Spatial data interpolation method, device and electronic equipment

By combining deep neural network super-resolution interpolation, inverse distance weight interpolation and Kriging interpolation, the performance differences of spatial data interpolation methods in the existing technology under different data sets are solved, and higher accuracy and adaptability are achieved. The spatial data is closer to the actual data.

CN114418843BActive Publication Date: 2024-11-26THE 54TH RESEARCH INSTITUTE OF CHINA ELECTRONICS TECHNOLOGY GROUP CORPORATION
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202111604032.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-12-24
Publication Date
2024-11-26
Estimated Expiration
2041-12-24

AI Technical Summary

Technical Problem

In the prior art, a single type of spatial data interpolation method shows large performance differences in different input data sets, which is difficult to meet the needs of high precision and adaptability.

Method used

Using a combination of deep neural network super-resolution interpolation, inverse distance weight interpolation and Kriging interpolation, the results of super-resolution interpolation processing of deep neural network are added with the inverse distance weight interpolation processing results, and the results are interpolated with Kriging interpolation processing. The result is stacked, and high-resolution spatial data is output.

Benefits of technology

It improves the accuracy and adaptability of spatial data, reduces prediction errors, and makes the generated spatial data closer to the actual data, solving the problem of smoothing effect of a single interpolation algorithm and insufficient processing of high-frequency feature.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114418843B_ABST
    Figure CN114418843B_ABST
Patent Text Reader

Abstract

The present invention is applicable to the field of spatial data processing technology, and provides a spatial data interpolation method, device and electronic device, the method comprising: obtaining first spatial data; inputting the first spatial data into a pre-trained spatial data interpolation model for interpolation processing, and outputting second spatial data; wherein the resolution of the second spatial data is higher than the resolution of the first spatial data, and the spatial data interpolation model performs deep neural network super-resolution interpolation processing, inverse distance weighted interpolation processing and Kriging interpolation processing on the first spatial data, respectively, and adds the result of the deep neural network super-resolution interpolation processing and the result of the inverse distance weighted interpolation processing, and then stacks them with the result of the Kriging interpolation processing to obtain the second spatial data and output it. The present invention can improve the accuracy of spatial data, and has good accuracy and adaptability.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of spatial data processing, and in particular relates to a spatial data interpolation method, device and electronic equipment. Background Art

[0002] Spatial data is a digital expression of the surface morphology of the terrain and has a wide range of applications in surveying and mapping, hydrology, meteorology, landforms, geology, soil, engineering construction, communications, and military affairs.

[0003] In practical applications, in order to analyze or display spatial features more precisely, higher-resolution spatial data is required. However, ultra-high-precision measured spatial data is usually difficult to obtain. Although there are many types of interpolation methods in the prior art to improve the accuracy of spatial data, they each have different advantages and disadvantages. When the input spatial data sets are different, the performance effects of each method vary greatly. Therefore, a single type of interpolation method cannot meet actual needs in terms of accuracy and adaptability. Summary of the invention

[0004] In view of this, embodiments of the present invention provide a spatial data interpolation method, device and electronic device to improve the precision of spatial data, and have good accuracy and adaptability.

[0005] A first aspect of an embodiment of the present invention provides a spatial data interpolation method, comprising:

[0006] Acquire first spatial data;

[0007] The first spatial data is input into a pre-trained spatial data interpolation model for interpolation processing, and the second spatial data is output; wherein the resolution of the second spatial data is higher than the resolution of the first spatial data, and the spatial data interpolation model performs deep neural network super-resolution interpolation processing, inverse distance weighted interpolation processing and Kriging interpolation processing on the first spatial data, respectively, and the result of the deep neural network super-resolution interpolation processing is added to the result of the inverse distance weighted interpolation processing, and then stacked with the result of the Kriging interpolation processing to obtain the second spatial data and output it.

