An indoor situation generation method, device and medium combining thin plate spline interpolation and natural neighbor interpolation

Abstract

The application discloses an indoor situation generation method and device combining thin-plate spline interpolation and natural neighborhood interpolation and a medium, and relates to the field of electric digital data processing.The method comprises the following steps: acquiring original sampling data; performing grid division, merging and extracting data falling into the same grid to obtain a first sample data set; generating preliminary interpolation data by using thin-plate spline interpolation; extracting supplementary sampling point data from the preliminary interpolation data in an area not covered by the original sampling data to obtain a supplementary sample data set; performing grid division and merging to obtain a second sample data set; combining the supplementary sample data set and the second sample data set, filtering the supplementary data, and obtaining a target sample data set; and generating final situation data by using natural neighborhood interpolation.The application can effectively improve the situation generation quality under the condition of uneven sample distribution by filling in sparse area samples through thin-plate spline interpolation and finely processing the samples through natural neighborhood interpolation.

Description

Technical Field

[0001] This invention relates to the field of electronic digital data processing technology, and in particular to radio spectrum situation generation, specifically an indoor situation generation method, device and medium that combines thin plate spline interpolation and natural neighbor cell interpolation. Background Technology

[0002] With the rapid development of mobile communication, the Internet of Things, and various short-range wireless communication technologies, the electromagnetic environment in indoor spaces has become increasingly complex. In various indoor scenarios such as offices, conference rooms, shopping malls, underground parking lots, and subway stations, numerous wireless access points, user terminals, Bluetooth devices, and RFID devices operate simultaneously, resulting in high spectrum resource reuse and widespread signal superposition and interference. Accurately acquiring the signal strength distribution at various locations within an indoor target area—that is, generating indoor spectrum situational awareness—is an indispensable foundation for many applications, including wireless network planning and optimization, rapid identification of interference sources, dynamic allocation of spectrum resources, and precise location of illegal transmission sources.

[0003] The generation of indoor spectral situation data is typically based on the deployment of fixed monitoring nodes or the collection of spectral data along planned paths. However, due to limitations in indoor building layout, equipment deployment costs, and personnel activity range, sampling points often only cover partial areas. Easily accessible locations such as corridor entrances have dense sample collections, while corners and areas behind obstacles have little or no sampling data. This results in a significant sparse and uneven spatial distribution of samples, making it impossible to directly form a complete situation image from the collected data. Data completion techniques must be used to infer values ​​at unknown locations.

[0004] Data completion essentially involves inferring attribute values ​​for unknown locations based on the spatial distribution patterns of existing discrete samples, falling under the category of spatial interpolation. Existing methods include inverse distance weighting, Kriging interpolation, and propagation model correction. Inverse distance weighting tends to shift predictions towards densely sampled areas when samples are not uniformly distributed; Kriging requires a sufficient number of samples to fit the semivariogram, making it difficult to establish robust statistical relationships in sparse regions; and propagation model-based methods often suffer from discrepancies between theoretical attenuation formulas and actual measurements due to severe multipath reflection and diffraction indoors. These conventional methods struggle to balance global trends and local fluctuations in scenarios with both missing and dense samples.

[0005] Therefore, there is an urgent need for a situation generation scheme that can adapt to the uneven distribution of indoor data collection, effectively handle blank areas, and maintain local accuracy. Summary of the Invention

[0006] This invention overcomes the shortcomings of the prior art and provides an indoor situation generation method, equipment and medium that combines thin plate spline interpolation and natural neighbor interpolation.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: Firstly, the present invention provides an indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation, comprising the following steps:

[0008] S1. Obtain the original sampling data within the target area, wherein the original sampling data includes the location information of the sampling points and their corresponding signal strength values;

[0009] S2. Divide the target area into grids according to the first spatial resolution, merge the original sampled data falling into the same grid, extract the merged data of each grid, and obtain the first sample dataset.

[0010] S3. Based on the first sample dataset, use the thin plate spline interpolation method to generate preliminary interpolation data covering the target area;

[0011] S4. According to the preset sampling and extraction rules, in the area of ​​the target area that does not cover the original sampling data, supplementary sampling point data is extracted from the preliminary interpolation data to obtain a supplementary sample dataset;

[0012] S5. Divide the original sampling data into grids according to the second spatial resolution, and merge the original sampling data falling into the same grid to obtain the second sample dataset;

[0013] S6. Merge the supplementary sample dataset with the second sample dataset, and filter the supplementary sample data that are close in location and have signal strength differences exceeding a preset threshold to obtain the target sample dataset.

[0014] S7. Based on the target sample dataset, the target region is interpolated using the natural neighborhood interpolation method to generate the final situational data of the target region.

[0015] In a preferred embodiment of the present invention, the first spatial resolution is greater than the second spatial resolution, wherein the first spatial resolution is 5 m-20 m and the second spatial resolution is 0.5 m-2 m.

[0016] In a preferred embodiment of the present invention, in step S2, the merging process is as follows: for each grid, the maximum value of the signal intensity values ​​of all sampling points in the grid is taken as the merged signal intensity value of the grid, and the geometric center position of the grid is taken as the spatial position corresponding to the merged signal intensity value.

[0017] In a preferred embodiment of the present invention, in step S4, the preset sampling extraction rule includes: scanning each spatial position row by row and column by column in the area not covered by the original sampling data with a third spatial resolution as the step size; for each scanned position, calculating the Euclidean distance between the position and the nearest original sampling point; if the Euclidean distance is greater than a preset distance threshold, then extracting the signal strength value corresponding to the position in the preliminary interpolation data as the signal strength value of the supplementary sampling point, and using the position as the location information of the supplementary sampling point.

[0018] In a preferred embodiment of the present invention, in step S6, the filtering process specifically involves: for each supplementary sampling point, calculating the Euclidean distance between the location of the supplementary sampling point and the locations of each original sampling point in the second sample dataset; if the minimum Euclidean distance is less than a preset conflict distance threshold, and the absolute value of the difference between the signal strength value of the supplementary sampling point and the signal strength value of the corresponding nearest original sampling point is greater than a preset difference threshold, then the supplementary sampling point is removed from the merged dataset.

