A spatial sampling point planning method considering spatiotemporal variability

Abstract

This paper proposes a spatial sampling point planning method that accounts for both spatiotemporal variability. The method includes: constructing a geographic statistical data set from multiple time layers, gridding the study area based on the data set and generating a set of candidate points; initializing sampling based on the candidate point set and constructing an INLA-SPDE-based spatiotemporal prediction model; performing multi-time-layer spatiotemporal prediction and multi-time-layer position sampling on undeployed areas using the spatiotemporal prediction model, updating the model parameters and sampling set of the spatiotemporal prediction model, and outputting a sampling point layout plan based on a cutoff condition. By incorporating the spatiotemporal variation characteristics of the target variable into spatial sampling point planning, the present invention enables sampling points to accurately reproduce the spatiotemporal variation of the target variable in the monitoring domain, effectively addressing the issue of insufficient attention to the spatiotemporal variability of the target variable and improving the spatiotemporal representativeness of the sampling points for the monitoring domain. The method has high practical value and application prospects.

Description

Technical Field

[0001] The present invention relates to the field of geographic statistical spatial data collection, and in particular to a spatial sampling point planning method taking into account temporal and spatial variability. Background Art

[0002] Spatial sampling point planning is one of the key steps in geostatistical spatial data collection and is crucial to ensuring the accuracy and comprehensiveness of the target variables in the sampling space.

[0003] Traditional sampling point planning methods often aim to maximize coverage and achieve uniform spatial density, including equally spaced sampling, equally probable sampling, and random sampling. These methods have limited application scenarios and certain drawbacks. For example, spaced sampling can easily lead to clustering and repeated sampling of sampling points, equally probable sampling is poorly adaptable to spatial variability, and random sampling can lead to uneven distribution of sampling points. Optimization methods based on intelligent algorithms such as genetic algorithms, ant colony algorithms, and particle swarm algorithms have also been widely used in spatial sampling planning, but they suffer from high computational complexity and difficulty in parameter setting. Furthermore, these methods all consider only the spatial characteristics of the target variable, rarely considering the spatiotemporal variability of the sampling targets within the monitoring domain. This makes it difficult for the collected data to capture the spatiotemporal trends of the sampling targets, resulting in poor representation of the sampling points within the spatial domain.

[0004] Therefore, how to plan sampling points in the study area so that the collected data can effectively capture the spatiotemporal variation characteristics of the sampling targets is a technical problem that needs to be solved urgently in this field. Summary of the Invention

[0005] To address the technical problem of planning sampling points within a study area so that the collected data can effectively capture the spatiotemporal variation characteristics of the sampling target, this paper proposes a spatial sampling point planning method that takes into account both spatiotemporal variability and its corresponding apparatus, equipment, and storage medium. This method supports applications such as spatiotemporal data analysis of sampling target variables and site selection for monitoring station deployment.

[0006] In order to achieve the above-mentioned purpose, the technical solution adopted by the present invention includes the following four aspects:

[0007] In a first aspect, the present invention provides a spatial sampling point planning method that takes into account temporal and spatial variability, comprising the following steps:

[0008] S1: Construct a set of geographic statistical data from multiple time layers, grid the study area based on the data set and generate a set of candidate points;

[0009] S2: Initialize sampling based on the candidate point set and build a spatiotemporal prediction model based on INLA-SPDE;

[0010] S3: Perform multi-time layer spatiotemporal prediction and multi-time layer position sampling on the undeployed area through the spatiotemporal prediction model to generate a sampling point set, and update the model parameters of the spatiotemporal prediction model based on the sampling point set. When the cutoff condition is met, the sampling point layout plan is output.

[0011] Furthermore, step S1 specifically includes:

[0012] S1.1: Extract relevant data content from geographic statistical data sets at multiple time levels based on the analysis objectives;

[0013] S1.2: Based on the data set constructed in step S1.1, perform data regularization based on target requirements, divide the study area into several grids, and generate candidate points in each grid to obtain a candidate point set.

[0014] Furthermore, step S2 specifically includes:

[0015] S2.1: Initialize sampling from the candidate point set generated in step S1 to construct an initial sampling point set;

[0016] S2.2: Based on the initial sampling point set constructed in step S2.1, construct a spatiotemporal prediction model based on INLA-SPDE.

