A method, apparatus, device, and medium for dynamic updating of 3D geographic scenes based on spatiotemporal interpolation

Abstract

This invention discloses a method, apparatus, device, and medium for dynamic updating of 3D geographic scenes based on spatiotemporal interpolation. The method includes acquiring multi-source heterogeneous data, including satellite imagery, UAV remote sensing imagery, and ground sensor data; calculating the weight of each data source based on its spatial and temporal resolution; performing spatial and temporal interpolation calculations on spatial points of the 3D scene, combining the weights of the multi-source data; performing nonlinear compensation on the spatial and temporal interpolations to obtain the updated spatial point values ​​of the compensated 3D scene; comparing the interpolation results at different time points, detecting local changes, and updating the changed areas of the 3D scene. This invention ensures fusion quality through reasonable weighting by comprehensively configuring weights based on spatial resolution, temporal resolution, data quality, and similarity to reference data, making the fusion of basic 3D scene data more closely resemble real geographic features.

Description

Technical Field

[0001] This invention relates to the field of dynamic updating technology for three-dimensional geographic scenes, and in particular to a method, apparatus, device, and medium for dynamic updating of three-dimensional geographic scenes based on spatiotemporal interpolation. Background Technology

[0002] With the rapid development of remote sensing technology and 3D Geographic Information Systems (3DGIS), the construction of 3D geographic scenes has been widely applied across various industries. For complex application scenarios, such as real-world 3D scenes, city maps, and real-time monitoring of the natural environment, the updating of dynamic scenes places higher demands on data fusion and processing. Existing technologies primarily use the Triangulated Internet of Things (TIN) algorithm to perform elevation matching between the model and terrain, achieving full analysis and coordinate transformation of old and new data sources, combined with linear interpolation methods to estimate data changes over time or space, and directly rendering the updated scene. However, existing technologies have significant shortcomings in the following aspects:

[0003] (1) A single data source usually has a specific coverage area (e.g., satellite remote sensing data mainly covers large-scale areas, while UAV remote sensing data is concentrated in small areas). In scenarios that require both global and local fine-grained modeling, a single data source cannot simultaneously meet the multi-scale requirements. (2) The full reconstruction method reloads and renders the updated 3D model and terrain data. This method requires a large amount of computing resources and has a long update time. (3) The linear interpolation method can only be used for simple changing scenarios. It cannot accurately describe real-time scenarios for complex nonlinear dynamic processes (such as disaster occurrence and ecological changes).

[0004] For example, invention application CN118689889A discloses a method and apparatus for updating full-element map data based on multi-technology fusion, which uses a drone to collect and update full-element data of a regional map according to a predetermined route. However, this invention uses a single drone data source, which limits it to updating only a small area of ​​3D map data; and for updating full-element data of a regional map, it cannot achieve rapid updating of local data, requires a large amount of computing resources, and has a long update time.

[0005] Based on the technical shortcomings exposed by the above methods, a dynamic update method for 3D geographic scenes based on spatiotemporal interpolation is needed. This method should be adaptable to high-quality fusion and dynamic weight allocation of multi-source data, and applicable to scenarios such as urban planning, disaster emergency response, and environmental monitoring. Summary of the Invention

[0006] The purpose of this invention is to provide a method, apparatus, device, and medium for dynamic updating of three-dimensional geographic scenes based on spatiotemporal interpolation, so that the fusion of basic data of three-dimensional scenes is more in line with the real geographic features, and the scene update provides high-quality basic data.

[0007] This invention provides a method, apparatus, device, and medium for dynamically updating three-dimensional geographic scenes based on spatiotemporal interpolation.

[0008] First aspect: A method for dynamic updating of 3D geographic scenes based on spatiotemporal interpolation, including:

[0009] S1. Collect multi-source heterogeneous data, including satellite imagery, UAV remote sensing imagery, and ground sensor data;

[0010] S2. Calculate the weight of a single data source based on the spatial and temporal resolution of the data;

[0011] S3. For spatial points in a 3D scene, perform spatial interpolation and temporal interpolation calculations by combining the weights of multi-source data;

[0012] S4. Perform nonlinear compensation on spatial interpolation and temporal interpolation to obtain the updated spatial point values ​​of the compensated 3D scene.

