A human-computer interaction type digital ancient book display method and system based on aerial imaging

Abstract

The application discloses a man-machine interactive digital ancient book display method and system based on aerial imaging, relates to the technical field of man-machine interaction, and comprises the following steps: acquiring a user viewpoint position set and a page edge point position set of an aerial imaging area, calculating a space orientation projection between the page edge point and the user viewpoint, and generating a reflection orientation change set; identifying a boundary area causing a reflection path jump based on the reflection orientation change set, extracting a viewpoint angle corresponding to the boundary area, and screening a stable observation direction with the smallest reflection response change amplitude as a main reflection direction of the page; calculating an included angle projection value through the main reflection direction of the page and the viewpoint angle, and generating a page continuous focal plane group according to the included angle projection value; and the application realizes the space consistency of parallax correction and the continuous stability of focal depth change of the aerial imaging page in the process of real-time change of imaging distance, so that the imaging continuity and observation consistency of the digital ancient book in the interactive display are ensured.

Description

Technical Field

[0001] This invention relates to the field of human-computer interaction technology, specifically to a human-computer interactive digital ancient book display method and system based on aerial imaging. Background Technology

[0002] Aerial imaging technology, through precise design of the optical path, enables images to form visible images in a spatial location without a physical screen. It features unobstructed, non-contact, and immersive observation, and is gradually being applied to the three-dimensional presentation of digital ancient books in cultural display and human-computer interaction scenarios. This type of display method usually combines user viewpoint perception, spatial projection, and image reconstruction, allowing users to freely observe the pages of ancient books from different positions. At the same time, it places higher demands on page edges, imaging distance, and display clarity, thus making the issue of spatial consistency under multi-view dynamic observation conditions a fundamental prerequisite for the continued development of this type of application.

[0003] Currently, in the process of human-computer interactive digital ancient book display based on aerial imaging, the continuous movement of the user's viewpoint in the front-back, left-right and height directions will cause changes in the spatial orientation between the page edge and the observation direction. The reflection path exhibits non-linear characteristics with the change of viewpoint, and sudden changes in directional response are likely to occur at specific locations. This causes parallax shift and page position instability when the aerial image is switched between multiple perspectives, especially in the page edge area where imaging misalignment and inconsistency in observation are more likely to be exposed.

[0004] Secondly, under the condition of multi-view observation and synchronous change of imaging distance, the focal position of the aerial imaging page will continuously drift with the observation direction. The imaging optical path lengths corresponding to different viewpoints are different. If there is a lack of overall constraint on the law of change of depth of focus, local areas of the page are prone to sudden changes in depth of focus, uneven sharpness, or short-term defocusing, which will affect the user's recognition and reading continuity of the ancient book page content during continuous observation.

[0005] Furthermore, during the spatial reprojection and optical path inversion process of the aerial imaging page, when the imaging path crosses the reflection path change area or the effective imaging boundary of the page, the projection path may experience spatial interruption or invalid mapping points, resulting in local image loss or abrupt changes. Without dynamic correction and continuity constraints on the reprojection path, it will be difficult to maintain the overall continuity and stable presentation of the aerial image under the combined conditions of multi-view dynamic observation and real-time changes in imaging distance. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention provides a human-computer interactive digital ancient book display method based on aerial imaging, the method comprising:

[0007] S11, obtain the set of user viewpoint positions and the set of page edge point positions in the aerial imaging area, and calculate the spatial orientation projection between the page edge point and the user viewpoint to generate a set of reflection orientation changes;

[0008] S12, based on the set of reflection orientation changes, identify the boundary regions that cause the reflection path to jump, extract the viewpoint angles corresponding to the boundary regions, and select the stable observation direction with the smallest reflection response change as the main reflection direction of the page.

[0009] S13, calculate the included angle projection value by using the main reflection direction of the page and the viewpoint angle, generate a continuous focal plane group of the page based on the included angle projection value, and construct the focal depth estimation structure corresponding to the page area;

[0010] S14, calculate the focal position change direction value based on the continuous focal plane group and focal depth estimation structure of the page, and generate the focal depth dynamic adjustment parameter value through the focal position change direction value;

[0011] S15: Generate a spatial reprojection path for the ancient book page image based on the dynamic adjustment parameter value of the depth of focus, identify projection interruption events, and generate reprojection path correction data based on the projection interruption events to update the spatial reprojection path.

[0012] Furthermore, the steps for extracting the viewpoint angle corresponding to the boundary region and selecting the stable observation direction with the smallest change in reflection response as the main reflection direction of the page are as follows:

[0013] S121, Based on the set of reflection azimuth changes, perform direction difference calculation processing on each reflection azimuth change data in the set to generate direction difference data to characterize the strength of direction change;

[0014] S122, calculate the direction change amplitude value at each position based on the continuous direction difference change value, and mark the position where the direction change amplitude value is greater than the jump recognition threshold as a jump candidate position;

[0015] S123, perform continuous filtering on the location indexes that are not marked as boundary areas, and include the location indexes in the direction difference data that do not meet the jump recognition conditions into the continuous observation area set;

[0016] S124, based on the direction difference data corresponding to each position index in the continuous observation area set, calculate the local change amplitude value of the direction difference data, identify the direction vector with the smallest local change amplitude value, and take the direction vector as the main reflection direction of the page.

[0017] Furthermore, the logic for identifying the boundary region that triggers the reflection path transition based on the candidate transition position is as follows:

[0018] a1, based on the direction difference data, performs adjacent difference calculations according to the order of the user's viewpoints to generate continuous direction difference change values;

[0019] a2, based on the change value of the continuous directional difference, the positions where the change value of the continuous directional difference is greater than the predetermined jump recognition threshold are marked as jump candidate positions;

[0020] a3, based on the jump candidate position, performs index aggregation processing on adjacent jump candidate positions to generate jump position groups;

[0021] a4, based on the jump position group, maps the jump position group to the position index in the page space, and outputs the page area corresponding to the position index as the boundary area.

[0022] Furthermore, the steps for performing index aggregation processing on adjacent jump candidate positions are as follows:

[0023] a31, based on the interval between candidate jump positions, calculate the distance change value between adjacent candidate jump positions, and determine whether the distance between candidate positions is continuous based on the distance change value;

[0024] a32, based on the distance change value, merge the candidate positions of the jump that have a continuous spacing of less than the set spacing difference into the same jump position subgroup in order;

[0025] a33, calculate the internal index distribution difference for each jump position subgroup, and filter out candidate positions with abnormal deviations based on the distribution difference. Renumber the jump position subgroups after removing abnormal deviations and generate jump position groups.

