Dynamic detection method and system for flatness of structured light fusion laser calibration

Abstract

This invention relates to the field of television backplane inspection technology, and particularly to a method and system for dynamic flatness detection using structured light fusion laser calibration. The method includes: projecting phase-coded light stripes onto the surface of the television backplane and acquiring a deformation image; simultaneously emitting a pulsed modulated laser beam and acquiring a reflection displacement signal; calculating the reflection displacement signal to generate three-dimensional coordinates of a laser reference point; generating an initial structured light point cloud based on the deformation image; registering the initial structured light point cloud to generate a spatiotemporally fused point cloud; generating a reference plane; calculating the thermal deformation compensation vector for each point based on a pre-stored material thermal expansion coefficient, and performing a reverse offset operation of vector superposition based on the thermal deformation compensation vector; calculating the distance from each point in the spatiotemporally fused point cloud to the reference plane, and generating a flatness detection result based on the standard deviation of the distance distribution. This invention effectively solves the technical problem of decreased flatness measurement accuracy caused by non-uniform thermal deformation and conveyor belt jitter coupling interference during dynamic inspection of television backplanes.

Description

Technical Field

[0001] This invention relates to the field of television backplane inspection technology, and in particular to a method and system for dynamic detection of flatness using structured light fusion laser calibration. Background Technology

[0002] In modern industrial manufacturing, flatness inspection of electronic device casings, such as television back panels, is a core step in ensuring product assembly accuracy and appearance quality. Flatness characterizes the degree of geometric deviation of the television back panel surface from an ideal reference plane, and its measurement relies on the calculation of spatial topological relationships of high-precision three-dimensional point cloud coordinates. Current mainstream technologies generally adopt a detection architecture that combines structured light projection with multi-sensor fusion: by projecting an coded grating onto the television back panel surface, reconstructing the point cloud model through phase calculation, and then achieving flatness measurement through thermal expansion compensation and fitting with a reference plane.

[0003] Existing technical solutions suffer from three major drawbacks. First, traditional temperature compensation mechanisms rely on global coordinate scaling based on the overall average expansion coefficient, which cannot respond to non-uniform thermal deformation caused by local material thickness differences. When a temperature gradient is generated between the thin-walled area and the metal frame due to thermal capacity imbalance, the linear compensation model leads to point cloud surface distortion, especially in the edge region, resulting in systematic measurement bias. Second, existing solutions rely on mechanical positioning fixtures or static measurement environments to establish a spatial coordinate system. When the TV back panel is affected by conveyor belt vibration during production line transportation, the relative pose of the sensor and the TV back panel shifts instantaneously, causing inaccurate data fusion of multi-frame point clouds and inconsistent measurement results. Finally, industry standards and mainstream patents use the distance between the highest and lowest points as a flatness criterion. This method only feeds back extreme deviation values ​​and ignores the surface fluctuation distribution characteristics, resulting in a severe deficiency in capturing high-frequency micro-amplitude wave deformations such as periodic fluctuations caused by thermal stress release, leading to a systematic underestimation of micro-deformation degradation.

[0004] The aforementioned defects form a technical bottleneck, namely, the coupling interference of non-uniform thermal expansion and vibration under motion conditions on the point cloud topology, while the coarse-grained flatness index cannot effectively trace the true deformation mode, ultimately leading to the degradation of the accuracy of the online detection system in high-throughput production environments.

[0005] The information disclosed in this background section is intended only to enhance the understanding of the general background of this disclosure and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0006] This invention provides a method and system for dynamic detection of flatness in structured light fusion laser calibration, which can effectively solve the problems in the background art.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] A method for dynamic detection of flatness in structured light fusion laser calibration, the method comprising:

[0009] Phase-coded light stripes are projected onto the surface of the TV back panel and deformed images are acquired. Simultaneously, pulse-modulated laser beams are emitted to at least three non-collinear positioning points deployed in the metal corner area and the central thin-wall area and reflection displacement signals are acquired.

[0010] Based on the pulse flight time-domain demodulation of the reflected displacement signal and the compensation for atmospheric scattering distortion based on light intensity attenuation, the three-dimensional coordinates of the laser reference point are generated.

[0011] An initial structured light point cloud is generated based on the deformed image. A spatiotemporal fusion point cloud is generated by registration using the three-dimensional coordinates of the laser reference point as a dynamic spatial reference. A reference plane is generated based on the spatiotemporal fusion point cloud, wherein the normal vector constraint direction of the reference plane is collinear with the assembly reference plane of the composite structure of the TV back panel thin wall and the frame.

[0012] Acquire the surface temperature field distribution data of the TV back panel and determine the temperature change of each point in the spatiotemporal fusion point cloud. Execute the partitioned thermal expansion coefficient call based on the difference in the material thermal expansion coefficients of the metal corner area and the central thin-walled area.

[0013] The thermal deformation compensation vector of each point is calculated based on the thermal expansion coefficient of the material and the temperature change, and a reverse offset operation of vector superposition is performed on the spatial position coordinates of the spatiotemporal fusion point cloud based on the thermal deformation compensation vector.

[0014] The distance from each point in the compensated spatiotemporal fusion point cloud to the reference plane is calculated. Based on the wave deformation mode of the thin-walled region of the TV back panel, a flatness detection result is generated according to the standard deviation of the distance distribution.

[0015] Furthermore, the three-dimensional coordinates of the laser reference point with timestamps are generated, including:

[0016] The reflection displacement signal is demodulated in the time domain to separate the pulse flight time sequence and the accompanying light intensity attenuation characteristic data corresponding to each positioning point.

[0017] The pulse flight time sequence is synchronized and aligned with the start time of a clock with the same exposure period to generate a timestamp for each positioning point;

[0018] Based on the light intensity attenuation characteristic data, the signal distortion caused by atmospheric scattering is compensated, and the initial three-dimensional coordinates of each positioning point in the laser measurement coordinate system are calculated in combination with the emission angle of the pulse-modulated laser beam.

[0019] Based on the spatial distribution characteristics of the at least three non-collinear positioning points, positional constraints are established, and the accuracy of the initial three-dimensional coordinates is optimized based on least squares adjustment to generate the three-dimensional coordinates of the laser reference point with the timestamp.

[0020] Furthermore, the accuracy of the initial three-dimensional coordinates is optimized based on least squares adjustment, including:

[0021] Based on the spatial distribution characteristics of the at least three non-collinear positioning points, spatial position constraints are established with the measured distance and theoretical geometric relationship between adjacent positioning points as the benchmark.

[0022] Calculate the spatial position offset of the initial three-dimensional coordinates under the spatial position constraint conditions, and generate a set of coordinate correction values ​​for each positioning point;

[0023] By gradually reducing the spatial position offset through multiple position adjustment operations, the spatial position of each positioning point satisfies the matching requirements of the theoretical geometric relationship.