[0017] Furthermore, step S3 specifically includes:

[0018] S3.1: Based on the spatiotemporal prediction model constructed in step S2, use the sampling points to perform multi-time layer spatiotemporal prediction on the undeployed area, obtain the predicted value and confidence interval of each candidate point, and calculate the prediction error of the candidate point set;

[0019] S3.2: Extracting sampling positions of multiple time layers based on the calculated prediction error and the set sampling frequency and sampling quantity;

[0020] S3.3: Construct a new set of sampling points based on the extracted multi-time layer sampling positions;

[0021] S3.4: Based on the constructed sampling point set, determine whether the stopping requirement is met according to the set cutoff condition. If the condition is met, complete the sampling and execute step S3.6. If the condition is not met, execute step S3.5.

[0022] S3.5: Reconstruct the spatiotemporal prediction model using the constructed sampling point set, update the model parameters, and repeat steps S3.1 to S3.5;

[0023] S3.6: Output the set of sampling points that meet the cutoff conditions, that is, the layout plan.

[0024] Furthermore, in step S2, in the initial stage of spatial data collection, a group of initial sampling points are selected through a preset initialization sampling strategy to construct an initial sampling point set.

[0025] Furthermore, the preset initialization sampling strategy includes random sampling, uniform sampling and suppressive sampling.

[0026] Furthermore, in step S3.3, the constructed new sampling point set does not include the initial sampling point set.

[0027] In a second aspect, the present invention provides a spatial sampling point planning device that takes into account temporal and spatial variability, comprising the following modules:

[0028] Data set processing module, used to construct geographic statistical data sets from multiple time layers, grid the study area according to the data set and generate candidate point sets;

[0029] The spatiotemporal model building module is used to initialize sampling based on the candidate point set and build a spatiotemporal prediction model based on INLA-SPDE;

[0030] The spatiotemporal adaptive sampling module is used to perform multi-time layer spatiotemporal prediction and multi-time layer position sampling on the undeployed area through the spatiotemporal prediction model, generate a set of sampling points, and update the model parameters of the spatiotemporal prediction model according to the sampling point set. When the cutoff condition is met, the sampling point layout plan is output.

[0031] In a third aspect, the present invention provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the spatial sampling point planning method when executing the program.

[0032] In a fourth aspect, the present invention provides a storage medium having a computer program stored thereon, which implements the steps of the spatial sampling point planning method when executed by a processor.

[0033] The technical solution provided by the present invention has the following beneficial effects:

[0034] The present invention provides a spatial sampling point planning method that takes into account spatiotemporal variability. The method first constructs a geographic statistical data set from multiple time layers, grids the study area according to the data set, and generates a candidate point set; then initializes sampling based on the candidate point set, and constructs a spatiotemporal prediction model based on INLA-SPDE; finally, the spatiotemporal prediction model is used to perform multi-time layer spatiotemporal prediction and multi-time layer position sampling on the undeployed area to generate a sampling point set, and updates the model parameters of the spatiotemporal prediction model based on the sampling point set. When the cutoff condition is met, the sampling point layout plan is output. In the above scheme, by integrating the spatiotemporal variation characteristics of the target variable into the spatial sampling point planning method, the sampling points can accurately restore the spatiotemporal variation of the target variable in the monitoring domain, effectively solving the problem of less attention paid to the spatiotemporal variability of the target variable, and improving the spatiotemporal representativeness of the sampling points to the monitoring domain. The spatial sampling point planning framework proposed in this invention is applicable to a variety of application scenarios and needs. The constructed spatiotemporal variability modeling method, sampling point optimization algorithm and evaluation index model are highly versatile and can be widely used in the fields of monitoring site selection for different targets, spatiotemporal analysis and prediction of geographic statistical data, etc., with high practical value and application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:

[0036] Figure 1 This is an overall flow chart of a spatial sampling point planning method that takes into account temporal and spatial variability in an embodiment of the present invention. DETAILED DESCRIPTION

[0037] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, specific embodiments of the present invention are now described in detail with reference to the accompanying drawings.

[0038] In practice, each task will have its own specific main goal. This paper proposes a general spatial sampling point planning framework to explain the technical process of spatial sampling points and layout points. The ultimate goal is to select a limited number of sampling points and layout points in an infinite number of spatial locations. The generated sampling layout scheme can effectively capture the spatiotemporal variation characteristics of the target variable.

