A forest dynamic growth model construction method and system

Abstract

The application discloses a kind of forest dynamic growth model construction method and system, belong to forestry digital twin technical field.The present application is by constructing real scene three-dimensional model and carrying out single fine modeling, the fine simulation and multidimensional prediction of forest growth process are realized;Forest simulation model and forest basic information are integrated using digital twin platform, extract stand structure parameters, establish arbor layer growth dynamic model, and then the dynamic simulation of shrub layer species diversity and natural regeneration density is carried out.The present application breaks through the single scale limit, integrates the coordinated evolution of arbor layer, shrub layer and natural regeneration layer for the first time, supports multidimensional prediction and visualization, and provides scientific support for forestry digital twin.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of forestry digital twinning, more particularly to a forest dynamic growth model construction method and system. BACKGROUND

[0002] At present, digital twinning technology has been applied in the fields of industry, water conservancy, smart city and the like, and through real-time monitoring of the state parameters of the twinned object, basis is provided for formulating reasonable management measures, and the intelligent development of related industries is promoted.

[0003] However, in the forestry industry, due to the combined action of the individual growth of trees and grasses and external environmental factors such as light, moisture and soil, the growth process is extremely complex, and the existing digital twinning technology is difficult to be directly applied. The current forest dynamic growth model is mostly at the single tree scale, and is modeled based on the measured data and real camera data of the growth state of the single tree in multiple seasons, and it is difficult to construct a whole dynamic growth model of the arbor layer from the stand scale. Moreover, most of them only focus on the growth dynamic change of the arbor layer of the forest, ignoring the dynamic change of the species diversity of the shrub layer and the natural regeneration of the stand, which makes it difficult to realize the dynamic simulation of different levels of the forest ecosystem, and leads to a lack of important technical support for truly realizing the function of forestry digital twinning.

[0004] Therefore, how to propose a forest dynamic growth model construction method and system to realize the fine simulation, multi-dimensional prediction and visual presentation of the forest growth process is a problem to be solved by those skilled in the art. SUMMARY

[0005] Therefore, the present application provides a forest dynamic growth model construction method and system,

[0006] In order to achieve the above purpose, the present application adopts the following technical scheme:

[0007] On the one hand, the present application discloses a forest dynamic growth model construction method, comprising the following steps:

[0008] Obtaining forest multi-source data; constructing a real scene three-dimensional model based on the forest multi-source data, and finely modeling single trees based on the real scene three-dimensional model to obtain a forest simulation model; integrating the forest simulation model and forest basic information into a digital twinning platform;

[0009] Obtaining stand structure parameters from the digital twinning platform according to the forest basic information;

[0010] Using the relationship between the productivity increment and the stand structure parameters to establish a growth dynamic model of the arbor layer at the stand scale;

[0011] calculating a stand comprehensive structure index and a shrub layer diversity comprehensive index based on the output of the arbor layer growth dynamic model, and simulating the species diversity dynamics of the shrub layer according to the stand comprehensive structure index and the shrub layer diversity comprehensive index;

[0012] simulating the natural regeneration density dynamics based on the stand structure parameters.

[0013] Preferably, the forest multi-source data comprises airborne LiDAR data, hyperspectral data, orthophoto, ground-based LiDAR data, and artificial survey data.

[0014] The stand structure parameters comprise diameter at breast height, tree height, crown width, crown length, first branch height, stand density, openness, competition index, layer index, spatial density index, angular scale, size ratio of diameter, mixed degree, cross-sectional area at breast height, stand density, herb coverage, total crown area, main crown layer crown area, and lower crown layer crown area.

[0015] Preferably, the real scene three-dimensional model is constructed based on the forest multi-source data, comprising:

[0016] Multi-angle images are acquired by tilt photography aerial flight, and ground control point data is combined to perform aerial triangulation and pricking, to complete control network adjustment and aerial triangulation precision check.

[0017] Based on the verified control point data, a modeling software is used to reconstruct the model, to generate an initial three-dimensional model of the whole stand including terrain, vegetation, and buildings.

[0018] The initial three-dimensional model of the whole stand is geometrically corrected to obtain the real scene three-dimensional model.