[0024] When the spatial offset of all positioning points is less than a preset threshold, the optimized three-dimensional coordinates are output as the three-dimensional coordinates of the timestamped laser reference point.

[0025] Furthermore, the reference plane is generated according to the weighted least squares method, including:

[0026] The curvature representation value of each point in the spatiotemporal fused point cloud is calculated based on the rate of change of the normal vector.

[0027] Based on the distribution range of the curvature characterization values, the spatiotemporal fusion point cloud is divided into high curvature regions and low curvature regions;

[0028] A first weighting coefficient is assigned to the high curvature region, and a second weighting coefficient is assigned to the low curvature region, wherein the first weighting coefficient is greater than the second weighting coefficient;

[0029] Based on the spatial coordinates of the spatiotemporal fusion point cloud, a plane fitting optimization target is constructed by combining the first weighting coefficient and the second weighting coefficient;

[0030] The planar spatial pose parameters are iteratively adjusted until the planar fitting optimization objective reaches a convergent state, and the reference plane is output.

[0031] Furthermore, a plane fitting optimization objective is constructed, including:

[0032] Based on the spatial coordinates of the spatiotemporal fusion point cloud, and combined with the first weighting coefficient and the second weighting coefficient, the objective of the plane fitting optimization is defined as minimizing the cumulative weighted distance deviation from each point in the spatiotemporal fusion point cloud to the reference plane.

[0033] The weighting of the cumulative weighted distance deviation is based on the division results of the high curvature region and the low curvature region, so that the points in the high curvature region are weighted by the first weighting coefficient and the points in the low curvature region are weighted by the second weighting coefficient.

[0034] The optimization constraints of the planar spatial pose parameters are established based on the weighted distance deviation accumulation to ensure that the planar fitting optimization target reflects the influence of the curvature characterization value on the measurement reliability.

[0035] The plane fitting optimization objective is associated with the iterative adjustment process to drive the adjustment direction of the plane spatial pose parameters until the convergence state is reached.

[0036] Furthermore, the spatiotemporal fusion point cloud is compensated for in three-dimensional coordinate offset by incorporating the pre-stored material thermal expansion coefficient, including:

[0037] Acquire the surface temperature field distribution data of the TV back panel, and determine the temperature change of each point in the spatiotemporal fusion point cloud based on the surface temperature field distribution data;

[0038] Based on the pre-stored material thermal expansion coefficient and the temperature change, calculate the thermal deformation compensation vector of each point in the spatiotemporal fusion point cloud;

[0039] Based on the thermal deformation compensation vector, the spatial position coordinates of the spatiotemporal fusion point cloud are adjusted by reverse offset to generate the compensated spatiotemporal fusion point cloud to eliminate measurement errors caused by temperature.

[0040] Furthermore, the spatial position coordinates of the spatiotemporal fused point cloud are adjusted by reverse offset based on the thermal deformation compensation vector, including:

[0041] Establish a mapping relationship between the thermal deformation compensation vector and the spatial position coordinates of each point in the spatiotemporal fusion point cloud, forming a coordinate compensation mapping relationship;

[0042] According to the coordinate compensation mapping relationship, a vector superposition operation is performed on the spatial position coordinates of the spatiotemporal fusion point cloud, so that the coordinates of each point are shifted in the opposite direction of the thermal deformation compensation vector;

[0043] The spatiotemporal fusion point cloud is updated based on the spatial position coordinates after the vector superposition operation to generate the compensated spatiotemporal fusion point cloud, wherein the relative pose of the compensated spatiotemporal fusion point cloud and the reference plane reflects the true geometric relationship after temperature compensation.

[0044] Furthermore, flatness test results are generated based on the range and standard deviation of the distance distribution, including:

[0045] Based on the distance distribution from each point in the compensated spatiotemporal fusion point cloud to the reference plane, the difference between the maximum and minimum values ​​of the distance distribution is calculated as the range;

[0046] The standard deviation is calculated based on the dispersion of the distance values ​​of each point in the distance distribution, which characterizes the overall fluctuation range of the compensated spatiotemporal fusion point cloud relative to the reference plane.

[0047] The range is used as an overall deviation metric, and the standard deviation is used as a local fluctuation metric. The flatness test result is generated by combining these two parameters, and the flatness test result is used to quantify the flatness quality of the TV back panel surface.

[0048] A structured light fusion laser calibration flatness dynamic detection system, the system comprising:

[0049] The laser modulation module projects phase-coded light stripes onto the surface of the TV back panel and acquires deformed images, while simultaneously emitting pulsed modulated laser beams to at least three non-collinear positioning points deployed in the metal corner area and the central thin-wall area and acquiring reflection displacement signals.

[0050] The coordinate generation module generates three-dimensional coordinates of the laser reference point by demodulating the reflection displacement signal in the time domain based on the pulse flight time and compensating for atmospheric scattering distortion based on light intensity attenuation.

[0051] The reference plane module generates an initial structured light point cloud based on the deformed image, and performs registration to generate a spatiotemporally fused point cloud using the three-dimensional coordinates of the laser reference point as a dynamic spatial reference; a reference plane is then generated based on the spatiotemporally fused point cloud.

[0052] The temperature acquisition module acquires surface temperature field distribution data of the TV back panel and determines the temperature change of each point in the spatiotemporal fusion point cloud. It also executes the partitioned thermal expansion coefficient call based on the difference in the thermal expansion coefficient of the materials between the metal corner area and the central thin-walled area.

[0053] The offset compensation module calculates the thermal deformation compensation vector for each point based on the material's thermal expansion coefficient and temperature change, and performs a reverse offset operation of vector superposition on the spatial position coordinates of the spatiotemporal fusion point cloud based on the thermal deformation compensation vector.

[0054] The planar detection module calculates the distance from each point in the compensated spatiotemporal fusion point cloud to the reference plane, and generates planarity detection results based on the wave deformation mode of the thin-walled region of the TV back panel and the standard deviation of the distance distribution.

[0055] Furthermore, the reference plane module includes:

[0056] The curvature calculation unit calculates the curvature representation value of each point in the spatiotemporal fused point cloud based on the rate of change of the normal vector.

[0057] The region division unit divides the spatiotemporal fusion point cloud into high curvature regions and low curvature regions based on the distribution range of curvature representation values;

[0058] The weighting unit assigns a first weighting coefficient to the high curvature region and a second weighting coefficient to the low curvature region.

[0059] The target construction unit uses the spatial coordinates of the spatiotemporal fusion point cloud as a reference and combines the first and second weighting coefficients to construct a plane fitting optimization target.

[0060] The fitting optimization unit iteratively adjusts the planar spatial pose parameters until the planar fitting optimization objective reaches a convergent state, and outputs a reference plane.

