WO2023020057A1

WO2023020057A1 - Fire spreading simulation acceleration method and system based on model simplification

Info

Publication number: WO2023020057A1
Application number: PCT/CN2022/094898
Authority: WO
Inventors: 汪涛; 陈小莹; 杨逍; 黄俊康
Original assignee: 中山大学
Priority date: 2021-08-16
Filing date: 2022-05-25
Publication date: 2023-02-23
Also published as: CN113722894B; CN113722894A

Abstract

Disclosed in the present invention are a fire spreading simulation acceleration method and system based on model simplification. The method comprises: acquiring forest fire point monitoring data and extracting an affected area boundary of a forest fire, so as to obtain affected area boundary information; determining speed sets of key parameters of a spreading model according to the affected area boundary information, and constructing an initial spreading model; simplifying the initial spreading model to obtain a simplified discrete spreading model; and dynamically simulating an affected area on the basis of the simplified discrete spreading model. The system comprises: an information extraction model, an initial model construction model, a simplification module and a dynamic simulation model. By means of the present invention, the time required for completing dynamic simulation of an affected area can be reduced while the prediction precision is ensured. The fire spreading simulation acceleration method and system based on model simplification in the present invention can be widely applied in the field of geographic information processing.

Description

A fire spread simulation acceleration method and system based on model simplification

technical field

The invention relates to the field of geographic information processing, in particular to a method and system for accelerating fire spread simulation based on model simplification.

Background technique

Worldwide, sprawling disasters occur frequently, threatening the ecological environment, life and property safety. The following takes forest fire as an example to introduce the spread control. Large-scale forest fires are difficult to extinguish manually and often require effective containment to contain the spread or protect human and animal habitats. The premise of constructing the isolation zone is to predict the time-varying process of the disaster-affected area as quickly and accurately as possible, and at the same time design the control strategy and quantify its impact on the disaster-affected area.

technical problem

In the existing research on spreading disasters, the spreading control model based on discrete differential inclusion is usually used to simulate the dynamic spreading process of forest fires. If there are too many discrete points, it will lead to an excessive amount of calculation when using this discrete model to perform numerical simulation of the disaster-affected area.

technical solution

In order to solve the above-mentioned technical problems, the object of the present invention is to provide a fire spread simulation acceleration method and system based on model simplification, which can greatly increase the speed of dynamic spread simulation on the premise of ensuring the accuracy of the model as much as possible.

The first technical solution adopted in the present invention is: a method for accelerating the simulation of fire spread based on model simplification, comprising the following steps:

S1. Obtain forest fire fire point monitoring data and extract the boundary of the affected area of the forest fire to obtain the boundary information of the affected area;

S2. Determine the speed set of the key parameters of the spreading model according to the boundary information of the disaster-affected area and construct the initial spreading model;

S3. Simplifying the initial spreading model to obtain a simplified discrete spreading model;

S4. Perform dynamic simulation of the disaster-affected area based on the simplified discrete sprawl model.

Further, the step of obtaining forest fire fire point monitoring data and extracting the disaster-affected area boundary of forest fire to obtain the disaster-affected area boundary information specifically includes:

S11. Download the surface temperature data obtained by satellite detection from the forest fire fire point monitoring website and read the original thermal infrared data set in the surface temperature data;

S12. Screen fire point information according to the original thermal infrared data set;

S13. Extracting the actual boundary of the affected area of the forest fire according to the fire point information.

Further, the step of determining the speed set of the key parameters of the sprawl model according to the boundary information of the disaster area and constructing the initial sprawl model specifically includes:

S21. Select the research area Ω and generate n discrete points in the research area Ω: x ₁ , x ₂ , x ₃ ,…, x _n ∈ Ω, and set the corresponding velocity set for these n discrete points: F(x ₁ ), F(x ₂ ), F(x ₃ ),…,F(x _n ).

Further, the step of simplifying the initial spreading model to obtain the simplified discrete spreading model specifically includes:

S31. Roughly classify the discrete points in the small triangle area that have not been correctly classified into k groups of data subsets according to the size of the Euclidean distance between them;

S32. Traverse the k sets of data subsets, select 4 discrete points in each set of data subsets as a 4-point group, and the number of 4-point groups is k groups;

S33, judging whether each group of 4 points is in the same small triangle area;

S34. Keep the 4-point groups included in the same small triangle and set the 4-point groups that are not in the same small triangle as null values;

S35, traverse all non-empty 4-point groups, remove redundant 4-point groups belonging to the same small triangle, and the number of remaining 4-point groups is k' groups;

S36. Traversing n discrete points in the area, and judging that the discrete point x _i and a certain 4-point group belong to the same small triangle, then adding the discrete point _xi to the class represented by the 4-point group, judging that the discrete point x _i does not Belong to any one of the k' 4-point groups, then add the point to the residue array, and the residue array is used to store the discrete points that need to be traversed in the next iteration;

S37. Add the convex hull boundary points of the class represented by each group of k' 4-point groups to the residue array and remove the duplicates, and add the convex hull boundary points of the class represented by each group of k' 4-point groups to final_TB array and deduplication.

