A layered grid image-assisted calibration system

Abstract

The present invention relates to the field of image calibration technology, and specifically to a layered grid image auxiliary calibration system, comprising a brightness intersection screening module, a node synchronization correction module, a reference difference extraction module, a spatial offset allocation module, and a template link update module. In the present invention, by directly extracting the variation characteristics of pixel brightness, in-depth screening of image structural information is achieved, spatial nodes are immediately paired and synchronously compensated after acquisition, the spatial state between grid reference points is dynamically captured and carefully compared, the system automatically allocates spatial errors to each grid point position, parameter linkage compensation maintains the overall spatial coordination between grid points, template data is reconstructed in real time and chain mapping is performed, spatial data is continuously synchronized during the calibration process, spatial parameter correction maintains high adaptability under noise disturbance, adapts to multiple scene environments and can respond continuously and dynamically, and enhances calibration stability and data link integrity under different lighting, occlusion and non-ideal conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of image calibration, and in particular to a layered grid image auxiliary calibration system. Background Art

[0002] The field of image calibration technology involves the correction and registration of the spatial geometric relationship of the acquired digital image. This technical field mainly includes the correction of image geometric distortion based on reference templates, physical correction using calibration plates, determination of spatial transformation parameters through feature point matching, and improvement of calibration accuracy based on imaging optical path design. Among them, the traditional layered grid image auxiliary calibration system refers to setting up multiple layers of grid templates with different spatial resolutions to compare the spatial position of the calibrated image with the layered grid to achieve auxiliary correction of image distortion or spatial offset. Usually, a printed layered grid plate is placed on a fixed reference plane, and an image containing the layered grid is obtained with the help of a camera. The intersection of the grid lines or the grid point position is used as a judgment reference, and mathematical methods are used to analyze and calculate the geometric distortion parameters of the image, thereby completing the spatial position determination and parameter measurement required for image calibration.

[0003] Existing technologies mainly rely on single static template parameters and local feature points for spatial calibration. When faced with illumination changes, local noise or incomplete image structure, they lack a flexible parameter adjustment mechanism and are unable to adjust the compensation strategy according to the actual spatial distribution. It is difficult to ensure the consistency and global consistency of the calibration process in a changing environment, which easily leads to the calibration results being sensitive to environmental fluctuations. In complex application scenarios, spatial geometric errors accumulate, parameter updates and link feedback lag behind actual image changes, and it is difficult to maintain a coherent output of high-standard spatial alignment. Summary of the Invention

[0004] In order to solve the technical problems that the existing technology mainly relies on single static template parameters and local feature points for spatial calibration, lacks a flexible parameter adjustment mechanism when facing illumination changes, local noise or incomplete image structure, cannot adjust the compensation strategy according to the actual spatial distribution, is difficult to ensure the consistency and global consistency of the calibration process in a changing environment, easily leads to the calibration results being sensitive to environmental fluctuations, spatial geometric errors accumulating in complex application scenarios, parameter updates and link feedback lagging behind the actual changes in the image, and it is difficult to maintain a coherent output of high-standard spatial alignment, the embodiment of the present invention provides a layered grid image assisted calibration system. The technical solution is as follows:

[0005] In one aspect, a layered grid image-assisted calibration system is provided, comprising:

[0006] The brightness intersection screening module detects the brightness changes of each pixel in the horizontal and vertical directions based on the pixel brightness data of the original image, compares the difference between the pixel and the neighborhood mean, screens the intersection candidates according to the change standard, calculates the distribution density in the area, removes noise points whose density does not meet the standard, and obtains the grid intersection density feature;

[0007] The node synchronization correction module extracts spatial coordinates based on the dense grid intersection features, measures the Euclidean distance of node pairs, determines consistency according to the synchronization standard, marks unsynchronized nodes, calculates spatial offsets, and merges the offsets based on the node hierarchy and cross-slice relationship to obtain spatial synchronization compensation parameters.

[0008] The reference difference extraction module extracts the coordinates of the reference points in the grid structure based on the spatial synchronization compensation parameters, obtains the spatial distances and angles between all reference points, compares them with the template parameters, identifies the spatial distribution difference intervals between the combinations, and obtains the reference feature offset amplitude;

[0009] The spatial offset allocation module calls the current coordinates of all reference points based on the offset amplitude of the reference feature, combines the offset of each point with the spatial distribution factor, allocates the offset to the corresponding grid coordinates, and obtains the linkage spatial compensation data.

[0010] On the other hand, the grid intersection density features include intersection concentration, intersection distribution uniformity, and regional brightness contrast; the spatial synchronization compensation parameters include synchronization adjustment amplitude, pairing offset, and node distribution weight; the benchmark feature offset amplitude includes spatial consistency index, feature point offset, and benchmark point spatial distribution ratio; the linkage spatial compensation data includes coordinate correction factor, grid compensation amplitude, and linkage correlation coefficient.

[0011] On the other hand, the brightness intersection screening module includes:

[0012] The brightness change extraction submodule analyzes the original image pixel brightness data, calculates the brightness difference between each pixel and its adjacent pixels in the horizontal and vertical directions, determines the brightness fluctuation, and obtains the fluctuation intensity parameter based on the characteristics of the adjacent pixel group.

[0013] The intersection candidate screening submodule screens key pixel points based on the fluctuation intensity parameter, compares their distribution characteristics, determines the pixel positions that meet the intersection characteristics, classifies the pixels that meet the requirements, and obtains feature aggregation coordinates;

[0014] The candidate density calculation submodule calculates the pixel density distribution of each partition according to the feature aggregation coordinates, determines the aggregation state of candidate pixels in each area, eliminates and sorts areas that do not meet the distribution rules, and obtains grid intersection density features.

[0015] On the other hand, the node synchronization correction module includes:

[0016] The node pairing construction submodule groups all intersections according to a hierarchical structure based on the dense grid intersection features, combines the node spatial coordinates, constructs node pairing combinations of different levels, screens nodes with associated relationships, and establishes a hierarchical pairing structure to obtain a hierarchical combination index;

[0017] The spatial consistency judgment submodule compares the spatial interval of each group of nodes based on the hierarchical combination index, analyzes the corresponding spatial distribution status of each node pair, determines whether the synchronization standard is met, marks the offset status of inconsistent paired nodes, and obtains the pairing space offset parameter set;

[0018] The offset merging calculation submodule matches the hierarchical area and intersection area corresponding to each group of nodes according to the paired spatial offset parameter set, analyzes the consistency of multiple groups of offset directions in the intersection position, merges the associated offsets according to regional affiliation, and obtains the spatial synchronization compensation parameters.