[0008] Optionally, performing deep neural network super-resolution interpolation processing on the first spatial data includes:

[0009] Performing a convolution operation on the first spatial data to obtain first processed data;

[0010] Inputting the first processed data into the residual structure for processing, and performing a convolution operation on the first processed data after the residual structure processing to obtain second processed data;

[0011] adding the first processed data and the second processed data to obtain third processed data;

[0012] The third processed data is upsampled to obtain a deep neural network super-resolution interpolation processing result.

[0013] Optionally, the residual structure is composed of a plurality of residual blocks connected in sequence; each residual block processes the input first processed data as follows:

[0014] The first processed data input is sequentially subjected to convolution operation, activation function processing and convolution operation, and the result is added to the first processed data before being input into the residual block and then output.

[0015] Optionally, each convolution layer that performs a convolution operation on the data is a convolution layer with 64 channels, a stride of 1, and a convolution kernel of 3*3.

[0016] Optionally, performing deep neural network super-resolution interpolation processing on the first spatial data further includes:

[0017] Before performing the convolution operation on the first spatial data, the first spatial data is preprocessed to remove noise interference in the first spatial data.

[0018] Optionally, both the inverse distance weighted interpolation process and the kriging interpolation process are four-fold interpolation.

[0019] Optionally, the training process of the spatial data interpolation model includes:

[0020] Acquire multiple third space data as verification data sets;

[0021] Downsampling each third space data to obtain a training data set;

[0022] Based on the training data set and the validation data set, model parameters of the spatial data interpolation model are trained to obtain a trained spatial data interpolation model.

[0023] A second aspect of an embodiment of the present invention provides a spatial data interpolation device, comprising:

[0024] An acquisition module, used for acquiring first spatial data;

[0025] An interpolation module is used to input the first spatial data into a pre-trained spatial data interpolation model for interpolation processing and output second spatial data; wherein the resolution of the second spatial data is higher than that of the first spatial data, and the spatial data interpolation model performs deep neural network super-resolution interpolation processing, inverse distance weighted interpolation processing and kriging interpolation processing on the first spatial data respectively, and adds the result of the deep neural network super-resolution interpolation processing and the result of the inverse distance weighted interpolation processing, and then stacks them with the result of the kriging interpolation processing to obtain the second spatial data and output it.

[0026] A third aspect of an embodiment of 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 when the processor executes the computer program, the steps of the data interpolation method of the first aspect described above are implemented.

[0027] A fourth aspect of an embodiment of the present invention provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the data interpolation method of the first aspect are implemented.

[0028] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0029] The spatial data interpolation model established by the embodiment of the present invention removes the influence of low-frequency changes on the deep neural network super-resolution interpolation algorithm by adding the result of the deep neural network super-resolution interpolation algorithm with the result of the inverse distance weight interpolation algorithm, so that it can better process the high-frequency characteristics of spatial data, focus on training the residual part that cannot be amplified by the inverse distance weight algorithm, and better explore the correlation between spatial data; then, the result is stacked with the output result of the Kriging interpolation algorithm, which solves the smoothing effect problem of the Kriging interpolation algorithm, making the prediction error smaller, and the generated spatial data is closer to the actual data. Compared with any single interpolation algorithm in the prior art, the embodiment of the present invention can better improve the accuracy of spatial data and has good accuracy and adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0031] Figure 1 is a schematic diagram of an implementation flow of a spatial data interpolation method provided by an embodiment of the present invention;

[0032] Figure 2 is a schematic diagram of the structure of a spatial data interpolation model provided by an embodiment of the present invention;

[0033] Figure 3 It is a schematic diagram of the structure of the DNCNN spatial data preprocessing network provided by an embodiment of the present invention;

[0034] Figure 4 is a schematic structural diagram of a spatial data interpolation device provided by an embodiment of the present invention;

[0035] Figure 5It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0036] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0037] In order to illustrate the technical solution of the present invention, a specific embodiment is provided below for illustration.