S38. Repeat steps S31-S37 until the residue array is completely consistent with the final_TB array, obtain discrete points and their velocity sets required by the simplified model, and generate a simplified discrete spread model.

Further, the step of judging whether each 4-point group is in a small triangular area together specifically includes:

S331. Find 3 discrete points in each 4-point group, judge whether these 3 points x ₁ , x ₂ , x ₃ are not collinear, judge that the points x ₁ , x ₂ , x ₃ are not collinear, go to step S312, otherwise, continue to search for 3 discrete points that are not collinear;

S332. Select a point other than the three non-collinear points in the 4-point group as the fourth point x ₄ , and obtain the approximate reconstruction according to the preset reconstruction formula

S333. Calculate the error between speed sets

If it is judged that the error is less than the preset threshold, then find the group of 4 points belonging to the same small triangle.

Further, the expression of the preset reconstruction formula is as follows:

β ₁ +β ₂ +β ₃ =1.

In the above formula, β ₁ , β ₂ , and β ₃ represent convex combination coefficients, which satisfy the following expressions:

x ₄ =β ₁ ·x ₁ +β ₂ ·x ₂ +β ₃ ·x ₃ ,β ₁ +β ₂ +β ₃ =1.

The second technical solution adopted by the present invention is: a fire spread simulation acceleration system based on model simplification, comprising:

The information extraction module is used to obtain forest fire fire point monitoring data and extract the boundary of the affected area of the forest fire to obtain the boundary information of the affected area;

The initial model construction module is used to determine the speed set of the key parameters of the sprawl model according to the boundary information of the disaster area and construct the initial sprawl model;

The simplification module is used to simplify the initial spread model to obtain a simplified discrete spread model;

The dynamic simulation module performs dynamic simulation of the disaster-affected area based on the simplified discrete sprawl model.

Beneficial effect

The beneficial effects of the method and system of the present invention are: the present invention initially constructs the spread and simplifies the model, reduces redundant data, and doubles the time required to complete the dynamic simulation of the disaster-affected area while ensuring the prediction accuracy. Provide more time for follow-up disaster prediction, control strategy design, and disaster relief scheduling, which can reduce the damage caused by disasters and reduce losses.

Description of drawings

Fig. 1 is a kind of step flow chart of the fire spread simulation acceleration method based on model simplification of the present invention;

Fig. 2 is a schematic diagram of a selected region, a discrete point and a velocity set thereof according to a specific embodiment of the present invention;

Fig. 3 is a schematic diagram of a discrete point and its velocity set amplification according to a specific embodiment of the present invention;

Fig. 4 is a schematic diagram of a velocity set at a single discrete point according to a specific embodiment of the present invention;

Fig. 5 is a schematic diagram of discrete points and their velocity sets in the same small triangle according to a specific embodiment of the present invention;

Fig. 6 is a schematic diagram of a simplified initial spreading model according to a specific embodiment of the present invention;

Fig. 7 is a schematic diagram of discrete points and their velocity sets in the area before simplification according to a specific embodiment of the present invention;

Fig. 8 is a schematic diagram of discrete points and their velocity sets in the area after simplification according to a specific embodiment of the present invention;

Fig. 9 is a structural block diagram of a fire spread simulation acceleration system based on model simplification in the present invention.

Embodiments of the present invention

The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. For the step numbers in the following embodiments, it is only set for the convenience of illustration and description, and the order between the steps is not limited in any way. The execution order of each step in the embodiments can be adapted according to the understanding of those skilled in the art sexual adjustment.

With reference to Fig. 1, the present invention provides a kind of fire spread simulation acceleration method based on model simplification, and this method comprises the following steps:

Further as a preferred embodiment of this method, the step of obtaining the forest fire fire point monitoring data and extracting the disaster-affected area boundary of the forest fire to obtain the disaster-affected area boundary information specifically includes:

Specifically, read the DN value in the thermal infrared data set and calculate the radiance of the 22nd and 31st bands, and then use the radiance to calculate the brightness temperature values of the 22nd and 31st bands, according to the brightness of the 22nd and 31st bands Fire point information is filtered by the temperature value.