[0019] On the other hand, the benchmark difference extraction module includes:

[0020] The coordinate data acquisition submodule collects the spatial position of each reference point in the grid structure based on the spatial synchronization compensation parameters, determines the ownership relationship of the nodes in the hierarchical structure, locates the reference nodes according to the current hierarchical structure, and obtains a reference position set;

[0021] The spatial parameter calculation submodule calculates the spatial distance between each reference point based on the reference position set, selects any three points to form a vector for angle analysis, and combines the distance and angle statistical parameters to determine the spatial distribution state of the node combination and obtain the spatial characteristic factor;

[0022] The geometric difference determination submodule compares the spatial characteristic factors with the standard parameters in the template structure, analyzes the variation range of the combination in spatial distribution and angle characteristics, screens the combination interval with differences, and obtains the reference feature offset amplitude.

[0023] On the other hand, the comparison with the standard parameters within the template structure is carried out using the formula:

[0024] ;

[0025] Calculate the spatial difference value, filter the combination interval with differences, and obtain the benchmark feature offset amplitude, where: Representative reference point and The spatial difference between Represents the current reference point and The Euclidean distance between Represents the reference point in the template structure and The theoretical Euclidean distance between Represents the current reference point The angle formed, Represents the reference point in the template structure The theoretical angle formed by Represents the current reference point Brightness contrast factor of the three-point combination.

[0026] On the other hand, the spatial offset allocation module includes:

[0027] The coordinate joint arrangement submodule arranges the current coordinates of all reference points based on the offset amplitude of the reference feature, determines the spatial ownership of each point in the grid structure, uniformly numbers and arranges the coordinate and offset information, and obtains the node ownership sequence based on the image grid distribution;

[0028] The distribution factor construction submodule calculates the spatial distance relationship between each benchmark point and the surrounding grid points based on the node attribution sequence, analyzes the spatial linkage between the nodes and the grid points, determines the influence of each benchmark point in the local area, and obtains the influence factor;

[0029] The offset weight distribution submodule performs a linear weighted distribution operation based on the influencing factors and the offset amplitude and distribution factor of each grid point associated with the reference point, distributes the offset signal to the target grid point, and obtains the linkage space compensation data.

[0030] On the other hand, the offset amplitude and distribution factor for each grid point associated with the reference point are calculated using the formula:

[0031] ;

[0032] Calculate the grid point linkage compensation weight value, distribute the offset signal to the target grid point, and obtain the linkage space compensation data, where: Representative The linkage compensation weight value of each target grid point, Representatives and The grid point associated The spatial coordinate difference value of the reference point relative to the template standard parameter, Representative The reference point is relative to the The spatial distribution factor of the grid points, Representatives and The number of reference points associated with each grid point.

[0033] In another aspect, the system further comprises:

[0034] The template link update module reconstructs the template space parameters based on the linkage space compensation data and the coordinates of each grid point, replaces the calibration system cache data, builds a linked list structure according to the grid structure template and the boundary alignment fragment, and outputs the template space update configuration;

[0035] The template space update configuration includes template structure parameters, space mapping relationship, and boundary alignment data.

[0036] On the other hand, the template link update module includes:

[0037] The parameter structure reconstruction submodule determines the spatial distribution of each grid point coordinate based on the linkage spatial compensation data, calculates the structural correspondence between adjacent grid points, adjusts the spatial parameter arrangement order of each group of grid points, optimizes the spatial topological layout, and generates structural mapping parameters;

[0038] The cache data replacement submodule compares the structure mapping parameters with the structure parameters in the current calibration system cache, analyzes the differences in the stored contents, adjusts the inconsistent grid coordinates and mapping relationships, updates the structure parameters, and obtains the template parameter integration;

[0039] The chain configuration generation submodule divides the boundary area in the grid template based on the template parameter integration, determines the chain pointing order between each grid point, combines the parameter path of each grid point, integrates the chain list structure, and outputs the template space update configuration.

[0040] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0041] By directly extracting the changing characteristics of pixel brightness, in-depth screening of image structural information is achieved. Spatial nodes are paired and synchronously compensated immediately after acquisition. The spatial state between grid reference points is dynamically captured and carefully compared. The system automatically allocates spatial errors to each grid point position. Parameter linkage compensation maintains the overall spatial coordination between grid points. Template data is reconstructed and chain-mapped in real time. Spatial data is continuously synchronized during the calibration process. Spatial parameter correction maintains high adaptability under noise disturbances. It can adapt to multiple scene environments and respond dynamically and continuously, driving the rapid adjustment of spatial geometric relationships and the fine control of template parameters, enhancing calibration stability and data link integrity under different lighting, occlusion and non-ideal conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0043] Figure 1 A schematic diagram of the system of the present invention;

[0044] Figure 2 Schematic diagram of the system framework of the present invention;

[0045] Figure 3 This is a flow chart of the brightness intersection screening module of the present invention;

[0046] Figure 4 This is a flow chart of the node synchronization correction module of the present invention;

[0047] Figure 5 is a flow chart of the benchmark difference extraction module of the present invention;

[0048] Figure 6 This is a flow chart of the spatial offset allocation module of the present invention;

[0049] Figure 7 This is a flow chart of the template link update module of the present invention. DETAILED DESCRIPTION

[0050] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0051] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0052] In the embodiments of the present invention, the terms "image" and "picture" may be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same. The terms "of," "corresponding," and "corresponding" may be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same.

[0053] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0054] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0055] The embodiment of the present invention provides a layered grid image auxiliary calibration system, such as Figure 1 As shown, the system includes:

[0056] The brightness intersection screening module detects the horizontal and vertical brightness changes of each pixel based on the pixel brightness data of the original image, compares the difference between the brightness of each pixel and the mean of adjacent pixels, and screens candidate intersection points based on the brightness change standard. It then calculates the density of pixel distribution in the candidate area and removes noise areas whose density does not meet the screening criteria to obtain the grid intersection density feature.

[0057] The node synchronization correction module is based on the dense grid intersection feature and the hierarchical structure to build node pairs at each level, extract the spatial coordinate information of the nodes, measure the Euclidean distance between the node pairs, determine the spatial consistency between the pairs according to the synchronization range standard, mark the unsynchronized nodes, obtain the spatial offset parameters, and merge the offsets according to the spatial relationship between the node hierarchy area and the node intersection area to obtain the spatial synchronization compensation parameters.

[0058] The benchmark difference extraction module extracts the current coordinates of the benchmark points in the grid structure based on the spatial synchronization compensation parameters, obtains the spatial distance and angle parameters of all benchmark points, compares the current data with the template standard parameters, identifies the geometric difference interval of the spatial distribution between the combinations, and obtains the benchmark feature offset amplitude;

[0059] The spatial offset allocation module calls the current coordinates of all reference points based on the offset amplitude of the reference feature, combines the offset amplitude of each point with the spatial distribution factor, and distributes all offsets to the corresponding grid coordinates in a linear weighted manner to obtain the linkage spatial compensation data;

[0060] The template link update module reconstructs the template space parameter structure based on the linkage space compensation data and the coordinate information of each grid point, replaces the current calibration system cache data, aligns the fragments with the boundary according to the grid structure template, builds a linked list structure, and outputs the template space update configuration.

