Ionospheric image generation method and device based on CNN multi-factor channel input
The relationship between multi-factor channel images and global ionosphere TEC images is constructed through the convolutional neural network method, and the global ionosphere TEC images are directly generated, which solves the shortcomings of traditional models in expressing small and medium-sized structures, and images independent variable factors to improve the The accuracy and feasibility of the ionosphere model are achieved.
Patent Information
- Application Number
- CN202411441585.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-16
- Publication Date
- 2024-12-31
- Estimated Expiration
- 2044-10-16
AI Technical Summary
Traditional empirical ionosphere models are difficult to effectively express the small and medium-scale structure of the ionosphere, and independent variable factors are difficult to express mathematically, resulting in low accuracy and high modeling difficulty.
Using a multi-factor channel input method based on convolutional neural network (CNN), the relationship between multi-factor channel images and global ionosphere TEC images is directly generated, avoiding the problem of excessive smoothing of mathematical function expression .
The problem of missing small-scale structures in the ionosphere is solved by mathematical models, and independent variable factors are expressed as image channels, which improves the fineness and feasibility of the ionosphere model.
Smart Images
Figure CN118967484B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of ionosphere modeling, and in particular relates to a method and a device for generating an ionosphere image based on multi-factor channel input of CNN. Background Art
[0002] Ionospheric error is an important error source for GNSS (Global Navigation Satellite System) positioning. Empirical ionospheric models are widely used in ionospheric error correction of major satellite navigation systems. However, traditional empirical ionospheric models usually use simple mathematical functions of longitude and latitude to express the horizontal distribution of ionospheric TEC (Total Electron Content). For example, the Klobuchar model used by GPS (Global Positioning System), the NeQuick model used by Galileo, and the BDGIM model used by BeiDou-3. The actual ionospheric TEC distribution has many irregular phenomena such as traveling disturbances, plasma bubbles, and tongue-shaped ionospheric anomalies. The ionospheric model expressed by the mathematical function is too smooth, and only retains the large-scale (thousands of kilometers) ionospheric horizontal structure, and it is difficult to retain the medium- and small-scale (hundreds of kilometers and below) ionospheric structure. Therefore, it has natural defects in expressing ionospheric TEC. The accuracy of the ionospheric model expressed by a simple mathematical model is low. If a complex mathematical model is used, although the ionospheric refinement can be improved to a certain extent, it increases the difficulty of modeling and there is a possibility of over-parameterization. Summary of the invention
[0003] In order to overcome the shortcomings of the above-mentioned prior art, the present invention provides an ionospheric image generation method and device based on CNN multi-factor channel input, which constructs an ionospheric model by generating ionospheric images, can express ionospheric irregularities, and avoid over-smoothing of global ionospheric modeling.
[0004] According to one aspect of the present invention, a method for generating an ionospheric image based on multi-factor channel input of a CNN is provided, comprising:
[0005] Construct a multi-factor channel image at the time to be calculated;
[0006] The multi-factor channel image at the time to be calculated is input into the trained global ionosphere empirical model based on the convolutional neural network CNN, and the global ionosphere TEC image at the time to be calculated is output; wherein the training of the global ionosphere empirical model based on the convolutional neural network CNN includes:
[0007] Obtain global ionospheric TEC images of a time series with a preset length;
[0008] Construct a multi-factor channel image for each moment of a time series of a preset length;
[0009] The multi-factor channel images at all times are paired with the global ionospheric TEC images at each time to form a sample data set;
[0010] A global ionosphere empirical model is constructed based on a convolutional neural network (CNN), wherein the input layer of the global ionosphere empirical model is a multi-factor channel image, and the output layer is a global ionosphere TEC image at the same time as the multi-factor channel image;
[0011] The sample data set is used to train the constructed global ionosphere empirical model to obtain a trained global ionosphere empirical model based on a convolutional neural network (CNN).
[0012] As a further technical solution, the construction of the multi-factor channel image includes: selecting the required ionospheric independent variable factors, each ionospheric independent variable factor corresponds to at least one factor channel, and forming a multi-factor channel image.
