Forest fire assessment method based on moisture content of surface litter

Abstract

The present invention provides a forest fire assessment method based on the moisture content of surface litter, which relates to the field of machine learning technology, including obtaining an analysis data set based on regularly collected surface litter data and meteorological data of a target forest and transmitting it to a target monitoring center; the target monitoring center extracts features and predicts the moisture content of the received analysis data set to obtain key prediction results; based on the key prediction results, a fire prediction is performed on the target forest to obtain a fire prediction assessment result, and a response is made. By collecting the surface litter data and meteorological data of the target forest and transmitting them to the target monitoring center; the target monitoring center extracts features and predicts the moisture content of the received data to obtain key prediction results; based on the key prediction results, a fire prediction is performed on the target forest to obtain a fire prediction assessment result and a response is made, the probability of forest fires can be effectively reduced, and the safety of forest resources and the ecological environment can be protected.

Description

Technical Field

[0001] The present invention relates to the technical field of machine learning, and in particular to a forest fire assessment method based on the moisture content of surface litter. Background Art

[0002] In recent years, with global warming, the frequency and destructiveness of forest fires have increased. This poses a serious threat to forest resources and the ecological environment. Traditional forest fire assessment methods often rely on meteorological factors, topography and other factors, ignoring the important indicator of surface litter moisture content, resulting in inaccurate forest fire assessment. Therefore, how to improve the accuracy of forest fire assessment, reduce the probability of forest fires, protect the safety of forest resources and the ecological environment, and provide a strong guarantee for the sustainable development of forests has become one of the current research centers.

[0003] Therefore, the present invention provides a forest fire assessment method based on the moisture content of surface litter. Summary of the invention

[0004] The present invention provides a forest fire assessment method based on the moisture content of surface litter, which is used to collect surface litter data and meteorological data of a target forest and transmit them to a target monitoring center; the target monitoring center performs feature extraction and moisture content prediction on the received data to obtain key prediction results; based on the key prediction results, a fire prediction is performed on the target forest, a fire prediction assessment result is obtained and a response is made, which can effectively reduce the probability of forest fires and protect the safety of forest resources and the ecological environment.

[0005] The present invention provides a forest fire assessment method based on the moisture content of surface litter, comprising:

[0006] Step 1: Based on the regularly collected surface litter data and meteorological data of the target forest, the analysis data set is obtained and transmitted to the target monitoring center;

[0007] Step 2: The target monitoring center extracts features and predicts moisture content of the received analysis data set to obtain key prediction results;

[0008] Step 3: Based on the key prediction results, fire prediction is performed on the target forest, fire prediction assessment results are obtained, and a response is made.

[0009] Preferably, based on the regularly collected surface litter data and meteorological data of the target forest, an analysis data set is obtained and transmitted to the target monitoring center, including:

[0010] Using the set monitoring equipment placed at the preset monitoring points, the surface litter data and meteorological data of each monitoring area of ​​the target forest are collected regularly;

[0011] The collected surface litter data and meteorological data are regarded as the first evaluation data;

[0012] The preset device information of the set monitoring device is extracted and combined with the corresponding first evaluation data collected, and transmitted to the target monitoring center as an analysis data set.

[0013] Preferably, the surface litter data refers to the moisture content, type and density of the litter.

[0014] Preferably, the target monitoring center performs feature extraction and moisture content prediction on the received analysis data set to obtain key prediction results, including:

[0015] The target monitoring center performs data preprocessing on the first evaluation data in the currently received analysis data set to obtain first key data;

[0016] Using the preset device information in the currently received analysis data set, including the device number of the set monitoring device, as a screening condition, extracting the first terrain condition from the set monitoring-terrain list;

[0017] Performing terrain annotation on the first key data using the first terrain condition;

[0018] Using the preset device information in the currently received analysis data set, the data collection time included in the data is time-stamped for the first key data;

[0019] Extracting features from the meteorological data in the annotated first key data to obtain key meteorological features;

[0020] Using the key meteorological characteristics and the first terrain condition, the moisture content of the surface litter at the next moment is predicted to obtain a moisture content prediction result;

[0021] The moisture content prediction result is combined with key meteorological characteristics and output as a key prediction result.

[0022] Preferably, the key meteorological feature and the first terrain condition are used to predict the moisture content of the surface litter at the next moment to obtain a moisture content prediction result, including:

[0023] According to the slope direction in the current first terrain condition, a matching prediction regression model is selected as the first regression prediction model;

[0024] Obtaining, from the model prediction accuracy table of the current first regression prediction model, a first prediction accuracy corresponding to the data collection moment of the first key data annotation;

[0025] The key meteorological feature, the first terrain condition and the first key data are input into the first regression prediction model to obtain a first prediction result;

[0026] If the first prediction accuracy is greater than the set prediction accuracy threshold, the first prediction result is output as the water content prediction result;

[0027] If the first prediction accuracy is not greater than the set prediction accuracy threshold, the key meteorological feature, the first terrain condition and the first key data are used as inputs into a pre-established water content prediction model to obtain a second prediction result;

[0028] Compare the first prediction result with the second prediction result to obtain an absolute difference in moisture content prediction;

[0029] When the absolute difference of the water content prediction is greater than the upper limit of the real-time prediction difference threshold, outputting the second prediction result as the water content prediction result;

[0030] When the absolute difference in the moisture content prediction is not greater than the upper limit of the real-time prediction difference threshold, the first prediction result and the second prediction result are comprehensively analyzed to obtain the moisture content prediction result and output it.

