CN-121999591-A - Fire grade determining method and device, electronic equipment and program product

CN121999591ACN 121999591 ACN121999591 ACN 121999591ACN-121999591-A

Abstract

The invention discloses a fire disaster grade determining method and device, electronic equipment and program products, and relates to the technical field of artificial intelligence, wherein the determining method comprises the steps of acquiring time sequence data of a position to be detected in a target area on the current time period, and determining a target weather feature vector based on a preset database, wherein the preset database comprises data of a plurality of positions in the target area; the time sequence data is decomposed to obtain decomposed data, the time sequence data, the target meteorological feature vector and the decomposed data are input into a first preset model to obtain a fire risk probability value of the position to be detected in a first preset period, the current period is earlier than the first preset period, and the fire risk probability value is used for determining the fire grade of the position to be detected. The invention solves the technical problem that fire disaster grade can not be accurately determined in the complex mountain area environment in the related technology.

Inventors

SUN WENHAO
RAO WEISHEN
LIU CHANG
ZHANG SIHANG
SHI ZHIYUAN
CHI XINGJIANG
ZHAO LIUXUE
DONG GUANGZHE
JIAO ZHENGGUO
ZHAO WEI
HAN XIAOKUN
WANG GUOLONG

Assignees

国网北京市电力公司
国家电网有限公司
国网电力工程研究院有限公司

Dates

Publication Date: 20260508
Application Date: 20260121

Claims (11)

1. A method for determining fire level, comprising: Acquiring time sequence data of a position to be detected in a target area in a current time period, and determining a target weather feature vector based on a preset database and the time sequence data, wherein the preset database comprises data of a plurality of positions in the target area; decomposing the time sequence data to obtain decomposed data; The time sequence data, the target weather feature vector and the decomposition data are input into a first preset model to obtain a fire risk probability value of the position to be detected in a first preset period, wherein the current period is earlier than the first preset period, the first preset model at least comprises an input layer, a hidden layer and an output layer, the input layer is used for receiving the input data, the hidden layer is used for extracting features of the input data, the output layer is used for outputting the fire risk probability value through a full connection layer and an activation function according to the features, and the fire risk probability value is used for determining fire grade of the position to be detected.
2. The fire level determination method according to claim 1, further comprising, before acquiring time series data of a position to be detected in the target area over a current period of time: acquiring data of a plurality of positions in the target area over a second preset period, and determining a feature vector of a fire risk factor of each position based on the data for each position, wherein the second preset period is earlier than the current period; Inputting each feature vector into a second preset classification model to obtain a fire risk classification result of the position, wherein a network model of the second preset classification model at least comprises a gating structure, a first classification model and a second classification model, wherein the gating structure is used for extracting information of each time step, outputting the fire risk classification result based on all the information, and the time steps are determined based on the feature vectors, and the fire risk classification result at least comprises a category and a probability value of the category; and constructing the preset database based on all the categories, all the probability values and all the data of the positions.
3. The fire level determination method according to claim 2, further comprising, before acquiring data of a plurality of the positions within the target area over a second preset period: Acquiring historical fire point time sequence data, wherein the historical fire point time sequence data at least comprises historical fire point data of a plurality of fire points and associated data sets corresponding to the historical fire point data, and the associated data sets at least comprise meteorological data, geographic information data and power transmission channel facility data; Processing the historical fire point time sequence data to obtain historical time sequence processing data; constructing a historical feature vector of the fire risk factor for each fire point based on the historical time sequence processing data; training the initial classification model based on the historical feature vector to obtain the second preset classification model.
4. A fire level determination method as defined in claim 3, wherein the step of constructing a historical feature vector of the fire risk factor for each of the fires based on the historical time series processing data includes: Determining a plurality of fire risk factors, wherein each fire risk factor at least comprises a plurality of fields; extracting the characteristics of the historical time sequence processing data to obtain characteristic data; Determining a field characteristic of each field based on all the characteristic data; for each of the fires, constructing the historical feature vector based on all of the field features.
5. The method of fire level determination according to claim 2, wherein the data includes at least first weather sequence data, and the step of constructing the preset database based on the data of all the categories, all the probability values, and all the positions further includes: Judging whether each category indicates that a fire disaster occurs at the position or not, and mapping the probability value of the category under the condition that the category indicates that the fire disaster occurs at the position to obtain a fire risk level corresponding to the first weather sequence data; For each fire risk level, judging whether the fire risk level is within a preset level threshold range, and adding the first meteorological sequence data to the preset database under the condition that the fire risk level is within the preset level threshold range, so as to construct the preset database.
6. The method of claim 1, wherein the time series data includes at least second weather sequence data, and wherein the step of determining the target weather feature vector based on the predetermined database and the time series data includes: calculating a similarity value between the second meteorological sequence data and each first meteorological sequence data in the preset database; comparing all the similarity values to obtain a comparison result; Determining a plurality of target weather sequence data based on the comparison result; distributing weights to each target weather sequence data based on the similarity value corresponding to each target weather sequence data to obtain weighted target weather sequence data; And fusing all the weighted target meteorological sequence data to obtain the target meteorological feature vector.
7. The fire level determination method according to claim 1, wherein the step of decomposing the time series data to obtain decomposed data comprises: Determining fire risk data corresponding to a plurality of fire risk factors from the time sequence data; decomposing all the fire risk data by adopting a data decomposition technology to obtain a trend item, a season item and a residual error item; the decomposition data is determined based on the trend term, the season term, and the residual term.
8. The fire level determination method according to claim 1, wherein after inputting the time series data, the target weather feature vector, and the decomposition data to a first preset model to obtain a fire probability value of the position to be detected over a first preset period, further comprising: mapping the fire probability value to obtain a target fire grade of the fire probability value; and determining a fire treatment strategy for the position to be detected based on the target fire grade.
9. A fire level determination apparatus, comprising: the determining unit is used for acquiring time sequence data of a position to be detected in a target area in a current time period, and determining a target weather feature vector based on a preset database and the time sequence data, wherein the preset database comprises data of a plurality of positions in the target area; The decomposition unit is used for decomposing the time sequence data to obtain decomposition data; The input unit is used for inputting the time sequence data, the target weather feature vector and the decomposition data into a first preset model to obtain a fire risk probability value of the position to be detected in a first preset period, wherein the current period is earlier than the first preset period, the first preset model at least comprises an input layer, a hidden layer and an output layer, the input layer is used for receiving the input data, the hidden layer is used for extracting the features of the input data, the output layer is used for outputting the fire risk probability value through a full connection layer and an activation function according to the features, and the fire risk probability value is used for determining the fire grade of the position to be detected.
10. A computer program product comprising a non-volatile computer readable storage medium storing a computer program which when executed by a processor implements the method of fire level determination of any one of claims 1 to 8.
11. An electronic device comprising one or more processors and a memory for storing one or more programs, wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method of determining fire rating of any of claims 1-8.