[0013] As a further technical solution, after selecting the required ionospheric independent variable factor, it also includes: performing sine and cosine transformation on each ionospheric independent variable factor to form two factor channels corresponding to one ionospheric independent variable factor.
[0014] As a further technical solution, the construction of the multi-factor channel image further includes: selecting an ionospheric factor that can be expressed in an image as a factor channel to form the multi-factor channel image.
[0015] As a further technical solution, the global ionosphere empirical model of the channel based on the convolutional neural network CNN includes an input layer, an output layer and three hidden layers, each of the hidden layers includes a 3D convolution and the convolution kernel size is n×n×the number of channels, wherein the third hidden layer is fully connected with the output layer, and n is 3, 5, 7 or 9.
[0016] According to one aspect of the present invention, there is provided an ionospheric image generating device based on multi-factor channel input of CNN, comprising:
[0017] The first main module is used to construct a multi-factor channel image at the time to be calculated;
[0018] The second main module is used to input the multi-factor channel image at the time to be calculated into the trained global ionosphere empirical model based on the convolutional neural network CNN, and output the global ionosphere TEC image at the time to be calculated; wherein the training of the global ionosphere empirical model based on the convolutional neural network CNN includes:
[0019] Obtain global ionospheric TEC images of a time series with a preset length;
[0020] Construct a multi-factor channel image for each moment of a time series of a preset length;
[0021] The multi-factor channel images at all times are paired with the global ionospheric TEC images at each time to form a sample data set;
[0022] A global ionosphere empirical model is constructed based on a convolutional neural network (CNN), wherein the input layer of the global ionosphere empirical model is a multi-factor channel image, and the output layer is a global ionosphere TEC image at the same time as the multi-factor channel image;
[0023] The sample data set is used to train the constructed global ionosphere empirical model to obtain a trained global ionosphere empirical model based on a convolutional neural network (CNN).
[0024] As a further technical solution, the first main module is also used to execute the following instructions: select the required ionospheric independent variable factors, each ionospheric independent variable factor corresponds to at least one factor channel, and form a multi-factor channel image.
[0025] As a further technical solution, the first main module is also used to execute the following instructions: perform sine and cosine transformation on each ionospheric independent variable factor to form two factor channels corresponding to one ionospheric independent variable factor.
[0026] As a further technical solution, the first main module is also used to execute the following instructions: select an ionospheric factor that can be expressed in an image as a factor channel to form a multi-factor channel image.
[0027] According to one aspect of the present specification, a non-transitory computer-readable storage medium is provided, wherein the non-transitory computer-readable storage medium stores computer instructions, wherein the computer instructions enable the computer to execute the method described above.
[0028] Compared with the prior art, the present invention has the following beneficial effects:
[0029] The present invention constructs an ionospheric model by generating an ionospheric image, thereby solving the problem of missing small and medium-scale structures when the mathematical model expresses regional or global ionospheric TEC; at the same time, the present invention also solves the problem that some ionospheric independent variable factors are difficult to express mathematically by expressing the independent variable factors as image channels.
[0030] The present invention constructs the relationship between the multi-factor channel image and the global ionospheric TEC image through a convolutional neural network (CNN) method, directly generates the global ionospheric TEC image, and avoids the over-smoothing problem caused by using mathematical functions to express ionospheric TEC.
[0031] The multi-factor channel used in the present invention can not only input the independent variable parameters required by the traditional ionospheric model such as longitude and latitude, but also input other factors that can be expressed graphically, such as surface elevation maps.
[0032] The present invention adopts a convolutional neural network to generate a global or regional ionospheric TEC image at one time, which solves the problem that the traditional mathematical function model can only generate a global image point by point. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, a brief introduction is given below to the drawings used in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0034] Figure 1 A schematic flow chart of a method for generating an ionospheric image based on multi-factor channel input of a CNN according to an embodiment of the present invention.
[0035] Figure 2 A schematic diagram of the training process of a global ionosphere empirical model based on a convolutional neural network (CNN) provided in an embodiment of the present invention.
