A method for detecting moisture content of forest fuels based on LiDAR technology

Abstract

The present invention belongs to the technical field of moisture content detection and discloses a method for detecting the moisture content of forest combustibles based on LiDAR technology, comprising: obtaining LiDAR scanning data of a target area, extracting point cloud information from the scanning data, wherein the point cloud information includes signal strength and spatial coordinates; calculating the reflectivity and scattering characteristics of each point based on the point cloud information, and obtaining a reflectivity distribution map and a scattering characteristic map; obtaining reflectivity and scattering characteristic reference values corresponding to the type and growth stage of the combustibles from a pre-established database, comparing the reference values with actual measured values, and calculating a reflectivity deviation value and a scattering characteristic deviation value; obtaining precipitation data and transpiration data of the target area from a meteorological database, and calculating a moisture content change trend in a time dimension based on the precipitation data and transpiration data, and obtaining a moisture content time change curve.

Description

Technical Field

[0001] The present invention belongs to the technical field of moisture content detection, and in particular relates to a method for detecting the moisture content of forest combustibles based on LiDAR technology. Background Art

[0002] A key technical challenge in using LiDAR technology to measure forest fuel moisture content is establishing an accurate correlation model between fuel moisture content and LiDAR signals. Due to the complex and ever-changing forest environment, fuels vary greatly in type and growth stage, and their surface reflectivity and scattering properties vary significantly. Furthermore, the spatial distribution of fuels is highly uneven and random, further complicating modeling.

[0003] Traditional modeling methods rely primarily on extensive field sampling and experimental measurements. This involves manually collecting combustible samples, measuring their moisture content, and simultaneously acquiring corresponding LiDAR data. This data is then used to train a correlation model. However, this approach is time-consuming and labor-intensive, and it struggles to cover a sufficient range of combustible species and growth conditions, limiting the model's generalization and applicability.

[0004] Therefore, it is urgent to explore new modeling approaches that fully utilize the rich information contained in LiDAR data and characterize the dynamic changes in fuel moisture content from both temporal and spatial dimensions. In the temporal dimension, it is necessary to consider the influence of factors such as precipitation and evaporation on fuel moisture content, which exhibits distinct seasonal and cyclical characteristics. In the spatial dimension, it is necessary to consider the impact of factors such as topography and slope on moisture content and characterize the spatial distribution of fuel moisture content. Only by deeply analyzing the inherent relationship between LiDAR signals and moisture content in a dynamic spatiotemporal context can we establish more accurate and robust correlation models and provide reliable decision-making basis for forest fire early warning and prevention. Summary of the Invention

[0005] To solve the problems existing in the existing technology, the present invention provides a forest combustible moisture content detection method based on LiDAR technology. By integrating LiDAR scanning, combustible material classification, meteorological data analysis and machine learning technologies, it achieves accurate prediction of the combustible moisture content in the target area, providing an important basis for forest fire prevention and management, and has significant practical value and social benefits.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] A method for detecting moisture content of forest fuels based on LiDAR technology, the method comprising:

[0008] Acquire LiDAR scanning data of the target area, and extract point cloud information from the LiDAR scanning data, wherein the point cloud information includes: signal strength and spatial coordinates;

[0009] Calculating the reflectivity and scattering characteristics of each point based on the point cloud information to obtain a reflectivity distribution map and a scattering characteristic map;

[0010] Determine the type of combustible material based on the reflectivity distribution map and scattering characteristic map combined with the preset combustible material classification model;

[0011] Obtaining reflectivity and scattering characteristic reference values corresponding to the type and growth stage of the combustible material from a pre-established database, comparing the reference values with actual measured values, and calculating reflectivity deviation values and scattering characteristic deviation values;

[0012] Obtaining precipitation data and transpiration data of a target area from a meteorological database, and calculating a moisture content variation trend in a time dimension based on the precipitation data and transpiration data to obtain a moisture content time variation curve;

[0013] The corrected reflectivity and scattering characteristic values are combined with the moisture content time variation curve, and the results are input into a pre-trained machine learning model. A random forest algorithm is used to train a correlation model between the reflectivity, scattering characteristic, and moisture content, thereby obtaining a trained moisture content prediction model.

[0014] The moisture content of forest fuels is detected based on the trained moisture content prediction model.