[0061] The grid intersection density features include intersection concentration, intersection distribution uniformity, and regional brightness contrast. The spatial synchronization compensation parameters include synchronization adjustment amplitude, pairing offset, and node distribution weight. The benchmark feature offset amplitude includes spatial consistency index, feature point offset, and benchmark point spatial distribution ratio. The linkage space compensation data includes coordinate correction factor, grid compensation amplitude, and linkage correlation coefficient. The template space update configuration includes template structure parameters, spatial mapping relationship, and boundary alignment data.

[0062] In the brightness intersection screening module, brightness change refers to the amplitude of the change between the horizontal and vertical brightness values ​​of a pixel in the original image and the average brightness of its surrounding pixels (up, down, left, right, or diagonal), and is often used to reflect the image characteristics of the grid intersection area. The brightness change standard is the judgment standard used to determine whether the brightness change is significant. It is usually a set brightness difference parameter threshold. Pixels below this standard are not considered intersection candidates. Intersection candidate points refer to the set of pixels at the grid intersection position that are initially screened out through brightness change and judgment standards. They are the basis for subsequent density statistics and further identification. Density refers to the density level of pixel distribution in a certain area. It is usually quantified by counting the number of intersection candidate points per unit area and is used to reflect whether the intersection characteristics of the area are obvious. The screening standard refers to the specific requirements for density determination. Only candidate point areas with density that meets this standard will be identified as grid intersection areas. Noise areas refer to areas with pixel anomalies caused by lighting, reflection, image defects, etc. These areas do not have the regular characteristics of grid intersections and need to be excluded during the screening process.

[0063] In the node synchronization correction module, a hierarchical structure refers to a structure in which a grid image is spatially or functionally divided into different levels or regions. Common examples include multiple grids and nested multiple regions, facilitating independent or collaborative processing of nodes at each level. Node pairing refers to pairing nodes (grid intersections) within the same or different levels according to spatial rules, facilitating subsequent measurement of spatial relationships and deviations between pairs. The synchronization range standard refers to the distance tolerance used to determine whether the distance between paired nodes is normal (i.e., meets preset requirements). Exceeding this range indicates a missynchronization. Spatial consistency between pairs refers to whether the spatial distance between paired nodes in the same group falls within the synchronization range standard; otherwise, it is considered spatially inconsistent. The spatial offset parameter refers to the quantified spatial coordinate deviation of a node pair identified as missynchronized, reflecting the offset between the actual and ideal positions of the node. The node hierarchical region refers to the specific hierarchical location of a node within the hierarchical grid structure and is used to determine attribution and merging rules. The node intersection area refers to the intersection or overlap of nodes at different levels or regions, which is a special location area for node pairing and offset merging.

[0064] In the benchmark difference extraction module, the current coordinates of the benchmark point refer to the actual coordinate data in space of the specific node (benchmark point) in the grid structure that serves as a reference or identifier during each calibration; the pairwise spatial distance refers to the Euclidean distance between all benchmark points, calculated in a pairwise combination; the angle parameter refers to the angle (vector angle) formed by any three benchmark points, reflecting the spatial geometric relationship between the benchmark points; the template standard parameter refers to the theoretical distance, angle, spatial layout and other parameters of each benchmark point in the template grid structure when the system is initialized, which serves as a reference standard for determining offsets or differences; the geometric difference interval refers to the statistical interval of all spatial offsets and angle changes obtained after comparing the currently collected benchmark point spatial parameters with the template standard parameters, which is used to identify anomalies or trends.

[0065] In the spatial offset allocation module, the offset amplitude refers to the numerical expression of the spatial coordinate difference (such as distance increase or decrease, angle change, etc.) generated by each reference point relative to the template standard parameters; the spatial distribution factor refers to the parameter that assigns weights according to the spatial position of the reference point or grid point in the overall structure, which is used to determine the degree of influence of the offset of each reference point on the final grid point adjustment; the linear weighted method refers to the distribution and summation of the offset amplitudes of all reference points through linear weights, and the reasonable transfer of the offset to the relevant grid points according to the weights; the corresponding grid point coordinates refer to the coordinate position of each grid point that actually needs to be spatially adjusted, which is the landing point of all compensation and correction.

[0066] In the template link update module, the grid structure template refers to the basic template in the layered grid image assisted calibration system, which defines the spatial relationship and structural topology of each grid point and each reference point under ideal conditions; the boundary alignment fragment refers to the sub-unit fragment divided by the grid structure at the boundary and junction, which is used to finely control the boundary accuracy during the template space adjustment and calibration process; the linked list structure refers to a structure that organizes and stores all grid point spatial parameters in a data chain manner (such as a linked list, a chain mapping), which facilitates the system to quickly access, update and maintain spatial configuration data.

[0067] like Figure 2 and Figure 3 As shown, the brightness intersection screening module includes:

[0068] The brightness change extraction submodule analyzes each pixel of the original image based on the pixel brightness data, calculates the brightness difference between it and its horizontal and vertical adjacent pixels, determines the brightness fluctuation, and obtains the fluctuation intensity parameter based on the characteristics of the adjacent pixel group;

[0069] Extract the brightness values ​​of all pixels from the original image, and traverse the image matrix row by row and column by column. During the traversal process, obtain the brightness values ​​of the adjacent pixels on the left and right of the horizontal direction for each current pixel, and obtain the brightness values ​​of the adjacent pixels above and below it in the vertical direction. By comparing the average brightness difference of the current pixel brightness value with the average brightness difference of its two horizontally adjacent pixels, and the average brightness difference of its two vertically adjacent pixels, evaluate the brightness change degree of the pixel in the two directions, and then combine the brightness change values ​​in the two directions to express the overall fluctuation intensity of the brightness of the pixel in space. The direction weight setting needs to be considered in the processing process. If the horizontal grid boundary contrast is strong in the image structure, the horizontal weight can be set to be higher than the vertical weight. In the actual acquisition scene, taking a 640×480 pixel image as an example, assuming that the brightness of a pixel in the center of the image is the same as that of its If the difference between the horizontal neighboring points is obvious, while the vertical change is small, it can be identified that the horizontal fluctuation of the pixel is dominant. After completing the traversal operation on the entire image, the brightness fluctuation intensity results of each pixel are saved in the matrix diagram. Then, a pixel area of ​​fixed size is selected around each pixel, and the average value of the fluctuation intensity in the area is calculated. The fluctuation intensity of the current pixel is compared with the value obtained by multiplying the average value of the area plus the brightness standard deviation. When the brightness fluctuation of the pixel is significantly higher than the average regional level, the pixel is marked as a point with significant brightness change. Combined with the image acquisition scenario, if the brightness of a pixel in the screen shot is significantly higher than that of the adjacent pixels and the brightness fluctuation amplitude of its surrounding area is limited, it can be identified as a pixel with obvious brightness change characteristics through the above method, and its position in the image is recorded to complete the extraction and preliminary screening of all brightness change intensity parameters of the image.