[0019] In a preferred embodiment of the present invention, the thin plate spline interpolation method includes: constructing an interpolation function based on radial basis functions and global affine transformation; determining the weight coefficients and affine transformation coefficients of the interpolation function by solving a system of linear equations composed of sample point constraints and affine transformation constraints; thereby predicting the signal intensity at any location within the target region.

[0020] In a preferred embodiment of the present invention, the natural neighborhood interpolation method includes:

[0021] Construct a first Voronoi diagram based on all sample points in the target sample dataset;

[0022] For the location to be interpolated, the location to be interpolated is added as a new point to the target sample dataset, a second Voronoi diagram is constructed, and the neighboring sample points of the location to be interpolated are determined;

[0023] Calculate the area ratio that the Voronoi cell corresponding to the position to be interpolated takes from the Voronoi cells corresponding to each neighboring sample point, and use this area ratio as the interpolation weight of each neighboring sample point.

[0024] The signal strength values ​​of each neighboring sample point are weighted and summed according to the interpolation weight to obtain the predicted signal strength value of the position to be interpolated.

[0025] In a preferred embodiment of the present invention, the original sampling data further includes sampling time and frequency information. When there are multiple sampling values ​​at the same location and frequency, the arithmetic mean of the multiple sampling values ​​at the same location is taken as the representative value of the signal strength at that location and frequency.

[0026] In a second aspect, embodiments of the present invention provide an electronic device, comprising: at least one processor; and a memory communicatively connected to at least one of the processors;

[0027] The memory stores a computer program executed by at least one of the processors, which enables the at least one processor to perform the indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation as described above.

[0028] Thirdly, embodiments of the present invention provide a computer-readable storage medium storing computer instructions, which are used to cause a processor to execute the indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation as described above.

[0029] This invention addresses the shortcomings of the prior art and has the following beneficial effects:

[0030] (1) This invention provides an indoor situation generation method that combines thin-plate spline interpolation and natural neighbor interpolation. By first generating preliminary interpolation data covering the entire area using thin-plate spline interpolation, and then extracting supplementary sampling points in uncovered areas from the data, and performing fine processing using natural neighbor interpolation, a synergistic mechanism of global situation supplementation and local refinement can be formed. This fully utilizes the extrapolation capability of thin-plate splines in areas with sparse samples to fill spatial gaps, while leveraging the precise weighting advantage of natural neighbor interpolation in local areas without relying on a preset model. Compared to the interpolation blind zone caused by insufficient neighboring samples when using natural neighbor interpolation alone, this invention can generate continuous situation images in areas with missing samples. Furthermore, compared to the non-physical fluctuations caused by long-distance interference from dense samples when using thin-plate spline interpolation alone, this invention corrects the extrapolation bias through subsequent local weighting, thereby ensuring both the spatial coverage integrity and numerical rationality of the final situation.

[0031] (2) This invention sets a first spatial resolution for coarse-grained raster merging before thin plate spline interpolation and a second spatial resolution for fine-grained raster merging before natural neighborhood interpolation. The two resolutions independently control the sample density of their respective stages. Coarse-grained merging effectively weakens the excessive constraint of the dense area of ​​the original sampling points on the global surface of the thin plate spline, and avoids the interpolation surface shape being dominated by local high-density samples, thereby enabling the thin plate spline to better capture the overall trend of the entire region. Fine-grained merging retains the details of local spatial changes in the original sampling data, providing sufficiently fine benchmark samples for natural neighborhood interpolation. The decoupling method of this invention can overcome the contradiction between global trend extraction and local detail preservation under a single resolution, so that the two interpolation stages can each operate at the optimal scale, thereby improving the adaptability of the overall situation generation.

[0032] (3) This invention introduces a dual filtering condition based on spatial distance and signal intensity difference during the merging of supplementary samples and original samples. Supplementary samples that are geographically close but have excessive intensity deviations are removed. This mechanism is designed to address the prediction bias that may occur when thin-plate spline interpolation extrapolates at the edge of dense sample areas. When the numerical difference between the supplementary sample and the nearby real sampling points exceeds a reasonable range, it indicates that the supplementary sample is not reliable enough. Filtering it can prevent erroneous data from contaminating the input sample set for subsequent natural neighborhood interpolation. Compared to indiscriminately including all extrapolated data into the interpolation system, this invention can ensure the data quality of the reference sample through conflict filtering, so that the natural neighborhood interpolation in the supplementary area still uses the real measurement data as a reliable reference, thereby maintaining the consistency between the interpolation results and the actual measurement environment. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0034] Figure 1 This is a flowchart illustrating the indoor situation generation method of the present invention that combines thin plate spline interpolation and natural neighbor interpolation;

[0035] Figure 2 This is a diagram of the original electromagnetic situation obtained from the simulation of this invention;

[0036] Figure 3 This is a sample data distribution diagram obtained after sampling the simulation situation data according to the present invention;

[0037] Figure 4 This is a schematic diagram of the distribution of the original spectrum data collected in an indoor scene according to the present invention;

[0038] Figure 5 This is an indoor spectrum situation diagram generated by the method proposed in this invention;

[0039] Figure 6 This is a comparison diagram of the electromagnetic situation reconstructed by the ordinary natural neighborhood method and the proposed method, wherein (a) is the reconstruction result of the ordinary natural neighborhood method and (b) is the reconstruction result of the proposed method.

[0040] Figure 7 A schematic diagram of an electronic device structure that can be used to implement the method proposed in this invention is shown.

[0041] In the diagram: 10. Electronic device; 11. Processor; 12. Read-only memory; 13. Random access memory; 14. Bus; 15. I / O interface; 16. Input unit; 17. Output unit; 18. Storage unit; 19. Communication unit. Detailed Implementation

[0042] 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, and 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.

