Automatic focusing method, system, device and storage medium for projector

Abstract

The present invention relates to the field of projection focus technology, and discloses a method, system, device, and storage medium for automatic focusing of a projector. The method comprises: collecting multi-dimensional high-dimensional perception data of a distributed sampling position sequence covering the entire focusing range of the projector at multiple scales; performing adaptive wavelet packet transform and nonlinear principal component analysis on the high-dimensional perception data to obtain perception fusion features, and performing multi-level fuzzy clustering on the perception fusion features to obtain a potential optimal focus area; based on the information entropy corresponding to the potential optimal focus area, adaptively sampling the potential optimal focus area to obtain a focus data point set, and performing focus fitting on the focus data point set to obtain a fitted focus data point set; performing particle swarm optimization on the fitted focus data point set to obtain a focus point position; and performing focus adjustment on the focusing lens to obtain an automatic focus result of the projector. The present application improves the automatic focus accuracy of the projector under the influence of multiple projection factors.

Description

Technical Field

[0001] The present invention relates to the field of projection focusing technology, and in particular to an automatic focusing method, system, device and storage medium for a projector. Background Art

[0002] With the development of projection and information technology, projectors, as an important display device, are widely used in education, business, entertainment, and other fields. During the projection process, projection clarity is one of the key factors affecting the user's viewing experience. Achieving a clear projected image depends on precise focus. Therefore, improving the focus accuracy of the projector is a key link in improving the user's projection viewing experience.

[0003] Nowadays, projector focusing methods mainly include electric focusing and automatic focusing. The electric focusing method uses a motor to drive the lens movement, and the user can adjust it through a remote control or control panel, while the automatic focus analyzes the corresponding focusing parameters to achieve automatic adjustment. However, these focusing methods have problems such as balancing focusing speed and accuracy, insufficient environmental adaptability, and high system complexity under various factors such as different projection distances and different ambient light conditions. In other words, the automatic focusing effect of existing projectors is poor under the influence of various projection factors. Summary of the Invention

[0004] The main purpose of the present invention is to solve the problem that the existing projector has poor auto-focusing effect under the influence of various projection factors.

[0005] A first aspect of the present invention provides an autofocus method for a projector, which is applied to an autofocus system of a projector. The autofocus system of the projector includes a perception module, a drive module, and a control module. The method includes: collecting multi-dimensional and high-dimensional perception data of a distributed sampling position sequence covering the entire focusing range of the projector by the perception module; performing adaptive wavelet packet transform and nonlinear principal component analysis on the high-dimensional perception data by the control module to obtain perception fusion features, and performing multi-level fuzzy clustering on the perception fusion features to obtain a potential optimal focus area; based on the information entropy corresponding to the potential optimal focus area, adaptively sampling the potential optimal focus area to obtain a focus data point set, and using a preset focus curve model to perform focus fitting on the focus data point set to obtain a fitted focus data point set; performing particle swarm optimization on the fitted focus data point set to obtain a focus point position; and based on the focus point position, adjusting the focus of the focus lens by the drive module to obtain an autofocus result of the projector.

[0006] Optionally, in a first implementation method of the first aspect of the present invention, before collecting multi-scale multiple high-dimensional perception data of the non-uniformly distributed sampling position sequence covering the entire focusing range of the projector through the perception module, it also includes: obtaining historical focusing data of the projector, and based on a preset focusing time series, performing focusing performance periodicity and trend characteristic analysis on the historical focusing data to obtain a characteristic analysis result; performing adaptive particle swarm optimization of multiple focusing parameters on the characteristic analysis result to obtain scene path parameters of optimal spiral scanning, and performing path curve trajectory smoothing on the scene path parameters to obtain a focusing path scanning trajectory; performing multi-scale decomposition on the focusing path scanning trajectory to obtain focusing sampling point sets of different resolutions, and performing control sequence conversion of the stepping motor in the driving module on the focusing sampling point set to obtain a distributed sampling position sequence covering the entire focusing range of the projector.

[0007] Optionally, in a second implementation method of the first aspect of the present invention, the control module performs adaptive wavelet packet transform and nonlinear principal component analysis on the high-dimensional perception data to obtain perception fusion features, including: performing piecewise nonlinear transformation on the high-dimensional perception data by the control module to obtain normalized high-dimensional perception data, and performing wavelet multi-resolution analysis on the normalized high-dimensional perception data to obtain high-dimensional perception data in the time-frequency domain; performing time-frequency domain threshold denoising on the high-dimensional perception data in the time-frequency domain, and extracting high-order statistics of the denoised high-dimensional perception data to obtain multidimensional nonlinear perception features, and performing nonlinear dimensionality reduction on the nonlinear perception features to obtain nonlinear perception features after dimensionality reduction; performing fuzzy conversion of perception semantic features on the nonlinear perception features after dimensionality reduction to obtain a perception fuzzy feature set, and performing multidimensional perception feature fusion on the perception fuzzy feature set to obtain perception fusion features.

