Vegetation attenuation prediction method based on machine learning, medium and electronic equipment
Through the vegetation attenuation prediction method based on machine learning, using ray tracing algorithm and GPU parallel computing, the dependence on electrical parameters and spatial distribution information in the existing technology is solved, and accurate vegetation attenuation prediction without this information is achieved, improving the prediction accuracy degree and efficiency.
Patent Information
- Application Number
- CN202410959914.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-17
- Publication Date
- 2024-11-08
AI Technical Summary
Existing vegetation attenuation prediction methods require accurate electrical parameters, three-dimensional shapes and spatial distribution information, or rely on insufficient measured data, resulting in inaccurate or unfeasible prediction results.
Using a machine learning-based method, by obtaining the attenuation mode parameter values, radiation source coordinates and electromagnetic wave frequency, using ray tracing algorithms and GPU parallel calculations, the vegetation attenuation prediction model is trained to achieve accurate prediction without the need for electrical parameters, three-dimensional shapes and spatial distribution information. .
The vegetation attenuation simulation calculation is achieved within a few seconds, and a large amount of reliable data is generated, which improves the accuracy and feasibility of vegetation attenuation prediction.
Smart Images

Figure CN118917532A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of vegetation attenuation prediction, and in particular to a vegetation attenuation prediction method, medium and electronic equipment based on machine learning. Background Art
[0002] Vegetation modeling is a common problem in the wireless communication industry. Since vegetation can attenuate the power of radio wave propagation, the presence of vegetation in urban areas has a significant impact on radio planning. In recent years, various methods for studying the attenuation of radio wave propagation in vegetation have emerged. According to the type of calculation method, they are mainly divided into theoretical model, empirical model and semi-empirical model calculation methods.
[0003] However, various model calculation methods have the following problems: for theoretical models, although relatively accurate prediction results can be obtained, the theoretical model calculation method requires precise knowledge of information such as the electrical parameters, three-dimensional shape, and spatial distribution of vegetation. In actual use, the above information is difficult to obtain; the empirical model calculation method has the characteristics of simple calculation and wide versatility. However, the measured data for fitting empirical models are often insufficient and have great limitations, resulting in poor accuracy of empirical formulas, which can often only provide approximate prediction results; the semi-empirical model calculation method can be regarded as a combination of theoretical model and empirical model calculation methods, and analysis and modeling are performed based on experimental results. Although the semi-empirical model calculation method does not require knowledge of information such as the electrical parameters, three-dimensional shape, and spatial distribution of vegetation, it requires certain measured data before it can be used, which also limits its feasibility in actual use. Summary of the invention
[0004] The present invention aims to solve one of the technical problems in the related art to at least a certain extent. To this end, the first purpose of the present invention is to provide a vegetation attenuation prediction method based on machine learning, which can realize vegetation attenuation prediction without information such as electrical parameters, three-dimensional shape and spatial distribution of vegetation, and the method can obtain a large amount of vegetation attenuation data, and can realize accurate vegetation attenuation prediction through a ray tracing algorithm based on the massive vegetation attenuation data.
[0005] A second object of the present invention is to provide a computer-readable storage medium.
[0006] A third object of the present invention is to provide an electronic device.
[0007] To achieve the above object, the present invention is implemented through the following technical solutions:
[0008] A vegetation attenuation prediction method based on machine learning, comprising:
[0009] Acquire vegetation attenuation data, wherein the vegetation attenuation data includes attenuation mode parameter values, radiation source coordinates, receiver coordinates, and electromagnetic wave frequencies emitted by the radiation source;
[0010] The vegetation attenuation prediction model is trained, and the vegetation attenuation data is input into the vegetation attenuation prediction model to predict the corresponding electromagnetic wave power loss value, so as to realize vegetation attenuation prediction.
[0011] Preferably, the attenuation mode parameter value is comprehensively determined based on the vegetation type and the relative position of the receiver and the vegetation.
[0012] Preferably, the vegetation types include single trees, green belts and woods.
[0013] Preferably, the relative position of the receiver and the vegetation includes that the receiver is inside or outside the vegetation.