[0043] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein. Therefore, the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0044] Application Overview:

[0045] To address the challenge of data completion, researchers in this field have gradually introduced natural neighborhood interpolation into spectral situation generation. This method, based on Thiessen polygons, assigns weights to neighboring samples by varying the polygon area after the insertion point is added. It requires no pre-defined statistical model, and in regions with relatively uniform sample distribution, the interpolation results can effectively reflect local changes, exhibiting strong adaptability and computational stability. However, the predictive ability of this method is highly dependent on the existence of sufficient neighboring sample points around the predicted location. If a large area of ​​missing samples exists within the target region, the insertion point lacks sufficient neighboring samples to construct an effective polygon intersection, making the interpolation process impossible and turning the blank areas into processing blind spots.

[0046] Meanwhile, thin-plate spline interpolation has also been introduced into this field as an alternative. This method analogizes the interpolation surface to the bending deformation of a thin plate under the action of support points, constructing a smooth function across the entire region by minimizing bending energy. It has strong extrapolation capabilities when dealing with regions lacking samples, requiring only a small number of samples to extrapolate the approximate trend of the blank areas. However, practical applications show that the thin-plate spline interpolation surface is constrained by all sample points. When local samples are dense while other regions are sparse, the dense points strongly influence the global morphology, causing the extrapolated values ​​for sparse regions to deviate from the true propagation characteristics. Furthermore, the predicted values ​​for locations far from the sample points often exhibit non-physical fluctuations due to the characteristics of the basis function, resulting in insufficient reliability.

[0047] Therefore, neither natural neighborhood nor thin-plate splines alone can adequately address the complex non-uniformity problem of both dense and blank samples in indoor scenes. Based on this, this invention organically integrates the global extrapolation capability of thin-plate spline interpolation with the local precise weighting capability of natural neighborhood interpolation.

[0048] It should be noted that natural neighborhood interpolation is suitable for areas with uniform sample distribution and sufficient density, and its interpolation accuracy increases with the number of neighboring samples. Thin plate spline interpolation, on the other hand, is suitable for scenarios with sparse samples but needing to maintain the overall trend extrapolation. Its output is constrained by the global samples. The local weighting mechanism of natural neighborhood interpolation and the global minimum curvature mechanism of thin plate spline interpolation are fundamentally different in mathematical principle. The former is highly sensitive to local sample density, while the latter is significantly affected by distant samples. The two are naturally contradictory in terms of locality and globality. Direct combination will inevitably lead to discontinuity in the values ​​of the supplemented data and the original samples at the boundary, or even physical contradictions. How to define the reliability range of the supplemented area, choose the sampling granularity, and set the conflict filtering conditions requires a systematic design based on the physical characteristics of signal strength attenuation with distance and abrupt changes caused by obstacles in indoor electromagnetic propagation.

[0049] Based on this, the present invention proposes an indoor situation generation method that combines thin-plate spline interpolation and natural neighbor interpolation. By setting different spatial resolutions, coarse-grained trend extraction and fine-grained local interpolation are decoupled, and a dual conflict filtering mechanism based on distance and signal strength differences is introduced, so that the two can form a synergy of global potential compensation and local refinement. Experimental data also show that this combination not only visually fills in the blank areas, but also reduces the reconstruction error by an average of about 0.5 dB compared with the conventional natural neighbor method, and the improvement is significantly improved as the sample sparsity increases.

[0050] Example 1:

[0051] like Figure 1 As shown, an indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation includes the following steps:

[0052] S1. Obtain the raw sampling data within the target area. The raw sampling data includes the location information of the sampling points and their corresponding signal strength values.

[0053] S2. Divide the target area into grids according to the first spatial resolution, merge the original sampled data falling into the same grid, extract the merged data of each grid, and obtain the first sample dataset.

[0054] S3. Based on the first sample dataset, use the thin plate spline interpolation method to generate preliminary interpolation data covering the target area;

[0055] S4. According to the preset sampling and extraction rules, in the areas of the target area that do not cover the original sampling data, supplementary sampling point data are extracted from the preliminary interpolation data to obtain a supplementary sample dataset.

[0056] S5. Divide the original sampling data into grids according to the second spatial resolution, and merge the original sampling data that fall into the same grid to obtain the second sample dataset.

[0057] S6. Merge the supplementary sample dataset with the second sample dataset, and filter the supplementary sample data that are close in location and have signal strength differences exceeding a preset threshold to obtain the target sample dataset.

[0058] S7. Based on the target sample dataset, the target area is interpolated using the natural neighborhood interpolation method to generate the final situational data of the target area.

[0059] The concept of this invention lies in leveraging the extrapolation capability of thin-plate spline interpolation in sparse sample regions to first generate trend-based interpolation data covering the entire region. Supplementary sampling points for missing regions are then extracted from this data. These supplementary samples are then merged with the original samples and refined using natural neighborhood interpolation. The two interpolation methods are not simply concatenated; rather, different spatial resolutions are set to control coarse-grained trend extraction and fine-grained local interpolation, respectively. A conflict filtering mechanism removes supplementary samples with insufficient reliability, ensuring that the merged sample dataset fills spatial gaps without introducing erroneous information contradicting the measured data. This fully utilizes the strong fitting capability of thin-plate splines for global smooth trends and the precise weighting capability of natural neighborhoods for local sample distribution, forming a synergy between global trend supplementation and local refinement. This effectively solves the problem of incomplete and inaccurate trend generation caused by extremely uneven sample distribution in indoor scenes.

[0060] Example 2:

[0061] like Figure 1 As shown, the indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation includes the following steps:

[0062] S1. Obtain the raw sampling data within the target area.

[0063] Specifically, the target area is the indoor space where the spectrum situation is to be generated, including but not limited to offices, conference rooms, shopping malls, underground parking lots, rail transit station halls, industrial plants, and various laboratories.

[0064] The raw sampling data is acquired through at least one of the following methods: deploying multiple fixed spectrum monitoring nodes in the target area, with each node continuously collecting signal strength data at its location at preset time intervals; or, having personnel equipped with portable spectrum analyzers move along a planned path within the target area to collect data, recording the spatial coordinates of each sampling point during the collection process using a global positioning system or an indoor positioning system.

[0065] In this embodiment, the original sampling data includes the location information of the sampling points and their corresponding signal strength values. The location information is represented by the horizontal and vertical coordinates in a two-dimensional Cartesian coordinate system. In addition, the original sampling data also includes the sampling time, frequency information, and signal bandwidth corresponding to each sampling point. When there are multiple sampling values ​​at the same location and frequency, the arithmetic mean of the multiple sampling values ​​is taken as the representative signal strength value at that location and frequency.