[0013] S5. Compare the interpolation results at different time points, detect local changes, and update the changed areas of the 3D scene.

[0014] Furthermore, in S2, a quality coefficient and a similarity metric are used to dynamically evaluate the weight of each data source, as expressed by the formula:

[0015]

[0016] Where R i Let T be the spatial resolution of the i-th data source, representing the number of pixels within the spatial area; where T i Let Q be the time resolution of the i-th data source, representing the time interval for data acquisition; i Sim(D) is a data quality coefficient used to consider factors such as noise and missing data in the data source. i D ref ) is a data similarity metric used to measure the degree of matching with reference data; α, β, and γ are used to adjust the weight of the metric.

[0017] Furthermore, in step S3, spatial interpolation is calculated using the following formula:

[0018]

[0019] Where ω i f is the weight of the i-th data source. i (x,y,z) represents the attribute value of the i-th point in the 3D data (x,y,z), Δ i (x,y,z) is the bias correction function for the i-th data source based on multi-scale decomposition and error statistics.

[0020] Furthermore, the deviation correction function of the data source is expressed by the following formula:

[0021]

[0022] Where C il φ is the correction coefficient for the i-th data source at the l-th scale. l (x,y,z) is the l-th scaling basis function.

[0023] Furthermore, in step S3, time interpolation is calculated using a modified Lagrange interpolation formula, which is expressed as:

[0024]

[0025] δ k This is the dynamic adjustment coefficient for the k-th time period. Let L be a nonlinear basis function for time period k. k (t) is the basis function of Lagrange interpolation.

[0026] Furthermore, in step S4, nonlinear compensation is performed on the spatial and temporal interpolations, as expressed by the following formula:

[0027]

[0028] Where C(t,x,t,z) is the compensated 3D scene update value, S(x,y,z) is the spatial interpolation result, T(t) is the temporal interpolation result, and ω i f is the weight of the i-th data source. i (x,y,z) represents the attribute value of the i-th point in the three-dimensional data (x,y,z), and θ and μ are adjustment coefficients. The hyperbolic tangent function tanh(x) is introduced to compress the extreme values ​​of the data.

[0029] Furthermore, in step S5, the interpolation results at different time points are compared to detect local changes, as expressed by the formula:

[0030]

[0031] Where R p For the p-th block region, ∈ p For region R p Change detection base threshold, σ p For region R p The uncertainty coefficient of change.

[0032] The second aspect: a dynamic update device for three-dimensional geographic scenes based on spatiotemporal interpolation, comprising:

[0033] The data acquisition module is used to collect heterogeneous data from multiple sources.

[0034] The weight configuration module is used to configure the weights of each data source;

[0035] The interpolation calculation module is used to perform spatial and temporal interpolation calculations by combining the weights of multi-source data.

[0036] The compensation calculation module is used to obtain the updated spatial point values ​​of the compensated 3D scene based on spatial interpolation and temporal interpolation.

[0037] The detection and update module is used to detect local changes in a 3D scene and update the changed areas of the 3D scene.

[0038] Third aspect: An electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, performs the steps of the method provided in the first aspect.

[0039] Fourth aspect: A non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method provided in the first aspect.

[0040] The beneficial effects of this invention are:

[0041] 1. This invention ensures fusion quality through reasonable weighting. By comprehensively configuring weights based on spatial resolution, temporal resolution, data quality, and similarity to reference data, it allows heterogeneous data from multiple sources such as satellites, drones, and sensors to "fully utilize their respective strengths." High spatial resolution data is weighted appropriately in local fine-grained modeling, while high temporal resolution data leverages its advantages in dynamic monitoring. This solves the accuracy loss problem of traditional one-size-fits-all fusion of multi-source data, making the fusion of basic 3D scene data more closely resemble real geographical features, thus laying a solid foundation for accurate modeling and updates.

[0042] 2. This invention utilizes a joint interpolation model to model spatial and temporal dynamic changes, breaking the limitations of traditional separate spatial and temporal interpolation. On one hand, it collaboratively processes spatiotemporal data, reducing redundant calculations and error propagation, and improving interpolation efficiency (e.g., shortening modeling time and adapting to high-frequency dynamic update scenarios). On the other hand, it more accurately captures the coupling law of "spatial distribution + temporal evolution" in 3D scenes (e.g., the spatial growth of urban buildings over time, and the spatiotemporal development of terrain disasters), making the interpolation results more closely match the real dynamics of the geographical scene, providing high-quality basic data for subsequent change detection and scene updates.