[0061] The technical solution of this invention can achieve the following technical effects:

[0062] By synchronously acquiring phase-encoded light stripes and pulsed laser reference point data, and using the time-stamped three-dimensional coordinates of the laser as a dynamic spatial reference to register the structured light point cloud, combined with temperature compensation and curvature adaptive weighted plane fitting, the technical problem of decreased flatness measurement accuracy caused by non-uniform thermal deformation and conveyor belt jitter coupling interference during the dynamic detection of the TV back panel was solved.

[0063] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0064] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0065] Figure 1 This is a flowchart illustrating a method for dynamic detection of flatness using structured light fusion laser calibration.

[0066] Figure 2 A flowchart illustrating the process of generating the three-dimensional coordinates of a laser reference point;

[0067] Figure 3 A flowchart illustrating the process of generating a reference plane;

[0068] Figure 4 A flowchart illustrating the process of performing three-dimensional coordinate compensation. Detailed Implementation

[0069] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0070] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0071] Example 1;

[0072] like Figure 1 As shown, this application provides a method for dynamic detection of flatness in structured light fusion laser calibration, the method comprising:

[0073] Phase-coded light stripes are projected onto the surface of the TV back panel and deformed images are acquired. Simultaneously, pulse-modulated laser beams are emitted to at least three non-collinear positioning points deployed in the metal corner area and the central thin-wall area and reflection displacement signals are acquired.

[0074] Based on the pulse flight time-domain demodulation of the reflection displacement signal and the compensation for atmospheric scattering distortion based on light intensity attenuation, the three-dimensional coordinates of the laser reference point are generated.

[0075] An initial structured light point cloud is generated based on the deformed image. The point cloud is then registered using the three-dimensional coordinates of the laser reference point as a dynamic spatial reference to generate a spatiotemporal fusion point cloud. A reference plane is generated based on the spatiotemporal fusion point cloud, wherein the normal vector constraint direction of the reference plane is collinear with the assembly reference plane of the composite structure of the TV back panel thin wall and the frame.

[0076] Acquire surface temperature field distribution data of TV back panel and determine the temperature change of each point in spatiotemporal fusion point cloud. Execute partitioned thermal expansion coefficient call based on the difference in thermal expansion coefficient between the metal corner area and the central thin-walled area.

[0077] The thermal deformation compensation vector of each point is calculated based on the material's thermal expansion coefficient and temperature change, and a reverse offset operation of vector superposition is performed on the spatial position coordinates of the spatiotemporal fusion point cloud based on the thermal deformation compensation vector.

[0078] The distance from each point in the compensated spatiotemporal fusion point cloud to the reference plane is calculated. Based on the wave deformation mode of the thin-walled region of the TV backplane, the flatness detection results are generated according to the standard deviation of the distance distribution.

[0079] Specifically, firstly, during the testing process, a phase-coded light source and a pulse-modulated laser are configured to project light stripes with unique phase codes and emit at least three non-collinear pulse-modulated laser beams onto the surface of the TV back panel under test. This ensures that a clear light stripe pattern and precise laser sampling points are formed on the entire surface of the TV back panel. A high-precision projection device is used to project the phase-coded light stripes onto the surface of the TV back panel. The pattern of the light stripes can be adaptively adjusted according to the material and surface characteristics of the TV back panel to ensure maximum sensitivity to light stripe deformation.Simultaneously with projection, the system employs at least three non-collinear lasers to emit pulsed modulated laser beams towards specific positioning points on the TV backplane surface. The non-collinear arrangement of the lasers ensures spatial diversity during dynamic measurements, thereby enhancing the stability and repeatability of the final measurement. Stripe projection and the acquisition of laser emission and reflection data are performed synchronously, i.e., completed within the same exposure cycle. This synchronous acquisition strategy effectively eliminates external light interference and dynamic errors caused by the reflective characteristics of the TV backplane surface. A high-speed camera or other suitable image acquisition device simultaneously acquires the distorted image of the phase-coded light stripes, and the image is then transmitted via photoelectric sensors positioned on the TV backplane surface. The detection device acquires the reflection displacement signal of pulsed laser light. The key is synchronous acquisition to ensure temporal and spatial consistency between the light stripe deformation data and the laser displacement data. Then, using the known speed of light and laser time of flight, the three-dimensional spatial coordinates of the laser reference point corresponding to the reflection signal are accurately calculated, and a timestamp is added to each coordinate. This timestamp information aids in subsequent spatiotemporal registration, ensuring the time traceability of the data stream. Next, based on the acquired deformation image, light stripe decoding is performed to generate initial structured light point cloud data. Using the calculated three-dimensional coordinates of the laser reference point as a dynamic spatial reference, coordinate transformation techniques, such as rigid body transformation, are employed. By using a homogeneous transformation matrix, the initial structured light point cloud is precisely registered to a common spatiotemporal coordinate system, thus forming an integrated spatiotemporal fused point cloud. In this process, the high-precision positioning function of the laser reference significantly improves the overall accuracy and time synchronization of the structured light point cloud. Then, the spatial curvature distribution of the generated spatiotemporal fused point cloud is calculated. This step specifically involves analyzing the rate of change of the normal vector in the point cloud, calculating the corresponding weights in each region, and further using the weighted least squares method to construct a stable reference plane. When dealing with large-area flow or environmentally affected measurement surfaces, this ensures that the obtained plane is as close as possible to the target surface. The initial surface is used to improve the reliability of the flatness results. In addition, to offset the impact of temperature changes during the inspection process, real-time surface temperature field distribution data is collected by infrared detection or by deploying temperature sensors on the surface of the TV back panel. The pre-stored material thermal expansion coefficient is embedded, and the three-dimensional coordinates of the spatiotemporal fusion point cloud are dynamically adjusted according to temperature changes to perform offset compensation, ensuring that external temperature changes during the inspection process do not cause result deviations. Finally, the shortest distance from each point in the spatiotemporal fusion point cloud after compensation processing to the reference plane is calculated, and the flatness inspection results of the TV back panel are generated by statistically analyzing the distance distribution of the entire point cloud, including the range and standard deviation.

[0080] The technical solution of this invention effectively solves the technical problem of decreased flatness measurement accuracy of TV back panel due to non-uniform thermal deformation and conveyor belt jitter coupling interference during dynamic testing.

[0081] Furthermore, such as Figure 2As shown, generating the three-dimensional coordinates of the laser reference point with a timestamp includes:

[0082] The reflection displacement signal is demodulated in the time domain to separate the pulse flight time sequence and the accompanying light intensity attenuation characteristic data corresponding to each positioning point.

[0083] Synchronize and align the pulse flight time sequence with the clock start time of the same exposure period to generate a timestamp for each positioning point;

[0084] Based on the light intensity attenuation characteristic data, the signal distortion caused by atmospheric scattering is compensated, and the initial three-dimensional coordinates of each positioning point in the laser measurement coordinate system are calculated by combining the emission angle of the pulse-modulated laser beam.