[0066] Furthermore, the raw sampling data is stored in the electronic device's storage unit in the format of a structured data table. Each record contains six fields: sampling point number, x-axis, y-axis, signal strength value, sampling time, and center frequency. Understandably, the spatial distribution of the raw sampling data is uneven; that is, the sampling point density is higher in open areas with frequent human activity within the target area, while the sampling point density is lower or even completely absent in corners, behind obstacles, and in areas with dense equipment.

[0067] S2. Divide the target area into grids according to the first spatial resolution, merge the original sampled data falling into the same grid, extract the merged data of each grid, and obtain the first sample dataset.

[0068] Specifically, step S2 is used to spatially downsample the original sampled data to extract the overall signal intensity distribution characteristics of the target region at a coarser granular level. In local space, sampling points may be densely packed, meaning that multiple sampling points may be contained within the same grid. These sampling points, due to their close spatial distance and similar propagation environments, exhibit differences in signal intensity values ​​within normal fluctuation ranges. If all sampling points are directly used for thin-plate spline interpolation, the large number of sampling points in dense areas will impose excessive local constraints on the interpolation surface, causing the global shape of the interpolation surface to be dominated by the dense areas and failing to accurately reflect the signal distribution trend in sparse areas. Therefore, it is necessary to merge the sample data to extract the overall characteristics of the local space.

[0069] Furthermore, step S2 is implemented as follows: the target area is divided into multiple square grids with the boundary of the target area as the range and the first spatial resolution as the grid side length. The first spatial resolution is a positive number with a value range of 5 m to 20 m.

[0070] In this embodiment, the first spatial resolution is 10 m. For each grid, all original sampling points within the spatial range covered by the grid are acquired. If the grid contains at least one sampling point, the maximum signal strength value among all sampling points within the grid is taken as the merged data of the grid, and the spatial location corresponding to the merged data is the geometric center of the grid. If the grid does not contain any sampling points, no merged data is generated for the grid. The above maximum value merging operation is performed on all grids containing sampling points to obtain a first sample dataset. Each record in the first sample dataset includes the x-coordinate of the grid's geometric center, the y-coordinate of the grid's geometric center, and the merged signal strength value of the grid.

[0071] It should be noted that the maximum value merging method can preserve the peak characteristics of the signal intensity in the local space, which is of great significance for the subsequent thin plate spline interpolation to obtain the overall envelope trend of the signal coverage in the region. In the indoor propagation environment, the signal intensity fluctuates in the local area due to the multipath effect. Taking the maximum value instead of the average value can effectively avoid the underestimation of the signal caused by local multipath fading, so that the thin plate spline interpolation results can more accurately reflect the upper envelope shape of the signal coverage.

[0072] Alternatively, in other embodiments, the merging process can also be implemented by taking the median or the average.

[0073] S3. Based on the first sample dataset, use the thin plate spline interpolation method to generate preliminary interpolation data covering the target area.

[0074] Specifically, step S3 takes the first sample dataset obtained in step S2 as input and uses the thin-plate spline interpolation method to generate preliminary interpolation data covering the entire target area. The preliminary interpolation data is used to characterize the overall distribution trend of signal intensity within the target area, and its spatial resolution is equal to the second spatial resolution, which is 1 m. The thin-plate spline interpolation method treats the interpolation function as the bending deformation of an infinite thin plate under given support point constraints, and constructs a smooth continuous surface passing through all given data points by minimizing the bending energy.

[0075] Furthermore, the specific implementation process of the thin plate spline interpolation method used in this embodiment is as follows.

[0076] For any spatial location X=(x,y) within the target region, the thin-plate spline interpolation function Φ(X) is constructed as follows:

[0077] .

[0078] Where X is the spatial position vector of the point to be interpolated; x is the x-coordinate of the point to be interpolated; and y is the y-coordinate of the point to be interpolated. Let be the spatial location vector of the i-th sample point in the first sample dataset; n is the total number of sample points in the first sample dataset; Let X be the point to be interpolated and the i-th sample point. The Euclidean distance between them; For the radial basis function of thin plate spline interpolation, when When defined, U(0) = 0; , , These are the global affine transformation coefficients; Let be the weight coefficient corresponding to the i-th sample point.

[0079] Global affine transformation coefficients , , With weighting coefficients This is obtained by solving the following system of linear equations:

[0080] For each sample point Establish the following equation:

[0081] .

[0082] in, This represents the signal strength value of the j-th sample point in the first sample dataset. Let x be the x-coordinate of the j-th sample point. Let be the ordinate of the j-th sample point.

[0083] To eliminate the underdetermined problem, add the following three constraints:

[0084] .

[0085] Combining the above n equations with the 3 constraints, we form a (n+3)×(n+3) system of linear equations, as shown below:

[0086] .

[0087] Where K is an n×n matrix, and the element in the i-th row and j-th column is... P is an n×3 matrix, and its i-th row is... w is an n×1 weight vector, and its i-th element is ;a is a 3×1 affine transformation coefficient vector, whose elements are z is an n×1 signal intensity value vector, and its j-th element is... ; 0 is a 3×3 zero matrix; Let P be the transpose of P.

[0088] Solving the above system of linear equations yields the weighting coefficients. and global affine transformation coefficients , , The value of is substituted into the thin-plate spline interpolation function Φ(X) to calculate the signal strength interpolation result for any spatial location X within the target area. In this embodiment, Gaussian elimination is used to solve the above linear equations.

[0089] In this embodiment, step S3 calculates the interpolation results grid by grid within the target area at the second spatial resolution to obtain preliminary interpolation data covering the entire target area. Figure 2 and Figure 3 The simulation situation result diagram and the sample data distribution diagram obtained after sampling are shown respectively. Figure 2 The continuous color levels in the image represent the original electromagnetic state obtained from the simulation. Figure 3 The discrete dots in the diagram represent the sample data points used for interpolation after sampling.