[0008] Optionally, in a third implementation manner of the first aspect of the present invention, performing multi-level fuzzy clustering on the perceptual fusion features to obtain a potential optimal focus area includes: performing density peak clustering on the perceptual fusion features to obtain an initial clustering result with an adaptive density threshold, and performing multi-scale kernel density estimation on the initial clustering result to obtain multiple candidate optimal focus areas; performing gradient calculation of multi-dimensional geometric features on each of the candidate optimal focus areas to obtain a surface focus change feature, and performing tensor field analysis on the surface focus change feature to obtain a focus key point and a focus ridge position; dynamically assigning a weight to each of the candidate optimal focus areas based on the focus key point and the focus ridge position to obtain a focus area weight, and performing fuzzy cognitive map analysis and region sorting on each of the candidate optimal focus areas based on the focus area weight to obtain a potential optimal focus area.

[0009] Optionally, in a fourth implementation manner of the first aspect of the present invention, the adaptive sampling of the potential optimal focusing area based on the information entropy corresponding to the potential optimal focusing area to obtain a focused data point set includes: adaptively gridding the potential optimal focusing area to obtain a non-uniformly distributed focused sampling point set, and locating each focused sampling point in the focused sampling point set to obtain a focused sampling point position; measuring the light intensity of the focused sampling point position to obtain focused sampling point light intensity data, and evaluating a feature evaluation value of each focused sampling point based on the focused sampling point light intensity data; and selecting a feature point set for each focused sampling point based on the feature evaluation value and a preset feature evaluation threshold to obtain a focused data point set.

[0010] Optionally, in a fifth implementation manner of the first aspect of the present invention, the focused data point set is focused and fitted using a preset focusing curve model to obtain a fitted focused data point set, including: using a preset focusing curve model to perform low-dimensional feature mapping on the focused data point set to obtain multiple low-dimensional point set features, and calculating the feature similarity of each of the low-dimensional point set features to construct a point set weighted graph between the focused data point sets; based on the point set weighted graph, determining the temporal relationship value between each of the low-dimensional point set features, and based on the temporal relationship value, performing feature fusion on each of the low-dimensional point set features to obtain multi-scale point set features that fuse different time scales; performing residual analysis and feature fitting on the multi-scale point set features to obtain a fitted focused data point set.

[0011] Optionally, in a sixth implementation manner of the first aspect of the present invention, the particle swarm optimization is performed on the fitted focus data point set to obtain the focus point position, including: quantum encoding and state evaluation of the fitted focus data point set to obtain multiple focus point candidate solutions, and evaluating the focus point suitability of each of the focus point candidate solutions; based on the focus point suitability, a global search of multiple rotation angles is performed on the fitted focus data point set to obtain a candidate focus set; and a local search of the Levy flight path and search point comparison are performed on the candidate focus set to obtain the focus point position.

[0012] A second aspect of the present invention provides an automatic focusing system for a projector, the automatic focusing system for the projector comprising: a perception module, configured to collect, through the perception module, multi-dimensional and high-dimensional perception data of a distributed sampling position sequence covering the entire focusing range of the projector at multiple scales; a control module, configured to perform adaptive wavelet packet transform and nonlinear principal component analysis on the high-dimensional perception data through the control module to obtain perception fusion features, and perform multi-level fuzzy clustering on the perception fusion features to obtain a potential optimal focusing area; based on the information entropy corresponding to the potential optimal focusing area, adaptively sample the potential optimal focusing area to obtain a focus data point set, and perform focus fitting on the focus data point set using a preset focus curve model to obtain a fitted focus data point set; perform particle swarm optimization on the fitted focus data point set to obtain a focus point position; and a driving module, configured to adjust the focus of the focusing lens through the driving module based on the focus point position to obtain an automatic focusing result of the projector.

[0013] A third aspect of the present invention provides an autofocus device for a projector, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the autofocus device of the projector performs the various steps of the above-mentioned autofocus method for the projector.

[0014] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the various steps of the above-mentioned projector auto-focusing method.

[0015] The above-mentioned projector autofocus method, system, device and storage medium. In an embodiment of the present invention, a perception module collects multi-dimensional high-dimensional perception data of a distributed sampling position sequence covering the entire focusing range of the projector at multiple scales; a control module performs adaptive wavelet packet transform and nonlinear principal component analysis on the high-dimensional perception data to obtain perception fusion features, and performs multi-level fuzzy clustering on the perception fusion features to obtain a potential optimal focus area; based on the information entropy corresponding to the potential optimal focus area, the potential optimal focus area is adaptively sampled to obtain a focus data point set, and a preset focus curve model is used to perform focus fitting on the focus data point set to obtain a fitted focus data point set; particle swarm optimization is performed on the fitted focus data point set to obtain a focus point position; based on the focus point position, the driving module adjusts the focus of the focus lens to obtain an autofocus result of the projector. Compared to the existing technology, this application collects multi-dimensional and high-dimensional perception data at multiple scales of the projector's distributed sampling position sequence, extracts and fuses the perception features of the multi-dimensional and high-dimensional perception data, and then uses the perception fusion features to find the potential optimal focus area, samples focus data points in the potential optimal focus area, and performs cluster fitting and particle swarm optimization on each focus data point to determine the current focus point position of the projector. Finally, the driving module is used to control the projector to achieve rapid focus preparation. This improves the projector's autofocus accuracy under the influence of various projection factors.