[0043] 3. This invention overcomes the bottleneck of traditional methods by using nonlinear correction to learn complex nonlinear change patterns. It corrects dynamic details that traditional interpolation struggles to capture, compensating for the shortcomings of linear interpolation in characterizing nonlinear features of geographic scenes. For example, in nonlinear uplift of terrain due to geological movements and nonlinear succession of wetland ecosystems with the seasons, traditional methods easily lose key details due to linear assumptions. This strategy, through nonlinear correction, allows for more accurate dynamic updates of 3D scenes to reproduce real geographic processes, enhancing the application value of the scene.

[0044] 4. This invention abandons the redundant full-data reconstruction mode and adopts a local change detection and targeted update mechanism, reducing the update time from 2 hours to 12 minutes. It can respond to dynamic changes in scenarios at high frequency, meeting the timeliness requirements of real-time monitoring and emergency response; local updates avoid redundant calculations of all data, significantly reducing the consumption of computing power and storage resources. When dealing with large-scale scenarios, it maintains dynamic updates with lower computational costs, adapts to different hardware environments, and lowers the threshold for project implementation and maintenance. Attached Figure Description

[0045] Figure 1 This is a flowchart illustrating the dynamic updating method for three-dimensional geographic scenes based on spatiotemporal interpolation according to the present invention.

[0046] Figure 2 This is a schematic diagram of the structure of the three-dimensional geographic scene dynamic update device based on spatiotemporal interpolation of the present invention;

[0047] Figure 3 This is a schematic diagram of the structure of the electronic device of the present invention. Detailed Implementation

[0048] Embodiments of the present invention are described in detail below. Examples of these embodiments are illustrated in the accompanying drawings, wherein the same or similar symbols denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0049] Dynamic updates of 3D geographic scenes face challenges: First, the coverage of a single data source is limited and cannot meet the needs of multi-scale modeling; second, full reconstruction methods are computationally intensive and time-consuming to update; and finally, linear interpolation methods are not suitable for complex nonlinear dynamic scenes.

[0050] To address the above problems, this invention provides a method for dynamically updating three-dimensional geographic scenes based on spatiotemporal interpolation. Figure 1 This is a flowchart illustrating a method for dynamically updating a 3D geographic scene based on spatiotemporal interpolation, as provided in an embodiment of the present invention. The method includes:

[0051] S1. Collect multi-source heterogeneous data.

[0052] Multi-source heterogeneous data types include satellite imagery, UAV remote sensing imagery, and ground sensor data.

[0053] Satellite imagery is characterized by its wide coverage, making it suitable for acquiring geographic information over large-scale areas, such as overall urban layouts and extensive topographical features. While its spatial resolution is relatively low (e.g., 1 meter in this example), its temporal resolution can be quite long, making it suitable for providing data support at the macro level.

[0054] UAV remote sensing offers high spatial resolution (e.g., 0.1 meters in this example), enabling detailed data collection over small areas, such as building details and road textures. Its relatively short temporal resolution allows for more frequent acquisition of up-to-date data for localized areas.

[0055] Ground-based sensor data offers extremely high temporal resolution (e.g., 0.1 hours in this example), enabling real-time or high-frequency data collection at specific locations, such as changes in ground elevation and environmental parameters. However, its spatial coverage is relatively small (e.g., a spatial resolution of 5 meters in this example), typically used for precise monitoring of key local points.

[0056] By employing multi-source heterogeneous data, multi-scale requirements can be met. Satellite imagery can cover the entire area, while UAV remote sensing and ground sensors can focus on specific areas. The combination of these three technologies can achieve multi-scale modeling from macro to micro, solving the problem that a single data source cannot simultaneously meet the requirements for global and local fine-grained modeling.

[0057] By employing multi-source heterogeneous data, data reliability can be improved. The characteristics of different data sources complement each other. For example, satellite imagery provides macro background, UAV remote sensing supplements details, and ground sensors verify local accuracy. This can reduce the error and limitations of a single data source and improve the overall quality and reliability of the data.

