Snow depth inversion method and device based on multi-source data fusion

Abstract

The invention discloses an accumulated snow depth inversion method and device based on multi-source data fusion, and the method comprises the steps: obtaining first snow depth data through a snow depth product, obtaining second snow depth data actually measured by a site, and obtaining independent variable data; the independent variable data comprises brightness temperature data, ground surface coverage data, forest coverage rate data, altitude data, latitude and longitude data, ground surface classification data and geographic coordinates; fusing the first snow depth data and the second snow depth data to obtain fused snow depth data; combining different independent variable combinations, and fusing the snow depth data as a data set to train a random forest model to obtain a plurality of snow depth prediction initial models; and evaluating the prediction precision of the plurality of snow depth prediction initial models, and taking the model with the highest prediction precision as a target snow depth prediction model. According to the method, the snow depth product and the site actually measured snow depth are fused, and the fused high-resolution snow depth data are obtained to serve as the target variable of the random forest algorithm, so that the model precision is improved.

Description

Technical Field

[0001] The present application relates to the technical field of snow depth detection, and in particular to a snow depth inversion method and device based on multi-source data fusion. Background Art

[0002] Snow is a vital component of the cryosphere, a major source of water for many rivers and a crucial source of groundwater. It serves as a natural reservoir for rivers and lakes worldwide. However, excessive snowfall and meltwater can lead to severe snowstorms and floods, threatening human life and negatively impacting agriculture, transportation, and other areas. Snow depth is a key snowpack parameter, and quantitative monitoring plays a crucial role in snowmelt runoff forecasting, water resource management, and flood control.

[0003] Extensive research has been conducted in the field of snow remote sensing, using microwave remote sensing to measure snow depth. Snow depth retrieval algorithms primarily include passive microwave and active microwave methods. However, various factors affect the accuracy of snow retrieval. For example, the low spatial distribution of passive microwave data, the deep snow saturation effect of microwave detection, vegetation, and moisture content prevent retrieval results based on passive microwave remote sensing data from accurately reflecting the actual snow cover properties. Active microwave remote sensing is highly sensitive to snow depth, but currently there are no active microwave satellites that meet the spatial and temporal requirements for snow depth observation, making it impossible to perform long-term series retrieval of snow depth over large areas. Summary of the Invention

[0004] The present disclosure provides a snow depth inversion method, apparatus, device and storage medium based on multi-source data fusion to at least solve the above technical problems existing in the prior art.

[0005] According to a first aspect of the present application, a snow depth inversion method based on multi-source data fusion is provided, the method comprising:

[0006] Using a snow depth product, first snow depth data, second snow depth data measured at a site, and independent variable data are obtained; the independent variable data include brightness temperature data, surface cover data, forest cover data, altitude data, longitude and latitude data, surface classification data, and geographic coordinates; the first snow depth data is snow depth data of a target area that is continuously distributed in time and space;

[0007] fusing the first snow depth data and the second snow depth data to obtain fused snow depth data;

[0008] The random forest model was trained using different combinations of independent variables and fused snow depth data as data sets to obtain multiple initial snow depth prediction models.

[0009] The prediction accuracy of the multiple snow depth prediction initial models is evaluated, and the model with the highest prediction accuracy is used as the target snow depth prediction model.

[0010] In one possible implementation, the first snow depth data and the second snow depth data are fused using an optimal interpolation method to obtain fused snow depth data.

[0011] In one embodiment, different combinations of independent variables are used as input data, including:

[0012] Based on the brightness temperature data, the remaining independent variables are selected for different combinations to obtain different independent variable combinations.

[0013] In one embodiment, different combinations of independent variables include:

[0014] Brightness temperature data, land cover data, forest cover data, altitude data, latitude and longitude data, land surface classification data, and geographic coordinates; or

[0015] Brightness temperature data, forest cover data, altitude data, latitude and longitude data, land surface classification data, and geographic coordinates; or

[0016] Brightness temperature data, land cover data, altitude data, latitude and longitude data, land surface classification data, and geographic coordinates; or

[0017] Brightness temperature data, land cover data, forest cover data, land surface classification data, and geographic coordinates.

