Cloud layer simulation method and device based on multi-source data fusion

Abstract

The invention relates to a cloud simulation method and device based on multi-source data fusion. The method comprises the following steps: step 1, collecting multi-source data from a satellite, a radar and a ground meteorological station; 2, data of different data sources are preprocessed, for example, image enhancement and filtering technologies are adopted for satellite images, radar data are calibrated, and ground observation data are cleaned and standardized; and step 3, establishing a hierarchical fusion model, fusing satellite image data and radar data in a spatial dimension, then fusing ground observation data and the preliminary model in a feature dimension, and adopting a weighted fusion algorithm. And 4, constructing a cloud layer physical model considering physical factors such as heating power, power and water vapor transportation based on the fused data, and realizing visualization by using a volume rendering method. And step 5, establishing a real-time updating mechanism to ensure that the model reflects the newest cloud layer state and can be applied to a plurality of fields such as weather forecast, environment monitoring, aviation navigation and the like. According to the invention, high-precision cloud layer simulation is realized, and the method has a wide application prospect.

Description

Technical Field

[0001] The present invention relates to meteorological simulation technology, in particular to the application of computer graphics and data science in the field of meteorology. More specifically, the present invention is a method and device for cloud simulation using multi-source data fusion technology. The present invention generates a more accurate and dynamic cloud model by integrating satellite images, radar data and ground observation data. The technology has broad application prospects in weather forecasting, environmental monitoring, aviation navigation, urban planning, and virtual reality and augmented reality applications. Background Art

[0002] Traditional cloud simulation methods mainly rely on a single data source, such as satellite images or ground observation data. Although these methods can simulate the morphology and dynamic changes of clouds to a certain extent, they have some obvious limitations:

[0003] 1. Limited data coverage: A single data source can often only provide information at a certain point in time or at a certain location, which makes it difficult to fully reflect the overall picture and dynamic changes of the cloud layer.

[0004] 2. Data accuracy and reliability issues: The accuracy and reliability of different data sources may vary. The error of a single data source may be amplified, affecting the accuracy of the simulation results.

[0005] 3. Impact of environmental factors: The formation and changes of clouds are affected by a variety of environmental factors, such as temperature, humidity, air pressure, and wind speed. It is difficult for a single data source to fully capture the impact of these complex factors on clouds.

[0006] In order to overcome the above problems, the present invention proposes a cloud simulation method of multi-source data fusion. The method realizes all-round monitoring by integrating satellite images, radar data and ground observation data. Through advanced data processing and layered fusion technology, an accurate cloud model with three-dimensional structural information is formed, and weights can be reasonably allocated according to the characteristics of different data sources. A real-time update mechanism is also established to ensure that the model always reflects the latest cloud status, which has wide application value and strong adaptability in many fields such as weather forecasting, environmental monitoring, and aviation navigation. Summary of the invention

[0007] The present invention aims to overcome the above-mentioned shortcomings of the prior art and provide a cloud simulation method and device based on multi-source data fusion. The present invention can generate an accurate and dynamic cloud model, which is suitable for application fields such as weather forecasting, environmental monitoring, aviation navigation, urban planning, virtual reality and augmented reality.

[0008] A first aspect of the present invention relates to a cloud simulation method based on multi-source data fusion, comprising the following steps:

[0009] 1. In the data collection phase, multi-source data are obtained from satellites, radars and ground meteorological observation stations, and then preliminarily processed:

[0010] The satellite data collection system is used to regularly collect satellite image data of different bands, focusing on the band information that can reflect the characteristics of cloud height, thickness, texture, etc. The collection frequency is once an hour to ensure the effective tracking of large-scale changes in the cloud layer. At the same time, the collected satellite image data is preliminarily geolocated and time-stamped for subsequent matching and fusion with data from other data sources.

[0011] With the help of radar monitoring equipment, vertical structure information of clouds is continuously obtained, including detailed data such as the distribution of precipitation particles inside the clouds and echo intensity. The collection frequency is once per minute, so as to accurately grasp the microscopic structural changes of the clouds.