[0038] The single type of interpolation method in the prior art has different advantages and disadvantages. When the input spatial data sets are different, the performance effects of the single interpolation method vary greatly. For example: the advantage of the Inverse Distance Weight (IDW) method is that it complies with the basic principles of spatial data and has a fast calculation speed, but it ignores the correlation between the spatial distribution of the data and the terrain characteristics; the Kriging interpolation method (Krige) gives the best linear unbiased interpolation under the condition of minimum interpolation variance, and the interpolation result is closer to the actual situation, but it has a smoothing effect problem and is not sensitive enough to the high-frequency characteristics of spatial data. Therefore, the embodiment of the present invention combines the advantages and disadvantages of the above single interpolation methods to design an interpolation method with higher accuracy and wider application range.

[0039] See also Figure 1 The spatial data interpolation method provided by the embodiment of the present invention includes:

[0040] Step S101, obtaining first spatial data.

[0041] Step S102: input the first spatial data into a pre-trained spatial data interpolation model for interpolation processing, and output second spatial data; wherein the resolution of the second spatial data is higher than that of the first spatial data, and the spatial data interpolation model performs deep neural network super-resolution interpolation processing, inverse distance weighted interpolation processing, and kriging interpolation processing on the first spatial data, respectively, and adds the result of the deep neural network super-resolution interpolation processing to the result of the inverse distance weighted interpolation processing, and then stacks them with the result of the kriging interpolation processing to obtain the second spatial data and output it.

[0042] In an embodiment of the present invention, the first spatial data may be original low-resolution spatial data. By inputting the low-resolution spatial data into a pre-trained spatial data interpolation model for interpolation processing, the resolution of the spatial data may be improved to obtain second spatial data with relatively high resolution.

[0043] For details, see Figure 2 As shown in the figure, the spatial data interpolation model has three interpolation processing branches, namely Kriging interpolation, deep neural network super-resolution interpolation and inverse distance weighted interpolation. The three interpolation algorithms interpolate the input low-resolution spatial data respectively. Then, the output result of the deep neural network super-resolution interpolation algorithm is first added to the output result of the inverse distance weighted interpolation algorithm, so that the output result of the deep neural network super-resolution interpolation algorithm focuses on the residual part that cannot be amplified by training the inverse distance weight, better explores the correlation between spatial data, and avoids overfitting of the deep neural network; then, by conecting and stacking with the output result of the Kriging interpolation algorithm, while retaining the best linear unbiased interpolation characteristics of the Kriging interpolation method, the smoothing effect problem of the Kriging interpolation algorithm is solved, making the algorithm more applicable and the generated spatial data closer to the actual data.

[0044] It can be seen that the spatial data interpolation model established by the embodiment of the present invention removes the influence of low-frequency changes on the deep neural network super-resolution interpolation algorithm by adding the results of the deep neural network super-resolution interpolation algorithm with the results of the inverse distance weighted interpolation algorithm, so that it can better process the high-frequency features of spatial data, focus on training the residual part that cannot be amplified by the inverse distance weighted algorithm, and better explore the correlation between spatial data; then, the result is stacked with the output result of the Kriging interpolation algorithm, which solves the smoothing effect problem of the Kriging interpolation algorithm, making the prediction error smaller, and the generated spatial data is closer to the actual data. Compared with any single interpolation algorithm in the prior art, the embodiment of the present invention can better improve the accuracy of spatial data and has good accuracy and adaptability.

[0045] Optionally, in step S102, deep neural network super-resolution interpolation processing is performed on the first spatial data, which can be described in detail as follows:

[0046] Step S1021, performing a convolution operation on the first spatial data to obtain first processed data;

[0047] Step S1022, inputting the first processed data into the residual structure for processing, and performing a convolution operation on the first processed data after the residual structure processing to obtain second processed data;

[0048] Step S1023, adding the first processed data and the second processed data to obtain third processed data;

[0049] Step S1024, up-sample the third processed data to obtain a deep neural network super-resolution interpolation processing result.

[0050] Optionally, performing deep neural network super-resolution interpolation processing on the first spatial data in step S102 may also include:

[0051] Before executing step S1021 to perform a convolution operation on the first spatial data, the first spatial data is preprocessed to remove noise interference in the first spatial data.