[0018] In one embodiment, the random forest model is trained to obtain multiple snow depth prediction initial models, including:

[0019] Dividing the data set into training samples and test samples;

[0020] The random forest model is trained with the training samples, and the trained random forest model is cross-validated with the test samples, and the random forest model with a validation result higher than a preset threshold is output as an initial model for snow depth prediction; wherein, the training samples use different independent variable combinations as input data and use fused snow depth data as output data.

[0021] In one possible implementation, the random forest model is trained for different seasons to obtain multiple initial snow depth prediction models;

[0022] On the basis of the same season, the prediction accuracy of the multiple snow depth prediction initial models is evaluated using the root mean square error, bias and correlation coefficient of the multiple snow depth prediction initial models.

[0023] In one possible implementation, after obtaining the target snow depth prediction model, the method further includes:

[0024] Obtain independent variable data for the target area and target time period;

[0025] The independent variable data is input into the target snow depth prediction model to obtain the snow depth.

[0026] According to a second aspect of the present application, a snow depth inversion method and apparatus based on multi-source data fusion is provided, the apparatus comprising:

[0027] An acquisition module is configured to use a snow depth product to acquire first snow depth data, second snow depth data measured at a site, and independent variable data; the independent variable data includes brightness temperature data, surface cover data, forest cover data, altitude data, longitude and latitude data, surface classification data, and geographic coordinates; the first snow depth data is snow depth data for a target area that is continuously distributed in time and space;

[0028] a fusion module, configured to fuse the first snow depth data and the second snow depth data to obtain fused snow depth data;

[0029] The training module is used to train the random forest model using different combinations of independent variables and fused snow depth data as data sets to obtain multiple initial snow depth prediction models;

[0030] The output module is used to evaluate the prediction accuracy of multiple snow depth prediction initial models and use the model with the highest prediction accuracy as the target snow depth prediction model.

[0031] According to a third aspect of the present application, an electronic device is provided, including:

[0032] at least one processor; and

[0033] a memory communicatively connected to the at least one processor; wherein,

[0034] The memory stores instructions that can be executed by the at least one processor. The instructions are executed by the at least one processor to enable the at least one processor to perform the method described in this application.

[0035] According to a fourth aspect of the present application, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute the method described in the present application.

[0036] According to a fifth aspect of the present application, a computer program product is provided, comprising a computer program or instructions, which implement the method described in the present application when executed by a processor.

[0037] Using the technical solution of this application, the snow depth product and the actual snow depth measured at the site are integrated as the target variable of the random forest algorithm to improve the model accuracy.

[0038] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The above and other objects, features and advantages of the exemplary embodiments of the present application will become readily understood by reading the detailed description below with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present application are shown in an illustrative and non-limiting manner, in which:

[0040] In the drawings, the same or corresponding reference numerals denote the same or corresponding parts.

[0041] Figure 1 The schematic diagram of the implementation process of the snow depth inversion method based on multi-source data fusion in the embodiment of the present application is shown. Figure 1 ;

[0042] Figure 2 The schematic diagram of the implementation process of the snow depth inversion method based on multi-source data fusion in the embodiment of the present application is shown. Figure 2 ;

[0043] Figure 3 The schematic diagram of the implementation process of the snow depth inversion method based on multi-source data fusion in the embodiment of the present application is shown. Figure 3 ;

[0044] Figure 4 A schematic structural diagram of a snow depth inversion device based on multi-source data fusion in an embodiment of the present application is shown;

[0045] Figure 5 A schematic diagram of the structure of an electronic device in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0046] In order to make the purpose, features, and advantages of this application more obvious and easy to understand, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of this application.

[0047] Existing snow depth inversion algorithms include semi-empirical algorithms, model-based snow depth inversion algorithms, data fusion and assimilation algorithms, machine learning algorithms, and active microwave snow depth inversion algorithms.