[0085] Establish positional constraints based on the spatial distribution characteristics of at least three non-collinear positioning points, optimize the accuracy of the initial three-dimensional coordinates based on least squares adjustment, and generate three-dimensional coordinates of the laser reference point with timestamps.

[0086] As a preferred embodiment of the above, the acquired laser reflection displacement signal first needs to undergo detailed time-domain demodulation processing. This process mainly includes spectral analysis of the signals fed back from each positioning point to separate the pulse flight time sequence of each positioning point. Simultaneously, to more completely describe the signal characteristics, light intensity attenuation characteristic data accompanying the pulse flight time is also extracted. After the demodulation step is completed and the time sequence of each positioning point is obtained, this sequence needs to be synchronized with the start time of the set exposure cycle clock. To ensure data consistency and timeliness, the timestamp of each positioning point must be accurately recorded to ensure multi-source data fusion in dynamic detection scenarios. After obtaining the timestamp synchronization, the signal distortion caused by atmospheric scattering is compensated based on the extracted light intensity attenuation characteristic data. The compensation step is adjusted based on the energy loss characteristics caused by laser propagation in the atmosphere, which improves the signal's true reflective ability to a certain extent. Combined with the specific emission angle of the pulse-modulated laser beam, the initial three-dimensional coordinates of each positioning point in the laser measurement coordinate system are calculated. The emission angle can be calibrated by a pre-established model or by using a reference plate to achieve accurate modeling. To further improve the measurement accuracy, a position constraint relationship network is constructed using the spatial distribution characteristics of at least three non-collinear positioning points. On this basis, the accuracy of the initial three-dimensional coordinates is optimized using the least squares adjustment method. This method provides higher fault tolerance while combining multi-dimensional data and adapts to complex and dynamically changing measurement environments, thereby generating three-dimensional coordinates of the laser reference point with precise timestamps.

[0087] Furthermore, optimizing the accuracy of the initial 3D coordinates based on least squares adjustment includes:

[0088] Based on the spatial distribution characteristics of at least three non-collinear positioning points, spatial position constraints are established with the measured distance and theoretical geometric relationship between adjacent positioning points as the benchmark.

[0089] Calculate the spatial position offset of the initial three-dimensional coordinates under spatial position constraints, and generate a set of coordinate correction values ​​for each positioning point;

[0090] By gradually reducing the spatial position offset through multiple position adjustment operations, the spatial position of each positioning point meets the matching requirements of the theoretical geometric relationship.

[0091] When the spatial offset of all positioning points is less than the preset threshold, the optimized three-dimensional coordinates are output as the three-dimensional coordinates of the laser reference point with timestamp.

[0092] As a preferred embodiment of the above, firstly, spatial position constraints need to be constructed based on the spatial distribution characteristics of at least three non-collinear positioning points. The main task of this step is to determine the measured distance between adjacent positioning points and compare it with the theoretical geometric expectation. In specific implementation, a high-precision laser rangefinder or other equivalent measuring device can be used to accurately obtain the measured distance between adjacent points. After obtaining these measured values, based on theoretical geometric relationships, such as based on known position settings or an established 3D model of the TV back panel, these data are corrected to establish spatial position constraints based on these real distances. This step creates a solid standard baseline for subsequent coordinate optimization, giving the optimization process a clear reference system. Secondly, the offset of the initial 3D coordinates under the above spatial position constraints is calculated. For each positioning point, the spatial offset of its initial coordinates relative to the ideal coordinates under the constraints is calculated. This offset is calculated using the 3D spatial difference between the actual measured value and the ideal position. Through this calculation, the position that each positioning point needs to be adjusted can be identified, thereby generating a set of coordinate corrections. This set will serve as the basis for coordinate adjustment in the subsequent optimization process. On this basis, a series of multiple position adjustment operations are then performed. Specifically, the optimization process involves gradually reducing the spatial offset to continuously adjust the coordinates of each positioning point. An iterative method is employed, allowing for recalculation and updating of the offset after each iteration. This ensures that as the number of iterations increases, the coordinates of all points approach their theoretical ideal state. During this process, least squares adjustment is used for coordinate correction, employing the criterion of minimizing the sum of squared overall errors to guide specific coordinate adjustments, ensuring the final positioning point's location approximates the theoretical standard. To guarantee the efficiency and effectiveness of the adjustment, historical data can be used to estimate the initial deviation, and real-time feedback can be incorporated during the iteration process for path optimization. For certain specific application scenarios, machine learning algorithms can be introduced to further improve the efficiency and accuracy of the iteration process through the analysis of large amounts of data. When the spatial offset of all positioning points is reduced to less than a preset threshold, the optimization process is considered successful. At this point, the optimized 3D coordinate output can be used as the latest version of the timestamped laser reference point 3D coordinates for subsequent flatness detection tasks. The preset threshold is an error limit set according to actual application requirements; its selection needs to be adjusted based on the specific measurement accuracy requirements to ensure the system's flexibility and broad applicability.

[0093] Furthermore, such as Figure 3 As shown, the reference plane is generated according to the weighted least squares method, including:

[0094] Calculate the curvature representation value of each point in the spatiotemporal fused point cloud based on the rate of change of the normal vector;

[0095] Based on the distribution range of curvature characterization values, the spatiotemporal fusion point cloud is divided into high curvature region and low curvature region;

[0096] A first weighting coefficient is assigned to the high curvature region, and a second weighting coefficient is assigned to the low curvature region, wherein the first weighting coefficient is greater than the second weighting coefficient;

[0097] Based on the spatial coordinates of the spatiotemporal fusion point cloud, a plane fitting optimization objective is constructed by combining the first and second weighting coefficients.

[0098] Iteratively adjust the planar spatial pose parameters until the planar fitting optimization objective reaches a convergent state, and output the reference plane.