[0026] Furthermore, a continuous focal plane group for the page is generated based on the included angle projection value, and a depth-of-focus estimation structure corresponding to the page region is constructed, including:

[0027] S131, Based on the main reflection direction of the page and the viewpoint angle, generate a filtered set of included angle projections;

[0028] S132, Perform segment grouping processing based on the included angle projection set, divide the included angle projection set into multiple included angle projection segments, and construct projection mapping relationship based on multiple included angle projection segments to generate a continuous focal plane group for the page;

[0029] S133, perform geometric optical path mapping processing based on the included angle projection value, calculate the viewing distance change of each included angle projection value on the virtual imaging plane, and use the viewing distance change as the relative focal length parameter value;

[0030] S134: Input the relative focal length parameter values ​​into the predetermined focal depth distribution function to generate the corresponding initial focal plane depth estimation value, and integrate the focal depth estimation structure corresponding to the page area based on all the initial focal plane depth estimation values.

[0031] Furthermore, the logic for generating the filtered set of included angle projections is as follows:

[0032] b1, set the origin of the spatial Cartesian coordinate system to be located at the center of the page, and extract the main reflection direction vector of the page;

[0033] b2, convert all viewpoint angles contained in the continuous observation area set generated in step S123 into spatial direction vectors, and uniformly map them to the spatial Cartesian coordinate system set in step b1;

[0034] b3 sequentially calculates the angle projection value between the main reflection direction vector of the page and the direction vector of each viewpoint, verifies whether all angle projection values ​​are within the preset imaging field of view, removes data points that exceed the range, and outputs the set of angle projections after filtering and projection calculation.

[0035] Furthermore, the steps for generating a continuous focal plane group for the page include:

[0036] S132.1 Based on the included angle projection set, extract adjacent included angle projection values ​​according to the arrangement order of each included angle projection value in the set, and calculate the numerical difference between adjacent included angle projection values ​​to generate an included angle projection difference set for group processing.

[0037] S132.2 Calculate the change in difference between consecutive angle projection differences, and identify the difference segments whose change in difference is continuously lower than a set change intensity threshold based on the change in difference, so as to generate a set of angle projection segments for segment grouping.

[0038] S132.3, Perform projection curvature fitting processing on each included angle projection segment, and generate a continuous focal plane group for the page based on the projection curvature fitting processing.

[0039] Furthermore, the steps for generating the dynamic adjustment parameter values ​​for the depth of focus are as follows:

[0040] S141, Based on the focal position data corresponding to each focal plane in the continuous focal plane group of the page, calculate the direction difference of the focal position as the included angle projection value changes, and record the direction difference as a set of focal position direction difference values;

[0041] S142, Based on the numerical change trend in the set of focal position direction difference values, identify the continuous rising segment and the continuous falling segment of the focal position change direction, and take the segment that meets the continuity condition as the candidate focal position change direction segment.

[0042] S143, Based on the candidate focus change direction segments, calculate the direction change amplitude value of each direction segment, and extract the dominant focus change direction according to the direction change amplitude value to generate the focus position change direction value;

[0043] S144: Input the focus position change direction value into the predetermined adjustment function to generate a dynamic focus adjustment parameter value for adjusting the focus stability of the page.

[0044] Furthermore, the steps for updating the spatial reprojection path are as follows:

[0045] S151, based on the included angle projection set, the continuous focal plane group of the page and the dynamic adjustment parameter value of the depth of focus, performs spatial mapping processing on each included angle projection value to generate an initial spatial reprojection path composed of multiple spatial mapping points;

[0046] S152, based on the spatial mapping point sequence of the initial spatial reprojection path, detect whether each mapping point has a spatial intersection with the boundary region, and mark the spatial mapping point that crosses the boundary region as a projection interruption event point;

[0047] S153, based on the included angle projection value corresponding to each projection interruption event point, perform nearest neighbor search in the continuous focal plane group of the page to determine the effective mapping interval corresponding to the projection interruption event point, and generate reprojection path correction data according to the effective mapping interval.

[0048] S154, the reprojection path correction data is matched with the corresponding spatial mapping points in the initial spatial reprojection path, and the spatial coordinates of the spatial mapping points are adjusted to generate the updated spatial reprojection path.

[0049] A human-computer interactive digital ancient book display system based on aerial imaging is provided to implement any of the aforementioned human-computer interactive digital ancient book display methods based on aerial imaging. The system includes:

[0050] The viewpoint acquisition module S21 acquires the set of user viewpoint positions and the set of page edge point positions in the aerial imaging area, calculates the spatial orientation projection between the page edge point and the user viewpoint, and generates a set of reflection orientation changes.

[0051] The path recognition module S22 identifies the boundary regions that cause the reflection path to jump based on the set of reflection orientation changes, extracts the viewpoint angle corresponding to the boundary regions, and selects the stable observation direction with the smallest reflection response change as the main reflection direction of the page.

[0052] The focal plane generation module S23 calculates the included angle projection value based on the main reflection direction of the page and the viewpoint angle, generates a continuous focal plane group of the page based on the included angle projection value, and constructs the focal depth estimation structure corresponding to the page area.

[0053] The focal depth derivation module S24 calculates the focal position change direction value based on the continuous focal plane group and focal depth estimation structure of the page, and generates the focal depth dynamic adjustment parameter value through the focal position change direction value;

[0054] The projection update module S25 generates a spatial reprojection path for the ancient book page image based on the dynamic adjustment parameter value of the depth of focus, identifies projection interruption events, and generates reprojection path correction data based on the projection interruption events to update the spatial reprojection path.

[0055] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0056] This invention constrains the reflection response of an aerial imaging page under different observation positions by using a set of user viewpoint positions, a set of page edge point positions, and a set of reflection orientation changes formed under multi-view dynamic observation conditions. This ensures that the reflection path remains continuous and controllable as the viewpoint changes, thereby effectively correcting the parallax shift of aerial images caused by viewpoint changes during the actual observation process of the user in front, behind, left, right, and at different heights. This avoids the problem of positional misalignment or unstable imaging in the page edge area due to changes in viewing angle.

[0057] Furthermore, this invention also filters stable observation directions within non-jumping areas by identifying boundary regions, continuous observation regions, and the main reflection direction of the page based on the set of reflection orientation changes. Combined with the set of included angle projections and the continuous focal plane group of the page, the focal position changes continuously and consistently as the imaging distance changes with the viewpoint angle. Thus, under the condition of real-time changes in imaging distance, the focal depth distribution of the page area is continuously adjusted to avoid sudden changes in focal depth or local defocusing caused by changes in the observation angle.

[0058] Furthermore, this invention dynamically constrains the spatial reprojection path based on the continuous focal plane group of the page, the depth-of-focus estimation structure, and the resulting dynamic adjustment parameter values ​​for the depth of focus. When a projection interruption event is detected, reprojection path correction data is generated, ensuring that the updated spatial reprojection path always remains consistent with the effective imaging area and reflection characteristics of the page. Thus, under the combined conditions of multi-view dynamic observation and real-time changes in imaging distance, a synergistic and stable effect between aerial image parallax correction and continuous depth-of-focus optimization is achieved, guaranteeing the spatial continuity and observation consistency of the digital ancient book page during interactive display.