[0090] Understandably, the thin-plate spline interpolation method minimizes bending energy, allowing the interpolation surface to remain smooth overall while passing through all given sample points. This enables the method to extrapolate data outward based on the trend of boundary sample points in areas with missing samples, providing a preliminary data source that reflects the overall distribution characteristics of the entire region for subsequent steps.

[0091] S4. According to the preset sampling and extraction rules, in the areas of the target area that do not cover the original sampling data, supplementary sampling point data are extracted from the preliminary interpolation data to obtain a supplementary sample dataset.

[0092] Specifically, step S4 is used to extract data points from the preliminary interpolation data generated in step S3 to fill the spatial gaps in the original sampling data. The target area for extracting supplementary sampling point data is the area not covered by the original sampling data, that is, the area within the target area consisting of spatial locations where the distance from all original sampling points is greater than a preset distance threshold.

[0093] Furthermore, the preset sampling and extraction rules are as follows: Within the target area, all grid positions are scanned row by row and column by column with the third spatial resolution as the step size. For each scanned position, the Euclidean distance between the position and the nearest original sampling point is calculated. If the distance is greater than the preset distance threshold, the position is identified as an extraction position in the uncovered area. The signal strength value corresponding to the position in the preliminary interpolation data is used as the signal strength value of the supplementary sampling point, and the spatial coordinates of the position are used as the position information of the supplementary sampling point, thereby generating a supplementary sampling point data record.

[0094] Preferably, the preset distance threshold ranges from 2 m to 5 m. In this embodiment, the preset distance threshold is 3 m. The third spatial resolution is an integer multiple of the second spatial resolution. In this embodiment, the third spatial resolution and the second spatial resolution are equal, both being 1 m.

[0095] Furthermore, extraction operations are performed on all scan locations that meet the above conditions, and all generated supplementary sampling point data records are summarized to form a supplementary sample dataset. It should be noted that the principle for setting the sampling extraction rules in step S4 is: in areas where the original sampling data is missing, supplementary sampling points are extracted with an appropriate spatial density so that the spatial distribution of the supplementary sampling points can reflect the variation characteristics of the preliminary interpolation data in the missing areas, while avoiding redundant data due to excessive extraction density.

[0096] The extraction density is controlled by the third spatial resolution. A smaller third spatial resolution results in denser extraction of supplementary sampling points and a larger amount of information in the supplementary sample dataset, but also increases data redundancy. Conversely, a larger third spatial resolution results in sparser extraction of supplementary sampling points, improving computational efficiency but potentially losing details of local variations in the initial interpolation data. Experiments have shown that a third spatial resolution of 1 m–5 m achieves a good balance between information integrity and data redundancy.

[0097] S5. Divide the original sampling data into grids according to the second spatial resolution, and merge the original sampling data that fall into the same grid to obtain the second sample dataset.

[0098] Specifically, step S5 is used to perform rasterization and merging on the original sampled data to obtain a benchmark sample dataset suitable for natural neighborhood interpolation. Step S5 is implemented as follows: the target region is divided into multiple square grids with the boundary of the target region as the range and the second spatial resolution as the grid side length.

[0099] Furthermore, the second spatial resolution is a positive number, ranging from 0.5 m to 2 m. In this embodiment, the second spatial resolution is 1 m. For each grid, all original sampling points within the spatial range covered by the grid are acquired. If the grid contains at least one sampling point, the signal strength values ​​of all sampling points within the grid are merged.

[0100] In this embodiment, the merging process uses the maximum value method, that is, the maximum signal strength value among all sampling points within the grid is taken as the merged signal strength value of the grid, and the spatial location corresponding to the merged signal strength value is the geometric center position of the grid; if the grid does not contain any sampling points, no merged data is generated for that grid. The above merging operation is performed on all grids containing sampling points to obtain a second sample dataset. Each record in the second sample dataset includes the x-coordinate of the grid's geometric center, the y-coordinate of the grid's geometric center, and the merged signal strength value of the grid. Based on the Nyquist sampling theorem and empirical data of indoor propagation environments, when the grid side length is within the range of half to one wavelength of the signal wavelength, rasterization can effectively eliminate random fluctuation components in the sampled data while retaining spatial variation information reflecting the essential characteristics of the propagation environment. For the common 2.4 GHz frequency band indoors, the signal wavelength is approximately 0.125 m. Considering the balance between computational efficiency and data accuracy in practical engineering applications, a second spatial resolution of 1 m to 2 m is preferred. In this embodiment, 1 m is used.

[0101] S6. Merge the supplementary sample dataset with the second sample dataset, and filter the supplementary sample data that are close in location and have signal strength differences exceeding a preset threshold to obtain the target sample dataset.

[0102] Specifically, step S6 merges the supplementary sample dataset generated in step S4 with the second sample dataset generated in step S5, and handles any conflicts in the merged data. A conflict here refers to a situation where the distance between a supplementary sampling point in the supplementary sample dataset and the location of an original merged sampling point in the second sample dataset is less than a preset conflict distance threshold, and the absolute value of the difference between their signal strength values ​​is greater than a preset difference threshold. In this case, the data from the supplementary sampling point is considered to conflict with the original sampled data and is filtered out.

[0103] Further, step S6 is implemented as follows: All records in the supplementary sample dataset and the second sample dataset are merged into a temporary sample dataset. For each supplementary sampling point record in the temporary sample dataset, the Euclidean distance between the location of the supplementary sampling point and the locations of all original merged sampling points in the second sample dataset is calculated. The minimum Euclidean distance value is obtained. If the minimum Euclidean distance value is less than a preset conflict distance threshold, and the absolute value of the difference between the signal strength value of the supplementary sampling point and the signal strength value of the corresponding nearest original merged sampling point is greater than a preset difference threshold, then the supplementary sampling point record is deleted from the temporary sample dataset. After performing the above judgment and filtering operations on all supplementary sampling point records, all remaining records in the temporary sample dataset constitute the target sample dataset. The preset conflict distance threshold ranges from 1 m to 3 m. In this embodiment, the preset conflict distance threshold is 2 m. The preset difference threshold ranges from 5 dB to 10 dB. In this embodiment, the preset difference threshold is 6 dB.

