Programmable projector image edge fusion system

Abstract

The invention relates to the technical field of image fusion, in particular to a programmable projector image edge fusion system. The method comprises the following steps: establishing a system hardware architecture consisting of a programmable optical edge fusion board POEP, a fixed mechanical structure and a control computer, and acquiring multi-channel projection configuration parameters of a target projection scene; network communication connection with each POEP is established through a control computer, and a POEP basic control parameter set is generated; dividing a fusion region and a non-fusion region by combining pixel array distribution of POEP, and generating initial fusion control image data; optical edge modulation is carried out on the initial fusion control image data, and evaluation is carried out to generate effect optimization feedback information; and updating fusion control image data, re-issuing the fusion control image data to the POEP, and repeatedly executing optical edge modulation and evaluation feedback until the projection picture reaches a preset edge fusion standard. According to the method, the optical edge fusion parameters can be quickly corrected, so that the fusion effect is quickly recovered.

Description

Technical Field

[0001] This invention relates to the field of image fusion technology, and in particular to a programmable projector image edge fusion system. Background Technology

[0002] In fields such as virtual reality, exhibitions, and professional simulations (such as flight simulators), multi-channel projection systems have become a core visual output solution due to their advantages of ultra-large size and immersive display. Edge blending technology, as a key support for multi-channel projection, eliminates splicing artifacts by creating a 10%-25% overlap between adjacent projected images and adjusting the brightness gradient of the overlap area. This directly determines the integrity and visual continuity of the displayed image.

[0003] Furthermore, Chinese patent CN108269231A discloses an image fusion method and device for dome screen systems. This invention utilizes a grid model of the spherical screen to determine the projection area of ​​each projector, enabling adjustable projection positions. Polar coordinate geometric correction is used to perform edge blending of each projection area, achieving precise stitching of images from projection areas of different positions and sizes. Color unification and light leakage compensation are employed for projector color calibration and image fusion with blanking operations, effectively avoiding color and brightness differences between edge-blended images. The steps provided by this method result in a complete, color-uniform, and moderately bright stereoscopic projected image on the dome screen, providing viewers with a comfortable and immersive viewing experience. The image fusion device using this method can produce a standard blended image with excellent display effects, while the dome screen system using this method can achieve superior projection effects, enhancing the viewer's experience. This invention optimizes the blending accuracy and display effect in dome scenes, but it focuses on basic compensation for geometric correction, color, and light leakage. It does not offer an effective solution to the core problem of "black light superimposed on bright bands" in black scenes, and it does not involve programmable optical control design, so it cannot achieve flexible adjustment and maintenance of blending parameters and is difficult to adapt to edge blending needs in multiple scenes and complex environments. Summary of the Invention

[0004] To address the aforementioned technical problems in existing projector image edge blending processes, this invention provides a programmable projector image edge blending system that uses a POEP hardware module to perform optical edge modulation on the projected image. Combined with grayscale gradient rules in the blending region, this achieves a uniform brightness transition in the overlapping area, completely eliminating bright and dark bands. Furthermore, relying on programmable optical design and network communication control, the POEP pixel transmittance and blending parameters can be flexibly adjusted, overcoming the limitations of traditional solutions that cannot adapt flexibly. This meets the blending requirements of multi-channel and complex environments, thereby ensuring uniform brightness, natural transitions, and image integrity in the projected image. The system includes: A system hardware architecture consisting of a programmable optical edge blending board (POEP), a fixed mechanical structure, and a control computer is constructed. Based on the system hardware architecture, multi-channel projection configuration parameters of the target projection scene are obtained, including the number of projectors, the projection coverage area, and the image overlap ratio of adjacent projectors. Based on the multi-channel projection configuration parameters of the target projection scene, determine the adaptation and installation parameters between each projector lens and POEP, and construct the spatial correspondence between POEP and projector; establish network communication connection with each POEP through the control computer, and generate the POEP basic control parameter set based on the hardware characteristic parameters of POEP; Based on the overlap ratio of adjacent projector images and the pixel array distribution of POEP, the fusion region and non-fusion region corresponding to each POEP are divided, and the grayscale gradient rules within the fusion region are defined to generate initial fusion control image data. The initial blending control image data is transmitted to the corresponding POEP via network communication. The POEP's hardware module receives and stores the data, and simultaneously controls each pixel of the POEP to adjust its transmittance according to the grayscale value of the initial blending control image data. Optical edge modulation is then applied to the image projected by the projector to generate the optically edge-modulated projected image. The edge blending effect of the projected image is evaluated through visual observation, and effect optimization feedback information is generated based on the edge blending effect. Based on the effect optimization feedback information, the grayscale gradient rules of the blending area and the pixel transmittance parameters of the POEP are adjusted by the control computer. The blending control image data is updated and re-sent to the POEP. The optical edge modulation and evaluation feedback are repeated until the projected image reaches the preset edge blending standard, thereby completing the programmable projector image edge blending.

[0005] This invention utilizes a programmable optical edge blending board (POEP) with hardware capabilities and a grayscale gradient rule design for the blending area to perform optical edge modulation on the projected image. This allows for a natural and uniform brightness transition in the overlapping area, completely eliminating interference from bright and dark bands and significantly optimizing the blending display effect in black fields and various brightness scenarios. Regarding parameter control and maintenance, relying on the programmable optical control design and network communication architecture, the core parameters of the POEP, such as pixel transmittance and grayscale gradient rules for the blending area, can be flexibly adjusted via a control computer. Compared to the shortcomings of traditional solutions that lack flexible adjustment mechanisms, this invention achieves dynamic optimization and convenient maintenance of blending parameters, reducing subsequent operating costs. Furthermore, the solution employs a complete process of "acquiring configuration parameters - determining installation parameters - generating control data - modulation evaluation - optimization iteration," enabling personalized configuration based on the number of projectors, coverage area, and overlap ratio of the target projection scene. It also establishes a spatial correspondence between the POEP and the projector, adapting to diverse edge blending needs in multi-channel and complex environments, breaking the limitations of traditional solutions that can only adapt to specific scenarios. In addition, through visual observation and evaluation and closed-loop optimization mechanisms, the brightness uniformity, transition state and image integrity of the projected image are continuously calibrated to ensure that the final projected image meets the preset standards and presents a high-quality effect with natural image splicing, color transition and overall display coordination. It is suitable for various projection scenarios such as dome screens, large exhibition halls, and conference centers, thereby bringing users an immersive and high-quality visual experience.

[0006] Preferably, the process of obtaining the multi-channel projection configuration parameters of the target projection scene includes: Select a customized transmissive OLED or LCD screen as the POEP and determine the POEP's hardware parameters, including pixel resolution, grayscale display level, physical size ratio, interface type, and IP address allocation rules. Design fixed mechanical structures corresponding to different projector models. These fixed mechanical structures have position adjustment functions, which can realize the fine adjustment of the displacement of the POEP in the front-back, left-right, and up-down directions and the correction of the angle deflection. Configure the hardware performance parameters and software operating environment of the control computer, and install edge blending debugging software with multi-POEP collaborative control function. This edge blending debugging software supports the definition of blending parameters, image generation, data transmission and effect preview, thereby building a system hardware architecture consisting of a programmable optical edge blending board (POEP), a fixed mechanical structure and a control computer. Based on the corresponding interface within the system hardware architecture, obtain the spatial layout data and installation parameters corresponding to the target projection scene, including the size, installation position, curvature parameters of the projection screen, as well as the installation coordinates, projection angle, and lens parameters of each projector; Based on the spatial layout data of the projection screen and the installation parameters of each projector, the range and shape of the image overlap area of ​​adjacent projectors are calculated, and the number of projectors, the projection coverage area, and the image overlap ratio of adjacent projectors in the multi-channel projection configuration are determined. This completes the construction of the system hardware architecture and the collection of basic parameters to obtain the multi-channel projection configuration parameters of the target projection scene.

[0007] This invention overcomes the limitations of traditional solutions and the lack of basic data at the underlying architecture level by customizing hardware selection, designing adjustable mechanical structures, deploying dedicated software, and acquiring parameters. By selecting a transmissive OLED or LCD screen as the POEP (Point of Interest), and specifying its core hardware parameters such as pixel resolution and grayscale levels, high-precision hardware support is provided for subsequent optical modulation, avoiding the shortcomings of traditional general-purpose hardware that cannot meet the requirements for fine edge blending. The fixed mechanical structure with multi-directional displacement fine-tuning and angle correction functions solves the blending misalignment problem caused by the relative positional deviation between the POEP and the projector in traditional installation methods, significantly improving the flexibility and accuracy of hardware installation. The multi-POEP collaborative control software on the control computer integrates functions such as parameter definition, image generation, and effect preview, breaking through the bottleneck of fragmented software functions and poor coordination in traditional solutions, laying the foundation for subsequent integrated operation. By comprehensively collecting the spatial layout of the projection screen and the installation parameters of the projector, and calculating the overlapping area range and multi-channel configuration parameters, the integrity and accuracy of the basic data are ensured. This avoids the problem of poor fusion effect caused by missing parameters or estimation deviations in traditional solutions. The system can adapt to projection scenarios with different sizes, curvatures, and installation layouts, and provides a solid architecture and data support for edge fusion in multiple scenarios and complex environments.

[0008] Preferably, the step of establishing a network communication connection with each POEP via a control computer and generating a POEP basic control parameter set based on the POEP's hardware characteristic parameters includes: A temporary connection is established with each POEP via the standard video interface of the control computer, a full grayscale test signal is sent, and the actual transmittance data of each pixel of the POEP at different grayscale values ​​is collected. Based on the collected actual transmittance data, and combined with the analysis of gray-level response characteristics, the gray-level transmittance response curve corresponding to POEP is analyzed. At the same time, the transmittance deviation caused by the difference in gray-level response characteristics is compensated, thereby establishing a mapping relationship model between gray value and actual transmittance. Obtain the physical distribution data of the POEP pixel array, including pixel pitch and effective display area boundary coordinates. Combine this with the optical projection parameters of the projector lens to calculate the spatial correspondence between POEP pixels and projected image pixels. Based on the brightness requirements of the target projection scene and the output brightness parameters of the projector, determine the optimal transmittance range of POEP to avoid insufficient brightness due to excessive darkness or excessive brightness that cannot suppress black light leakage. By integrating the mapping relationship model between grayscale values ​​and actual transmittance, the spatial correspondence between POEP pixels and projected image pixels, and the optimal transmittance range, a set of basic control parameters for POEP, including pixel control parameters, communication protocol parameters, and data storage parameters, is generated.

[0009] This invention addresses the core pain points of traditional solutions, namely the lack of programmable optical control and rigid, fixed fusion parameters, by establishing a mapping relationship and parameter system. Through full grayscale testing and actual transmittance acquisition, combined with grayscale response characteristic analysis to generate response curves and compensate for deviations, a mapping relationship model between grayscale values ​​and actual transmittance is established. This effectively avoids the transmittance runaway problem caused by individual differences in POEP pixels and nonlinear responses, providing a scientific basis for subsequent optical modulation. The spatial correspondence between POEP pixels and projected image pixels is calculated, achieving pixel-level control and avoiding modulation deviations in overlapping areas caused by spatial correspondence ambiguity in traditional fusion, fundamentally improving the precision of edge fusion. By combining scene brightness requirements and projector output parameters to determine the optimal transmittance range, this solves the problem of excessively dark or bright images caused by the blind adjustment of brightness in traditional solutions, and specifically suppresses the phenomenon of "black light superimposed on bright bands" in black scenes, addressing the core pain points that traditional solutions have not solved. The POEP basic control parameter set, generated by integrating multi-dimensional parameters, covers key information such as pixel control, communication protocol, and data storage. It realizes the systematic and standardized management of fused parameters, providing the possibility for flexible adjustment and maintenance in the future. It completely changes the situation of traditional solutions where parameters are not adjustable and maintenance is difficult, and greatly improves the practicality and operability of the system.

[0010] Preferably, the model for establishing the mapping relationship between grayscale values ​​and actual transmittance includes: By controlling the computer to send gradient test images from the lowest gray level to the highest gray level to the POEP, and maintaining each gray level for a preset stabilization time, the pixel response of the POEP is ensured to be completely stable. By using an optical brightness measurement device, multiple uniformly distributed test points were selected within the effective display area of ​​the POEP, and the transmitted light brightness data of each test point at different gray levels were collected. Based on the transmitted light brightness data and the incident light brightness data projected onto the POEP by the projector, the actual transmittance of each test point at different gray levels is calculated, and the original gray-transmittance data set of each test point is established. Based on the gray-level response characteristics, the original gray-level transmittance data of all test points were fitted and analyzed to eliminate random errors, generate the overall average gray-level transmittance response curve of POEP, and identify the nonlinear segments in the average gray-level transmittance response curve. For the nonlinear segment in the average gray-transmittance response curve, a compensation function is established to compensate for the transmittance deviation caused by the difference in gray-level response characteristics. The output gray value is then corrected by the control software to ensure that the actual transmittance of POEP maintains a linear correspondence with the theoretical set value, thereby establishing a mapping relationship model between gray value and actual transmittance.

[0011] This invention establishes a high-precision grayscale-transmittance mapping model through a refined process of gradient testing, multi-point acquisition, and fitting compensation, ensuring the reliability of programmable optical control from a core technical perspective. Sending full-range gradient test images to the POEP and ensuring stable response time guarantees the authenticity and reliability of the test data, avoiding data deviations caused by unstable responses in traditional single-tests. Data is collected from multiple test points uniformly selected within the effective display area, covering the entire display range of the POEP, overcoming the limitation of traditional single-point testing in reflecting overall characteristics, and making the subsequently generated response curve more representative. The actual transmittance is obtained by calculating the ratio of transmitted light intensity to incident light intensity. Combined with fitting analysis to eliminate random errors, an average grayscale-transmittance response curve is generated, and nonlinear segments are identified, capturing the optical characteristics of the POEP. A compensation function is established for the nonlinear segments, and the output grayscale value is corrected to maintain a linear correspondence between the actual transmittance and the theoretical set value, completely solving the modulation deviation problem caused by the nonlinearity of the POEP's own grayscale response, providing core technical support for the execution of subsequent grayscale gradient rules in the fusion region. The high-precision mapping model formed by this process ensures the transmittance control accuracy of POEP at different gray levels. It can not only effectively suppress the "black light superimposed on bright band" in the black field, but also ensure the uniform and natural brightness transition in the overlapping area, significantly improving the overall effect of edge blending. At the same time, it provides a basis for subsequent parameter adjustment, further enhancing the system's adaptability to complex scenes.

[0012] Preferably, the step of establishing a compensation function for the nonlinear segment in the average grayscale-transmittance response curve and correcting the output grayscale value through control software includes: The gray level range corresponding to POEP in the nonlinear segment is divided into multiple continuous sub-intervals. The nonlinear coefficient of the gray-transmittance response curve in each sub-interval is calculated, and the key sub-intervals with nonlinear coefficients higher than the preset threshold are identified. Based on the original data of the key sub-intervals, a local compensation function corresponding to each sub-interval is constructed by using polynomial fitting or piecewise linear interpolation to ensure a smooth transition of transmittance changes after compensation. Establish a parameter index table for the local compensation function, associate and store gray values ​​with the corresponding compensation coefficients, and send test gray values ​​containing compensation corrections to POEP via the control computer. Re-collect transmittance data for each test point to verify whether the compensation effect meets the preset requirements. If the compensated transmittance deviation exceeds the allowable range, adjust the parameters of the local compensation function or the fitting method, and repeat the verification and adjustment until a mapping relationship model between the grayscale value and the actual transmittance that meets the requirements is formed.

[0013] This invention addresses the modulation deviation problem caused by the nonlinear grayscale-transmittance response of POEP through sub-interval division, local compensation function construction, and closed-loop verification optimization, providing high-precision assurance for programmable optical control. By dividing the nonlinear segment into continuous sub-intervals and selecting key sub-intervals, targeted compensation is applied to the core deviation area, avoiding the insufficient local precision problem caused by traditional overall compensation and significantly improving the degree of compensation. The local compensation function is constructed using polynomial fitting or piecewise linear interpolation, ensuring a smooth transition of transmittance after compensation and effectively avoiding possible grayscale abrupt changes during compensation, laying the foundation for a natural brightness transition in overlapping areas. A parameter index table is established to achieve the associated storage of grayscale values ​​and compensation coefficients. Combined with the verification feedback mechanism of the control computer, repeated testing and adjustment of compensation parameters and fitting methods ensure that transmittance deviation is controlled within the allowable range, completely solving the problems of poor black light suppression and bright band residue caused by the lack of a compensation mechanism in traditional solutions. The high-precision mapping model formed by this process ensures that the transmittance control of POEP is highly consistent with the theoretical set value. This not only enhances the reliability of programmable optical control but also provides core support for the execution of grayscale gradient rules in the subsequent fusion area. This allows the system to flexibly adapt to the brightness modulation requirements of different scenarios, further improving the overall effect and scene adaptability of edge fusion.