[0059] In summary, this invention constrains and coordinates the changes in reflection orientation, the main reflection direction of the page, the continuous focal plane group of the page, and the depth-of-focus estimation structure under multi-view dynamic observation conditions. This allows the aerial imaging page to maintain spatial consistency of parallax correction and continuous stability of depth-of-focus changes simultaneously as the imaging distance changes in real time, thereby ensuring the imaging continuity and observation consistency of digital ancient books in interactive displays. Attached Figure Description

[0060] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0061] Figure 1 A flowchart of a human-computer interactive digital ancient book display method based on aerial imaging provided in Embodiment 1 of the present invention;

[0062] Figure 2 This is a module diagram of a human-computer interactive digital ancient book display system based on aerial imaging, provided in Embodiment 2 of the present invention. Detailed Implementation

[0063] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Example

[0064] Please see Figure 1 As shown in the figure, this embodiment discloses a human-computer interactive digital ancient book display method based on aerial imaging, the method including:

[0065] S11, obtain the set of user viewpoint positions and the set of page edge point positions in the aerial imaging area, and calculate the spatial orientation projection between the page edge point and the user viewpoint to generate a set of reflection orientation changes;

[0066] This step is used to establish the spatial geometric basis of the aerial imaging page under multi-view observation conditions. By generating a set of reflection orientation changes through the three-dimensional correspondence between the viewpoint position and the edge points of the page, it serves as the input basis for subsequent identification of reflection path jumps and extraction of the main reflection direction.

[0067] S111, Acquisition and structured recording of user viewpoint position;

[0068] In one specific embodiment, multiple sets of position sensing units, including but not limited to a TOF depth sensor array or a multi-point infrared ranging device, are set around the aerial imaging device to collect the three-dimensional position coordinates of the user's head when viewing the pages of the ancient book, and record them at a fixed sampling frequency to form a set of user viewpoint positions.

[0069] Each viewpoint in the user viewpoint position set contains three-dimensional coordinate values, represented as: viewpoint x-axis coordinate, viewpoint y-axis coordinate, and viewpoint z-axis coordinate;

[0070] It should be noted that the unit of each coordinate is millimeters. The viewpoint coordinates obtained from multiple consecutive sampling times are arranged in chronological order to characterize the user's actual observation trajectory within the aerial imaging area. This set is used to describe the user's dynamic observation position in the forward / backward, left / right, and height directions.

[0071] S112, Obtaining and geometrically calibrating the position of page edge points;

[0072] Four boundary curves, including the left, right, top, and bottom boundaries, are predefined inside the projection carrier of the aerial imaging device. Using the geometric calibration results fixed inside the projection device, each edge point is mapped to a unified three-dimensional spatial coordinate system.

[0073] In practical implementation, by calibrating the projection light path and imaging plane, the position coordinates of each page edge point in three-dimensional space can be obtained, including: left boundary point sequence, right boundary point sequence, upper boundary point sequence, and lower boundary point sequence;

[0074] It should be noted that each page edge point is recorded in three-dimensional coordinates as a set of page edge point positions, used to describe the boundary shape of static ancient book pages.

[0075] S113, Calculation of spatial orientation projection and generation of the set of reflection orientation changes;

[0076] Calculate the three-dimensional direction vector between the page edge point and the user's viewpoint, and perform normalization processing on the direction vector to obtain the spatial orientation vector representing the position of the viewpoint pointing to the page boundary;

[0077] Specifically, for the spatial coordinates of any page edge point Spatial coordinates of the user's viewpoint The three-dimensional direction vector d is calculated as follows:

[0078]

[0079] In the formula, Spatial coordinates and spatial coordinates The differences in x-axis coordinates, y-axis coordinates, and z-axis coordinates;

[0080] The length of the three-dimensional direction vector is normalized and converted into a unit vector u;

[0081] Represented as:

[0082]

[0083] In the formula, Let d be the length of the three-dimensional direction vector. The calculation formula is expressed as:

[0084]

[0085] It should be noted that the unit direction vector is used to represent viewing direction information and does not include the distance between the viewpoint and the page;

[0086] The unit direction vectors generated for the spatial coordinates of all user viewpoints and all page edge points are recorded according to the order of the user viewpoints, forming a spatial sequence in which the direction changes with the viewpoint. The change in direction is calculated one by one in this spatial sequence to obtain the set of reflection orientation changes.

[0087] In an exemplary scenario, when a user moves from left to right within the aerial imaging area, the unit direction vectors corresponding to each page edge point show a continuous changing trend as the viewpoint moves. If some directions change abruptly between adjacent viewpoints, it will be reflected in the set of reflection orientation changes as the position where the absolute value of the change amplitude exceeds the predetermined reflection change threshold.

[0088] The logic for generating the reflection change threshold is as follows:

[0089] The movement of the user's viewpoint relative to the page's imaging plane is usually continuous, and the change between unit direction vectors calculated from different viewpoints should also be kept within a continuous and smooth range.

[0090] Based on this, the reflection change threshold in this embodiment is obtained by statistical analysis of the directional change data of a user moving smoothly in the horizontal or vertical direction under unobstructed conditions:

[0091] Multiple continuous viewpoint positions are collected on a preset test path. The change amplitude of the unit direction vector between adjacent viewpoints is calculated. The higher confidence interval of the statistical distribution of the change amplitude, preferably the 90% or 95% quantile interval of the distribution, is used as the upper limit of normal continuous change to define the reflection change threshold for identifying jumps.

[0092] Under this threshold setting method, when the amplitude of change in a certain direction exceeds the statistical upper limit during the actual viewpoint movement, it can be determined that the reflection path corresponding to that position has undergone discontinuous change.

[0093] S12, based on the set of reflection orientation changes, identify the boundary regions that cause the reflection path to jump, extract the viewpoint angles corresponding to the boundary regions, and select the stable observation direction with the smallest reflection response change as the main reflection direction of the page.

[0094] In this step, the processing flow includes multiple processes such as direction difference calculation, jump candidate position identification, jump position aggregation, boundary region mapping, and continuity filtering of non-jump regions, to ensure that the main reflection direction of the page that describes the most stable viewing direction of the page can be obtained in the end.

[0095] Specifically, the steps for extracting the viewpoint angle corresponding to the boundary region and selecting the stable viewing direction with the smallest change in reflection response as the main reflection direction of the page are as follows:

[0096] S121, Based on the set of reflection azimuth changes, perform direction difference calculation processing on each reflection azimuth change data in the set to generate direction difference data to characterize the strength of direction change;

[0097] In one specific embodiment, the multiple sets of direction vectors recorded in the aforementioned set of reflection azimuth changes are processed one by one according to the viewpoint order, and the difference between each set of direction vectors is calculated using the cosine angle method to obtain the direction difference data.