[0104] It should be noted that the significance of setting up a conflict filtering mechanism in step S6 is that the interpolation results of thin-plate spline interpolation in dense sample regions are strongly constrained by the dense sample points, resulting in high interpolation accuracy within these regions. However, the interpolation results in sparse sample regions mainly rely on extrapolation, which has relatively low reliability. When the data extrapolated from the edge of the original sample coverage area by thin-plate spline interpolation differs significantly from the original sample data near the boundary, this extrapolated data reflects the prediction bias generated by thin-plate spline interpolation during the transition from dense to sparse regions. Its reliability is insufficient to serve as input samples for natural neighborhood interpolation. Therefore, it is removed through the conflict filtering mechanism to avoid erroneous data contaminating subsequent interpolation results.

[0105] Optionally, in other embodiments, the number of supplementary sampling points that are filtered out can be counted, and when the filtering out ratio exceeds a preset ratio threshold, a prompt can be made to adjust the value of the first spatial resolution or the third spatial resolution.

[0106] S7. Based on the target sample dataset, the target area is interpolated using the natural neighborhood interpolation method to generate the final situational data of the target area.

[0107] Specifically, step S7 takes the target sample dataset obtained in step S6 as input and uses the natural neighborhood interpolation method to interpolate each grid position within the target area, generating final situational data covering the entire target area. The natural neighborhood interpolation method constructs the weight relationship between neighboring sample points and the point to be interpolated based on Thiessen polygons, and specifically includes the following sub-steps:

[0108] S71. Construct the first Voronoi diagram based on all sample points in the target sample dataset.

[0109] Specifically, for each sample point in the target sample dataset Its corresponding Voronoi unit Defined as the set of all spatial locations X within the target area that satisfy the following condition: X and The Euclidean distance between X and any other sample point is less than or equal to X. The Euclidean distance between them, where j≠i.

[0110] In a formal way, , where Ω represents the target region.

[0111] The Voronoi units of all sample points together constitute the first Voronoi diagram, which divides the target region into multiple non-overlapping polygonal regions, each containing one and only one sample point.

[0112] S72. For each grid position to be interpolated in the target area. ,Will Add it as a new point to the target sample dataset and reconstruct the second Voronoi diagram.

[0113] Specifically, in the second Voronoi diagram... The corresponding Voronoi unit From the first Voronoi diagram and A new polygon formed by "capturing" a portion of the area from several adjacent Voronoi elements. Let... The sample points corresponding to adjacent Voronoi units constitute The set of neighboring sample points, denoted as The number of sample points in the neighboring sample point set is m.

[0114] S73, Calculation The weight of the impact of the insertion on each neighboring sample point.

[0115] Specifically, for each sample point in the set of neighboring sample points ,set up for The corresponding Voronoi element in the first Voronoi diagram is... The corresponding Voronoi element is shown in the second Voronoi diagram. The area of ​​the overlapping portion of the covered area is the sample point. right interpolation weights for:

[0116] .

[0117] in, The Voronoi cell corresponding to the j-th neighboring sample point is The area of ​​the overlapping portion of the covered area, where m is the number of sample points in the set of neighboring sample points, and m is the sum of the weights of all neighboring sample points. .

[0118] S74. Calculate the interpolation grid position by weighting and summing the signal strength values ​​of each neighboring sample point according to the above weights. Final signal strength prediction :

[0119] .

[0120] in, Let i be the signal strength value of the i-th neighboring sample point. is the interpolation weight of the i-th neighboring sample point.

[0121] Furthermore, the above natural neighborhood interpolation calculation is repeated for all grid locations to be interpolated in the target area to generate final situational data covering the entire target area. The final situational data is stored in the form of a grid matrix, and the value of each grid element represents the predicted signal strength at that location.

[0122] In this embodiment, a scan-line algorithm is used to calculate the overlapping area in the Voronoi diagram. This algorithm scans the pixels in the target area line by line, counts the number of pixels covered by each Voronoi unit, and uses pixel count instead of area calculation.

[0123] Alternatively, in other embodiments, a precise area calculation method based on polygon clipping can also be used.

[0124] Figure 4 The diagram shows the distribution of raw spectral data collected in an indoor scene. The blue dots represent the map outline and internal partition structure of the indoor space, while the discrete yellow or light blue dots represent the collected sample data points. It can be seen from the figure that the spatial distribution of the sample data is obviously uneven, that is, the sample points are more dense in the central area, while the corner areas are rarely covered by sample points. Figure 5 The diagram shows the indoor spectrum situation generated using the method of this embodiment. The blue-purple line represents the movement path during data acquisition, the color-filled area represents the generated signal strength situation, the red area represents the signal strength is high, the green area represents the signal strength is low, and the solid circle in the figure represents the candidate point for locating the signal source recommended based on the situation results. The deviation between the position of the candidate point and the actual position of the transmitter is less than 1.5 m.

[0125] contrast Figure 4 and Figure 5 As can be seen, the situational results generated by the method in this embodiment can still form a continuous and smooth signal strength distribution in the corner areas where the original samples are missing, and the spatial variation trend of the signal strength conforms to the basic laws of indoor propagation environment, that is, the closer to the signal source, the higher the signal strength, and the signal strength is attenuated behind the partition wall.

[0126] To further illustrate the present invention, this embodiment verifies the above-mentioned indoor situation generation method combining thin plate spline interpolation and natural neighbor interpolation through simulation experiments.

[0127] The experiment was conducted within a rectangular target area of ​​300 km × 200 km, containing multiple transmitting devices with a transmission power of 10 W operating in the VHF band. The target area was discretized into 1 km × 1 km grids. The signal strength value at each grid location was calculated using the ITU-R P.1546 propagation model recommended by the International Telecommunication Union (ITU), obtaining the coverage situation data of the transmitting devices within the area, which was used as the true reference situation. Based on this true reference situation data, data from a portion of the grid points were extracted according to a preset sampling method to simulate the sparse and unevenly distributed sample data in actual data acquisition. Then, the situation was reconstructed from the simulated sampled data using both conventional natural neighborhood interpolation and the method proposed in this embodiment. The reconstructed results were compared grid-by-grid with the true reference situation, and the reconstruction error was statistically analyzed.