[0014] Preferably, the step of dividing the POEP into blended and non-blended regions based on the overlap ratio of adjacent projector images and the pixel array distribution of the POEP, and defining the grayscale gradient rules within the blended region, includes: Based on the obtained overlap ratio of adjacent projector images and combined with the pixel resolution of the projected image, the pixel coordinate range of the overlapping area in the projected image is calculated, and the pixel boundary of the fusion area corresponding to each POEP is determined. Based on the requirements of dark field fusion, a black light suppression target for the fusion area is set. Combined with the minimum transmittance parameter of POEP, the minimum gray value at the edge of the fusion area is determined to ensure that the amount of black light superposition in the overlapping area is lower than the threshold perceptible to the human eye. Based on the characteristics of human vision, we designed grayscale gradient curve types within the fusion area, including linear gradient, gamma curve gradient, or custom nonlinear gradient, so that the brightness transition in the overlapping area is smooth without obvious jumps. Based on the grayscale gradient curve, calculate the target grayscale value of each pixel within the fusion region, and set the pixels in the non-fusion region to the highest grayscale level to generate the pixel data matrix of the initial fusion control image. The pixel data matrix of the initial fusion control image is smoothed to eliminate abrupt grayscale changes between pixels and avoid jagged edges or mottled light and shadow on the projected image, thus forming the initial fusion control image data.

[0015] This invention addresses the core pain points of traditional solutions, such as "black light superimposed on bright bands" and abrupt brightness transitions in dark scenes, by dividing the fusion region, customizing grayscale gradient rules, and implementing smoothing processes. By combining the overlap ratio and pixel resolution to calculate the boundary of the fusion region, the distinction between fusion and non-fusion regions is achieved, avoiding modulation range deviations caused by blurred boundaries in traditional methods and providing a clear basis for subsequent optical modulation. Based on the requirements of dark scene fusion, black light suppression targets and minimum grayscale values ​​are set. Combined with the minimum transmittance parameter of POEP, black light superposition in overlapping areas under dark scene conditions is suppressed at the source, addressing the core problem unsolved by traditional solutions and ensuring that the superposition amount is below the threshold perceptible to the human eye. Diverse grayscale gradient curves are designed based on the characteristics of human vision, breaking the limitations of traditional single gradient modes and ensuring smooth brightness transitions in overlapping areas, significantly improving visual comfort. The highest grayscale level is set in the non-fusion region to ensure image integrity, while the pixel data matrix is ​​smoothed to eliminate abrupt grayscale changes and jagged edges, preventing mottled light and shadow on the projected image and further optimizing the display effect. The initial fusion control image data generated in this step ensures a natural transition in the fusion area while maintaining the image integrity of the non-fusion area. This provides high-quality data support for subsequent optical edge modulation, enabling the system to adapt to different scene requirements such as dark and bright scenes, and significantly improving the scene adaptability and display quality of edge fusion.

[0016] Preferably, the optical edge modulation of the image projected by the projector includes: Based on the pixel array spatial topology model and gray-level-transmittance mapping relationship model of POEP, a dynamic light modulation logic library is constructed. This logic library contains pixel transmittance control rules corresponding to different image edge types, including hard edges, soft edges and gradient edges. The image edge feature intelligent extraction algorithm captures the edge contour, gradient change and texture features of the projected image in real time, generates an image edge feature map, and performs global texture detail enhancement processing on the projected image based on the image edge feature map. The high-frequency edge information and low-frequency background information corresponding to the projected image are separated by dynamic threshold segmentation technology to generate multi-scale feature layered data. Skeleton extraction and connectivity analysis are performed on high-frequency edge information in multi-scale feature-layered data to construct an edge feature spatial topology network and clarify the branching relationships and continuity characteristics of the edges. Based on the edge feature spatial topology network, edge segments with correlation are aggregated into complete edge regions. At the same time, a background feature masking mechanism is used to delineate pure background regions that are not related to the edge regions. In addition, combined with the fusion objectives of the projection scene, including black light suppression in dark fields and smooth transition in bright fields, the modulation priority of the complete edge regions is sorted, and the core modulation edge regions and auxiliary modulation edge regions are marked. The topological information of the core modulation edge region, auxiliary modulation edge region and pure background region is mapped to the pixel array coordinate system of POEP to generate a localization result map containing region type and modulation priority. The edge region and non-edge region that need to be modulated are located based on the localization result map. The corresponding rules in the dynamic light modulation logic library are called according to the image edge feature map, and the gray values ​​of each pixel of POEP are dynamically adjusted to change the phase distribution of the projected light in the edge region, thereby realizing the directional modulation of the edge light field. A nonlinear optical enhancement mechanism is introduced to perform phase superposition and amplitude calibration on the modulated light in the edge region, enhance the light intensity transition difference between the edge and non-edge regions, and suppress the light field distortion in the non-edge region to ensure the sharpness of the main subject in the image. By integrating the directional modulated light in the edge region with the original projected light in the non-edge region, the light field interference noise during the modulation process is eliminated, resulting in a projected image with natural edge transitions and complete details after optical edge modulation.

[0017] This invention achieves refined and intelligent optical modulation of projected image edges through dynamic optical modulation logic construction, edge feature extraction, and layered modulation optimization, comprehensively overcoming the limitations of traditional solutions such as poor fusion effects and poor scene adaptability. A dynamic optical modulation logic library containing control rules for different edge types is constructed, providing targeted solutions for diverse edge scenes and avoiding the inadequacy of adaptation caused by traditional single modulation rules. Real-time extraction of image edge contours, gradient changes, and other features is performed, along with global detail enhancement. Dynamic threshold segmentation separates high-frequency edges from low-frequency background information, capturing core image features and laying the foundation for layered modulation. Skeleton extraction and aggregation of high-frequency edge information, combined with a background masking mechanism, divides the core, auxiliary modulation, and pure background regions, achieving the allocation of modulation resources and avoiding resource waste and effect imbalance caused by indiscriminate modulation. Mapping regional topology information to the POEP pixel coordinate system ensures the specificity of modulation positions. Combined with edge-oriented light field modulation and nonlinear optical enhancement mechanisms, this not only strengthens the transition difference between edge and non-edge regions but also suppresses light field distortion and interference noise, completely solving the problems of bright bands, dark bands, and image blurring in overlapping areas found in traditional solutions. The projected image generated in this step has a natural edge transition and complete details. It not only effectively suppresses the "black light superimposed on bright band" in black fields, but also ensures the brightness uniformity in bright field scenes. It greatly improves the fineness of edge blending and visual effect, while enhancing the system's adaptability to complex image scenes and providing reliable technical support for multi-scene and high-requirement projection needs.

[0018] Preferably, the step of adjusting the grayscale gradient rules of the fusion region and the pixel transmittance parameters of the POEP based on the effect optimization feedback information, updating the fusion control image data and re-sending it to the POEP, and repeatedly executing optical edge modulation and evaluation feedback until the projected image reaches the preset edge fusion standard includes: The evaluation indicators of edge blending effect of the projected image are obtained through visual observation, including the brightness uniformity of the overlapping area, the blurring degree of the blending boundary, and the brightness consistency between the non-overlapping area and the overlapping area. For areas where brightness uniformity is not up to standard, locate the corresponding POEP pixel range and analyze the reasons for the transmittance deviation in that area, including POEP installation position deviation, grayscale-transmittance mapping model error or projector projection angle deviation. If the POEP installation position is misaligned, the spatial correspondence between the POEP and the projector lens can be corrected by adjusting the displacement or angle parameters of the fixing mechanical structure, and the pixel control parameters of the area can be re-determined. If the error is due to the gray-level-transmittance mapping model, supplement the transmittance test data of the area, update the gray-level-transmittance mapping model, and adjust the target gray-level value of the corresponding pixel. If the projector's projection angle is off, combine the geometric correction data of the multi-channel projection to adjust the division range of the blending area and the grayscale gradient rules so that the blending effect adapts to the change in projection angle. Update the fusion control image data according to the adjusted parameters, send it to the corresponding POEP and load it to take effect, and perform visual observation and evaluation again until all evaluation indicators meet the preset edge fusion standards.

[0019] This invention, through multi-dimensional evaluation, attribution, and closed-loop adjustment mechanisms, completely solves the problems of uncontrollable blending effects and blind parameter adjustments in traditional solutions, significantly improving the final quality and stability of edge blending. It clearly defines core evaluation indicators such as brightness uniformity and the degree of blurring at blending boundaries, providing quantitative standards for effect judgment and avoiding biases caused by traditional subjective evaluations, ensuring the direction of optimization. For areas with substandard brightness, it locates the POEP pixel range and deeply analyzes three core deviation causes: installation location, mapping model, and projection angle, breaking through the limitations of traditional solutions that offer only general adjustments and cannot address specific problems, thus achieving root cause analysis. Differentiated adjustment strategies are adopted for different deviation causes: correcting installation deviations through mechanical structures, optimizing the mapping model through supplementary data, and adjusting blending rules to adapt to changes in projection angle, forming a comprehensive and targeted solution that effectively solves the shortcomings of traditional solutions where single adjustment methods cannot cover multiple problems. Through a closed-loop iterative process of "adjustment-issuance-evaluation," parameters are continuously optimized until all indicators meet the standards, ensuring that the projected image always meets the preset standards and completely eliminating problems that traditional solutions struggle to solve, such as "black light superimposed on bright bands" in black areas and harsh brightness transitions. This step not only ensures the integrity and stability of edge blending, but also enhances the system's adaptability to complex situations such as installation errors and equipment characteristic fluctuations, enabling the system to stably output high-quality projection images in multiple scenarios and complex environments, significantly improving the practicality and reliability of the solution.

[0020] Preferably, for areas where brightness uniformity is substandard, the method of locating the corresponding POEP pixel range and analyzing the reasons for the transmittance deviation in these areas includes: Using a brightness acquisition device, multiple sampling points are selected in the overlapping and non-overlapping areas of the projected image, and the actual brightness values ​​of each sampling point are collected to establish a brightness distribution data matrix. Calculate the standard deviation of the brightness values ​​of each sampling point in the overlapping area. If the standard deviation exceeds the preset threshold, it is determined that the brightness uniformity of the area is not up to standard, and the area with the largest standard deviation is identified as the key optimization area. Compare the brightness difference between adjacent sampling points in the overlapping and non-overlapping areas. If the brightness difference exceeds the acceptable range of human vision, the transition of the fusion boundary is deemed unqualified, and the corresponding POEP pixel column or row is recorded. The relationship between the target grayscale value and the actual transmittance of the POEP pixels corresponding to the key optimization area is analyzed. If the deviation between the actual transmittance and the theoretically calculated value exceeds the allowable range, it is determined to be an error in the grayscale-transmittance mapping model. The actual installation posture of POEP is captured by an image acquisition device and compared with the preset installation parameters. If the position deviation or angle deviation exceeds the adaptation range, it is determined to be a POEP installation position deviation. Retrieve the geometric correction parameters of the multi-channel projection and analyze the difference between the actual projection angle of the projector and the preset angle. If the difference causes the shape of the overlapping area to be deformed, it is determined to be a projection angle deviation of the projector.

[0021] This invention provides solid data support for previous closed-loop optimization through scientific sampling, quantitative analysis, and attribution, solving the core pain points of traditional solutions, such as vague identification of deviation causes and lack of targeted optimization. By using a brightness acquisition device to sample multiple points in overlapping and non-overlapping areas, a brightness distribution data matrix is ​​constructed, achieving comprehensive and objective collection of brightness data. This avoids the data bias caused by traditional single-point sampling or subjective judgment, laying a reliable foundation for subsequent analysis. The standard deviation of brightness is calculated to determine whether uniformity meets the standard and to locate key optimization areas, transforming the abstract concept of "poor performance" into specific pixel ranges and data indicators, making the optimization goals clearer and more explicit. By comparing the brightness differences between adjacent areas, the quality of the fusion boundary transition is judged, and the corresponding POEP pixel positions are recorded, capturing the root cause of abrupt boundary transitions and providing a clear basis for subsequent targeted adjustments. By comparing actual and theoretical light transmittance, POEP installation posture, projector projection angle, and other data, the causes of deviations were classified and determined. This breaks through the limitations of traditional solutions that cannot distinguish deviation types and make blind adjustments. It provides a scientific basis for the formulation of subsequent differentiated adjustment strategies, avoids the waste of time and resources caused by ineffective adjustments, greatly improves the efficiency and degree of closed-loop optimization, and ensures a rapid improvement in edge blending effect.

[0022] Preferably, if the deviation is due to the projection angle of the projector, adjusting the division range of the fusion region and the grayscale gradient rules based on the geometric correction data of the multi-channel projection to make the fusion effect adapt to the change in projection angle includes: The deformation of the overlapping area shape is calculated based on the projection angle deviation of the projector, including specific parameters of tensile deformation, compressive deformation or tilting deformation. Adjust the pixel boundary coordinates of the fusion area according to the deformation amount to make the fusion area completely match the actual overlapping area after deformation, and avoid bright or dark bands caused by the offset of the fusion range. To optimize the grayscale gradient curve parameters for the shape of the fused region after deformation, the gradient transition length is extended in the stretched deformation region and shortened in the compressed deformation region to ensure a uniform brightness transition rate. By combining the distance changes between each pixel in the fused area after deformation and the projector lens, the corresponding target grayscale value is corrected to compensate for the light intensity attenuation difference caused by the distance change. Interpolation is performed on the pixel data of the deformed region to ensure the continuity of grayscale gradient, and the division range and grayscale gradient rules of the fusion region are adjusted to avoid pixel-level brightness discontinuities, thereby generating adjusted fusion control image data.

[0023] This invention addresses the distortion of the blended area caused by projector projection angle deviation. Through deformation compensation, gradient rule optimization, and pixel data calibration, it effectively solves the limitations of traditional solutions that cannot adapt to changes in projection angle and result in unbalanced blending effects. Based on the projection angle deviation, it calculates the stretching and compression deformation parameters of the overlapping area, enabling quantitative analysis of the deformation state. This avoids the insufficient compensation problem caused by the vague understanding of deformation in traditional solutions and provides data support for adjustments. Adjusting the pixel boundary coordinates of the blended area according to the deformation amount ensures a perfect match between the blended area and the actual overlapping area, fundamentally solving the problem of bright and dark bands caused by the offset of the blending range and adapting to the impact of changes in projection angle. Optimizing the grayscale gradient curve parameters for different deformation types—extending the transition length in the stretched area and shortening the transition length in the compressed area—ensures the uniformity of the brightness transition rate and avoids the abrupt transitions or insufficient blending problems caused by traditional fixed gradient rules. Combining the distance change between pixels and the lens to correct the target grayscale value compensates for differences in light intensity attenuation. Simultaneously, interpolation processing is performed on the deformed area to eliminate brightness banding, further optimizing the display effect. This step, through a series of targeted adjustments, enables the blending effect to actively adapt to the projection angle deviation of the projector, completely breaking the shortcomings of traditional solutions that have strict requirements on projection angle and poor scene adaptability. Even if there are slight deviations in the installation of the equipment or changes in the projection angle, it can still ensure that the brightness transition of the overlapping area is natural and there are no obvious bright or dark bands. This greatly improves the system's environmental adaptability and fault tolerance, ensuring high-quality edge blending effects under multiple scenarios and complex installation conditions.

[0024] It has the following beneficial effects: (1) By building a customized system hardware architecture and collecting multi-channel projection configuration parameters, the problem of poor fusion effect caused by poor hardware adaptability and lack of basic data in traditional solutions was solved from the bottom layer. The system hardware architecture uses a programmable optical edge blending board (POEP) as the core modulation component, combined with a fixed mechanical structure with position adjustment function and a dedicated control computer, breaking the limitation of the traditional solution lacking programmable optical control design, and providing hardware support for subsequent flexible adjustment of fusion parameters and realization of optical modulation. The fixed mechanical structure supports multi-directional displacement fine adjustment and angle deflection correction, effectively avoiding the fusion misalignment caused by hardware position deviation in the traditional installation method, and improving the flexibility and degree of system installation. The dedicated software on the control computer integrates parameter definition, data transmission and other functions, laying the foundation for integrated operation. By comprehensively collecting configuration parameters such as the number of projectors, projection coverage area, and image overlap ratio, combined with the spatial layout of the projection scene and equipment installation data, the integrity and accuracy of the basic data are ensured, avoiding the imbalance of fusion effect caused by parameter estimation deviation in the traditional solution. The hardware architecture and core parameters acquired in this step provide a solid foundation for establishing spatial correspondences and defining fusion rules, enabling the system to adapt to projection scenes of different scales and layouts, and fundamentally improving the scene adaptability and overall reliability of edge fusion.