[0048] Semi-empirical algorithms are based on fitting the linear relationship between snow depth and brightness temperature gradient. Chang et al. initially used NimBus-7-SMMR data and, based on the radiative transfer equation for homogeneous snow, set the snow particle size and density to fixed values ​​(0.3 mm and 300 kg / m³). Snow depth was retrieved by calculating the relationship between the brightness temperature difference between 18 GHz and 36 GHz and the measured snow depth. This semi-empirical algorithm is essentially a linear brightness temperature gradient method. Due to a lack of research on the interference of snow physical properties on microwave signals, the model's inversion accuracy varies significantly across regions, its universality is poor, and brightness temperature difference saturation occurs in deep snow. Model-based snow depth inversion algorithms, such as the dense media radiative transfer (DMRT) model, retrieve snow depth by simulating microwave radiative transfer within the snow layer. This model calculates the effective propagation constant of electromagnetic waves in snow and uses it to correct the snow's extinction coefficient, scattering coefficient, and albedo. DMRT effectively simulates the linear relationship between snow depth and passive microwave brightness temperature difference under varying snow properties. However, the model calculation is complex and requires many input parameters, making it difficult to achieve snow depth inversion over a large area and for a long time.

[0049] Data fusion and assimilation algorithms combine process models with observation models for snow depth estimation. Based on Bayesian theory and controlled by satellite observations, they improve the a priori snow depth state simulated by the process model. This method combines the strengths of satellite data, ground station data, and model simulation data, leveraging the complementary advantages of observational data from different sources to improve the quality of snow depth data. However, the success of this method is highly dependent on the quality of both the observations and the model. Land surface process models often require a large amount of input data, including meteorological and surface parameters. Some parameters are often difficult to obtain, which in turn reduces the applicability of data assimilation algorithms.

[0050] Machine learning methods utilize machine learning models to invert snow depth based on passive microwave brightness temperature data, meteorological station data, and other auxiliary data. Machine learning methods can effectively describe the complex, nonlinear relationship between snow depth and passive microwave brightness temperature. This method has wide applicability and high inversion accuracy, and can, to a certain extent, overcome the limitations of linear algorithms in different regions. However, currently, most machine learning algorithms utilize station data. Due to the scarcity of stations at high altitudes, machine learning algorithms that use station snow depth as the ground truth often have lower accuracy at high altitudes.

[0051] Active microwave snow depth retrieval algorithms exploit the correlation between different radar backscatter coefficients and snow depth to retrieve snow depth. Yueh et al. and King et al., using airborne and ground-based X- and Ku-band scatterometers to measure multipolarization backscatter from snow, have also demonstrated that high-frequency X- and Ku-band radars are more suitable for snow depth retrieval. However, there are currently no satellites capable of meeting the temporal and spatial requirements for global snow depth observation.

[0052] The snow depth inversion method based on multi-source data fusion provided in this application fuses the mountain snow depth product and the measured snow depth at the site to obtain the fused high-resolution snow depth data. Then, a random forest model is constructed with passive microwave brightness temperature and remote sensing auxiliary data as prediction variables and the high-resolution snow depth product as the target variable, and the model is used to invert the snow depth.

[0053] The following describes a snow depth inversion method and device based on multi-source data fusion provided by the present application in conjunction with the accompanying drawings.

[0054] like Figure 1 As shown, the present application provides a snow depth inversion method based on multi-source data fusion, the method comprising:

[0055] S101, using a snow depth product to obtain first snow depth data, second snow depth data measured at a site, and independent variable data; the independent variable data includes brightness temperature data, surface cover data, forest cover data, altitude data, longitude and latitude data, surface classification data, and geographic coordinates; the first snow depth data is snow depth data of a target area that is continuously distributed in time and space;