[0128] Furthermore, the experiment employed three different sampling methods for simulated sampling.

[0129] The first sampling method is as follows: five consecutive data points are taken every 30 grid points in both the horizontal and vertical directions. That is, after collecting five grid points in each sampling row, skip 25 grid points and collect five more grid points. The sampling rows are arranged at the same regular intervals in the vertical direction.

[0130] The second sampling method is to take two consecutive data points every 30 grid points in both the horizontal and vertical directions.

[0131] The third sampling method is to take two consecutive data points every 40 grid points in both the horizontal and vertical directions.

[0132] The sample data obtained under all three sampling methods are sparse and unevenly distributed. The first sampling method has the largest total number of sample points, while the third sampling method has the smallest total number of sample points.

[0133] Figure 2 The simulation results show the actual reference situation. Figure 3The diagram shows the distribution of sample data obtained after sampling the simulation situation data using the first sampling method. From Figure 2 As can be seen, in the actual reference situation, the signal strength decreases from the source location outwards, forming a continuous, gradually changing distribution within the target area; from Figure 3 As can be seen, the sample data obtained after sampling forms regular strip-shaped blank areas in space, that is, the sampling points exist only in some rows and some columns, and there are a large number of unsampled blank areas between rows and between columns.

[0134] Figure 6 Images (a) and (b) show the results of situation reconstruction of sample data from the first sampling method using the conventional natural neighborhood interpolation method and the method proposed in this embodiment, respectively. Figure 6 It can be observed that: Figure 6 (a) The conventional natural neighborhood method cannot generate effective interpolation results in the strip-shaped blank area of ​​the sample data, resulting in multiple blanks and breaks. The reason is that the natural neighborhood interpolation method requires a sufficient number of neighboring samples around the point to be interpolated to construct the intersection of Voronoi polygons. In the strip-shaped blank area, the number of neighboring samples of the point to be interpolated is insufficient, and the interpolation equation cannot be solved. Figure 6 (b) The method proposed in this embodiment generates a continuous and complete situational image in the same area, with signal strength exhibiting a smooth, gradual transition in space, without any blank or broken areas. Further comparison... Figure 6 (b) and Figure 2 It can be seen that the situational awareness results generated by the method proposed in this embodiment are consistent with the actual reference situation in terms of overall distribution trend. The boundary positions of the high signal strength region near the source and the low signal strength region far away from the source have a high degree of agreement with the actual reference situation.

[0135] Table 1 presents the reconstruction error statistics for the conventional natural neighborhood interpolation method and the method of this embodiment under three sampling modes. The reconstruction error is calculated as follows: Situational reconstruction is performed on the sample data of each sampling mode using two different methods. Grid points within the overlapping area where both methods can generate valid results are taken as statistical samples. The absolute value of the difference between the reconstructed signal strength value and the true reference signal strength value at each grid point is calculated. The arithmetic mean of the absolute errors of all statistical grid points is taken as the reconstruction error of this method. The statistical results are shown in Table 1.

[0136] Take 5 consecutive data points every 30 grid points. 1.73 1.62 0.11 Take two consecutive data points every 30 grid points. 2.00 1.40 0.60 Take two consecutive data points every 40 grid points. 2.89 2.08 0.81

[0137] The statistical results in Table 1 show that, under all three sampling methods, the reconstruction error of the method proposed in this embodiment is smaller than that of the conventional natural neighborhood interpolation method. Specifically, for sampling method 1, the reconstruction error of the method in this embodiment is 1.62 dB, which is 0.11 dB less than the 1.73 dB of the conventional natural neighborhood method; for sampling method 2, the reconstruction error of the method in this embodiment is 1.40 dB, which is 0.60 dB less than the 2.00 dB of the conventional natural neighborhood method; and for sampling method 3, the reconstruction error of the method in this embodiment is 2.08 dB, which is 0.81 dB less than the 2.89 dB of the conventional natural neighborhood method.

[0138] Further analysis of Table 1 reveals that as the sparsity of the sample data increases, the error reduction of the method proposed in this embodiment compared to the conventional natural neighborhood method shows an increasing trend. In sampling method 1, the continuous sampling point length is 5 grid points with an interval of 30 grid points, resulting in a relatively high sample density. Conventional natural neighborhood methods have sufficient neighboring samples in most areas, and the error difference between the two methods is relatively small. In sampling method 2, the continuous sampling point length is reduced to 2 grid points, significantly reducing the sample density. Conventional natural neighborhood methods fail to interpolate in the blank areas, expanding the region where interpolation fails. However, the method proposed in this embodiment supplements the sample data in the blank areas through thin-plate spline interpolation, effectively maintaining the interpolation accuracy in the blank areas. The error difference between the two methods increases to 0.60 dB. In sampling method 3, the sampling interval is expanded to 40 grid points, further reducing the sample density and increasing the area of ​​the blank areas. Conventional natural neighborhood methods experience a sharp drop in interpolation accuracy in sparse areas, with the error increasing to 2.89 dB. In contrast, the error of the method proposed in this embodiment only increases from 1.40 dB to 2.08 dB, increasing the error difference between the two methods to 0.81 dB.

[0139] In summary, the experimental results show that the combined strategy of first generating supplementary samples using thin-plate spline interpolation in sparse sample regions and then refining the process using natural neighborhood interpolation can effectively overcome the shortcomings of conventional natural neighborhood interpolation methods in generating effective interpolation results in sparse sample regions. At the same time, the conflict filtering mechanism eliminates supplementary samples with insufficient reliability, avoiding the negative impact of erroneous data on the accuracy of natural neighborhood interpolation. This ensures that the reconstructed situation has complete spatial coverage in sample-missing regions and maintains continuity consistent with the real propagation environment.

[0140] Example 3:

[0141] Figure 7A schematic diagram of an electronic device structure that can be used to implement Embodiment 1 of the present invention is shown. The electronic device 10 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0142] like Figure 7 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory 12 or a random access memory 13, which is communicatively connected to the at least one processor 11. The memory stores a computer program that can be executed by the at least one processor 11, and the computer program is executed by the at least one processor 11 to enable the at least one processor 11 to perform the method provided by the present invention.

