structure and a non-contact optical structure arranged relative to the splicing mirror body; The contact sensing and detection structure includes an edge sensor, which is positioned at the seam between adjacent sub-mirrors. The non-contact optical structure includes at least a beam splitter element arranged sequentially along the beam transmission direction, an image acquisition component, and / or an event sensor for capturing events of changes in the splicing mirror boundary. The processor is connected to the edge sensor, the image acquisition component, and / or the event sensor, and is used to detect changes in the splicing mirror boundary based on the data acquired by the edge sensor and the sensor mounting arm length parameter, and to detect changes in the splicing mirror boundary based on the data acquired by the image acquisition component and / or the data acquired by the event sensor.

[0007] Preferably, the edge sensor includes multiple sets of sensing probes, which are mounted on a mounting base and arranged in an array. The sensing probes adopt a differential structure with upward and downward opposing beams. The processor is used to detect changes in the edge of the splicing mirror based on the data collected by the edge sensor and the sensor mounting arm length parameter, including: Displacement readings are collected by the multiple sets of sensor probes. Based on the installation position of the multiple sets of sensor probes in the local coordinate system of the sensor, and combined with the small-angle rigid body kinematic approximation when adjacent sub-mirrors undergo relative translation and relative tilt around the local coordinate axis, a positive kinematic mapping relationship between displacement readings and spatial pose is established. Based on the aforementioned positive kinematic mapping relationship, a matrix for symmetric difference operations is constructed; The matrix obtained by the symmetric difference operation is used to project the displacement readings collected by the multiple sets of sensor probes onto the orthogonal spatial degrees of freedom, and output the absolute translation and relative tilt of the adjacent sub-mirrors. The changes in the splicing mirror boundary are detected based on the absolute translation, the relative tilt, and the sensor mounting arm length parameter.

[0008] Preferably, the processor is used to determine the sensor mounting arm length parameter including: The output data of each sensor probe is collected as the multiple sets of sensor probes rotate around the axis at different positions of the test device. Based on the output data of each sensor probe, positions whose output values ​​do not change with rotation are selected. The geometric parameters corresponding to the determined position are used as the length parameters of the sensor mounting arm.

[0009] Preferably, the processor detects changes in the mosaic mirror boundary based on the absolute translation, the relative tilt, and the sensor mounting arm length parameter, including: Based on the absolute translation, the relative tilt, and the sensor mounting arm length parameters, an error compensation model is constructed to correct the edge sensor readings and obtain the corrected sensor output value. The corrected sensor output value is decomposed into multiple components according to the dimension; among them, the equation corresponding to one set of components includes the coupling compensation term of the mounting arm length parameter to the tilt amount, and the equation corresponding to the other components is constructed based on the ideal reading model without the influence of the mounting arm length parameter. By combining the equations corresponding to each component, a system of equations containing different error compensation terms is formed; The actual translation and tilt states of each sub-mirror are decoupled and deduced from the system of equations containing different error compensation terms.

[0010] Preferably, the non-contact optical structure includes a first optical detection structure; The first optical detection structure includes a light source, a beam splitter, an imaging lens, and a camera. The camera is positioned equidistantly from the front and back focal lengths of the imaging lens. The light beam emitted by the light source passes through the beam splitter and illuminates the splicing mirror. The light beam reflected by the splicing mirror is incident on the beam splitter in the opposite direction and is reflected and deflected by the beam splitter to the imaging lens. After passing through the imaging lens, it is transmitted to the camera. The processor is used to receive the light intensity image modulated by the pose of the sub-mirror acquired by the camera, and calculate the wavefront phase based on the light intensity image to complete the detection of changes in the splicing mirror boundary.

[0011] Preferably, the processor is used to receive a light intensity image modulated by the pose of the receiver lens acquired by the camera, and to calculate the wavefront phase based on the light intensity image, including: Images of defocused light intensity at equidistant positions before and after the focal point are acquired. The energy distribution characteristics of the image points are extracted and analyzed, and the centroid offset of the image point distribution and the gradient of image contrast decrease are calculated. The macroscopic alignment error and boundary curvature change of the system are evaluated based on the centroid offset of the image point distribution and the gradient of image contrast descent, so as to initially screen out sub-mirror regions with misalignment problems. For sub-mirror regions with misalignment issues, based on the light intensity transmission equation and under the condition that the defocus amount is less than the preset defocus amount, the differential relationship between the axial derivative of the light intensity along the optical axis and the wavefront phase is simplified to a curvature-sensing Poisson equation; wherein, the light intensity difference term on the right side of the curvature-sensing Poisson equation is calculated from the acquired defocused light intensity image. The wavefront phase to be determined is represented as a linear combination of orthogonal basis functions of a preset order, and substituted into the curvature sensing Poisson equation to establish the correspondence between the wavefront coefficient and the light intensity difference term. Based on the aforementioned correspondence, the spatial domain signal is mapped to the frequency domain using the Fast Fourier Transform, and a discrete wavefront solution matrix equation is constructed. The inverse operation of the matrix equation is performed to calculate the wavefront coefficient vector, thereby reconstructing the wavefront phase.

[0012] Preferably, the non-contact optical structure includes a second optical detection structure; The second optical detection structure includes a light source, a beam splitter, an imaging lens, and the event sensor. The light beam emitted by the light source passes through the beam splitter and illuminates the splicing mirror. The light beam reflected by the splicing mirror is incident on the beam splitter in the opposite direction and is reflected and deflected by the beam splitter to the imaging lens. After passing through the imaging lens, it is transmitted to the event sensor. The processor is used to receive event data collected by the event sensor; and to extract dynamic variation features of the splicing mirror boundary based on the event data in order to detect changes in the splicing mirror boundary.

[0013] Preferably, the non-contact optical structure includes a third optical detection structure; The third optical detection structure includes a light source, a beam splitter, a semi-transparent mirror, an imaging mirror group, and a camera; the semi-transparent mirror is arranged on the conjugate optical path of the splicing mirror. The light beam emitted by the light source passes through the beam splitter and illuminates the splicing mirror. The light beam reflected by the splicing mirror is then incident on the semi-transparent and semi-reflective mirror. The light beam transmitted from the semi-transparent and semi-reflective mirror passes through the imaging lens group and enters the camera. The processor is used to receive light intensity images captured by the camera and to detect changes in the boundary of the stitching mirror based on the light intensity images.

[0014] Preferably, the semi-transparent and semi-reflective mirror is mounted on a piezoelectric ceramic; The processor is used to receive light intensity images captured by the camera, detect changes in the boundary of the splicing mirror based on the light intensity images, and control the piezoelectric ceramic to adjust its attitude and position according to the boundary change detection results.

[0015] Preferably, after detecting changes in the boundary of the splicing mirror, the processor is further configured to: The original electrical signals output by the multiple sets of sensor probes are acquired and converted to obtain the basic distance reading. A linear mapping matrix model between the distance reading and the local spatial pose state of the sub-mirror is established by combining the local coordinate distribution of the probe. The equations corresponding to the linear mapping matrix model are solved by the least squares method to extract the optimal local pose state. After obtaining the optimal local pose state, the Abbe error caused by the height difference between the measurement point and the neutral plane of the sub-mirror is compensated, and the geometric coupling error is eliminated to obtain the true edge sensor translation reading. The readings of all seam nodes on the entire mirror surface are combined into a global measurement feedback column vector; The camera in the first optical detection structure acquires defocused light intensity images, and the light intensity transmission equation is solved to obtain the true global pose state vector of each sub-mirror. In calibration mode, a preset disturbance is injected, and multiple sets of data are collected to form paired data in which the sensor feedback measurement value and the state true value correspond one-to-one; wherein, each set of data is composed of the global measurement feedback column vector and the global pose state vector. A pre-defined linear mapping equation is solved using a regularization algorithm to obtain the global sensitivity calibration matrix of the sensor. The linear mapping equation represents the relationship between the product of the global sensitivity calibration matrix and the global pose state vector and the global measurement feedback column vector. Based on the global sensitivity calibration matrix, and combined with the tilt and displacement state calculation results of each sub-mirror, control and adjustment commands for the actuator are generated to regulate the attitude of the sub-mirrors.