[0058] By using multi-source heterogeneous data, we can adapt to complex scenarios. In complex scenarios such as urban planning, disaster emergency response, and environmental monitoring, multi-source data can more comprehensively reflect the dynamic changes of the geographical scene and provide a richer information foundation for subsequent dynamic updates.

[0059] S2. Calculate the weight of a single data source based on the spatial and temporal resolution of the data.

[0060] Weighted calculation dynamically allocates the contribution ratio of multi-source data by quantifying four major indicators: spatial resolution, temporal resolution, data quality, and similarity measurement, thus solving the problem that a single data source cannot adapt to multi-scale requirements in existing technologies.

[0061] The weight calculation results directly affect the accuracy of subsequent spatiotemporal interpolation and are one of the keys to achieving local updates rather than full reconstruction. The formula is expressed as:

[0062]

[0063] Where R i R represents the spatial resolution of the i-th data source, calculated as the number of pixels within the spatial area, measuring the spatial fineness of the data; i The smaller the resolution, the higher the weighting percentage.

[0064] Where T i Let T be the time resolution of the i-th data source, representing the time interval between data acquisitions; for example, in this case, ground sensor T3 = 0.1, T... i The smaller the value (the higher the update frequency), the higher the weight percentage.

[0065] Q i This is a data quality coefficient used to consider factors such as noise and missing data in the data source; for example, in the drone data example, Q2 = 0.95, and the higher the value, the higher the weight.

[0066] Sim(D i D ref This is a data similarity metric used to measure the degree of matching with reference data, such as the similarity between satellite imagery and reference data. The higher the similarity, the higher the weight.

[0067] α, β, and γ are used to adjust the weights of the indicators. The weights of the indicators are set manually (such as α = 0.4, β = 0.3, and γ = 0.2 in the example). The importance of each indicator can be flexibly adjusted according to the needs of the scenario.

[0068] S3. For spatial points in the 3D scene, perform spatial interpolation and temporal interpolation calculations by combining the weights of multi-source data.

[0069] Based on multi-source data weights, the three-dimensional attribute values ​​(such as elevation) of multi-source data such as satellites, drones, and sensors are integrated into precise values ​​under a unified spatial coordinate system in the spatial dimension; in the temporal dimension, the dynamic change patterns of the same spatial point at different times are captured.

[0070] The formula for spatial interpolation is:

[0071]

[0072] Where ω i f is the weight of the i-th data source. i (x,y,z) represents the attribute value of the i-th point in the 3D data (x,y,z), Δ i (x,y,z) is the bias correction function for the i-th data source based on multi-scale decomposition and error statistics.

[0073] This formula follows the principle that high-weighted data dominates the interpolation results; for example, UAV data (high spatial resolution) dominates in local detail interpolation. A theoretical innovation for bias correction is introduced by setting a bias correction function, expressed as follows:

[0074]

[0075] Where C il φ is the correction coefficient for the i-th data source at the l-th scale. l (x,y,z) is the l-th scaling basis function.

[0076] By using a deviation correction function to address inherent errors in the data source, the error is decomposed into components of different spatial scales (such as meters and centimeters). Layered correction is achieved through linear combination, thereby reducing systematic errors.

[0077] Based on the spatial interpolation calculations described above, multi-source data consistency constraints are achieved through weight ω. i And deviation correction Δ i (x,y,z) forces attribute values ​​from different data sources to meet weighted consistency at spatial points, for example, to solve the scale conflict problem between satellite (macro) and UAV (local) data.

[0078] Traditional Lagrange interpolation is only applicable to linear variation scenarios. Its theoretical assumption is that the time series satisfies polynomial smoothness, and it cannot describe abrupt changes.

[0079] This invention performs time interpolation calculations using a modified Lagrange interpolation formula, expressed as follows:

[0080]

[0081] Where, δ k This is the dynamic adjustment coefficient for the k-th time period. Let L be a nonlinear basis function for time period k. k (t) is the basis function of Lagrange interpolation.

[0082] Using the dynamic adjustment coefficient δ k Based on time series analysis of variance, the weights of each time period can be adjusted in real time, δ k Increasing the value of the newest data will amplify its impact; nonlinear basis functions It can fit dynamic processes.