[0031] Preferably, when the absolute difference of the moisture content prediction is not greater than the upper limit of the real-time prediction difference threshold, the first prediction result and the second prediction result are comprehensively analyzed to obtain and output the moisture content prediction result, including:

[0032] When the absolute difference of the moisture content prediction is not greater than the upper limit of the real-time prediction difference threshold and is greater than the lower limit of the real-time prediction difference, the first prediction result and the second prediction result are combined and calculated to obtain a first moisture content, and the first moisture content is output as the moisture content prediction result;

[0033] The calculation formula of the first moisture content is as follows:

[0034] ; In the formula, Expressed as the first moisture content; Represented as the first prediction result; Represented as the second prediction result; It means setting the prediction accuracy threshold; It is expressed as the first prediction accuracy;

[0035] When the absolute difference of the moisture content prediction is not greater than the lower limit of the real-time prediction difference, determining the slope reference range according to the slope of the current monitoring area;

[0036] Select monitoring areas with the same slope direction as the current monitoring area, slopes within the slope reference range and similar surface litter height as the reference area from the target forest;

[0037] Calculate the slope difference between each reference area and the current monitoring area, and sort the reference areas from small to large according to the slope difference to obtain a reference area list;

[0038] Obtain the historical actual water content and historical predicted water content of each reference area at the last data collection moment and calculate them to obtain the historical predicted rate difference;

[0039] According to the reference area list, the reference slope-prediction curve is constructed with the slope of the reference area as the independent variable and the historical prediction rate difference as the dependent variable;

[0040] Extracting the curve characteristics of the reference slope-prediction curve and calculating it to obtain the slope influence coefficient;

[0041] The slope influence coefficient is comprehensively analyzed with meteorological data of different regions to obtain a second moisture content.

[0042] Preferably, the calculation formula of the second moisture content is as follows:

[0043] ; In the formula, Expressed as the second moisture content; Represented as the first prediction result; Represented as the second prediction result; It is expressed as the first prediction accuracy; It is expressed as the slope influence coefficient; It is expressed as the historical prediction rate difference of the reference area with the smallest slope difference with the current monitoring area; It is expressed as the slope of the reference area with the smallest slope difference with the current monitoring area; It is expressed as the slope of the current monitoring area; It is expressed as the i-th reference impact index, where i=1, 2, 3; the reference impact index includes the similarity of surface litter and meteorological characteristics between the monitoring area and the reference area with the smallest slope difference with the current monitoring area, and the moisture content change rate of the current monitoring area; It is represented as the influence weight of the i-th reference influencing index on the analyzed moisture content; e is represented as the natural base number.

[0044] Preferably, performing fire prediction on the target forest based on the key prediction results, obtaining fire prediction assessment results, and responding includes:

[0045] The key prediction result, the first key data and the corresponding data collection time are used as input to a fire prediction and evaluation model established by a pre-trained neural network to obtain a fire prediction and evaluation result;

[0046] Determine the current fire risk factor based on the fire occurrence probability and fire spread speed in the fire prediction assessment results;

[0047] Determining the fire risk level from a preset risk level table according to the fire risk coefficient;

[0048] When the fire risk level is low risk, the fire risk level, the first key data and the corresponding preset equipment information are combined to generate a fire risk warning report;

[0049] Transmitting the fire risk warning report to the forest management department to adopt the corresponding fire warning plan for processing;

[0050] When the fire risk level is medium risk or high risk, the fire risk level, the first key data and the corresponding preset equipment information are combined to generate a fire risk emergency report;

[0051] The fire risk emergency report is transmitted to the forest management department, and corresponding fire emergency plans are adopted to deal with it according to the fire risk level.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] By collecting the surface litter fall data and meteorological data of the target forest and transmitting them to the target monitoring center; the target monitoring center extracts features and predicts the moisture content of the received data to obtain key prediction results; based on the key prediction results, fire prediction is carried out on the target forest, and fire prediction assessment results are obtained and responded to, which can effectively reduce the probability of forest fires and protect the safety of forest resources and the ecological environment.