[0036] Figure 3 A schematic diagram of the structure of an ionospheric image generation device with multi-factor channel input based on CNN provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0037] The traditional empirical ionosphere model has low accuracy if a simple mathematical model is used to express the horizontal distribution of ionosphere TEC. If a complex mathematical model is used, although the ionosphere fineness can be improved to a certain extent, the difficulty of modeling is increased, and there is a possibility of over-parameterization. To this end, the present invention provides a global ionosphere modeling method that can express ionosphere irregularities and avoid over-smoothing. Through a convolutional neural network method, a relationship between a multi-factor channel image and a global ionosphere TEC image is constructed, and a global ionosphere TEC image is directly generated, which solves the problem of missing small and medium-scale structures when the mathematical model expresses regional or global ionosphere TEC, and also solves the problem that some ionosphere independent variable factors are difficult to express mathematically.
[0038] The terms "including" and "having" and any variations thereof in the specification and claims of the present invention and the above-mentioned drawings are intended to cover non-exclusive inclusions, for example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to the steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products or apparatuses.
[0039] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices. The flowcharts shown in the accompanying drawings are merely exemplary and do not necessarily include all the contents and operations / steps, nor do they necessarily have to be executed in the order described. For example, some operations / steps may be further decomposed, while some operations / steps may be combined or partially combined, so the actual execution order may change according to the actual situation.
[0040] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. In addition, the technical features in the various embodiments or single embodiments provided by the present invention are arbitrarily combined with each other to form a new technical solution. This combination is not restricted by the sequence of steps and / or the structural composition mode, but must be based on the ability of ordinary technicians in this field to achieve. When the combination of technical solutions is contradictory or cannot be achieved, it should be considered that this combination of technical solutions does not exist and is not within the scope of protection required by the present invention.
[0041] The embodiment of the present invention provides an ionospheric image generation method based on multi-factor channel input of CNN, comprising: step 1, constructing a multi-factor channel image at the time to be calculated. Step 2, inputting the multi-factor channel image at the time to be calculated into a trained global ionosphere empirical model based on a convolutional neural network CNN, and outputting a global ionosphere TEC image at the time to be calculated. The embodiment of the present invention directly generates a global ionosphere TEC image by constructing a relationship between the multi-factor channel image and the global ionosphere TEC image, thereby solving the problem of missing small and medium-scale structures when the mathematical model expresses regional or global ionosphere TEC, and the problem that some ionosphere independent variable factors are difficult to express mathematically.
[0042] It should be noted that the multi-factor channel image refers to a multi-channel image composed of ionospheric independent variable factors, such as the geographical longitude and latitude that express the ionospheric position information, which can be used as a separate factor channel. Other factors that can be expressed graphically, such as surface elevation maps, can also be used as a separate factor channel.
[0043] In the ionospheric image generation method based on multi-factor channel input of CNN provided by an embodiment of the present invention, the training of the global ionosphere empirical model based on the convolutional neural network CNN includes: obtaining a global ionosphere TEC image of a time series of preset length; constructing a multi-factor channel image for each moment of the time series of preset length; pairing the multi-factor channel images at all moments with the global ionosphere TEC image according to the moment to form a sample data set; constructing a global ionosphere empirical model based on the convolutional neural network CNN, the input layer of the global ionosphere empirical model is the multi-factor channel image, and the output layer is the global ionosphere TEC image at the same moment as the multi-factor channel image; and using the sample data set to train the constructed global ionosphere empirical model to obtain a trained global ionosphere empirical model based on the convolutional neural network CNN.
[0044] As a preferred embodiment, the ionospheric image generation method based on multi-factor channel input of CNN provided in the embodiment of the present invention constructs a global ionosphere empirical model based on convolutional neural network CNN based on CODE global ionosphere map, such as Figure 1 As shown, the specific process includes:
[0045] Step (1) Download the CODE Global Ionosphere Map (GIM) of a solar activity cycle (11 years from 2010 to 2020) in IONEX format. Read the CODE GIM product in IONEX format and save it as a long-term series of global ionosphere images with an image size of 71×73.