[0083] The calculations of spatial interpolation and temporal interpolation form a spatiotemporal coupling relationship. The spatial interpolation result provides the spatial reference for the temporal interpolation at the current moment, while the temporal interpolation result provides the historical trend of spatial interpolation. The two form a two-way constraint through weights.

[0084] S4. Perform nonlinear compensation on spatial interpolation and temporal interpolation to obtain the updated spatial point values ​​of the compensated 3D scene.

[0085] The obtained spatial and temporal interpolation results are input into a nonlinear mapping function. A nonlinear compensation algorithm is then used to correct the interpolation results, resulting in more accurate 3D scene update values. Nonlinear compensation is applied to both spatial and temporal interpolation, as expressed by the following formula:

[0086]

[0087] Where C(t,x,t,z) is the compensated 3D scene update value, S(x,y,z) is the spatial interpolation result, T(t) is the temporal interpolation result, and ω i f is the weight of the i-th data source. i (x,y,z) represents the attribute value of the i-th point in the three-dimensional data (x,y,z), and θ and μ are adjustment coefficients. The hyperbolic tangent function tanh(x) is introduced to compress the extreme values ​​of the data.

[0088] Spatial interpolation S(x,y,z) reflects the distribution of attribute values ​​in the spatial dimension of a 3D scene. Temporal interpolation T(t) reflects the trend of data change over time. Data source weight ω i This is used to measure the contribution of different data sources to nonlinear compensation. Adjustment coefficients θ and μ are used to regulate the intensity and method of nonlinear compensation and can be set according to the actual application scenario. tanh(x) is the hyperbolic tangent function, which compresses the extreme values ​​of the data, making the compensated results more stable and reasonable.

[0089] First, the spatial interpolation result S(x,y,z) and the temporal interpolation result T(t), multiplied by an adjustment factor θ, are added together to obtain the preliminary compensation basis. Then, ω is calculated. i f i (x,y,z), where f i (x,y,z) represents the attribute value of the i-th point in the 3D data. This product reflects the combined influence of the data source weights and attribute values. Then, this product is substituted into the hyperbolic tangent function tanh(x) to compress the data extrema. The compressed result is multiplied by an adjustment coefficient μ and incremented by 1 to obtain a correction factor. Finally, the temporal interpolation result T(t) is multiplied by this correction factor and then added to the spatial interpolation result S(x,y,z) to obtain the final compensated 3D scene update value C(t,x,t,z).

[0090] The output is the compensated 3D scene update value C(t,x,t,z). This value takes into account information from both spatial and temporal dimensions and corrects for possible nonlinear changes through a nonlinear compensation algorithm. It can more accurately reflect the actual state of the 3D geographic scene and provide a more reliable data foundation for subsequent local change detection and scene updates.

[0091] S5. Compare the interpolation results at different time points, detect local changes, and update the changed areas of the 3D scene.

[0092] By comparing spatiotemporal interpolation results at different time points, localized areas of change in a 3D geographic scene are accurately identified, and updates are performed only on these changed areas, avoiding full reconstruction and thus improving update efficiency and accuracy. The formula is as follows:

[0093]

[0094] Where R p For the p-th block region, ∈ p For region R p Change detection base threshold, σ p For region R p The uncertainty coefficient of change.

[0095] Calculate the spatiotemporal interpolation result C1(t1,x,t,z) at the current time point and the interpolation result C2(t2,x,t,z) at historical time points.

[0096] The 3D scene is divided into multiple block regions R p (p is the region number), each region corresponds to an independent change detection threshold ∈ p and uncertainty coefficient σ p The threshold is set according to the needs of the scenario (such as terrain complexity and data accuracy).

[0097] By calculating the absolute value of the difference between the current and historical interpolation results within each block region, if the value exceeds a threshold, it is determined that there is a significant change in the region and it is included in the set of regions to be updated, D.

[0098] Example 2: The present invention also provides a three-dimensional geographic scene dynamic update device based on spatiotemporal interpolation, such as... Figure 2 As shown, the device includes a data acquisition module, a weight configuration module, an interpolation calculation module, a compensation calculation module, and a detection and update module, among which:

[0099] The data acquisition module is used to collect multi-source heterogeneous data; the weight configuration module is used to configure the weights of each data source; the interpolation calculation module is used to perform spatial and temporal interpolation calculations by combining the weights of multi-source data; the compensation calculation module is used to obtain the updated spatial points of the compensated 3D scene based on spatial and temporal interpolation; and the detection and update module is used to detect local changes in the 3D scene and update the changed areas of the 3D scene.