[0070] The intersection candidate screening submodule selects key pixel points based on the fluctuation intensity parameter, compares their distribution characteristics, determines the pixel positions that meet the intersection characteristics, classifies the pixels that meet the requirements, and obtains the feature aggregation coordinates;

[0071] For each strong point, a pixel neighborhood within a certain range around it is constructed, and the number of other strong fluctuation points in the neighborhood is counted, and the statistical value is compared with the preset candidate density judgment standard. If there are many similar strong fluctuation points in the neighborhood of the pixel point, the pixel is considered to have the potential feature of becoming an intersection point. On the contrary, if there are few similar points around the pixel point, it is determined to be edge disturbance or background noise and is eliminated. After the processing is completed, a preliminary set of candidate pixel points is obtained. The entire image is further divided into equally spaced grid areas. Density analysis is performed on the candidate pixels screened out in each area, and the number of candidate points contained in each grid is counted. It is compared horizontally with the average density of the entire image. When a certain When the number of candidate points in an area is significantly higher than the average level of the entire image, the area is identified as having obvious aggregation and is marked as an intersection aggregation area. On the contrary, if the number of candidate points in an area is low, it is determined to be an atypical area and discarded. In actual operation, assuming that a 640×480 pixel image is divided into 10×10 pixel grids, there are more than three thousand areas in total. There are fifteen candidate points in a specific area, while the average of the entire image is six points. In this case, the area clearly has intersection features. The pixel points that are determined to be intersection aggregation areas will be reorganized according to their image coordinate order to form a set of stable and highly structured feature aggregation coordinate data, which will serve as the input basis for subsequent node recognition processing.

[0072] The candidate density calculation submodule calculates the pixel density distribution of each partition based on the feature aggregation coordinates, determines the aggregation state of candidate pixels in each area, eliminates and sorts out areas that do not meet the distribution rules, and obtains the grid intersection density features;

[0073] The distribution rule refers to the standardized judgment conditions used to determine whether the aggregation state of candidate pixels in each partition (or area) meets the grid intersection characteristics, including:

[0074] Regional candidate point quantity standard

[0075] Determine whether the number of candidate points in each partition reaches the preset density benchmark, and compare it with the average candidate density and standard deviation of the entire map. If it is higher than a certain judgment benchmark value, the area is considered to have intersection clustering characteristics;

[0076] Spatial aggregation requirements

[0077] In addition to the number, the spatial proximity of candidate points within the partition must also be considered. Only when the candidate points show a certain spatial aggregation state (that is, there are many strong points close to each other) can they be considered to be effectively clustered. Otherwise, even if the number is close, they may be regarded as disordered distribution or noise points.

[0078] Remove isolated points and sparse areas

[0079] If the number of candidate points in a certain area is low, or if there are candidate points but they are too scattered and do not have the clustering properties of grid intersections, these areas will be eliminated and will not be counted as intersection density features.

[0080] The distribution rule is a set of judgment criteria for the number, degree of aggregation and distribution pattern of candidate points in space. It is used to screen out areas that do not conform to the typical spatial distribution characteristics of grid intersections, and only retain dense, aggregated and highly structured intersection areas to obtain high-quality grid intersection dense feature data.

[0081] To perform a more accurate regional density analysis on the image, first, the image is uniformly partitioned according to the overall image size. The number of candidate points contained in each partition is counted and the statistical results of each region are recorded. Then, the statistical data of all regions of the entire image are analyzed centrally to calculate the density distribution of candidate points in the entire region. By comparing the number of candidate points in each region with the average density and density standard deviation of the entire image, those regions with significantly higher than average density are identified and assigned a higher density score. The score represents the degree of clustering of candidate points in the region. When a score is higher than the density judgment benchmark value, the region is retained; if it is lower, it is eliminated. For example, if the average density of the image region is five points and the density fluctuation is two points, then a region with nine candidate points can be considered to have obvious clustering characteristics. The region score is recorded when it is above the average density and higher than the set benchmark. On the contrary, if a region has only two points, the score is significantly lower than the standard and is eliminated. Only the pixels in all high-density regions are retained to form a new intersection pixel set. In this process, all isolated candidate pixels without adjacent points are removed to form recognized grid intersection dense feature data for subsequent hierarchical node recognition.

[0082] like Figure 2 and Figure 4 As shown, the node synchronization correction module includes:

[0083] The node pairing construction submodule groups all intersections according to the hierarchical structure based on the dense grid intersection features, combines the node spatial coordinates, constructs node pairing combinations of different levels, filters nodes with associated relationships, and establishes a hierarchical pairing structure to obtain a hierarchical combination index;

[0084] All the extracted intersection coordinate data are classified and sorted according to a preset hierarchical structure. The structure is set according to the spatial hierarchical division method of the grid pattern in the image. For example, in a three-layer nested grid, the coordinates of each intersection are classified into the corresponding first-level, second-level or third-level area according to its position. In the classification process, the coordinate values ​​of all intersections are read, and then the level to which each intersection belongs is determined according to the absolute position range of each intersection in the image. If the horizontal and vertical coordinates of an intersection are both in the center one-third of the image, it is classified into the middle layer. The rest are set as the peripheral layer or the core layer according to the image distribution. After the stratification is completed, the intersection set in each layer is matched point by point. The method is to select a reference intersection in the intersection set of the layer in turn, and then use the set The maximum pairing radius search range is used as the boundary to find its adjacent intersections, and pairwise combinations are constructed. The Euclidean distance between the two points is calculated as the reference value of the pairing spacing. All combinations are then screened and only those node pairs with regular arrangement in spatial relationships or spacing that meets the grid structure characteristics are retained. For example, if the maximum pairing radius is set to 80 pixels, if there are three points adjacent to a reference point with distances of 42, 78, and 90 pixels respectively, only the first two groups of pairings are retained. Subsequently, all valid node pairings in each layer are structured and organized, indexed by the level number, and the intersection number and coordinate value corresponding to each pairing are recorded to generate a level combination index that can be called by subsequent modules to describe all valid intersection pairings in the hierarchical structure.

[0085] The spatial consistency judgment submodule compares the spatial interval of each group of nodes based on the hierarchical combination index, analyzes the relative spatial distribution of each node pair, determines whether the synchronization standard is met, marks the offset status of inconsistent paired nodes, and obtains the pairing space offset parameter set;

[0086] The spatial coordinates of each pair of nodes are read, and the horizontal and vertical distances between the two points are obtained by subtraction. The corresponding actual spacing values ​​are then counted and compared with the set standard synchronization range to determine whether the spacing between the two points is within an acceptable error range. If the spacing deviation exceeds this range, the pairing is marked as out of sync. In practice, if the synchronization range is set to allow a maximum deviation of ±5 pixels, then any two points are considered synchronized if the spacing does not exceed 5 pixels compared to the ideal spacing. Conversely, if the spacing between a pair of nodes should be 60 pixels but is measured to be 67 pixels, the pairing is considered out of sync and marked as an offset pairing. The spatial offset of the node pair is recorded, and the level and pairing number to which it belongs are synchronized. This process is repeated for all pairs to complete the spatial consistency judgment of all node pairs in the entire image. All pairs that do not meet the synchronization standard and their offset differences are summarized to form a set of pairing spatial offset parameters, which are used to describe the node pairing data with spatial misalignment at each level.