[0046] Step (2), for the first moment in step (1), construct a multi-factor channel image. The independent variable factors selected in this embodiment are: geographic longitude (B), geographic latitude (L), and annual cumulative day (D). In order to avoid the discontinuity of the independent variable factors, a 6-channel image is constructed as a multi-factor channel image through sine and cosine transformation. The 6 channels are cos(B), sin(B), cos(L), sin(L), cos(D / 365), sin(D / 365), and the channel image size is 71×73.
[0047] Step (3), repeat step (2) to generate multi-factor channel images at all times.
[0048] Step (4): Pair the generated multi-factor channel images at all times with the global ionosphere images by time, and randomly select 80% of them as training sets and 20% as validation sets.
[0049] Step (5), constructing a 6-channel input CNN as a global ionosphere empirical model based on a convolutional neural network CNN, the input image size of the model is 71×73. The CNN network architecture adopted by the global ionosphere empirical model based on the convolutional neural network CNN includes three hidden layers, each of which includes 3D convolution, normalization, and nonlinear excitation. Among them, the convolution kernel size is 5×5×the number of channels, the convolution step is 1, the number of feature operators in each layer is set to 32, and the edge supplement is set to 2.
[0050] The output layer of the global ionosphere empirical model based on the convolutional neural network CNN is the global ionosphere map (GIM). Specifically, the third hidden layer is fully connected with the output layer, and the output layer is the global ionosphere TEC image at the same time as the multi-factor channel image.
[0051] Step (6), input the training set and validation set in step (4) into the global ionosphere empirical model based on the convolutional neural network CNN constructed in step (5), train until the model converges, and save the network parameters.
[0052] Based on the contents of the above embodiments, as a preferred embodiment, the ionospheric image generation method based on multi-factor channel input of CNN provided in the embodiment of the present invention uses the trained global ionosphere empirical model based on convolutional neural network CNN to generate ionosphere images, such as Figure 2 As shown, specifically including:
[0053] Step S1, construct a multi-factor channel image at the time to be calculated. The selected ionospheric independent variable factors include geographic longitude (B), geographic latitude (L) and annual cumulative day (D). In addition, to avoid the discontinuity of the independent variable factors, sine and cosine transformations are used to construct images of 6 channels as multi-factor channel images.
[0054] Furthermore, taking 0:00 on July 1, 2021 as an example, a 6-channel image is constructed as input, namely cos(B), sin(B), cos(L), sin(L), cos(D / 365), sin(D / 365). Among them, D is the annual cumulative day corresponding to 0:00 on July 1, 2021, which is 182.0 days.
[0055] Step S2, inputting the multi-factor channel image at the time to be calculated into the trained global ionosphere empirical model based on convolutional neural network CNN, and outputting the global ionosphere TEC image at the time to be calculated.
[0056] Specifically, the 6-channel image constructed in step S1 is used as the input of the global ionosphere empirical model based on the convolutional neural network CNN to generate the global ionosphere TEC image at the required time. Similarly, the global ionosphere images at other times can also be calculated similarly.
[0057] The implementation basis of each embodiment of the present invention is to implement programmed processing through a device with a processor function. Therefore, in engineering practice, the technical solutions and functions of each embodiment of the present invention are encapsulated into various modules. Based on this reality, on the basis of the above embodiments, an embodiment of the present invention provides an ionospheric image generation device based on multi-factor channel input of CNN, which is used to execute the ionospheric image generation method based on multi-factor channel input of CNN in the above method embodiment.
[0058] See also Figure 3 The device comprises: a first main module, which is used to construct a multi-factor channel image at a time to be calculated; a second main module, which is used to input the multi-factor channel image at the time to be calculated into a trained global ionosphere empirical model based on a convolutional neural network (CNN), and output a global ionosphere TEC image at the time to be calculated; wherein the training of the global ionosphere empirical model based on the convolutional neural network (CNN) comprises: obtaining a global ionosphere TEC image of a time series of preset length; constructing a multi-factor channel image for each moment of the time series of preset length; pairing the multi-factor channel images at all moments with the global ionosphere TEC image according to the moment to form a sample data set; constructing a global ionosphere empirical model based on the convolutional neural network (CNN), wherein the input layer of the global ionosphere empirical model is the multi-factor channel image, and the output layer is the global ionosphere TEC image at the same moment as the multi-factor channel image; and using the sample data set to train the constructed global ionosphere empirical model to obtain a trained global ionosphere empirical model based on the convolutional neural network (CNN).