[0016] Preferably, the processor is further configured to: The electronic timing statistics characteristics are integrated with the optical energy distribution characteristics; Using the fused data, a joint anomaly observation matrix is ​​constructed within a synchronous sliding time window; Based on the joint anomaly observation matrix, dimensionality reduction and fault detection are performed using principal component analysis. The construction of the joint anomaly observation matrix within the synchronous sliding time window includes: Within the synchronous sliding time window, collect the working data of all edge sensors, calculate the deviation between the actual detection reading and the theoretical prediction reading of each edge sensor, and form electronic state characteristics based on various deviations. Acquire specular optical imaging spot images, and extract spot center moment, centroid offset, and optical bounding energy range parameters to form optical energy distribution characteristics; The electronic state characteristics and optical energy distribution characteristics at the same time point are time-matched and spliced ​​together to form the joint features corresponding to a single sampling. The joint features of multiple consecutive time points are continuously collected and stacked sequentially to construct the joint anomaly observation matrix.

[0017] Preferably, the processor is used for dimensionality reduction and fault detection based on the joint anomaly observation matrix and through principal component analysis, including: Calculate the covariance matrix of the preprocessed joint variable observation matrix; The covariance matrix is ​​subjected to eigenvalue decomposition to obtain the corresponding eigenvalues ​​and orthogonal eigenvectors in sequence. The corresponding contribution rate is calculated based on each feature value, and then sorted from high to low contribution rate. Extract the first preset number of principal components whose cumulative contribution rate exceeds a preset threshold to form a projection matrix; Based on the projection matrix, the preprocessed joint variation observation matrix is ​​projected onto the principal component space to obtain the score matrix; If an anomaly is detected in the principal component score, the orthogonal feature vector corresponding to the principal component is retrieved, and the dominant feature dimension with the largest absolute value among the orthogonal feature vectors is extracted to determine the dominant influencing factor that caused the anomaly.

[0018] To address the aforementioned technical problems, the present invention also provides a method for detecting changes in the boundary of a splicing mirror, applied to the aforementioned device for detecting changes in the boundary of a splicing mirror, the method comprising: Acquire data from edge sensors, sensor mounting arm length parameters, image acquisition components, and / or event sensors; The changes in the splicing mirror boundary are detected based on the data collected by the edge sensor and the sensor mounting arm length parameter, as well as based on the data collected by the image acquisition component and / or the event sensor.

[0019] The splicing mirror boundary change detection device provided by the present invention includes: a contact sensing and detection structure arranged relative to the splicing mirror body, a non-contact optical structure, and a processor; the contact sensing and detection structure includes an edge sensor, which is arranged at the splicing seam of adjacent sub-mirrors; the non-contact optical structure includes at least a beam splitting element arranged sequentially along the beam transmission direction, an image acquisition component, and / or an event sensor for capturing splicing mirror boundary change events. The processor connects to edge sensors, image acquisition components, and / or event sensors, enabling it to detect changes in the splicing mirror boundary based on data acquired by the edge sensors and sensor mounting arm length parameters, as well as data acquired by the image acquisition components and / or event sensors. This achieves the detection of changes in the splicing mirror boundary. Furthermore, when using a contact-based sensing structure, the introduction of mounting arm length parameters to correct edge sensor readings effectively separates the translational and tilt components coupled to the sensor readings due to mirror tilt, eliminating Abbe errors caused by the misalignment of the mounting reference plane and the optical neutral plane. This ensures a true reflection of the actual displacement changes at the splicing mirror edge, thereby improving the accuracy and reliability of contact-based detection. The non-contact optical structure incorporates an event sensor, which typically captures transient events of edge changes with microsecond-level time resolution, suitable for real-time detection in dynamic environments. Additionally, the contact-based sensing structure and the non-contact optical structure can serve as backups for each other. When the performance of one detection method deteriorates due to environmental interference, the other can still maintain basic detection functions, improving the reliability of the device in harsh outdoor environments.

[0020] In addition, the present invention also provides a method for detecting changes in the boundary of a splicing mirror, which has the same or corresponding technical features as the splicing mirror boundary change detection device mentioned above, and has the same effect. Attached Figure Description

[0021] To more clearly illustrate the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a schematic diagram of a splicing mirror boundary change detection device provided in an embodiment of the present invention; Figure 2 A schematic diagram of a detection structure for changes in the boundary of a splicing mirror provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of an edge heterogeneous sensing architecture based on an edge sensor, provided in an embodiment of the present invention. Figure 4 A partial schematic diagram of a third optical detection structure based on conjugate optics provided in an embodiment of the present invention; Figure 5 This is a flowchart of a method for detecting changes in the boundary of a splicing mirror, provided in an embodiment of the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of the present invention.

[0024] The core of this invention is to provide a device and method for detecting changes in the boundary of a mosaic mirror, so as to solve the technical problem of detecting changes in the boundary of a mosaic mirror.

[0025] To enable those skilled in the art to better understand the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Figure 1 This is a schematic diagram of a splicing mirror boundary change detection device provided in an embodiment of the present invention, as shown below. Figure 1 As shown, it includes: a processor 4, and a contact sensing and detection structure 2 and a non-contact optical structure 3 arranged relative to the body of the splicing mirror 1.

[0026] The contact sensing and detection structure 2 includes an edge sensor, which is located at the seam between adjacent sub-mirrors. The non-contact optical structure 3 includes at least a beam splitting element arranged sequentially along the beam transmission direction, an image acquisition component, and / or an event sensor for capturing events of changes in the splicing mirror boundary. The processor 4 is connected to the edge sensor, the image acquisition unit and / or the event sensor, and is used to detect changes in the boundary of the splicing mirror 1 based on the data acquired by the edge sensor and the sensor mounting arm length parameter, and to detect changes in the boundary of the splicing mirror 1 based on the data acquired by the image acquisition unit and / or the data acquired by the event sensor.

[0027] This invention can detect changes in the boundary of a video wall frame based on a contact-type sensing detection structure, or it can detect changes in the boundary of a video wall frame based on a non-contact optical structure. The non-contact optical structure can be one or more of the following: a first optical detection structure for wavefront phase detection, a second optical detection structure for detecting changes in the boundary of the video wall frame, and a third optical detection structure based on conjugate optics. The following description uses a contact-type sensing detection structure combined with a first optical detection structure as an example to illustrate the detection structure for changes in the boundary of the video wall frame. Figure 2 This is a schematic diagram of a detection structure for changes in the boundary of a splicing mirror provided in an embodiment of the present invention, as shown below. Figure 2 As shown, the system includes a splicing mirror 1, a contact sensing and detection structure 2 and a first optical detection structure 30 arranged relative to the body of the splicing mirror 1, and a multi-degree-of-freedom relative pose simulation mechanism. The contact sensing and detection structure 2 includes an edge sensor 201, which is a high-dimensional edge sensor constructed from low-dimensional ranging elements (sensing probes).