[0025] (2) By constructing a spatial correspondence and a standardized set of basic control parameters, the core pain points of traditional solutions, such as the lack of a system parameter framework and insufficient fusion control accuracy, are solved. Based on the multi-channel projection configuration parameters, the matching installation parameters of POEP and projector are determined, clarifying the installation orientation, relative distance, and attitude angle, establishing a spatial correspondence between the two, effectively avoiding pixel modulation deviation caused by spatial ambiguity, and providing spatial positioning assurance for subsequent pixel-level control. By connecting each POEP through network communication, its transmittance range, gray-level response characteristics, and other hardware parameters are collected. Combined with full gray-level testing and data fitting analysis, a set of basic control parameters including pixel control and communication protocols is generated, breaking the limitations of traditional solutions with scattered parameters and no unified standard. This parameter set fully considers the differences in the hardware characteristics of POEPs, establishes a mapping relationship between gray values ​​and transmittance, provides a scientific basis for subsequent fusion area modulation, and supports flexible adjustment and maintenance of parameters, solving the problems of fixed and rigid parameters and difficult maintenance in traditional solutions. Through this series of operations, step S2 provides a standardized and regulated control basis for subsequent optical edge modulation, which not only improves the precision of fusion control, but also lays the parameter support for suppressing the "black light superimposed on bright band" in the black field, and greatly enhances the system's adaptability to complex environments.

[0026] (3) By dividing the fusion region and defining scientific grayscale gradient rules, the core problems of "black light superimposed on bright bands" and abrupt brightness transitions in traditional solutions are specifically addressed. Combining the overlap ratio of adjacent projector images and the distribution of POEP pixel arrays, the pixel boundaries between the fusion region and the non-fusion region are divided, avoiding the modulation range deviation caused by the blurred boundary division in traditional solutions. This ensures that the fusion modulation only acts on the target region, taking into account both the fusion effect and the image integrity of the non-fusion region. Based on the requirements of dark field fusion, a black light suppression target is set, and the minimum grayscale value at the edge of the fusion region is determined by combining the minimum transmittance parameter of POEP. This suppresses the black light superposition in the overlapping region under dark field conditions from the source, hitting the core pain point that traditional solutions have not solved, and ensuring that the superposition amount is lower than the threshold perceptible to the human eye. By referring to the visual characteristics of the human eye, diverse gradient rules such as linear gradient and gamma curve gradient are designed, breaking the limitations of the traditional single gradient mode. This makes the brightness transition in the overlapping region smooth without obvious jumps, significantly improving visual comfort. The generated initial fusion control image data undergoes boundary calibration and smoothing to eliminate abrupt grayscale changes between pixels, preventing jagged edges or mottled shadows in the projected image and providing high-quality data support for subsequent optical edge modulation. This step ensures a natural transition in the fusion area while maintaining image integrity in the non-fusion areas, enabling the system to flexibly adapt to different scene requirements such as dark and bright scenes, significantly improving the display quality and scene adaptability of edge fusion.

[0027] (4) By using a closed-loop optical edge modulation and iterative optimization mechanism, the problems of uncontrollable fusion effect and difficulty in achieving preset standards in traditional solutions are completely solved, realizing the optimization and high quality of edge fusion. The initial fusion control image data is sent to POEP, and the optical edge modulation of the projected image is performed by dynamically adjusting the pixel transmittance. This directly affects the light field control of the overlapping area, effectively suppressing the "black light superimposed on bright band" in black fields, while ensuring the brightness uniformity in bright field scenes. This breaks through the limitation of traditional solutions that can only perform basic color and light leakage compensation. The core indicators such as brightness uniformity and transition state are evaluated by visual observation, and targeted effect optimization feedback information is generated, so that parameter adjustment has a clear basis and avoids the drawbacks of blind adjustment in traditional solutions. Based on the feedback information, the grayscale gradient rules and POEP pixel transmittance parameters are flexibly adjusted, the control data is updated and resent for execution, forming a closed-loop iterative process of "modulation-evaluation-optimization", continuously correcting deviations until the projected image reaches the preset standard. This closed-loop mechanism ensures continuous optimization of the fusion effect, eliminating bright and dark bands in overlapping areas while maintaining image integrity and clarity in non-fusion areas. Ultimately, it achieves a projection effect with natural edge transitions and harmonious overall display. Furthermore, the process supports flexible parameter adjustment and maintenance, resolving the inability of traditional solutions to adapt to changing scenarios. This enables the system to stably output high-quality projection images in multiple scenarios and complex environments, significantly improving the practicality and reliability of the solution. Attached Figure Description

[0028] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a flowchart illustrating the steps of the programmable projector image edge blending system in this embodiment. Figure 2 This is a schematic diagram of the structural connections of the programmable projector image edge blending system of the present invention; Figure 3 This is a schematic diagram illustrating the control principle of the programmable projector image edge blending system of the present invention. Detailed Implementation

[0029] The present invention will be further described below with reference to the accompanying drawings and embodiments, but this should not be construed as limiting the present invention.

[0030] To achieve the above objectives, please refer to Figure 1-2 This invention provides a programmable projector image edge blending system, comprising: S01: Construct a system hardware architecture consisting of a programmable optical edge blending board (POEP), a fixed mechanical structure, and a control computer. Based on the system hardware architecture, obtain multi-channel projection configuration parameters for the target projection scene, including the number of projectors, the projection coverage area, and the image overlap ratio of adjacent projectors. In this embodiment of the invention, three customized transmissive OLED screens were selected as POEPs (resolution 1920×1200, grayscale 256 levels, transmittance 0.08%-99.3%), along with three sets of aluminum alloy fixing mechanical structures (the inner diameter of the ring frame is adapted to the projector lens diameter of 10cm, supporting ±3cm displacement fine adjustment and ±2° angle correction), and one control computer (quad-core 3.2GHz processor, 8GB memory, dual gigabit network interfaces) to build the system hardware architecture. Transmittance testing was conducted by connecting the POEPs via the control computer's HDMI interface, and network communication was established via the RJ45 interface. The mechanical structures fixed the POEPs 15cm in front of the projector lens (meeting the 10-20cm installation requirement). Obtain the multi-channel projection configuration parameters for the target projection scene (simulated training curved screen): The projection screen size was measured to be 6m wide × 2.5m high using a laser rangefinder, and the number of projectors was determined to be 3 (DLP type, each projector with a width of 2m, covering a 6m screen); the projection coverage area was the entire curved screen area (physical coordinates (0,0)-(6m,2.5m)); the projected images of adjacent projectors were captured using image analysis tools, and the overlap area width was calculated to be 0.4m, with an overlap ratio of 0.4m / 2m × 100% = 20%. The final configuration parameters were determined as follows: 3 projectors, coverage area of ​​6m × 2.5m, and overlap ratio of 20%.

[0031] S02: Based on the multi-channel projection configuration parameters of the target projection scene, determine the adaptation and installation parameters between each projector lens and POEP, and construct the spatial correspondence between POEP and projector; establish network communication connection with each POEP through the control computer, and generate the POEP basic control parameter set based on the hardware characteristic parameters of POEP; In this embodiment of the invention, based on the multi-channel projection configuration parameters (3 projectors arranged horizontally with a spacing of 2m), the matching installation parameters between each projector lens and POEP are determined: POEP1 corresponds to the left projector, with the installation orientation directly in front of the lens, a relative distance of 15cm from the lens, and the posture angle perpendicular to the lens optical axis (horizontal angle 0°, vertical angle 0°); POEP2 corresponds to the middle projector, with the same installation orientation and distance as POEP1, and the posture angles are 0° horizontally and 0° vertically; POEP3 corresponds to the right projector, with the same parameters as before. A spatial coordinate calibration method was adopted to construct the spatial correspondence between the POEP and the projector: with the optical center of the left projector lens as the origin (0,0,0), the physical coordinates (0.15xmm+150mm, 0.15ymm, 150mm) of POEP pixel 1 (x,y) are respectively (0.15mm is the pixel pitch, 150mm is the installation distance). The corresponding projection coordinates after projection onto the screen are ((0.15x+150)×(6000mm / 180mm), 0.15y×(2500mm / 112.5mm)) (180mm and 112.5mm are the physical dimensions of the POEP), forming a coordinate mapping table. TCP / IP communication (baud rate 115200bps) was established with each POEP through the network interface of the control computer, based on the POEP hardware characteristics (transmittance range 0.08%-99.3%, grayscale response curve T=0.0001G). 3 +0.02G 2 +0.1G+0.05, pixel array 1920×1200), generate POEP basic control parameter set: pixel control parameters (grayscale-transmittance mapping table), communication parameters (IP address 192.168.1.101-103), storage parameters (BMP format, EPROM storage path), ensuring that the parameters cover all hardware control requirements.

[0032] S03: Based on the overlap ratio of adjacent projector images and the pixel array distribution of POEP, divide the fusion region and non-fusion region corresponding to each POEP, define the grayscale gradient rules within the fusion region, and generate initial fusion control image data. In this embodiment of the invention, based on the overlap ratio of 20% between adjacent projector images (corresponding to a projection width of 0.4m) and combined with the POEP pixel array distribution (1920 pixels correspond to a projection width of 2m, pixel width = 2000mm / 1920 ≈ 1.04mm), the pixel width of the fusion region is calculated to be approximately 0.4m / 1.04mm ≈ 384 POEP pixels. Divide the blended and non-blended areas of each POEP: POEP1 (left projector) has a blended area of ​​384 pixels on the right (x-coordinate 1536-1920, y-coordinate 0-1200) and a non-blended area of ​​1536 pixels on the left (0-1535, 0-1200); POEP2 (middle projector) has a blended area of ​​384 pixels on the left (0-383, 0-1200) and 384 pixels on the right (1536-1920, 0-1200) and a non-blended area of ​​1152 pixels in the middle (384-1535, 0-1200); POEP3 (right projector) has a blended area of ​​384 pixels on the left (0-383, 0-1200) and a non-blended area of ​​1536 pixels on the right (384-1920, 0-1200). Define the grayscale gradient rules within the blending area: Use a gamma curve gradient (γ=0.8). In the POEP1 blending area, the grayscale value linearly transitions from 255 (99.3% transmittance) to 198 (75% transmittance) from left to right. The gradient formula is G=255-57×(x / 384)^0.8 (where x is the number of pixels from the left boundary of the blending area). In the POEP2 blending area, the grayscale value changes from 255 to 198 from right to left on the left and from 255 to 198 on the right. In the POEP3 blending area, the grayscale value changes from 255 to 198 from right to left. The grayscale value of pixels in non-fusion areas is uniformly set to 255. An initial fusion control image data matrix is ​​generated at a resolution of 1920×1200 pixels. The matrix elements are grayscale values ​​of 0-255. For example, the grayscale value of pixel POEP1 (1600, 600) is 255-57×(64 / 384)^0.8≈255-57×0.33≈255-18.8≈236, ensuring that the data can be directly used for POEP loading and display.

[0033] S04: The initial blending control image data is transmitted to the corresponding POEP via network communication. The POEP's hardware module receives and stores the data, and simultaneously controls each pixel of the POEP to adjust its transmittance according to the grayscale value of the initial blending control image data. Optical edge modulation is then applied to the image projected by the projector to generate an optically edge-modulated projected image. The edge blending effect of the projected image is evaluated through visual observation, and effect optimization feedback information is generated based on the edge blending effect. Based on the effect optimization feedback information, the grayscale gradient rules of the blending area and the pixel transmittance parameters of the POEP are adjusted by the control computer. The blending control image data is updated and resent to the POEP. The optical edge modulation and evaluation feedback are repeated until the projected image reaches the preset edge blending standard, thereby completing the programmable projector image edge blending.

[0034] In this embodiment of the invention, the initial fusion control image data is transmitted to three POEPs via the RJ45 network interface of the control computer, based on the TCP / IP communication protocol (baud rate 115200bps, data transmission block size 1024 bytes). POEP1 receives the fusion data (1920×1200 pixel matrix, BMP format, file size 2.25MB) from the corresponding left projector. Its built-in hardware module (including EPROM memory and network receiver chip) receives the data, verifies it, and stores it in the EPROM at a specified address (address range 0x000000-0x232FFF). After storage, it sends a "storage successful" signal back to the control computer. After receiving feedback, the control computer sends a "load display" command. Each pixel in the POEP adjusts its transmittance based on its stored grayscale values: pixels in the POEP1 blending area (1536-1920, 0-1200) are adjusted to approximately 88% transmittance based on a grayscale value of 236 (e.g., pixels 1600, 600); pixels in the non-blending area are adjusted to 99.3% transmittance based on a grayscale value of 255. When the projector projects an image (a simulated training scene, including an airplane model and instrument panel) through the POEP, the POEP uses pixel transmittance changes to optically modulate the projected light at the edges. The brightness of the projected light in the blending area gradually changes according to a gamma curve, generating an optically edge-modulated projected image. There is no obvious brightness discontinuity in the overlapping areas of adjacent projectors in the image. Next, visual observation combined with optical measurement equipment is used to evaluate the edge blending effect of the projected image. Evaluation criteria are set as follows: brightness uniformity in the overlapping area ≥95%, no bright or dark bands wider than 2mm, and image integrity in the non-overlapping area ≥99%. The brightness uniformity of the overlapping area (the blending area between the left and middle projectors, 0.4m wide) was measured: 20 sampling points were evenly selected within the area, and the measured brightness values ​​ranged from 188 to 198 cd / m². 2The calculated uniformity is (188 / 198) × 100% ≈ 94.9% (close to the standard, slightly below the standard); the transition state of the fusion boundary was observed: a dark band with a width of about 3mm was found on the left side of the overlapping area (brightness 185cd / m²). 2 It is lower than the surrounding 188 cd / m 2 There was no obvious bright band on the right side; the image integrity of the non-overlapping area was checked: the edge of the airplane model wing in the non-fusion area of ​​the left projector was intact, the instrument panel scale was clear, and the integrity reached 99.5% (meeting the standard). Based on the evaluation results, effect optimization feedback information was generated: the brightness uniformity of the overlapping area was poor by 0.1%, and there was a 3mm dark band on the left. The gray value of the pixels on the left side of the POEP1 fusion area needs to be adjusted to increase the light transmittance of the corresponding area to eliminate the dark band, while improving the overall brightness uniformity of the overlapping area. Then, based on the effect optimization feedback information, the gray gradient rule of the fusion area was adjusted by controlling the computer: the gamma value of the POEP1 fusion area was adjusted from 0.8 to 0.7, and the gradient formula was updated to G=255-57×(x / 384)^0.7, which improved the gray value on the left side of the fusion area (e.g., the original gray value of pixel 1536,600 was 255, and it remained 255 after adjustment; the original gray value of pixel 1550,600 was 252, and it became 253 after adjustment), and the corresponding light transmittance increased from ≈97% to ≈98%. Simultaneously, the target grayscale values ​​of the 10 pixel columns (1536-1545, 0-1200) on the left side of the POEP1 fusion area are adjusted, with the grayscale value of each pixel column increasing by 1. For example, pixel 1536,600 remains at 255, while pixel 1537,600 increases from 254 to 255. After updating the fusion control image data matrix, it is redeployed to POEP1 via the network and loaded to take effect. Optical edge modulation and evaluation feedback are performed again: the brightness uniformity of the overlapping area is measured to be approximately 95.9% (meets the standard), the width of the dark band on the left side is reduced to 1mm (below the 2mm standard), and the image integrity of the non-overlapping area remains at 99.5%. The adjustment is repeated twice: the grayscale difference in the gradient formula of the POEP1 fusion area is increased from 57 to 58 in the second adjustment, further improving the grayscale value of the edge pixels. Finally, the brightness uniformity of the overlapping area reaches 96.2%, the dark band is completely eliminated, and all evaluation indicators meet the preset standards. Stop adjusting and complete the edge blending of the programmable projector image. The final projected image has a natural edge transition, with no black bands in dark areas and no obvious brightness jumps in bright areas.

[0035] Furthermore, the acquisition of multi-channel projection configuration parameters for the target projection scene includes: Select a customized transmissive OLED or LCD screen as the POEP and determine the POEP's hardware parameters, including pixel resolution, grayscale display level, physical size ratio, interface type, and IP address allocation rules. In this embodiment of the invention, a customized transmissive OLED screen is selected as the POEP, and the hardware parameter standards are clearly defined: the pixel resolution is set to 1920×1200 (higher than the minimum requirement of 1280×800) to ensure accurate image detail presentation; the grayscale display level is fixed at 256 levels, with complete transparency at grayscale 255 and light leakage rate of less than 0.1% at grayscale 0, meeting the requirements for precise brightness control; the physical size ratio is 16:10, the width is set to 18cm (not exceeding the upper limit of 20cm), and the height corresponds to 11.25cm, adapting to the light spot range of mainstream projection lenses; the interface type is configured with a standard HDMI interface (for factory transmittance and image quality checks) and an RJ45 network interface (for actual control data transmission); the IP address allocation rule is set according to "network segment + device number", and POEPs within the same system are assigned network segments of 192.168.1.101-192.168.1.200, with each POEP corresponding to a unique IP address, which is permanently stored through a hardware module to ensure that the address is unique and identifiable during remote control.