[0098] Represented as:

[0099]

[0100] In the formula, For the first and the Directional difference between viewpoint positions For the first The unit direction vector corresponding to each viewpoint position For the first The unit direction vector corresponding to each viewpoint position These represent the magnitudes of the direction vectors; since they are unit vectors, their values ​​are both 1. It is the dot product of two direction vectors;

[0101] It should be noted that the orientation difference data is used to reflect the degree of deviation in orientation response when the viewpoint changes.

[0102] S122, calculate the direction change amplitude value at each position based on the continuous direction difference change value, and mark the position where the direction change amplitude value is greater than the jump recognition threshold as a jump candidate position, and identify the boundary region that causes the reflection path jump based on the jump candidate position;

[0103] Specifically, the logic for identifying the boundary region that triggers the reflection path transition based on the candidate transition position is as follows:

[0104] a1, based on the direction difference data, performs adjacent difference calculations according to the order of the user's viewpoints to generate continuous direction difference change values;

[0105] In practice, the trend of directional difference changes is obtained by performing adjacent difference operations on the directional difference data sequence, as shown below:

[0106]

[0107] In the formula, For the first and the The amount of change between the differences in each direction For the first and the The directional difference between viewpoint positions For the first The directional difference between the position of the first point and the position of the second point;

[0108] a2, based on the change value of the continuous directional difference, the positions where the change value of the continuous directional difference is greater than the predetermined jump recognition threshold are marked as jump candidate positions;

[0109] It should be noted that the logic for generating the predetermined transition recognition threshold is as follows:

[0110] Select a set of user viewpoint movement paths that do not need to cross page boundaries or reflective areas, make the user viewpoint move at a stable speed in the horizontal or vertical direction, record the positions of multiple consecutive viewpoints, calculate the direction difference between the unit direction vectors of every two adjacent viewpoints, and perform statistical analysis on all direction difference results.

[0111] Secondly, based on the statistical distribution of the direction difference results, the range of direction difference variation under normal continuous viewpoint movement conditions is identified, and the upper bound of its probability distribution high confidence interval is determined. Preferably, the upper bound of the high confidence interval is selected within the range of 90%–95%, and this upper bound is used as the jump recognition threshold.

[0112] a3, based on the jump candidate position, performs index aggregation processing on adjacent jump candidate positions to generate jump position groups;

[0113] In a specific implementation, multiple candidate positions for abrupt changes may appear in groups due to continuous changes. This step is used to group adjacent candidate positions that have common change characteristics into a group of abrupt change positions.

[0114] Specifically, the steps for performing index aggregation on adjacent jump candidate positions are as follows:

[0115] a31, based on the interval between candidate jump positions, calculate the distance change value between adjacent candidate jump positions, and determine whether the distance between candidate positions is continuous based on the distance change value;

[0116]

[0117] In the formula, For the nth jump candidate position and the nth jump candidate position The distance change value between the candidate jump positions For the first A jump candidate position This is the nth candidate position for the jump;

[0118] like If the difference is less than the predetermined jump interval, it is considered continuous;

[0119] It should be noted that the logic for generating the jump interval difference is as follows:

[0120] Record the viewpoint sequence of the user moving smoothly along a fixed direction under unobstructed conditions within the page imaging area, and calculate the direction difference data corresponding to each viewpoint. Statistically analyze the location index of local peaks in the direction difference data. Since the sampling frequency of the user's viewpoint is fixed, the index spacing between adjacent peaks will present a relatively stable range under normal continuous change conditions. Based on the statistical results of this stable range, determine the most frequently occurring index spacing as the reference spacing. Based on this reference spacing, set an allowable fluctuation range around it to define the jump spacing difference. Preferably, the fluctuation range needs to be less than one percent of the reference spacing.

[0121] a32, based on the distance change value, merge the candidate positions of the jump that have a continuous spacing of less than the set spacing difference into the same jump position subgroup in order;

[0122] a33, calculate the internal index distribution difference for each jump position subgroup, and filter out candidate positions with abnormal deviation points based on the distribution difference. Renumber the jump position subgroups after removing abnormal deviation points and generate jump position groups.

[0123] It should be noted that if there are isolated positions in a subgroup whose index distribution is inconsistent with the continuity of the group, they will be removed in this step.

[0124] a4, based on the jump position group, maps the jump position group to the position index in the page space, and outputs the page area corresponding to the position index as the boundary area.

[0125] It should be noted that: the jump in direction usually corresponds to the position where the reflection path in the page imaging area changes rapidly, by mapping the jump index to the page space.

[0126] S123, perform continuous filtering on the location indexes that are not marked as boundary areas, and include the location indexes in the direction difference data that do not meet the jump recognition conditions into the continuous observation area set;

[0127] In one specific embodiment, adjacency judgment processing is performed on each position index that has not entered the jump position group based on the index order of the direction difference data;

[0128] The index spacing between location indices is used as the basis for continuity, and a threshold for determining whether adjacent indices can be considered as continuous adjacent spacing is set.

[0129] It should be noted that the adjacent spacing threshold is derived from the distribution range of the index spacing statistics of continuous non-jump regions in the experiment. The spacing value falling within the stable interval is selected as the adjacent spacing threshold. The stable interval is the interval that has not jumped for N consecutive intervals. Preferably, N is an integer greater than or equal to 2.

[0130] Location indices that meet the adjacent spacing threshold are sequentially merged into continuous index segments; location indices that do not meet the adjacent spacing threshold are used as the starting point of a new continuous index segment, and all continuous index segments are summarized to form a continuous observation area set.

[0131] S124, based on the direction difference data corresponding to each position index in the continuous observation area set, calculate the local change amplitude value of the direction difference data, identify the direction vector with the smallest local change amplitude value, and take the direction vector as the main reflection direction of the page;

[0132] In one specific embodiment, each location index in the continuous observation area set is selected, and the direction difference data of the location index is calculated to be different from the direction difference data of its adjacent indexes before and after it to form a local difference amount;

[0133] It should be noted that the local difference is used to characterize the degree of local variation of a location within a continuous region.

[0134] The local change amplitude value is calculated based on the local difference value. Specifically, the absolute difference of the difference data in adjacent directions is used as the local change amplitude value.

[0135] To ensure that the fluctuation of directional difference within different continuous regions does not affect the filtering effect of the main reflection direction, this step introduces a local amplitude filtering rule after calculating the local change amplitude value. That is, among the local change amplitude values ​​of all position indices in the continuous observation area set, the position index with the smallest amplitude value is found, and the direction vector corresponding to the position index is output as the main reflection direction of the page.

[0136] For example, when a user moves from left to right and maintains a steady observation, certain positions within a continuous area without abrupt changes often exhibit minimal changes in their directional difference data. These directional vectors can be considered the most stable observation direction for the page's optical path. This step involves selecting this directional vector through a process of calculating the local change amplitude.