Further as a preferred embodiment of this method, the step of determining the speed set of the key parameters of the spreading model according to the boundary information of the disaster-affected area and constructing the initial spreading model specifically includes:

S21. Select the research area Ω and generate n discrete points in the research area Ω: x ₁ , x ₂ , x ₃ ,..., x _n ∈ Ω, and set the corresponding speed according to the n discrete points for the boundary information of the disaster area Set: F(x ₁ ),F(x ₂ ),F(x ₃ ),…,F(x _n );

Specifically, select a 128km×128km research area Ω=[-64,64]×[-64,64],

Randomly select n discrete points x={x ₁ ,x ₂ ,x ₃ ,…,x _n } in the area Ω, define the velocity set at point x _i as F( _xi ), 1≤i≤n,n ≈400, the generated effect is shown in Figure 2. where the box represents the boundary of the study area Ω, and each point cluster represents a discrete point and its velocity set. Refer to Figure 3 for the enlarged view of the discrete point and its velocity set, where the box indicates the selected area, the intersection indicates the discrete point, and the dots around the intersection indicate the velocity set of the discrete point.

In addition, with the boundary information of the affected area as the actual data support, the velocity set at the discrete point is set. Refer to Figure 4 for the schematic diagram of the velocity set. The central point represents the position of the discrete point, and the peripheral points represent the maximum spread of the modulus in a limited number of directions. The end point of the velocity vector, the length of the dotted line represents the modulus length of the velocity vector, therefore, the velocity vector represented by the peripheral point can be expressed by the following formula:

f(x)=(v*cosθ,v*sinθ), f(x)∈F(x).

In the above formula, v is the modulus length of the velocity vector, that is, the length of the dotted line between the center point and the peripheral point, and θ is the angle between the dotted line and the coordinate axis.

Further as a preferred embodiment of the method, referring to Fig. 6, the step of simplifying the initial spreading model to obtain the simplified discrete spreading control model specifically includes:

S33, judging whether each group of 4 points is in the same small triangle area;

Specifically, among the selected k 4-point groups, there may be 4-point groups with null values, and an operation of removing null values is required.

Specifically, after the above steps are completed, the n discrete points in the original model are either correctly divided into 4-point groups belonging to the same small triangle as it, or they have not been correctly classified and are stored in the residue array. The class represented by each group in these k' 4-point groups already contains all the points that belong to the same small triangle as the 4-point group

S38. Repeat steps S31-S37 until the rdsidue array is completely consistent with the final_TB array, obtain discrete points and their velocity sets required by the simplified model, and generate a simplified discrete-spread model.

Specifically, refer to FIG. 7 for n discrete points and their velocity sets in the region before simplification, and refer to FIG. 8 for m discrete points and their velocity sets after simplification.

Further as a preferred embodiment of this method, the problem of simplifying the initial spreading model specifically includes:

S39. Simplify the model, and find the most representative m discrete points in the region, where m<<n. The velocity sets at the m discrete points are used to approximate and reconstruct the velocity sets at all n points in the region, and the sum of errors between the approximately reconstructed velocity sets and the real velocity sets is minimized. Therefore, the mathematical description of the simplified problem of the sprawl control model is:

x _k ∈Δy _k1 y _k2 y _k3 , 1≤k≤n.

Further as a preferred embodiment of this method, said judging whether each group of 4 points is in a small triangular area

This step specifically includes:

Specifically, the judgment condition for judging whether it is non-collinear is

Rank is 2.

S332. In the 4-point group, select a point other than the three points that are not collinear as the fourth point x ₄ , and approximate reconstruction according to the preset reconstruction formula

constructed

S333. Calculate the error between speed sets

If it is judged that the error is less than the preset threshold, then find the

Group of 4 points of the same small triangle.

Further as a preferred embodiment of this method, the expression of the preset reconstruction formula is as follows:

β ₁ +β ₂ +β ₃ =1.

As shown in Figure 9, a fire spread simulation acceleration system based on model simplification includes:

The content in the above-mentioned method embodiments is applicable to this system embodiment. The specific functions realized by this system embodiment are the same as those of the above-mentioned method embodiments, and the beneficial effects achieved are also the same as those achieved by the above-mentioned method embodiments.

The above is a specific description of the preferred implementation of the present invention, but the invention is not limited to the described embodiments, and those skilled in the art can also make various equivalent deformations or replacements without violating the spirit of the present invention. , these equivalent modifications or replacements are all within the scope defined by the claims of the present application.