[0012] Various meteorological data are collected from ground meteorological observation stations distributed in different geographical locations, such as basic meteorological parameters such as temperature, humidity, air pressure, wind speed, and visual observation descriptive information on clouds, such as the color and shape characteristics of clouds. The collection interval is once every 10 minutes, providing data support for a comprehensive understanding of the ground environment conditions where the clouds are located.

[0013] 2. Data preprocessing and optimization: Perform corresponding processing on data from different data sources to improve data quality:

[0014] For satellite image data, image enhancement algorithms are used to highlight the features of clouds in the image and enhance their contrast and clarity. At the same time, filtering technology is used to remove noise interference in the image, making the edges and details of the clouds more clearly discernible. For example, the median filtering algorithm is used to remove salt and pepper noise to make the outline of the cloud smoother.

[0015] Calibrate the radar data and perform error correction on the radar measurement data based on the known radar echo characteristics of standard targets to eliminate measurement deviations caused by equipment errors and environmental factors. For example, by comparing the difference between the radar echo of a building of known height and the actual height, the vertical measurement accuracy of the radar can be calibrated.

[0016] The ground observation data is cleaned and standardized to remove outliers and erroneous data, and the data from different observation stations are converted to the same unit and format for subsequent data fusion calculations. For example, temperature data is converted to degrees Celsius and wind speed data is converted to meters per second.

[0017] 3. Multi-source data fusion construction, establish a layered fusion model to fuse data from different data sources:

[0018] A multi-source data fusion model is established, which adopts a layered fusion strategy. First, the satellite image data and radar data are fused in the spatial dimension, and the vertical structure information of the radar data is matched with the plane position information of the satellite image using image registration technology to form a preliminary cloud model with three-dimensional structure information.

[0019] Then, the ground observation data is fused with the above preliminary model in the feature dimension. According to the ground meteorological parameters and visual observation information, the physical characteristics and appearance characteristics of the cloud model are corrected and improved. For example, the water vapor content and condensation state of the cloud layer are adjusted according to the ground temperature and humidity data to make the cloud model more consistent with the actual meteorological conditions.

[0020] The weighted fusion algorithm is used to assign different weight coefficients to different data sources according to their reliability and importance. For example, satellite image data has a high reliability in reflecting the macroscopic morphology of clouds, and is assigned a weight of 0.4; radar data has an advantage in reflecting the internal structure of clouds, and is assigned a weight of 0.4; ground observation data has important value in reflecting the impact of the ground environment on clouds, and is assigned a weight of 0.2.

[0021] 4. Cloud model construction and presentation: construct and visualize cloud model based on fused data:

[0022] Based on the fused data, a physical model of clouds is constructed. This model considers various physical factors involved in the formation, development and dissipation of clouds, such as thermal effects, dynamic effects and water vapor transport. For example, according to the principles of thermodynamics, the temperature and humidity changes inside the cloud, as well as the condensation and evaporation rates of water vapor, are calculated to simulate the growth and dissipation process of clouds.

[0023] The physical model is converted into a visualized cloud model using the ray marching algorithm in the volume rendering method. Light is emitted from the camera position to each pixel, and the light gradually passes through the cloud volume data structure. During the ray marching process, the interaction between light and cloud material is accurately calculated based on the physical characteristics of the cloud layer reflected by the fused data, and the propagation direction, intensity, color, and transparency of the light are dynamically changed according to the internal state of the cloud layer. As the light continues to march and accumulates its influence, the color and transparency values ​​corresponding to each pixel are finally determined, presenting a realistic cloud visualization effect.