[0015] Preferably, obtaining LiDAR scanning data of the target area and extracting point cloud information from the scanning data includes:

[0016] Obtain LiDAR scanning data of the target area, preprocess the LiDAR scanning data and extract point cloud data;

[0017] According to the spatial range of the scanning area, the point cloud data is spatially divided into multiple sub-areas;

[0018] According to the point cloud data of each sub-area, the three-dimensional spatial coordinates of the point cloud and the corresponding signal strength value are extracted to obtain the point cloud coordinates and strength information of the sub-area;

[0019] The point cloud coordinates and intensity information of each sub-area are merged to obtain the complete point cloud data of the target area, including the three-dimensional coordinates and signal intensity of the points.

[0020] Preferably, calculating the reflectivity and scattering characteristics of each point based on the point cloud information to obtain a reflectivity distribution map and a scattering characteristics map includes:

[0021] Obtaining the point cloud information, and calculating the reflectivity and scattering characteristic value of each point in the point cloud information;

[0022] Generate a reflectivity distribution map based on the calculated reflectivity value of each point, where different colors or grayscales represent different reflectivity value ranges;

[0023] Generate a scattering characteristic map based on the calculated scattering characteristic value of each point, where different colors or grayscales in the scattering characteristic map represent different scattering characteristic value ranges;

[0024] Perform image segmentation processing on the reflectivity distribution map to obtain the distribution position and range of different reflectivity areas;

[0025] Perform image segmentation processing on the scattering characteristic map to obtain the distribution position and range of different scattering characteristic areas;

[0026] Perform image registration on the reflectivity distribution map and the scattering characteristic map so that the positions of corresponding points in the two images coincide;

[0027] The registered reflectivity distribution map and scattering characteristic map are fused to generate a fused point cloud attribute map. Different colors or grayscales represent different combinations of reflectivity and scattering characteristics.

[0028] Preferably, the type of combustible material is determined based on the reflectivity distribution diagram and the scattering characteristic diagram in combination with a preset combustible material classification model, including:

[0029] Obtain a reflectivity distribution map, and for each point in the reflectivity distribution map, extract the reflectivity value of the point;

[0030] According to the preset combustible material classification model, determine whether the reflectivity value of the point is within the preset range;

[0031] If the reflectivity value of the point is within the preset range, the combustible type corresponding to the point is determined through the combustible type classification model;

[0032] If the reflectivity value of the point exceeds the preset range, the point is marked as an unknown type;

[0033] Classify and judge the combustible type for each point in the reflectivity distribution map;

[0034] According to the corresponding relationship between reflectivity value and combustible type, a combustible type classification model is constructed;

[0035] The reflectivity distribution map after classification judgment is used as the combustible material type distribution map for subsequent combustible material identification and positioning.

[0036] Preferably, obtaining reflectivity and scattering characteristic reference values corresponding to the combustible material type and growth stage from a pre-established database, comparing the reference values with actual measured values, and calculating the reflectivity deviation value and the scattering characteristic deviation value includes:

[0037] Obtaining information on the type of combustible material and information on its growth stage, and obtaining reference reflectivity and reference scattering characteristic values corresponding to the type and growth stage from a pre-established database;

[0038] Obtain image information of the combustible area, obtain spectral data of each pixel in the area based on the image information, and calculate the actual reflectivity and actual scattering characteristic value of each pixel through the spectral data;

[0039] Calculate the reflectivity deviation value of each pixel based on the actual reflectivity and reference reflectivity of each pixel. If the reflectivity deviation value is greater than the preset threshold, mark it.

[0040] Calculate the scattering characteristic deviation value of each pixel point based on the actual scattering characteristic value and the reference scattering characteristic value of each pixel point, and mark it if the scattering characteristic deviation value is greater than a preset threshold;

[0041] A support vector machine algorithm is used to train the image based on the reflectivity deviation value and the scattering characteristic deviation value of the marker, thereby obtaining a trained reflectivity deviation and scattering characteristic deviation classification model.

[0042] Using a K-nearest neighbor algorithm, based on a reflectivity deviation and scattering characteristic deviation classification model, the unlabeled reflectivity deviation values and the unlabeled scattering characteristic deviation values are classified to determine the classification results of the reflectivity deviation values and the scattering characteristic deviation values;

[0043] The random forest algorithm is used to judge whether the classification results of the reflectivity deviation value and the scattering characteristic deviation value meet the preset conditions according to the classification results of the reflectivity deviation value and the scattering characteristic deviation value, and to obtain the accuracy of the classification results of the reflectivity deviation value and the scattering characteristic deviation value.