[0087] The offset merging calculation submodule matches the hierarchical area and intersection area corresponding to each group of nodes based on the paired spatial offset parameter set, analyzes the consistency of multiple groups of offset directions in the intersection position, merges the associated offsets according to regional affiliation, and obtains the spatial synchronization compensation parameters;

[0088] Match and identify the hierarchical area to which each group of nodes belongs and whether it is located in the intersection area. First, traverse all node pairs marked as offsets, extract their hierarchical labels, and then call the area information of each intersection from the structure template to determine whether it is in the intersection area of ​​multiple hierarchical areas, that is, the intersection area. If the node belongs to the intersection area, it needs to participate in the multi-group pairing offset analysis, classify all offset nodes in the same intersection area in terms of direction, and group the node sets with similar offset directions into similar offset clusters. In each offset cluster, numerically integrate all offsets and calculate them by area. The average offset amplitude of all nodes in the cluster is calculated. If the offset direction of the nodes in a cluster is all upward right, and the offsets are 3, 5, and 4 pixels respectively, the merged result is a 4-pixel upward right offset. This offset result is then registered as the merged offset of the area according to the area number. In non-intersection areas, if a node only participates in a single level, its offset is directly merged and recorded as the offset data of this area. The offset merging results of all areas are unified into a set of spatial synchronization compensation parameters, which are used to describe the spatial correction amount of each level and each area in the entire image when the node is misaligned.

[0089] like Figure 2 and Figure 5 As shown, the benchmark difference extraction module includes:

[0090] The coordinate data acquisition submodule collects the spatial position of each reference point in the grid structure based on the spatial synchronization compensation parameters, determines the ownership relationship of the nodes in the hierarchical structure, locates the reference nodes according to the current hierarchical structure, and obtains the reference position set;

[0091] By reading the final compensated coordinate values ​​corresponding to each grid intersection, the horizontal and vertical spatial position data of all reference points in the calibrated grid structure are extracted point by point, and then a reference point coordinate index table is established, and the position of each reference point in the image coordinate system is recorded as a two-dimensional coordinate group. Then, the attribution judgment operation is performed on the reference points, that is, all reference points are traversed and their level and area in the hierarchical grid structure are judged point by point. According to the hierarchical structure area boundary set in the image template, the coordinate position of the current reference point is compared point by point to see whether it falls within the coordinate boundary range of each level area. If its horizontal and vertical coordinates are both within a certain level boundary interval, it is determined that the reference point belongs to that level. The regional coordinate lower limit and upper limit judgment method is adopted in the judgment process. For example, when the horizontal and vertical coordinates of the first level area in the hierarchical structure are 10 The vertical coordinates are (0, 300, 150, 350). If the coordinates of a reference point are (220, 280), it is determined that the point belongs to the first-level area. Similarly, all reference points are classified into the corresponding hierarchical areas according to their coordinates. Then, within each hierarchical area, all reference points in the area are extracted and a reference node set is established based on the coordinate sorting. Each node in the set retains its unique number, belonging level and current coordinate position information. In a 640×480 standard test image, it is assumed that there are 15 reference points in total. After judging their hierarchical affiliation, 5 of them are identified as belonging to the center level, another 5 points as belonging to the peripheral layer, and the remaining 5 points as the boundary transition level. Based on this, three reference node sets are established respectively as the input set for subsequent spatial analysis to obtain a complete reference position set.

[0092] The spatial parameter calculation submodule calculates the spatial distance between each reference point based on the reference position set, selects any three points to form a vector for angle analysis, and combines the distance and angle statistical parameters to determine the spatial distribution state of the node combination and obtain the spatial characteristic factor;

[0093] First, take the reference node set of each level as the unit, read the coordinate positions of all reference points layer by layer, then build the horizontal and vertical spacing relationship for every two reference points, calculate the horizontal and vertical distances between all pairs of reference points by point-by-point matching, and calculate the straight-line distance between the two points on this basis to record the actual spatial spacing value between the node pairs, and then partition each group of interval values ​​according to the distance range, and divide them into three categories: less than 50 pixels, 50-100 pixels in the middle, and greater than 100 pixels, to determine whether they fall within the set standard template distance tolerance range. If the distance between two points is 62 pixels and should be 60 pixels in the reference template, then the distance difference falls within the ±5 pixel allowable interval and is marked as a valid distribution. Then, any three points are selected from the valid node set to form a spatial vector group, and the three points are divided into a vector group according to the distance range. The angle structure is constructed based on the position relationship. The angle value formed between any two vectors is calculated respectively. Then all the angle values ​​are statistically analyzed, the average angle and the maximum and minimum angle values ​​are calculated, and a judgment is made as to whether a structure close to equilateral, right angle or acute angle is formed. The judgment standard is based on whether the angle falls within the corresponding angle tolerance range. For example, when the three angles are 60, 59, and 61 degrees respectively, and the angle judgment tolerance is set to ±5 degrees, the three-point combination is marked as forming an approximately equilateral triangle relationship. If it is 30, 60, or 90 degrees, it is judged to be a right-angle structure. Finally, all the point-to-point distance statistics and the three-point angle statistics are merged and sorted, and the spatial relationship characteristics of each group of reference point combinations are marked, including distance intervals, angle patterns, and combination structure forms, to form a set of spatial distribution state description data, and the data is summarized by region to generate spatial feature factors.

[0094] The geometric difference determination submodule compares the spatial characteristic factors with the standard parameters in the template structure, analyzes the variation range of the combination in spatial distribution and angular characteristics, screens the combination intervals with differences, and obtains the benchmark feature offset amplitude;

[0095] Comparison with standard parameters within the template structure is done using the formula:

[0096] ;

[0097] Calculate the spatial difference value, filter the combination interval with differences, and obtain the benchmark feature offset amplitude, where: Representative reference point and The spatial difference between Represents the current reference point and The Euclidean distance between Represents the reference point in the template structure and The theoretical Euclidean distance between Represents the current reference point The angle formed, Represents the reference point in the template structure The theoretical angle formed by Represents the current reference point Brightness contrast factor of the three-point combination;

[0098] The spatial difference value is an indicator used to quantitatively reflect the degree of difference between the spatial relationship between the actual reference points in the current grid structure and the spatial relationship between the theoretical reference points in the template structure. It reflects the degree of deviation in the spatial distribution of the reference points during the calibration process of the grid structure and can be used for subsequent screening of reference point combinations with obvious differences.