[0036] Furthermore, a fixed mechanical structure corresponding to different projector models is designed. This fixed mechanical structure has a position adjustment function, which can realize the displacement fine adjustment and angle deflection correction of POEP in the front-back, left-right, and up-down directions. In this embodiment of the invention, corresponding fixing mechanical structures are designed for DLP projectors (lens diameter 10cm) and LCD projectors (lens diameter 8cm). The main body of the structure is made of aluminum alloy and has a ring frame design, with the inner diameter matching the lens diameter of each projector. The frame thickness is 5cm. The mechanical structure has built-in sliding guide rails and angle adjustment axes. The front-to-back adjustment range is 0-10cm (keeping the distance between the POEP and the lens 10-20cm), and the left-to-right and up-to-down adjustment ranges are ±3cm each. Fine-tuning of displacement is achieved through a knob-type adjustment mechanism. Angle deflection correction is achieved through three adjusting screws on the edge of the frame, with horizontal and vertical deflection adjustment ranges of ±2° each, ensuring that the POEP plane is perpendicular to the projection optical axis. A detachable connecting base is designed at the bottom of the structure. Professional simulation projectors are directly fixed to the front panel of the projector through the connecting base, while other projectors are rigidly connected to the projector support structure through the connecting base, ensuring that the POEP does not loosen or shift after installation.

[0037] Furthermore, the hardware performance parameters and software operating environment of the control computer are configured, and edge blending debugging software with multi-POEP collaborative control function is installed. This edge blending debugging software supports the definition of blending parameters, image generation, data transmission and effect preview, thereby building a system hardware architecture consisting of a programmable optical edge blending board (POEP), a fixed mechanical structure and a control computer. In this embodiment of the invention, the control computer hardware performance parameters are configured such that the processor is a quad-core processor with a main frequency of 3.2GHz, memory capacity of 8GB, hard disk storage capacity of 500GB, and equipped with dual gigabit network interfaces (connected to the POEP network and scene control network respectively); the software operating environment is set to a 64-bit operating system, and irrelevant background processes are closed to ensure exclusive use of software operating resources. Edge blending debugging software with multi-POEP collaborative control function is installed. This software supports simultaneous connection of 32 POEP devices. The blending parameter definition module can set the shape of the blending area (rectangular, trapezoidal, arc) and gamma value (0.1-2.0 range), the image generation module can create 8-bit depth BMP format blended images, the data transmission module sends image data to the POEP's built-in memory via TCP / IP protocol, and the effect preview module can display the simulated image after multi-POEP blending in real time. A wired connection is established between the POEP's RJ45 interface and the control computer's network interface. The POEP is fixed in front of the projector lens using a fixed mechanical structure, and a system hardware architecture consisting of 3 POEPs, 3 sets of fixed mechanical structures, and 1 control computer is built. After connecting all components, a power-on test is performed to ensure normal communication.

[0038] Furthermore, based on the corresponding interface within the system hardware architecture, spatial layout data and installation parameters corresponding to the target projection scene are obtained, including the size, installation position, curvature parameters of the projection screen, as well as the installation coordinates, projection angle, and lens parameters of each projector. In this embodiment of the invention, spatial layout data and installation parameters of the target projection scene are obtained through the HDMI and network interfaces based on the system hardware architecture: the projection screen is a curved screen with dimensions of 6m wide × 2.5m high, the horizontal distance from the installation position to the projector is 8m, and the radius of curvature is 5m; three projectors are installed horizontally, with installation coordinates of (0,0,0), (2m,0,0), and (4m,0,0) (with the left projector as the origin), and the projection angle is horizontal with a vertical tilt angle of 0°; the lens parameters are uniformly set to a focal length of 120mm, a throw ratio of 1.5:1, and an image aspect ratio of 16:10. All data are obtained through on-site measurement tools. The projection screen size is measured using a laser rangefinder, the installation coordinates are located using a total station, and the lens parameters are extracted from the projector technical manual to ensure that the data accurately reflects the actual installation status on site.

[0039] Furthermore, based on the spatial layout data of the projection screen and the installation parameters of each projector, the range and shape of the image overlap area of ​​adjacent projectors are calculated, and the number of projectors, the projection coverage area, and the image overlap ratio of adjacent projectors in the multi-channel projection configuration are determined, thereby completing the construction of the system hardware architecture and the acquisition of basic parameters to obtain the multi-channel projection configuration parameters of the target projection scene.

[0040] In this embodiment of the invention, the image overlap area of ​​adjacent projectors is calculated using a geometric calculation method based on the spatial layout data of the projection screen and the projector installation parameters: the horizontal overlap range between the projected images of the left and middle projectors is 0.8m, the overlap shape is rectangular, and the overlap ratio = 0.8m / 2m (projection width of a single projector) = 40%. The overlap ratio is corrected to 20% by adjusting the projection angle of the projectors; the image overlap area parameters between the middle and right projectors are the same as those on the left. The multi-channel projection configuration parameters are determined: the number of projectors is 3, the projection coverage area is 6m wide × 2.5m high (completely covering the curved screen), and the image overlap ratio of adjacent projectors is uniformly set to 20%. The calculated overlap area range, number of projectors, coverage area size, and other parameters are compiled into a configuration file and stored in the control computer, completing the system hardware architecture construction and basic parameter collection, providing data support for subsequent fusion parameter settings.

[0041] Furthermore, the step of establishing network communication connections with each POEP via a control computer and generating a basic control parameter set for the POEP based on its hardware characteristic parameters includes: A temporary connection is established with each POEP via the standard video interface of the control computer, a full grayscale test signal is sent, and the actual transmittance data of each pixel of the POEP at different grayscale values ​​is collected. In this embodiment of the invention, temporary connections are established between the control computer's HDMI standard video interface and three POEPs respectively. During connection, interface protocol matching is ensured, and the transmission bandwidth is stabilized at 10Gbps or higher. The control computer sends a full grayscale test signal, with grayscale values ​​increasing sequentially from 0 to 255, each grayscale level held for 5 seconds. Simultaneously, a transmittance measuring instrument is used to collect the actual transmittance data of each pixel of the POEPs. The sampling frequency of the measuring instrument is set to 10Hz, and 10 sets of data are collected for each pixel at each grayscale level. The average value is taken as the transmittance value of that pixel at the corresponding grayscale level. For example, the average transmittance data of the (100, 200)th pixel of POEP1 is 0.08% when the gray value is 0, 49.7% when the gray value is 128, and 99.3% when the gray value is 255. A total of 1920×1200×256×3 sets of data were collected by the three POEPs. All data are classified and stored in the format of "POEP number-pixel coordinate-gray value-transmittance" to ensure that the transmittance data of each pixel is complete and traceable.

[0042] Furthermore, based on the collected actual transmittance data and combined with the analysis of gray-level response characteristics, the gray-level transmittance response curve corresponding to POEP is analyzed, and the transmittance deviation caused by the difference in gray-level response characteristics is compensated, thereby establishing a mapping relationship model between gray value and actual transmittance. In this embodiment, the gray-level-transmittance response curve of the POEP is analyzed using a polynomial fitting method based on the collected actual transmittance data, with the fitting order set to 3. Taking POEP1 as an example, the transmittance data corresponding to gray values ​​of 0-255 is fitted to obtain the response curve equation: T=0.0001G 3 +0.02G 2 +0.1G+0.05 (T is transmittance, G is grayscale value). By comparing the response curves of different POEPs, it was found that the transmittance of POEP2 in the grayscale value range of 50-100 was on average 3.2% lower than that of POEP1, indicating a difference in grayscale response characteristics. A linear compensation algorithm was used to compensate for the deviation, with the compensation formula being G'=G+ΔG (ΔG is the compensated grayscale value). For POEP2 with a grayscale value of 50, ΔG=3, and the transmittance corresponding to the corrected grayscale value of 53 is consistent with that of POEP1 with a grayscale value of 50. Finally, a mapping relationship model between grayscale value and actual transmittance was established. The model input is the target transmittance, and the output is the corresponding calibrated grayscale value. For example, when the target transmittance is 50%, the model outputs a grayscale value of 129, ensuring that the three POEPs output consistent grayscale control signals under the same target transmittance.

[0043] Furthermore, obtain the physical distribution data of the POEP pixel array, including pixel pitch and effective display area boundary coordinates, and combine them with the optical projection parameters of the projector lens to calculate the spatial correspondence between POEP pixels and projected image pixels. In this embodiment of the invention, the physical distribution data of the pixel array is obtained from the POEP technical manual: the pixel pitch is 0.15mm, and the effective display area boundary coordinates are taken with the lower left corner of the POEP as the origin, with the horizontal coordinate 0-180mm (corresponding to 1920 pixels) and the vertical coordinate 0-112.5mm (corresponding to 1200 pixels). Optical projection parameters are extracted from the projector technical manual: focal length 120mm, projection ratio 1.5:1, and lens distortion coefficient ≤0.1%. A geometric projection transformation algorithm is used to calculate the spatial correspondence between the POEP pixels and the projected image pixels, establishing a spatial coordinate system with the intersection of the projector lens optical axis and the center of the POEP as the mapping origin. For example, the center pixel (960, 600) of POEP1 corresponds to the center pixel (1920, 1080) of the projected image; the right edge pixel (1800, 600) of POEP1 is determined to correspond to the pixel (3456, 1080) of the projected image by calculating the relationship between the projection angle and distance. At the same time, considering the influence of lens distortion, the correspondence of edge pixels is corrected, and the correction amount does not exceed 1 projection pixel. Finally, a two-way mapping table of "POEP pixel coordinates - projected image pixel coordinates" is generated to ensure that each POEP pixel can accurately correspond to a specific area of ​​the projected image.

[0044] Furthermore, based on the brightness requirements of the target projection scene and combined with the output brightness parameters of the projector, the optimal transmittance range of POEP is determined to avoid insufficient brightness due to excessive darkness or inability to suppress black light leakage due to excessive brightness. In this embodiment of the invention, by obtaining the brightness requirements of the target projection scene, the target brightness range of the projected image is determined to be 150-200 cd / m² based on the scene application scenario (simulation training). 2 The black level brightness requirement is ≤0.5cd / m². 2 Output brightness parameters extracted from the projector's technical manual: 3000 lm in standard mode; the brightness of the image projected onto a 6m x 2.5m screen is 198 cd / m². 2 Based on this parameter, the optimal transmittance range for POEP is calculated: when the projected image requires 150 cd / m². 2 At the same time, the POEP transmittance is approximately 75.8% (150 / 198). When it is necessary to suppress black light leakage, the POEP transmittance in a black state needs to be ≤0.1% (corresponding to 0.08% transmittance at a grayscale value of 0). The optimal transmittance range for POEP is determined to be 0.08%-99.3%, with the transmittance controlled between 30%-99.3% during normal display to avoid excessive darkness that would cause the screen brightness to drop below 150 cd / m². 2 When displaying black levels, the transmittance is controlled at 0.08%-0.1% to effectively suppress black light leakage and prevent the generation of bright bands.

[0045] Furthermore, by integrating the mapping relationship model between grayscale values ​​and actual transmittance, the spatial correspondence between POEP pixels and projected image pixels, and the optimal transmittance range, a set of basic control parameters for POEP, including pixel control parameters, communication protocol parameters, and data storage parameters, is generated.

[0046] In this embodiment of the invention, a set of basic control parameters for POEP is generated by integrating the previously obtained mapping relationship model, spatial correspondence, and optimal transmittance range. Pixel control parameters include the calibrated grayscale value-transmittance mapping table for each POEP pixel and the corresponding coordinates of the projected image pixel. For example, the control parameters for POEP1 pixel (100, 200) are "target transmittance 0.08%-99.3%, corresponding grayscale value 0-255, corresponding projection coordinates (200, 400)". The communication protocol parameters are set to TCP / IP protocol, data transmission baud rate 115200bps, and parity bit even parity to ensure error-free transmission of control signals. The data storage parameters specify that the fused image data storage path is the POEP built-in EPROM, the storage format is BMP, and the size of a single image file does not exceed 2.25MB (1920×1200×8bit). The basic control parameter sets of the three POEPs are named "POEP1-Control Parameter", "POEP2-Control Parameter" and "POEP3-Control Parameter" respectively, and stored in the control computer. At the same time, they are synchronized to the built-in memory of each POEP, providing a basic control basis for subsequent parameter fusion configuration.

[0047] Furthermore, the model for establishing the mapping relationship between grayscale values ​​and actual transmittance includes: By controlling the computer to send gradient test images from the lowest gray level to the highest gray level to the POEP, and maintaining each gray level for a preset stabilization time, the pixel response of the POEP is ensured to be completely stable. In this embodiment of the invention, a stable connection is established between the control computer's RJ45 network interface and three POEPs, with the connection bandwidth maintained above 100Mbps to ensure packet-free image data transmission. The control computer sequentially sends gradient test images ranging from grayscale value 0 (lowest) to 255 (highest) to each POEP. The image resolution is consistent with the POEP pixel resolution (1920×1200), and each image is filled with a solid color (containing only a single grayscale value). The preset stabilization time for each grayscale level is 8 seconds, which was determined through prior testing—after a grayscale value switch, the POEP pixel transmittance reaches a stable value within 6 seconds, with an extension of 2 seconds to ensure a completely stable response. The sending order is arranged in ascending order of grayscale value, starting from grayscale 0. After sending each image, an 8-second wait is made before sending the next one. The three POEPs synchronously receive and display the test images. Throughout the process, the POEP display status is monitored through the effect preview module of the control computer to ensure no image misalignment or display abnormalities.

[0048] Furthermore, by using an optical brightness measurement device, multiple uniformly distributed test points were selected within the effective display area of ​​the POEP, and transmittance data of each test point at different gray levels were collected. In this embodiment of the invention, an optical brightness measurement device (measurement range 0.01-1000 cd / m²) is used. 2 To ensure accuracy within ±1%, nine evenly distributed test points were selected within the effective display area of ​​the POEP (18cm × 11.25cm). The coordinates of these test points, with the lower left corner of the POEP as the origin, were (3cm, 2.25cm), (9cm, 2.25cm), (15cm, 2.25cm), (3cm, 5.625cm), (9cm, 5.625cm), (15cm, 5.625cm), (3cm, 8.999cm), (9cm, 8.999cm), and (15cm, 8.999cm), covering both the edge and center areas. During measurement, the device probe was vertically aligned with the POEP test points, 5cm away from the POEP surface, to avoid ambient light interference. While the POEP displayed the test image for each grayscale level, transmittance data was collected from the nine test points. Data was collected three times for each test point, and the average value was taken as the transmittance value for that test point at the corresponding grayscale level. For example, when the POEP1 displays a test image with a grayscale value of 50, the average transmitted light intensity at the center test point (9cm, 5.625cm) is 28.5 cd / m². 2 The edge test point (3cm, 2.25cm) has a value of 27.9 cd / m². 2 All data are recorded in the format of "POEP number-grayscale value-test point coordinates-transmitted light intensity".

[0049] Furthermore, based on the transmitted light brightness data and the incident light brightness data projected onto the POEP by the projector, the actual transmittance of each test point at different gray levels is calculated, and a raw data set of gray level-transmittance for each test point is established. In this embodiment of the invention, by measuring the incident light brightness data projected onto the POEP by the projector in advance, the POEP is removed while keeping the projector's position and parameters unchanged. At nine test points on the original POEP mounting plane, the incident light brightness is collected using an optical brightness measurement device, yielding an average value of 142 cd / m². 2(This value is a fixed parameter and will be reused in subsequent calculations). Based on the transmitted light intensity data and the incident light intensity data, the transmittance of each test point is calculated using the formula "Actual transmittance = (Transmitted light intensity / Incident light intensity) × 100%". For example, when the gray value of POEP1 is 50, the transmittance of the center test point = (28.5 / 142) × 100% ≈ 20.1%, and the transmittance of the edge test point = (27.9 / 142) × 100% ≈ 19.6%. A raw gray-transmittance data set is established for each test point in order of gray values ​​from 0 to 255. Each set contains 256 data pairs (gray value G, transmittance T). For example, the raw data set for the center test point of POEP1 contains data such as (0, 0.07%), (50, 20.1%), (100, 41.3%), and (255, 99.2%), ensuring that each data pair comes from actual measurement calculations and has no estimated values.

[0050] Furthermore, based on the gray-level response characteristics, the original gray-level transmittance data sets of all test points are fitted and analyzed to eliminate random errors, generate the overall average gray-level transmittance response curve of POEP, and identify the nonlinear segments in the average gray-level transmittance response curve. In this embodiment of the invention, the gray-transmittance raw data set of 9 test points is fitted and analyzed using the least squares method, and the fitting function is a third-order polynomial (T=aG). 3 +bG 2 +cG+d), random errors are eliminated by calculating residuals (residuals controlled within ±0.5%). For each grayscale value, the average transmittance of 9 test points is taken to generate the overall average grayscale-transmittance response curve of POEP. Taking POEP1 as an example, the curve equation obtained by fitting is: T=0.0001G 3 +0.02G 2 +0.1G+0.05. By comparing the slopes of each segment of the curve, the non-linear segment was identified. It was found that the slope of the gray value 0-20 range was 0.11 (good linearity), the slope of the gray value 21-80 range decreased from 0.11 to 0.09 (non-linear segment), and the slope of the gray value 81-255 range remained stable at 0.095 (good linearity). It was determined that the non-linear segment was the gray value 21-80 range. The maximum deviation between the actual transmittance and the theoretical linear value in this range was 1.8% (when the gray value was 50, the theoretical linear value was 20.0%, and the actual average value was 19.6%).