[0052] In the embodiments of the present invention, see Figure 2 As shown in Figure 2, the deep neural network super-resolution interpolation part can be composed of two parts: the DNCNN spatial data preprocessing network and the spatial data super-resolution network. The structure of the DNCNN spatial data preprocessing network can be found in Figure 3 As shown in the figure, the main function is to remove the noise effects other than the terrain features in the spatial data. The spatial data super-resolution network consists of several COVN convolutional layers, residual structures, and subpixel upsampling. Its main function is to amplify the terrain change characteristics of the spatial data.

[0053] Optional, see Figure 2 As shown, the residual structure in step S1022 is composed of a plurality of residual blocks ResBlock connected in sequence. In this embodiment, the residual structure is composed of 16 residual blocks ResBlock. The residual structure can solve the gradient vanishing problem of deep neural networks.

[0054] Each residual block ResBlock includes a COVN convolution layer, an activation function layer and a COVN convolution layer.

[0055] Each residual block ResBlock processes the first processed data as follows:

[0056] The first processed data input is sequentially subjected to convolution operation, activation function processing and convolution operation, and the result is added to the first processed data before being input into the residual block and then output.

[0057] Optionally, each convolution layer that performs convolution operation on the data in the above steps can be a convolution layer with 64 channels, a step size of 1, and a convolution kernel of 3*3.

[0058] Optionally, the inverse distance weighted interpolation processing and the Kriging interpolation processing in the above steps can both be four-fold interpolation to achieve four-fold interpolation of spatial data.

[0059] Optionally, the training process of the spatial data interpolation model includes:

[0060] Acquire multiple third space data as verification data sets;

[0061] Downsampling each third space data to obtain a training data set;

[0062] Based on the training data set and the validation data set, model parameters of the spatial data interpolation model are trained to obtain a trained spatial data interpolation model.

[0063] In an embodiment of the present invention, the third spatial data can be high-resolution spatial data as a validation data set. After downsampling the third spatial data, a low-resolution training data set is obtained. Specifically, the 30m elevation data can be used as a basis, and the data can be cut into 100,000 256*256 regional elevation block areas, and the most intuitive loss1 loss (the average difference between the elevation value and the corresponding point) is taken as the training loss, and the optimizer is the adam optimizer. During training, the training machine is read as a true map, and it is normalized, and linearly scaled to a small block of 32*32 as training data, and the inverse distance weight interpolation is used as an intermediate result. According to the size of the training set and the size of the video memory of the training computer, the training batch is set to 16, and a total of 60 epochs are trained. The initial learning rate is 0.0001, and the learning rate drops to one-third of the original every 15 epochs. Training is performed through the above steps to obtain a trained network.

[0064] The multi-method fusion spatial data interpolation method designed in the embodiment of the present invention can convert low-resolution spatial data into high-resolution spatial data, improve the precision of spatial data, and improve the delicacy of spatial data display, and can be widely used in various application scenarios. For example: in engineering construction, it can be used for earthwork volume calculation, line of sight analysis, etc.; in wireless communication, it can be used for communication site analysis, coverage analysis, etc.; in the military field, it can be used for battlefield environment 3D modeling, digital twin technology, etc.

[0065] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0066] See also Figure 4 As shown, an embodiment of the present invention provides a spatial data interpolation device 40, comprising:

[0067] The acquisition module 41 is used to acquire first spatial data.

[0068] The interpolation module 42 is used to input the first spatial data into a pre-trained spatial data interpolation model for interpolation processing and output second spatial data; wherein the resolution of the second spatial data is higher than the resolution of the first spatial data, and the spatial data interpolation model performs deep neural network super-resolution interpolation processing, inverse distance weighted interpolation processing and kriging interpolation processing on the first spatial data, respectively, and adds the result of the deep neural network super-resolution interpolation processing to the result of the inverse distance weighted interpolation processing, and then stacks it with the result of the kriging interpolation processing to obtain the second spatial data and output it.