[0028] An edge sensor is installed at the seam between the two mirror blocks. The first optical detection structure includes a light source 301, a beam splitter 302, an imaging lens 303, and a camera 304. The camera 304 is positioned equidistantly before and after the focal length of the imaging lens 303. The light beam emitted by the light source 301 passes through the beam splitter 302 and illuminates the splicing mirror 1. The light beam reflected by the splicing mirror 1 is incident back onto the beam splitter 302 and deflected by the beam splitter 302 to the imaging lens 303, and then transmitted to the camera 304. The beam splitter 302 can be a beam splitter mirror. In addition, a collimating lens 305 can be added between the light source 301 and the beam splitter 302 to collimate the light beam. The pose simulation mechanism 5 is a micro-displacement stage made of six-degree-of-freedom piezoelectric ceramic. The micro-displacement stage drives the movement of the splicing mirror to simulate the relative translation (piston) and tilt (tip / tilt) between actual adjacent sub-mirrors.

[0029] The following section details the detection of changes in the splicing mirror boundary based on a contact-type sensing detection structure. To obtain the multi-degree-of-freedom relative displacement between adjacent sub-mirrors, this invention proposes an edge variation sensing architecture based on an edge sensor. This architecture constructs a high-dimensional edge sensor from low-dimensional ranging elements (sensing probes). Specifically, the edge sensor includes multiple sets of sensing probes, which are mounted on a mounting base and arranged in an array. The sensing probes employ a differential structure with upward and downward opposing beams. The number of sets of sensing probes is not limited and is determined based on the actual situation.

[0030] Existing single-pass calibration platforms often focus only on translational response in a single dimension. However, in actual large-aperture splicing systems, sub-mirrors undergo not only translation but also complex relative tilting due to variations in zenith angle, rotation angle, and temperature. Since the mounting reference plane of the edge sensor is typically located on the back or side of the sub-mirror, its mounting position does not coincide with the optical reflection neutral plane of the sub-mirror, resulting in an objective "lever arm" distance. When tilt disturbances occur, this lever effect introduces significant Abbe error, causing sensor readings to be mixed with translational and tilt coupling signals. The high-order nonlinear error of this coupling signal cannot be extracted by existing single-dimensional linear calibration methods, leading to calibration matrix distortion. In other words, existing single-pass calibration platforms suffer from neglecting Abbe error and difficulty in separating multi-degree-of-freedom coupling crosstalk. Therefore, in this invention, when detecting changes in the splicing mirror boundary based on a contact-type sensing detection structure, the detection is performed using data collected by the edge sensor and the sensor mounting arm length parameter.

[0031] In some embodiments, the processor detects changes in the splicing mirror boundary based on data collected by the edge sensor and the sensor mounting arm length parameter, including: Displacement readings are collected by multiple sets of sensor probes. Based on the installation positions of multiple sets of sensor probes in the local coordinate system of the sensor, and combined with the small-angle rigid body kinematic approximation when adjacent sub-mirrors undergo relative translation and relative tilt around the local coordinate axis, a positive kinematic mapping relationship between displacement readings and spatial pose is established. Based on the positive kinematic mapping relationship, a matrix for symmetric difference operations is constructed; The matrix obtained by symmetric difference operation is used to project the displacement readings collected by multiple sets of sensor probes to the orthogonal spatial degrees of freedom, and output the absolute translation and relative tilt of adjacent sub-mirrors. Changes in the splicing mirror boundary are detected based on parameters such as absolute translation, relative tilt, and sensor mounting arm length.

[0032] To accurately determine the sensor mounting arm length parameters, in some embodiments, the processor determines the sensor mounting arm length parameters by including: The output data of each sensor probe is collected as the test device rotates around the axis at different positions. Based on the output data of each sensor probe, positions whose output values ​​do not change with rotation are selected. The geometric parameters corresponding to the determined position are used as the sensor mounting arm length parameters.

[0033] In addition, the processor detects changes in the mosaic mirror boundary based on parameters such as absolute translation, relative tilt, and sensor mounting arm length, including: Based on the parameters of absolute translation, relative tilt and sensor mounting arm length, an error compensation model is constructed to correct the edge sensor readings and obtain the corrected sensor output value. The corrected sensor output value is decomposed into multiple components according to the dimension; among them, the equation corresponding to one set of components includes the coupling compensation term of the mounting arm length parameter to the tilt amount, and the equation corresponding to the other components is constructed based on the ideal reading model without the influence of the mounting arm length parameter. By combining the equations corresponding to each component, a system of equations containing different error compensation terms is formed; The actual translation and tilt states of each sub-mirror are decoupled and deduced from the system of equations containing different error compensation terms.

[0034] To enable those skilled in the art to better understand the process of detecting changes in the boundary of a splicing mirror based on a contact-type sensing detection structure, the following explanation uses an edge sensor comprising four sets of sensing probes as an example. Figure 3 This is a schematic diagram of an edge heterogeneous sensing architecture based on an edge sensor, provided as an embodiment of the present invention. Figure 3 As shown, the edge sensor 201 in this edge heterogeneous sensing architecture is a high-dimensional edge sensor built based on low-dimensional edge sensors (i.e., multiple sensing probes). Figure 3 In this edge sensor 201, four pairs of differential displacement probes are arranged in a rectangular array on the mounting base. The use of a top-and-bottom opposing differential structure effectively suppresses thermal drift interference caused by drastic environmental changes. Let the mounting positions of the four pairs of differential sensors in the sensor's local coordinate system be (…). In order, they are respectively ,in, The semi-lateral spacing of the probe array. This represents the semi-longitudinal spacing of the probe array.

[0035] Suppose that the displacement readings acquired in real time by the four pairs of low-dimensional differential probes are respectively When adjacent sub-mirrors undergo relative translation (Piston, denoted as a local small displacement), And the relative tilt about the local coordinate axis (Tip / Tilt, denoted as local small deflection angles respectively). (A small local deflection angle on the X-axis) and When there is a small local deflection angle on the Y-axis, based on the small-angle rigid body kinematics approximation, a low-dimensional measurement reading is obtained. The positive kinematic mapping relationship with the pose in high-dimensional space is as follows: ; By constructing a symmetric difference matrix, four pairs of low-dimensional data can be projected onto orthogonal space degrees of freedom, achieving inverse high- and low-dimensional solutions. The solution formula is as follows: ; ; ; Using the aforementioned high- and low-dimensional solution models, the high-dimensional edge sensor can output the absolute translation of adjacent sub-mirrors in real time. and relative tilt and The high-dimensional solution result will serve as a direct input prerequisite for the subsequent multi-degree-of-freedom Abbe error decoupling matrix (leverage effect compensation), providing underlying multi-dimensional data support for accurately removing complex pose crosstalk.

[0036] Let the position of the edge sensor be... Under ideal conditions where installation errors are not considered, the readings of the edge sensor This can be expressed as the phase difference between adjacent sub-mirrors at that point: ; in, Indicates the wavelength of the light source, ( ) indicates the first The center coordinates of the horn mirror. ) represents the adjacent first The center coordinates of the horn mirror. Indicates the adjacent first The relative translation amount (piston) of the horn mirror. Indicates the adjacent first The relative translation amount of the horn mirror. Indicates the first The surface of the horn mirror Residual higher-order aberrations (such as Zernike coefficients). Indicates the first The surface of the horn mirror Residual higher-order aberrations. Indicates the first The tilt angle of the horn mirror around the X-axis, Indicates the first The tilt angle of the horn mirror around the Y-axis; Indicates the first The tilt angle of the horn mirror around the X-axis, Indicates the first The tilt angle of the horn mirror around the Y-axis.