[0044] Preferably, obtaining precipitation data and transpiration data of the target area from a meteorological database, and calculating the moisture content variation trend in the time dimension based on the precipitation data and transpiration data to obtain the moisture content time variation curve includes:

[0045] According to the geographical location information of the target area, the historical precipitation data and evaporation data of the area are obtained from the meteorological database;

[0046] Preprocess the acquired precipitation data and evaporation data to remove outliers and missing values;

[0047] Using the time series analysis method, the precipitation and evaporation at each time point are calculated to obtain the precipitation and evaporation data series in the time dimension;

[0048] According to the precipitation and transpiration data series, the moisture content at each time point is calculated to obtain the moisture content data series in the time dimension;

[0049] Conduct trend analysis on the moisture content data series and use moving average or exponential smoothing methods to determine the changing trend of moisture content;

[0050] According to the changing trend of moisture content, a moisture content time variation curve is generated to intuitively display the dynamic change process of moisture content.

[0051] Preferably, the corrected reflectivity and scattering characteristic values are combined with the moisture content time variation curve and input into a pre-trained machine learning model to train the correlation model between the reflectivity, scattering characteristic and moisture content. The trained moisture content prediction model includes:

[0052] Obtaining raw data of reflectivity and scattering characteristics, and correcting the raw data according to a preset correction rule to obtain corrected reflectivity and scattering characteristic values;

[0053] Obtaining curve data of moisture content changing over time, and correlating the curve data with the corrected reflectivity and scattering characteristic values to form a training data set;

[0054] The associated model is trained according to the input and output values of the training data set to obtain a trained moisture content prediction model.

[0055] Preferably, completing the forest fuel moisture content detection based on the trained moisture content prediction model includes:

[0056] According to the input values of the reflectivity and scattering characteristics to be measured, the trained moisture content prediction model is used to predict the corresponding moisture content prediction value;

[0057] Determine whether the predicted moisture content exceeds the preset threshold range. If so, trigger the early warning mechanism to indicate abnormal moisture content.

[0058] According to the deviation between the predicted moisture content and the actual moisture content, the moisture content prediction model is retrained to continuously optimize the prediction accuracy of the model;

[0059] The moisture content prediction results are associated with the time curve to generate a time variation curve of the moisture content prediction, providing data support for subsequent moisture content trend analysis.

[0060] Compared with the prior art, the present invention has the following beneficial effects:

[0061] The present invention discloses a method for detecting the moisture content of forest combustibles based on LiDAR technology. The method obtains LiDAR point cloud data of the target area, extracts reflectivity and scattering characteristic information, and determines the type of combustibles in combination with a preset combustible classification model. Subsequently, the actual measured value is compared with the reference value in the database to calculate the deviation value. At the same time, the time trend of the moisture content is calculated using meteorological data. Finally, the corrected reflectivity, scattering characteristic value and moisture content change curve are input into a pre-trained machine learning model, and a random forest algorithm is used to establish a moisture content prediction model. The present invention achieves accurate prediction of the moisture content of combustibles in the target area by integrating LiDAR scanning, combustible classification, meteorological data analysis and machine learning technologies, providing an important basis for forest fire prevention and management, and has significant practical value and social benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0063] Figure 1 This is a flow chart of a method for detecting moisture content of forest fuels based on LiDAR technology according to an embodiment of the present invention;

[0064] Figure 2 This is a schematic diagram of the steps of building a moisture content prediction model according to an embodiment of the present invention;

[0065] Figure 3 This is a schematic diagram of the LSTM unit structure according to an embodiment of the present invention;

[0066] Figure 4 Schematic diagram of a combustible moisture content prediction model based on informer according to an embodiment of the present invention. DETAILED DESCRIPTION

[0067] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0068] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0069] Example 1

[0070] like Figure 1 As shown, an embodiment of the present invention provides a method for detecting moisture content of forest fuels based on LiDAR technology, the method comprising:

[0071] Acquire LiDAR scanning data of the target area, and extract point cloud information from the LiDAR scanning data, wherein the point cloud information includes: signal strength and spatial coordinates;

[0072] Calculating the reflectivity and scattering characteristics of each point based on the point cloud information to obtain a reflectivity distribution map and a scattering characteristic map;

[0073] Determine the type of combustible material based on the reflectivity distribution map and scattering characteristic map combined with the preset combustible material classification model;

[0074] Obtaining reflectivity and scattering characteristic reference values corresponding to the type and growth stage of the combustible material from a pre-established database, comparing the reference values with actual measured values, and calculating reflectivity deviation values and scattering characteristic deviation values;

[0075] Obtaining precipitation data and transpiration data of a target area from a meteorological database, and calculating a moisture content variation trend in a time dimension based on the precipitation data and transpiration data to obtain a moisture content time variation curve;

[0076] The corrected reflectivity and scattering characteristic values are combined with the moisture content time variation curve, and the results are input into a pre-trained machine learning model. A random forest algorithm is used to train a correlation model between the reflectivity, scattering characteristic, and moisture content, thereby obtaining a trained moisture content prediction model.