[0054] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0055] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0057] Figure 1 The present invention is a flowchart of a forest fire assessment method based on the moisture content of surface litter in an embodiment of the present invention. DETAILED DESCRIPTION

[0058] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0059] The embodiment of the present invention provides a forest fire assessment method based on the moisture content of surface litter, such as Figure 1 As shown, including:

[0060] Step 1: Based on the regularly collected surface litter data and meteorological data of the target forest, the analysis data set is obtained and transmitted to the target monitoring center;

[0061] Step 2: The target monitoring center extracts features and predicts moisture content of the received analysis data set to obtain key prediction results;

[0062] Step 3: Based on the key prediction results, fire prediction is performed on the target forest, fire prediction assessment results are obtained, and a response is made.

[0063] In this embodiment, the target forest refers to a forest that requires fire assessment; the surface litter data and meteorological data of the target forest collected regularly refer to the surface litter data and meteorological data of the target forest collected at set time intervals, for example, the surface litter data and meteorological data of the target forest are collected once every hour; the surface litter data refers to the moisture content, type and density of the litter; the meteorological data refers to meteorological parameters such as temperature, humidity, wind speed and wind direction; the analysis data set is composed of the first evaluation data and the preset equipment information of the set monitoring equipment, wherein the first evaluation data refers to the surface litter data and meteorological data of the target forest collected regularly; the set monitoring equipment refers to the tool used to collect surface litter data and meteorological data, such as a temperature sensor, a humidity sensor, a portable spectroradiometer, etc.; the preset equipment information includes the equipment number of the set monitoring equipment, the installation location coordinates, the data collection time and the collection coverage range.

[0064] In this embodiment, the target monitoring center refers to a cloud platform responsible for receiving, processing and analyzing data sets, feature extraction, moisture content prediction, and performing fire prediction and assessment based on these prediction results; key prediction results refer to the moisture content prediction results and key meteorological characteristics, such as temperature and humidity, obtained by the target monitoring center after feature extraction and moisture content prediction based on the analysis data set; wherein the moisture content prediction result refers to the predicted value of the moisture content of the surface litter at the next data collection moment; the fire prediction and assessment model refers to a mathematical model for evaluating the fire risk of the target forest based on the key prediction results; the fire prediction and assessment result refers to the assessment result of the fire risk of the target forest obtained after analyzing the key prediction results based on the fire prediction and assessment model.

[0065] The beneficial effects of the above technical solution are: by collecting surface litter data and meteorological data of the target forest and transmitting them to the target monitoring center; the target monitoring center extracts features and predicts the moisture content of the received data to obtain key prediction results; based on the key prediction results, fire prediction is performed on the target forest, and fire prediction evaluation results are obtained and responded to, which can effectively reduce the probability of forest fires and protect the safety of forest resources and the ecological environment.

[0066] The embodiment of the present invention provides a forest fire assessment method based on the moisture content of surface litter, which obtains an analysis data set based on regularly collected surface litter data and meteorological data of a target forest and transmits it to a target monitoring center, including:

[0067] Using the set monitoring equipment placed at the preset monitoring points, the surface litter data and meteorological data of each monitoring area of ​​the target forest are collected regularly;

[0068] The collected surface litter data and meteorological data are regarded as the first evaluation data;

[0069] The preset device information of the set monitoring device is extracted and combined with the corresponding first evaluation data collected, and transmitted to the target monitoring center as an analysis data set.

[0070] In this embodiment, the monitoring area refers to an area selected in the target forest where surface litter data and meteorological data are regularly collected in order to monitor and analyze the fire conditions of the forest; the analysis data set is composed of first evaluation data and preset equipment information of the set monitoring equipment, wherein the first evaluation data refers to the surface litter data and meteorological data of the target forest collected regularly; the set monitoring equipment refers to a tool for collecting surface litter data and meteorological data, such as a temperature sensor, a humidity sensor, a portable spectroradiometer, etc.; the preset equipment information includes the equipment number of the set monitoring equipment, the installation location coordinates, the data collection time and the collection coverage range, wherein the collection coverage range refers to the monitoring area monitored by the currently set monitoring equipment; the target monitoring center refers to a cloud platform responsible for receiving, processing and analyzing data sets, feature extraction, moisture content prediction, and performing fire prediction and evaluation based on these prediction results.

[0071] The beneficial effect of the above technical solution is: by using the set monitoring tools to regularly collect surface litter data and meteorological data of the target forest, and transmit them to the target monitoring center in a timely manner, data support can be provided for subsequent forest fire assessments.

[0072] The embodiment of the present invention provides a forest fire assessment method based on the moisture content of surface litter. The target monitoring center extracts features and predicts the moisture content of the received analysis data set to obtain key prediction results, including:

[0073] The target monitoring center performs data preprocessing on the first evaluation data in the currently received analysis data set to obtain first key data;

[0074] Using the preset device information in the currently received analysis data set, including the device number of the set monitoring device, as a screening condition, extracting the first terrain condition from the set monitoring-terrain list;

[0075] Performing terrain annotation on the first key data using the first terrain condition;

[0076] Using the preset device information in the currently received analysis data set, the data collection time included in the data is time-stamped for the first key data;

[0077] Extracting features from the meteorological data in the annotated first key data to obtain key meteorological features;

[0078] Using the key meteorological characteristics and the first terrain condition, the moisture content of the surface litter at the next moment is predicted to obtain a moisture content prediction result;

[0079] The moisture content prediction result is combined with key meteorological characteristics and output as a key prediction result.