[0037] However, due to the limitations of sensor mounting structures in large-aperture splicing systems, reference edge sensors typically need to extend into the sub-mirror seam or be mounted on the back of the sub-mirror, resulting in a significant height difference between the sensor's measurement point and the sub-mirror's optical reflection neutral plane. When the sub-mirror tilts (Tip / Tilt), this height difference acts like a lever, amplifying the tilt angle deflection and coupling it into a translational displacement error in the sensor reading, i.e., the lever effect or Abbe error. The mounting arm length parameter is introduced. and ,in, This represents the physical height difference in the X dimension between the actual measurement point of the edge sensor and the neutral surface of the sub-mirror. This represents the physical height difference in the Y dimension between the actual measurement point of the edge sensor and the neutral surface of the sub-mirror.

[0038] Actual high-order edge sensor readings Revised to: ; Substituting these values ​​into the matrix and rearranging, we obtain the final calibration model, which includes multi-degree-of-freedom coupling error compensation: ; This model allows for the calibration of different types of edge sensors using a high-precision displacement stage, effectively separating the additional tilt and pure translation displacement caused by the change in the zenith angle of the sub-mirror.

[0039] The corrected sensor output value is decomposed into multiple components according to dimensions; one set of components (i.e. ,express The equation corresponding to the first element in the equation contains a coupling compensation term for the installation arm length parameter and the tilt amount: ; The remaining components ( express The vector formed by the second and third elements in the vector is a three-dimensional vector. The corresponding equation is constructed based on an ideal reading model that does not take into account the influence of the mounting arm length parameter. ; By combining the equations corresponding to each component, a system of equations containing different error compensation terms is formed. That is, by simultaneously solving the above equations... expressions and From the expression, we can derive the following system of equations: ; in, express The second element in express The third element in. correspond , correspond , correspond That is, based on the decoupling of the system of equations containing different error compensation terms, the actual translation and tilt states of each sub-mirror are deduced.

[0040] This means that the detection of changes in the splicing mirror boundary was achieved based on a contact-based sensing structure. The following section explains the process of detecting changes in the splicing mirror boundary using a non-contact optical structure.

[0041] Existing sensor calibration requires displacement gauges and interferometers as references. These devices are bulky and often only allow for localized calibration on separate test platforms, making effective integration with the actual optical path system of the spliced ​​mirrors impossible. This results in the calibration process being detached from the actual wavefront control system. If new interference factors are introduced into the optical system, such as residual stress generated during installation, the original offline calibration data will become inaccurate, lacking an integrated absolute optical measurement benchmark. Therefore, this invention provides an absolute calibration truth value for pose by employing high-precision wavefront curvature sensing technology (i.e., the first optical detection structure) as the core optical feedback link. By acquiring the light field energy distribution of the defocused surface, a dual-path evaluation mechanism combining an image point distribution-based evaluation method and a frequency domain-based evaluation method is introduced to establish a physical mapping relationship between light intensity changes and wavefront phase.

[0042] The following describes the process of detecting changes in the splicing mirror boundary based on the first optical detection structure. Please refer to the description of the first optical detection structure above; it will not be repeated here. The processor receives the light intensity image modulated by the pose of the sub-mirrors from the camera in the first optical detection structure, and calculates the wavefront phase based on the light intensity image to complete the detection of changes in the splicing mirror boundary.

[0043] In some embodiments, the processor is configured to receive a light intensity image modulated by the pose of the receiver mirror acquired by the camera, and to calculate the wavefront phase based on the light intensity image, including: Images of defocused light intensity at equidistant positions before and after the focal point are acquired. The energy distribution characteristics of the image points are extracted and analyzed, and the centroid offset of the image point distribution and the gradient of image contrast decrease are calculated. The macroscopic alignment error and boundary curvature change of the system are evaluated based on the centroid offset of the image point distribution and the gradient of image contrast descent, so as to initially screen out sub-mirror regions with misalignment problems. For sub-mirror regions with misalignment issues, based on the light intensity transmission equation and under the condition that the defocus amount is less than the preset defocus amount, the differential relationship between the axial derivative of the light intensity along the optical axis and the wavefront phase is simplified to a curvature-sensing Poisson equation; wherein, the light intensity difference term on the right side of the curvature-sensing Poisson equation is calculated from the acquired defocused light intensity image. The wavefront phase to be determined is represented as a linear combination of orthogonal basis functions of a preset order, and then substituted into the curvature sensing Poisson equation to establish the correspondence between the wavefront coefficient and the light intensity difference term. Based on the correspondence, the spatial domain signal is mapped to the frequency domain by the fast Fourier transform, and a discrete wavefront solution matrix equation is constructed. The wavefront coefficient vector is calculated by inverting the matrix equation to reconstruct the wavefront phase.

[0044] To enable those skilled in the art to better understand the process of detecting changes in the boundary of a splicing mirror based on a first optical detection structure provided by the present invention, the process will be described again below with reference to specific embodiments.

[0045] First, a macroscopic state assessment is conducted using an evaluation method based on image point distribution.

[0046] Let the phase of a single sub-mirror be... ,in Number the sub-mirrors, ( () represents the spatial coordinates of the pupil plane. As a central mirror, The unit is the imaginary unit. The complex optical field of a single sub-mirror is shown below: ; in, The light field intensity mask is set to 1 inside the element pupil and 0 outside. The total complex optical field of the block-mounted mirror is: ; express The total complex optical field at the pupil plane of the modular mirror. This indicates the defocus distance along the optical axis of the light beam.

[0047] Block splicing mirror light intensity Represented as: ; in, for The complex conjugate of the total complex optical field of the block-mounted mirror. By establishing Translation of sub-mirror ,tilt The relationship can establish a mapping relationship between the light intensity of the defocused star point image and the confocal and misaligned states of the sub-mirrors.

[0048] The system acquires data at equal distances before and after the focal point (defocusing amount is...). The in-focus light intensity image at point () is denoted as . The light intensity of the defocused light intensity image after focusing is denoted as The system extracts and analyzes the energy distribution characteristics of the acquired image points. When the stitching sub-mirrors experience macroscopic translational or tilting misalignment, the image point distribution will exhibit obvious morphological distortions (such as spot splitting, symmetry breaking, or local energy diffusion). By calculating the centroid offset of the image point distribution and the image contrast reduction gradient, the system can quickly assess the macroscopic alignment error and boundary curvature changes of the system, providing transient feature feedback for fault diagnosis and early warning.

[0049] Secondly, a frequency-domain evaluation method is employed for high-precision absolute phase reconstruction. Based on the image point distribution assessment, the system switches to a frequency-domain evaluation method for quantitative reconstruction of the wavefront phase. According to the light intensity transmission equation, the axial derivative of the light intensity along the optical axis is related to the wavefront phase. The following differential relation is satisfied: ; In defocus Under smaller conditions, the above equation can be expanded into the form of the Laplace operator multiplied by the light intensity gradient. Under the uniform illumination approximation of a large-aperture splicing system, neglecting the abrupt changes in light intensity gradient caused by edge effects, the formula can be simplified to the curvature sensing Poisson equation: ; in, This represents the average light intensity distribution across the pupil.

[0050] The wavefront phase to be determined Represented as orthogonal basis functions Linear combination: ; Substituting into the Poisson equation, we get: ; To achieve rapid computer-aided solution, the system maps the aforementioned spatial domain signals to the frequency domain using techniques such as Fast Fourier Transform (FFT), constructing discrete wavefront solution matrix equations. Let the number of pixels acquired by the camera (CCD) be... The frequency domain evaluation and solution matrix is ​​represented as follows: ; By using matrix inversion operations based on frequency domain evaluation, the system not only effectively avoids the low-frequency divergence problem inherent in direct spatial domain integration, but also accurately calculates the wavefront coefficient column vector. The final high-precision reconstructed wavefront phase As an absolute optical standard.