[0069] Optionally, the interpolation module 42 is specifically used for:

[0070] Performing a convolution operation on the first spatial data to obtain first processed data;

[0071] Inputting the first processed data into the residual structure for processing, and performing a convolution operation on the first processed data after the residual structure processing to obtain second processed data;

[0072] adding the first processed data and the second processed data to obtain third processed data;

[0073] The third processed data is upsampled to obtain a deep neural network super-resolution interpolation processing result.

[0074] Optionally, the residual structure consists of multiple residual blocks connected in sequence.

[0075] Each residual block processes the first processed data as follows:

[0076] The first processed data input is sequentially subjected to convolution operation, activation function processing and convolution operation, and the result is added to the first processed data before being input into the residual block and then output.

[0077] Optionally, each convolution layer that performs a convolution operation on the data is a convolution layer with 64 channels, a stride of 1, and a convolution kernel of 3*3.

[0078] Optionally, the interpolation module 42 is further used for:

[0079] Before performing the convolution operation on the first spatial data, the first spatial data is preprocessed to remove noise interference in the first spatial data.

[0080] Optionally, both the inverse distance weighted interpolation process and the kriging interpolation process are four-fold interpolation.

[0081] Optionally, the interpolation module 42 is further used for:

[0082] Acquire multiple third space data as verification data sets;

[0083] Downsampling each third space data to obtain a training data set;

[0084] Based on the training data set and the validation data set, model parameters of the spatial data interpolation model are trained to obtain a trained spatial data interpolation model.

[0085] Figure 5 FIG. 5 is a schematic diagram of an electronic device 50 provided by an embodiment of the present invention. Figure 5 As shown, the electronic device 50 of this embodiment includes: a processor 51, a memory 52, and a computer program 53 stored in the memory 52 and executable on the processor 51, such as a spatial data interpolation program. When the processor 51 executes the computer program 53, the steps in the above-mentioned various spatial data interpolation method embodiments are implemented, such as Figure 1 Alternatively, when the processor 51 executes the computer program 53, the functions of each module in the above-mentioned device embodiments are realized, for example Figure 4 The functions of the modules 41 to 42 are shown.

[0086] Exemplarily, the computer program 53 may be divided into one or more modules / units, one or more modules / units are stored in the memory 52, and executed by the processor 51 to complete the present invention. One or more modules / units may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program 53 in the electronic device 50. For example, the computer program 53 may be divided into an acquisition module 41 and an interpolation module 42 (modules in the virtual device), and the specific functions of each module are as follows:

[0087] The acquisition module 41 is used to acquire first spatial data.

[0088] The interpolation module 42 is used to input the first spatial data into a pre-trained spatial data interpolation model for interpolation processing and output second spatial data; wherein the resolution of the second spatial data is higher than the resolution of the first spatial data, and the spatial data interpolation model performs deep neural network super-resolution interpolation processing, inverse distance weighted interpolation processing and kriging interpolation processing on the first spatial data, respectively, and adds the result of the deep neural network super-resolution interpolation processing to the result of the inverse distance weighted interpolation processing, and then stacks it with the result of the kriging interpolation processing to obtain the second spatial data and output it.

[0089] The electronic device 50 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The electronic device 50 may include, but is not limited to, a processor 51 and a memory 52. ​​Those skilled in the art will appreciate that Figure 5It is only an example of the electronic device 50 and does not constitute a limitation of the electronic device 50. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the electronic device 50 may also include input and output devices, network access devices, buses, etc.

[0090] The processor 51 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0091] The memory 52 may be an internal storage unit of the electronic device 50, such as a hard disk or memory of the electronic device 50. The memory 52 may also be an external storage device of the electronic device 50, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 50. Further, the memory 52 may also include both an internal storage unit of the electronic device 50 and an external storage device. The memory 52 is used to store computer programs and other programs and data required by the electronic device 50. The memory 52 may also be used to temporarily store data that has been output or is to be output.