[0024] 5. Model update and application expansion, to achieve real-time optimization of the model and apply it to multiple fields:

[0025] Establish a real-time update mechanism for the model and continuously receive new multi-source data. When new data is passed in, start the data processing process, and use the data fusion algorithm to fuse the new data with the existing cloud model data. In this process, adjust and optimize the various parameters of the model according to the physical characteristics and change laws of the cloud layer. Whether it is the internal characteristics of the cloud layer such as shape, density, temperature, humidity, or its external manifestations such as position and movement trend in space, they can be updated in time with the addition of new data to ensure that the model can always accurately reflect the latest actual status of the cloud layer. The optimized cloud model is applied to fields such as weather forecasting, environmental monitoring, aviation navigation, urban planning, virtual reality and augmented reality.

[0026] The second aspect of the present invention relates to a cloud simulation device based on multi-source data fusion, comprising a memory and one or more processors, wherein the memory stores executable code, and when the one or more processors execute the executable code, it is used for a cloud simulation method based on multi-source data fusion of the present invention.

[0027] A third aspect of the present invention relates to a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements a cloud simulation method based on multi-source data fusion of the present invention.

[0028] The beneficial effects of the present invention are mainly manifested in:

[0029] (1) Improved data comprehensiveness. By integrating satellite images, radar data, and ground observation data, the problem of limited coverage of a single data source has been overcome, and the cloud layer can be monitored in all directions, fully reflecting the overall appearance and dynamic changes of the cloud layer.

[0030] (2) Enhanced data accuracy and reliability. Preprocessing of different data sources, such as calibrating radar data and cleaning and standardizing ground observation data, reduces errors, improves data accuracy and reliability, and avoids the situation where errors from a single data source are amplified. A weighted fusion algorithm is used to reasonably allocate weights according to the characteristics of different data sources, further improving the accuracy of the fused data.

[0031] (3) Comprehensive consideration of environmental factors. Various meteorological data from ground-based meteorological observation stations are collected to fully capture the impact of environmental factors such as temperature, humidity, air pressure, and wind speed on clouds, making cloud simulation more consistent with actual conditions.

[0032] (4) Real-time update capability. A real-time update mechanism has been established to continuously receive new data and perform integration and model optimization to ensure that the model always reflects the latest cloud status and provides timely and accurate cloud information for various application fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1is a flow chart of the method of the present invention.

[0034] Figure 2 It is a data acquisition system architecture diagram of the present invention.

[0035] Figure 3 This is a schematic diagram of the ray marching algorithm. DETAILED DESCRIPTION

[0036] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution 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.

[0037] The present invention will be further described below in conjunction with the accompanying drawings.

[0038] Example 1

[0039] like Figure 1 , this embodiment relates to a cloud simulation method based on multi-source data fusion, comprising the following steps:

[0040] Step 1: If Figure 2 Collect data from different data sources. Using the satellite data acquisition system, regularly collect satellite image data of different bands, focusing on the band information that can reflect the characteristics of cloud height, thickness, texture, etc., with a collection frequency of once an hour to ensure effective tracking of large-scale changes in the cloud layer. At the same time, the collected satellite image data is preliminarily geolocated and time-stamped for subsequent matching and fusion with data from other data sources. With the help of radar monitoring equipment, the vertical structure information of the cloud layer is continuously obtained, including detailed data such as the distribution of precipitation particles and echo intensity inside the cloud layer. The collection frequency is once a minute, so as to accurately grasp the microscopic structural changes of the cloud layer. Various meteorological data are collected from ground meteorological observation stations distributed in different geographical locations, such as basic meteorological parameters such as temperature, humidity, air pressure, wind speed, and visual observation description information of the cloud layer, such as the color and shape characteristics of the cloud layer. The collection interval is once every 10 minutes, providing data support for a comprehensive understanding of the ground environment conditions where the cloud layer is located.