[0014] Preferably, when the vegetation type is a single tree, the attenuation mode parameter value is 0;
[0015] When the vegetation type is a green belt and the receiver is inside the vegetation, the attenuation mode parameter value is 1;
[0016] When the vegetation type is a green belt and the receiver is outside the vegetation, the attenuation mode parameter value is 2;
[0017] When the vegetation type is a forest and the receiver is inside the vegetation, the attenuation mode parameter value is 3;
[0018] When the vegetation type is forest and the receiver is outside the vegetation, the attenuation mode parameter value is 4.
[0019] Preferably, the vegetation attenuation prediction model performs vegetation attenuation prediction by a ray tracing algorithm.
[0020] Preferably, the prediction step of the ray tracing algorithm includes:
[0021] Determining the radiation source and receiver positions according to the radiation source coordinates and the receiver coordinates, and determining the vegetation type and the relative position of the receiver and the vegetation according to the attenuation mode parameter value;
[0022] Tracking the propagation path of each electromagnetic wave ray emitted by the radiation source, and determining the transmission distance, receiving angle and time delay parameters of each propagation path;
[0023] The electromagnetic wave power loss value of the corresponding propagation path is determined according to the electromagnetic wave frequency, propagation path, transmission distance, receiving angle and time delay parameters.
[0024] Preferably, the ray tracing algorithm is accelerated by a GPU parallel computing method to obtain a vegetation attenuation data set for implementing vegetation attenuation prediction model training.
[0025] To achieve the above-mentioned purpose, the second aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the above-mentioned vegetation attenuation prediction method based on machine learning is implemented.
[0026] To achieve the above-mentioned purpose, the third aspect of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the above-mentioned vegetation attenuation prediction method based on machine learning is implemented.
[0027] The present invention has at least the following technical effects:
[0028] The present invention provides a vegetation attenuation prediction method based on machine learning, which first obtains vegetation attenuation data, wherein the vegetation attenuation data includes attenuation mode parameter values, radiation source coordinates, receiver coordinates, and the frequency of electromagnetic waves emitted by the radiation source, and then inputs the vegetation attenuation data into a trained vegetation attenuation prediction model, and the vegetation attenuation prediction model predicts the input data through a ray tracing algorithm to obtain a corresponding electromagnetic wave power loss value, thereby realizing vegetation attenuation prediction, and thus the method can realize vegetation attenuation prediction without information such as electrical parameters, three-dimensional shapes, and spatial distribution of vegetation. In addition, when the present invention uses a ray tracing algorithm to perform vegetation attenuation calculations, a GPU parallel computing method can be used to accelerate the ray tracing algorithm, so that a simulation calculation can be completed within a few seconds, thereby facilitating the generation of massive and reliable vegetation attenuation data to form a vegetation attenuation data set, thereby facilitating the training of a vegetation attenuation prediction model and improving the accuracy of vegetation attenuation prediction.
[0029] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] Figure 1 The present invention is a flowchart of a vegetation attenuation prediction method based on machine learning.
[0031] Figure 2 Schematic diagram of the vegetation attenuation prediction model structure according to an embodiment of the present invention. DETAILED DESCRIPTION
[0032] The present embodiment is described in detail below, and examples of the embodiment are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.
[0033] A vegetation attenuation prediction method, medium, and electronic device based on machine learning according to this embodiment are described below with reference to the accompanying drawings.
[0034] Figure 1 FIG. 4 is a flow chart of a vegetation attenuation prediction method based on machine learning according to an embodiment of the present invention. Figure 1 As shown, the method includes:
[0035] Step S1: Obtain vegetation attenuation data, where the vegetation attenuation data includes attenuation mode parameter values, radiation source coordinates, receiver coordinates, and the frequency of electromagnetic waves emitted by the radiation source.
[0036] The attenuation mode parameter value is determined based on the vegetation type and the relative position of the receiver to the vegetation. The vegetation type includes a single tree, a green belt, and a forest. The relative position of the receiver to the vegetation includes whether the receiver is inside or outside the vegetation.
[0037] In this embodiment, the attenuation mode parameter values are set as follows. First, the vegetation is divided into three types: single tree, green belt and forest. Secondly, the relative position between the receiver and the vegetation will also seriously affect the calculation of the vegetation attenuation. Therefore, for vegetation of green belt and forest type, considering the two situations that the receiver is located inside and outside the green belt and forest, the attenuation mode classification is shown in Table 1.