[0039] Please refer to Figure 1 In an embodiment of the present invention, a spatial sampling point planning method taking into account temporal and spatial variability mainly includes the following steps:

[0040] S1: Construct a set of geographic statistical data from multiple time layers, grid the study area based on the data set and generate a set of candidate points;

[0041] S2: Initialize sampling based on the candidate point set and build a spatiotemporal prediction model based on INLA-SPDE based on the initial sampling points;

[0042] S3: Perform multi-time layer spatiotemporal prediction and multi-time layer position sampling on the undeployed area through the spatiotemporal prediction model to generate a sampling point set, and update the model parameters of the spatiotemporal prediction model based on the sampling point set. When the cutoff condition is met, the sampling point layout plan is output.

[0043] Based on but not limited to the above method, the specific implementation process of step S1 is as follows:

[0044] S1.1: Extract relevant data content from geographic statistical data sets at multiple time levels based on the analysis objectives;

[0045] For example, in the field of environmental monitoring, it is necessary to analyze the spatiotemporal variations of air pollutants such as PM2.5 and PM10, and therefore extract concentration data for these pollutants from a collection of geographic statistical data at multiple time layers. In the field of weather forecasting, it is necessary to analyze the spatiotemporal variations of meteorological elements such as ground temperature and rainfall, and therefore extract data for these meteorological elements from a collection of geographic statistical data at multiple time layers.

[0046] S1.2: Based on the data set constructed in step S1.1, perform data regularization based on target requirements, divide the study area into several grids, and generate candidate points in each grid to obtain a candidate point set.

[0047] The candidate point sets obtained in S1.2 can be used as candidate locations for sampling points. Subsequent optimization algorithms then filter and optimize the set to produce the final set of sampling points. Grid segmentation and candidate point set generation are crucial steps in spatial sampling planning, directly impacting the accuracy and efficiency of the final sampling results. The spatiotemporal resolution of the segmentation is adjusted based on the analysis objectives.

[0048] Based on but not limited to the above method, the specific implementation process of step S2 is as follows:

[0049] S2.1: Initialize sampling from the candidate point set generated in step S1 to construct an initial sampling point set;

[0050] In the initial stage of spatial data collection, this step selects a set of initial sampling points according to certain rules or methods to construct an initial sampling set. For example, the optional initial sampling strategies include random sampling, uniform sampling, suppressive sampling and other non-adaptive sampling methods.

[0051] S2.2: Based on the initial sampling point set constructed in step S2.1, construct a spatiotemporal prediction model based on INLA-SPDE.

[0052] This step combines the spatial and temporal dimensions of the monitoring domain to establish a model capable of simultaneously predicting data in both space and time. INLA (Integrated Nested Laplace Approximation) is a Bayesian statistical inference method, and SPDE (Stochastic Partial Differential Equation) is a stochastic partial differential equation that effectively links continuous Gaussian random fields and discrete Gaussian Markov random fields, accelerating the INLA algorithm's approximation of the posterior distribution of the Bayesian spatiotemporal model. Combining these can construct an efficient and accurate spatiotemporal prediction model, providing spatiotemporal prediction in this embodiment.

[0053] Based on but not limited to the above method, the specific implementation process of step S3 is as follows:

[0054] S3.1: Based on the spatiotemporal prediction model constructed in step S2, use the sampling points to perform multi-time layer spatiotemporal prediction on the undeployed area, obtain the predicted value and confidence interval of each candidate point, and calculate the prediction error of the candidate point set;

[0055] S3.2: Extracting sampling positions of multiple time layers based on the calculated prediction error and the set sampling frequency and sampling quantity;

[0056] S3.3: Construct a new set of sampling points based on the extracted multi-time layer sampling positions;

[0057] S3.4: Based on the constructed sampling point set, determine whether the stopping requirement is met according to the set cutoff condition. If the condition is met, complete the sampling and execute step S3.6. If the condition is not met, execute step S3.5.

[0058] S3.5: Reconstruct the spatiotemporal prediction model using the constructed sampling point set, update the model parameters, and repeat steps S3.1 to S3.5;

[0059] S3.6: Output the set of sampling points that meet the cutoff conditions, that is, the layout plan.