[0041] Step 2: Perform corresponding preprocessing and optimization for different data sources to improve data quality. For satellite image data, use image enhancement techniques such as histogram equalization algorithm to enhance the contrast and clarity of the image. And use filtering technology to remove noise interference in the image, such as using the median filtering algorithm. For the grayscale value of a certain pixel point (x, y) in the image, take the median of the grayscale values ​​of the pixels in its neighborhood as the new grayscale value of the point to eliminate salt and pepper noise, etc., so that the edges and details of the cloud layer are clearer. For radar data, error correction is performed on radar measurement data based on the radar echo characteristics of known standard targets. For ground observation data, data cleaning and standardization are performed to remove outliers and erroneous data, and data from different observation stations are uniformly converted to the same unit and format, such as converting wind speed from kilometers per hour to meters per second.

[0042] Step 3: Establish a hierarchical fusion model to fuse data from different data sources. The specific steps are as follows:

[0043] (3.1) Satellite image data and radar data are fused in the spatial dimension, and the vertical structure information of radar data is matched with the plane position information of satellite image using image registration technology.

[0044] Suppose the coordinates of a point in the satellite image are (x s ,y s ), the coordinates of the matching point in the corresponding radar data are (x r ,y r ,z r )(z r is the vertical height information), through the coordinate conversion function F(x s ,y s )=(x r ,y r ,z r ) to achieve matching and form a preliminary cloud model with three-dimensional structural information. In the fusion process, the weighted fusion method can be used, assuming that the weight of satellite image data in the fusion is ω s , the radar data weight is ω r , the fusion result D fusion It can be expressed as D fusion =ω s D s +ω r D r (Where D s is the satellite image data, D r for radar data).

[0045] (3.2) The ground observation data is integrated with the above preliminary model in the feature dimension. According to the ground meteorological parameters and visual observation information, the physical characteristics and appearance characteristics of the cloud model are corrected and improved.

[0046] The water vapor content and condensation state of the cloud layer are adjusted according to the ground temperature and humidity data. The relationship between the water vapor content C and the temperature T and humidity W can be approximately expressed as C = f(T, W) (the specific functional relationship can be obtained by fitting the experimental data), making the cloud model more in line with the actual meteorological conditions.

[0047] Step 4: Build a cloud model based on the fused data and visualize it. The specific steps are as follows:

[0048] (4.1) Based on the fused data, a physical model of the cloud layer is constructed.

[0049] The model considers various physical factors involved in the formation, development and dissipation of clouds, such as thermal effects, dynamic effects and water vapor transport. According to the principles of thermodynamics, the temperature and humidity changes inside the cloud, as well as the condensation and evaporation rates of water vapor are calculated. For example, the water vapor evaporation rate E is related to temperature T, relative humidity RH and air pressure P. a The relationship can be expressed as E = g(T, RH, P a ) (function g is derived through theoretical derivation and experimental verification), thereby simulating the growth and dissipation process of clouds. In terms of dynamic action, the movement of clouds is affected by the atmospheric flow field. The drift and deformation of clouds can be simulated based on data such as wind speed V and wind direction θ. Suppose the displacement speed of clouds in the x and y directions is V respectively. x and V y , then V x =V×cos(θ),V y =V×sin(θ).

[0050] (4.2) Use volume rendering methods to transform the physical model into a visual cloud model.

[0051] Compared with the traditional volume rendering implementation method, the ray marching algorithm is more flexible and can effectively handle various optical phenomena and materials, such as volumetric clouds, fluids, refraction, etc. Therefore, the present invention uses the ray marching algorithm to visualize the constructed physical model to achieve a natural and realistic volumetric cloud visual effect. Its working principle is as follows Figure 3As shown: The ray marching algorithm starts from the camera position, emits rays from the camera near plane to each pixel through the cloud layer, and accumulates the medium density at each step along the ray direction. For sampling points with non-zero density, the density between the sampling point and the sun is accumulated while stepping in the direction of the sun. The opacity of the current rendered pixel is determined by the accumulated density on the ray, while the color value is affected by the color of the sunlight and the density between the sunlight and the sampling point, thus achieving a natural and realistic volumetric cloud visual effect.