[0051] Furthermore, a compensation function is established for the nonlinear segment in the average gray-level transmittance response curve to compensate for the transmittance deviation caused by the difference in gray-level response characteristics. The output gray value is then corrected by the control software to ensure that the actual transmittance of POEP maintains a linear correspondence with the theoretical set value, thereby establishing a mapping relationship model between gray value and actual transmittance.

[0052] In this embodiment of the invention, a compensation function is established for the nonlinear segment with grayscale values ​​of 21-80. A piecewise linear interpolation method is used to divide the nonlinear segment into three sub-intervals (21-40, 41-60, and 61-80), and a linear compensation formula is fitted to each sub-interval. For example, the compensation formula for sub-interval 21-40 is G'=G+0.02G-0.4 (where G' is the corrected grayscale value), for sub-interval 41-60 it is G'=G+0.015G-0.2, and for sub-interval 61-80 it is G'=G+0.01G-0.1. By embedding this compensation function in the control software, when the target transmittance needs to be output, the theoretical grayscale value is first calculated based on the average response curve, and then the theoretical grayscale value of the nonlinear segment is corrected. For example, when the target transmittance is 20%, the average response curve calculates a theoretical grayscale value of 50 (within the range of 41-60). The corrected grayscale value G' = 50 + 0.015 × 50 - 0.2 = 50 + 0.75 - 0.2 = 50.55, rounded to 51. The actual transmittance of POEP1 with a grayscale value of 51 after correction is 20.0%, which is completely consistent with the theoretical setting. Finally, a mapping relationship model between grayscale value and actual transmittance is established. The model input is the target transmittance, and the output is the corrected grayscale value, ensuring that the deviation between the actual transmittance and the theoretical value is ≤0.3% across the entire grayscale range (0-255).

[0053] Furthermore, the step of establishing a compensation function for the nonlinear segment in the average grayscale-transmittance response curve and correcting the output grayscale value through control software includes: The gray level range corresponding to POEP in the nonlinear segment is divided into multiple continuous sub-intervals. The nonlinear coefficient of the gray-transmittance response curve in each sub-interval is calculated, and the key sub-intervals with nonlinear coefficients higher than the preset threshold are identified. In this embodiment of the invention, the nonlinear segment (grayscale value 21-80) of POEP1 is divided into 5 continuous sub-intervals, each sub-interval containing 12 grayscale levels, specifically 21-32, 33-44, 45-56, 57-68, and 69-80. The nonlinear coefficient was calculated using the formula "K=|(Tactual - Tlinear) / Tlinear|×100%" (where Tactual is the average transmittance within the sub-interval, and Tlinear is the average theoretical linear transmittance within that interval). The nonlinear coefficient for each sub-interval was calculated as follows: Sub-interval 21-32: Tactual mean 12.3%, Tlinear mean 12.8%, K=|(12.3-12.8) / 12.8|×100%≈3.91%; Sub-interval 33-44: Tactual 18.5%, Tlinear 19.2%, K≈3.65%; Sub-interval 45-56: Tactual 24.7%, Tlinear 25.6%, K≈3.52%; Sub-interval 57-68: Tactual 30.9%, Tlinear 32.0%, K≈3.44%; Sub-interval 69-80: Tactual 37.1%, Tlinear 38.4%, K≈3.39%. The preset nonlinear coefficient threshold is 3.5%. The key sub-intervals with nonlinear coefficients higher than the threshold are selected as follows: 21-32 (3.91%), 33-44 (3.65%), and 45-56 (3.52%). The sub-intervals 57-68 and 69-80 are not included in the key sub-interval range because K is lower than the threshold. Subsequent compensation optimization will only be performed on these three key sub-intervals.

[0054] Furthermore, based on the original data of the key sub-intervals, a local compensation function corresponding to each sub-interval is constructed by using polynomial fitting or piecewise linear interpolation to ensure a smooth transition of transmittance changes after compensation. In this embodiment of the invention, a local compensation function for each sub-interval is constructed using piecewise linear interpolation based on the original data of three key sub-intervals (e.g., sub-intervals 21-32 contain 12 sets of data such as (21, 10.1%) and (32, 14.5%)). For sub-intervals 21-32: Let the grayscale value before compensation be G, and the grayscale value after compensation be G'. The compensation function G'=1.04G-0.84 is obtained through two-point fitting. For example, when G=21, G'=1.04×21-0.84=21.84≈22; For sub-intervals 33-44: The fitting function G'=1.03G-0.52. When G=33, G'=1.03×33-0.52=33.47≈33; For sub-intervals 45-56: The fitting function G'=1.02G-0.28. When G=45, G'=1.02×45-0.28=45.62≈46. To ensure a smooth transition in transmittance after compensation, the transmittance after compensation at the boundaries of adjacent sub-intervals is calculated: for sub-interval 21-32, the boundary G=32, and after compensation G'=33.44, corresponding to a transmittance of 14.8%; for sub-interval 33-44, the boundary G=33, and after compensation G'=33.47, corresponding to a transmittance of 14.9%. The difference between the two is 0.1%, which meets the requirement of a smooth transition (difference ≤ 0.2%) and avoids abrupt changes in transmittance.

[0055] Furthermore, a parameter index table for the local compensation function is established, and gray values ​​are associated with and stored with the corresponding compensation coefficients. The test gray values ​​containing compensation corrections are sent to the POEP via the control computer, and the transmittance data of each test point is re-acquired to verify whether the compensation effect meets the preset requirements. In this embodiment of the invention, a parameter index table for the local compensation function is established. The table structure includes three columns: "sub-interval range - compensation function coefficient - grayscale value mapping". For example, the coefficient corresponding to sub-interval 21-32 is "1.04, -0.84", and grayscale value 21 maps to 22, 22 maps to 23, etc.; the coefficient for sub-interval 33-44 is "1.03, -0.52", and grayscale value 33 maps to 33, 34 maps to 34, etc. The index table is stored in the control computer in sub-interval order. The control computer sends test grayscale values ​​containing compensation corrections to POEP1, covering 3 key sub-intervals and boundary grayscale levels, sending a total of 36 test grayscale values ​​(12 for each sub-interval). Simultaneously, transmittance data at nine test points were re-acquired using optical brightness measurement equipment, and the transmittance deviation after compensation was calculated: the maximum deviation after compensation was 0.4% in sub-interval 21-32 (theoretical transmittance 11.5%, actual 11.9% when G=25), 0.35% in sub-interval 33-44 (theoretical 17.2%, actual 17.55% when G=38), and 0.32% in sub-interval 45-56 (theoretical 23.0%, actual 23.32% when G=50). The preset compensation effect requirement was a deviation ≤0.5%, which was met in all three key sub-intervals. Since the deviation in non-key sub-intervals was already ≤0.3%, the overall compensation effect met the standard.

[0056] Furthermore, if the compensated transmittance deviation exceeds the allowable range, the parameters of the local compensation function or the fitting method are adjusted, and the verification and adjustment are repeated until a mapping relationship model between the gray value and the actual transmittance that meets the requirements is formed.

[0057] In this embodiment of the invention, assuming that after the initial compensation, the transmittance deviation is 0.6% when G=28 in sub-intervals 21-32 (exceeding the allowable range of 0.5%), the parameters of the local compensation function for this sub-interval need to be adjusted. The cause of the deviation is analyzed: the original compensation function G'=1.04G-0.84 is insufficient for compensating the low grayscale end. The coefficient is adjusted to G'=1.05G-1.02, and the value of G' when G=28 is recalculated as G'=1.05×28-1.02=28.38≈28. Transmittance data is re-collected, and the actual transmittance is 13.2%, the theoretical value is 13.1%, and the deviation is 0.07%, which meets the requirements. If the deviation still exceeds the range after adjusting the parameters, the fitting method is changed to second-order polynomial fitting, for example, the fitting function for sub-intervals 21-32 is G'=0.0005G. 2+1.03G-0.9, calculating G' when G=28: G'=0.0005×784+1.03×28-0.9=0.392+28.84-0.9=28.332≈28, corresponding to a transmittance deviation of 0.05%. Repeat the process of "adjusting parameters / changing methods - sending test grayscale values ​​- collecting data - verifying deviations" until the deviations in the three key sub-intervals are all ≤0.5% after compensation, and the deviations in the non-key sub-intervals are ≤0.3%. This ultimately forms a mapping model between grayscale values ​​and actual transmittance, covering the full grayscale level 0-255, ensuring that the grayscale value control accuracy corresponding to any target transmittance meets system requirements.

[0058] Furthermore, the step of dividing the POEP into blended and non-blended regions based on the overlap ratio of adjacent projector images and the pixel array distribution of the POEP, and defining the grayscale gradient rules within the blended region, includes: Based on the obtained overlap ratio of adjacent projector images and combined with the pixel resolution of the projected image, the pixel coordinate range of the overlapping area in the projected image is calculated, and the pixel boundary of the fusion area corresponding to each POEP is determined. In this embodiment of the invention, given that the overlap ratio of adjacent projector images is 20% and the pixel resolution of the projected image is 3840×2160 (the projection width of a single projector corresponds to 3840 pixels), the pixel coordinate range of the overlapping area in the projected image is calculated: the projection width of a single projector is 3840 pixels, and the number of overlapping pixels corresponding to the 20% overlap ratio is 3840×20%=768 pixels. The pixel coordinates of the projected image of the left projector are (0,0)-(3840,2160), and the coordinates of its right overlapping area are (3840-768,0)-(3840,2160), i.e. (3072,0)-(3840,2160). The coordinates of the projected image of the middle projector are (3072,0)-(6912,2160), and its left overlapping area is the same as the right overlapping area of ​​the left projector (3072,0)-(3840,2160), while its right overlapping area is (6912-768,0)-(6912,2160), i.e. (6144,0)-(6912,2160). Based on the spatial correspondence between POEP and the projected image pixels (POEP1 pixel (1000, 500) corresponds to the left projector image (2000, 1000)), the pixel boundaries of the blending area of ​​each POEP are determined: the pixel boundary of the blending area of ​​POEP1 (corresponding to the left projector) is (1000 + (3072 - 2000) × (1920 / 3840), 0) - (1920, 1200), i.e. (1480, 0) - (1920, 1200), and the left blending area boundary of POEP2 (corresponding to the middle projector) is (0, 0) - (440, 1200) and the right boundary is (1480, 0) - (1920, 1200), ensuring that the POEP blending area and the projection overlap area correspond accurately.

[0059] Furthermore, based on the requirements of dark field fusion, a black light suppression target for the fusion area is set. Combined with the minimum transmittance parameter of POEP, the minimum gray value at the edge of the fusion area is determined to ensure that the amount of black light superposition in the overlapping area is lower than the threshold perceptible to the human eye. In this embodiment of the invention, based on the requirements for dark field fusion, the black light suppression target of the fusion region is set as ≤0.3 cd / m² of black light superposition in the overlapping region. 2 (The human eye's perceptible threshold is 0.5 cd / m) 2 The projector's black level output brightness is known to be 0.2 cd / m². 2 (Single unit), the overlapping area is the superposition of two projectors. Without suppression, the amount of black light superposition = 0.2 × 2 = 0.4 cd / m². 2 It is necessary to reduce the concentration of 0.1 cd / m³ by POEP. 2The minimum transmittance parameter for POEP is 0.08% (at a grayscale value of 0). Transmittance is linearly related to black light attenuation. The required minimum transmittance is calculated as follows: target black light superposition amount 0.3 cd / m². 2 =0.2×2×T (T is the transmittance of POEP), solving for T, we get T=0.3 / (0.4)=75%. Based on the POEP gray-level-transmittance mapping relationship model (the relationship between gray value G and transmittance T is T=0.0001G), 3 +0.02G 2 (+0.1G+0.05), calculate the grayscale value corresponding to 75% target transmittance in reverse: substituting into the formula, we get G≈198. Determine the minimum grayscale value at the edge of the fusion region to be 198, that is, set the outermost pixel of the fusion region (closest to the non-overlapping area) to a grayscale value of 198, ensuring that the black light superposition in the overlapping area is reduced to 0.3 cd / m². 2 It is below the threshold that the human eye can perceive.

[0060] Furthermore, based on the characteristics of human vision, the grayscale gradient curve types within the fusion area are designed, including linear gradient, gamma curve gradient, or custom nonlinear gradient, so that the brightness transition in the overlapping area is smooth without obvious jumps. In this embodiment of the invention, based on the characteristics of human visual perception (more sensitive to brightness changes in low-brightness areas), the grayscale gradient curve type within the fusion area is designed as a gamma curve gradient, with the gamma value set to 0.8 (a gamma value < 1 enhances the smoothness of the gradient in low-brightness areas). The starting point of the gradient curve is the grayscale value of 255 (highest grayscale level, 99.3% transmittance, no black light suppression) of the inner pixel (closest to the projection center) within the fusion area, and the ending point is the grayscale value of 198 (lowest grayscale value, 75% transmittance) of the outer pixel within the fusion area. The gradient range covers 440 pixels (width direction) of the POEP fusion area. The gamma curve gradient formula is G = Gmax - (Gmax - Gmin) × (x / X)^γ (G is the grayscale value of a pixel, Gmax = 255, Gmin = 198, x is the distance from the pixel to the inner edge, X = 440, γ = 0.8). For example, when x=110 (1 / 4 gradient distance), G=255-(255-198)×(110 / 440)^0.8=255-57×(0.25)^0.8≈255-57×0.336≈255-19.15≈235.85≈236; when x=220 (1 / 2 gradient distance), G≈255-57×(0.5)^0.8≈255-57×0.574≈255-32.72≈222.28≈222, ensuring that the grayscale change rate of the low brightness end (close to 198) is lower than that of the high brightness end, which conforms to the characteristics of human visual perception and avoids obvious jumps in brightness transition.

[0061] Furthermore, based on the grayscale gradient curve, the target grayscale value of each pixel within the fusion region is calculated, and the pixels in the non-fusion region are set to the highest grayscale level, generating the pixel data matrix of the initial fusion control image; In this embodiment of the invention, the target grayscale value of each pixel within the fusion region is calculated according to the gamma gradient curve formula. Taking the left fusion region of POEP2 (pixel coordinates (0,0)-(440,1200)) as an example, the x-axis is the width direction (0-440), and the y-axis is the height direction (0-1200). All pixels with the same y-coordinate have the same grayscale value under the same x-coordinate. When x=0 (inner edge), G=255; when x=100, G=255-57×(100 / 440)^0.8≈255-57×0.305≈255-17.39≈237.61≈238; when x=440 (outer edge), G=198. A total of 440 grayscale values ​​corresponding to x-coordinates are calculated to form the grayscale matrix of the fusion region. The pixels in the non-fusion region ((440,0)-(1480,1200)) of POEP2 are set to the highest gray level of 255 to generate the pixel data matrix of the initial fusion control image. The matrix size is 1920×1200 (POEP resolution), and each element is a gray value of 0-255. For example, the matrix element (500,600) (non-fusion region) has a value of 255, and the element (200,600) (fusion region x=200) has a value of 225.

[0062] Furthermore, the pixel data matrix of the initial fusion control image is smoothed to eliminate abrupt grayscale changes between pixels, thus avoiding jagged edges or mottled shadows on the projected image and forming the initial fusion control image data.

[0063] In this embodiment of the invention, a 3×3 Gaussian filtering algorithm is used to smooth the pixel data matrix of the initial fusion control image. The Gaussian kernel function parameter σ=1.2 (controlling the smoothing degree), and the filtering weight matrix is ​​[[0.075,0.124,0.075],[0.124,0.204,0.124],[0.075,0.124,0.075]]. The boundary pixels between the fusion and non-fusion regions (such as the x=440 pixel column of POEP2) are processed with special attention. The weighted average of the gray values ​​of this column and its adjacent columns (x=439, x=441) is calculated: x=440 pixel gray value 198, x=439 pixel gray value 199, x=441 pixel gray value 255. After filtering, the gray value of x=440 pixel = 198×0.204 + 199×0.124 + 25. 5×0.124+198×0.124+255×0.075≈40.39+24.68+31.62+24.55+19.13≈140.37 (Incorrect, recalculate: Correct 3×3 filtering requires taking the surrounding 8 pixels plus the center pixel. Here, taking the boundary x=440 as an example, there are no pixels on the right, so mirror filling is used. The x=441 pixel is copied to x=442, and the calculated gray value is ≈200). After processing, check for abrupt gray-level changes between pixels: the maximum gray-level difference between adjacent pixels in the fusion area is 2 (the original maximum difference is 4), and the difference at the boundary decreases from 57 to 3, eliminating jagged edges and mottled light and shadow. Encode the smoothed pixel data matrix in BMP format to form the initial fusion control image data, with a data size of 1920×1200×8bit=2.25MB.