Description

Fire grade determining method and device, electronic equipment and program product Technical Field The invention relates to the technical field of artificial intelligence, in particular to a fire disaster grade determining method and device, electronic equipment and program products. Background As the power transmission network is continuously extended to complicated terrains such as mountain areas, mountain area power transmission channels face mountain fire threats, and the mountain fire may cause serious accidents such as tripping of the power transmission line and damage to towers, so that safe and stable operation of the power network and reliable supply of power are threatened. Currently, mountain fire monitoring mainly relies on manual inspection and observation tower to observe, but manual inspection is difficult to cover a large-area mountain area power transmission channel, and in the initial stage of mountain fire occurrence, when the fire is smaller and the smog is less, the mountain fire is difficult to discover in time, and the observation of the observation tower is also limited by factors such as topography, weather and the like, so that all-weather and omnibearing monitoring cannot be realized. With the progress of technology, satellite remote sensing technology is gradually applied to mountain fire monitoring, however, the spatial resolution of satellite remote sensing monitored data is relatively low, for some small-scale fires or fires hidden in complex terrains such as valleys, accurate detection may not be possible, the report rate of satellite remote sensing on small-area mountain fires is high, and the monitoring effect is often severely affected in overcast and rainy weather or when a cloud layer is thick. In addition, by using the meteorological data, a fire risk meteorological grade is established to evaluate the risk of mountain fire occurrence, but other factors influencing the mountain fire occurrence are ignored, so that the accuracy of the prediction result is not high enough. In view of the above problems, no effective solution has been proposed at present. Disclosure of Invention The embodiment of the invention provides a fire grade determining method and device, electronic equipment and program products, which at least solve the technical problem that fire grade cannot be accurately determined in a complicated mountain area environment in the related technology. According to one aspect of the embodiment of the application, a fire hazard level determining method is provided, which comprises the steps of obtaining time sequence data of a position to be detected in a target area in a current time period, determining a target weather feature vector based on a preset database and the time sequence data, wherein the preset database comprises data of a plurality of positions in the target area, decomposing the time sequence data to obtain decomposed data, inputting the time sequence data, the target weather feature vector and the decomposed data into a first preset model to obtain a fire hazard probability value of the position to be detected in the first preset time period, wherein the current time period is earlier than the first preset time period, the first preset model at least comprises an input layer, a hidden layer and an output layer, the input layer is used for receiving the input data, the hidden layer is used for extracting features of the input data, the output layer is used for outputting the fire hazard probability value according to the features through a full connection layer and an activation function, and the fire hazard probability value is used for determining the fire hazard level of the position to be detected. Further, before time sequence data of the position to be detected in the target area on the current time period are acquired, data of a plurality of positions in the target area on a second preset time period are acquired, feature vectors of fire risk factors of the positions are determined based on the data for each position, the second preset time period is earlier than the current time period, each feature vector is respectively input into a second preset classification model to obtain a fire risk classification result of the position, a network model of the second preset classification model at least comprises a gating structure, the gating structure is used for extracting information of each time step, the fire risk classification result is output based on all the information, the time steps are determined based on the feature vectors, the fire risk classification result at least comprises a category and a probability value of the category, and a preset database is built based on all the category, all the probability values and the data of all the positions. Further, before acquiring the data of the plurality of positions in the target area over the second preset period, acquiring historical fire time sequence data, wherein the historical fire t