[0137] S13, calculate the included angle projection value by using the main reflection direction of the page and the viewpoint angle, generate a continuous focal plane group of the page based on the included angle projection value, and construct the focal depth estimation structure corresponding to the page area;

[0138] Specifically, a continuous focal plane group for the page is generated based on the included angle projection value, including:

[0139] S131, Based on the main reflection direction of the page and the viewpoint angle, generate a filtered set of included angle projections;

[0140] Specifically, the logic for generating the filtered set of included angle projections is as follows:

[0141] b1, set the origin of the spatial Cartesian coordinate system to be located at the center of the page, and extract the main reflection direction vector of the page;

[0142] Let the main reflection direction vector of the page be represented as:

[0143]

[0144] In the formula, represents the components of the reflection direction vector on the three coordinate axes;

[0145] It should be noted that the main reflection direction vector of the page has already been normalized in the aforementioned steps, satisfying the following conditions:

[0146] .

[0147] b2, convert all viewpoint angles contained in the continuous observation area set generated in step S123 into spatial direction vectors, and uniformly map them to the spatial Cartesian coordinate system set in step b1;

[0148] The continuous observation area set output in step S123 contains multiple viewpoint angles. Let any one of these viewpoint angles be:

[0149]

[0150] In the formula, The angle of the k-th viewpoint in a set of continuously observed regions. Let the viewpoint direction vector of the k-th viewpoint angle be mapped to the Cartesian coordinate system of step b1;

[0151] To ensure that the direction vectors of each viewpoint are unit vectors, normalization is performed.

[0152] Represented as:

[0153]

[0154] In the formula, Let be the unit vector of the viewpoint direction vector at the k-th viewpoint angle. The viewpoint direction vector of the k-th viewpoint angle The modulus length is calculated using the same formula as in step S113.

[0155] b3, calculate the angle projection value between the main reflection direction vector of the page and the direction vector of each viewpoint in sequence, check whether all angle projection values ​​are within the preset imaging field of view, remove data points that exceed the range, and output the set of angle projections after filtering and projection calculation;

[0156] In one specific embodiment, based on the page's main reflection direction vector The unit vector of the viewpoint direction vector Generate the included angle projection value, represented as:

[0157]

[0158] In the formula, Let be the angle projection value of the k-th viewpoint. It is the dot product of two vectors;

[0159] Let the imaging field of view threshold be The field of view threshold is derived from the actual half-angle of the field of view of the device's optical structure and is obtained through statistical measurement of the device's imaging optical path.

[0160] Set the filter criteria as follows:

[0161]

[0162] All that meet the above conditions The filtered set of included angle projections is represented as:

[0163]

[0164] In the formula, It is the set of angle projections.

[0165] S132, Perform segment grouping processing based on the included angle projection set, divide the included angle projection set into multiple included angle projection segments, and construct projection mapping relationship based on multiple included angle projection segments to generate a continuous focal plane group for the page;

[0166] Specifically, the steps for generating a continuous focal plane group for a page include:

[0167] S132.1 Based on the included angle projection set, extract adjacent included angle projection values ​​according to the arrangement order of each included angle projection value in the set, and calculate the numerical difference between adjacent included angle projection values ​​to generate an included angle projection difference set for group processing.

[0168] Let the set of included angle projections after filtering by S131 be represented as:

[0169]

[0170] The q-th angle projection value arranged in viewpoint order This represents the total number of angle projection values ​​within the set.

[0171] The formula for calculating the numerical difference between adjacent included angle projection values ​​is as follows:

[0172]

[0173] The difference between adjacent included angle projection values ​​is used to characterize the strength of the change in included angle projection. Arranged in order of viewpoint Each included angle projection value;

[0174] All The set of projection differences of the included angle is represented as:

[0175]

[0176] In the formula, This is the set of projection differences of the included angle;

[0177] It should be noted that the set of angle projection differences is used for subsequent segment change intensity judgment and segment grouping.

[0178] S132.2 Calculate the change in difference between consecutive angular projection differences, and identify the difference segments whose changes in difference are continuously lower than a set change intensity threshold based on the change in difference, so as to generate a set of angular projection segments for segment grouping; it should be noted that this step is used to determine the stationary and non-stationary segments when the angular projection value changes with the viewpoint, so as to determine which segments are suitable for constructing a continuous focal plane group.

[0179] Let the change in the difference between the projection differences of consecutive included angles be:

[0180]

[0181] In the formula, This represents the change in the difference between the projections of the qth included angle;

[0182] Let the threshold of change intensity be The filtering logic for identifying difference segments where the difference change is continuously lower than a set change intensity threshold based on the amount of difference change is as follows:

[0183]

[0184] It should be noted that: by collecting multiple sets of viewpoint movement data under conditions of no reflected optical path and no obstruction in the imaging area of ​​the device, the statistical distribution of the change in the continuous angle projection difference is calculated, and the upper limit of the 93%–95% confidence interval is preferred as the change intensity threshold.

[0185] Continuous angle projection difference segments that meet the above conditions are grouped into the same segment to generate an angle projection segment set.

[0186] Represented as:

[0187]

[0188] In the formula, For the m-th angle projection segment, Let M be the set of angled projection segments, and M be the number of angled projection segments.

[0189] S132.3, Perform projection curvature fitting processing on each included angle projection segment, and generate a continuous focal plane group of the page based on the projection curvature fitting processing;

[0190] Projection section of included angle All included angle projection values ​​are subjected to curvature fitting, and the fitting function is expressed as:

[0191]

[0192] In the formula, For the included angle projection segment The projection curvature fitting function, These are the fitting coefficients obtained by fitting using the least squares method. The input is the included angle projection value;

[0193] It should be noted that when the user's viewpoint moves along a continuous trajectory, the change in the focal plane usually exhibits second-order continuity. Using quadratic curve fitting can simultaneously express the shift trend and rate of change of the focal position.

[0194] After obtaining the fitting function for each segment, the continuous focal plane group of the page is represented as:

[0195]

[0196] This is a continuous focal plane group for the page.

[0197] Specifically, the steps for constructing the depth-of-focus estimation structure corresponding to the page area are as follows:

[0198] S133, perform geometric optical path mapping processing based on the included angle projection value, calculate the viewing distance change of each included angle projection value on the virtual imaging plane, and use the viewing distance change as the relative focal length parameter value;

[0199] The formula for calculating the change in sight distance is expressed as:

[0200]

[0201] In the formula, This represents the change in viewing distance corresponding to the q-th angle projection value. Given a virtual imaging plane reference distance, This is the projection value of the qth included angle;

[0202] It should be noted that: the change in viewing distance reflects the change in the optical path length corresponding to the included angle projection value, and is the core quantity used in the geometric optical path model to describe the trend of focal length change; the virtual imaging plane reference distance These are engineering parameters obtained by measuring the fixed imaging position of the projection optical path during the factory optical calibration process of the aerial imaging system.

[0203] The total changes in viewing distance obtained will be input as the relative focal length parameter value in step S134.