[0052] Step 5: Establish real-time update and optimization of the model and apply it to multiple fields. Establish a real-time update mechanism for the model to continuously receive new multi-source data. When new data is passed in, start the data processing process and use the data fusion algorithm to fuse the new data with the existing cloud model data. In this process, adjust and optimize the various parameters of the model according to the physical characteristics and change laws of the cloud layer. Whether it is the internal characteristics of the cloud layer such as shape, density, temperature, humidity, or its external manifestations such as position and movement trend in space, they can be updated in time with the addition of new data to ensure that the model can always accurately reflect the latest actual status of the cloud layer. Apply the optimized cloud model to fields such as weather forecasting, environmental monitoring, aviation navigation, urban planning, virtual reality and augmented reality.

[0053] Example 2

[0054] The present embodiment relates to a cloud layer simulation device based on multi-source data fusion, comprising a memory and one or more processors, wherein the memory stores executable code, and when the one or more processors execute the executable code, they are used to implement a cloud layer simulation method based on multi-source data fusion of Example 1.

[0055] Example 3

[0056] The present embodiment relates to a computer-readable storage medium, characterized in that a program is stored thereon, and when the program is executed by a processor, a cloud layer simulation method based on multi-source data fusion of embodiment 1 is implemented.

[0057] The above description is only an embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.

Claims

1. A cloud simulation method based on multi-source data fusion, characterized by: The following steps are involved: Step 1: Data collection and collation: obtaining multi-source data from satellites, radars and ground meteorological observation stations, and performing preliminary processing on them; Step 2: Data preprocessing and optimization: perform corresponding processing on data from different data sources to improve data quality; Step 3: Multi-source data fusion construction, establish a hierarchical fusion model to fuse data from different data sources; Step 4: Cloud model construction and presentation: construct and visualize the cloud model based on the fused data; Step 5: Model update and application expansion, to achieve real-time optimization of the model and apply it to multiple fields.

2. A cloud layer simulation method based on multi-source data fusion as claimed in claim 1, characterized in that: step; Specifically, it includes: using the satellite data collection system to regularly collect satellite image data of different bands, focusing on the band information that can reflect the characteristics of cloud height, thickness, texture, etc., with a collection frequency of once an hour to ensure effective tracking of large-scale changes in the cloud layer; at the same time, the collected satellite image data is preliminarily geolocated and time-stamped for subsequent matching and fusion with data from other data sources; With the help of radar monitoring equipment, vertical structure information of clouds is continuously obtained, including detailed data on the distribution of precipitation particles inside the clouds and echo intensity. The collection frequency is once a minute, so as to accurately grasp the microstructural changes of the clouds. Various meteorological data are collected from ground meteorological observation stations distributed in different geographical locations, as well as visual observation description information of the clouds. The collection interval is once every 10 minutes.

3. The cloud layer simulation method based on multi-source data fusion according to claim 1, characterized in that: In step 2, corresponding preprocessing and optimization are performed on data from different data sources to improve data quality. For satellite image data, image enhancement technology is used to enhance the contrast and clarity of the image. Filtering technology is used to remove noise interference in the image. For the grayscale value of a pixel point (x, y) in the image, the median of the grayscale values ​​of the pixels in its neighborhood is taken as the new grayscale value of the point to eliminate salt and pepper noise, etc., making the edges and details of the clouds clearer. For radar data, error correction is performed on radar measurement data based on the radar echo characteristics of known standard targets. For ground observation data, data cleaning and standardization are performed to remove outliers and erroneous data, and data from different observation stations are uniformly converted into the same unit and format.