[0080] In this embodiment, data preprocessing refers to cleaning, formatting, removing outliers and missing values ​​from the first evaluation data, and unifying the data format; the first key data refers to the data obtained after cleaning, formatting, removing outliers and missing values ​​from the first evaluation data, and unifying the data format; the preset equipment information includes setting the equipment number of the monitoring equipment, the installation location coordinates, the data collection time, and the collection coverage range.

[0081] In this embodiment, the set monitoring-terrain list refers to a list consisting of the equipment number of the set monitoring equipment and the corresponding terrain conditions; the first terrain condition refers to the slope aspect and altitude; the key meteorological characteristics refer to those obtained by feature extraction from the meteorological data in the first key data, including temperature, humidity, wind speed, etc.; the moisture prediction result refers to the predicted value of the moisture content of the surface litter at the next data collection moment; the next moment refers to the next data collection moment, for example, the current data collection moment is 13:00 and the next moment is 14:00; the key prediction result refers to the moisture content prediction result and key meteorological characteristics, such as temperature and humidity, obtained by the target monitoring center after feature extraction and moisture content prediction based on the analysis data set.

[0082] In this embodiment, for example, there is an analysis data set 1, and the first evaluation data in the analysis data set 1 are data 1, data 2, and data 3. Data preprocessing is performed on data 1, data 2, and data 3 respectively, and first key data t1, t2, and t3 are obtained accordingly;

[0083] According to the device number of the corresponding set monitoring device that collected data 1, data 2, and data 3 contained in the preset device information in the analysis data set 1, the terrain conditions are extracted from the set monitoring-terrain list, and conditions 1, condition 2, and condition 3 are obtained correspondingly, and all of them are regarded as the first terrain conditions;

[0084] Using the conditions 1, 2 and 3, the first key data t1, t2 and t3 are respectively annotated with corresponding terrain;

[0085] According to the data collection time included in the preset device information in the analysis data set 1, the first key data t1, t2 and t3 are time-marked.

[0086] The beneficial effects of the above technical solution are: obtaining the first key data by preprocessing the first evaluation data; and obtaining key evaluation features by extracting and screening the features after terrain annotation of the first key data, which can help improve the accuracy of data analysis and the subsequent prediction and evaluation level of forest fires.

[0087] The embodiment of the present invention provides a forest fire assessment method based on the moisture content of surface litter, which uses the key meteorological characteristics and the first terrain condition to predict the moisture content of surface litter at the next moment, and obtains a moisture content prediction result, including:

[0088] According to the slope direction in the current first terrain condition, a matching prediction regression model is selected as the first regression prediction model;

[0089] Obtaining, from the model prediction accuracy table of the current first regression prediction model, a first prediction accuracy corresponding to the data collection moment of the first key data annotation;

[0090] The key meteorological feature, the first terrain condition and the first key data are input into the first regression prediction model to obtain a first prediction result;

[0091] If the first prediction accuracy is greater than the set prediction accuracy threshold, the first prediction result is output as the water content prediction result;

[0092] If the first prediction accuracy is not greater than the set prediction accuracy threshold, the key meteorological feature, the first terrain condition and the first key data are used as inputs into a pre-established water content prediction model to obtain a second prediction result;

[0093] Compare the first prediction result with the second prediction result to obtain an absolute difference in moisture content prediction;

[0094] When the absolute difference of the water content prediction is greater than the upper limit of the real-time prediction difference threshold, outputting the second prediction result as the water content prediction result;

[0095] When the absolute difference in the moisture content prediction is not greater than the upper limit of the real-time prediction difference threshold, the first prediction result and the second prediction result are comprehensively analyzed to obtain the moisture content prediction result and output it.

[0096] In this embodiment, the slope aspect includes two types: sunny slope and shady slope; the first regression prediction model refers to a prediction regression model selected based on the current slope aspect, and different slope aspects are matched with different prediction regression models; the prediction regression model refers to a model obtained by training a random forest using pre-processed historical surface litter data and historical meteorological data of the forest as training data, and is used to predict the moisture content of forest surface litter; the model prediction accuracy table is composed of the daily data collection time of the current prediction regression model and the corresponding prediction accuracy, wherein the prediction accuracy is obtained by evaluating the accuracy of the model test results obtained by testing the prediction regression model using the historical surface litter moisture content and meteorological data of the forest;

[0097] The calculation formula for prediction accuracy is as follows:

[0098] ; In the formula, f represents the prediction accuracy; It is expressed as the predicted value of the test moisture content of the xth test sample, x=1, 2, 3, , b; It is expressed as the actual value of moisture content of the xth test sample; e is expressed as the natural base number.