[0143] Furthermore, the processor 11 can perform various appropriate actions and processes based on the computer program stored in the read-only memory 12 or the computer program loaded from the storage unit 18 into the random access memory 13. The random access memory 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, the read-only memory 12, and the random access memory 13 are interconnected via a bus 14. The I / O interface 15 is also connected to the bus 14.

[0144] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0145] Furthermore, processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as an indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation.

[0146] In some specific embodiments, the indoor situation generation method combining plate spline interpolation and natural neighbor interpolation can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via read-only memory 12 and / or communication unit 19. When the computer program is loaded into random access memory 13 and executed by processor 11, one or more steps of the indoor situation generation method combining plate spline interpolation and natural neighbor interpolation described above can be performed. Alternatively, in other embodiments, processor 11 can be configured by any other suitable means (e.g., by means of firmware) to perform the indoor situation generation method combining plate spline interpolation and natural neighbor interpolation.

[0147] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard parts (ASSPs), systems-on-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0148] Computer programs used to implement the methods of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The computer programs can be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0149] In the context of this invention, a computer-readable storage medium stores computer instructions for causing a processor to execute and implement the indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation provided by this invention. The computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. The computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, the computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory, read-only memory 12, erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0150] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a cathode ray tube (CRT)) or a liquid crystal display (LCD monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0151] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0152] Optionally, the computing system may include clients and servers. Clients and servers are generally geographically separated and typically interact via a communication network. The client-server relationship is established by computer programs running on the respective computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system. This addresses the shortcomings of traditional physical hosts and Virtual Private Server (VPS) services, such as high management difficulty and weak business scalability.

[0153] The above description is based on the preferred embodiments of the present invention. It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered exemplary and non-limiting in all respects. The scope of the invention is defined by the appended claims rather than the foregoing description, and all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0154] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A method for generating indoor situational awareness by combining thin-plate spline interpolation and natural neighbor interpolation, characterized in that, Includes the following steps: S1. Obtain the original sampling data within the target area, wherein the original sampling data includes the location information of the sampling points and their corresponding signal strength values; S2. Divide the target area into grids according to the first spatial resolution, merge the original sampled data falling into the same grid, extract the merged data of each grid, and obtain the first sample dataset. S3. Based on the first sample dataset, use the thin plate spline interpolation method to generate preliminary interpolation data covering the target area; S4. According to the preset sampling and extraction rules, in the area of ​​the target area that does not cover the original sampling data, supplementary sampling point data is extracted from the preliminary interpolation data to obtain a supplementary sample dataset; S5. Divide the original sampling data into grids according to the second spatial resolution, and merge the original sampling data falling into the same grid to obtain the second sample dataset; S6. Merge the supplementary sample dataset with the second sample dataset, and filter the supplementary sample data that are close in location and have signal strength differences exceeding a preset threshold to obtain the target sample dataset. S7. Based on the target sample dataset, the target region is interpolated using the natural neighborhood interpolation method to generate the final situational data of the target region.

2. The indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation according to claim 1, characterized in that, The first spatial resolution is greater than the second spatial resolution, the first spatial resolution is 5m-20m, and the second spatial resolution is 0.5m-2m.

3. The indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation according to claim 1, characterized in that, In step S2, the merging process is as follows: for each grid, the maximum value of the signal intensity values ​​of all sampling points in the grid is taken as the merged signal intensity value of the grid, and the geometric center position of the grid is taken as the spatial position corresponding to the merged signal intensity value.

4. The indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation according to claim 1, characterized in that, In step S4, the preset sampling extraction rule includes: scanning each spatial position row by row and column by column in the area not covered by the original sampling data with a third spatial resolution as the step size; for each scan position, calculating the Euclidean distance between the position and the nearest original sampling point; if the Euclidean distance is greater than a preset distance threshold, then extracting the signal strength value corresponding to the position in the preliminary interpolation data as the signal strength value of the supplementary sampling point, and using the position as the location information of the supplementary sampling point.

5. The indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation according to claim 1, characterized in that, In step S6, the filtering process specifically involves: for each supplementary sampling point, calculating the Euclidean distance between the location of the supplementary sampling point and the locations of each original sampling point in the second sample dataset; if the minimum Euclidean distance is less than a preset conflict distance threshold, and the absolute value of the difference between the signal strength value of the supplementary sampling point and the signal strength value of the corresponding nearest original sampling point is greater than a preset difference threshold, then the supplementary sampling point is removed from the merged dataset.

6. The indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation according to claim 1, characterized in that, The thin-plate spline interpolation method includes: constructing an interpolation function based on radial basis functions and global affine transformation; determining the weight coefficients and affine transformation coefficients of the interpolation function by solving a system of linear equations composed of sample point constraints and affine transformation constraints; thereby predicting the signal intensity at any location within the target area.

7. The indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation according to claim 1, characterized in that, The natural neighborhood interpolation method includes: Construct a first Voronoi diagram based on all sample points in the target sample dataset; For the location to be interpolated, the location to be interpolated is added as a new point to the target sample dataset, a second Voronoi diagram is constructed, and the neighboring sample points of the location to be interpolated are determined; Calculate the area ratio that the Voronoi cell corresponding to the position to be interpolated takes from the Voronoi cells corresponding to each neighboring sample point, and use this area ratio as the interpolation weight of each neighboring sample point. The signal strength values ​​of each neighboring sample point are weighted and summed according to the interpolation weight to obtain the predicted signal strength value of the position to be interpolated.

8. The indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation according to claim 1, characterized in that, The original sampling data also includes sampling time and frequency information. When there are multiple sampling values ​​at the same location and frequency, the arithmetic mean of the multiple sampling values ​​is taken as the representative value of the signal strength at that location and frequency.

9. An electronic device, characterized in that, include: At least one processor; and a memory communicatively connected to at least one of the processors; The memory stores a computer program executed by at least one of the processors, which enables the at least one processor to perform the indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the indoor situation generation method combining thin-plate spline interpolation and natural neighbor interpolation as described in any one of claims 1-8.

Images