[0059] The ionospheric image generation device based on CNN multi-factor channel input provided by the embodiment of the present invention aims at the problems existing in the traditional ionospheric empirical model. Figure 3 By using several modules in the ionospheric model to generate ionospheric images, the ionospheric model is constructed to solve the problem of missing small and medium-scale structures when the mathematical model expresses regional or global ionospheric TEC. At the same time, by expressing the independent variable factors as image channels, the problem of the difficulty in mathematically expressing some ionospheric independent variable factors is also solved.
[0060] It should be noted that the device embodiments provided by the present invention are used to implement the methods in the above method embodiments as well as the methods in other method embodiments provided by the present invention. The only difference is that the corresponding functional modules are set, and the principle is basically the same as the principle of the above device embodiments provided by the present invention. As long as the technical personnel in the field refer to the specific technical solutions in other method embodiments on the basis of the above device embodiments, obtain the corresponding technical means and the technical solutions composed of these technical means by combining the technical features, and on the premise of ensuring the practicality of the technical solutions, improve the modules in the above device embodiments to obtain the corresponding device class embodiments, which are used to implement the methods in other method class embodiments. For example:
[0061] Based on the content of the above device embodiment, as a preferred embodiment, in the ionospheric image generation device based on multi-factor channel input of CNN provided in the embodiment of the present invention, the first main module is also used to execute the following instructions:
[0062] The required ionospheric independent variable factors are selected, each ionospheric independent variable factor corresponds to at least one factor channel, and a multi-factor channel image is formed.
[0063] Based on the content of the above device embodiment, as a preferred embodiment, in the ionospheric image generation device based on multi-factor channel input of CNN provided in the embodiment of the present invention, the first main module is also used to execute the following instructions:
[0064] Each ionospheric independent variable factor is subjected to sine and cosine transformation to form two factor channels corresponding to one ionospheric independent variable factor.
[0065] Based on the content of the above device embodiment, as a preferred embodiment, in the ionospheric image generation device based on multi-factor channel input of CNN provided in the embodiment of the present invention, the first main module is also used to execute the following instructions:
[0066] The ionospheric factor that can be expressed in a graphical way is selected as a factor channel to form a multi-factor channel image.
[0067] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0068] In addition, when the logic instructions in the above method flow are implemented in the form of software functional units and sold or used as independent products, they are stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or partly in other words, the part that contributes to the prior art or part of the technical solution is embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for a computer device (a personal computer, a server, or a network device) to perform all or part of the steps of the method described in each method embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or CD-ROM various media for storing program codes.
[0069] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0070] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0071] These computer program instructions can also be loaded onto a computer or other programmable data processing device 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 implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0072] The system embodiments described above are merely illustrative, wherein the units described as separate components are or are not physically separated, and the components shown as units are or are not physical units, located in one place, or distributed on multiple network units. Some or all of the modules are selected according to actual conditions to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement without creative effort.
[0073] In summary, the present invention constructs the relationship between the multi-factor channel image and the global ionospheric TEC image through the convolutional neural network method, and directly generates the global ionospheric TEC image. The innovation of the present invention lies in that the ionospheric model is constructed by generating an ionospheric image, which solves the problem of missing small and medium-scale structures when the mathematical model expresses regional or global ionospheric TEC; at the same time, the present invention also solves the problem that some ionospheric independent variable factors are difficult to express mathematically by expressing the independent variable factors as image channels.
[0074] The present invention has the following advantages:
[0075] 1. Directly generate ionospheric TEC images, avoiding excessive smoothing caused by using mathematical functions to express ionospheric TEC.
[0076] 2. In addition to the independent variable parameters required by traditional ionospheric models such as longitude and latitude, the multi-channel factors used can also input other factors that can be expressed graphically, such as surface elevation maps.
[0077] 3. The convolutional neural network can generate global or regional ionospheric TEC images at one time, which solves the problem that the traditional mathematical function model can only generate global images point by point.