[0099] In this embodiment, the model prediction accuracy table is regularly updated according to the prediction accuracy obtained by regularly testing and analyzing the prediction regression model. For example, the prediction regression model is tested and analyzed every month, and the model prediction accuracy table is updated according to the prediction accuracy obtained by the analysis.

[0100] In this embodiment, the first prediction accuracy refers to the prediction accuracy that matches the data collection time marked with the current first key data, obtained from the model prediction accuracy table of the current first regression prediction model; the first prediction result refers to the predicted value of the moisture content of the surface litter at the next data collection time output after inputting the key meteorological characteristics, the first terrain conditions and the first key data into the first regression prediction model; the prediction accuracy threshold is set in advance and is generally 0.8.

[0101] In this embodiment, the second prediction result refers to the key meteorological characteristics, the first terrain conditions and the first key data as inputs into the moisture prediction model, and the output is the predicted value of the moisture content of the surface litter at the next data collection moment, wherein the preset time period is predetermined; the absolute difference in moisture content prediction refers to the predicted absolute difference between the first prediction result and the second prediction result; the upper limit of the real-time prediction difference threshold refers to 80% of the average value of the current first prediction result and the second prediction result; the lower limit of the real-time prediction difference threshold refers to 20% of the average value of the current first prediction result and the second prediction result; the moisture prediction model refers to a model obtained by training a neural network using the pre-processed historical surface litter data and historical meteorological data of the forest as training data, and is used to predict the moisture content of forest surface litter.

[0102] In this embodiment, for example, there is a first prediction result 1 obtained at data collection time 1, and the first prediction accuracy of the current data collection time 1 obtained according to the model prediction accuracy table is 0.85, which is greater than the set prediction accuracy threshold 0.8. At this time, the first prediction result 1 is output as the water content prediction result.

[0103] In this embodiment, for example, there are a first prediction result r1 and a second prediction result r2, and ,in, It is expressed as the upper limit of the real-time prediction difference threshold;

[0104] At this time, the second prediction result r2 is output as the water content prediction result.

[0105] The beneficial effect of the above technical solution is: by utilizing the key meteorological characteristics and the first terrain condition, the moisture content of the current surface litter is predicted, and the moisture content prediction result is obtained, which can provide data support for subsequent forest fire assessment.

[0106] The embodiment of the present invention provides a forest fire assessment method based on the moisture content of surface litter. When the absolute difference of the moisture content prediction is not greater than the upper limit of the real-time prediction difference threshold, the first prediction result and the second prediction result are comprehensively analyzed to obtain and output the moisture content prediction result, including:

[0107] When the absolute difference of the moisture content prediction is not greater than the upper limit of the real-time prediction difference threshold and is greater than the lower limit of the real-time prediction difference, the first prediction result and the second prediction result are combined and calculated to obtain a first moisture content, and the first moisture content is output as the moisture content prediction result;

[0108] The calculation formula of the first moisture content is as follows:

[0109] ; In the formula, Expressed as the first moisture content; Represented as the first prediction result; Represented as the second prediction result; It means setting the prediction accuracy threshold; It is expressed as the first prediction accuracy;

[0110] When the absolute difference of the moisture content prediction is not greater than the lower limit of the real-time prediction difference, determining the slope reference range according to the slope of the current monitoring area;

[0111] Select monitoring areas with the same slope direction as the current monitoring area, slopes within the slope reference range and similar surface litter height as the reference area from the target forest;

[0112] Calculate the slope difference between each reference area and the current monitoring area, and sort the reference areas from small to large according to the slope difference to obtain a reference area list;

[0113] Obtain the historical actual water content and historical predicted water content of each reference area at the last data collection moment and calculate them to obtain the historical predicted rate difference;

[0114] According to the reference area list, the reference slope-prediction curve is constructed with the slope of the reference area as the independent variable and the historical prediction rate difference as the dependent variable;

[0115] Extracting the curve characteristics of the reference slope-prediction curve and calculating it to obtain the slope influence coefficient;

[0116] The slope influence coefficient is comprehensively analyzed with meteorological data of different regions to obtain a second moisture content.

[0117] In this embodiment, the upper limit of the real-time prediction difference threshold refers to 80% of the average value of the current first prediction result and the second prediction result; the lower limit of the real-time prediction difference threshold refers to 20% of the average value of the current first prediction result and the second prediction result.

[0118] In this embodiment, the first moisture content refers to the moisture content obtained by combining the first prediction result and the second prediction result when the absolute difference of the moisture content prediction is not greater than the upper limit of the real-time prediction difference threshold and greater than the lower limit of the real-time prediction difference; the slope reference range refers to ; The reference area refers to a monitoring area with the same slope direction as the current monitoring area, a slope within the slope reference range and a similar surface litter height.