[0099] Fiducial points in the current image , , its Euclidean distance Calculated as:

[0100] ;

[0101] Reference point in template structure The theoretical distance is the system initialization calibration value, recorded as ;

[0102] Angle parameter part:

[0103] Set a benchmark ,but:

[0104] ;

[0105] ;

[0106] Inner product:

[0107] ;

[0108] Mould length:

[0109] ;

[0110] ;

[0111] The current angle is calculated as:

[0112] ;

[0113] The theoretical angle between the three points in the template is set to ;

[0114] Brightness contrast factor part:

[0115] For each of the three points Neighborhood area sampling, extracting grayscale value sets ,get:

[0116] Maximum brightness: 243;

[0117] Minimum brightness: 175;

[0118] The average brightness (sum of 25 points divided by 25) is about 210.4;

[0119] Then the brightness contrast factor is calculated as:

[0120] ;

[0121] Parameter normalization results (for substitution into the formula):

[0122] 、 , after normalization: , ;

[0123] 、 , after normalization: , ;

[0124] (keep original value);

[0125] Substitute into the formula to calculate:

[0126] ;

[0127] ;

[0128] The calculated spatial difference value is , the value is higher than the set threshold , indicating the combination of reference points in the current image The geometric relationship that deviates significantly from the template structure should be judged as a significant difference combination. The result is used to further obtain the offset amplitude of the benchmark feature as the basis for spatial compensation. The formula is introduced by the brightness contrast factor As a regulating factor for angle deviation, the image brightness characteristics are effectively introduced into the spatial structure calculation; at the same time, the squared distance term and the squared angle term are superimposed to construct a joint spatial and brightness difference model, thereby enhancing the system's spatial judgment robustness under non-ideal lighting conditions, which is of supporting significance for subsequent calibration template updates.

[0129] like Figure 2 and Figure 6 As shown, the spatial offset allocation module includes:

[0130] The coordinate joint arrangement submodule arranges the current coordinates of all reference points based on the offset amplitude of the reference features, determines the spatial ownership of each point in the grid structure, unifies the coordinate and offset information, and obtains the node ownership sequence based on the image grid distribution.

[0131] The current coordinate data of all reference points involved in the calibration are extracted from the image matrix and collected uniformly. During the extraction process, the number of each reference point, its corresponding horizontal and vertical coordinate position in the image, and its offset amplitude value are recorded. The coordinates and offset values ​​are then sequentially combined into a structured data set. The spatial ownership relationship of each reference point in the entire grid structure is then judged one by one. The judgment process is completed by comparing the current reference point coordinates with the boundary coordinate range of each area in the grid structure. During the judgment process, for any point, its coordinate value is read and it is found whether it falls within the horizontal and vertical boundaries of a certain grid unit. For example, if the coordinates of a point are (210, 180), and the grid partitions are horizontally from 200 to 250 and vertically from 150 to 200, then the point belongs to that area. After completing the ownership judgment, a unique number is assigned to each point and the row priority is prioritized. The numbering order is unified in a column-first or column-first manner, from left to right and from top to bottom, and the coordinates of the point corresponding to the number and its corresponding offset amplitude are recorded. All the numbered point data are unified and sorted, and each reference point is associated with the number of the grid area where it is located. Then the global distribution of the image grid points is read, and a complete grid grid structure matrix is ​​constructed according to the horizontal and vertical spacing between the grid points. Each reference point is classified into the corresponding grid matrix unit according to the regional range. For example, if the image size is 640×480 and the grid spacing is set to 40 pixels, a 16×12 grid area is constructed. If a point is located in the area of ​​the 6th column horizontally and the 5th row vertically, its attribution sequence is (5, 6). Finally, all the reference points are unified and integrated according to the number, coordinates, offset amplitude and grid attribution area to generate a node attribution sequence.

[0132] The distribution factor construction submodule calculates the spatial distance relationship between each benchmark point and the surrounding grid points based on the node attribution sequence, analyzes the spatial linkage between the nodes and the grid points, determines the influence of each benchmark point in the local area, and obtains the impact factor;

[0133] The image coordinate value of each reference point is read one by one, and the spatial distance relationship between the current reference point and the surrounding grid points is calculated by combining the information of all grid points in the grid point area to which it belongs. This process traverses a certain number of grid points around the grid unit to which the point belongs, for example, eight adjacent grid points are selected in a 3×3 range, and the horizontal and vertical differences between each pair of reference points and grid points are calculated respectively. The straight-line distance between the point and the grid point is obtained by combining the differences in the two directions, and the spatial distance between each grid point and the current reference point is saved in a mapping list. The distance values ​​are then classified and divided into direct influence areas and indirect influence areas according to the set influence radius threshold. The default direct influence radius is 60 pixels, that is, if the distance between a reference point and a grid point is less than or equal to 60 pixels, the grid point is considered to be affected by it. If the distance to the directly affected target grid point is within the range of 60 to 120 pixels, it is determined to be a grid point within the indirect influence range. Then, a different action factor is assigned to each grid point according to the distance. The closer the distance, the greater the action factor, and vice versa. For example, the action factor is set to 1.0 in the direct influence area, and linearly decreases with distance to the minimum value of 0.2 in the indirect influence area. If the distance from a reference point to a grid point is 45 pixels, the action factor is assigned to 0.95, and if it is 90 pixels, the action factor is assigned to 0.4. After completing the assignment of the influence factors of all grid points, a set of action factor structure data is established for each reference point to describe its ability to contribute to the offset linkage of the surrounding grid points in the local area. Finally, the action factor data of all reference points are integrated to generate a complete action factor result set.

[0134] The offset weight allocation submodule performs a linear weighted allocation operation based on the influencing factors and the offset amplitude and distribution factor of each grid point associated with the reference point, allocates the offset signal to the target grid point, and obtains the linkage spatial compensation data;

[0135] For the offset amplitude and distribution factor of each grid point associated with the reference point, the formula is used:

[0136] ;

[0137] Calculate the grid point linkage compensation weight value, distribute the offset signal to the target grid point, and obtain the linkage space compensation data, where: Representative The linkage compensation weight value of each target grid point, Representatives and The grid point associated The spatial coordinate difference value of the reference point relative to the template standard parameter, Representative The reference point is relative to the The spatial distribution factor of the grid points, Representatives and The number of reference points associated with each grid point;

[0138] The linkage compensation weight value refers to the weighted coefficient used to reflect the degree of influence of each reference point on the final coordinate adjustment of each target grid point when it receives offset compensation signals from all its associated reference points during the spatial calibration process in the layered grid image assisted calibration system. It determines the influence and contribution ratio of different reference points in the grid point calibration compensation.