[0077] The moisture content of forest fuels is detected based on the trained moisture content prediction model.

[0078] In this embodiment, obtaining LiDAR scanning data of a target area and extracting point cloud information from the scanning data include:

[0079] Obtain LiDAR scanning data of the target area, preprocess the LiDAR scanning data and extract point cloud data;

[0080] According to the spatial range of the scanning area, the point cloud data is spatially divided into multiple sub-areas;

[0081] According to the point cloud data of each sub-area, the three-dimensional spatial coordinates of the point cloud and the corresponding signal strength value are extracted to obtain the point cloud coordinates and strength information of the sub-area;

[0082] The point cloud coordinates and intensity information of each sub-area are merged to obtain the complete point cloud data of the target area, including the three-dimensional coordinates and signal intensity of the points.

[0083] Specifically, spatial division of point cloud data includes: spatial analysis is the process of revealing the spatial pattern contained in spatial data and the relationship between elements, which is the most distinctive content in GIS. In order to solve the problem of data imitation and precision in current spatial analysis, the present invention first introduces LiDAR point cloud data into the spatial analysis process, with the purpose of generating high-precision DEM and DSM data through data classification and extraction; then extracts combustible element data related to spatial analysis from vector DLG data, and generates raster data of combustible elements after rasterization; then uses DEM data and DSM data generated from LiDAR point cloud data, combined with rasterized combustible data, and after normalization and other processing, generates a spatial analysis raster map under a single influencing factor; finally, based on the influence weights of various data on spatial analysis, a raster analysis raster map of multiple influencing factors is generated through raster overlay analysis, thereby improving the accuracy and efficiency of complex spatial analysis algorithms.

[0084] In this embodiment, the reflectivity and scattering characteristics of each point are calculated based on the point cloud information to obtain a reflectivity distribution map and a scattering characteristic map, including:

[0085] Obtaining the point cloud information, and calculating the reflectivity and scattering characteristic value of each point in the point cloud information;

[0086] Generate a reflectivity distribution map based on the calculated reflectivity value of each point, where different colors or grayscales represent different reflectivity value ranges;

[0087] Generate a scattering characteristic map based on the calculated scattering characteristic value of each point, where different colors or grayscales in the scattering characteristic map represent different scattering characteristic value ranges;

[0088] Perform image segmentation processing on the reflectivity distribution map to obtain the distribution position and range of different reflectivity areas;

[0089] Perform image segmentation processing on the scattering characteristic map to obtain the distribution position and range of different scattering characteristic areas;

[0090] Perform image registration on the reflectivity distribution map and the scattering characteristic map so that the positions of corresponding points in the two images coincide;

[0091] The registered reflectivity distribution map and scattering characteristic map are fused to generate a fused point cloud attribute map. Different colors or grayscales represent different combinations of reflectivity and scattering characteristics.

[0092] In this embodiment, the types of combustibles determined based on the reflectivity distribution diagram and the scattering characteristic diagram in combination with the preset combustible classification model include:

[0093] Obtain a reflectivity distribution map, and for each point in the reflectivity distribution map, extract the reflectivity value of the point;

[0094] According to the preset combustible material classification model, determine whether the reflectivity value of the point is within the preset range;

[0095] If the reflectivity value of the point is within the preset range, the combustible type corresponding to the point is determined through the combustible type classification model;

[0096] If the reflectivity value of the point exceeds the preset range, the point is marked as an unknown type;

[0097] Classify and judge the combustible type for each point in the reflectivity distribution map;

[0098] According to the corresponding relationship between reflectivity value and combustible type, a combustible type classification model is constructed;

[0099] The reflectivity distribution map after classification judgment is used as the combustible material type distribution map for subsequent combustible material identification and positioning.

[0100] Specifically, the preset combustible species classification model is established based on a large amount of experimental data. For example, the reflectivity of wood may be between 0.5 and 0.8, while that of plastic may be between 0.3 and 0.6. When encountering a point with a reflectivity of 0.7, the system will determine that it falls within the preset range for wood. Subsequently, the combustible species classification model will further analyze it and may determine it to be pine. If a point with a reflectivity of 0.2 is encountered, which is outside the preset range of common combustibles, the system will mark it as an unknown species, which helps to identify potential anomalies or new combustibles.