[0016] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or understood by practicing the present invention. The purposes and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0017] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 1 is a schematic diagram of a first embodiment of an automatic focusing method for a projector according to an embodiment of the present invention;

[0019] Figure 2 2 is a schematic diagram of a second embodiment of an automatic focusing method for a projector according to an embodiment of the present invention;

[0020] Figure 3 Schematic diagram of an embodiment of an auto-focus system for a projector according to an embodiment of the present invention;

[0021] Figure 4 Schematic diagram of an embodiment of an autofocus device for a projector according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0023] The terms "including," "having," and any variations thereof, as used in the embodiments of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or device comprising a series of steps or units is not limited to the listed steps or units, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to the process, method, product, or device.

[0024] To facilitate understanding of this embodiment, the specific process of the embodiment of the present invention is described below. Figure 1 A first embodiment of the projector auto-focus method according to the present invention includes:

[0025] 101. Collect multi-dimensional and high-dimensional perception data of the distributed sampling position sequence covering the entire focusing range of the projector through the perception module;

[0026] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, techniques, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results.

[0027] Fundamental AI technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0028] In this embodiment, the autofocus method of the projector is applied to the autofocus system of the projector, wherein the autofocus system of the projector includes a sensing module, a driving module and a control module. The sensing module includes a photodetector, an ambient light sensor, a spectral sensor, a temperature sensor and a high-resolution camera, etc., and the corresponding sensors and detection data can be increased or decreased according to the actual environment needs; the distributed sampling position sequence here refers to a series of position points selected by the system for data collection within the entire moving range of the focusing lens.

[0029] In practical applications, based on the distributed sampling position sequence covering the entire focusing range of the projector, the perception module uses relevant sensors to collect perception data from a series of position points. For example, a high-precision photodetector is used to measure the reflected laser light intensity (the reflected laser light intensity is a key indicator for determining whether the projector has reached the optimal focal length). The perception module uses ambient light sensors and spectral sensors to collect ambient spectral distribution data. The perception module monitors the driving current waveform of the stepper motor (because the change in its driving current when the stepper motor adjusts the focal length can reflect the motor's operating status and the precise focusing position. By real-time monitoring of the current waveform, the rotation of the motor can be precisely controlled to achieve subsequent precise focusing). The perception module uses temperature sensors to monitor and record the temperature of key components of the system (to ensure that the focusing operation is performed within the appropriate temperature range to avoid affecting the focusing accuracy due to excessively high or low temperatures). The perception module also uses a high-resolution camera to capture images of the projected light spot, thereby obtaining a multi-dimensional high-dimensional data set containing reflected laser light intensity, ambient spectral distribution, driving current waveform, system temperature field, and light spot morphology parameters.

[0030] 102. Perform adaptive wavelet packet transform and nonlinear principal component analysis on high-dimensional perception data through the control module to obtain perception fusion features, and perform multi-level fuzzy clustering on the perception fusion features to obtain potential optimal focusing areas;

[0031] In this embodiment, the control module here is used to process the data collected by the perception module, generate corresponding focusing control instructions, and transmit the value driving module to control the operation of the corresponding adjustment motor; the control module performs piecewise nonlinear transformation on the high-dimensional perception data to obtain normalized high-dimensional perception data, and performs wavelet multi-resolution analysis on the normalized high-dimensional perception data to obtain high-dimensional perception data in the time-frequency domain; performs time-frequency domain threshold denoising on the high-dimensional perception data in the time-frequency domain, and extracts high-order statistics of the denoised high-dimensional perception data to obtain multidimensional nonlinear perception features, and performs nonlinear dimensionality reduction on the nonlinear perception features to obtain nonlinear perception features after dimensionality reduction; performs fuzzy conversion of perception semantic features on the nonlinear perception features after dimensionality reduction to obtain perception fuzzy features. set, and fuse the multi-dimensional perceptual features of the perceptual fuzzy feature set to obtain the perceptual fusion feature; perform density peak clustering on the perceptual fusion feature to obtain the initial clustering result of the adaptive density threshold, and perform multi-scale kernel density estimation on the initial clustering result to obtain multiple candidate optimal focusing areas; perform gradient calculation of the multi-dimensional geometric features on each candidate optimal focusing area to obtain the surface focusing change feature, and perform tensor field analysis on the surface focusing change feature to obtain the focusing key points and focusing ridge line positions; based on the focusing key points and focusing ridge line positions, perform dynamic weight allocation on each candidate optimal focusing area to obtain the focusing area weight, and based on the focusing area weight, perform fuzzy cognitive map analysis and regional sorting on each candidate optimal focusing area to obtain the potential optimal focusing area.