[0099] As a preferred embodiment of the above, firstly, based on the rate of change of the normal vector of each data point in the spatiotemporal fusion point cloud, the curvature representation value of each point is calculated. The rate of change of the normal vector can be obtained by calculating the difference between the normal vectors of adjacent points. By calculating these curvature representation values, the surface features of different regions in the point cloud can be identified, which provides a standard for identification and classification for subsequent operations. According to the calculated curvature representation values, the entire spatiotemporal fusion point cloud is divided into high curvature regions and low curvature regions. This division is based on the curvature value reaching a certain threshold for boundary, and more detailed region identification can be achieved by setting multiple levels of curvature thresholds. This process involves different geometric complexities in the point cloud. The distinction allows for differentiated treatment of complex curved surfaces and flat regions. After defining the region division, weights are assigned to different regions. For high-curvature regions, a higher first weight coefficient is assigned to reflect the region's importance and complexity. Generally, high-curvature regions are more susceptible to errors due to their drastic changes, thus receiving greater attention during optimization. For low-curvature regions, a relatively lower second weight coefficient is assigned because these regions are usually relatively stable and less affected by errors. Next, combining the different weight coefficients for high-curvature and low-curvature regions, and using the spatial coordinates of the spatiotemporal fusion point cloud as a benchmark, an optimization objective for plane fitting is constructed. By combining two types of weights to constrain the point cloud data, the optimization process can better adapt to the real geometric characteristics of the point cloud while effectively avoiding overfitting, thereby improving the accuracy and robustness of the reference plane fitting. Finally, by iteratively adjusting the planar spatial pose parameters, the plane fitting optimization objective reaches a convergent state. In practice, numerical optimization algorithms are usually used to gradually adjust the planar spatial pose parameters. After multiple iterations, it is confirmed that the optimization objective has reached the preset convergence criterion, and the final fitted reference plane can be output. This plane provides a standard reference for subsequent flatness evaluation, and detection is carried out under its guidance to ensure detection accuracy and reliability.

[0100] Furthermore, constructing the plane fitting optimization objective includes:

[0101] Based on the spatial coordinates of the spatiotemporal fusion point cloud, and combined with the first and second weighting coefficients, the plane fitting optimization objective is defined as minimizing the cumulative weighted distance deviation from each point in the spatiotemporal fusion point cloud to the reference plane.

[0102] The weighting of the cumulative weighted distance deviation is based on the division of high curvature region and low curvature region, so that the points in high curvature region are weighted by the first weighting coefficient and the points in low curvature region are weighted by the second weighting coefficient.

[0103] The optimization constraints of the planar spatial pose parameters are established based on the weighted cumulative distance deviation to ensure that the plane fitting optimization target reflects the influence of the curvature characterization value on the measurement reliability.

[0104] The plane fitting optimization objective is associated with the iterative adjustment process to drive the adjustment direction of the planar spatial pose parameters until convergence is achieved.

[0105] As a preferred embodiment of the above, firstly, based on the spatial coordinates of the spatiotemporal fusion point cloud and combined with the previously classified high-curvature and low-curvature regions, the optimization objective of plane fitting is defined as minimizing the cumulative weighted distance deviation from each point in the spatiotemporal fusion point cloud to the reference plane. In specific implementation, the distance deviation is measured by the distance of each point relative to the reference plane. The cumulative deviation of all points is weighted using different weighting coefficients. For points in high-curvature regions, due to the higher surface complexity, there may be greater geometric errors; therefore, a larger first weighting coefficient is assigned to emphasize their importance. Points in low-curvature regions have smoother surfaces and lower geometric errors. If the value is small, a lower second weight coefficient is assigned, thus appropriately weakening its influence during the plane fitting process. Next, based on the expression of the weighted distance deviation accumulation, optimization constraints are established to adjust the planar spatial pose parameters. First, in the specific calculation process, considering that points in high curvature regions contribute more geometric complexity, while points in low curvature regions have an auxiliary effect on overall stability, the optimization constraints must ensure that the plane fitting target can fully reflect this characteristic. By embedding the influence of curvature characterization values ​​into the optimization constraints, geometric realism and measurement reliability are given to the optimization constraints, making the adjustment direction of the fitting plane more adaptable to the actual distribution and characteristics of the spatiotemporal point cloud; for the real The dynamic adjustment of planar spatial pose parameters closely links the plane fitting optimization objective with the iterative adjustment process. The specific steps are as follows: In each iteration, based on the deviation data of the current plane fitting result, the adjustment direction of the planar spatial pose parameters is driven, continuously reducing the weighted cumulative distance deviation. In practical scenarios, to avoid computational stagnation caused by local optima, dynamic convergence conditions can be introduced, such as setting the change range of the cumulative distance deviation or a predefined target change threshold. As the iteration progresses, when the optimization objective reaches a convergence state, the final fitted reference plane can be output. To more clearly describe the specific application scenarios of this process, we can take the surface of a TV back panel as an example. Explanation: Assuming the spatiotemporal fusion point cloud contains a complex surface and some flat regions, the curvature value is calculated by adjusting the rate of change of the point cloud normal vector. It is found that the high curvature region of the surface has a large width and significant nonlinear changes. When fitting the reference plane, the system assigns a high weight to the high curvature region to ensure that the influence of this region on the reference plane is more significant during the fitting process, thereby achieving the optimization goal. In addition, it should be noted that during the iterative adjustment process, the adjustment of the planar spatial pose parameters not only needs to improve the accuracy of the fitting results, but also needs to consider the measurement characteristics in the actual application scenario, such as the impact of temperature changes on the curvature calculation results. Therefore, this optimization process is a dynamic operation that integrates multiple factors.

[0106] Furthermore, such as Figure 4 As shown, the three-dimensional coordinate offset compensation for spatiotemporal fusion point clouds is performed by combining the pre-stored material thermal expansion coefficient, including:

[0107] Acquire surface temperature field distribution data of the TV back panel, and determine the temperature change of each point in the spatiotemporal fusion point cloud based on the surface temperature field distribution data;

[0108] Based on the pre-stored material thermal expansion coefficient and temperature change, calculate the thermal deformation compensation vector of each point in the spatiotemporal fused point cloud.

[0109] The spatial coordinates of the spatiotemporal fusion point cloud are adjusted by reverse offset based on the thermal deformation compensation vector to generate a compensated spatiotemporal fusion point cloud to eliminate measurement errors caused by temperature.