[0119] In this embodiment, high similarity of surface litter means that the similarity between the surface litter of the current monitoring area and another monitoring area is greater than the set similarity threshold, where the surface litter similarity is equal to , where e represents the natural base, It is expressed as the ratio of the number of the same surface litter species in the current monitoring area and another monitoring area to the total number of surface litter species in the current monitoring area. It is expressed as the ratio of the total amount of surface litter in the current monitoring area to the total amount of surface litter in another monitoring area. The similarity threshold is set in advance, usually 0.8.

[0120] In this embodiment, for example, the current monitoring area is a sunny slope with a slope of 20°. At this time, the slope reference range is Right now ;

[0121] The monitoring area 1 is a sunny slope with a slope of 18° and a similarity of 0.85 with the surface litter in the current monitoring area;

[0122] The monitoring area 2 is a sunny slope with a slope of 21° and a similarity of 0.65 with the surface litter in the current monitoring area;

[0123] At this time, monitoring area 1 is used as the reference area of ​​the current monitoring area.

[0124] In this embodiment, the reference area list is obtained by sorting the reference areas in ascending order of slope difference, and is composed of reference areas and slope differences; the reference slope-prediction curve is a curve constructed with the slope in the reference area list as the independent variable and the historical prediction rate difference as the dependent variable; the slope influence coefficient is obtained by weighted averaging the curve features extracted from the reference slope-prediction curve, wherein the curve features include slope, acceleration, and the difference between the extreme value and the average value of the curve; the weights assigned to the curve features are obtained by solving a matrix constructed after pairwise comparison and relative importance scoring using the hierarchical analysis method; the historical prediction rate difference refers to the absolute difference between the historical actual water content of the reference area at the last data collection moment and the historical predicted water content.

[0125] The beneficial effect of the above technical solution is: by predicting the moisture content of the current surface litter, the moisture content prediction result can be obtained, which can provide data support for subsequent forest fire assessments, thereby helping to reduce the probability of forest fires and protect the safety of forest resources and the ecological environment.

[0126] The embodiment of the present invention provides a forest fire assessment method based on the moisture content of surface litter, and the calculation formula of the second moisture content is as follows:

[0127] ; In the formula, Expressed as the second moisture content; Represented as the first prediction result; Represented as the second prediction result; It is expressed as the first prediction accuracy; It is expressed as the slope influence coefficient; It is expressed as the historical prediction rate difference of the reference area with the smallest slope difference with the current monitoring area; It is expressed as the slope of the reference area with the smallest slope difference with the current monitoring area; It is expressed as the slope of the current monitoring area; It is expressed as the i-th reference impact index, where i=1, 2, 3; the reference impact index includes the similarity of surface litter and meteorological characteristics between the monitoring area and the reference area with the smallest slope difference with the current monitoring area, and the moisture content change rate of the current monitoring area; It is represented as the influence weight of the i-th reference influencing index on the analyzed moisture content; e is represented as the natural base number.

[0128] In this embodiment, the moisture content change rate is equal to ,in, It is expressed as the moisture content of the surface litter in the first key data of the current monitoring area; It is expressed as the historical actual moisture content of the current monitoring area at the last data collection moment; the historical actual moisture content refers to the moisture content of the surface litter collected by the set monitoring equipment.

[0129] In this embodiment, the similarity of surface litter is equal to ,in, It is expressed as the number of the same surface litter species in the current monitoring area and the reference area with the smallest slope difference with the current monitoring area; The total number of litter species on the ground in the current monitoring area; It is expressed as the total amount of litter on the ground in the current monitoring area; It is expressed as the total amount of surface litter in the reference area with the smallest slope difference with the current monitoring area.

[0130] In this embodiment, the meteorological feature similarity is equal to ,in, It is expressed as the absolute characteristic difference of the ath key meteorological characteristic at the current moment between the current monitoring area and the reference area with the smallest slope difference with the current monitoring area, a=1, 2, , L; L represents the total number of key meteorological features; It is expressed as the influence weight of the ath key meteorological feature on the moisture content; the weight assigned to the key meteorological feature is obtained by solving the matrix constructed after pairwise comparison and relative importance scoring of the key meteorological features using the hierarchical analysis method; the weight assigned to the reference impact index is obtained by solving the matrix constructed after pairwise comparison and relative importance scoring of the reference impact index using the hierarchical analysis method.

[0131] The beneficial effect of the above technical solution is that by calculating the second moisture content, data support can be provided for subsequent forest fire assessment, thereby helping to reduce the probability of forest fires and protect the safety of forest resources and the ecological environment.