[0060] It should be noted that the new sampling point set constructed in step S3.3 does not include the initial sampling point set.

[0061] In summary, the core technical process of the framework proposed in this paper includes the following two parts: (1) construction of a spatiotemporal prediction model based on INLA-SPDE; (2) spatiotemporal adaptive optimization sampling strategy. The detailed technology is as follows:

[0062] (1) A method for constructing a spatiotemporal prediction model based on INLA-SPDE, which belongs to the prior art and mainly includes the following steps:

[0063] Step 1: Assume that the spatial index of the geographic statistical data model is in two-dimensional space R 2 The study area is a fixed subset D∈R in the two-dimensional space. 2 Point s changes continuously in D, and Y(s,t) represents the variable attribute value of point s at time t. Y(s,t) is a random variable, which means it has the characteristics of spatial randomness and its value is uncertain. For a fixed time t, given n points s=s1,…,s n , n random variables Y(s0,t),…,Y(s n ,t)Because of the spatial autocorrelation, it shows that it has the characteristics of spatial structure.

[0064] Step 2: Let Y(s,t):s∈D denote a set of random variables, i.e., positions s=s1,…,s n The observation value Z(s,t) is a realization of the spatial random field.

[0065] Step 4: Determine the Gaussian random field and define its mean and covariance function. The covariance function is usually expressed using the Matern correlation function. The general form of the Matern function is:

[0066]

[0067]

[0068] where u=||s i -s j ||, represents the Euclidean distance between two points in space; is the decay rate of spatial correlation; v is the parameter that controls the degree of smoothness; K v is the second-order Bessel function, and Γ(·) is the gamma function.

[0069] Step 5: Let Z(s,t) represent the actual observed value of the variable at site s at time t, and its corresponding true value is characterized by the underlying Gaussian random process Y(s,t), and the two satisfy the following relationship:

[0070] Z(s,t)=β0+βX T (s,t)+Y(s,t)+ε(s,t)

[0071] Where s=s1,…,s n is the geographical location of the site; t is the time (day); X(X1,…,X n ) TSpatially varying covariates are also called predictors, (β0,…,β n ) is the regression coefficient of the covariate influencing factor, Y(s,t) is a Gaussian random field with a continuous spatial index; ε(s,t) is an error term that is independent of the spatial position, usually Gaussian white noise, that is, is the residual error term.

[0072] Step 6: There is noise in the observation process, so the spatial variation signal cannot be directly observed. Therefore, Y(s,t) is a latent Gaussian spatial random field (latent Gaussian process). Use the link function to build a first-order autoregressive model for the observed value and the latent variable attribute value Y(s,t):

[0073] Y(s,t)=C0+cX t +η(s,t)=ρY(s,t-1)+η(s,t)

[0074] Where η(s, t) is the residual random term, which is used to characterize the temporal and spatial random effects of the latent variables. It is assumed that η(s, t) is independent in time and satisfies the Gaussian process GP(0,∑ η ),in is the variance that does not vary with space, S η The covariance matrix representing the spatial correlation.

[0075] Step 7: In summary, the INLA spatiotemporal prediction model results are as follows:

[0076] Z t =β0+X T β+Y t +ε t

[0077] Y(s,t)=C0+cX t +ε(s,t)=ρY(s,t-1)+η(s,t)

[0078]

[0079]

[0080] Step 8: The parameters to be estimated in the spatiotemporal prediction model are Apart from All parameters except v have conjugate prior distributions, that is, after given prior distributions, β, ρ, The standard posterior distribution that meets the conditions can be obtained.

[0081] Step 9: Since the precision matrix of the continuous Gaussian random field is not a sparse matrix, the stochastic partial differential equation (SPDE) method is used. This method can approximate the continuous GF with a discrete GMRF.

[0082] (2) Spatiotemporal adaptive optimization sampling strategy

[0083] Traditional spatial sampling algorithms use data sets that only have spatial features between sampling points, not temporal features. Spatiotemporal sampling data sets contain not only the spatial features between sampling points at the same moment, but also the temporal features of the same location at different moments. Spatial adaptive sampling involves progressive spatial sampling at the same time slice, cycling through all time layers. The subsequent sampling relies on previous batches of data to optimize data collection and achieve the analysis goal.