[0056] It should be noted that snow depth products are products or systems used to measure the thickness of snow layers and are widely used in meteorology, transportation, agriculture and other fields. Snow depth can be accurately measured through different technical means, such as GNSS snow depth products, laser snow depth monitoring stations or ultrasonic snow depth monitoring stations. The snow depth product provided in this application can be used to detect snow depth in mountainous areas, for example, it can be the Sentinel-1 snow depth product. Figure 2 As shown, the first snow depth data obtained by the Sentinel-1 snow depth product is the snow depth data of the target area with continuous temporal and spatial distribution. In addition, this application also needs to obtain the site measured brightness temperature data, surface cover data, forest cover data, altitude data, longitude and latitude data, surface classification data and geographic coordinates; in this application, the surface classification data can be converted into grid data of preset resolution according to the geographic coordinates, and the grid data can match the microwave brightness temperature data.

[0057] In this application, brightness temperature data, surface cover data, forest cover data, altitude data, latitude and longitude data, surface classification data and geographic coordinates are selected as independent variables because brightness temperature data provides basic scattering information and can be used to invert snow depth. Surface cover data has a certain correlation with snow depth, and snow cover can be used to determine whether snow exists. For the target snow depth, if the surface cover of a pixel is 0, the snow depth value of the pixel is set to 0. Snow depth inversion algorithms usually perform better in open and sparsely forested areas, but face obvious difficulties in dense forest areas. Therefore, in this algorithm, forest cover is regarded as an input feature to help reduce the uncertainty associated with vegetation. Geographic location and altitude variables help improve the performance of machine learning models, so latitude and longitude and elevation will also be used as input data for the model. At the same time, considering that different surface types have an impact on microwave scattering, surface classification data is added as model input.

[0058] S102, fusing the first snow depth data and the second snow depth data to obtain fused snow depth data;

[0059] It is understandable that the Sentinel-1 snow depth product provides snow depth data of mountainous areas that are continuously distributed in time and space, and the average snow depth at the passive microwave pixel scale can be obtained. Since the prediction variable is at the passive microwave pixel scale, theoretically, the use of the Sentinel-1 average snow depth in mountainous areas as the target variable of the machine learning algorithm is more representative than the site snow depth, and can improve the accuracy of the model obtained after training. Therefore, the optimal interpolation method is used in this application to fuse the first snow depth data and the second snow depth data to obtain fused snow depth data. The Seninel-1 mountain snow depth product and the measured snow depth at the site are fused as the target variable of the random forest algorithm through the optimal interpolation method to improve the accuracy of the model. And because the snow accumulation season is intense, the present invention performs model training by season.

[0060] S103, training the random forest model using different independent variable combinations and fused snow depth data as data sets to obtain multiple snow depth prediction initial models;

[0061] It should be noted that the independent variable data are sorted according to their impact on snow depth data as follows: temperature data, surface cover data, forest cover data, altitude data, longitude and latitude data, surface classification data, and geographic coordinates. In this application, different combinations of independent variables can be used to obtain different independent variable combinations that affect snow depth. The random forest model is trained using different independent variable combinations to obtain multiple initial snow depth prediction models corresponding to different independent variable combinations.

[0062] It is understandable that in this application, data preprocessing can also be performed on the obtained independent variable combination, such as deleting duplicate data, normalizing, resampling and reprojecting the data, and then feature selection is performed on the preprocessed independent variable combination data, so that the feature number is input into the random forest model for training.

[0063] like Figure 3 As shown, the characteristic of the first snow depth data in this application is the spatiotemporal analysis of the long-term series of snow depth in the target area, such as the intra-annual variation characteristics of snow depth on the Qinghai-Tibet Plateau and the inter-annual variation characteristics of snow depth in the mountainous areas of the Qinghai-Tibet Plateau. Among them, the inter-annual variation characteristics of snow depth in the mountainous areas of the Qinghai-Tibet Plateau include the inter-annual variation trend and significance analysis of the average snow depth during the snow accumulation period and the inter-annual variation trend and significance analysis of the monthly average snow depth.

[0064] S104: Evaluate the prediction accuracy of the multiple initial snow depth prediction models, and use the model with the highest prediction accuracy as the target snow depth prediction model.