[0139] No. Target grid association The parameters are as follows:

[0140] The first reference point: The coordinates recorded by the system are (102.5, 97.0, 100.2) mm, the standard coordinates of the template are (100.0, 100.0, 100.0) mm, the measured spatial coordinate difference is 3.02 mm, the distance from the target grid point is 48.2 mm, and the normalized distribution factor is 0.021;

[0141] The second reference point: The coordinate difference is measured to be 4.75 mm, the distance is 61.4 mm, and the normalized distribution factor is 0.016;

[0142] The third reference point: The coordinate difference is measured to be 2.15 mm, the distance is 40.1 mm, and the normalized distribution factor is 0.025;

[0143] Substituting the above values ​​into the formula:

[0144] ;

[0145] The numerator is calculated as follows:

[0146] ;

[0147] ;

[0148] ;

[0149] The sum of the numerators is:

[0150] ;

[0151] The sum of the denominators is:

[0152] ;

[0153] The calculation results are:

[0154] ;

[0155] The results show that the linkage compensation weight value corresponding to the first target grid point is 0.1819, which belongs to the lower median area of ​​the recommended system linkage weight interval [0.1, 0.3] and is suitable for adjusting low-intensity local offsets. The formula introduces the spatial difference Distance decay factor The product of is used to construct a quantitative assessment of the degree of influence of real space deformation, and combined with a unified normalized denominator design to avoid the imbalance caused by extremely small or extremely large weight factors, forming a robust linkage weighted structure in the process of spatial signal compensation.

[0156] like Figure 2 and Figure 7 As shown, the template link update module includes:

[0157] The parameter structure reconstruction submodule determines the spatial distribution of each grid point coordinate based on the linkage spatial compensation data, calculates the structural correspondence between adjacent grid points, adjusts the spatial parameter arrangement order of each group of grid points, optimizes the spatial topological layout, and generates structural mapping parameters;

[0158] Traverse all grid point coordinates, extract the final position data of each grid point after compensation adjustment, and then perform the attribution judgment operation on each grid point, classify it into the grid area of ​​its corresponding row and column and record its relative row and column number. On this basis, check the arrangement position of the grid points in the whole image, and judge whether there are spatial structural anomalies such as cross dislocation, linear offset or matrix collapse by comparing the horizontal and vertical position change trends of two adjacent grid points. At the same time, count the relative distance change values ​​between adjacent grid points and analyze whether the change amplitude falls within the tolerance range of the structural correspondence. If the horizontal or vertical spacing change of a group of grid points exceeds the preset structural consistency threshold, it will be marked as a structurally inconsistent combination. The structural consistency threshold is based on the image resolution and the standard template. The average spacing of the grid points is generally set to within 5 pixels. The relative offset trend of all grid point combinations marked as abnormal is recorded, and the grid points are rearranged in the order of row and column numbers. The combinations with the same offset direction are placed in front or back. The coordinates of each group of grid points are reordered to ensure that the order of grid points in the same row or column meets the increasing or decreasing arrangement requirements. At the same time, the spatial identification index table is updated. Then, according to the adjusted arrangement order of each group of grid points, a set of data structure mapping contents describing the topological relationship between grid points is generated. The structure mapping parameters record the sequential link, arrangement direction and spacing adjustment status between each grid point and its adjacent grid points, and output the spatial topology reconstruction data of all grid points as the input for subsequent cache structure comparison.

[0159] The cache data replacement submodule compares the structure mapping parameters with the structure parameters in the current calibration system cache, analyzes the differences in the stored content, adjusts the inconsistent grid coordinates and mapping relationships, updates the structure parameters, and obtains the template parameter integration;

[0160] Call the structure parameter table stored in the current calibration system cache, compare the cache parameters with the grid position and link relationship in the latest structure mapping one by one, use the grid number as the index, retrieve the coordinates and adjacent structure description of each grid point in the current cache, and then compare them with the grid data of the corresponding number in the reconstructed mapping parameters one by one. First, compare the horizontal and vertical coordinate values. If the difference exceeds 2 pixels, it is recorded as a position change. Then compare it with the link direction and index order of the four adjacent grid points above, below, left, and right. If the link relationship is inconsistent or missing, it is marked as a topological relationship change. For all changed grid points The number and change type are automatically recorded, and the proportion of changed grid points is counted. When the number of changes exceeds 15% of the total number of grid points, a cache update operation is triggered. Subsequently, all grid points marked as position or relationship changes are removed from the cache, and the newly generated coordinate data and link parameters are written to the corresponding address of the cache location. After the update operation is completed, the full-map grid structure index table is rebuilt, and a difference mapping log is generated for the new and old cache data. Finally, the template parameter integration is obtained. This data set contains the latest coordinates, link structure and number index of all grid points in the current image structure, forming a unified structure description configuration set.

[0161] The chain configuration generation submodule is based on template parameter integration, divides the boundary area in the grid template, determines the chain pointing order between each grid point, combines the parameter path of each grid point, integrates the chain list structure, and outputs the template space update configuration;

[0162] Extract the spatial arrangement structure of all grid points in the complete grid template, identify the grid points at the edge of the image or the boundary of the region, and divide the grid points into boundary patch identification areas. The division operation is completed by judging whether the horizontal and vertical coordinates of the grid points are close to the image boundary or whether they belong to the intersection area of ​​adjacent grid segments. Usually, the grid points within 30 pixels of the edge of the image are taken as the boundary candidate areas, and then the grid points with obvious connection change characteristics are selected from the candidates as the boundary patch nodes. Then, the four-way link parameters are extracted from the data structure of each grid point to judge whether there is a valid link target on its left, right, above or below. If so, the corresponding link target grid point number is recorded. At the same time, according to the location of the grid point, the four-way link parameters are extracted from the data structure of each grid point. The grid arrangement order is used to establish a grid point chain pointing path, and the link sequence of each grid point is written into the chain structure table. For example, if a grid point is numbered A, its right side is linked to numbered B, and its bottom side is linked to numbered C, then it is recorded in the chain structure as A→B, A→C. The chain link sequence of all grid points is constructed in sequence. After all grid point paths are constructed, all path tables are integrated, and the parameter path of each grid point is organized in a chain data structure. The path structure contains the coordinates of the point, the link node number, the link direction mark, and the flag bit of whether it is a boundary node. The complete template space update configuration is output, which is used to guide the start and execution of the next round of image processing and structure correction process.

[0163] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the associated objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0164] In this disclosure, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, "at least one of a, b, or c" can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or plural.

[0165] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0166] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0167] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0168] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of the device or unit, which can be electrical, mechanical or other forms.

[0169] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of the solution of this embodiment according to actual needs.

[0170] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0171] If the functions are implemented as software functional units and sold or used as independent products, they 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 the portion 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 methods described in various embodiments of the present invention. The aforementioned storage media include various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROM), random access memories (RAM), magnetic disks, or optical disks.