[0101] The classification and judgment of combustible types for each point in the reflectance distribution map include: first, using the successive projection algorithm (SPA) to screen the basic bands, and comparing and analyzing the classification accuracy of two deep learning models, one-dimensional convolutional neural network (1DCNN) and long short-term memory artificial neural network (LSTM) under the conditions of original spectrum, characteristic bands and partial characteristic bands, to explore the information carrying capacity of characteristic bands for the original spectrum; then, to address the misclassification problem, an advanced band screening method is used to retrain the misclassified samples of various types of combustibles under the conditions of basic variable combinations, and this cycle is repeated until there is no significant increase in classification accuracy, and the spectral characteristics and misclassification rules of the misclassified samples are studied; finally, the classification accuracy of different methods is compared.

[0102] Specifically, the constructed fuel type classification model includes: (1) primary fuel type. Referring to the IGBP land cover classification system, the primary fuel type is divided into three fuel types: forest, shrub and herb. (2) secondary fuel type. The primary fuel type is further divided into secondary types. Among them, forests are divided into coniferous forests, broad-leaved forests and mixed forests according to leaf shape; shrubs are divided into forested shrubs and shrub types; and herbs are divided into woody vegetation-grass mosaic and grass types. (3) tertiary fuel type. NTC can reflect the distribution of ground fuels. Taking NTC (non-tree cover percentage) as the basis for division, the secondary fuel type is further divided into tertiary types. Drawing on the Prometheus fuel classification system, in coniferous forests, broad-leaved forests and mixed forest types, NTC ≤ 30% is classified as low shrub-grass cover type, and NTC > 30% is classified as high shrub-grass cover type; in forested shrub and shrub types, NTC ≤ 60% is classified as low shrub-grass cover type, and NTC > 60% is classified as high shrub-grass cover type; for woody vegetation-grass mosaic and grass types, NTC > 50% or 50% ≥ NTC > 10% is used to divide high cover and low cover types.

[0103] In this embodiment, obtaining reflectivity and scattering characteristic reference values corresponding to the combustible material type and growth stage from a pre-established database, comparing the reference values with actual measured values, and calculating the reflectivity deviation value and the scattering characteristic deviation value includes:

[0104] Obtaining information on the type of combustible material and information on its growth stage, and obtaining reference reflectivity and reference scattering characteristic values corresponding to the type and growth stage from a pre-established database;

[0105] Obtain image information of the combustible area, obtain spectral data of each pixel in the area based on the image information, and calculate the actual reflectivity and actual scattering characteristic value of each pixel through the spectral data;

[0106] Calculate the reflectivity deviation value of each pixel based on the actual reflectivity and reference reflectivity of each pixel. If the reflectivity deviation value is greater than the preset threshold, mark it.

[0107] Calculate the scattering characteristic deviation value of each pixel point based on the actual scattering characteristic value and the reference scattering characteristic value of each pixel point, and mark it if the scattering characteristic deviation value is greater than a preset threshold;

[0108] A support vector machine algorithm is used to train the image based on the reflectivity deviation value and the scattering characteristic deviation value of the marker, thereby obtaining a trained reflectivity deviation and scattering characteristic deviation classification model.

[0109] Using a K-nearest neighbor algorithm, based on a reflectivity deviation and scattering characteristic deviation classification model, the unlabeled reflectivity deviation values and the unlabeled scattering characteristic deviation values are classified to determine the classification results of the reflectivity deviation values and the scattering characteristic deviation values;

[0110] The random forest algorithm is used to judge whether the classification results of the reflectivity deviation value and the scattering characteristic deviation value meet the preset conditions according to the classification results of the reflectivity deviation value and the scattering characteristic deviation value, and to obtain the accuracy of the classification results of the reflectivity deviation value and the scattering characteristic deviation value.

[0111] Specifically, a pre-established database stores reference reflectance and scattering properties for various combustible materials at different growth stages. For example, the reference reflectance for pine trees in the seedling stage might be between 0.3 and 0.4, while that for mature trees might be between 0.2 and 0.3. These reference values provide a baseline for subsequent deviation calculations.

[0112] In this embodiment, precipitation data and transpiration data of the target area are obtained from the meteorological database. Based on the precipitation data and transpiration data, the moisture content variation trend in the time dimension is calculated to obtain the moisture content time variation curve including:

[0113] According to the geographical location information of the target area, the historical precipitation data and evaporation data of the area are obtained from the meteorological database;

[0114] Preprocess the acquired precipitation data and evaporation data to remove outliers and missing values;

[0115] Using the time series analysis method, the precipitation and evaporation at each time point are calculated to obtain the precipitation and evaporation data series in the time dimension;

[0116] According to the precipitation and transpiration data series, the moisture content at each time point is calculated to obtain the moisture content data series in the time dimension;

[0117] Conduct trend analysis on the moisture content data series and use moving average or exponential smoothing methods to determine the changing trend of moisture content;

[0118] According to the changing trend of moisture content, a moisture content time variation curve is generated to intuitively display the dynamic change process of moisture content.