[0100] Using the aforementioned device, the various modules work collaboratively to address the problems of low efficiency in full-scale reconstruction and insufficient accuracy in linear interpolation in traditional technologies. Through the coordinated operation of five modules—data acquisition, weight configuration, interpolation calculation, compensation correction, and change detection—the device achieves high-precision and high-efficiency dynamic updates of 3D geographic scenes, making it suitable for various scenarios such as urban planning and disaster emergency response.

[0101] The data acquisition module collects heterogeneous data from multiple sources, providing basic information for subsequent processing. Supported data sources include satellite imagery (large-scale spatial coverage), UAV remote sensing (high spatial resolution), and ground sensors (high-frequency temporal sampling), covering multi-scale and multi-temporal data requirements.

[0102] The weight configuration module dynamically calculates the weights of each data source based on data characteristics, thereby achieving weighted fusion of multi-source data.

[0103] Weight calculation formula:

[0104]

[0105] R i T represents the spatial resolution (number of pixels). i For time resolution (time interval); Q i Sim(D) is the data quality coefficient (measuring noise and missing data). i D ref ) is a similarity measure with reference data, and α, β, and γ are weight adjustment coefficients.

[0106] For example, the original weights of satellite imagery, drones, and ground sensors are 0.6618, 4.2925, and 3.31, respectively. After normalization, they are 0.08, 0.52, and 0.40, reflecting the weight advantages of high spatial resolution (drones) and high temporal resolution (sensors) data.

[0107] The interpolation calculation module combines the data source weights to perform spatial interpolation and temporal interpolation respectively, constructing a spatiotemporal dynamic model of the 3D scene.

[0108] The spatial interpolation formula is:

[0109]

[0110] Where fi (x, y, z) are the data source attribute values, Δ i (x,y,z) is the deviation correction function, which is corrected through multi-scale decomposition and error statistics calculation (e.g., in the example, the deviation correction of UAV data is Δ2 = 0.5 meters, which improves the interpolation accuracy).

[0111] Time interpolation is obtained using a modified Lagrange interpolation formula, expressed as:

[0112]

[0113] Where L k (t) is the Lagrange basis function, δ k For dynamic adjustment coefficients, It is a nonlinear basis function (for example, the interpolation result for the time period t0 to t1 in the example is 52.0 meters).

[0114] The compensation calculation module compensates for the spatiotemporal interpolation results through a nonlinear mapping function, correcting nonlinear changes that traditional linear interpolation cannot capture.

[0115] The nonlinear compensation formula is expressed as:

[0116]

[0117] Where θ and μ are adjustment coefficients (θ = 0.3 and μ = 0.7 in the example), the hyperbolic tangent function tanh(x) is used to compress data extrema.

[0118] The detection and update module compares the interpolation results at multiple time points, detects local changes, and updates only the changed areas to avoid full reconstruction.

[0119] Change detection formula

[0120]

[0121] Where R p For the p-th block region, ∈ p For region R p Change detection base threshold, σ p For region R p The uncertainty coefficient of change. (In the example, the threshold for region A is 0.5 meters, the coefficient is 0.1, and the difference of 2.18 meters > 0.55 meters, so it is determined to be a region of change).

[0122] This invention utilizes change detection to reload data, perform interpolation and compensation only in the changed areas. In the example, the update time was reduced from 2 hours in the traditional method to 12 minutes, and the elevation error was reduced from 8% to 1.5%.

[0123] This invention's device adapts to multi-source data such as satellites, drones, and sensors through a weight configuration module, solving the problem that a single data source cannot meet multi-scale requirements. The detection and update module achieves local updates, reducing computation by over 90% compared to traditional full reconstruction, while improving accuracy through nonlinear compensation. It is suitable for scenarios requiring high-frequency updates, such as urban planning: dynamically updating the elevation data of newly constructed buildings; real-time monitoring of terrain changes in disaster areas such as landslides and floods; and environmental monitoring tracking nonlinear processes such as vegetation cover and landform evolution in ecological areas.

[0124] Application examples:

[0125] For example, dynamically update the three-dimensional elevation data of a certain urban area to detect changes in newly constructed buildings (area A).