[0051] In other words, the detection of changes in the splicing mirror boundary was achieved through the first optical detection structure. The process of detecting changes in the splicing mirror boundary based on the second optical detection structure is explained below.

[0052] The second optical detection structure includes a light source, a beam splitter, an imaging lens, and an event sensor. The light beam emitted by the light source passes through the beam splitter and illuminates the stitching mirror. The beam reflected by the stitching mirror is reflected back to the beam splitter, and then reflected and deflected by the beam splitter to the imaging lens. After passing through the imaging lens, it is transmitted to the event sensor. The only structural difference between the second and first optical detection structures is that the camera in the first structure is replaced by an event sensor.

[0053] The processor is used to receive event data collected by the event sensor; and to extract dynamic variation features of the splicing mirror boundary based on the event data in order to detect changes in the splicing mirror boundary.

[0054] The processor receives event data collected by the event sensor, analyzes and processes the acquired event data, extracts the dynamic change features generated at the boundary of the splicing mirror, and relies on the extracted boundary change features to complete the identification and monitoring of the boundary changes of the splicing mirror, thereby achieving effective detection of the boundary changes of the splicing mirror.

[0055] The process of detecting changes in the boundary of the splicing mirror based on the third optical detection structure is explained below. Figure 4 This is a partial schematic diagram of a third optical detection structure based on conjugate optics provided in an embodiment of the present invention. The third optical detection structure includes a light source 301, a beam splitter 302, a semi-transparent mirror 310, an imaging mirror group 311, and a camera 304; the semi-transparent mirror 310 is arranged in the conjugate optical path of the splicing mirror 1. The light beam emitted by the light source 301 is transmitted through the beam splitter 302 and then illuminates the splicing mirror 1. The light beam reflected by the splicing mirror 1 is incident on the semi-transparent and semi-reflective mirror 310. The light beam transmitted from the semi-transparent and semi-reflective mirror 310 passes through the imaging mirror group 311 and enters the camera 304. Processor 4 is used to receive light intensity images acquired by camera 304 and to detect boundary changes of splicing mirror 1 based on the light intensity images.

[0056] In addition, the semi-transparent and semi-reflective mirror 310 is mounted on the piezoelectric ceramic 50; The processor 4 is used to receive the light intensity image acquired by the camera 304, detect the boundary changes of the splicing mirror 1 based on the light intensity image, and control the piezoelectric ceramic 50 to adjust its attitude and position according to the boundary change detection result.

[0057] In this embodiment, multiple small semi-transparent and semi-reflective mirrors are arranged at the conjugate optical path positions of the splicing mirror, with each small semi-transparent and semi-reflective mirror corresponding to the main mirror of the splicing mirror. These small semi-transparent and semi-reflective mirrors reflect most of the incident light beam while allowing a small portion to pass through. The transmitted beam is then transmitted to the camera via an imaging lens to complete image acquisition. The entire semi-transparent and semi-reflective mirror is mounted and fixed above the piezoelectric ceramic, which allows for precise adjustment of minute displacements and angles. This structure eliminates the need for an additional independent detection and sensing optical path. The reflective mirror structure equipped with the piezoelectric ceramic can achieve self-sensing of its own state. This self-sensing system provides real-time feedback on the optical path status, thereby directly aligning and adjusting the attitude of the splicing mirror's optical path.

[0058] The above describes the process of detecting changes in the edge of a video wall using a contact-based sensing structure and a non-contact optical structure. With long-term operation of the video wall system, the edge sensors may experience sensitivity degradation due to factors such as aging or sudden local temperature changes. Existing offline calibration methods cannot achieve online recalibration without disassembling the video wall system, affecting the long-term stability of large-aperture systems. Therefore, in some embodiments, after detecting changes in the video wall boundary, the processor is also used for: The raw electrical signals output by multiple sensor probes are acquired and converted to obtain the basic distance reading. A linear mapping matrix model between the distance reading and the local spatial pose state of the sub-mirror is established by combining the local coordinate distribution of the probe. The equations corresponding to the linear mapping matrix model are solved by the least squares method to extract the optimal local pose state. After obtaining the optimal local pose state, the Abbe error caused by the height difference between the measurement point and the neutral plane of the sub-mirror is compensated, and the geometric coupling error is eliminated to obtain the true edge sensor translation reading. The readings of all seam nodes on the entire mirror surface are combined into a global measurement feedback column vector; The camera in the first optical detection structure acquires defocused light intensity images, and the light intensity transmission equation is solved to obtain the true global pose state vector of each sub-mirror. In calibration mode, a preset disturbance is injected, and multiple sets of data are collected to form paired data in which the sensor feedback measurement value and the state true value correspond one-to-one; wherein, each set of data is composed of the global measurement feedback column vector and the global pose state vector. The pre-defined linear mapping equation is solved using a regularization algorithm, and the global sensitivity calibration matrix of the sensor is obtained by inverse solution. The linear mapping equation represents the relationship between the product of the global sensitivity calibration matrix and the global pose state vector and the global measurement feedback column vector. Based on the global sensitivity calibration matrix and combined with the tilt and displacement state calculation results of each sub-mirror, control and adjustment commands for the actuator are generated to regulate the attitude of the sub-mirrors.

[0059] The process described above is referred to as a dynamically calibrated feedback control strategy. To enable those skilled in the art to better understand this process, it will be described below with reference to specific embodiments.

[0060] The purpose of this process is to perform in-situ calibration of the underlying edge sensor using high-precision optical wavefront ground truth, thereby obtaining an accurate global sensor sensitivity matrix. The specific calibration data calculation process is as follows: The edge sensor's four pairs of differential probes first output raw electrical signals, which are then converted into four local baseline distance readings. Let the high-dimensional local spatial pose state vector that the sensor node needs to extract be... .

[0061] Combining the half-spacing distribution of the four pairs of probes in the local coordinate system Establish a positive kinematic linear mapping matrix model between low-dimensional observations and high-dimensional poses. ,in For high-frequency measurement noise, the structure mapping matrix Expressed as: ; To eliminate measurement noise caused by redundant probes, the system employs the least squares method to solve the generalized inverse of the overdetermined equation, extracting the optimal local spatial pose state: ; In obtaining the optimal local state Subsequently, it is necessary to further compensate for the Abbe error caused by the height difference between the measurement point of action and the neutral plane of the sub-mirror. Define the preset projection matrix of the release lever arm of the sensor. for The true edge sensor translation readings after eliminating geometric coupling errors were calculated. : ; Calculated from all seam nodes of the entire mirror surface Combined, they are constructed into a global measurement feedback column vector. Simultaneously, utilizing the wavefront curvature sensing optical link, the light intensity transmission equation is calculated by solving the defocused light intensity images before and after focus acquired by the CCD camera, thereby obtaining the true global pose state vector of each sub-mirror. (Includes the true Piston, displacement, and tilt angle values ​​for each sub-mirror). In calibration mode, the system injects specific disturbances via actuators and collects multiple sets of state truth values. With sensor feedback measurement value The paired data. Solving the linear mapping equation using the Tikhonov regularization algorithm. This allows for the inverse solution to obtain the accurate global sensitivity calibration matrix of the sensor. .

[0062] To ensure state stability during the calibration process, the system employs the following closed-loop state equation for control: ; In the formula, is the global pose state vector of the sub-mirror; This is the control influence matrix of the underlying actuator system, used to characterize the pose change caused by a unit actuation amount; Let be the feedback gain matrix for closed-loop control. This equation drives the system state to converge smoothly, and is the sensitivity matrix. The high-precision solution provides a stable dynamic environment.