[0204] S134, input the relative focal length parameter values ​​into the predetermined focal depth distribution function to generate the corresponding initial focal plane depth estimation value, and integrate the focal depth estimation structure corresponding to the page area based on all the initial focal plane depth estimation values.

[0205] In one specific embodiment, based on all the obtained relative focal length parameter values ;

[0206] These relative focal length parameter values ​​are input into a predetermined focal depth distribution function to generate an initial focal plane depth estimate.

[0207] Represented as:

[0208]

[0209] In the formula, This is the initial focal plane depth estimate corresponding to the q-th relative focal length parameter value. , and The fitting parameters for the depth-of-focus distribution function are obtained by statistical modeling of the optical path samples of the imaging system.

[0210] It should be noted that the depth of focus distribution function is used to characterize the change of depth of focus of the imaging optical path under different viewpoint directions. It is a function model obtained based on the statistical analysis of actual optical path observation samples. The estimated depth of each initial focal plane is affected by the relative focal length parameter value.

[0211] All initial focal plane depth estimates are integrated in index order to construct the focal depth estimation structure corresponding to the page area.

[0212] Represented as:

[0213]

[0214] In the formula, This is the depth-of-focus estimation structure for the corresponding page area.

[0215] S14, calculate the focal position change direction value based on the continuous focal plane group and focal depth estimation structure of the page, and generate the focal depth dynamic adjustment parameter value through the focal position change direction value;

[0216] This step is used to analyze the directional characteristics of the focal position as a function of the included angle projection value, based on the constructed continuous focal plane group and depth-of-focus estimation structure of the page. By calculating the directional difference and the magnitude of the directional change, the dominant direction in the focal position change process is identified, and then dynamic adjustment parameter values ​​for the depth of focus are generated, providing a data basis for the dynamic control of the page's depth of focus stability.

[0217] Specifically, the steps for generating dynamic focus depth adjustment parameter values ​​are as follows:

[0218] S141, Based on the focal position data corresponding to each focal plane in the continuous focal plane group of the page, calculate the direction difference of the focal position as the included angle projection value changes, and record the direction difference as a set of focal position direction difference values;

[0219] Let the focal position data corresponding to the k-th focal plane in the continuous focal plane group on the page be:

[0220]

[0221] Calculation and acquisition and the focal position of the next focal plane} The direction difference is expressed as:

[0222]

[0223] In the formula, The difference in direction between the kth focal positions, " represents the vector dot product symbol;

[0224] It should be noted that the set of focal position direction differences consists of all Build in index order.

[0225] S142, Based on the numerical change trend in the set of focal position direction difference values, identify the continuous rising segment and the continuous falling segment of the focal position change direction, and take the segment that meets the continuity condition as the candidate focal position change direction segment.

[0226] After performing adjacent difference calculations on the set of focal position orientation differences generated in step S141, divide the results by the predetermined focal difference standard value. The change in direction difference is obtained. ;

[0227] Represented as:

[0228]

[0229] It should be noted that the logic for generating cross-standard values ​​is as follows:

[0230] Under unobstructed conditions, the user's viewpoint moves along a preset straight observation path at a stable speed. Multiple sets of continuous focal position data are acquired through the viewpoint acquisition method in step S11, and the corresponding focal position direction difference is calculated. Adjacent difference calculations are performed on the direction difference in all standard viewpoint movement data, and all calculation results are combined into a set of direction difference change benchmarks. Statistical analysis is performed on the set of direction difference change benchmarks to construct its numerical distribution, and the mean of the 95% confidence interval is selected as the standard value of focal difference.

[0231] Based on the change in direction difference The logic for identifying consecutive rising and falling segments is as follows:

[0232] Obtain the change in direction difference Below the continuity threshold of directional change The quantity, if the change in direction difference Below the continuity threshold of directional change If there are two or more segments, the segment is marked as satisfying the continuity condition;

[0233] Segments that meet the continuity condition are included in the candidate focal change direction segment set;

[0234] Continuity threshold of directional change The threshold is: the threshold for continuity of directional change. The selection process is based on the continuity of historical directional changes. Specifically, the change in stable directional difference is obtained through stable segments, and the average of all stable directional difference changes is calculated as the final threshold for the continuity of directional changes. .

[0235] S143, Based on the candidate focus change direction segments, calculate the direction change amplitude value of each direction segment, and extract the dominant focus change direction according to the direction change amplitude value to generate the focus position change direction value;

[0236] For each candidate focus change direction segment, the magnitude of the direction change is calculated to identify the dominant direction of the page focus change.

[0237] For each candidate focus change direction segment, the magnitude of the direction change is calculated to identify the dominant direction of the page focus change.

[0238] The formula for calculating the magnitude of directional change is as follows:

[0239]

[0240] In the formula, For the first The magnitude of the directional change of the segment;

[0241] The rules for selecting the dominant direction are:

[0242]

[0243] This can be understood as follows: Find the minimum value among all directional change amplitude values, and regard the corresponding directional segment as the dominant focal change directional segment; within this directional segment, select the directional vector corresponding to the index position with the smallest directional change amplitude as the focal position change directional value, denoted as... .

[0244] S144, input the focus position change direction value to the predetermined adjustment function to generate a dynamic focus adjustment parameter value for adjusting the focus stability of the page;

[0245] The extracted focus position change direction value is input into the predetermined adjustment function, which is represented as:

[0246]

[0247] In the formula, This is the value of the dynamic adjustment parameter for depth of focus. This represents the direction of change in the focal position. It is a regulation function;

[0248] It should be noted that the adjustment function is based on the factory calibration samples of the aerial imaging device and is determined by statistical fitting of the relationship between different viewpoint directions and changes in depth of focus.

[0249] S15, generate a spatial reprojection path for the ancient book page image based on the dynamic adjustment parameter value of the depth of focus, identify projection interruption events, generate reprojection path correction data based on the projection interruption events, and update the spatial reprojection path.

[0250] This step is used to perform a three-dimensional optical path inversion process on the ancient book page image under dynamic depth of focus adjustment. A complete reprojection path is constructed through spatial mapping relationship. When a projection interruption event caused by the optical path crossing the boundary area is detected, the path is corrected to ensure that the final spatial reprojection path is continuous, effective and can correctly correspond to the actual geometric area of ​​the page.

[0251] Specifically, the steps for updating the spatial reprojection path are as follows:

[0252] S151, based on the included angle projection set, the continuous focal plane group of the page and the dynamic adjustment parameter value of the depth of focus, performs spatial mapping processing on each included angle projection value to generate an initial spatial reprojection path composed of multiple spatial mapping points;

[0253] The initial spatial reprojection path generation logic includes:

[0254] Projection set of included angles Based on any included angle projection value, the depth of field is estimated according to the output of step S134 above. Select the focal plane depth estimate that matches the projection value index of the included angle. , represented as:

[0255]

[0256] in, For the first Estimated focal depth at each focal point;

[0257] Let the set of dynamic adjustment parameter values ​​for focal depth be:

[0258] It should be noted that: The generation logic of each sub-data is as follows in step S144. same;

[0259] For each focal plane depth, an adjusted depth-of-focus value is generated; expressed as:

[0260]

[0261] In the formula, This is the adjusted depth of focus value.