[0126] Collect multi-source heterogeneous data, including satellite imagery, UAV remote sensing data, and ground sensor data. For example, collect data from satellite imagery, UAV data, and ground sensor data separately:

[0127] Satellite imagery: spatial resolution R1 = 1 meter, temporal resolution T1 = 168 hours, quality coefficient Q1 = 0.9, similarity Sim1 = 0.8.

[0128] UAV remote sensing: spatial resolution R2 = 0.1 meters, temporal resolution T2 = 24 hours, quality coefficient Q2 = 0.95, similarity Sim2 = 0.9.

[0129] Ground sensor: spatial resolution R3 = 5 meters, temporal resolution T3 = 0.1 hours, quality coefficient Q3 = 0.8, similarity Sim3 = 0.7.

[0130] Set α=0.4, β=0.3, and γ=0.2.

[0131] Based on the spatial and temporal resolution of the data, the weight of each data source is calculated. Substituting into the formula, we get: satellite imagery 0.6618, UAV remote sensing 4.2925; ground sensor 3.31. After normalization, we get:

[0132] ω1=0.6618 / (0.6618+4.2925+3.31)=0.08;

[0133] ω² = 4.2925 / 8.2643 = 0.52;

[0134] ω3 = 3.31 / 8.2643 = 0.40;

[0135] Spatiotemporal interpolation is performed on spatial points in a 3D scene by combining multi-source data.

[0136] Spatial interpolation calculation: Input data: satellite elevation f1 = 50.0 meters, UAV elevation f2 = 52.3 meters, sensor elevation f3 = 51.8 meters, bias correction Δ2 = 0.5 meters (UAV data), bias correction for satellite data and sensor data is 0. Calculation results:

[0137] s(x,y,z)=0.08×50.0+0.52×(52.3+0.5)+0.40×51.8=4.0+27.456+20.72=52.176 meters

[0138] Time interpolation calculation: The improved Lagrange interpolation formula is used, with a time period from t0 to t1 (7 days), and dynamic adjustment coefficients σ1 = 0.6 and σ2 = 0.4. The interpolation result T(t) = 52.0 meters.

[0139] Spatial and temporal interpolation values ​​are input into the nonlinear mapping function to obtain the compensated result. Setting μ = 0.7 and θ = 0.3, the nonlinear compensation is calculated by substituting the aforementioned temporal, spatial, and weight values ​​into the formula.

[0140] C(t,x,t,z)=tanh(0.7×52.176+0.3×52.2)=tanh(52.177)≈52.18 meters

[0141] By comparing the interpolation results at different time points, local changes are detected, and the changed areas are updated. Based on the above calculation steps, the original elevation was 50 meters, and after the update, it is 52.18 meters. According to the formula:

[0142]

[0143] The threshold for area A is set at 0.5 meters, with an uncertainty coefficient of 1.1.

[0144] Judgment: |52.18-50.0|=2.18>0.5×1.1=0.55, therefore this region needs to be updated.

[0145] Based on the steps above, region A needs to be dynamically updated.

[0146] Simultaneously, performance verification was conducted, and the data shown in Table 1 supports this:

[0147] Table 1 Comparison of Elevation Error and Update Time between Traditional Method and Invention Method

[0148] index Traditional methods This invention Elevation error 8% (Full Reconstruction) 1.5% (Partial Update) Update time 2 hours 12 minutes

[0149] The above examples demonstrate that the present invention achieves high-precision and high-efficiency dynamic updating of three-dimensional scenes through multi-source data fusion and spatiotemporal interpolation, which can improve the accuracy and efficiency of computers in processing three-dimensional scenes and realize efficient updating of three-dimensional scenes.

[0150] The present invention also provides an electronic device, Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention, such as... Figure 3 As shown, the electronic device may include a processor, a communications interface, a memory, and a communication bus, wherein the processor, communications interface, and memory communicate with each other via the communication bus. The processor can invoke logical instructions from the memory, for example, to execute the following method (method steps).

[0151] S1. Collect multi-source heterogeneous data, including satellite imagery, UAV remote sensing imagery, and ground sensor data.

[0152] S2. Calculate the weight of a single data source based on the spatial and temporal resolution of the data.