[0032] In practical applications, the control module performs piecewise nonlinear transformation on the collected multi-dimensional high-dimensional perception data, that is, mapping high-dimensional perception data of different scales into a unified numerical range through piecewise functions. For example, the data of each dimension is mapped using the following nonlinear transformation formula:

[0033] ;

[0034] in, are the coefficients of the piecewise function, is the segmentation point, thus obtaining the normalized high-dimensional perception data, and inputting the normalized data into the wavelet transform for multi-resolution analysis. The multi-resolution analysis calculation formula is:

[0035] ;

[0036] Here, set is the coefficient after wavelet transformation, is the original signal, is the wavelet function, and Respectively represent the translation and scaling parameters in the time-frequency domain, thereby decomposing the normalized high-dimensional perception data into components on different time scales, and obtaining high-dimensional perception data in the time-frequency domain, so as to facilitate subsequent time-frequency domain analysis; then, the high-dimensional perception data in the time-frequency domain is subjected to threshold denoising, that is, by setting a threshold , retaining the wavelet coefficients above the threshold, and extracting the high-order statistics of the denoised high-dimensional perception data, to obtain multidimensional nonlinear perception features, such as skewness (used to measure the symmetry of the data distribution) and kurtosis. For example, the skewness calculation formula is used to extract the skewness of the denoised high-dimensional perception data, where the skewness calculation formula is:

[0037] ;

[0038] here, is the denoised data point, is the average value, and n is the number of samples. Based on principal component analysis, the variables that may be correlated in the nonlinear perceptual features are converted into a set of linearly uncorrelated variables through orthogonal transformation to obtain the nonlinear perceptual features after dimensionality reduction, so as to reduce the dimension of the data and extract the most informative perceptual features; and then the fuzzy transformation of the perceptual semantic features of the nonlinear perceptual features after dimensionality reduction is performed through the membership function. For example, for the feature , its membership It can be defined as:

[0039] ;

[0040] in, and It is the parameter of the membership function, which controls the shape of the membership curve, thereby converting the precise numerical value corresponding to the nonlinear perceptual feature after dimensionality reduction into the membership in the fuzzy set, and then fusing the multi-dimensional perceptual features of the perceptual fuzzy feature set through weighted averaging and other methods to integrate the information of multiple features, and converting the original high-dimensional perceptual data into a set of reduced-dimensional and fused perceptual features, and finally obtaining the perceptual fusion feature.

[0041] Based on the density distribution corresponding to the perceptual fusion features, the perceptual fusion features are clustered by density peaks. That is, the local density of each perceptual fusion feature data point is calculated by the clustering formula, the density peak point is identified, and the initial clustering result of the adaptive density threshold is obtained. The clustering formula is:

[0042] ;

[0043] in, Represents data points The local density of Therefore The set of data points in the neighborhood centered at is the neighborhood radius; and the density estimation formula is used to perform multi-scale kernel density estimation on the initial clustering results (i.e., using a Gaussian kernel to smooth the data points) to identify multiple candidate optimal focus areas. The density estimation formula is as follows:

[0044] ;

[0045] in, It's on point The estimated density at is the number of data points, is the bandwidth of the kernel; then, the gradient of the multi-dimensional geometric features is calculated for each candidate focus area to identify the characteristics of the focus change, obtain the surface focus change characteristics, and calculate the second-order tensor of the gradient for the surface focus change characteristics to identify the focus key points and the focus ridge position; then, based on the focus key points and the focus ridge position, the candidate best focus areas are dynamically weighted, and the weight allocation formula is as follows:

[0046] ;

[0047] in, It is a region The weight of is a parameter that controls the weight distribution, It is a region The gradient of the candidate focus area is calculated, and based on the focus area weight, the relationship network between the attributes of the candidate focus area is constructed to evaluate the quality of the candidate area. The fuzzy cognitive map analysis formula is as follows:

[0048] ;

[0049] in, It is a region The comprehensive evaluation value of Positive impact area The attribute set of Negative impact area The attribute set of and They are the positive and negative influence degrees between attributes, and the regional ranking of the candidate best focusing areas after fuzzy cognitive evaluation. Finally, the candidate best focusing area with the highest evaluation score is selected as the potential best focusing area.

[0050] 103. Based on the information entropy corresponding to the potential optimal focus area, adaptively sample the potential optimal focus area to obtain a focus data point set, and perform focus fitting on the focus data point set using a preset focus curve model to obtain a fitted focus data point set;

[0051] In this embodiment, the potential optimal focusing area is adaptively gridded to obtain a non-uniformly distributed focused sampling point set, and each focused sampling point in the focused sampling point set is located to obtain the focused sampling point position; the light intensity of the focused sampling point position is measured to obtain focused sampling point light intensity data, and based on the focused sampling point light intensity data, a feature evaluation value of each focused sampling point is evaluated; based on the feature evaluation value, a feature point set is selected for each focused sampling point based on a preset feature evaluation threshold to obtain a focused data point set.