Claims

A fire spread simulation acceleration method based on model simplification, characterized in that it comprises the following steps:

S1. Obtain forest fire fire point monitoring data and extract the boundary of the affected area of the forest fire to obtain the boundary information of the affected area;

S2. Determine the speed set of the key parameters of the spreading model according to the boundary information of the disaster-affected area and construct the initial spreading model;

S3. Simplifying the initial spreading model to obtain a simplified discrete spreading model;

S4. Perform dynamic simulation of the disaster-affected area based on the simplified discrete sprawl model.
According to claim 1, a method for accelerating the simulation of fire spread based on model simplification, characterized in that, the step of obtaining forest fire fire point monitoring data and extracting the disaster-affected area boundary of forest fire to obtain the disaster-affected area boundary information, comprises Specifically include:

S11. Download the surface temperature data obtained by satellite detection from the forest fire fire point monitoring website and read the original thermal infrared data set in the surface temperature data;

S12. Screen fire point information according to the original thermal infrared data set;

S13. Extracting the actual boundary of the affected area of the forest fire according to the fire point information.
A fire spread simulation acceleration method based on model simplification according to claim 2, characterized in that the step of determining the speed set of the key parameters of the spread model and constructing the initial spread model according to the boundary information of the disaster area includes:

Select the research area Ω and generate n discrete points in the research area Ω: x 1 , x 2 , x 3 ,…, x n ∈ Ω, and set the corresponding velocity set for these n discrete points: F(x 1 ) ,F(x 2 ),F(x 3 ),…,F(x n ).
A fire spread simulation acceleration method based on model simplification according to claim 3, wherein the step of simplifying the initial spread model to obtain a simplified discrete spread model specifically includes:

S31. Roughly classify the discrete points in the small triangle area that have not been correctly classified into k groups of data subsets according to the size of the Euclidean distance between them;

S32. Traverse the k sets of data subsets, select 4 discrete points in each set of data subsets as a 4-point group, and the number of 4-point groups is k groups;

S33, judging whether each group of 4 points is in the same small triangle area;

S34. Keep the 4-point groups included in the same small triangle and set the 4-point groups that are not in the same small triangle as null values;

S35, traverse all non-empty 4-point groups, remove redundant 4-point groups belonging to the same small triangle, and the number of remaining 4-point groups is k' groups;

S36. Traversing n discrete points in the area, and judging that the discrete point x i and a certain 4-point group belong to the same small triangle, then adding the discrete point xi to the class represented by the 4-point group; judging that the discrete point xi does not Belong to any one of the k' 4-point groups, then add the point to the residue array, and the residue array is used to store the discrete points that need to be traversed in the next iteration;

S37. Add the convex hull boundary points of the class represented by each group of k' 4-point groups to the residue array and remove the duplicates, and add the convex hull boundary points of the class represented by each group of k' 4-point groups to In the final_TB array and deduplication, the final_TB array is used to store known convex hull boundary points;

S38. Repeat steps S31-S37 until the residue array is completely consistent with the final_TB array, obtain discrete points and their velocity sets required by the simplified model, and generate a simplified discrete spread model.
According to claim 4, a method for simulating fire spread based on model simplification, wherein the step of judging whether each 4-point group is in the same small triangular area includes:

S331. Find 3 discrete points in each 4-point group, judge whether these 3 points x 1 , x 2 , x 3 are not collinear, judge that the points x 1 , x 2 , x 3 are not collinear, go to step S312, otherwise, continue to search for 3 discrete points that are not collinear;

S332. Select a point other than the three non-collinear points in the 4-point group as the fourth point x 4 , and obtain the approximate reconstruction according to the preset reconstruction formula

S333. Calculate the error between speed sets
If it is judged that the error is less than the preset threshold, then find the group of 4 points belonging to the same small triangle.
A method for accelerating fire spread simulation based on model simplification according to claim 5, wherein the expression of the preset reconstruction formula is as follows:

In the above formula, β 1 , β 2 , and β 3 represent convex combination coefficients, which satisfy the following expressions:

x 4 =β 1 ·x 1 +β 2 ·x 2 +β 3 ·x 3 ,β 1 +β 2 +β 3 =1.
A fire spread simulation acceleration system based on model simplification, characterized in that it includes:

The information extraction module is used to obtain forest fire fire point monitoring data and extract the boundary of the affected area of the forest fire to obtain the boundary information of the affected area;

The initial model construction module is used to determine the speed set of the key parameters of the sprawl model according to the boundary information of the disaster area and construct the initial sprawl model;

The simplification module is used to simplify the initial spread model to obtain a simplified discrete spread model;

The dynamic simulation module performs dynamic simulation of the disaster-affected area based on the simplified discrete sprawl model.