[0063] In this embodiment, a sensitivity calibration matrix is ​​used to complete the data calculation, obtaining the tilt amount corresponding to each mirror body. Then, the calculated mirror body tilt state is used as the calibration basis to accurately determine the adjustment amount and direction of each actuator, thereby enabling corresponding control of each actuator and achieving closed-loop correction of the mirror body attitude.

[0064] During the operation of the splicing mirror, sub-mirrors often exhibit image divergence. The causes of this divergence can include: misalignment or detection malfunction of the sub-mirror edge sensors, leading to misalignment of the sub-mirror boundaries and consequently image divergence; or misalignment or distortion of the sub-mirror's own mirror surface attitude, causing optical path defocusing and disordered light propagation. To address this issue, fault tracing is necessary. To facilitate fault tracing, in some embodiments, the processor is also used for: The electronic timing statistics characteristics are integrated with the optical energy distribution characteristics; Using the fused data, a joint anomaly observation matrix is ​​constructed within a synchronous sliding time window; Based on the joint anomaly observation matrix, dimensionality reduction and fault detection are performed through principal component analysis; The construction of the joint anomaly observation matrix within the synchronous sliding time window includes: Within the synchronous sliding time window, collect the working data of all edge sensors, calculate the deviation between the actual detection reading and the theoretical prediction reading of each edge sensor, and form electronic state characteristics based on various deviations. Acquire specular optical imaging spot images, and extract spot center moment, centroid offset, and optical bounding energy range parameters to form optical energy distribution characteristics; The electronic state characteristics and optical energy distribution characteristics at the same time point are time-matched and spliced ​​together to form the joint features corresponding to a single sampling. Multiple sets of joint features at consecutive time points are continuously collected and stacked sequentially to construct a joint anomaly observation matrix.

[0065] Furthermore, the processor is used for dimensionality reduction and fault detection based on a joint anomaly observation matrix and through principal component analysis, including: Calculate the covariance matrix of the preprocessed joint variable observation matrix; Perform eigenvalue decomposition on the covariance matrix to obtain the corresponding eigenvalues ​​and orthogonal eigenvectors in sequence; The corresponding contribution rate is calculated based on each feature value, and then sorted from high to low contribution rate. Extract the first preset number of principal components whose cumulative contribution rate exceeds a preset threshold to form a projection matrix; Based on the projection matrix, the preprocessed joint variation observation matrix is ​​projected onto the principal component space to obtain the score matrix; When an anomaly is detected in the principal component score, the orthogonal feature vectors corresponding to the principal components are retrieved, and the dominant feature dimension with the largest absolute value among the orthogonal feature vectors is extracted to determine the dominant influencing factor that caused the anomaly.

[0066] To enable those skilled in the art to better understand the process of fault tracing, the following description is provided in conjunction with specific embodiments.

[0067] This embodiment employs a fault tracing mechanism based on the fusion of opto-mechanical-electrical features using Principal Component Analysis (PCA).

[0068] To overcome the challenge of multiple solutions in anomaly detection using single sensor signals or single optical features, this invention proposes a joint classification and source tracing method based on principal component analysis (PCA). This method fuses the underlying electronic temporal statistical features with the top-level optical energy distribution features, constructing a high-dimensional joint anomaly observation matrix within a synchronous sliding time window. Dimensionality reduction and pattern extraction are then performed using PCA, ultimately achieving precise fault location and source tracing.

[0069] First, a dual-link feature extraction and joint anomaly observation matrix are constructed. The system operates within a synchronized sliding time window. Within, complementary mathematical features from electronics and optics are extracted and combined into a heterogeneous high-dimensional observation state.

[0070] Extract full caliber The residual values ​​of each edge sensor constitute the electronic state vector. ,in, Defined as the first The difference between the real-time corrected readings of each edge sensor and the kinematic model-based predictions: ; Statistical features were extracted from the synchronously acquired far-field star spot images, and the higher-order central moments and centroid evolution indices of the light spots were calculated to construct the optical morphological moment feature vector. : ; in, To characterize the second moment of symmetric dispersion, To characterize the third moment of asymmetric distortion, It is a second-order cross moment. It is a third-order cross moment, reflecting the asymmetric deformation characteristics of the light spot. The amount of centroid drift. for Enclosing energy radius.

[0071] The electronic readings and optical moment features are time-stamped and combined, and then concatenated to form a single-sample opto-mechanical-electronic joint feature vector. : ; Set the number of observation samples to By stacking consecutive joint eigenvectors, a joint anomaly observation matrix of the system is constructed. To eliminate the discrepancy between electrical scalar measurements and optical pixel dimensions, Z-score standardization is performed to obtain the normalized matrix. .

[0072] The system for joint normalized matrices Perform principal component analysis. Calculate its covariance matrix. : ; Perform eigenvalue decomposition on the covariance matrix to find the eigenvalues ​​and their corresponding orthogonal eigenvectors (i.e., loading vectors). Extracting the top performers whose cumulative contribution rate exceeds a preset threshold (e.g., 95%). The principal components constitute the projection matrix. Calculate the score matrix Due to the original matrix Forced fusion and Principal component space Each feature projection objectively reflects the strong coupling mapping relationship between a specific optical distortion and the underlying physical node.

[0073] When the system detects the first When a significant anomaly occurs in the score of a principal component, the system automatically retrieves the orthogonal feature vector corresponding to that principal component. The invention extracts several feature dimensions with the largest absolute values ​​(i.e., the dominant features that contribute the most to the current anomaly). By analyzing the mathematical property distribution of high-load features, this invention does not rely on a single physical cause presupposition, but rather generalizes and classifies the abnormal state of the system into the following typical failure modes, thereby achieving source tracing: Mode I (Global Gradual Mismatch): If the feature dimension with the largest absolute value of the loading coefficient in the anomalous principal component is concentrated in the optical second central moment ( ) and the radius of the surrounding energy ( If the indicators that characterize the symmetric dispersion of energy are used, then the source tracing system is currently in a low-frequency, slow global rigid body displacement or surface degradation state (such as overall defocusing caused by slow environmental drift or global hysteresis of support stiffness).

[0074] Mode II (Locally Independent Microanomaly): If high load factors are concentrated in specific edge sensor residual terms ( The statistical skewness characteristics of the macroscopic optical feature terms are relatively low, so the source of the anomaly is highly limited to the underlying electromechanical link where the specific sensor node is located (such as local electrical transient glitches, single-point physical connection mutations, or single actuator failures).

[0075] Mode III (Dynamic Asymmetric Distortion Perturbation): If the high load factor manifests as the third-order optical central moment ( ) and centroid drift ( If the synchronous surge of ) is accompanied by high-frequency fluctuations in the underlying electronic characteristics, then the source determination system is experiencing dynamic high-frequency disturbances that cause asymmetric spatial aberrations (such as complex flutter caused by external high-frequency excitation or local sub-mirror structure resonance).

[0076] Through the above-mentioned quantitative analysis and generalized clustering based on load coefficients, this invention achieves a reliable mapping from principal component scores to bottom-level node location and macroscopic failure modes, greatly improving the inclusiveness and generalization adaptability of the fault diagnosis mechanism to complex and variable field conditions.

[0077] The splicing mirror boundary change detection device provided by the present invention has the following advantages: (1) Effective decoupling of Abbe error (lever effect) in multiple degrees of freedom: Compared with traditional single-axis calibration platforms, this invention addresses the height difference between the sensor measurement point and the neutral plane of the sub-mirror during actual installation by introducing a high-order matrix model that includes parameters for releasing the lever arm, thereby effectively compensating for the lever effect generated when the sub-mirror is tilted. This method can better separate the kinematic coupling errors of translation (Piston) and tilt (Tip / Tilt), and more accurately restore the actual response characteristics of the sensor.