[0064] Furthermore, the optical edge modulation of the image projected by the projector includes: Based on the pixel array spatial topology model and gray-level-transmittance mapping relationship model of POEP, a dynamic light modulation logic library is constructed. This logic library contains pixel transmittance control rules corresponding to different image edge types, including hard edges, soft edges and gradient edges. In this embodiment of the invention, a pixel array spatial topology model based on POEP (pixel pitch 0.15mm, effective display area 18cm×11.25cm, pixel coordinates (x,y) corresponding to physical positions (0.15xmm, 0.15ymm)) and a grayscale-transmittance mapping relationship model (T=0.0001G) are used. 3 +0.02G 2(+0.1G+0.05), constructing a dynamic light modulation logic library. The logic library contains transmittance control rules for three types of image edges: hard edges (such as text borders, geometric outlines), with a transmittance control precision of ±0.5%, the rule being "edge pixel gray value = 255 - (edge ​​contrast × 10), gray value ≤ 200 when contrast > 80%"; soft edges (such as gradient background transition areas), with a control precision of ±1%, the rule being "edge pixel gray value gradually changes according to the gamma curve (γ=0.9), gray value difference between adjacent pixels ≤ 3"; gradient edges (such as the boundary between fused and non-fused areas), with a control precision of ±0.8%, the rule being "edge pixel gray value = 198 + (distance from fusion center × 0.12), distance range 0-440 pixels." Each rule is associated with a specific parameter calculation example, such as when the contrast of a hard edge is 85%, the gray value = 255 - (85 × 10) = 170, corresponding to a transmittance T = 0.0001 × 170. 3 +0.02×170 2 +0.1×170+0.05≈491.3+578+17+0.05≈1086.35 (Error, exceeding 100% transmittance. Correction rule is "when contrast ratio > 60%, grayscale value ≤ 220". When contrast ratio is 85%, grayscale value = 255 - (85×0.4) = 255 - 34 = 221, corresponding to transmittance ≈ 95.2%). Ensure that the rule can be directly used for pixel control.

[0065] Furthermore, through an intelligent image edge feature extraction algorithm, the edge contours, gradient changes, and texture features of the projected image are captured in real time to generate an image edge feature map. Based on the image edge feature map, the projected image is subjected to global texture detail enhancement processing. In order to separate the high-frequency edge information and low-frequency background information corresponding to the projected image through dynamic threshold segmentation technology, multi-scale feature layered data is generated. In this embodiment of the invention, the Canny edge detection algorithm (threshold 100-200, Gaussian filter σ=1.5) is used as an intelligent image edge feature extraction algorithm to capture the edge contours of the projected image (resolution 3840×2160) in real time, such as the wing contours of the "airplane simulation model" and the line contours of the "dashboard scale" in the projected image. This generates an image edge feature map containing edge position, direction, and intensity. The map is presented in grayscale form, with edge pixel values ​​of 255 and background pixel values ​​of 0. Based on the map, global texture detail enhancement processing is performed using a histogram equalization algorithm (contrast enhancement coefficient 1.2) to expand the image grayscale range from 0-200 to 0-255, enhancing the texture clarity of the "dashboard scale". A dynamic thresholding technique (the threshold is adaptively adjusted based on the average pixel grayscale value, ranging from 120 to 180) separates high-frequency edge information from low-frequency background information: high-frequency edge information consists of pixels with an intensity >150 in the edge feature map (such as the edge of an aircraft wing), while low-frequency background information consists of pixels with an intensity ≤150 (such as the background color of a dashboard). Multi-scale feature layered data is generated, divided into three scales: Scale 1 (1920×1080, capturing large outlines), Scale 2 (960×540, capturing medium details), and Scale 3 (480×270, capturing subtle textures). Each layer of data is labeled with edge intensity and the average background grayscale value to ensure thorough separation of high-frequency and low-frequency information.

[0066] Furthermore, skeleton extraction and connectivity analysis are performed on high-frequency edge information in multi-scale feature-layered data to construct an edge feature spatial topology network and clarify the branching relationships and continuity characteristics of the edges. Based on the edge feature spatial topology network, edge segments with related relationships are aggregated into complete edge regions. At the same time, a background feature masking mechanism is used to delineate pure background regions that are not related to the edge regions. In addition, combined with the fusion objectives of the projection scene, including black light suppression in dark fields and smooth transition in bright fields, the modulation priority of the complete edge regions is sorted, and the core modulation edge regions and auxiliary modulation edge regions are marked. In this embodiment of the invention, the Zhang-Suen skeleton extraction algorithm is used to extract the edge skeleton from the high-frequency edge information (such as the wing outline edge in scale 1) in the multi-scale feature layered data. Edges with a width ≥ 3 pixels are compressed into a 1-pixel skeleton. At the same time, a connectivity analysis is performed using a connected component labeling algorithm (neighborhood 8-connectivity determination) to identify three discontinuous wing edge segments (lengths of 200, 150, and 80 pixels, respectively). An edge feature spatial topology network is constructed, where network nodes are the endpoints of edge segments, and edges represent the spatial relationships between segments (distance < 10 pixels is considered a relationship). Based on the network, the three edge segments are aggregated into a complete wing edge region (total length 430 pixels). Simultaneously, a background feature masking mechanism is activated, setting regions with a background grayscale variance < 20 as pure background regions (such as the dashboard background color region, variance 15), and defining pure background regions (area 120,000 pixels) that are not related to the wing edge region. Combined with the projection scene fusion objectives (dark field black light suppression, bright field transition smoothing), the modulation priority of the complete edge region is sorted: the brightness in the dark field > 0.4 cd / m² is prioritized. 2 Edge regions (such as the area where light is reflected in the dark field at the edge of the wing) are marked as core modulation edge regions (priority 1), edge regions with a gray level difference > 50 in the bright field (such as the transition area between the instrument scale and the background color) are marked as auxiliary modulation edge regions (priority 2), and pure background regions are marked as unmodulated regions (priority 3). The sorting results are stored as a region priority matrix with the same size as the projected image (3840×2160), and each element is a priority value of 1-3.

[0067] Furthermore, the topological information of the core modulation edge region, auxiliary modulation edge region, and pure background region is mapped to the pixel array coordinate system of POEP to generate a localization result map containing region type and modulation priority. Based on the localization result map, the edge region and non-edge region that need to be modulated are located. In this embodiment of the invention, based on the spatial correspondence between the POEP pixel array and the projected image (POEP pixel (x,y) corresponds to the projected image pixel (2x,2y), since the POEP resolution is 1920×1200 and the projected image is 3840×2160), the topological information of the core modulation edge region (such as the wing edge reflected light region, with projection coordinates (1000,800)-(1430,1200)), the auxiliary modulation edge region (such as the instrument panel scale transition area, with projection coordinates (2000,500)-(2200,700)), and the pure background region (such as the instrument panel background color, with projection coordinates (1800,300)-(1950,450)) is mapped to the POEP pixel array coordinate system. The mapping relationships are calculated as follows: the projected coordinates of the core modulation edge region (1000, 800) correspond to the POEP coordinates (500, 400), and the projected coordinates (1430, 1200) correspond to the POEP coordinates (715, 600). Therefore, the coordinate range of the core region in POEP is (500, 400) - (715, 600). The projected coordinates of the auxiliary region (2000, 500) - (2200, 700) correspond to the POEP coordinates (1000, 250) - (1100, 350). The pure background region corresponds to the POEP coordinates (900, 150) - (975, 225). A localization result map is generated with a size of 1920×1200. The core region is labeled "Type 1 - Priority 1", the auxiliary region is labeled "Type 2 - Priority 2", and the pure background region is labeled "Type 3 - Priority 3". The unlabeled areas are non-modulated regions. The edge and non-edge regions to be modulated are accurately located using the map.

[0068] Furthermore, based on the image edge feature map, the corresponding rules in the dynamic light modulation logic library are called, and the gray values ​​of each pixel of POEP are dynamically adjusted to change the phase distribution of the projected light in the edge region, thereby achieving directional modulation of the edge light field. In this embodiment of the invention, the corresponding rules of the dynamic light modulation logic library are called according to the edge type (hard edge in the core region and soft edge in the auxiliary region) in the image edge feature map. The core modulation edge region (hard edge, contrast 85%) calls the hard edge rule: gray value = 255 - (85 × 0.4) = 221, corresponding to a transmittance of 95.2%. The control computer sends a control signal of gray value 221 to the pixels (500, 400) - (715, 600) in this region of the POEP to dynamically adjust the pixel transmittance. The auxiliary modulation edge region (soft edge, gradient range 100 pixels) calls the soft edge rule: gradient according to the gamma curve (γ = 0.9), the gray value of the inner pixel (1000, 250) is 255, the gray value of the outer pixel (1100, 350) is 220, calculate the gray value of the middle pixel, such as the gray value of (1050, 300) = 255 - (255 - 220) × (50 / 100)^0.9≈236, and sends the corresponding gray control signal to the pixels in this region. The phase distribution of the projected light in the edge region is changed by adjusting the gray value: the transmittance of the core region is 95.2%, which delays the phase of the projected light by 0.12π. The gradually changing transmittance of the auxiliary region makes the phase linearly transition from 0.05π to 0.1π, thereby realizing the directional modulation of the edge light field and ensuring that the light field in the edge region is adapted to the background light field.

[0069] Furthermore, a nonlinear optical enhancement mechanism is introduced to perform phase superposition and amplitude calibration on the modulated light in the edge region, thereby enhancing the light intensity transition difference between the edge and non-edge regions, while suppressing the light field distortion in the non-edge region and ensuring the sharpness of the main subject in the image. In this embodiment of the invention, a nonlinear optical enhancement mechanism is introduced, and a phase superposition algorithm is used to process the modulated light in the edge region: the phase superposition coefficient of the core modulation edge region is set to 1.2, and the original phase of 0.12π is superimposed to 0.12π×1.2=0.144π, thereby enhancing the light field intensity in the edge region; the auxiliary modulation edge region adopts segmented superposition, with a superposition coefficient of 1.1 (phase 0.05π×1.1=0.055π) in the inner 1 / 2 region and a superposition coefficient of 1.2 (phase 0.1π×1.2=0.12π) in the outer 1 / 2 region, thereby strengthening the light intensity transition difference between the edge and non-edge regions. Simultaneously, amplitude calibration is performed, employing an amplitude normalization algorithm to control the modulation light amplitude in the core area within the range of 0.9-1.0 (1.0 for non-edge areas), and the amplitude in the auxiliary area within the range of 0.92-0.98, suppressing light field distortion in non-edge areas. In the pure background area, the amplitude fluctuation is controlled within ±0.02 using an amplitude stabilization algorithm, avoiding mottled background light field caused by edge modulation, ensuring the clarity of the main subjects such as the "dashboard background color," and ensuring that the grayscale uniformity deviation of the modulated background area is ≤1%.

[0070] Furthermore, by integrating the directional modulated light in the edge region with the original projected light in the non-edge region, the light field interference noise during the modulation process is eliminated, generating a projected image with natural edge transitions and complete details after optical edge modulation.

[0071] In this embodiment of the invention, an optical field interference cancellation algorithm is used to integrate the directional modulated light from the edge region and the original projected light from the non-edge region. The algorithm analyzes the frequency difference between the two light fields (500-800Hz for the edge region and 300-500Hz for the non-edge region) and sets an interference noise filtering threshold of 450Hz to filter interference signals in the frequency overlap range. For example, the superimposed light fields of the core modulation edge region (650Hz frequency) and the adjacent non-edge region (480Hz frequency) are filtered to remove noise signals in the 450-500Hz range, retaining the effective components of the 650Hz edge light field and the 480Hz background light field. After integration, the resulting optically edge-modulated projection image is generated: the light intensity transition in the core edge region (wing edge) is smooth, and the grayscale difference is reduced from 20 to 5; the auxiliary edge region (dashboard scale) has no obvious jumps, and the width of the transition area is reduced from 15 pixels to 8 pixels; the pure background region (dashboard background color) has no mottled light shadows, and the grayscale uniformity reaches 99%. Verified using optical brightness measurement equipment, the image edges exhibit natural transitions and complete detail, with a black light overlay of 0.28 cd / m² in overlapping areas during dark scenes. 2 (less than 0.3 cd / m 2 (Target) The edge transition in the bright field has no visual abrupt changes and meets the fusion requirements.

[0072] Furthermore, the step of optimizing the edge blending based on feedback information, adjusting the grayscale gradient rules of the blending region and the pixel transmittance parameters of the POEP by controlling the computer, updating the blending control image data and re-sending it to the POEP, and repeatedly executing optical edge modulation and evaluation feedback until the projected image reaches the preset edge blending standard includes: The evaluation indicators of edge blending effect of the projected image are obtained through visual observation, including the brightness uniformity of the overlapping area, the blurring degree of the blending boundary, and the brightness consistency between the non-overlapping area and the overlapping area. In this embodiment of the invention, the edge blending effect evaluation index of the projected image is obtained by combining visual observation with optical measurement equipment. The evaluation criteria are set as follows: brightness uniformity ≥ 95%, blending boundary blurring degree (transition zone width) ≤ 10 pixels, and brightness consistency deviation between non-overlapping and overlapping areas ≤ 5%. The brightness uniformity of the overlapping area (the blending area between the left and middle projectors, 768 pixels wide) is measured: 20 test points are evenly selected within the area, and the measured brightness values ​​range from 185 to 198 cd / m². 2Uniformity calculation = (minimum / maximum) × 100% = (185 / 198) × 100% ≈ 93.4% (not up to standard); Blur level of fusion boundary: The width of the transition area was measured using image analysis tools. The width of the transition area in the core edge region was 12 pixels (not up to standard), and the width of the transition area in the auxiliary edge region was 8 pixels (up to standard); Brightness of the non-overlapping area (non-fusion area of ​​the left projector) was 195 cd / m². 2 The average brightness of the overlapping area is 191.5 cd / m². 2 Consistency deviation = |195-191.5| / 195×100%≈1.8% (meets the standard). Record the non-compliant indicators: brightness uniformity 93.4%, core edge area transition zone width 12 pixels, and clarify the direction for subsequent adjustments.

[0073] Furthermore, for areas where brightness uniformity is substandard, the corresponding POEP pixel range is located, and the reasons for the transmittance deviation in this area are analyzed, including POEP installation position deviation, grayscale-transmittance mapping model error, or projector projection angle deviation. In this embodiment of the invention, the corresponding POEP pixel range is located for the overlapping area (projection coordinates 3072-3840, 0-2160) where the brightness uniformity is substandard: based on spatial correspondence, this projection area corresponds to pixels (1480, 0) - (1920, 1200) of POEP1. The cause of the transmittance deviation is analyzed: first, the POEP installation position is checked; the distance between POEP1 and the projector lens is measured to be 18cm (standard 10-20cm, meets the requirements) using a laser positioning device, but there is a horizontal offset of 2mm (the lens optical axis deviates 2mm from the POEP center); second, the grayscale-transmittance mapping model is verified; pixels (1600, 600) of POEP1 are selected, and a grayscale value of 198 (target transmittance 75%) is input; the measured transmittance is 74.8% (deviation 0.2%, within the allowable range); finally, the projector projection angle is checked; the horizontal projection angle deviation of the left projector is measured to be 0.5° (standard 0°, deviation exists). The main causes of the deviation were identified as follows: the POEP installation was horizontally offset by 2mm, the projector projection angle was deviated by 0.5°, and the grayscale-transmittance model had no significant error.

[0074] Furthermore, if the POEP installation position is misaligned, the spatial correspondence between the POEP and the projector lens can be corrected by adjusting the displacement or angle parameters of the fixing mechanical structure, and the pixel control parameters of the area can be redefined. In this embodiment of the invention, to address the issue of a 2mm horizontal offset in the installation of POEP1, the horizontal displacement parameters of the fixing mechanical structure are adjusted: each rotation of the mechanical structure's horizontal adjustment knob corresponds to a 1mm displacement; rotating it clockwise twice shifts POEP1 2mm towards the lens optical axis. After adjustment, a remeasurement using a laser positioning instrument shows that the alignment deviation between the POEP center and the lens optical axis is ≤0.5mm, meeting the installation standard. Based on the corrected spatial correspondence, the control parameters for pixels (1480,0)-(1920,1200) of POEP1 are redefined: the original projection coordinate 3072 corresponds to POEP pixel 1480, and after correction, it corresponds to 1481. The coordinate mapping table between pixels in this area and the projected image is updated to ensure that each POEP pixel accurately corresponds to the target position in the overlapping projection area, avoiding local transmittance deviations caused by positional offsets.

[0075] Furthermore, if the error is due to the gray-level-transmittance mapping model, supplement the transmittance test data of the area, update the gray-level-transmittance mapping model, and adjust the target gray value of the corresponding pixel. In this embodiment of the invention, it is assumed that in subsequent verification, it is found that in the (800,400)-(900,500) pixel region of POEP2, when the input grayscale value is 220 (target transmittance 92%), the measured transmittance is 89.5% (deviation 2.5%, exceeding the allowable range by 0.5%), which is determined to be an error in the grayscale-transmittance mapping model. Supplementary transmittance test data for this region is collected: 10 pixels in this region are selected, and test signals with grayscale values ​​of 180-240 (intervals of 10) are sent. Transmittance data is collected 3 times for each grayscale value, such as an average measured transmittance of 89.5% for a grayscale value of 220 and 91.2% for a grayscale value of 230. Based on the supplementary data, the mapping model is updated, and the coefficients of the original third-order polynomial are adjusted to T=0.00009G. 3 +0.021G 2 +0.09G+0.06, recalculate the gray value corresponding to 92% target transmittance: substitute into the formula to get G≈225, adjust the target gray value of the pixels in this area from 220 to 225 to ensure that the deviation between the measured transmittance and the target value is ≤0.5%.