[0052] In practical applications, first, the potential optimal focus area is adaptively meshed, that is, the potential optimal focus area is subdivided into smaller sampling units, so that the grid density is dynamically adjusted according to the importance and change rate of the area to achieve a non-uniformly distributed focus sampling point set. The adaptive meshing function is:

[0053] ;

[0054] in, is the grid density function, is a point in the region, Yes The light intensity value, is the average light intensity in the area, is the parameter that controls the non-uniform distribution of the grid; and the focus sampling point set obtained by division is positioned through coordinate transformation. The coordinate transformation formula is: , , among which, is the point in the original coordinate system, is the point in the coordinate system after grid division, and is the offset adjusted according to the grid density to ensure that each sampling point has accurate focus sampling point coordinates in the new grid; then the light intensity of each focus sampling point is measured by the photoelectric detector to obtain the focus sampling point light intensity data. for point The light intensity value at the location is, then the formula for light intensity measurement is The light intensity data of each sampling point is calculated, where k is the response coefficient of the photodetector. Based on the data obtained from the light intensity measurement, the characteristic evaluation value of each focused sampling point is evaluated using the local variance formula. The local variance formula is:

[0055] ;

[0056] in, for point The local variance at Yes The coordinates of the sampling points in the neighborhood, It is the degree of change of light intensity around the sampling points in the neighborhood, which is an indicator of feature evaluation, so as to evaluate the feature evaluation value of each focused sampling point. Then, based on the feature evaluation value, each focused sampling point is screened through a preset feature evaluation threshold, and each focused sampling point with a value greater than the preset feature evaluation threshold is extracted to obtain a focused data point set.

[0057] 104. Perform particle swarm optimization on the fitted focus data point set to obtain the focus point position;

[0058] In this embodiment, a preset focusing curve model is used to perform low-dimensional feature mapping on the focused data point set to obtain multiple low-dimensional point set features, and the feature similarity of each low-dimensional point set feature is calculated to construct a point set weighted graph between the focused data point sets; based on the point set weighted graph, the temporal relationship value between each low-dimensional point set feature is determined, and based on the temporal relationship value, the features of each low-dimensional point set are fused to obtain multi-scale point set features that fuse different time scales; residual analysis and feature fitting are performed on the multi-scale point set features to obtain the fitted focused data point set.

[0059] In practical applications, the focusing curve formula in the preset focusing curve model is used to perform low-dimensional feature mapping on the focused data point set, where the focusing curve formula is:

[0060] ;

[0061] in, It is principal components, It is The data point in The weight coefficients on the principal components, is the value of the original data point, n is the dimension of the data point, so that the feature mapping obtains multiple low-dimensional point set features, and the feature similarity between each low-dimensional point set feature is calculated by Euclidean distance or cosine similarity; then based on the feature similarity, a point set weighted graph between the focused data point sets is constructed, where the weighted graph is an adjacency matrix ,in Represents a point set Hedianji The similarity weight between them is constructed as follows:

[0062] ;

[0063] in, is a parameter that controls weight decay; based on the point set weighted graph, the temporal relationship value between each low-dimensional point set feature is determined by topological sorting of nodes in the network or by the shortest path algorithm, and based on the temporal relationship value, the features of each low-dimensional point set are fused by weighted averaging or fusion network. The feature fusion formula is:

[0064] ;

[0065] in, is the fused feature vector, is the weight determined according to the temporal relationship value, is the number of point set features; finally, the fused multi-scale point set features are subjected to residual analysis through curve fitting or regression, and the parameters of the scale point set features after residual analysis are adjusted by the least squares method to minimize the sum of squared residuals to obtain the fitted focused data point set. Then, the best fitting curve is generated based on the fitted focused data point set. This curve represents the ideal focusing path, and the residuals generated during the fitting process are used to evaluate and optimize the focusing strategy to ensure the accuracy and effectiveness of the focusing process.

[0066] 105. Based on the focus point position, the focus lens is adjusted by the driving module to obtain an automatic focus result of the projector.

[0067] In this embodiment, the driving module here is a focusing driving module such as a stepper motor in a projector; the driving module converts the focusing instruction into a control signal of the stepper motor or servo motor, accurately controls the rotation angle and speed of the motor, and accurately drives the focusing lens to a predetermined position.

[0068] In actual applications, the optimal focus position information calculated by the control module is transmitted to the driver module. After receiving the focus point position, the driver module converts the focus adjustment command into a control signal for a stepper motor or servo motor, precisely controlling the motor's rotation angle and speed to accurately drive the focus lens to the predetermined position. This movement is fine-tuned, aiming to precisely position the lens to the optimal focus position calculated by the control module. The lens movement must be both precise and fast to ensure that the projector can quickly respond and adjust to the optimal focus state, ultimately achieving the projector's autofocus result.