[0078] (2) Provides a high-precision optical truth reference to improve calibration accuracy: Unlike the indirect measurement method of traditional offline calibration that relies on commercial displacement gauges or interferometers, this invention uses a frequency domain wavefront curvature sensing algorithm to calculate the optical wavefront phase by extracting the light intensity distribution before and after focal length. Using the macroscopic pupil phase distribution as a non-contact reference provides a reliable "optical truth" reference for the calibration of edge sensors, which helps to reduce additional errors caused by factors such as residual stress from mechanical installation.

[0079] (3) Possesses dynamic closed-loop calibration and online status monitoring capabilities: Combining the actuator influence matrix and the Tikhonov regularized feedback control strategy, this method can not only be used to obtain calibration coefficients offline, but also promote the smooth convergence of the system state when the co-phase system is running. When the edge sensor experiences zero-point drift due to aging or temperature changes, the system can assist in online recalibration and abnormal node localization, thereby helping to improve the long-term operational stability of the large-aperture splicing system.

[0080] (4) Compatible with high-order optical sensors and supports multi-dimensional spatial information calculation: The multi-degree-of-freedom mapping model established in this invention is not only applicable to error decoupling of conventional one-dimensional sensors, but can also be extended to new high-order optical edge sensors. Based on this technical solution, the system can directly calculate and obtain the relative angle between adjacent sub-mirrors ( ) and translation ( This provides multi-dimensional spatial information, including [database information]. This changes the traditional situation where sensing nodes mainly feed back a single scalar signal, providing a richer information foundation and theoretical support for multi-dimensional state perception and collaborative control of large-aperture co-phase systems.

[0081] The splicing mirror boundary change detection device provided by this invention not only detects changes in the splicing mirror boundary but also improves upon the technical problems faced by traditional edge sensor calibration under complex conditions such as varying zenith angles for large-aperture splicing mirrors. These problems include large multi-degree-of-freedom coupling errors, lack of in-situ high-precision benchmarks, and difficulties in online diagnostics. By designing a multi-degree-of-freedom pose simulation mechanism and combining wavefront curvature sensing frequency domain solution algorithms with control closed loops, this invention achieves an effective mapping from macroscopic optical phase to microscopic node displacement.

[0082] The above describes a device for detecting changes in the boundary of a splicing mirror. This embodiment also provides a method for detecting changes in the boundary of a splicing mirror, which is applied to the above-described device for detecting changes in the boundary of a splicing mirror. Figure 5 A flowchart of a method for detecting changes in the boundary of a splicing mirror provided in an embodiment of the present invention is shown below. Figure 5 As shown, the method includes: S10: Acquire data from edge sensors, sensor mounting arm length parameters, image acquisition components, and / or event sensors; S11: Detect changes in the splicing mirror boundary based on data collected by the edge sensor and the sensor mounting arm length parameter, and detect changes in the splicing mirror boundary based on data collected by the image acquisition component and / or the event sensor.

[0083] The splicing mirror boundary change detection method provided in this embodiment is applied to the splicing mirror boundary change detection device mentioned above. It has the same or corresponding technical features as the splicing mirror boundary change detection method mentioned above, which is applied to the splicing mirror boundary change detection device mentioned above. The embodiments of the splicing mirror boundary change detection device have been described in detail above. The embodiments of the splicing mirror boundary change detection method will not be described again here, and the effects are the same as above.

[0084] The foregoing has provided a detailed description of the splicing mirror boundary change detection device and method provided by the present invention. The various embodiments in the specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. Regarding the methods disclosed in the embodiments, since they correspond to the devices disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the device section description. It should be noted that those skilled in the art can make various improvements and modifications to the present invention without departing from its principles, and these improvements and modifications also fall within the protection scope of the present invention.

[0085] It should also be noted that, in this specification, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. 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.

Claims

1. A device for detecting changes in the boundary of a splicing mirror, characterized in that, include: The processor, and the contact sensing and detection structure and the non-contact optical structure arranged relative to the splicing mirror body; The contact sensing and detection structure includes an edge sensor, which is positioned at the seam between adjacent sub-mirrors. The non-contact optical structure includes at least a beam splitter element arranged sequentially along the beam transmission direction, an image acquisition component, and / or an event sensor for capturing events of changes in the splicing mirror boundary. The processor is connected to the edge sensor, the image acquisition component, and / or the event sensor, and is used to detect changes in the splicing mirror boundary based on the data acquired by the edge sensor and the sensor mounting arm length parameter, and to detect changes in the splicing mirror boundary based on the data acquired by the image acquisition component and / or the data acquired by the event sensor.

2. The splicing mirror boundary change detection device according to claim 1, characterized in that, The edge sensor includes multiple sets of sensing probes, which are mounted on a mounting base and arranged in an array. The sensing probes adopt a differential structure with upward and downward opposing beams. The processor is used to detect changes in the edge of the splicing mirror based on the data collected by the edge sensor and the sensor mounting arm length parameter, including: Displacement readings are collected by the multiple sets of sensor probes. Based on the installation position of the multiple sets of sensor probes in the local coordinate system of the sensor, and combined with the small-angle rigid body kinematic approximation when adjacent sub-mirrors undergo relative translation and relative tilt around the local coordinate axis, a positive kinematic mapping relationship between displacement readings and spatial pose is established. Based on the aforementioned positive kinematic mapping relationship, a matrix for symmetric difference operations is constructed; The matrix obtained by the symmetric difference operation is used to project the displacement readings collected by the multiple sets of sensor probes onto the orthogonal spatial degrees of freedom, and output the absolute translation and relative tilt of the adjacent sub-mirrors. The changes in the splicing mirror boundary are detected based on the absolute translation, the relative tilt, and the sensor mounting arm length parameter.

3. The splicing mirror boundary change detection device according to claim 2, characterized in that, The processor is used to determine the sensor mounting arm length parameters, including: The output data of each sensor probe is collected as the multiple sets of sensor probes rotate around the axis at different positions of the test device. Based on the output data of each sensor probe, positions whose output values ​​do not change with rotation are selected. The geometric parameters corresponding to the determined position are used as the length parameters of the sensor mounting arm.

4. The splicing mirror boundary change detection device according to claim 2 or 3, characterized in that, The processor detects changes in the mosaic mirror boundary based on the absolute translation, the relative tilt, and the sensor mounting arm length parameter, including: Based on the absolute translation, the relative tilt, and the sensor mounting arm length parameters, an error compensation model is constructed to correct the edge sensor readings and obtain the corrected sensor output value. The corrected sensor output value is decomposed into multiple components according to the dimension; among them, the equation corresponding to one set of components includes the coupling compensation term of the mounting arm length parameter to the tilt amount, and the equation corresponding to the other components is constructed based on the ideal reading model without the influence of the mounting arm length parameter. By combining the equations corresponding to each component, a system of equations containing different error compensation terms is formed; The actual translation and tilt states of each sub-mirror are decoupled and deduced from the system of equations containing different error compensation terms.

5. The splicing mirror boundary change detection device according to claim 2, characterized in that, The non-contact optical structure includes a first optical detection structure; The first optical detection structure includes a light source, a beam splitter, an imaging lens, and a camera. The camera is positioned equidistantly from the front and back focal lengths of the imaging lens. The light beam emitted by the light source passes through the beam splitter and illuminates the splicing mirror. The light beam reflected by the splicing mirror is incident on the beam splitter in the opposite direction and is reflected and deflected by the beam splitter to the imaging lens. After passing through the imaging lens, it is transmitted to the camera. The processor is used to receive the light intensity image modulated by the pose of the sub-mirror acquired by the camera, and calculate the wavefront phase based on the light intensity image to complete the detection of changes in the splicing mirror boundary.