[0038] Table 1 Attenuation mode classification table
[0039] Vegetation Type Relative Position Decay Mode Single tree \ 0 Green belt internal 1 Green belt external 2 forest internal 3 forest external 4
[0040] Specifically, when the vegetation type is a single tree, the attenuation mode parameter value is 0; when the vegetation type is a green belt and the receiver is inside the vegetation, the attenuation mode parameter value is 1; when the vegetation type is a green belt and the receiver is outside the vegetation, the attenuation mode parameter value is 2; when the vegetation type is a forest and the receiver is inside the vegetation, the attenuation mode parameter value is 3; when the vegetation type is a forest and the receiver is outside the vegetation, the attenuation mode parameter value is 4.
[0041] Step S2: training a vegetation attenuation prediction model, inputting vegetation attenuation data into the vegetation attenuation prediction model to predict the corresponding electromagnetic wave power loss value, so as to realize vegetation attenuation prediction.
[0042] After obtaining the vegetation attenuation data, this method designs the following neural network model, namely the vegetation attenuation prediction model, to predict the power attenuation of electromagnetic waves by vegetation. Figure 2 As shown in the figure, the vegetation attenuation prediction model uses the attenuation mode parameter value mod, the three-dimensional coordinates of the radiation source (x t ,y y ,z t ), the three-dimensional coordinates of the receiver (x r ,y r ,z r ) and the electromagnetic wave frequency f are taken as input, and then the power attenuation value of the electromagnetic wave can be output through n hidden layer operations, that is, the power loss value L caused by the attenuation of the electromagnetic wave by vegetation pre .
[0043] In one embodiment of the present invention, the vegetation attenuation prediction model can predict vegetation attenuation through a ray tracing algorithm. The prediction steps of the ray tracing algorithm include: determining the radiation source and receiver positions according to the radiation source coordinates and the receiver coordinates, and determining the vegetation type and the relative position of the receiver and the vegetation according to the attenuation mode parameter value; tracing the propagation path of each electromagnetic wave ray emitted by the radiation source, and determining the transmission distance, receiving angle and time delay parameters of each propagation path; determining the electromagnetic wave power loss value of the corresponding propagation path according to the electromagnetic wave frequency, propagation path, transmission distance, receiving angle and time delay parameters.
[0044] Specifically, the radiation source and receiver positions can be determined according to the radiation source coordinates and the receiver coordinates, and the vegetation type and the relative position of the receiver and the vegetation can be determined according to the attenuation mode parameter value. Then, the electromagnetic waves radiated from the source point of the radiation source are regarded as independent rays. These rays propagate in space and may reflect, refract or diffract when encountering obstacles, that is, there are propagation paths such as reflection, refraction or diffraction. Further, the propagation path of the ray is tracked, and its transmission distance, receiving angle, time delay and other parameters are recorded. Then, based on these parameters and the propagation path of the ray and the transmission characteristics of the electromagnetic wave (such as attenuation, phase change, etc.), the electromagnetic wave power loss value corresponding to the ray propagation path is calculated in combination with the Fresnel reflection law.
[0045] Furthermore, due to the difficulty in obtaining measured data in the background technology, the amount of measured data is often insufficient. For the neural network model, a sufficient and high-quality training data set is crucial, and the importance of the data set is even greater than the structure of the neural network model. In order to solve the problem of insufficient measured data, this method also uses high-performance GPU (Graphics Processing Unit) parallel technology to accelerate the ray tracing algorithm so that a simulation calculation can be completed within a few seconds, thereby facilitating the generation of massive and reliable vegetation attenuation data to form a vegetation attenuation data set, thereby facilitating the training of the vegetation attenuation prediction model and improving the accuracy of vegetation attenuation prediction.