[0119] In this embodiment, the corrected reflectivity and scattering characteristic values are combined with the moisture content time variation curve and input into a pre-trained machine learning model to train the correlation model between the reflectivity, scattering characteristic and moisture content. The trained moisture content prediction model includes:

[0120] Obtaining raw data of reflectivity and scattering characteristics, and correcting the raw data according to a preset correction rule to obtain corrected reflectivity and scattering characteristic values;

[0121] Obtaining curve data of moisture content changing over time, and correlating the curve data with the corrected reflectivity and scattering characteristic values to form a training data set;

[0122] The associated model is trained according to the input and output values of the training data set to obtain a trained moisture content prediction model.

[0123] Specifically, the modeling steps of the moisture content prediction model are as follows: Figure 2 As shown in the figure, the meteorological elements and moisture content data measured in the field are divided into training set and test set. The training set is used to train the deep learning model and fit the direct estimation method equation parameters, and the test set is used to evaluate the prediction accuracy of the model. The moisture content series prediction models are constructed for the two types of combustible materials respectively.

[0124] Reconstruction direct estimation method: First, the Simard equilibrium moisture content equation is selected, such as formula (1):

[0125]

[0126] Where: E is the equilibrium moisture content, H is the relative humidity, and T is the temperature.

[0127] According to Byram's water diffusion equation:

[0128]

[0129] Where: d m / d t represents the change in moisture content of litter in time period t; m represents the moisture content of fuel; E represents the equilibrium moisture content of litter; τ represents the time lag of litter, M t represents the moisture content at time t, and M0 represents the initial moisture content.

[0130] Let dt = Δt = 1h, discretize the formula, and obtain the litter moisture content calculation equation in discrete form:

[0131] M t =λ 2 M t-1 +λ(1-λ)E t-1 +(1-λ)E t (3)

[0132] Where: E t-1 and E t Represent the equilibrium moisture content at time t and t-1 respectively; M t and M t-1represents the moisture content at time t and time t-1, respectively; λ is the parameter to be estimated. Using the fitted Simard equilibrium moisture content equation, substitute it into the above equation. Using SPSS nonlinear regression, fit the model parameters to obtain the value of parameter λ in equation (4).

[0133] M t+j =λ 2 M t+j-1 +λ(1-λ)E t+j-1 +(1-λ)E t+j (4)

[0134] If the water content sequence with hourly steps in the future i is predicted, that is, the water content in the interval [t, t+i] is predicted. When j = 0, 1, ..., i, the recursive calculation results in M t , M t+1 ,…M t+i .

[0135] The prediction of moisture content series is actually a time series prediction problem. In the rolling prediction setting with a fixed window size, the input of the model at time point t is:

[0136]

[0137] Where: X t is the input of the model. x Indicates the current input sequence length, Represents the input vector of the i-th one in the window at time point t, each input is d x A vector of latitudes.

[0138] The output is the sequence corresponding to the prediction:

[0139]

[0140] Where: Y t is the output of the model. y Indicates the current output sequence length, Represents the output vector of the i-th one in the window at time point t, each output is d y A vector of latitudes.

[0141] In a standard recurrent neural network (RNN), the vanishing gradient problem makes it difficult for the network to learn long-term dependencies in time series over time. However, the Long Short-Term Memory (LSTM) network solves the vanishing gradient problem through its unique design structure, effectively retaining long-term information, enabling the model to retain information in very long sequences and make effective predictions. Figure 3 As shown in the figure, the LSTM unit mainly includes an input gate, a forget gate, an output gate, and a unit state that maintains the internal state.

[0142] When using LSTM to predict long sequences, because time t depends on the output at time t-1, the longer the sequence, the slower and less effective the prediction. Furthermore, because the LSTM model uses a backpropagation algorithm to calculate the loss function and optimize model parameters, convergence is difficult. To address the poor performance and slow convergence of the LSTM model in long sequence prediction, we attempted to apply the informer model to predict moisture content.

[0143] like Figure 4 As shown in the figure, the Informer model consists of five main parts: input, encoder, decoder, fully connected layer, and output. The model utilizes some core components of the Transformer model (such as the encoder and decoder). It uses a multi-head sparse self-attention mechanism to concurrently calculate the closeness of the relationship between the current time point and previous time points, assigning greater weight to important time point features. This reduces model complexity while preventing significant information loss. Convolutional pooling operations are also added between adjacent attention blocks to downsample features, reducing the length of the input sequence and significantly improving speed.