6. The splicing mirror boundary change detection device according to claim 5, characterized in that, The processor is used to receive a light intensity image modulated by the pose of the receiver lens acquired by the camera, and to calculate the wavefront phase based on the light intensity image, including: Images of defocused light intensity at equidistant positions before and after the focal point are acquired. The energy distribution characteristics of the image points are extracted and analyzed, and the centroid offset of the image point distribution and the gradient of image contrast decrease are calculated. The macroscopic alignment error and boundary curvature change of the system are evaluated based on the centroid offset of the image point distribution and the gradient of image contrast descent, so as to initially screen out sub-mirror regions with misalignment problems. For sub-mirror regions with misalignment issues, based on the light intensity transmission equation and under the condition that the defocus amount is less than the preset defocus amount, the differential relationship between the axial derivative of the light intensity along the optical axis and the wavefront phase is simplified to a curvature-sensing Poisson equation; wherein, the light intensity difference term on the right side of the curvature-sensing Poisson equation is calculated from the acquired defocused light intensity image. The wavefront phase to be determined is represented as a linear combination of orthogonal basis functions of a preset order, and substituted into the curvature sensing Poisson equation to establish the correspondence between the wavefront coefficient and the light intensity difference term. Based on the aforementioned correspondence, the spatial domain signal is mapped to the frequency domain using the Fast Fourier Transform, and a discrete wavefront solution matrix equation is constructed. The inverse operation of the matrix equation is performed to calculate the wavefront coefficient vector, thereby reconstructing the wavefront phase.

7. The device for detecting changes in the boundary of a splicing mirror according to claim 1, characterized in that, The non-contact optical structure includes a second optical detection structure; The second optical detection structure includes a light source, a beam splitter, an imaging lens, and the event sensor. The light beam emitted by the light source passes through the beam splitter and illuminates the splicing mirror. The light beam reflected by the splicing mirror is incident on the beam splitter in the opposite direction and is reflected and deflected by the beam splitter to the imaging lens. After passing through the imaging lens, it is transmitted to the event sensor. The processor is used to receive event data collected by the event sensor; and to extract dynamic variation features of the splicing mirror boundary based on the event data in order to detect changes in the splicing mirror boundary.

8. The splicing mirror boundary change detection device according to claim 1, characterized in that, The non-contact optical structure includes a third optical detection structure; The third optical detection structure includes a light source, a beam splitter, a semi-transparent mirror, an imaging mirror group, and a camera; the semi-transparent mirror is arranged on the conjugate optical path of the splicing mirror. The light beam emitted by the light source passes through the beam splitter and illuminates the splicing mirror. The light beam reflected by the splicing mirror is then incident on the semi-transparent and semi-reflective mirror. The light beam transmitted from the semi-transparent and semi-reflective mirror passes through the imaging lens group and enters the camera. The processor is used to receive light intensity images captured by the camera and to detect changes in the boundary of the stitching mirror based on the light intensity images.

9. The splicing mirror boundary change detection device according to claim 8, characterized in that, The semi-transparent and semi-reflective mirror is mounted on a piezoelectric ceramic. The processor is used to receive light intensity images captured by the camera, detect changes in the boundary of the splicing mirror based on the light intensity images, and control the piezoelectric ceramic to adjust its attitude and position according to the boundary change detection results.

10. The splicing mirror boundary change detection device according to claim 5, characterized in that, After detecting changes in the splicing mirror boundary, the processor is also used for: The original electrical signals output by the multiple sets of sensor probes are acquired and converted to obtain the basic distance reading. A linear mapping matrix model between the distance reading and the local spatial pose state of the sub-mirror is established by combining the local coordinate distribution of the probe. The equations corresponding to the linear mapping matrix model are solved by the least squares method to extract the optimal local pose state. After obtaining the optimal local pose state, the Abbe error caused by the height difference between the measurement point and the neutral plane of the sub-mirror is compensated, and the geometric coupling error is eliminated to obtain the true edge sensor translation reading. The readings of all seam nodes on the entire mirror surface are combined into a global measurement feedback column vector; The camera in the first optical detection structure acquires defocused light intensity images, and the light intensity transmission equation is solved to obtain the true global pose state vector of each sub-mirror. In calibration mode, a preset disturbance is injected, and multiple sets of data are collected to form paired data in which the sensor feedback measurement value and the state true value correspond one-to-one; wherein, each set of data is composed of the global measurement feedback column vector and the global pose state vector. A pre-defined linear mapping equation is solved using a regularization algorithm to obtain the global sensitivity calibration matrix of the sensor. The linear mapping equation represents the relationship between the product of the global sensitivity calibration matrix and the global pose state vector and the global measurement feedback column vector. Based on the global sensitivity calibration matrix, and combined with the tilt and displacement state calculation results of each sub-mirror, control and adjustment commands for the actuator are generated to regulate the attitude of the sub-mirrors.

11. The splicing mirror boundary change detection device according to claim 10, characterized in that, The processor is also used for: The electronic timing statistics characteristics are integrated with the optical energy distribution characteristics; Using the fused data, a joint anomaly observation matrix is ​​constructed within a synchronous sliding time window; Based on the joint anomaly observation matrix, dimensionality reduction and fault detection are performed using principal component analysis. The construction of the joint anomaly observation matrix within the synchronous sliding time window includes: Within the synchronous sliding time window, collect the working data of all edge sensors, calculate the deviation between the actual detection reading and the theoretical prediction reading of each edge sensor, and form electronic state characteristics based on various deviations. Acquire specular optical imaging spot images, and extract spot center moment, centroid offset, and optical bounding energy range parameters to form optical energy distribution characteristics; The electronic state characteristics and optical energy distribution characteristics at the same time point are time-matched and spliced ​​together to form the joint features corresponding to a single sampling. The joint features of multiple consecutive time points are continuously collected and stacked sequentially to construct the joint anomaly observation matrix.

12. The device for detecting changes in the boundary of a splicing mirror according to claim 11, characterized in that, The processor is used to perform dimensionality reduction and fault detection based on the joint anomaly observation matrix and through principal component analysis, including: Calculate the covariance matrix of the preprocessed joint variable observation matrix; The covariance matrix is ​​subjected to eigenvalue decomposition to obtain the corresponding eigenvalues ​​and orthogonal eigenvectors in sequence. The corresponding contribution rate is calculated based on each feature value, and then sorted from high to low contribution rate. Extract the first preset number of principal components whose cumulative contribution rate exceeds a preset threshold to form a projection matrix; Based on the projection matrix, the preprocessed joint variation observation matrix is ​​projected onto the principal component space to obtain the score matrix; If an anomaly is detected in the principal component score, the orthogonal feature vector corresponding to the principal component is retrieved, and the dominant feature dimension with the largest absolute value among the orthogonal feature vectors is extracted to determine the dominant influencing factor that caused the anomaly.

13. A method for detecting changes in the boundary of a spliced ​​mirror, characterized in that, The method, applied to the splicing mirror boundary change detection device according to any one of claims 1 to 12, comprises: Acquire data from edge sensors, sensor mounting arm length parameters, image acquisition components, and / or event sensors; The changes in the splicing mirror boundary are detected based on the data collected by the edge sensor and the sensor mounting arm length parameter, as well as based on the data collected by the image acquisition component and / or the event sensor.

Images