[0046] In summary, the present invention provides a vegetation attenuation prediction method based on machine learning, which first obtains vegetation attenuation data, wherein the vegetation attenuation data includes attenuation mode parameter values, radiation source coordinates, receiver coordinates, and the frequency of electromagnetic waves emitted by the radiation source, and then inputs the vegetation attenuation data into a trained vegetation attenuation prediction model, and the vegetation attenuation prediction model predicts the input data through a ray tracing algorithm to obtain a corresponding electromagnetic wave power loss value, thereby realizing vegetation attenuation prediction. Therefore, this method can realize vegetation attenuation prediction without information such as electrical parameters, three-dimensional shapes, and spatial distribution of vegetation. In addition, when the present invention uses a ray tracing algorithm to perform vegetation attenuation calculations, the ray tracing algorithm can be accelerated using a GPU parallel computing method, so that a simulation calculation can be completed within a few seconds, thereby facilitating the generation of massive and reliable vegetation attenuation data to form a vegetation attenuation data set, thereby facilitating the training of a vegetation attenuation prediction model and improving the accuracy of vegetation attenuation prediction.
[0047] Furthermore, the present invention also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the above-mentioned vegetation attenuation prediction method based on machine learning is implemented.
[0048] Furthermore, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the above-mentioned vegetation attenuation prediction method based on machine learning is implemented.
[0049] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.
[0050] Although the content of the present invention has been described in detail through the above preferred embodiments, it should be appreciated that the above description should not be considered as a limitation of the present invention. After reading the above content, it will be apparent to those skilled in the art that various modifications and substitutions of the present invention will occur. Therefore, the protection scope of the present invention should be limited by the appended claims.
Claims
1. A vegetation attenuation prediction method based on machine learning, characterized in that: Acquire vegetation attenuation data, wherein the vegetation attenuation data includes attenuation mode parameter values, radiation source coordinates, receiver coordinates, and electromagnetic wave frequencies emitted by the radiation source; The vegetation attenuation prediction model is trained, and the vegetation attenuation data is input into the vegetation attenuation prediction model to predict the corresponding electromagnetic wave power loss value, so as to realize vegetation attenuation prediction.
2. The vegetation attenuation prediction method based on machine learning according to claim 1, characterized in that: The attenuation mode parameter value is comprehensively determined according to the vegetation type and the relative position of the receiver and the vegetation.
3. The vegetation attenuation prediction method based on machine learning according to claim 2, characterized in that: The vegetation types include single trees, green belts and woods.
4. The vegetation attenuation prediction method based on machine learning according to claim 3, characterized in that: The relative position of the receiver and the vegetation includes whether the receiver is inside or outside the vegetation.
5. The vegetation attenuation prediction method based on machine learning according to claim 4, characterized in that: When the vegetation type is a single tree, the attenuation mode parameter value is 0; When the vegetation type is a green belt and the receiver is inside the vegetation, the attenuation mode parameter value is 1; When the vegetation type is a green belt and the receiver is outside the vegetation, the attenuation mode parameter value is 2; When the vegetation type is a forest and the receiver is inside the vegetation, the attenuation mode parameter value is 3; When the vegetation type is forest and the receiver is outside the vegetation, the attenuation mode parameter value is 4.
6. The vegetation attenuation prediction method based on machine learning according to claim 1, characterized in that: The vegetation attenuation prediction model performs vegetation attenuation prediction through a ray tracing algorithm.
7. The vegetation attenuation prediction method based on machine learning according to claim 6, characterized in that: The prediction step of the ray tracing algorithm includes: Determining the radiation source and receiver positions according to the radiation source coordinates and the receiver coordinates, and determining the vegetation type and the relative position of the receiver and the vegetation according to the attenuation mode parameter value; Tracking the propagation path of each electromagnetic wave ray emitted by the radiation source, and determining the transmission distance, receiving angle and time delay parameters of each propagation path; The electromagnetic wave power loss value of the corresponding propagation path is determined according to the electromagnetic wave frequency, propagation path, transmission distance, receiving angle and time delay parameters.
8. The vegetation attenuation prediction method based on machine learning according to claim 6, characterized in that: The ray tracing algorithm is accelerated by GPU parallel computing method to obtain a vegetation attenuation data set for implementing vegetation attenuation prediction model training.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the vegetation attenuation prediction method based on machine learning as described in any one of claims 1 to 8 is implemented.
10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the vegetation attenuation prediction method based on machine learning as described in any one of claims 1 to 8 is implemented.
Citation Information
Cited By
Vegetation scattering model construction method and system based on point cloud data and application
CN120635324A