[0065] It is understandable that multiple snow depth prediction initial models can all predict snow depth, but because different combinations of independent variables have different effects on snow depth, the prediction accuracy of multiple snow depth prediction initial models is different. In this application, after obtaining multiple snow depth prediction initial models, the prediction accuracy of multiple snow depth prediction initial models can be evaluated, and then they can be sorted from large to small according to the prediction accuracy. The model with the highest ranking is used as the target snow depth prediction model, and then the target snow depth prediction model is used to predict the snow depth.

[0066] In some embodiments, different combinations of independent variables are used as input data, including:

[0067] Based on the brightness temperature data, the remaining independent variables are selected for different combinations to obtain different independent variable combinations.

[0068] It should be noted that, therefore, in order to determine the most appropriate combination of independent variables, this application, based on only brightness temperature data as a predictor variable, sequentially adds other predictor variables to determine whether remote sensing auxiliary data will improve the accuracy of the random forest algorithm.

[0069] In some embodiments, different combinations of independent variables include:

[0070] Brightness temperature data, land cover data, forest cover data, altitude data, latitude and longitude data, land surface classification data, and geographic coordinates; or

[0071] Brightness temperature data, forest cover data, altitude data, latitude and longitude data, land surface classification data, and geographic coordinates; or

[0072] Brightness temperature data, land cover data, altitude data, latitude and longitude data, land surface classification data, and geographic coordinates; or

[0073] Brightness temperature data, land cover data, latitude and longitude data, land surface classification data and geographic coordinates;

[0074] Brightness temperature data, land cover data, forest cover data, land surface classification data, and geographic coordinates.

[0075] It is understandable that the independent variable combination in this application may also include brightness temperature data and surface coverage data, or brightness temperature data and forest coverage data, etc., which will not be repeated in this application.

[0076] In some embodiments, the random forest model is trained to obtain multiple snow depth prediction initial models, including:

[0077] Dividing the data set into training samples and test samples;

[0078] The random forest model is trained with the training samples, and the trained random forest model is cross-validated with the test samples, and the random forest model with a validation result higher than a preset threshold is output as an initial model for snow depth prediction; wherein, the training samples use different independent variable combinations as input data and use fused snow depth data as output data.

[0079] In this application, for different independent variable combinations, 80% of the samples were randomly selected as training samples, and the remaining 20% ​​of the samples were used as test samples for cross-validation.

[0080] In some embodiments, a random forest model is trained for different seasons to obtain multiple initial snow depth prediction models;

[0081] On the basis of the same season, the prediction accuracy of the multiple snow depth prediction initial models is evaluated using the root mean square error, bias and correlation coefficient of the multiple snow depth prediction initial models.

[0082] It should be noted that the first snow depth data in the present application is snow depth data of the target area and continuously distributed in time and space, and different target snow depth prediction models are trained according to different seasons, so that the snow depth can be predicted more accurately. It should be noted that, on the basis of the same season, the present application adopts different independent variable combinations to obtain the target snow depth prediction initial model with the highest prediction accuracy. After obtaining multiple snow depth prediction initial models, RMSE (root mean square error), MAE (mean square error) and correlation coefficient are used to evaluate the inversion performance of each model, and the model with the highest accuracy is selected as the final model of the winter snow depth inversion of the present invention. Among them, the prediction accuracy evaluation includes overall accuracy evaluation, accuracy evaluation of different snow depth ranges and accuracy evaluation of interannual variation of snow depth.

[0083] In some embodiments, after obtaining the target snow depth prediction model, the method further includes:

[0084] Obtain independent variable data for the target area and target time period;

[0085] The independent variable data is input into the target snow depth prediction model to obtain the snow depth.

[0086] After obtaining the trained target snow depth prediction model of this application, the independent variable data for the target area and target time period can be obtained. The independent variable data has the same attributes as the independent variable data used to train the target snow depth prediction model, thereby obtaining the predicted snow depth.