[0153] S3. For spatial points in the 3D scene, perform spatial interpolation and temporal interpolation calculations by combining the weights of multi-source data.

[0154] S4. Perform nonlinear compensation on spatial interpolation and temporal interpolation to obtain the updated spatial point values ​​of the compensated 3D scene.

[0155] S5. Compare the interpolation results at different time points, detect local changes, and update the changed areas of the 3D scene.

[0156] Furthermore, the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, and can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0157] This invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, is implemented to perform the methods provided in the above embodiments, including, for example:

[0158] S1. Collect multi-source heterogeneous data, including satellite imagery, UAV remote sensing imagery, and ground sensor data.

[0159] S2. Calculate the weight of a single data source based on the spatial and temporal resolution of the data.

[0160] S3. For spatial points in the 3D scene, perform spatial interpolation and temporal interpolation calculations by combining the weights of multi-source data.

[0161] S4. Perform nonlinear compensation on spatial interpolation and temporal interpolation to obtain the updated spatial point values ​​of the compensated 3D scene.

[0162] S5. Compare the interpolation results at different time points, detect local changes, and update the changed areas of the 3D scene.

[0163] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. 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 the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0164] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0165] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for dynamic updating of a three-dimensional geographic scene based on spatio-temporal interpolation, characterized in that, The method comprises the steps of: S1, collecting multi-source heterogeneous data, including satellite images, unmanned aerial vehicle remote sensing images and ground sensor data; S2, calculating the weight of a single data source according to the spatial resolution and time resolution of the data; S3, performing spatial interpolation and time interpolation calculation on the spatial points of the three-dimensional scene in combination with the multi-source data weight; S4, performing nonlinear compensation on the spatial interpolation and time interpolation to obtain the updated value of the spatial points of the compensated three-dimensional scene; S5, comparing the interpolation results of different time points, detecting local changes, and updating the changed area of the three-dimensional scene; In S2, the weight of each data source is dynamically evaluated using a quality coefficient and a similarity measure, and the formula is: ； wherein, is the weight of the th data source, is the spatial resolution of the th data source; is the temporal resolution of the th data source; is a data quality coefficient; is a data similarity measure; for adjusting the indicator weights; In S3, the spatial interpolation calculation is performed, and the formula is: ； in, For spatial interpolation results, For the first The weight of each data source, For the first Each point in 3D data The attribute values ​​in For the first time based on multi-scale decomposition and error statistics Deviation correction function for each data source; The bias correction function of the data source is represented by the formula: ； wherein, is a correction factor for the th data source on the th scale, is a scale basis function for the th scale. In S3, the time interpolation calculation is performed using an improved Lagrange interpolation formula, and the formula is: ； in, For time interpolation results, For the first Dynamic adjustment coefficient for each time period Time period Nonlinear basis functions, These are the basis functions for Lagrange interpolation.

2. The method of claim 1, wherein, In S4, the nonlinear compensation of the spatial interpolation and time interpolation is performed, and the formula is: ； in, The updated 3D scene values ​​after compensation. For spatial interpolation results, For time interpolation results, For the first The weight of each data source, For the first Each point in 3D data The attribute values ​​in To adjust the coefficient, It is the hyperbolic tangent function.

3. The method of claim 2, wherein, In S5, the interpolation results of different time points are compared, and the local change detection is performed, and the formula is: ； in, To update the region set, For the first Each segmented region For the region The basic threshold for detecting changes, For the region The uncertainty coefficient of change.

4. A device for dynamic updating of a three-dimensional geographical scene based on spatio-temporal interpolation applied to the method according to any one of claims 1 to 3, characterized in that, The device comprises: A data acquisition module for acquiring multi-source heterogeneous data; A weight configuration module for configuring the weight of each data source; An interpolation calculation module for performing spatial interpolation and time interpolation calculation in combination with the multi-source data weight; A compensation calculation module for obtaining the updated value of the spatial points of the compensated three-dimensional scene according to the spatial interpolation and time interpolation; A detection and update module for detecting local changes in the three-dimensional scene and updating the changed area of the three-dimensional scene.

5. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to implement the steps of the three-dimensional geographic scene dynamic update method based on spatio-temporal interpolation according to any one of claims 1-3.

6. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the three-dimensional geographic scene dynamic update method based on spatio-temporal interpolation according to any one of claims 1-3.

Images