[0092] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0093] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0094] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0095] In the embodiments provided by the present invention, it should be understood that the disclosed devices / electronic devices and methods can be implemented in other ways. For example, the device / electronic device embodiments described above are only schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0096] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0097] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0098] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electric carrier signals and telecommunication signals.

[0099] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A spatial data interpolation method, characterized in that: include: Acquire first spatial data; Inputting the first spatial data into a pre-trained spatial data interpolation model for interpolation processing, and outputting second spatial data; The resolution of the second spatial data is higher than that of the first spatial data, and the spatial data interpolation model performs deep neural network super-resolution interpolation processing, inverse distance weighted interpolation processing and kriging interpolation processing on the first spatial data respectively, and adds the result of the deep neural network super-resolution interpolation processing and the result of the inverse distance weighted interpolation processing, and then stacks them with the result of the kriging interpolation processing to obtain the second spatial data and output it; Performing deep neural network super-resolution interpolation processing on the first spatial data includes: Preprocessing the first spatial data to remove noise interference in the first spatial data; Performing a convolution operation on the first spatial data to obtain first processed data; Inputting the first processed data into a residual structure for processing, and performing a convolution operation on the first processed data after the residual structure processing to obtain second processed data; adding the first processed data and the second processed data to obtain third processed data; Upsampling the third processed data to obtain a deep neural network super-resolution interpolation processing result; The residual structure is composed of a plurality of residual blocks connected in sequence; Each residual block processes the first processed data as follows: The first processed data input is sequentially subjected to convolution operation, activation function processing and convolution operation, and the result is added to the first processed data before being input into the residual block and then output.

2. The spatial data interpolation method according to claim 1, characterized in that: Each convolution layer that performs convolution operations on the data is a convolution layer with 64 channels, a stride of 1, and a convolution kernel of 3*3.

3. The spatial data interpolation method according to claim 1, characterized in that: The inverse distance weighted interpolation process and the Kriging interpolation process are both four-fold interpolation.

4. The spatial data interpolation method according to any one of claims 1 to 3, characterized in that: The training process of the spatial data interpolation model includes: Acquire multiple third space data as verification data sets; Downsampling each third space data to obtain a training data set; Based on the training data set and the validation data set, model parameters of the spatial data interpolation model are trained to obtain a trained spatial data interpolation model.

5. A spatial data interpolation device, characterized in that: include: An acquisition module, used for acquiring first spatial data; An interpolation module is used to input the first spatial data into a pre-trained spatial data interpolation model for interpolation processing, and output second spatial data; wherein the resolution of the second spatial data is higher than the resolution of the first spatial data, and the spatial data interpolation model performs deep neural network super-resolution interpolation processing, inverse distance weighted interpolation processing and kriging interpolation processing on the first spatial data, respectively, and adds the result of the deep neural network super-resolution interpolation processing and the result of the inverse distance weighted interpolation processing, and then stacks them with the result of the kriging interpolation processing to obtain the second spatial data and output it; Performing deep neural network super-resolution interpolation processing on the first spatial data includes: Preprocessing the first spatial data to remove noise interference in the first spatial data; Performing a convolution operation on the first spatial data to obtain first processed data; Inputting the first processed data into a residual structure for processing, and performing a convolution operation on the first processed data after the residual structure processing to obtain second processed data; adding the first processed data and the second processed data to obtain third processed data; Upsampling the third processed data to obtain a deep neural network super-resolution interpolation processing result; The residual structure is composed of a plurality of residual blocks connected in sequence; Each residual block processes the first processed data as follows: The first processed data input is sequentially subjected to convolution operation, activation function processing and convolution operation, and the result is added to the first processed data before being input into the residual block and then output.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Citation Information

Patent Citations

  • Scale adaptation interpolation method and scale adaptation interpolation system

    CN104102845A

  • Resolution improvement method and system based on deep convolution prediction and interpolation

    CN110136061A

  • Distributed multi-algorithm fused meteorological data interpolation method

    CN110609923A

  • Layered image super-resolution method and system, computer equipment and application

    CN111968036A

  • Model training method, super-resolution perception method and related device

    CN113807395A