[0172] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A layered grid image assisted calibration system, characterized in that: The system comprises: The brightness intersection screening module detects the brightness changes of each pixel in the horizontal and vertical directions based on the pixel brightness data of the original image, compares the difference between the pixel and the neighborhood mean, screens the intersection candidates according to the change standard, calculates the distribution density in the area, removes noise points whose density does not meet the standard, and obtains the grid intersection density feature; The node synchronization correction module extracts spatial coordinates based on the dense grid intersection features, measures the Euclidean distance of node pairs, determines consistency according to the synchronization standard, marks unsynchronized nodes, calculates spatial offsets, and merges the offsets based on the node hierarchy and cross-slice relationship to obtain spatial synchronization compensation parameters. The reference difference extraction module extracts the coordinates of the reference points in the grid structure based on the spatial synchronization compensation parameters, obtains the spatial distances and angles between all reference points, compares them with the template parameters, identifies the spatial distribution difference intervals between the combinations, and obtains the reference feature offset amplitude; The spatial offset allocation module calls the current coordinates of all reference points based on the offset amplitude of the reference feature, combines the offset of each point with the spatial distribution factor, allocates the offset to the corresponding grid coordinates, and obtains the linkage spatial compensation data.

2. The layered grid image-assisted calibration system according to claim 1, characterized in that: The grid intersection density features include intersection concentration, intersection distribution uniformity, and regional brightness contrast; the spatial synchronization compensation parameters include synchronization adjustment amplitude, pairing offset, and node distribution weight; the benchmark feature offset amplitude includes spatial consistency index, feature point offset, and benchmark point spatial distribution ratio; the linkage space compensation data includes coordinate correction factor, grid point compensation amplitude, and linkage correlation coefficient.

3. The layered grid image-assisted calibration system according to claim 1, characterized in that: The brightness intersection screening module includes: The brightness change extraction submodule analyzes the original image pixel brightness data, calculates the brightness difference between each pixel and its adjacent pixels in the horizontal and vertical directions, determines the brightness fluctuation, and obtains the fluctuation intensity parameter based on the characteristics of the adjacent pixel group. The intersection candidate screening submodule screens key pixel points based on the fluctuation intensity parameter, compares their distribution characteristics, determines the pixel positions that meet the intersection characteristics, classifies the pixels that meet the requirements, and obtains feature aggregation coordinates; The candidate density calculation submodule calculates the pixel density distribution of each partition according to the feature aggregation coordinates, determines the aggregation state of candidate pixels in each area, eliminates and sorts areas that do not meet the distribution rules, and obtains grid intersection density features.

4. The layered grid image-assisted calibration system according to claim 1, characterized in that: The node synchronization correction module includes: The node pairing construction submodule groups all intersections according to a hierarchical structure based on the dense grid intersection features, combines the node spatial coordinates, constructs node pairing combinations of different levels, screens nodes with associated relationships, and establishes a hierarchical pairing structure to obtain a hierarchical combination index; The spatial consistency judgment submodule compares the spatial interval of each group of nodes based on the hierarchical combination index, analyzes the corresponding spatial distribution status of each node pair, determines whether the synchronization standard is met, marks the offset status of inconsistent paired nodes, and obtains the pairing space offset parameter set; The offset merging calculation submodule matches the hierarchical area and intersection area corresponding to each group of nodes according to the paired spatial offset parameter set, analyzes the consistency of multiple groups of offset directions in the intersection position, merges the associated offsets according to regional affiliation, and obtains the spatial synchronization compensation parameters.

5. The layered grid image-assisted calibration system according to claim 1, characterized in that: The benchmark difference extraction module includes: The coordinate data acquisition submodule collects the spatial position of each reference point in the grid structure based on the spatial synchronization compensation parameters, determines the ownership relationship of the nodes in the hierarchical structure, locates the reference nodes according to the current hierarchical structure, and obtains a reference position set; The spatial parameter calculation submodule calculates the spatial distance between each reference point based on the reference position set, selects any three points to form a vector for angle analysis, and combines the distance and angle statistical parameters to determine the spatial distribution state of the node combination and obtain the spatial characteristic factor; The geometric difference determination submodule compares the spatial characteristic factors with the standard parameters in the template structure, analyzes the variation range of the combination in spatial distribution and angle characteristics, screens the combination interval with differences, and obtains the reference feature offset amplitude.

6. The layered grid image-assisted calibration system according to claim 5, characterized in that: The comparison with the standard parameters in the template structure is done using the formula: ； Calculate the spatial difference value, filter the combination interval with differences, and obtain the benchmark feature offset amplitude, where: Representative reference point and The spatial difference between Represents the current reference point and The Euclidean distance between Represents the reference point in the template structure and The theoretical Euclidean distance between Represents the current reference point The angle formed, Represents the reference point in the template structure The theoretical angle formed by Represents the current reference point Brightness contrast factor of the three-point combination.

7. The layered grid image-assisted calibration system according to claim 1, characterized in that: The spatial offset allocation module includes: The coordinate joint arrangement submodule arranges the current coordinates of all reference points based on the offset amplitude of the reference feature, determines the spatial ownership of each point in the grid structure, uniformly numbers and arranges the coordinate and offset information, and obtains the node ownership sequence based on the image grid distribution; The distribution factor construction submodule calculates the spatial distance relationship between each benchmark point and the surrounding grid points based on the node attribution sequence, analyzes the spatial linkage between the nodes and the grid points, determines the influence of each benchmark point in the local area, and obtains the influence factor; The offset weight distribution submodule performs a linear weighted distribution operation based on the influencing factors and the offset amplitude and distribution factor of each grid point associated with the reference point, distributes the offset signal to the target grid point, and obtains the linkage space compensation data.

8. The layered grid image-assisted calibration system according to claim 7, characterized in that: The offset amplitude and distribution factor for each grid point associated with the reference point are calculated using the formula: ； Calculate the grid point linkage compensation weight value, distribute the offset signal to the target grid point, and obtain the linkage space compensation data, where: Representative The linkage compensation weight value of each target grid point, Representatives and The grid point associated The spatial coordinate difference value of the reference point relative to the template standard parameter, Representative The reference point is relative to the The spatial distribution factor of the grid points, Representatives and The number of reference points associated with each grid point.

9. The layered grid image-assisted calibration system according to claim 1, characterized in that: The system further comprises: The template link update module reconstructs the template space parameters based on the linkage space compensation data and the coordinates of each grid point, replaces the calibration system cache data, builds a linked list structure according to the grid structure template and the boundary alignment fragment, and outputs the template space update configuration; The template space update configuration includes template structure parameters, space mapping relationship, and boundary alignment data.

10. The layered grid image-assisted calibration system according to claim 9, characterized in that: The template link update module includes: The parameter structure reconstruction submodule determines the spatial distribution of each grid point coordinate based on the linkage spatial compensation data, calculates the structural correspondence between adjacent grid points, adjusts the spatial parameter arrangement order of each group of grid points, optimizes the spatial topological layout, and generates structural mapping parameters; The cache data replacement submodule compares the structure mapping parameters with the structure parameters in the current calibration system cache, analyzes the differences in the stored contents, adjusts the inconsistent grid coordinates and mapping relationships, updates the structure parameters, and obtains the template parameter integration; The chain configuration generation submodule divides the boundary area in the grid template based on the template parameter integration, determines the chain pointing order between each grid point, combines the parameter path of each grid point, integrates the chain list structure, and outputs the template space update configuration.

Images