[0078] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.
Claims
1. The ionospheric image generation method based on multi-factor channel input of CNN is characterized by: include: Construct a multi-factor channel image at the time to be calculated; The multi-factor channel image at the time to be calculated is input into the trained global ionosphere empirical model based on the convolutional neural network CNN, and the global ionosphere TEC image at the time to be calculated is output; wherein the training of the global ionosphere empirical model based on the convolutional neural network CNN includes: Obtain global ionospheric TEC images of a time series with a preset length; Construct a multi-factor channel image for each moment of a time series of a preset length; The multi-factor channel images at all times are paired with the global ionospheric TEC images at each time to form a sample data set; A global ionosphere empirical model is constructed based on a convolutional neural network (CNN), wherein the input layer of the global ionosphere empirical model is a multi-factor channel image, and the output layer is a global ionosphere TEC image at the same time as the multi-factor channel image; The sample data set is used to train the constructed global ionosphere empirical model to obtain a trained global ionosphere empirical model based on a convolutional neural network (CNN).
2. The ionospheric image generation method based on multi-factor channel input of CNN according to claim 1 is characterized in that: The construction of the multi-factor channel image includes: selecting required ionospheric independent variable factors, each ionospheric independent variable factor corresponds to at least one factor channel, and forming a multi-factor channel image.
3. The ionospheric image generation method based on multi-factor channel input of CNN according to claim 2 is characterized in that: After selecting the required ionospheric independent variable factors, the method further includes: performing sine and cosine transformation on each ionospheric independent variable factor to form two factor channels corresponding to one ionospheric independent variable factor.
4. The ionospheric image generation method based on multi-factor channel input of CNN according to claim 2 is characterized in that: The construction of the multi-factor channel image also includes: selecting an ionospheric factor that can be expressed in an image as a factor channel to form the multi-factor channel image.
5. The ionospheric image generation method based on multi-factor channel input of CNN according to claim 1 is characterized in that: The global ionosphere empirical model of the channel based on the convolutional neural network CNN includes an input layer, an output layer and three hidden layers, each of the hidden layers includes a 3D convolution and the convolution kernel size is n×n×the number of channels, wherein the third hidden layer is fully connected with the output layer, and n is 3, 5, 7 or 9.
6. An ionospheric image generation device based on CNN multi-factor channel input, characterized in that: include: The first main module is used to construct a multi-factor channel image at the time to be calculated; The second main module is used to input the multi-factor channel image at the time to be calculated into the trained global ionosphere empirical model based on the convolutional neural network CNN, and output the global ionosphere TEC image at the time to be calculated; wherein the training of the global ionosphere empirical model based on the convolutional neural network CNN includes: Obtain global ionospheric TEC images of a time series with a preset length; Construct a multi-factor channel image for each moment of a time series of a preset length; The multi-factor channel images at all times are paired with the global ionospheric TEC images at each time to form a sample data set; A global ionosphere empirical model is constructed based on a convolutional neural network (CNN), wherein the input layer of the global ionosphere empirical model is a multi-factor channel image, and the output layer is a global ionosphere TEC image at the same time as the multi-factor channel image; The sample data set is used to train the constructed global ionosphere empirical model to obtain a trained global ionosphere empirical model based on a convolutional neural network (CNN).
7. The ionospheric image generation device based on multi-factor channel input of CNN according to claim 6, characterized in that: The first main module is also used to execute the following instructions: select the required ionospheric independent variable factors, each ionospheric independent variable factor corresponds to at least one factor channel, and form a multi-factor channel image.
8. The ionospheric image generation device based on CNN multi-factor channel input according to claim 7, characterized in that: The first main module is also used to execute the following instructions: perform sine and cosine transformation on each ionospheric independent variable factor to form two factor channels corresponding to one ionospheric independent variable factor.
9. The ionospheric image generation device based on CNN multi-factor channel input according to claim 7, characterized in that: The first main module is also used to execute the following instructions: select an ionospheric factor that can be expressed in an image as a factor channel to form a multi-factor channel image.
10. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium stores computer instructions, which cause the computer to execute the method of any one of claims 1 to 5.