[0132] The embodiment of the present invention provides a forest fire assessment method based on the moisture content of surface litter, which performs fire prediction on a target forest based on the key prediction result, obtains a fire prediction assessment result, and responds, including:

[0133] The key prediction result, the first key data and the corresponding data collection time are used as input to a fire prediction and evaluation model established by a pre-trained neural network to obtain a fire prediction and evaluation result;

[0134] Determine the current fire risk factor based on the fire occurrence probability and fire spread speed in the fire prediction assessment results;

[0135] Determining the fire risk level from a preset risk level table according to the fire risk coefficient;

[0136] When the fire risk level is low risk, the fire risk level, the first key data and the corresponding preset equipment information are combined to generate a fire risk warning report;

[0137] Transmitting the fire risk warning report to the forest management department to adopt the corresponding fire warning plan for processing;

[0138] When the fire risk level is medium risk or high risk, the fire risk level, the first key data and the corresponding preset equipment information are combined to generate a fire risk emergency report;

[0139] The fire risk emergency report is transmitted to the forest management department, and corresponding fire emergency plans are adopted to deal with it according to the fire risk level.

[0140] In this embodiment, the key prediction results refer to the moisture content prediction results and key meteorological characteristics, such as temperature and humidity, obtained by the target monitoring center after feature extraction and moisture content prediction based on the analysis data set; the fire prediction and evaluation model refers to a model obtained by training a neural network using the data obtained after data preprocessing and feature extraction of historical forest fire data, surface litter data and meteorological data as training data, wherein feature extraction refers to extracting features related to the occurrence and spread of fire, such as meteorological features and moisture content of surface litter; the fire prediction and evaluation results are composed of the probability of fire occurrence and the speed of fire spread; the fire risk coefficient is calculated by weighting the probability of fire occurrence and the speed of fire spread respectively, multiplying them, and then adding them up to obtain the average, wherein the weights assigned to the probability of fire occurrence and the speed of fire spread are obtained by solving the matrix constructed after pairwise comparison and relative importance scoring using the hierarchical analysis method.

[0141] In this embodiment, the preset risk level table is composed of a fire risk coefficient range and a corresponding fire risk level, wherein the fire risk level includes three types: low risk, medium risk and high risk; the fire risk emergency report and the fire risk warning report are composed of the report generation time, the fire risk level, the first key data and the corresponding preset equipment information, wherein the preset equipment information includes the equipment number of the set monitoring equipment, the installation location coordinates, the data collection time and the collection coverage range; the forest management department refers to the department responsible for the management and maintenance of forest resources, which is used to take corresponding fire handling plans according to the fire risk level of the received fire risk report, wherein the fire risk report includes two types: fire risk emergency report and fire risk warning report; the fire handling plan includes two types: fire warning plan and fire emergency plan, wherein the fire warning plan is composed of fire risk level, a list of fire warning handling personnel, a list of fire warning handling materials and a fire warning handling plan; the fire emergency plan is composed of fire risk level, a list of fire emergency handling personnel, a list of fire emergency handling materials and a fire rescue plan.

[0142] The beneficial effects of the above technical solution are: by conducting fire prediction and assessment based on the fire prediction and assessment model established by using a neural network and generating a fire risk report, the forest management department is promptly notified to adopt corresponding fire emergency plans for fire warning, which can buy precious time for preventing and responding to fires, thereby helping to reduce the probability of forest fires and protect the safety of forest resources and the ecological environment.