[0144] The calculation of the multi-head sparse probabilistic self-attention mechanism is shown in formula (7), which contains three vectors: Qu-ery vector (Q), Key vector (K) and Value vector (V).

[0145]

[0146] Where: L Q , L K and L V They represent the linear transformation layers of the query set, key set, and value set in the informer module, respectively, and d is the input dimension.

[0147] In this embodiment, the detection of forest fuel moisture content based on the trained moisture content prediction model includes:

[0148] According to the input values of the reflectivity and scattering characteristics to be measured, the trained moisture content prediction model is used to predict the corresponding moisture content prediction value;

[0149] Determine whether the predicted moisture content exceeds the preset threshold range. If so, trigger the early warning mechanism to indicate abnormal moisture content.

[0150] According to the deviation between the predicted moisture content and the actual moisture content, the moisture content prediction model is retrained to continuously optimize the prediction accuracy of the model;

[0151] The moisture content prediction results are associated with the time curve to generate a time variation curve of the moisture content prediction, providing data support for subsequent moisture content trend analysis.

[0152] The embodiments described above are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by persons skilled in the art should fall within the scope of protection defined by the claims of the present invention.

Claims

1. A method for detecting moisture content of forest fuels based on LiDAR technology, characterized in that: The method comprises: Acquire LiDAR scanning data of the target area, and extract point cloud information from the LiDAR scanning data, wherein the point cloud information includes: signal strength and spatial coordinates; Calculating the reflectivity and scattering characteristics of each point based on the point cloud information to obtain a reflectivity distribution map and a scattering characteristic map; Determine the type of combustible material based on the reflectivity distribution map and scattering characteristic map combined with the preset combustible material classification model; Obtaining reflectivity and scattering characteristic reference values corresponding to the type and growth stage of the combustible material from a pre-established database, comparing the reference values with actual measured values, and calculating reflectivity deviation values and scattering characteristic deviation values; Obtaining precipitation data and transpiration data of a target area from a meteorological database, and calculating a moisture content variation trend in a time dimension based on the precipitation data and transpiration data to obtain a moisture content time variation curve; Combining the corrected reflectivity and scattering characteristic values with the moisture content time variation curve, and inputting the results into a pre-trained machine learning model, training a correlation model between the reflectivity, scattering characteristic, and moisture content, and obtaining a trained moisture content prediction model; The moisture content of forest fuels is detected based on the trained moisture content prediction model.

2. The method according to claim 1, characterized in that Obtaining LiDAR scanning data of a target area and extracting point cloud information from the scanning data includes: Obtain LiDAR scanning data of the target area, preprocess the LiDAR scanning data and extract point cloud data; According to the spatial range of the scanning area, the point cloud data is spatially divided into multiple sub-areas; According to the point cloud data of each sub-area, the three-dimensional spatial coordinates of the point cloud and the corresponding signal strength value are extracted to obtain the point cloud coordinates and strength information of the sub-area; The point cloud coordinates and intensity information of each sub-area are merged to obtain the complete point cloud data of the target area, including the three-dimensional coordinates and signal intensity of the points.

3. The method according to claim 1, characterized in that According to the point cloud information, the reflectivity and scattering characteristics of each point are calculated to obtain a reflectivity distribution map and a scattering characteristic map, including: Obtaining the point cloud information, and calculating the reflectivity and scattering characteristic value of each point in the point cloud information; Generate a reflectivity distribution map based on the calculated reflectivity value of each point, where different colors or grayscales represent different reflectivity value ranges; Generate a scattering characteristic map based on the calculated scattering characteristic value of each point, where different colors or grayscales in the scattering characteristic map represent different scattering characteristic value ranges; Perform image segmentation processing on the reflectivity distribution map to obtain the distribution position and range of different reflectivity areas; Perform image segmentation processing on the scattering characteristic map to obtain the distribution position and range of different scattering characteristic areas; Perform image registration on the reflectivity distribution map and the scattering characteristic map so that the positions of corresponding points in the two images coincide; The registered reflectivity distribution map and scattering characteristic map are fused to generate a fused point cloud attribute map. Different colors or grayscales represent different combinations of reflectivity and scattering characteristics.