[0076] Furthermore, if the projector's projection angle is deviated, the geometric correction data of the multi-channel projection is combined to adjust the division range of the blending area and the grayscale gradient rules so that the blending effect adapts to the change in projection angle. In this embodiment of the invention, to address the issue of a 0.5° deviation in the horizontal projection angle of the left-side projector, and in conjunction with the geometric correction data of the multi-channel projection (the horizontal offset of the projected image has been measured to be 30 pixels using geometric correction software), the division range of the fusion area is adjusted: the original overlapping area projection coordinates of 3072-3840 are adjusted to 3042-3770 (shifted 30 pixels to the left), corresponding to the pixel range of the fusion area of ​​POEP1 being adjusted from (1480,0)-(1920,1200) to (1465,0)-(1885,1200). Simultaneously, the grayscale gradient rules are adjusted: the original gamma curve gradient range of 440 pixels is adjusted to 425 pixels, and the gradient formula is updated to G=255-(255-198)×(x / 425)^0.8 (where x is the pixel distance within the adjusted fusion area), ensuring precise matching between the fusion area and the offset projection overlap area, eliminating the problem of discontinuous edge transitions caused by the projection angle deviation.

[0077] Furthermore, the fusion control image data is updated according to the adjusted parameters, sent to the corresponding POEP and loaded to take effect, and visual observation and evaluation are performed again until all evaluation indicators meet the preset edge fusion standards.

[0078] In this embodiment of the invention, the fusion control image data is regenerated based on the adjusted parameters (2mm horizontal displacement of POEP1, pixel range of the fusion area (1465,0)-(1885,1200), and grayscale gradient formula update): the grayscale values ​​of the corresponding areas in the pixel data matrix are updated, such as adjusting the grayscale value of the POEP1 (1465,600) pixel from 198 to 199. After encoding in BMP format, the data is sent to POEP1 through the network interface of the control computer and loaded to take effect. Visual observation and evaluation are performed again: the brightness uniformity of the overlapping area is measured to be approximately 95.9% (meets the standard), the width of the transition area in the core edge region is 9 pixels (meets the standard), and the brightness consistency deviation is 1.5% (meets the standard). When all evaluation indicators meet the preset standards (brightness uniformity ≥95%, transition area width ≤10 pixels, deviation ≤5%), the adjustment is stopped, and the edge fusion effect optimization is completed.

[0079] Furthermore, for areas where brightness uniformity is substandard, the corresponding POEP pixel range is located, and the reasons for the transmittance deviation in these areas are analyzed, including: Using a brightness acquisition device, multiple sampling points are selected in the overlapping and non-overlapping areas of the projected image, and the actual brightness values ​​of each sampling point are collected to establish a brightness distribution data matrix. In this embodiment of the invention, a brightness acquisition device (measurement range 0.01-1000 cd / m²) is used. 2Sampling points were selected in the overlapping area (left and middle projector blending area, projection coordinates 3072-3840, 0-2160) and non-overlapping area (left projector non-blending area 300-1000, 0-2160, middle projector non-blending area 4000-4800, 0-2160) of the projected image (accuracy ±1%). In the overlapping area, sampling points were selected at 50×50 pixel intervals, totaling 180 sampling points (3072, 0), (3122, 50), ... (3840, 2160), etc. In the non-overlapping area, sampling points were selected at 100×100 pixel intervals, with 50 sampling points selected in each area, for a total of 280 sampling points. During data acquisition, the device probe was vertically aligned with the projection screen, 10cm away, to avoid ambient light interference. Five actual brightness values ​​were collected for each sampling point, and the average value was used as the final data. For example, the average brightness of the overlapping area sampling point (3500, 1000) was 192 cd / m². 2 (3200,500) 186cd / m 2 The sampling points in the non-overlapping region (500, 1000) have a density of 195 cd / m³. 2 (4500, 1000) 194 cd / m 2 A brightness distribution data matrix of 180×3 (overlapping area) + 100×3 (non-overlapping area) is established according to the format of "area type-sampling point coordinates-brightness value" to ensure that the data of each sampling point is traceable.

[0080] Furthermore, the standard deviation of the brightness values ​​of each sampling point in the overlapping area is calculated. If the standard deviation exceeds the preset threshold, the brightness uniformity of the area is determined to be substandard, and the area with the largest standard deviation is identified as the key optimization area. In this embodiment of the invention, based on the brightness distribution data matrix, the standard deviation formula "σ=√[Σ(xi-x̄)" is used. 2 The standard deviation of the brightness values ​​of each sampling point within the overlapping region is calculated using the formula: x̄ = (192 + 186 + ... + 198) / 180 ≈ 191.5 cd / m². The average brightness of the overlapping region is calculated as follows: x̄ = (192 + 186 + ... + 198) / 180 ≈ 191.5 cd / m². 2 The calculation yields σ≈4.2cd / m 2 The preset standard deviation threshold for brightness uniformity is 3.5 cd / m². 2 (When σ≤3.5cd / m) 2 When uniformity is ≥95%, the current σ = 4.2 cd / m 2 If the value exceeds the threshold, the brightness uniformity of the area is deemed substandard. By comparing the standard deviations of each sub-region, it was found that the standard deviation of the overlapping sub-region (3100-3300, 800-1200) was the largest, reaching 6.8 cd / m². 2 This sub-region contains sampling points (3122,800) (luminance 182 cd / m²).2 ), (3200,1000) (184cd / m 2 The area with the largest difference from the regional average brightness is identified as the key area for optimization, and subsequent adjustments will be prioritized for this area.

[0081] Furthermore, compare the brightness difference between adjacent sampling points in the overlapping and non-overlapping areas. If the brightness difference exceeds the acceptable range of human vision, the transition of the fusion boundary is deemed unqualified, and the corresponding POEP pixel column or row is recorded. In this embodiment of the invention, by comparing the brightness difference between adjacent sampling points in the overlapping region and the non-overlapping region, adjacent sampling points are defined as sampling points at the edge of the overlapping region and sampling points at the edge of the non-overlapping region (spacing ≤ 50 pixels), and a total of 30 sets of adjacent points are selected. For example, the sampling point in the overlapping region (3072, 1000) (brightness 188 cd / m²) 2 ) and non-overlapping sampling points (3071, 1000) (195 cd / m 2 The difference is |188-195| = ​​7 cd / m 2 Overlapping region sampling points (3840, 1500) (193 cd / m²) 2 ) and non-overlapping region sampling points (3841, 1500) (194 cd / m 2 The difference is 1 cd / m 2 The preset acceptable range for human visual perception of brightness difference is ≤5 cd / m². 2 Among them, the brightness difference between 12 adjacent points exceeded the range (maximum 7 cd / m²). 2 The transition of the fusion boundary is deemed unqualified. The corresponding POEP pixel column or row is recorded: According to the spatial correspondence, the sampling point (3072, 1000) in the overlapping area corresponds to the pixel column (1480, 500) of POEP1. Therefore, the 1480th and 1481st pixel columns of POEP1 (a total of 10 columns) are recorded as the pixel columns corresponding to the unqualified transition, thus clarifying the pixel range for boundary optimization.

[0082] Furthermore, the relationship between the target grayscale value and the actual transmittance of the POEP pixels corresponding to the key optimization area is analyzed. If the deviation between the actual transmittance and the theoretically calculated value exceeds the allowable range, it is determined to be an error in the grayscale-transmittance mapping model. In this embodiment of the invention, by analyzing the POEP pixel range corresponding to the key optimization area (projected coordinates 3100-3300, 800-1200), based on spatial correspondence, this area corresponds to the (1490-1590, 400-600) pixels of POEP1. The target grayscale values ​​of the POEP pixels in this area are retrieved (based on the grayscale-transmittance mapping model, a target transmittance of 75% corresponds to a grayscale value of 198). The actual transmittance of 10 representative pixels in this area is collected using a transmittance measurement device, such as the actual transmittance of POEP1 (1500, 500) pixels being 73.2% and (1550, 550) pixels being 72.8%. The theoretically calculated transmittance is 75%, and the actual and theoretical deviations are |73.2-75|=1.8% and |72.8-75|=2.2%. The preset allowable deviation range is ≤0.5%. Since the current deviations all exceed this range, it is determined to be an error in the grayscale-transmittance mapping model, requiring subsequent updates to the model parameters to correct the deviation.

[0083] Furthermore, the actual installation posture of the POEP is captured by an image acquisition device and compared with the preset installation parameters. If the positional or angular deviation exceeds the adaptation range, it is determined to be a POEP installation position deviation. In this embodiment of the invention, the actual installation posture of POEP1 is captured from two angles, front and side, using an image acquisition device (4K resolution, 30fps). Three sets of images are taken, and image analysis software is used to extract the physical position parameters of POEP1: horizontal offset, vertical offset, and angle with the lens optical axis. Analysis of the front image yields a horizontal offset of 3mm and a vertical offset of 0.8mm; analysis of the side image yields an angle with the lens optical axis of 0.3°. The preset POEP installation parameters are: horizontal / vertical offset ≤ 1mm, angle ≤ 0.1°. The actual offset and angle both exceed the adaptation range, with the horizontal offset showing the largest deviation (exceeding 2mm). Comparing the actual installation parameters with the preset parameters, the difference exceeds the adaptation threshold, indicating a POEP installation position deviation. The deviation is mainly reflected in the horizontal direction, requiring adjustment of the mechanical structure for correction.

[0084] Furthermore, the geometric correction parameters of the multi-channel projection are retrieved, and the difference between the actual projection angle of the projector and the preset angle is analyzed. If the difference causes the shape of the overlapping area to be deformed, it is determined to be a projection angle deviation of the projector.

[0085] In this embodiment of the invention, geometric correction parameters for multi-channel projection are retrieved from the control computer. These parameters include the projection angles (horizontal and vertical angles), lens focal length, and projection ratio of each projector. The preset projection angles of the left projector are 0° horizontally and 0° vertically. The actual projection angles of the left projector are measured using an angle measuring device: 0.6° horizontally and 0.1° vertically, differing from the preset horizontal angles by 0.6° and vertical angle by 0.1°. The impact of these angle differences on the overlapping area is analyzed: geometric projection calculations show that a 0.6° horizontal angle deviation causes shape deformation in the overlapping area (originally 768 pixels wide), shrinking the left edge by 25 pixels and stretching the right edge by 15 pixels. The overlapping area changes from a rectangle to a trapezoid, exceeding the allowable range (≤10 pixels). This is determined to be a projector projection angle deviation, with the horizontal angle deviation being the primary cause of the overlapping area's shape deformation. Adjustments to the projector angle or the range of the blending area need to be made to accommodate the deformation.

[0086] Furthermore, if the deviation is due to the projector's projection angle, adjusting the division range of the blending region and the grayscale gradient rules based on the geometric correction data of the multi-channel projection to adapt the blending effect to changes in the projection angle includes: The deformation of the overlapping area shape is calculated based on the projection angle deviation of the projector, including specific parameters of tensile deformation, compressive deformation or tilting deformation. In this embodiment of the invention, based on the horizontal angle deviation of the left projector of 0.6°, a geometric projection deformation calculation method is adopted. A three-dimensional coordinate system is established with the optical center of the projector lens as the origin. The projection screen is a z=8m plane (8m away from the lens). The shape deformation of the overlapping area (original projection coordinates 3072-3840, 0-2160, original width 768 pixels, height 2160 pixels) is calculated. The horizontal angular deviation causes a trapezoidal distortion in the projected image along the horizontal direction. Using trigonometric functions, the left edge (originally x=3072) shifts inward due to the angular deviation, with an offset of 8m × tan(0.6°) ≈ 8 × 0.01047 ≈ 0.0838m. The corresponding pixel offset is 0.0838m ÷ (screen width 6m ÷ 3840 pixels) ≈ 0.0838 ÷ 0.0015625 ≈ 53.6 pixels, rounded down to 54 pixels. This represents a compression distortion of 54 pixels on the left edge. The right edge (originally x=3840) shifts outward, with an offset of 8m × tan(0.6°) ≈ 0.0838m, also with a pixel offset of 54 pixels. This represents a stretching distortion of 54 pixels on the right edge. The final deformation parameters for the overlapping area were determined as follows: the shape changed from a rectangle to a trapezoid, the left edge was compressed by 54 pixels (originally 3072 → 3018), the right edge was stretched by 54 pixels (originally 3840 → 3894), and there was no deformation in the height direction (still 2160 pixels). The width of the overlapping area after deformation was 3894 - 3018 = 876 pixels, which is 108 pixels wider than the original width.

[0087] Furthermore, the pixel boundary coordinates of the fusion region are adjusted according to the deformation amount to make the fusion region completely match the actual overlapping region after deformation, thus avoiding bright or dark bands caused by the offset of the fusion range. In this embodiment of the invention, the pixel boundary coordinates of the fusion region are adjusted according to the deformation amount (54 pixels compressed on the left and 54 pixels stretched on the right). The original fusion region corresponds to the pixel range of POEP1 (1480, 0) - (1920, 1200). Based on the spatial correspondence between POEP and the projected image (1 POEP pixel = 2 projected pixels), the left side compression of 54 projected pixels corresponds to a POEP pixel compression of 27 pixels (54 ÷ 2), and the right side stretching of 54 projected pixels corresponds to a POEP pixel stretching of 27 pixels. The adjusted pixel boundary coordinates of the fusion region of POEP1 are as follows: the left boundary is adjusted from 1480 to 1480-27=1453, the right boundary is adjusted from 1920 to 1920+27=1947, and the height boundary remains unchanged at 0-1200. The new fusion region pixel range is (1453, 0) - (1947, 1200), corresponding to the projected coordinates of the overlapping region after deformation: 3018-3894, 0-2160. Verification via coordinate mapping: POEP1 pixel 1453 corresponds to projection coordinate 3018 (1453×2-48=3018-48, correction: the original correspondence is POEP pixel x corresponds to projection pixel 2x, so 1453×2=3006. Since the original overlapping area shifted 54 pixels to the left, an offset of 12 needs to be added, so the final value is 3006+12=3018). This ensures that the fused area and the actual overlapping area after deformation are completely matched, avoiding the appearance of dark bands (uncovered) on the left side and bright bands (repeated coverage) on the right side of the overlapping area due to the offset of the fused range.

[0088] Furthermore, for the shape of the fused area after deformation, the parameters of the grayscale gradient curve are optimized. The gradient transition length is extended in the stretching deformation area and shortened in the compression deformation area to ensure a uniform brightness transition rate. In this embodiment of the invention, the parameters of the grayscale gradient curve are optimized based on the shape of the deformed fusion region (trapezoidal, narrow on the left and wide on the right). The original gradient curve is a gamma curve (γ=0.8) with a gradient transition length of 440 POEP pixels. It is adjusted to a segmented gradient: the gradient transition length of the left compressed region (POEP1 pixels 1453-1480, length 27 pixels) is shortened to 20 pixels, and the gradient formula is G=255-(255-198)×(x / 20)^0.8 (x is the number of pixels from the left boundary). This ensures that the brightness transition rate of this region is consistent with the original rate (original rate = 57 grayscale value / 440 pixels ≈ 0.13 grayscale value / pixel, adjusted rate = 57 / 20 ≈ 2.85). Correction: The left area only needs to transition from the non-blended area grayscale 255 to the middle gradient area grayscale 245, with a transition grayscale difference of 10 and a rate of 10 / 27≈0.37. Therefore, the formula is adjusted to G=255-10×(x / 27)^0.8. The middle normal area (1480-1920, length 440 pixels) retains the original gradient parameters. The right stretched area (1920-1947, length 27 pixels) extends the gradient transition length to 34 pixels, with the gradient formula G=245-47×(x / 34)^0.8 (transitioning from 245 to 198, grayscale difference 47). Through segmented optimization, the brightness transition rate of the entire blended area is stabilized at 0.35-0.4 grayscale values / pixel, avoiding bright bands caused by slow transitions in the stretched area and dark bands caused by excessively fast transitions in the compressed area.