[0110] As a preferred embodiment of the above, it is first necessary to acquire the temperature field distribution data of the TV back panel surface. This is achieved by deploying high-precision temperature sensors or infrared imaging equipment. The temperature sensors can be deployed in multiple key areas of the TV back panel to cover the entire surface and ensure the spatial resolution of the temperature data. The infrared imaging equipment can be used to capture surface temperature changes in real time. By synchronizing with real-time structured light and laser measurement data, the dynamic consistency of the temperature field distribution data is ensured. In practical applications, to further improve the accuracy of the temperature field distribution, multiple data acquisitions and corrections can be performed on the same area, and the data can be filled in using a temperature field interpolation algorithm. To fill the gaps, a continuous surface temperature field is constructed. After acquiring the temperature field distribution data, the temperature change at each point is calculated based on the specific temperature data. Under dynamic detection conditions, the surface of a TV back panel often experiences non-uniform temperature changes due to environmental factors or processing. For example, a local area on the surface of a TV back panel may experience a temperature increase due to a heat source during processing, resulting in non-uniform expansion and offset differences at the points. In this step, the temperature data is matched point-by-point with the spatiotemporal fusion point cloud to ensure that the temperature change of each data point is aligned to its actual spatial location. The determination of the temperature change can be combined with the initial ambient temperature. The calculations are performed based on the real-time temperature changes measured during the process. Based on the pre-stored coefficient of thermal expansion of the material and the temperature change at each data point, the thermal deformation compensation vector for each point in the spatiotemporally fused point cloud is calculated. The coefficient of thermal expansion is a physical parameter that varies with different materials and is usually pre-stored in the system database for use with specific TV back panel materials. For each point, the thermal expansion displacement in the corresponding direction is estimated using the coefficient of thermal expansion and the temperature change at that point. These compensation vectors reflect the degree of offset of the point cloud data due to thermal effects. To ensure the accuracy of the compensation vector calculations, finite element analysis can be used. Thermal expansion behavior is simulated and verified to make the compensation data closer to the actual expansion effect. Finally, the thermal deformation compensation vector calculated above is used to perform reverse offset adjustment on the spatiotemporal fusion point cloud to eliminate measurement errors caused by temperature. During the adjustment process, the compensation vector is applied point by point to the spatial position coordinates of the structured light point cloud and the laser point cloud and the positioning is corrected. The thermal additional error of the entire point cloud is reversed through batch processing to generate a new compensated spatiotemporal fusion point cloud. In the specific implementation, in order to verify the accuracy of the compensated point cloud, it can be compared with a standard reference plane to evaluate the data error after compensation and optimize the compensation process.For example, when inspecting the flatness of a large metal television back panel, due to the unevenness of the ambient temperature, the high coefficient of thermal expansion of the metal material causes significant displacement in some areas. During inspection, the system first collects real-time temperature field distribution data on the surface of the television back panel to determine the temperature change at each point. Based on this data, using the coefficient of thermal expansion of the metal material, a compensation vector is calculated for each region. Subsequently, a reverse adjustment is performed based on the compensation vector. The resulting compensated point cloud eliminates the displacement deviation caused by temperature, making the final flatness inspection more accurate.

[0111] Furthermore, the spatial coordinates of the spatiotemporally fused point cloud are adjusted by reverse offset based on the thermal deformation compensation vector, including:

[0112] Establish a mapping relationship between the thermal deformation compensation vector and the spatial coordinates of each point in the spatiotemporal fusion point cloud, forming a coordinate compensation mapping relationship;

[0113] Based on the coordinate compensation mapping relationship, a vector superposition operation is performed on the spatial position coordinates of the spatiotemporal fusion point cloud, so that the coordinates of each point are shifted in the opposite direction of the thermal deformation compensation vector.

[0114] The spatiotemporal fusion point cloud is updated based on the spatial position coordinates after vector superposition operation, and a compensated spatiotemporal fusion point cloud is generated. The relative pose of the compensated spatiotemporal fusion point cloud and the reference plane reflects the true geometric relationship after temperature compensation.

[0115] As a preferred embodiment of the above, firstly, it is necessary to establish a mapping relationship between the thermal deformation compensation vector and the spatial coordinates of each point in the spatiotemporal fusion point cloud. Specifically, by collecting and matching the temperature change data and the thermal expansion coefficient of the material at each point on the TV back panel, a thermal deformation compensation vector in the target direction is calculated for each point. These vectors reflect the positional shift of the point cloud data due to temperature changes in practical applications. Since the actual shift of each point may vary due to uneven temperature fields, this mapping relationship needs to support data compensation down to the single point level. The next step is to perform a vector superposition operation on the spatial coordinates of the spatiotemporal fusion point cloud based on the established coordinate compensation mapping relationship. The key to this process is to add the opposite direction value of the corresponding thermal deformation compensation vector to the current coordinate value of each point. This means that points that have been shifted by the expansion effect caused by temperature are directionally corrected, so that the position of the point returns to a state closer to reality. To improve the accuracy of the operation, when performing coordinate correction, the same point can be subjected to... Multiple compensation rehearsals were conducted, and the results were recorded and reviewed to ensure that the adjusted coordinates were highly consistent with the expected results. After completing the vector inverse superposition operation, a compensated spatiotemporal fusion point cloud could be generated based on the updated spatial coordinates. This update ensured that the coordinates of all points reflected their true geometric state after temperature compensation. When generating the compensated point cloud, its effectiveness could be verified by further comparing it with an ideal reference plane, and the overall geometric compliance of the point cloud could be optimized by adjusting the compensation strategy. For example, in the production process of a high-precision component, a local area of ​​a TV back panel faced complex temperature changes under heated conditions. The thermal deformation compensation vector calculated based on the previously measured temperature field distribution reflected its expansion effect in the high-heat region. By applying the above-mentioned vector inverse superposition, the coordinate deviation of the measurement points in this part was accurately corrected, ensuring that even under non-uniform temperature change conditions, the final detection point cloud was consistent with the initial design geometry, thereby ensuring that the overall assembly and functional characteristics were not affected.

[0116] Furthermore, flatness test results are generated based on the range and standard deviation of the distance distribution, including:

[0117] Based on the distance distribution from each point in the compensated spatiotemporal fusion point cloud to the reference plane, the difference between the maximum and minimum values ​​of the distance distribution is calculated as the range;

[0118] The standard deviation is calculated based on the degree of dispersion of the distance values ​​of each point in the distance distribution, which characterizes the overall fluctuation range of the compensated spatiotemporal fusion point cloud relative to the reference plane.

[0119] The range is used as an overall deviation measure and the standard deviation is used as a local fluctuation measure. The flatness test results are combined to generate the flatness test results, which are used to quantify the flatness quality of the TV back panel surface.

[0120] As a preferred embodiment of the above, firstly, the vertical distance from each point in the compensated spatiotemporal fusion point cloud to the reference plane is obtained. These distances constitute a complete distance distribution dataset. This distance data is extracted based on the spatiotemporal fusion point cloud after coordinate compensation, ensuring that the starting point of the calculation has accuracy after temperature effect adjustment. Before analyzing these distances, the specific deviation of each point needs to be confirmed to ensure the validity and accuracy of the basic data. Next, using the obtained distance distribution information, the range of the distribution is calculated. The range, i.e., the difference between the maximum and minimum values ​​in the distance distribution, directly reflects the range of the overall deviation of the TV back panel surface, providing an overall deviation metric in the flatness detection results. The focus of this step is... To accurately identify the maximum and minimum distances in the distance distribution, ensuring that the range values ​​accurately reflect the distance changes between the highest and lowest points on the TV back panel surface; subsequently, the standard deviation of the distance distribution is calculated. This standard deviation is an indicator of the dispersion of distance values, characterizing the overall fluctuation range of the compensated spatiotemporal fusion point cloud relative to the reference plane. In practice, the standard deviation is calculated by taking the square root of the mean of the squares of the deviations of each point in the entire distance distribution from the mean. This indicator can meticulously display the local fluctuation characteristics of the TV back panel surface; that is, the standard deviation provides an average measure of the surface unevenness, indicating the contribution of subtle surface fluctuations to the overall flatness. Finally, the range and standard deviation are combined to generate flatness detection results. The range serves as an overall deviation measure, while the standard deviation serves as a local fluctuation measure. Together, they constitute the final result of flatness inspection. The range allows inspectors to quickly perceive the range of overall height differences on the TV back panel surface, helping to identify potential large deviation problems. The standard deviation provides a statistical range analysis of small surface fluctuations. This combination enables the inspection results to systematically express the nature of different problems.