[0084] In the geographic statistical design based on spatiotemporal adaptation, the sampling locations are spatially sampled at different times in the order of time layers. Spatial sampling needs to be performed on different time layers each time. The latter sampling depends on the previous sampling set on different time layers for data optimization and step-by-step sampling.

[0085] The spatiotemporal adaptive optimization sampling strategy specifically includes the following steps:

[0086] Step 1: Specify a finite set Indicates that within the study area D, n ) There are n moments * Sampling position x i,t ∈D, all x i,t All points ∈D can be selected as potential points (candidate points)

[0087] Step 2: Use non-adaptive design to select an initial set of sample locations, X 0,t ={x0∈D;t=t1,…,t n};

[0088] Step 3: Extract t1…t n The corresponding data Y on the time layer X0 0,t ={y0∈D;t=t1,…,t n}

[0089] Step 4: Estimate the parameters of a hypothetical spatiotemporal prediction model and construct the spatiotemporal prediction model f(X 0,t );

[0090] Step 5: Maximize the prediction difference at different time layers t1…t n Using f(X 0,t), perform the first spatial sampling, and set the number of sampling points to be b. Then at time t1, the new sampling point set At time t2, the new sampling point set In t n New sampling point set at time in Represents the set point obtained at the first sampling time t1. Due to the different data at different time layers, f(X 0,t ) In the prediction of different time layers, the generated sampling results are different, and the estimation error of each position in different time layers is calculated. The collection contains a collection of samples at different time levels.

[0091] Step 6: After the first sampling, a gradually expanding set of sampling points is formed:

[0092] Right now (The union of the removed duplicates.) If there are n time layers, and b samples are taken from each time layer, and assuming that the sampling locations of each time layer are different, then after one sampling, bn new candidate points are added to the set.

[0093] Step 7: Use the newly constructed sampling point set E1 to construct the spatiotemporal prediction model f(X 1,t ).

[0094] Step 8: Repeat Steps 5, 6, and 7, and continuously enhance the set of candidate data points. Sampling ends when the number of sampling points reaches the preset standard or there are no more potential sampling points available.

[0095] Based on the above technical solution, the key technical points of the present invention are:

[0096] Incorporating the spatiotemporal correlation and spatiotemporal variation characteristics of the target variable into the spatial sampling point planning method addresses the problem of traditional spatial sampling's limited attention to the spatiotemporal variability of the target variable. The constructed INLA-SPDE spatiotemporal model, based on Bayesian statistics, provides selected sampling points with the ability to predict unsampled areas. Aiming to minimize the cumulative prediction error over long time series, a spatiotemporal adaptive sampling method is used to generate a set of sampling points, accurately capturing the spatiotemporal variation of the target variable. The sampling points provide good spatiotemporal representativeness of the monitoring domain.

[0097] In order to verify the beneficial effects of the technical solution of the present invention, this embodiment takes Wuhan City as the research object, and compares the method of the present invention with the traditional sampling point layout method that does not take into account the spatiotemporal changes of pollutants. The target variable is LST, and the data source is USGC daily LST data. The data period is from January 1, 2021 to December 31, 2021. The LST image data of Wuhan City is divided into 5km grids, and a total of 400 grids in a single time layer are used to extract the coordinates of the pixel centroid points as a set of candidate points and take values. The final time layer of the research data is 12 layers (January-December), with 400 single-layer candidate points, for a total of 4,800 coordinate points. In order to stabilize the variance that increases with the mean value and make the distribution of the LST data approximate to normal, the logarithmic transformation rule is used.

[0098] The following table shows the prediction errors for undeployed areas at different numbers of sampling points compared to traditional sampling point placement methods. Table 1 shows that, compared to traditional sampling point placement strategies, the proposed method outperforms traditional spatial sampling schemes in long-term spatiotemporal prediction of undeployed areas. The cumulative error over long time series and multiple time periods, focusing on spatiotemporal variations, is significantly better than that of traditional spatial sampling schemes. Furthermore, over long time series, the proposed method effectively represents the environmental level of the monitored area, with efficiency improvements ranging from a minimum of 4.37% to a maximum of 15.56%, depending on the number of sampling points.