[0019] Preferably, based on the real scene three-dimensional model, different three-dimensional modeling methods are used for fine modeling of single trees for different scenes.

[0020] The different three-dimensional modeling methods comprise a three-dimensional modeling method based on the real scene three-dimensional model, a three-dimensional modeling method based on point cloud data, a three-dimensional modeling method based on a digital photogrammetry system, a three-dimensional modeling method based on a fitting algorithm, and a three-dimensional modeling method based on artificial intelligence.

[0021] Preferably, a stand-scale arbor layer growth dynamic model is established by using the relationship between productivity increment and the stand structure parameters, comprising:

[0022] Selecting standard trees and measuring historical radial growth information of the standard trees;

[0023] Based on the historical radial growth information, a one-element collection equation is used to calculate the historical productivity increment of the trees and to calculate the stand productivity increment.

[0024] constructing a relationship model between the increment of stand productivity and the stand structure parameters, and correcting the relationship model with the current site index level;

[0025] calculating the increment of stand productivity under other site index levels in combination with the current site index level;

[0026] calculating the increment of stand productivity under other site index levels in combination with the current site index level;

[0027] Preferably, the comprehensive structure index of the stand and the comprehensive index of the shrub layer diversity are calculated based on the output of the growth dynamic model of the tree layer, and the dynamic simulation of the species diversity of the shrub layer is performed according to the comprehensive structure index of the stand and the comprehensive index of the shrub layer diversity, including:

[0028] constructing a shrub species database;

[0029] The comprehensive structure index S of the stand and the comprehensive index R of the shrub layer diversity are calculated based on the output of the growth dynamic model of the tree layer, and the formula is as follows:

[0030] S = 0.81 * DBH + 0.91 * crown length + 0.74 * forest layer index - 0.21 * competition index + 0.37 * spatial density index + 0.1 * angle size - 0.23 * stand density;

[0031] R = 0.70 * Shannon-Wiener index + 0.35 * Margalef index + 0.94 * Pielou index + 0.93 * Simpson index;

[0032] The number of simulated shrub layer species is obtained according to the relationship between the comprehensive structure index S of the stand and the comprehensive index R of the shrub layer diversity;

[0033] Randomly extracting the same number of shrubs as the number of simulated shrub layer species in the shrub species database for simulation, and completing the simulation of the diversity of the shrub layer.

[0034] Preferably, the natural regeneration density dynamic simulation is performed based on the stand structure parameters, including:

[0035] establishing a tree regeneration species database and completing real scene modeling to form a regeneration seedling model library;

[0036] The stand regeneration density Y is calculated, and the formula is as follows:

[0037] Y = 3600 + 3982.99 * ratio of dominant species forest tree diameter size - 1160.60 / ratio of main canopy layer area and lower canopy layer area - 493.696;

[0038] After obtaining the stand regeneration density, a regeneration seedling model is randomly extracted from a regeneration seedling model library according to the number of seedlings for simulation.

[0039] Preferably, the forest dynamic growth model construction method further comprises optimizing the dynamic simulation of the arbor layer, the shrub layer and the natural regeneration density.

[0040] In another aspect, the present application also discloses a forest dynamic growth model construction system, comprising:

[0041] A forest simulation construction module is configured to obtain forest multi-source data, construct a real scene three-dimensional model based on the forest multi-source data, perform fine modeling on individual trees based on the real scene three-dimensional model, and obtain a forest simulation model; and integrate the forest simulation model and tree basic information into a digital twin platform.

[0042] A parameter extraction module is configured to obtain stand structure parameters from the digital twin platform according to the tree basic information.

[0043] An arbor layer simulation module is configured to establish a stand-scale arbor layer growth dynamic model by using the relationship between productivity growth and the stand structure parameters.

[0044] A shrub layer simulation module is configured to calculate a stand comprehensive structure index and a shrub layer diversity comprehensive index based on the output of the arbor layer growth dynamic model, and perform dynamic simulation of shrub layer species diversity according to the stand comprehensive structure index and the shrub layer diversity comprehensive index.

[0045] A natural regeneration density simulation module is configured to perform dynamic simulation of natural regeneration density based on the stand structure parameters.