[0143] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A forest fire assessment method based on the moisture content of surface litter, characterized in that: include: Step 1: Based on the regularly collected surface litter data and meteorological data of the target forest, the analysis data set is obtained and transmitted to the target monitoring center; Step 2: The target monitoring center extracts features and predicts moisture content of the received analysis data set to obtain key prediction results; Step 3: Predict fires in the target forest based on the key prediction results, obtain fire prediction assessment results, and respond; The target monitoring center extracts features and predicts moisture content of the received analysis data set to obtain key prediction results, including: The target monitoring center performs data preprocessing on the first evaluation data in the currently received analysis data set to obtain first key data; Using the preset device information in the currently received analysis data set, including the device number of the set monitoring device, as a screening condition, extracting the first terrain condition from the set monitoring-terrain list; Performing terrain annotation on the first key data using the first terrain condition; Using the preset device information in the currently received analysis data set, the data collection time included in the data is time-stamped for the first key data; Extracting features from the meteorological data in the annotated first key data to obtain key meteorological features; Using the key meteorological characteristics and the first terrain condition, the moisture content of the surface litter at the next moment is predicted to obtain a moisture content prediction result; Combining the moisture content prediction result with key meteorological characteristics to output as a key prediction result; The key meteorological characteristics and the first terrain condition are used to predict the moisture content of the surface litter at the next moment, and the moisture content prediction result is obtained, including: According to the slope direction in the current first terrain condition, a matching prediction regression model is selected as the first regression prediction model; Obtaining, from the model prediction accuracy table of the current first regression prediction model, a first prediction accuracy corresponding to the data collection moment of the first key data annotation; The key meteorological feature, the first terrain condition and the first key data are input into the first regression prediction model to obtain a first prediction result; If the first prediction accuracy is greater than the set prediction accuracy threshold, the first prediction result is output as the moisture content prediction result; If the first prediction accuracy is not greater than the set prediction accuracy threshold, the key meteorological feature, the first terrain condition and the first key data are used as inputs into a pre-established water content prediction model to obtain a second prediction result; Compare the first prediction result with the second prediction result to obtain an absolute difference in moisture content prediction; When the absolute difference of the water content prediction is greater than the upper limit of the real-time prediction difference threshold, outputting the second prediction result as the water content prediction result; When the absolute difference of the moisture content prediction is not greater than the upper limit of the real-time prediction difference threshold, comprehensively analyzing the first prediction result and the second prediction result to obtain and output the moisture content prediction result; When the absolute difference of the moisture content prediction is not greater than the upper limit of the real-time prediction difference threshold, the first prediction result and the second prediction result are comprehensively analyzed to obtain the moisture content prediction result and output it, including: When the absolute difference of the moisture content prediction is not greater than the upper limit of the real-time prediction difference threshold and is greater than the lower limit of the real-time prediction difference, the first prediction result and the second prediction result are combined and calculated to obtain a first moisture content, and the first moisture content is output as the moisture content prediction result; The calculation formula of the first moisture content is as follows: ; In the formula, Expressed as the first moisture content; Represented as the first prediction result; Represented as the second prediction result; It means setting the prediction accuracy threshold; It is expressed as the first prediction accuracy; When the absolute difference of the moisture content prediction is not greater than the lower limit of the real-time prediction difference, determining the slope reference range according to the slope of the current monitoring area; Select monitoring areas with the same slope direction as the current monitoring area, slopes within the slope reference range and similar surface litter height as the reference area from the target forest; Calculate the slope difference between each reference area and the current monitoring area, and sort the reference areas from small to large according to the slope difference to obtain a reference area list; Obtain the historical actual water content and historical predicted water content of each reference area at the last data collection moment and calculate them to obtain the historical predicted rate difference; According to the reference area list, the reference slope-prediction curve is constructed with the slope of the reference area as the independent variable and the historical prediction rate difference as the dependent variable; Extracting the curve characteristics of the reference slope-prediction curve and calculating it to obtain the slope influence coefficient; The slope influence coefficient is comprehensively analyzed with meteorological data of different regions to obtain a second moisture content.

2. A forest fire assessment method based on the moisture content of surface litter according to claim 1, characterized in that: Based on the regularly collected surface litter data and meteorological data of the target forest, the analysis data set is obtained and transmitted to the target monitoring center, including: Using the set monitoring equipment placed at the preset monitoring points, the surface litter data and meteorological data of each monitoring area of ​​the target forest are collected regularly; The collected surface litter data and meteorological data are regarded as the first evaluation data; The preset device information of the set monitoring device is extracted and combined with the corresponding first evaluation data collected, and transmitted to the target monitoring center as an analysis data set.

3. A forest fire assessment method based on the moisture content of surface litter according to claim 2, characterized in that: Surface litter data refers to the moisture content, type and density of litter.

4. A forest fire assessment method based on moisture content of surface litter according to claim 1, characterized in that: The calculation formula for the second moisture content is as follows: ; In the formula, Expressed as the second moisture content; Represented as the first prediction result; Represented as the second prediction result; It is expressed as the first prediction accuracy; It is expressed as the slope influence coefficient; It is expressed as the historical prediction rate difference of the reference area with the smallest slope difference with the current monitoring area; It is expressed as the slope of the reference area with the smallest slope difference with the current monitoring area; It is expressed as the slope of the current monitoring area; It is expressed as the i-th reference impact index, where i=1, 2, 3; the reference impact index includes the similarity of surface litter and meteorological characteristics between the monitoring area and the reference area with the smallest slope difference with the current monitoring area, and the moisture content change rate of the current monitoring area; It is represented as the influence weight of the i-th reference influencing index on the analyzed moisture content; e is represented as the natural base number.

5. The forest fire assessment method based on the moisture content of surface litter according to claim 1, characterized in that: Based on the key prediction results, fire prediction is performed on the target forest, fire prediction assessment results are obtained, and responses are made, including: The key prediction result, the first key data and the corresponding data collection time are used as input to a fire prediction and evaluation model established by a pre-trained neural network to obtain a fire prediction and evaluation result; Determine the current fire risk factor based on the fire occurrence probability and fire spread speed in the fire prediction assessment results; Determining the fire risk level from a preset risk level table according to the fire risk coefficient; When the fire risk level is low risk, the fire risk level, the first key data and the corresponding preset equipment information are combined to generate a fire risk warning report; Transmitting the fire risk warning report to the forest management department to adopt the corresponding fire warning plan for processing; When the fire risk level is medium risk or high risk, the fire risk level, the first key data and the corresponding preset equipment information are combined to generate a fire risk emergency report; The fire risk emergency report is transmitted to the forest management department, and corresponding fire emergency plans are adopted to deal with it according to the fire risk level.

Images