[0121] Example 2;

[0122] Based on the same inventive concept as the planarity dynamic detection method for structured light fusion laser calibration in the foregoing embodiments, the present invention also provides a planarity dynamic detection system for structured light fusion laser calibration, the system comprising:

[0123] The laser modulation module projects phase-coded light stripes onto the surface of the TV back panel and acquires deformed images, while simultaneously emitting pulsed modulated laser beams to at least three non-collinear positioning points deployed in the metal corner area and the central thin-wall area and acquiring reflection displacement signals.

[0124] The coordinate generation module generates three-dimensional coordinates of the laser reference point by demodulating the reflection displacement signal in the time domain based on the pulse flight time and compensating for atmospheric scattering distortion based on light intensity attenuation.

[0125] The reference plane module generates an initial structured light point cloud based on the deformed image, and performs registration to generate a spatiotemporally fused point cloud using the three-dimensional coordinates of the laser reference point as a dynamic spatial reference; a reference plane is then generated based on the spatiotemporally fused point cloud.

[0126] The temperature acquisition module acquires surface temperature field distribution data of the TV back panel and determines the temperature change of each point in the spatiotemporal fusion point cloud. It also executes the partitioned thermal expansion coefficient call based on the difference in the thermal expansion coefficient of the materials between the metal corner area and the central thin-walled area.

[0127] The offset compensation module calculates the thermal deformation compensation vector for each point based on the material's thermal expansion coefficient and temperature change, and performs a reverse offset operation of vector superposition on the spatial position coordinates of the spatiotemporal fusion point cloud based on the thermal deformation compensation vector.

[0128] The planar detection module calculates the distance from each point in the compensated spatiotemporal fusion point cloud to the reference plane, and generates planarity detection results based on the wave deformation mode of the thin-walled region of the TV back panel and the standard deviation of the distance distribution.

[0129] The adjustment system described above in this invention can effectively realize the dynamic detection method of flatness for structured light fusion laser calibration, and the technical effects it can achieve are as described in the above embodiments, and will not be repeated here.

[0130] Furthermore, the reference plane module includes:

[0131] The curvature calculation unit calculates the curvature representation value of each point in the spatiotemporal fused point cloud based on the rate of change of the normal vector.

[0132] The region division unit divides the spatiotemporal fusion point cloud into high curvature regions and low curvature regions based on the distribution range of curvature representation values;

[0133] The weighting unit assigns a first weighting coefficient to the high curvature region and a second weighting coefficient to the low curvature region.

[0134] The target construction unit uses the spatial coordinates of the spatiotemporal fusion point cloud as a reference and combines the first and second weighting coefficients to construct a plane fitting optimization target.

[0135] The fitting optimization unit iteratively adjusts the planar spatial pose parameters until the planar fitting optimization objective reaches a convergent state, and outputs a reference plane.

[0136] Similarly, the above-mentioned optimization schemes for the system can also achieve the optimization effects corresponding to the methods in Embodiment 1, which will not be repeated here.

[0137] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely exemplary illustrations of the application as defined herein, and are to be considered as covering any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Thus, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for dynamic detection of flatness using structured light fusion laser calibration, characterized in that, The method includes: Phase-coded light stripes are projected onto the surface of the TV back panel and deformed images are acquired. Simultaneously, pulse-modulated laser beams are emitted to at least three non-collinear positioning points deployed in the metal corner area and the central thin-wall area and reflection displacement signals are acquired. Based on the pulse flight time-domain demodulation of the reflected displacement signal and the compensation for atmospheric scattering distortion based on light intensity attenuation, the three-dimensional coordinates of the laser reference point are generated. An initial structured light point cloud is generated based on the deformed image. A spatiotemporal fusion point cloud is generated by registration using the three-dimensional coordinates of the laser reference point as a dynamic spatial reference. A reference plane is generated based on the spatiotemporal fusion point cloud, wherein the normal vector constraint direction of the reference plane is collinear with the assembly reference plane of the composite structure of the TV back panel thin wall and the frame. Acquire the surface temperature field distribution data of the TV back panel and determine the temperature change of each point in the spatiotemporal fusion point cloud. Execute the partitioned thermal expansion coefficient call based on the difference in the material thermal expansion coefficients of the metal corner area and the central thin-walled area. The thermal deformation compensation vector of each point is calculated based on the thermal expansion coefficient of the material and the temperature change, and a reverse offset operation of vector superposition is performed on the spatial position coordinates of the spatiotemporal fusion point cloud based on the thermal deformation compensation vector. The distance from each point in the compensated spatiotemporal fusion point cloud to the reference plane is calculated. Based on the wave deformation mode of the thin-walled region of the TV back panel, a flatness detection result is generated according to the standard deviation of the distance distribution.

2. The method for dynamic detection of flatness of structured light fusion laser calibration according to claim 1, characterized in that, Generate the 3D coordinates of the laser reference point with timestamps, including: The reflection displacement signal is demodulated in the time domain to separate the pulse flight time sequence and the accompanying light intensity attenuation characteristic data corresponding to each positioning point. The pulse flight time sequence is synchronized and aligned with the start time of a clock with the same exposure period to generate a timestamp for each positioning point; Based on the light intensity attenuation characteristic data, the signal distortion caused by atmospheric scattering is compensated, and the initial three-dimensional coordinates of each positioning point in the laser measurement coordinate system are calculated in combination with the emission angle of the pulse-modulated laser beam. Based on the spatial distribution characteristics of the at least three non-collinear positioning points, positional constraints are established, and the accuracy of the initial three-dimensional coordinates is optimized based on least squares adjustment to generate the three-dimensional coordinates of the laser reference point with the timestamp.

3. The method for dynamic detection of flatness of structured light fusion laser calibration according to claim 2, characterized in that, The accuracy of the initial three-dimensional coordinates is optimized based on least squares adjustment, including: Based on the spatial distribution characteristics of the at least three non-collinear positioning points, spatial position constraints are established with the measured distance and theoretical geometric relationship between adjacent positioning points as the benchmark. Calculate the spatial position offset of the initial three-dimensional coordinates under the spatial position constraint conditions, and generate a set of coordinate correction values ​​for each positioning point; By gradually reducing the spatial position offset through multiple position adjustment operations, the spatial position of each positioning point satisfies the matching requirements of the theoretical geometric relationship. When the spatial offset of all positioning points is less than a preset threshold, the optimized three-dimensional coordinates are output as the three-dimensional coordinates of the timestamped laser reference point.