4. The method according to claim 1, wherein Based on the reflectivity distribution map and scattering characteristic map, combined with the preset combustible material classification model, the combustible material types include: Obtain a reflectivity distribution map, and for each point in the reflectivity distribution map, extract the reflectivity value of the point; According to the preset combustible material classification model, determine whether the reflectivity value of the point is within the preset range; If the reflectivity value of the point is within the preset range, the combustible type corresponding to the point is determined through the combustible type classification model; If the reflectivity value of the point exceeds the preset range, the point is marked as an unknown type; Classify and judge the combustible type for each point in the reflectivity distribution map; According to the corresponding relationship between reflectivity value and combustible type, a combustible type classification model is constructed; The reflectivity distribution map after classification judgment is used as the combustible material type distribution map for subsequent combustible material identification and positioning.

5. The method according to claim 1, characterized in that Obtaining reflectivity and scattering characteristic reference values corresponding to the combustible type and growth stage from a pre-established database, comparing the reference values with actual measured values, and calculating reflectivity deviation values and scattering characteristic deviation values includes: Obtaining information on the type of combustible material and information on its growth stage, and obtaining reference reflectivity and reference scattering characteristic values corresponding to the type and growth stage from a pre-established database; Obtain image information of the combustible area, obtain spectral data of each pixel in the area based on the image information, and calculate the actual reflectivity and actual scattering characteristic value of each pixel through the spectral data; Calculate the reflectivity deviation value of each pixel based on the actual reflectivity and reference reflectivity of each pixel. If the reflectivity deviation value is greater than the preset threshold, mark it. Calculate the scattering characteristic deviation value of each pixel point based on the actual scattering characteristic value and the reference scattering characteristic value of each pixel point, and mark it if the scattering characteristic deviation value is greater than a preset threshold; A support vector machine algorithm is used to train the image based on the reflectivity deviation value and the scattering characteristic deviation value of the marker, thereby obtaining a trained reflectivity deviation and scattering characteristic deviation classification model. Using a K-nearest neighbor algorithm, based on a reflectivity deviation and scattering characteristic deviation classification model, the unlabeled reflectivity deviation values and the unlabeled scattering characteristic deviation values are classified to determine the classification results of the reflectivity deviation values and the scattering characteristic deviation values; The random forest algorithm is used to judge whether the classification results of the reflectivity deviation value and the scattering characteristic deviation value meet the preset conditions according to the classification results of the reflectivity deviation value and the scattering characteristic deviation value, and to obtain the accuracy of the classification results of the reflectivity deviation value and the scattering characteristic deviation value.

6. The method according to claim 1, characterized in that Obtaining precipitation data and transpiration data of the target area from the meteorological database, calculating the moisture content variation trend in the time dimension based on the precipitation data and transpiration data, and obtaining a moisture content time variation curve includes: According to the geographical location information of the target area, the historical precipitation data and evaporation data of the area are obtained from the meteorological database; Preprocess the acquired precipitation data and evaporation data to remove outliers and missing values; Using the time series analysis method, the precipitation and evaporation at each time point are calculated to obtain the precipitation and evaporation data series in the time dimension; According to the precipitation and transpiration data series, the moisture content at each time point is calculated to obtain the moisture content data series in the time dimension; Conduct trend analysis on the moisture content data series and use moving average or exponential smoothing methods to determine the changing trend of moisture content; According to the changing trend of moisture content, a moisture content time variation curve is generated to intuitively display the dynamic change process of moisture content.

7. The method according to claim 1, characterized in that The corrected reflectivity and scattering characteristic values are combined with the moisture content time variation curve and input into a pre-trained machine learning model to train the correlation model between the reflectivity, scattering characteristic and moisture content. The trained moisture content prediction model includes: Obtaining raw data of reflectivity and scattering characteristics, and correcting the raw data according to a preset correction rule to obtain corrected reflectivity and scattering characteristic values; Obtaining curve data of moisture content changing over time, and correlating the curve data with the corrected reflectivity and scattering characteristic values to form a training data set; The associated model is trained according to the input and output values of the training data set to obtain a trained moisture content prediction model.

8. The method according to claim 1, characterized in that Based on the trained moisture content prediction model, the following steps are performed to complete the forest fuel moisture content detection: According to the input values of the reflectivity and scattering characteristics to be measured, the trained moisture content prediction model is used to predict the corresponding moisture content prediction value; Determine whether the predicted moisture content exceeds the preset threshold range. If so, trigger the early warning mechanism to indicate abnormal moisture content. According to the deviation between the predicted moisture content and the actual moisture content, the moisture content prediction model is retrained to continuously optimize the prediction accuracy of the model; The moisture content prediction results are associated with the time curve to generate a time variation curve of the moisture content prediction, providing data support for subsequent moisture content trend analysis.

Images