[0099] Table 1 Comparison of prediction results

[0100]

[0101] The advantage of this method lies in incorporating the spatiotemporal variation characteristics of the target variable into the spatial sampling point planning method, enabling the sampling points to accurately restore the spatiotemporal variation of the target variable in the monitoring domain, effectively addressing the issue of insufficient attention to the spatiotemporal variability of the target variable and improving the spatiotemporal representativeness of the sampling points for the monitoring domain. The spatial sampling point planning framework proposed by this method is applicable to a variety of application scenarios and needs. The constructed spatiotemporal variability modeling method, sampling point optimization algorithm, and evaluation index model are highly versatile and can be widely applied in the selection of monitoring sites for different targets, spatiotemporal analysis and prediction of geographic statistical data, and other fields, showing high practical value and application prospects.

[0102] In order to implement the above method, the embodiment of the present invention further provides a spatial sampling point planning device that takes into account the temporal and spatial variability, such as Figure 1 As shown, it includes the following modules:

[0103] Data set processing module, used to construct geographic statistical data sets from multiple time layers, grid the study area according to the data set and generate candidate point sets;

[0104] The spatiotemporal model construction module is used to initialize sampling based on the candidate point set and build an INLA-SPDE-based spatiotemporal prediction model based on the initial sampling points;

[0105] The spatiotemporal adaptive sampling module is used to perform multi-time layer spatiotemporal prediction and multi-time layer position sampling on the undeployed area through the spatiotemporal prediction model, generate a set of sampling points, and update the model parameters of the spatiotemporal prediction model according to the sampling point set. When the cutoff condition is met, the sampling point layout plan is output.

[0106] The embodiments of the present application may be provided as methods or computer program products. Thus, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of one or more computer-usable storage media containing computer-usable program code.

[0107] These computer program instructions can also be loaded onto a computer or other programmable data processing device or memory, so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable device to implement the process described in the framework flow. Figure 1 The steps of the functions specified in a process or multiple processes specifically include: S1: constructing a geographic statistical data set from multiple time layers, gridding the study area according to the data set and generating a candidate point set; S2: initializing sampling according to the candidate point set, and constructing a spatiotemporal prediction model based on INLA-SPDE; S3: performing multi-time layer spatiotemporal prediction and multi-time layer position sampling on the undeployed area through the spatiotemporal prediction model, generating a sampling point set, and updating the model parameters of the spatiotemporal prediction model according to the sampling point set. When the cutoff condition is met, the sampling point layout plan is output.

[0108] In addition, the logical instructions in the above-mentioned memory can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0109] On the other hand, an embodiment of the present invention further provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-mentioned spatial sampling point planning method, specifically including: S1: constructing a geographic statistical data set from multiple time layers, gridding the study area according to the data set and generating a candidate point set; S2: initializing sampling according to the candidate point set, and constructing a spatiotemporal prediction model based on INLA-SPDE; S3: performing multi-time layer spatiotemporal prediction and multi-time layer position sampling on the undeployed area through the spatiotemporal prediction model, generating a sampling point set, and updating the model parameters of the spatiotemporal prediction model according to the sampling point set. When the cutoff condition is met, the sampling point layout plan is output.

[0110] The present invention implements a spatial sampling point planning method, apparatus, device, and storage medium. By incorporating the spatiotemporal variation characteristics of the target variable into spatial sampling point planning, the sampling points can accurately reproduce the spatiotemporal variation of the target variable in the monitoring domain, effectively addressing the issue of insufficient attention to the spatiotemporal variability of the target variable and improving the spatiotemporal representativeness of the sampling points for the monitoring domain. The proposed spatial sampling point planning framework is applicable to a variety of application scenarios and requirements. The constructed spatiotemporal variability modeling method, sampling point optimization algorithm, and evaluation index model are highly versatile and can be widely applied in the selection of monitoring sites for different targets, spatiotemporal analysis and prediction of geographic statistical data, and other fields, showing high practical value and application prospects.

[0111] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or system comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or system. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or system comprising the element.

[0112] The serial numbers of the embodiments of the present invention are for descriptive purposes only and do not represent superiority or inferiority of the embodiments. In a unit claim that lists several means, several of these means may be embodied by the same item of hardware. The use of the terms first, second, and third, etc., does not denote any order and should be construed as identifiers.