[0069] In an embodiment of the present invention, multidimensional and high-dimensional perception data is collected at multiple scales from a projector's distributed sampling position sequence, and perception features are extracted and fused from the multidimensional and high-dimensional perception data. The fused perception features are then used to identify a potential optimal focus area. Focus data points are sampled within the potential optimal focus area, and cluster fitting and particle swarm optimization are performed on each focus data point to determine the current focus point position of the projector. Finally, a driving module is used to control the projector to achieve rapid focus. This improves the projector's autofocus accuracy under the influence of various projection factors.

[0070] See also Figure 2 A second embodiment of the projector auto-focus method according to the present invention includes:

[0071] 201. Obtain historical focus data of the projector, and perform focus performance periodicity and trend characteristic analysis on the historical focus data based on a preset focus time series to obtain characteristic analysis results;

[0072] In this embodiment, the focus data of this type of projector over a period of time in the past is obtained, including but not limited to the position and time of each focus adjustment, environmental parameters (such as temperature and humidity), and user operation records. Then, the time series characteristics describing the focus operation are established through the autoregressive integral sliding average model in the time series analysis (for example, it is found that the frequency of focus operations increases significantly when large meetings are held at the beginning of each quarter). Then, the periodic pattern of the focus operation is identified through Fourier transform. At the same time, by calculating the sliding average of the time series or using a linear regression model to perform trend feature analysis, the key features of the focus performance based on the preset focus time series are extracted, such as periodic intensity, trend direction, rate of change, etc.

[0073] 202. Performing adaptive particle swarm optimization of multiple focusing parameters on the feature analysis results to obtain optimal spiral scanning scene path parameters, and performing path curve trajectory smoothing on the scene path parameters to obtain a focusing path scanning trajectory;

[0074] In this embodiment, an adaptive particle swarm optimization algorithm is used to search for the optimal spiral scanning path parameters based on the feature analysis results, and a PSO algorithm is used to search for the optimal focusing point through a spiral path from the inside to the outside or from the outside to the inside to optimize the spiral scanning path parameters. Thus, after continuously updating the position and velocity of the particles, the global optimal solution is approached to obtain the optimal spiral scanning scene path parameters, and then the scene path parameters are smoothed by polynomial fitting or spline interpolation, thereby eliminating noise and discontinuity in the path corresponding to the scene path parameters. Finally, a focusing path scanning trajectory is generated based on the smoothed path parameters.

[0075] 203. Perform multi-scale decomposition on the focus path scanning trajectory to obtain focus sampling point sets of different resolutions, and perform control sequence conversion of the stepping motor in the driving module on the focus sampling point sets to obtain a distributed sampling position sequence covering the full focus range of the projector;

[0076] In this embodiment, a wavelet transform is used to perform multi-scale decomposition on the generated focus path scanning trajectory to extract focus sampling point sets of different resolutions (i.e., identify key points on the focus path). These sets of focus sampling points are then converted into a stepper motor control sequence by calculating the difference between each sampling point and a reference point and multiplying it by a scaling factor. The number of stepper motor steps corresponding to each sampling point is then determined. All stepper motor control sequences are then integrated to produce a distributed sampling position sequence covering the projector's full focus range. This sequence reflects all the positions the focus lens must pass through during autofocus, ensuring the integrity and continuity of the focusing process.

[0077] 204. Collect multi-dimensional and high-dimensional perception data of multiple scales of a distributed sampling position sequence covering the entire focusing range of the projector through a perception module;

[0078] 205. Performing adaptive wavelet packet transform and nonlinear principal component analysis on high-dimensional perception data through the control module to obtain perception fusion features, and performing multi-level fuzzy clustering on the perception fusion features to obtain potential optimal focusing areas;

[0079] 206. Based on the information entropy corresponding to the potential optimal focus area, adaptively sample the potential optimal focus area to obtain a focus data point set, and perform focus fitting on the focus data point set using a preset focus curve model to obtain a fitted focus data point set;

[0080] 207. Perform particle swarm optimization on the fitted focus data point set to obtain the focus point position;

[0081] 208. Based on the focus point position, the driving module adjusts the focus of the focus lens to obtain an automatic focus result of the projector.

[0082] In an embodiment of the present invention, multidimensional and high-dimensional perception data is collected at multiple scales from a projector's distributed sampling position sequence, and perception features are extracted and fused from the multidimensional and high-dimensional perception data. The fused perception features are then used to identify a potential optimal focus area. Focus data points are sampled within the potential optimal focus area, and cluster fitting and particle swarm optimization are performed on each focus data point to determine the current focus point position of the projector. Finally, a driving module is used to control the projector to achieve rapid focus. This improves the projector's autofocus accuracy under the influence of various projection factors.