[0046] A simulation improvement module is configured to optimize the dynamic simulation of the arbor layer, the shrub layer and the natural regeneration density.

[0047] Compared with the prior art, the forest dynamic growth model construction method and system provided by the present application can realize fine simulation and multi-dimensional prediction of the forest growth process by constructing a real scene three-dimensional model and performing fine modeling on individual trees; the forest simulation model and tree basic information are integrated into a digital twin platform to extract stand structure parameters, establish an arbor layer growth dynamic model, and further perform dynamic simulation of shrub layer species diversity and natural regeneration density. The present application breaks through the limitations of individual scale, integrates the coordinated evolution of the arbor layer, the shrub layer and the natural regeneration layer for the first time, supports multi-dimensional prediction and visualization, and provides scientific support for forestry digital twinning. BRIEF DESCRIPTION OF DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of the provided drawings.

[0049] Figure 1 The method flowchart provided by the present application;

[0050] Figure 2 The forest stand overall initial three-dimensional model construction flowchart provided by the present application;

[0051] Figure 3 The three-dimensional modeling method flowchart based on the real scene three-dimensional model construction result provided by the present application;

[0052] Figure 4 The three-dimensional modeling method flowchart based on point cloud data provided by the present application;

[0053] Figure 5 The three-dimensional modeling method flowchart based on the digital photogrammetry system provided by the present application. DETAILED DESCRIPTION

[0054] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0055] In one aspect, the present application discloses a forest dynamic growth model construction method, as shown in Figure 1 The method comprises the following steps:

[0056] S1. Obtain forest multi-source data; construct a real scene three-dimensional model based on the forest multi-source data, finely model single trees based on the real scene three-dimensional model, and obtain a forest simulation model; integrate the forest simulation model and forest basic information into a digital twin platform.

[0057] S11. The forest multi-source data comprises airborne laser radar data, hyperspectral data, orthophoto, ground-based laser radar data, and artificial survey data.

[0058] The airborne data is divided into laser radar data, hyperspectral data, and orthophoto, and is realized by relying on a UAV.

[0059] After the laser radar data is processed, a digital elevation model of the forest farm is obtained, and combined with the CCD image, vegetation point cloud and non-vegetation point cloud are obtained, the non-vegetation point cloud is used to reconstruct the house building, and the vegetation point cloud is used to segment single trees and extract single tree structure information.

[0060] After the hyperspectral data is processed, the spectral characteristics and vegetation index information of different tree species are mainly provided, which are used for single tree classification.

[0061] After the orthophoto is processed, more abundant texture information is provided, and the laser radar data is combined to analyze the stand information.

[0062] The ground-based laser radar is used to scan the forest at the sample plot scale, mainly to obtain crown layer information and structure information, obtain ground point cloud, and register with airborne point cloud data, which is used for construction of 1:1 three-dimensional tree species model in the later stage, and improves the accuracy of branch reconstruction.

[0063] The ground artificial investigation data includes single tree position, tree height and tree species information, and field photos, which provide verification data for single tree segmentation and single tree species classification results.

[0064] S12. Constructing a real three-dimensional model based on forest multi-source data, including:

[0065] Obtain multi-angle images by tilt photography aerial flight, and perform aerial triangulation and pricking points combined with ground control point data to complete control network adjustment and aerial triangulation precision check; based on the verified control point data, use modeling software to reconstruct the model to generate a whole initial three-dimensional model of the stand containing terrain, vegetation, buildings and other ground objects, for details, refer to Figure 2 .

[0066] The whole initial three-dimensional model of the stand established based on the modeling software has the problems of building deformation (texture pull flower, structure distortion, broken and missing parts, etc.) caused by incorrect image matching or poor geometric posture, suspended objects, missing parts, etc. In this embodiment, the whole initial three-dimensional model of the stand is geometrically corrected by using 3dsMax and similar software to obtain a real three-dimensional model.