[0113] The above are only preferred embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A spatial sampling point planning method that takes into account both temporal and spatial variability, characterized in that: The following steps are involved: S1: Construct a set of geographic statistical data from multiple time layers, grid the study area based on the data set and generate a set of candidate points; S2: Initialize sampling based on the candidate point set and build a spatiotemporal prediction model based on INLA-SPDE; S3: Perform multi-time-layer spatiotemporal prediction and multi-time-layer position sampling on the undeployed area through the spatiotemporal prediction model to generate a sampling point set. The model parameters of the spatiotemporal prediction model are updated based on the sampling point set. When the cutoff condition is met, the sampling point layout plan is output. Specifically include: S3.1: Based on the spatiotemporal prediction model constructed in step S2, use the sampling points to perform multi-time layer spatiotemporal prediction on the undeployed area, obtain the predicted value and confidence interval of each candidate point, and calculate the prediction error of the candidate point set; S3.2: Extracting sampling positions of multiple time layers based on the calculated prediction error and the set sampling frequency and sampling quantity; S3.3: Construct a new set of sampling points based on the extracted multi-time layer sampling positions; S3.4: Based on the constructed sampling point set, determine whether the stopping requirement is met according to the set cutoff condition. If the condition is met, complete the sampling and execute step S3.

6. If the condition is not met, execute step S3.

5. S3.5: Reconstruct the spatiotemporal prediction model using the constructed sampling point set, update the model parameters, and repeat steps S3.1 to S3.5; S3.6: Output the set of sampling points that meet the cutoff conditions, that is, the layout plan.

2. The spatial sampling point planning method according to claim 1, characterized in that: Step S1 specifically includes: S1.1: Extract relevant data content from geographic statistical data sets at multiple time levels based on the analysis objectives; S1.2: Based on the data set constructed in step S1.1, perform data regularization based on target requirements, divide the study area into multiple grids, and generate candidate points in each grid to obtain a candidate point set.

3. The spatial sampling point planning method according to claim 1, characterized in that: Step S2 specifically includes: S2.1: Initialize sampling from the candidate point set generated in step S1 to construct an initial sampling point set; S2.2: Based on the initial sampling point set constructed in step S2.1, construct a spatiotemporal prediction model based on INLA-SPDE.

4. The spatial sampling point planning method according to claim 1, characterized in that: In step S2, at the initial stage of spatial data collection, a set of initial sampling points is selected through a preset initialization sampling strategy to construct an initial sampling point set.

5. The spatial sampling point planning method according to claim 4, characterized in that: The preset initialization sampling strategies include random sampling, uniform sampling and suppressive sampling.

6. The spatial sampling point planning method according to claim 1, characterized in that: In step S3.3, the constructed new sampling point set does not include the initial sampling point set.

7. A spatial sampling point planning device that takes into account temporal and spatial variability, characterized in that: Includes the following modules: Data set processing module, used to construct geographic statistical data sets from multiple time layers, grid the study area according to the data set and generate candidate point sets; The spatiotemporal model building module is used to initialize sampling based on the candidate point set and build a spatiotemporal prediction model based on INLA-SPDE; The spatiotemporal adaptive sampling module is used to perform multi-time layer spatiotemporal prediction and multi-time layer position sampling on the undeployed area through the spatiotemporal prediction model, generate a sampling point set, and update the model parameters of the spatiotemporal prediction model based on the sampling point set. When the cutoff condition is met, the sampling point layout plan is output; Specifically include: S3.1: Based on the constructed spatiotemporal prediction model, use the sampling points to perform multi-time layer spatiotemporal prediction on the undeployed area, obtain the predicted value and confidence interval of each candidate point, and calculate the prediction error of the candidate point set; S3.2: Extracting sampling positions of multiple time layers based on the calculated prediction error and the set sampling frequency and sampling quantity; S3.3: Construct a new set of sampling points based on the extracted multi-time layer sampling positions; S3.4: Based on the constructed sampling point set, determine whether the stopping requirement is met according to the set cutoff condition. If the condition is met, complete the sampling and execute step S3.

6. If the condition is not met, execute step S3.

5. S3.5: Reconstruct the spatiotemporal prediction model using the constructed sampling point set, update the model parameters, and repeat steps S3.1 to S3.5; S3.6: Output the set of sampling points that meet the cutoff conditions, that is, the layout plan.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the spatial sampling point planning method according to any one of claims 1 to 6 are implemented.

9. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the spatial sampling point planning method according to any one of claims 1 to 6 are implemented.

Images