[0087] like Figure 4 As shown, an embodiment of the present application provides a snow depth inversion device based on multi-source data fusion, the device comprising:

[0088] Acquisition module 401 is used to obtain first snow depth data using a snow depth product, obtain second snow depth data measured at a site, and obtain independent variable data; the independent variable data includes brightness temperature data, surface cover data, forest cover data, altitude data, longitude and latitude data, surface classification data, and geographic coordinates; the first snow depth data is snow depth data of a target area that is continuously distributed in time and space;

[0089] A fusion module 402 is configured to fuse the first snow depth data and the second snow depth data to obtain fused snow depth data;

[0090] A training module 403 is used to train the random forest model using different independent variable combinations and integrated snow depth data as a data set to obtain multiple snow depth prediction initial models;

[0091] The output module 404 is used to evaluate the prediction accuracy of the multiple snow depth prediction initial models and use the model with the highest prediction accuracy as the target snow depth prediction model.

[0092] The snow depth inversion device based on multi-source data fusion provided in the present application comprises an acquisition module 401 which uses a snow depth product to acquire first snow depth data, acquires second snow depth data measured at a site, and acquires independent variable data; the independent variable data include brightness temperature data, surface coverage data, forest coverage data, altitude data, longitude and latitude data, surface classification data, and geographic coordinates; the first snow depth data is snow depth data of a target area and is continuously distributed in time and space; the fusion module 402 fuses the first snow depth data and the second snow depth data to obtain fused snow depth data; the training module 403 trains a random forest model using different independent variable combinations and fused snow depth data as data sets to obtain multiple snow depth prediction initial models; the output module 404 evaluates the prediction accuracy of multiple snow depth prediction initial models, and uses the model with the highest prediction accuracy as the target snow depth prediction model.

[0093] According to an embodiment of the present application, the present application also provides an electronic device and a readable storage medium.

[0094] The electronic device includes at least one processor and a memory in communication with the at least one processor. The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the snow depth inversion method based on multi-source data fusion described in this application. The computer instructions are used to cause the computer to perform the snow depth inversion method based on multi-source data fusion described in this application.

[0095] The present application also provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the snow depth inversion method based on multi-source data fusion of the present application.

[0096] Figure 5 A schematic block diagram of an example electronic device 800 that can be used to implement an embodiment of the present application is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present application described and / or claimed herein.

[0097] like Figure 5 As shown, the device 800 includes a computing unit 801, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 802 or a computer program loaded from a storage unit 808 into a random access memory (RAM) 803. Various programs and data required for the operation of the device 800 can also be stored in the RAM 803. The computing unit 801, the ROM 802, and the RAM 803 are connected to each other via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0098] Various components in device 800 are connected to I / O interface 805, including an input unit 806, such as a keyboard, mouse, etc.; an output unit 807, such as various types of displays, speakers, etc.; a storage unit 808, such as a magnetic disk, optical disk, etc.; and a communication unit 809, such as a network card, modem, wireless communication transceiver, etc. The communication unit 809 allows device 800 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0099] The computing unit 801 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 801 performs the various methods and processes described above, such as the snow depth inversion method based on multi-source data fusion. For example, in some embodiments, the snow depth inversion method based on multi-source data fusion can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 808. In some embodiments, part or all of the computer program can be loaded and / or installed onto the device 800 via the ROM 802 and / or the communication unit 809. When the computer program is loaded into the RAM 803 and executed by the computing unit 801, one or more steps of the snow depth inversion method based on multi-source data fusion described above can be performed. Alternatively, in other embodiments, the computing unit 801 may be configured in any other appropriate manner (eg, by means of firmware) to execute the snow depth inversion method based on multi-source data fusion.