[0089] Furthermore, by combining the distance changes between each pixel in the fused area after deformation and the projector lens, the corresponding target grayscale value is corrected to compensate for the light intensity attenuation difference caused by the distance change. In this embodiment of the invention, the corresponding target grayscale value is corrected by combining the distance changes between each pixel in the fused area after deformation and the projector lens. Measurements using a laser rangefinder show that the distance between pixel 1453 (left edge) and the lens is 18.2m, and the distance between pixel 1947 (right edge) and the lens is 18.8m, a distance difference of 0.6m. The light intensity attenuation is inversely proportional to the square of the distance; therefore, the attenuation difference = (18.8m). 2 -18.2 2 ) / 18.2 2≈(353.44-331.24) / 331.24≈22.2 / 331.24≈6.7%. Based on the grayscale-transmittance mapping model, the original target transmittance of 75% corresponds to a grayscale of 198. The light intensity attenuation caused by distance is corrected: the left pixel 1453 is closer and has higher light intensity, so the target transmittance decreases by 6.7% to 68.3%, and the corresponding grayscale value is calculated by reverse calculation through the model. When T=68.3%, G≈185; the right pixel 1947 is farther and has lower light intensity, so the target transmittance increases by 6.7% to 81.7%, and the corresponding grayscale value is ≈210. Establish a distance-grayscale correction table and calculate the correction amount for each pixel in the fusion area according to the distance. For example, for pixel 1700 (distance 18.5m), the correction amount is (18.5-18.2) / 0.6×(210-185)=0.3 / 0.6×25=12.5, and the target grayscale value is 198+12.5≈210.5≈211. This compensates for the difference in light intensity attenuation caused by the change in distance and ensures that the light intensity is uniform throughout the fusion area.

[0090] Furthermore, the pixel data in the deformed area is interpolated to ensure the continuity of the grayscale gradient, and the division range and grayscale gradient rules of the fusion area are adjusted to avoid pixel-level brightness discontinuities, thereby generating adjusted fusion control image data.

[0091] In this embodiment of the invention, the pixel data of the deformed region (left side 1453-1480, right side 1920-1947) is interpolated using a bilinear interpolation algorithm: the gray values ​​of the missing pixels in the left region (e.g., no gradient data between 1453-1480) are calculated by interpolation using adjacent valid pixels (245 at 1480, 255 at 1453). For example, the gray value of pixel 1460 = 255 - (255-245) × (7 / 27) ≈ 255 - 2.59 ≈ 252.41 ≈ 252; the gray values ​​of the newly added pixels in the right region (1920-1947) are calculated by interpolation using 245 at 1920 and 198 at 1947. For example, the gray value of pixel 1930 = 245 - (245-198) × (10 / 27) ≈ 245 - 17.4 ≈ 227.6 ≈ 228. After interpolation, the continuity of grayscale values ​​of all pixels is checked, and the grayscale difference between adjacent pixels is ≤3, with no pixel-level brightness discontinuities. The fusion region is adjusted to range (1453,0)-(1947,1200), and the grayscale gradient rule is updated to a piecewise gamma curve. Adjusted fusion control image data is generated, with a data matrix size of 1920×1200 (the portion exceeding the POEP resolution is automatically truncated, retaining only valid pixels), ensuring that the data can be directly sent to POEP for loading and activation.

[0092] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0093] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A programmable projector image edge blending system, characterized in that, include: A system hardware architecture consisting of a programmable optical edge blending board (POEP), a fixed mechanical structure, and a control computer is constructed. Based on the system hardware architecture, multi-channel projection configuration parameters of the target projection scene are obtained, including the number of projectors, the projection coverage area, and the image overlap ratio of adjacent projectors. Based on the multi-channel projection configuration parameters of the target projection scene, determine the adaptation and installation parameters between each projector lens and POEP, and construct the spatial correspondence between POEP and projector; establish network communication connection with each POEP through the control computer, and generate the POEP basic control parameter set based on the hardware characteristic parameters of POEP; Based on the overlap ratio of adjacent projector images and the pixel array distribution of POEP, the fusion region and non-fusion region corresponding to each POEP are divided, and the grayscale gradient rules within the fusion region are defined to generate initial fusion control image data. The initial blending control image data is transmitted to the corresponding POEP via network communication. The POEP's hardware module receives and stores the data, and simultaneously controls each pixel of the POEP to adjust its transmittance according to the grayscale value of the initial blending control image data. Optical edge modulation is then applied to the image projected by the projector to generate the optically edge-modulated projected image. The edge blending effect of the projected image is evaluated through visual observation, and effect optimization feedback information is generated based on the edge blending effect. Based on the effect optimization feedback information, the grayscale gradient rules of the blending area and the pixel transmittance parameters of the POEP are adjusted by the control computer. The blending control image data is updated and re-sent to the POEP. The optical edge modulation and evaluation feedback are repeated until the projected image reaches the preset edge blending standard, thereby completing the programmable projector image edge blending.

2. The programmable projector image edge blending system according to claim 1, characterized in that, The multi-channel projection configuration parameters for obtaining the target projection scene include: Select a customized transmissive OLED or LCD screen as the POEP and determine the POEP's hardware parameters, including pixel resolution, grayscale display level, physical size ratio, interface type, and IP address allocation rules. Design fixed mechanical structures corresponding to different projector models. These fixed mechanical structures have position adjustment functions, which can realize the fine adjustment of the displacement of the POEP in the front-back, left-right, and up-down directions and the correction of the angle deflection. Configure the hardware performance parameters and software operating environment of the control computer, and install edge blending debugging software with multi-POEP collaborative control function. This edge blending debugging software supports the definition of blending parameters, image generation, data transmission and effect preview, thereby building a system hardware architecture consisting of a programmable optical edge blending board (POEP), a fixed mechanical structure and a control computer. Based on the corresponding interface within the system hardware architecture, obtain the spatial layout data and installation parameters corresponding to the target projection scene, including the size, installation position, curvature parameters of the projection screen, as well as the installation coordinates, projection angle, and lens parameters of each projector; Based on the spatial layout data of the projection screen and the installation parameters of each projector, the range and shape of the image overlap area of ​​adjacent projectors are calculated, and the number of projectors, the projection coverage area, and the image overlap ratio of adjacent projectors in the multi-channel projection configuration are determined. This completes the construction of the system hardware architecture and the collection of basic parameters to obtain the multi-channel projection configuration parameters of the target projection scene.

3. The programmable projector image edge blending system according to claim 1, characterized in that, The process of establishing network communication connections with each POEP via a control computer and generating a basic control parameter set for the POEP based on its hardware characteristic parameters includes: A temporary connection is established with each POEP via the standard video interface of the control computer, a full grayscale test signal is sent, and the actual transmittance data of each pixel of the POEP at different grayscale values ​​is collected. Based on the collected actual transmittance data, and combined with the analysis of gray-level response characteristics, the gray-level transmittance response curve corresponding to POEP is analyzed. At the same time, the transmittance deviation caused by the difference in gray-level response characteristics is compensated, thereby establishing a mapping relationship model between gray value and actual transmittance. Obtain the physical distribution data of the POEP pixel array, including pixel pitch and effective display area boundary coordinates. Combine this with the optical projection parameters of the projector lens to calculate the spatial correspondence between POEP pixels and projected image pixels. Based on the brightness requirements of the target projection scene and the output brightness parameters of the projector, determine the optimal transmittance range of POEP to avoid insufficient brightness due to excessive darkness or excessive brightness that cannot suppress black light leakage. By integrating the mapping relationship model between grayscale values ​​and actual transmittance, the spatial correspondence between POEP pixels and projected image pixels, and the optimal transmittance range, a set of basic control parameters for POEP, including pixel control parameters, communication protocol parameters, and data storage parameters, is generated.

4. The programmable projector image edge blending system according to claim 3, characterized in that, The model for establishing the mapping relationship between grayscale values ​​and actual transmittance includes: By controlling the computer to send gradient test images from the lowest gray level to the highest gray level to the POEP, and maintaining each gray level for a preset stabilization time, the pixel response of the POEP is ensured to be completely stable. By using an optical brightness measurement device, multiple uniformly distributed test points were selected within the effective display area of ​​the POEP, and the transmitted light brightness data of each test point at different gray levels were collected. Based on the transmitted light brightness data and the incident light brightness data projected onto the POEP by the projector, the actual transmittance of each test point at different gray levels is calculated, and the original gray-transmittance data set of each test point is established. Based on the gray-level response characteristics, the original gray-level transmittance data of all test points were fitted and analyzed to eliminate random errors, generate the overall average gray-level transmittance response curve of POEP, and identify the nonlinear segments in the average gray-level transmittance response curve. For the nonlinear segment in the average gray-transmittance response curve, a compensation function is established to compensate for the transmittance deviation caused by the difference in gray-level response characteristics. The output gray value is then corrected by the control software to ensure that the actual transmittance of POEP maintains a linear correspondence with the theoretical set value, thereby establishing a mapping relationship model between gray value and actual transmittance.

5. The programmable projector image edge blending system according to claim 4, characterized in that, The process of establishing a compensation function for the nonlinear segment in the average grayscale-transmittance response curve and correcting the output grayscale value through control software includes: The gray level range corresponding to POEP in the nonlinear segment is divided into multiple continuous sub-intervals. The nonlinear coefficient of the gray-transmittance response curve in each sub-interval is calculated, and the key sub-intervals with nonlinear coefficients higher than the preset threshold are identified. Based on the original data of the key sub-intervals, a local compensation function corresponding to each sub-interval is constructed by using polynomial fitting or piecewise linear interpolation to ensure a smooth transition of transmittance changes after compensation. Establish a parameter index table for the local compensation function, associate and store gray values ​​with the corresponding compensation coefficients, and send test gray values ​​containing compensation corrections to POEP via the control computer. Re-collect transmittance data for each test point to verify whether the compensation effect meets the preset requirements. If the compensated transmittance deviation exceeds the allowable range, adjust the parameters of the local compensation function or the fitting method, and repeat the verification and adjustment until a mapping relationship model between the grayscale value and the actual transmittance that meets the requirements is formed.

6. The programmable projector image edge blending system according to claim 1, characterized in that, The process of dividing the POEP into blended and non-blended regions based on the overlap ratio of adjacent projector images and the pixel array distribution of the POEP, and defining the grayscale gradient rules within the blended region, includes: Based on the obtained overlap ratio of adjacent projector images and combined with the pixel resolution of the projected image, the pixel coordinate range of the overlapping area in the projected image is calculated, and the pixel boundary of the fusion area corresponding to each POEP is determined. Based on the requirements of dark field fusion, a black light suppression target for the fusion area is set. Combined with the minimum transmittance parameter of POEP, the minimum gray value at the edge of the fusion area is determined to ensure that the amount of black light superposition in the overlapping area is lower than the threshold perceptible to the human eye. Based on the characteristics of human vision, we designed grayscale gradient curve types within the fusion area, including linear gradient, gamma curve gradient, or custom nonlinear gradient, so that the brightness transition in the overlapping area is smooth without obvious jumps. Based on the grayscale gradient curve, calculate the target grayscale value of each pixel within the fusion region, and set the pixels in the non-fusion region to the highest grayscale level to generate the pixel data matrix of the initial fusion control image. The pixel data matrix of the initial fusion control image is smoothed to eliminate abrupt grayscale changes between pixels and avoid jagged edges or mottled light and shadow on the projected image, thus forming the initial fusion control image data.

7. The programmable projector image edge blending system according to claim 1, characterized in that, The optical edge modulation of the image projected by the projector includes: Based on the pixel array spatial topology model and gray-level-transmittance mapping relationship model of POEP, a dynamic light modulation logic library is constructed. This logic library contains pixel transmittance control rules corresponding to different image edge types, including hard edges, soft edges and gradient edges. The image edge feature intelligent extraction algorithm captures the edge contour, gradient change and texture features of the projected image in real time, generates an image edge feature map, and performs global texture detail enhancement processing on the projected image based on the image edge feature map. The high-frequency edge information and low-frequency background information corresponding to the projected image are separated by dynamic threshold segmentation technology to generate multi-scale feature layered data. Skeleton extraction and connectivity analysis are performed on high-frequency edge information in multi-scale feature-layered data to construct an edge feature spatial topology network and clarify the branching relationships and continuity characteristics of the edges. Based on the edge feature spatial topology network, edge segments with correlation are aggregated into complete edge regions. At the same time, a background feature masking mechanism is used to delineate pure background regions that are not related to the edge regions. In addition, combined with the fusion objectives of the projection scene, including black light suppression in dark fields and smooth transition in bright fields, the modulation priority of the complete edge regions is sorted, and the core modulation edge regions and auxiliary modulation edge regions are marked. The topological information of the core modulation edge region, auxiliary modulation edge region and pure background region is mapped to the pixel array coordinate system of POEP to generate a localization result map containing region type and modulation priority. The edge region and non-edge region that need to be modulated are located based on the localization result map. The corresponding rules in the dynamic light modulation logic library are called according to the image edge feature map, and the gray values ​​of each pixel of POEP are dynamically adjusted to change the phase distribution of the projected light in the edge region, thereby realizing the directional modulation of the edge light field. A nonlinear optical enhancement mechanism is introduced to perform phase superposition and amplitude calibration on the modulated light in the edge region, enhance the light intensity transition difference between the edge and non-edge regions, and suppress the light field distortion in the non-edge region to ensure the sharpness of the main subject in the image. By integrating the directional modulated light in the edge region with the original projected light in the non-edge region, the light field interference noise during the modulation process is eliminated, resulting in a projected image with natural edge transitions and complete details after optical edge modulation.

8. The programmable projector image edge blending system according to claim 1, characterized in that, The process of optimizing feedback information, adjusting the grayscale gradient rules of the fusion region and the pixel transmittance parameters of the POEP through a control computer, updating the fusion control image data and re-sending it to the POEP, and repeatedly executing optical edge modulation and evaluation feedback until the projected image reaches the preset edge fusion standard includes: The evaluation indicators of edge blending effect of the projected image are obtained through visual observation, including the brightness uniformity of the overlapping area, the blurring degree of the blending boundary, and the brightness consistency between the non-overlapping area and the overlapping area. For areas where brightness uniformity is not up to standard, locate the corresponding POEP pixel range and analyze the reasons for the transmittance deviation in that area, including POEP installation position deviation, grayscale-transmittance mapping model error or projector projection angle deviation. If the POEP installation position is misaligned, the spatial correspondence between the POEP and the projector lens can be corrected by adjusting the displacement or angle parameters of the fixing mechanical structure, and the pixel control parameters of the area can be re-determined. If the error is due to the gray-level-transmittance mapping model, supplement the transmittance test data of the area, update the gray-level-transmittance mapping model, and adjust the target gray-level value of the corresponding pixel. If the projector's projection angle is off, combine the geometric correction data of the multi-channel projection to adjust the division range of the blending area and the grayscale gradient rules so that the blending effect adapts to the change in projection angle. Update the fusion control image data according to the adjusted parameters, send it to the corresponding POEP and load it to take effect, and perform visual observation and evaluation again until all evaluation indicators meet the preset edge fusion standards.

9. The programmable projector image edge blending system according to claim 8, characterized in that, For areas where brightness uniformity is substandard, the corresponding POEP pixel range is located, and the reasons for the transmittance deviation in these areas are analyzed, including: Using a brightness acquisition device, multiple sampling points are selected in the overlapping and non-overlapping areas of the projected image, and the actual brightness values ​​of each sampling point are collected to establish a brightness distribution data matrix. Calculate the standard deviation of the brightness values ​​of each sampling point in the overlapping area. If the standard deviation exceeds the preset threshold, it is determined that the brightness uniformity of the area is not up to standard, and the area with the largest standard deviation is identified as the key optimization area. Compare the brightness difference between adjacent sampling points in the overlapping and non-overlapping areas. If the brightness difference exceeds the acceptable range of human vision, the transition of the fusion boundary is deemed unqualified, and the corresponding POEP pixel column or row is recorded. The relationship between the target grayscale value and the actual transmittance of the POEP pixels corresponding to the key optimization area is analyzed. If the deviation between the actual transmittance and the theoretically calculated value exceeds the allowable range, it is determined to be an error in the grayscale-transmittance mapping model. The actual installation posture of POEP is captured by an image acquisition device and compared with the preset installation parameters. If the position deviation or angle deviation exceeds the adaptation range, it is determined to be a POEP installation position deviation. Retrieve the geometric correction parameters of the multi-channel projection and analyze the difference between the actual projection angle of the projector and the preset angle. If the difference causes the shape of the overlapping area to be deformed, it is determined to be a projection angle deviation of the projector.

10. The programmable projector image edge blending system according to claim 8, characterized in that, If the deviation is due to the projector's projection angle, the geometric correction data from the multi-channel projection is used to adjust the division range of the blending region and the grayscale gradient rules to make the blending effect adapt to changes in the projection angle, including: The deformation of the overlapping area shape is calculated based on the projection angle deviation of the projector, including specific parameters of tensile deformation, compressive deformation or tilting deformation. Adjust the pixel boundary coordinates of the fusion area according to the deformation amount to make the fusion area completely match the actual overlapping area after deformation, and avoid bright or dark bands caused by the offset of the fusion range. To optimize the grayscale gradient curve parameters for the shape of the fused region after deformation, the gradient transition length is extended in the stretched deformation region and shortened in the compressed deformation region to ensure a uniform brightness transition rate. By combining the distance changes between each pixel in the fused area after deformation and the projector lens, the corresponding target grayscale value is corrected to compensate for the light intensity attenuation difference caused by the distance change. Interpolation is performed on the pixel data of the deformed region to ensure the continuity of grayscale gradient, and the division range and grayscale gradient rules of the fusion region are adjusted to avoid pixel-level brightness discontinuities, thereby generating adjusted fusion control image data.

Images