[0262] For ease of understanding, let the main reflection direction vector of the page be:

[0263]

[0264] Let the unit vector of the viewpoint direction vector corresponding to the included angle projection value be:

[0265]

[0266] The formula for calculating the spatial mapping point corresponding to the included angle projection value is:

[0267]

[0268] In the formula, For the first A spatial mapping point, These are the spatial coordinates of the page center point, determined by the system.

[0269] Map all spatial points The paths are combined sequentially to form the initial spatial reprojection path; represented as:

[0270]

[0271] This refers to the initial 3D reprojection path without boundary correction.

[0272] S152, based on the spatial mapping point sequence of the initial spatial reprojection path, detect whether each mapping point has a spatial intersection with the boundary region, and mark the spatial mapping point that crosses the boundary region as a projection interruption event point;

[0273] Let the set of boundary regions output by step S12 be represented as:

[0274]

[0275] In the formula, For the first page A region where the reflection path is discontinuous.

[0276] Let the effective imaging area of ​​the page be If the following occurs:

[0277]

[0278] This is considered a projection interruption event.

[0279] Count the projection interruption event points to construct a projection interruption event point set, represented as:

[0280]

[0281] In the formula, For the set of projection interruption event points, For the first One projection interruption event point.

[0282] S153, based on the included angle projection value corresponding to each projection interruption event point, perform nearest neighbor search in the continuous focal plane group of the page to determine the effective mapping interval corresponding to the projection interruption event point, and generate reprojection path correction data according to the effective mapping interval.

[0283] In one specific embodiment, obtain Corresponding angle projection value ,in, , The total number of included angle projection values;

[0284] Call the focal plane function:

[0285] Using the minimum curvature offset as the nearest neighbor condition, it can be expressed as:

[0286]

[0287] In the formula, For the first Depth prediction of the focal plane segment. For the first The adjusted depth of focus value of the interrupt event corresponding to each included angle projection value; This represents the focal plane segment index that is closest to the adjusted depth of focus value corresponding to the projection interruption event point, which is the effective mapping interval corresponding to the projection interruption event point.

[0288] according to The corrected depth value is generated and represented as:

[0289]

[0290] In the formula, To correct the depth value, For the focal plane segment The corresponding focal plane function;

[0291] Let the original mapping point corresponding to the interruption event point of the h-th angle projection value be . Then the unit vector of the viewpoint direction vector and obtain the corrected depth value. and the adjusted depth of focus Calculate the generated spatial correction offset;

[0292] Represented as:

[0293]

[0294] In the formula, Spatial correction offset;

[0295] An offset set is constructed based on all spatial correction offsets, and is represented as follows:

[0296]

[0297] The offset set is output as reprojection path correction data.

[0298] S154, Match the reprojection path correction data with the corresponding spatial mapping points in the initial spatial reprojection path, perform position adjustment processing on the spatial coordinates of the spatial mapping points, and generate the updated spatial reprojection path.

[0299] Represented as:

[0300]

[0301] If the remaining unbroken points remain unchanged, the updated spatial mapping points are represented as follows:

[0302]

[0303] The final updated reprojection path is:

[0304]

[0305] The updated spatial reprojection path.

[0306] Please see Figure 2 As shown, based on a unified inventive concept, this embodiment discloses a human-computer interactive digital ancient book display system based on aerial imaging, including:

[0307] The viewpoint acquisition module S21 acquires the set of user viewpoint positions and the set of page edge point positions in the aerial imaging area, calculates the spatial orientation projection between the page edge point and the user viewpoint, and generates a set of reflection orientation changes.

[0308] The path recognition module S22 identifies the boundary regions that cause the reflection path to jump based on the set of reflection orientation changes, extracts the viewpoint angle corresponding to the boundary regions, and selects the stable observation direction with the smallest reflection response change as the main reflection direction of the page.

[0309] The focal plane generation module S23 calculates the included angle projection value based on the main reflection direction of the page and the viewpoint angle, generates a continuous focal plane group of the page based on the included angle projection value, and constructs the focal depth estimation structure corresponding to the page area.

[0310] The focal depth derivation module S24 calculates the focal position change direction value based on the continuous focal plane group and focal depth estimation structure of the page, and generates the focal depth dynamic adjustment parameter value through the focal position change direction value;

[0311] The projection update module S25 generates a spatial reprojection path for the ancient book page image based on the dynamic adjustment parameter value of the depth of focus, identifies projection interruption events, and generates reprojection path correction data based on the projection interruption events to update the spatial reprojection path.

[0312] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A human-computer interaction method for digital ancient book display based on aerial imaging, characterized in that, The method includes: S11, obtain the set of user viewpoint positions and the set of page edge point positions in the aerial imaging area, and calculate the spatial orientation projection between the page edge point and the user viewpoint to generate a set of reflection orientation changes; S12, based on the set of reflection orientation changes, identify the boundary regions that cause the reflection path to jump, extract the viewpoint angles corresponding to the boundary regions, and select the stable observation direction with the smallest reflection response change as the main reflection direction of the page. S13, calculate the included angle projection value by using the main reflection direction of the page and the viewpoint angle, generate a continuous focal plane group of the page based on the included angle projection value, and construct the focal depth estimation structure corresponding to the page area; S14, calculate the focal position change direction value based on the continuous focal plane group and focal depth estimation structure of the page, and generate the focal depth dynamic adjustment parameter value through the focal position change direction value; S15: Generate a spatial reprojection path for the ancient book page image based on the dynamic adjustment parameter value of the depth of focus, identify projection interruption events, and generate reprojection path correction data based on the projection interruption events to update the spatial reprojection path.

2. The human-computer interactive digital ancient book display method based on aerial imaging according to claim 1, characterized in that, The steps for extracting the viewpoint angle corresponding to the boundary region, and selecting the stable viewing direction with the smallest change in reflection response as the main reflection direction of the page, are as follows: S121, Based on the set of reflection azimuth changes, perform direction difference calculation processing on each reflection azimuth change data in the set to generate direction difference data to characterize the strength of direction change; S122, calculate the direction change amplitude value at each position based on the continuous direction difference change value, and mark the position where the direction change amplitude value is greater than the jump recognition threshold as a jump candidate position; S123, perform continuous filtering on the location indexes that are not marked as boundary areas, and include the location indexes in the direction difference data that do not meet the jump recognition conditions into the continuous observation area set; S124, based on the direction difference data corresponding to each position index in the continuous observation area set, calculate the local change amplitude value of the direction difference data, identify the direction vector with the smallest local change amplitude value, and take the direction vector as the main reflection direction of the page.