[0083] The above describes the automatic focusing method of the projector in the embodiment of the present invention. The following describes the automatic focusing system of the projector in the embodiment of the present invention. Figure 3 In one embodiment of the present invention, an automatic focusing system for a projector includes:

[0084] A perception module 301 is configured to collect, through the perception module, multi-dimensional and high-dimensional perception data of a distributed sampling position sequence covering the entire focusing range of the projector at multiple scales;

[0085] The control module 302 is configured to perform adaptive wavelet packet transform and nonlinear principal component analysis on the high-dimensional perception data through the control module to obtain perception fusion features, and perform multi-level fuzzy clustering on the perception fusion features to obtain a potential optimal focus area; based on the information entropy corresponding to the potential optimal focus area, adaptively sample the potential optimal focus area to obtain a focus data point set; perform focus fitting on the focus data point set using a preset focus curve model to obtain a fitted focus data point set; and perform particle swarm optimization on the fitted focus data point set to obtain a focus point position;

[0086] The driving module 303 is configured to adjust the focus of the focus lens through the driving module based on the focus point position to obtain an auto-focus result of the projector.

[0087] In an embodiment of the present invention, multidimensional and high-dimensional perception data is collected at multiple scales from a projector's distributed sampling position sequence, and perception features are extracted and fused from the multidimensional and high-dimensional perception data. The fused perception features are then used to identify a potential optimal focus area. Focus data points are sampled within the potential optimal focus area, and cluster fitting and particle swarm optimization are performed on each focus data point to determine the current focus point position of the projector. Finally, a driving module is used to control the projector to achieve rapid focus. This improves the projector's autofocus accuracy under the influence of various projection factors.

[0088] above Figure 3 The auto-focus system of the projector in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The auto-focus device of the projector in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0089] Figure 4Figure 4 is a schematic diagram of the structure of an autofocus device for a projector provided by an embodiment of the present invention. The autofocus device 400 for a projector may vary significantly depending on configuration or performance. It may include one or more central processing units (CPUs) 410 (e.g., one or more processors), memory 420, and one or more storage media 430 (e.g., one or more mass storage devices) storing applications 433 or data 432. The memory 420 and storage media 430 may be either transient or persistent storage. The program stored in the storage medium 430 may include one or more modules (not shown), each of which may include a series of instructions for operating the autofocus device 400 for the projector. Furthermore, the processor 410 may be configured to communicate with the storage medium 430 to execute the instructions stored in the storage medium 430 on the autofocus device 400 for the projector.

[0090] The projector autofocus device 400 may further include one or more power supplies 440, one or more wired or wireless network interfaces 450, one or more input and output interfaces 460, and / or one or more operating systems 431, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that Figure 4 The illustrated structure of the auto-focus device of the projector does not limit the auto-focus device of the projector and may include more or fewer components than shown in the figure, or may combine certain components, or arrange the components differently.

[0091] The present invention also provides an automatic focusing device for a projector, wherein the computer device includes a memory and a processor, wherein the memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor executes each step of the automatic focusing method for the projector in the above-mentioned embodiments.

[0092] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the various steps of the projector autofocus method.

[0093] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0094] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0095] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments in which tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0096] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A projector autofocus method, applied to a projector autofocus system, characterized in that: The auto-focus system of the projector includes a sensing module, a driving module and a control module, and the method includes: The perception module collects a variety of high-dimensional perception data of multiple scales of a distributed sampling position sequence covering the entire focusing range of the projector; Performing adaptive wavelet packet transform and nonlinear principal component analysis on the high-dimensional perception data through the control module to obtain perception fusion features, and performing multi-level fuzzy clustering on the perception fusion features to obtain a potential optimal focus area; Based on the information entropy corresponding to the potential optimal focus area, adaptively sampling the potential optimal focus area to obtain a focus data point set, and performing focus fitting on the focus data point set using a preset focus curve model to obtain a fitted focus data point set; performing particle swarm optimization on the fitted focus data point set to obtain a focus point position; Based on the focus point position, the focus lens is adjusted by the driving module to obtain an automatic focus result of the projector.

2. The automatic focusing method of a projector according to claim 1, wherein: Before collecting, by the perception module, a plurality of high-dimensional perception data of a non-uniformly distributed sampling position sequence covering the entire focusing range of the projector at multiple scales, the method further includes: Acquiring historical focus data of the projector, and performing focus performance periodicity and trend characteristic analysis on the historical focus data based on a preset focus time series to obtain a characteristic analysis result; Performing adaptive particle swarm optimization of multiple focusing parameters on the feature analysis results to obtain optimal spiral scanning scene path parameters, and performing path curve trajectory smoothing on the scene path parameters to obtain a focusing path scanning trajectory; The focus path scanning trajectory is decomposed at multiple scales to obtain focus sampling point sets of different resolutions, and the focus sampling point sets are converted into a control sequence of the stepping motor in the driving module to obtain a distributed sampling position sequence covering the full focus range of the projector.