[0100] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0101] The program code for implementing the methods of the present application can be written in any combination of one or more programming languages. Such program code can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the program code is executed by the processor or controller, the functions / operations specified in the flow charts and / or block diagrams are implemented. The program code can be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0102] In the context of the present application, a machine-readable medium can be a tangible medium that can contain or store a program for use by an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared or semiconductor system, device or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium can include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0103] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0104] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0105] A computer system may include a client and a server. The client and server are generally remote from each other and typically interact through a communication network. The client-server relationship arises through computer programs running on the respective computers and having a client-server relationship with each other. The server may be a cloud server, a server in a distributed system, or a server integrated with a blockchain.

[0106] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A snow depth inversion method based on multi-source data fusion, characterized in that: The method comprises: Using a snow depth product, first snow depth data, second snow depth data measured at a site, and independent variable data are obtained; the independent variable data include brightness temperature data, surface cover data, forest cover data, altitude data, longitude and latitude data, surface classification data, and geographic coordinates; the first snow depth data is snow depth data of a target area that is continuously distributed in time and space; fusing the first snow depth data and the second snow depth data to obtain fused snow depth data; The random forest model was trained using different combinations of independent variables and fused snow depth data as data sets to obtain multiple initial snow depth prediction models. The prediction accuracy of the multiple snow depth prediction initial models is evaluated, and the model with the highest prediction accuracy is used as the target snow depth prediction model.

2. The method according to claim 1, characterized in that The first snow depth data and the second snow depth data are fused using an optimal interpolation method to obtain fused snow depth data.

3. The method according to claim 1, characterized in that Different combinations of independent variables are used as input data, including: Based on the brightness temperature data, the remaining independent variables are selected for different combinations to obtain different independent variable combinations.

4. The method according to claim 3, characterized in that Different combinations of independent variables, including: Brightness temperature data, land cover data, forest cover data, altitude data, latitude and longitude data, land surface classification data, and geographic coordinates; or Brightness temperature data, forest cover data, altitude data, latitude and longitude data, land surface classification data, and geographic coordinates; or Brightness temperature data, land cover data, altitude data, latitude and longitude data, land surface classification data, and geographic coordinates; or Brightness temperature data, land cover data, latitude and longitude data, land surface classification data and geographic coordinates; Brightness temperature data, land cover data, forest cover data, land surface classification data, and geographic coordinates.

5. The method according to claim 1, wherein The random forest model was trained to obtain multiple initial snow depth prediction models, including: Dividing the data set into training samples and test samples; The random forest model is trained with the training samples, and the trained random forest model is cross-validated with the test samples, and the random forest model with a validation result higher than a preset threshold is output as an initial model for snow depth prediction; wherein, the training samples use different independent variable combinations as input data and use fused snow depth data as output data.

6. The method according to claim 1, wherein The random forest model is trained for different seasons to obtain multiple initial models for snow depth prediction; On the basis of the same season, the prediction accuracy of the multiple snow depth prediction initial models is evaluated using the root mean square error, bias and correlation coefficient of the multiple snow depth prediction initial models.

7. The method according to claim 6, characterized in that After obtaining the target snow depth prediction model, it also includes: Obtain independent variable data for the target area and target time period; The independent variable data is input into the target snow depth prediction model to obtain the snow depth.

8. A snow depth inversion method and device based on multi-source data fusion, characterized in that: The device comprises: An acquisition module is configured to use a snow depth product to acquire first snow depth data, second snow depth data measured at a site, and independent variable data; the independent variable data includes brightness temperature data, surface cover data, forest cover data, altitude data, longitude and latitude data, surface classification data, and geographic coordinates; the first snow depth data is snow depth data for a target area that is continuously distributed in time and space; a fusion module, configured to fuse the first snow depth data and the second snow depth data to obtain fused snow depth data; The training module is used to train the random forest model using different combinations of independent variables and fused snow depth data as data sets to obtain multiple initial snow depth prediction models; The output module is used to evaluate the prediction accuracy of multiple snow depth prediction initial models and use the model with the highest prediction accuracy as the target snow depth prediction model.

9. An electronic device, characterized in that: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 8.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to cause a computer to execute the method according to any one of claims 1 to 8.

Images