4. The cloud layer simulation method based on multi-source data fusion according to claim 1, characterized in that: In step 3, a hierarchical fusion model is established to fuse data from different data sources. The specific steps are as follows: (3.1) Fuse satellite image data and radar data in the spatial dimension, and use image registration technology to match the vertical structure information of radar data with the plane position information of satellite images; Suppose the coordinates of a point in the satellite image are (x s ,y s ), the corresponding matching point coordinates in the radar data are (x r ,y r ,z r )(z r is the vertical height information), through the coordinate conversion function F(x s ,y s )=(x r ,y r ,z r ) to achieve matching and form a preliminary cloud model with three-dimensional structural information; in the fusion process, the weighted fusion method is adopted, and the weight of satellite image data in the fusion is ω s , the radar data weight is ω r , the fusion result D fusion Indicated as D fusion =ω s D s +ω r D r (Where D s is the satellite image data, D r for radar data) (3.2) Fuse the ground observation data with the above preliminary model in the feature dimension; modify and improve the physical characteristics and appearance features of the cloud model based on ground meteorological parameters and visual observation information; The water vapor content and condensation state of the cloud layer are adjusted according to the ground temperature and humidity data. The relationship between the water vapor content C and the temperature T and humidity W can be approximately expressed as C = f(T, W) (the specific functional relationship can be obtained by fitting the experimental data), making the cloud model more in line with the actual meteorological conditions.

5. The cloud layer simulation method based on multi-source data fusion according to claim 1, characterized in that: In step 4, a cloud layer model is constructed and visualized based on the fused data. The specific steps are as follows: (4.1) Based on the fused data, a physical model of the cloud layer is constructed; The model considers various physical factors involved in the formation, development and dissipation of clouds, such as thermal effects, dynamic effects and water vapor transport. Based on the principles of thermodynamics, it calculates the temperature and humidity changes inside the cloud, as well as the condensation and evaporation rates of water vapor. The water vapor evaporation rate E is related to temperature T, relative humidity RH and air pressure P. a The relationship can be expressed as E = g(T, RH, P a ), thereby simulating the growth and dissipation process of clouds; in terms of dynamic action, the movement of clouds is affected by the atmospheric flow field. The drift and deformation of clouds are simulated according to data such as wind speed V and wind direction θ. Suppose the displacement speed of clouds in the x and y directions is V respectively. x and V y , then V x =V×cos(θ),V y =V×sinθ; (4.2) Use volume rendering methods to transform the physical model into a visual cloud model; The ray-stepping algorithm in the volume rendering method is used to transform the physical model into a visual cloud model. Light is emitted from the camera position to each pixel, and the light gradually passes through the cloud volume data structure. During the ray-stepping process, the interaction between light and cloud material is accurately calculated based on the physical characteristics of the cloud reflected by the fused data, and the propagation direction, intensity, color and transparency of the light are dynamically changed according to the internal state of the cloud. As the light continues to step and accumulates its impact, the color and transparency value corresponding to each pixel is finally determined, presenting a realistic cloud visualization effect.

6. The cloud layer simulation method based on multi-source data fusion according to claim 1, characterized in that: In step 5, real-time updating and real-time optimization of the model are established and applied to multiple fields. When new data is input, the data processing process is started, and the data fusion algorithm is used to fuse the new data with the existing cloud model data. In this process, various parameters of the model are adjusted and optimized according to the physical characteristics and change laws of the cloud layer. Whether it is the internal characteristics of the cloud layer such as shape, density, temperature, humidity, or its external manifestations such as position and movement trend in space, they can be updated in time with the addition of new data to ensure that the model can always accurately reflect the latest actual status of the cloud layer. The optimized cloud model is applied to fields such as weather forecasting, environmental monitoring, aviation navigation, urban planning, virtual reality, and augmented reality.

7. A cloud simulation device based on multi-source data fusion, characterized in that: It comprises a memory and one or more processors, wherein the memory stores executable codes, and when the one or more processors execute the executable codes, they are used to implement a cloud simulation method based on multi-source data fusion as described in any one of claims 1-6.

8. A computer-readable storage medium, characterized in that: A program is stored thereon, and when the program is executed by a processor, a cloud layer simulation method based on multi-source data fusion as described in any one of claims 1-6 is implemented.

Images