3. The automatic focusing method of a projector according to claim 1, wherein: The step of performing adaptive wavelet packet transform and nonlinear principal component analysis on the high-dimensional perception data by the control module to obtain perception fusion features includes: The control module performs a piecewise nonlinear transformation on the high-dimensional perception data to obtain normalized high-dimensional perception data, and performs wavelet multi-resolution analysis on the normalized high-dimensional perception data to obtain high-dimensional perception data in the time-frequency domain; Performing time-frequency domain threshold denoising on the high-dimensional perception data in the time-frequency domain, extracting high-order statistics of the denoised high-dimensional perception data to obtain multidimensional nonlinear perception features, and performing nonlinear dimensionality reduction on the nonlinear perception features to obtain nonlinear perception features after dimensionality reduction; The nonlinear perceptual features after dimensionality reduction are subjected to fuzzy transformation of perceptual semantic features to obtain a perceptual fuzzy feature set, and the perceptual fuzzy feature set is subjected to multi-dimensional perceptual feature fusion to obtain a perceptual fusion feature.

4. The automatic focusing method of a projector according to claim 1, wherein: The multi-level fuzzy clustering is performed on the perception fusion features to obtain a potential optimal focus area, including: Performing density peak clustering on the perceptual fusion features to obtain an initial clustering result of an adaptive density threshold, and performing multi-scale kernel density estimation on the initial clustering result to obtain multiple candidate optimal focus areas; Performing gradient calculation of multi-dimensional geometric features on each candidate optimal focus area to obtain a surface focus change feature, and performing tensor field analysis on the surface focus change feature to obtain a focus key point and a focus ridge position; Based on the focus key points and the focus ridge positions, dynamic weight assignment is performed on each candidate optimal focus area to obtain a focus area weight. Based on the focus area weight, fuzzy cognitive map analysis and area sorting are performed on each candidate optimal focus area to obtain a potential optimal focus area.

5. The automatic focusing method of a projector according to claim 1, wherein: Adaptively sampling the potential optimal focus area based on the information entropy corresponding to the potential optimal focus area to obtain a focus data point set includes: Adaptively gridding the potential optimal focusing area to obtain a non-uniformly distributed focusing sampling point set, and locating each focusing sampling point in the focusing sampling point set to obtain a focusing sampling point position; Performing light intensity measurement on the focus sampling point position to obtain focus sampling point light intensity data, and evaluating a feature evaluation value of each focus sampling point based on the focus sampling point light intensity data; A feature point set is selected for each focused sampling point based on the feature evaluation value and a preset feature evaluation threshold to obtain a focused data point set.

6. The automatic focusing method of a projector according to claim 1, wherein: The focusing data point set is subjected to focus fitting by using a preset focusing curve model to obtain a fitted focusing data point set, including: Performing low-dimensional feature mapping on the focused data point set using a preset focusing curve model to obtain multiple low-dimensional point set features, and calculating the feature similarity of each of the low-dimensional point set features to construct a point set weighted graph between the focused data point sets; Based on the point set weighted graph, determining a temporal relationship value between each of the low-dimensional point set features, and based on the temporal relationship value, performing feature fusion on each of the low-dimensional point set features to obtain a multi-scale point set feature that fuses different time scales; Residual analysis and feature fitting are performed on the multi-scale point set features to obtain a fitted focused data point set.

7. The automatic focusing method of a projector according to claim 1, wherein: The performing particle swarm optimization on the fitted focus data point set to obtain the focus point position includes: performing quantum encoding and state evaluation on the fitted focus data point set to obtain a plurality of focus point candidate solutions, and evaluating the focus point applicability of each of the focus point candidate solutions; Based on the focus point applicability, a global search of multiple rotation angles is performed on the fitted focus data point set to obtain a candidate focus selection set; A local search of the Levy flight path and search point comparison are performed on the candidate focus selection set to obtain the focus point position.

8. An automatic focus system for a projector, characterized in that: The automatic focusing system of the projector comprises: A perception module, configured to collect, through the perception module, a variety of high-dimensional perception data of multiple scales of a distributed sampling position sequence covering the entire focusing range of the projector; a control module configured to perform adaptive wavelet packet transform and nonlinear principal component analysis on the high-dimensional perception data through the control module to obtain perception fusion features, and perform multi-level fuzzy clustering on the perception fusion features to obtain a potential optimal focus area; based on the information entropy corresponding to the potential optimal focus area, adaptively sample the potential optimal focus area to obtain a focus data point set; perform focus fitting on the focus data point set using a preset focus curve model to obtain a fitted focus data point set; and perform particle swarm optimization on the fitted focus data point set to obtain a focus point position; The driving module is used to adjust the focus of the focus lens based on the focus point position to obtain an automatic focus result of the projector.

9. An automatic focus device for a projector, characterized in that: The auto-focus device of the projector includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor calls the instructions in the memory to enable the auto-focus device of the projector to perform each step of the auto-focus method of the projector according to any one of claims 1 to 7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by a processor, the steps of the automatic focusing method for a projector according to any one of claims 1 to 7 are implemented.

Images