3. The human-computer interactive digital ancient book display method based on aerial imaging according to claim 2, characterized in that, The logic for identifying the boundary region that triggers the reflection path jump based on the candidate jump location is as follows: a1, based on the direction difference data, performs adjacent difference calculations according to the order of the user's viewpoints to generate continuous direction difference change values; a2, based on the change value of the continuous directional difference, the positions where the change value of the continuous directional difference is greater than the predetermined jump recognition threshold are marked as jump candidate positions; a3, based on the jump candidate position, performs index aggregation processing on adjacent jump candidate positions to generate jump position groups; a4, based on the jump position group, maps the jump position group to the position index in the page space, and outputs the page area corresponding to the position index as the boundary area.

4. The human-computer interactive digital ancient book display method based on aerial imaging according to claim 3, characterized in that, The steps for performing index aggregation on adjacent jump candidate positions are as follows: a31, based on the interval between candidate jump positions, calculate the distance change value between adjacent candidate jump positions, and determine whether the distance between candidate positions is continuous based on the distance change value; a32, based on the distance change value, merge the candidate positions of the jump that have a continuous spacing of less than the set spacing difference into the same jump position subgroup in order; a33, calculate the internal index distribution difference for each jump position subgroup, and filter out candidate positions with abnormal deviations based on the distribution difference. Renumber the jump position subgroups after removing abnormal deviations and generate jump position groups.

5. The human-computer interactive digital ancient book display method based on aerial imaging according to claim 4, characterized in that, A continuous focal plane group for the page is generated based on the included angle projection value, and a depth-of-focus estimation structure corresponding to the page region is constructed, including: S131, Based on the main reflection direction of the page and the viewpoint angle, generate a filtered set of included angle projections; S132, Perform segment grouping processing based on the included angle projection set, divide the included angle projection set into multiple included angle projection segments, and construct projection mapping relationship based on multiple included angle projection segments to generate a continuous focal plane group for the page; S133, perform geometric optical path mapping processing based on the included angle projection value, calculate the viewing distance change of each included angle projection value on the virtual imaging plane, and use the viewing distance change as the relative focal length parameter value; S134: Input the relative focal length parameter values ​​into the predetermined focal depth distribution function to generate the corresponding initial focal plane depth estimation value, and integrate the focal depth estimation structure corresponding to the page area based on all the initial focal plane depth estimation values.

6. The human-computer interactive digital ancient book display method based on aerial imaging according to claim 5, characterized in that, The logic for generating the filtered set of included angle projections is as follows: b1, set the origin of the spatial Cartesian coordinate system to be located at the center of the page, and extract the main reflection direction vector of the page; b2, convert all viewpoint angles contained in the continuous observation area set generated in step S123 into spatial direction vectors, and uniformly map them to the spatial Cartesian coordinate system set in step b1; b3 sequentially calculates the angle projection value between the main reflection direction vector of the page and the direction vector of each viewpoint, verifies whether all angle projection values ​​are within the preset imaging field of view, removes data points that exceed the range, and outputs the set of angle projections after filtering and projection calculation.

7. The human-computer interactive digital ancient book display method based on aerial imaging according to claim 6, characterized in that, The steps for generating a continuous focal plane group for a page include: S132.1 Based on the included angle projection set, extract adjacent included angle projection values ​​according to the arrangement order of each included angle projection value in the set, and calculate the numerical difference between adjacent included angle projection values ​​to generate an included angle projection difference set for group processing. S132.2 Calculate the change in difference between consecutive angle projection differences, and identify the difference segments whose change in difference is continuously lower than a set change intensity threshold based on the change in difference, so as to generate a set of angle projection segments for segment grouping. S132.3, Perform projection curvature fitting processing on each included angle projection segment, and generate a continuous focal plane group for the page based on the projection curvature fitting processing.

8. The human-computer interactive digital ancient book display method based on aerial imaging according to claim 7, characterized in that, The steps for generating dynamic focus depth adjustment parameter values ​​are as follows: S141, Based on the focal position data corresponding to each focal plane in the continuous focal plane group of the page, calculate the direction difference of the focal position as the included angle projection value changes, and record the direction difference as a set of focal position direction difference values; S142, Based on the numerical change trend in the set of focal position direction difference values, identify the continuous rising segment and the continuous falling segment of the focal position change direction, and take the segment that meets the continuity condition as the candidate focal position change direction segment. S143, Based on the candidate focus change direction segments, calculate the direction change amplitude value of each direction segment, and extract the dominant focus change direction according to the direction change amplitude value to generate the focus position change direction value; S144: Input the focus position change direction value into the predetermined adjustment function to generate a dynamic focus adjustment parameter value for adjusting the focus stability of the page.

9. A human-computer interactive digital ancient book display method based on aerial imaging according to claim 8, characterized in that, The steps to update the spatial reprojection path are as follows: S151, based on the included angle projection set, the continuous focal plane group of the page and the dynamic adjustment parameter value of the depth of focus, performs spatial mapping processing on each included angle projection value to generate an initial spatial reprojection path composed of multiple spatial mapping points; S152, based on the spatial mapping point sequence of the initial spatial reprojection path, detect whether each mapping point has a spatial intersection with the boundary region, and mark the spatial mapping point that crosses the boundary region as a projection interruption event point; S153, based on the included angle projection value corresponding to each projection interruption event point, perform nearest neighbor search in the continuous focal plane group of the page to determine the effective mapping interval corresponding to the projection interruption event point, and generate reprojection path correction data according to the effective mapping interval. S154, the reprojection path correction data is matched with the corresponding spatial mapping points in the initial spatial reprojection path, and the spatial coordinates of the spatial mapping points are adjusted to generate the updated spatial reprojection path.

10. A human-computer interactive digital ancient book display system based on aerial imaging, used to implement the human-computer interactive digital ancient book display method based on aerial imaging as described in any one of claims 1-9, characterized in that, The system includes: The viewpoint acquisition module S21 acquires the set of user viewpoint positions and the set of page edge point positions in the aerial imaging area, calculates the spatial orientation projection between the page edge point and the user viewpoint, and generates a set of reflection orientation changes. The path recognition module S22 identifies the boundary regions that cause the reflection path to jump based on the set of reflection orientation changes, extracts the viewpoint angle corresponding to the boundary regions, and selects the stable observation direction with the smallest reflection response change as the main reflection direction of the page. The focal plane generation module S23 calculates the included angle projection value based on the main reflection direction of the page and the viewpoint angle, generates a continuous focal plane group of the page based on the included angle projection value, and constructs the focal depth estimation structure corresponding to the page area. The focal depth derivation module S24 calculates the focal position change direction value based on the continuous focal plane group and focal depth estimation structure of the page, and generates the focal depth dynamic adjustment parameter value through the focal position change direction value; The projection update module S25 generates a spatial reprojection path for the ancient book page image based on the dynamic adjustment parameter value of the depth of focus, identifies projection interruption events, and generates reprojection path correction data based on the projection interruption events to update the spatial reprojection path.

Images