4. The method for dynamic detection of flatness of structured light fusion laser calibration according to claim 1, characterized in that, The reference plane is generated using the weighted least squares method, including: The curvature representation value of each point in the spatiotemporal fused point cloud is calculated based on the rate of change of the normal vector. Based on the distribution range of the curvature characterization values, the spatiotemporal fusion point cloud is divided into high curvature regions and low curvature regions; A first weighting coefficient is assigned to the high curvature region, and a second weighting coefficient is assigned to the low curvature region, wherein the first weighting coefficient is greater than the second weighting coefficient; Based on the spatial coordinates of the spatiotemporal fusion point cloud, a plane fitting optimization target is constructed by combining the first weighting coefficient and the second weighting coefficient; The planar spatial pose parameters are iteratively adjusted until the planar fitting optimization objective reaches a convergent state, and the reference plane is output.

5. The method for dynamic detection of flatness of structured light fusion laser calibration according to claim 4, characterized in that, Construct a plane fitting optimization objective, including: Based on the spatial coordinates of the spatiotemporal fusion point cloud, and combined with the first weighting coefficient and the second weighting coefficient, the objective of the plane fitting optimization is defined as minimizing the cumulative weighted distance deviation from each point in the spatiotemporal fusion point cloud to the reference plane. The weighting of the cumulative weighted distance deviation is based on the division results of the high curvature region and the low curvature region, so that the points in the high curvature region are weighted by the first weighting coefficient and the points in the low curvature region are weighted by the second weighting coefficient. The optimization constraints of the planar spatial pose parameters are established based on the weighted distance deviation accumulation to ensure that the planar fitting optimization target reflects the influence of the curvature characterization value on the measurement reliability. The plane fitting optimization objective is associated with the iterative adjustment process to drive the adjustment direction of the plane spatial pose parameters until the convergence state is reached.

6. The method for dynamic detection of flatness of structured light fusion laser calibration according to claim 1, characterized in that, Incorporating compensation based on the thermal expansion coefficient of pre-stored materials, including: Acquire the surface temperature field distribution data of the TV back panel, and determine the temperature change of each point in the spatiotemporal fusion point cloud based on the surface temperature field distribution data; Based on the pre-stored material thermal expansion coefficient and the temperature change, calculate the thermal deformation compensation vector of each point in the spatiotemporal fusion point cloud; Based on the thermal deformation compensation vector, the spatial position coordinates of the spatiotemporal fusion point cloud are adjusted by reverse offset to generate the compensated spatiotemporal fusion point cloud to eliminate measurement errors caused by temperature.

7. The method for dynamic detection of flatness of structured light fusion laser calibration according to claim 6, characterized in that, The spatial position coordinates of the spatiotemporal fusion point cloud are adjusted by reverse offset based on the thermal deformation compensation vector, including: Establish a mapping relationship between the thermal deformation compensation vector and the spatial position coordinates of each point in the spatiotemporal fusion point cloud, forming a coordinate compensation mapping relationship; According to the coordinate compensation mapping relationship, a vector superposition operation is performed on the spatial position coordinates of the spatiotemporal fusion point cloud, so that the coordinates of each point are shifted in the opposite direction of the thermal deformation compensation vector; The spatiotemporal fusion point cloud is updated based on the spatial position coordinates after the vector superposition operation to generate the compensated spatiotemporal fusion point cloud, wherein the relative pose of the compensated spatiotemporal fusion point cloud and the reference plane reflects the true geometric relationship after temperature compensation.

8. The method for dynamic detection of flatness of structured light fusion laser calibration according to claim 1, characterized in that, Generate flatness test results, including: Based on the distance distribution from each point in the compensated spatiotemporal fusion point cloud to the reference plane, the difference between the maximum and minimum values ​​of the distance distribution is calculated as the range; The standard deviation is calculated based on the dispersion of the distance values ​​of each point in the distance distribution, which characterizes the overall fluctuation range of the compensated spatiotemporal fusion point cloud relative to the reference plane. The range is used as an overall deviation metric, and the standard deviation is used as a local fluctuation metric. The flatness test result is generated by combining these two parameters, and the flatness test result is used to quantify the flatness quality of the TV back panel surface.

9. A dynamic flatness detection system based on structured light fusion laser calibration, characterized in that, The system includes: The laser modulation module projects phase-coded light stripes onto the surface of the TV back panel and acquires deformed images, while simultaneously emitting pulsed modulated laser beams to at least three non-collinear positioning points deployed in the metal corner area and the central thin-wall area and acquiring reflection displacement signals. The coordinate generation module generates three-dimensional coordinates of the laser reference point by demodulating the reflection displacement signal in the time domain based on the pulse flight time and compensating for atmospheric scattering distortion based on light intensity attenuation. The reference plane module generates an initial structured light point cloud based on the deformed image, and performs registration to generate a spatiotemporally fused point cloud using the three-dimensional coordinates of the laser reference point as a dynamic spatial reference; a reference plane is then generated based on the spatiotemporally fused point cloud. The temperature acquisition module acquires surface temperature field distribution data of the TV back panel and determines the temperature change of each point in the spatiotemporal fusion point cloud. It also executes the partitioned thermal expansion coefficient call based on the difference in the thermal expansion coefficient of the materials between the metal corner area and the central thin-walled area. The offset compensation module calculates the thermal deformation compensation vector for each point based on the material's thermal expansion coefficient and temperature change, and performs a reverse offset operation of vector superposition on the spatial position coordinates of the spatiotemporal fusion point cloud based on the thermal deformation compensation vector. The planar detection module calculates the distance from each point in the compensated spatiotemporal fusion point cloud to the reference plane, and generates planarity detection results based on the wave deformation mode of the thin-walled region of the TV back panel and the standard deviation of the distance distribution.

10. The structured light fusion laser calibration flatness dynamic detection system according to claim 9, characterized in that, The reference plane module includes: The curvature calculation unit calculates the curvature representation value of each point in the spatiotemporal fused point cloud based on the rate of change of the normal vector. The region division unit divides the spatiotemporal fusion point cloud into high curvature regions and low curvature regions based on the distribution range of curvature representation values; The weighting unit assigns a first weighting coefficient to the high curvature region and a second weighting coefficient to the low curvature region. The target construction unit uses the spatial coordinates of the spatiotemporal fusion point cloud as a reference and combines the first and second weighting coefficients to construct a plane fitting optimization target. The fitting optimization unit iteratively adjusts the planar spatial pose parameters until the planar fitting optimization objective reaches a convergent state, and outputs a reference plane.

Images