[0067] S13. Based on the real three-dimensional model, different three-dimensional modeling methods are used for fine modeling of single trees according to different scenes. The fine modeling method of single trees is as follows:

[0068] (1) The three-dimensional modeling method based on the real three-dimensional model construction result is as follows Figure 3As shown, the single-body modeling is completed through several links such as local separation editing, fine revision and reconstruction, and update merging, based on the reconstructed real three-dimensional model as a basic reference combined with ground shooting images. This modeling method fully utilizes the three-dimensional model results, and the operating personnel can complete single-body modeling in the real world based on the three-dimensional model restoration without the aid of auxiliary equipment.

[0069] (2) The three-dimensional modeling method based on point cloud data is as shown in Figure 4

[0070] 1) Obtain oblique data and perform preprocessing, denoising, distortion correction and other processing on oblique images to improve image quality.

[0071] Perform image dodging and color uniformity processing to ensure consistent image tone and facilitate subsequent texture mapping.

[0072] Obtain the position and attitude information of the image through the POS system carried by the unmanned aerial vehicle.

[0073] 2) Obtain laser point cloud data and perform preprocessing

[0074] Use a ground laser scanner (such as LiDAR) or an airborne laser scanner to obtain high-precision three-dimensional point cloud data. Laser scanning obtains the three-dimensional coordinates of ground objects through ranging principles, and the point cloud density is usually several dozen to several hundred points per square meter.

[0075] Perform denoising, filtering, classification and other processing on the original point cloud data to remove invalid points (such as vegetation, vehicles, etc.). Register the point cloud data with the oblique image data to ensure spatial consistency.

[0076] Edit the point cloud data, repair missing areas, and optimize the point cloud quality.

[0077] Generate an irregular triangular mesh model (TIN) based on the point cloud data to construct the geometric framework of the ground surface and buildings.

[0078] Generate a digital surface model (DSM) based on the point cloud data to reflect the elevation information of the ground surface.

[0079] 3) Single three-dimensional model solution

[0080] Combine oblique images and laser point cloud data, and use three-dimensional modeling software (such as ContextCapture, AgisoftMetashape) to reconstruct the model. Generate a three-dimensional model with real textures. Perform denoising, simplification, repair and other operations on the model to improve the model quality. Generate models with different levels of detail (such as LOD1, LOD2, LOD3) according to requirements.

[0081] 4) Ground three-dimensional model solution ​

[0082] A digital terrain model (DTM) is generated based on laser point cloud data, reflecting the elevation information of the ground surface.

[0083] The ground surface model is fused with the building model to generate a complete three-dimensional scene.

[0084] 5) Model data modification and improvement

[0085] The generated three-dimensional model is manually edited to repair defects and optimize the geometric structure. The model texture is optimized to ensure clear and non-distorted texture.

[0086] The geometric accuracy of the model is verified using ground control points (GCPs) and check points (Check Points).

[0087] This method uses high-precision laser point cloud as the data source for the model Mesh net, uses laser scanning of ground objects, and obtains high-precision three-dimensional point cloud data through laser ranging principles. On this basis, a triangular mesh model is constructed. Studies have shown that when the point cloud density is less than 10 points per square meter, the modeling precision is highest at LOD2. Therefore, if more detailed three-dimensional results are required, the point cloud density needs to be increased, for example, the super-fine modeling point cloud density needs to reach more than 200 points per square meter.

[0088] (3) Three-dimensional modeling method based on digital photogrammetry system

[0089] The digital photogrammetry system is one of the essential equipment for large-scale topographic map production. The theoretical core of digital photogrammetry is the collinearity equation. The so-called collinearity equation is that the photographic center, the image point coordinates (x, y) of the image, and the corresponding ground point are on the same straight line. Based on the principle of collinearity equation, the necessary point coordinates of the model can be measured in a stereo environment, and a three-dimensional stereo model can be constructed, as shown in the technical route Figure 5 The key step is to collect the feature lines of ground objects in a stereo environment and complete the construction of the ground object white model in combination with the fine digital elevation model.

[0090] The collection content mainly includes the contour lines of buildings, road boundary lines, water system boundaries, and terrain feature lines (such as ridge lines and valley lines). The feature lines of ground objects are superimposed with DEM in a three-dimensional environment to ensure that the elevation information of the feature lines is consistent with DEM. The feature lines are corrected in elevation to eliminate errors. The contours are extracted from the feature lines, and the height, width, and local refinement are assigned according to the DEM or additional height information, road grade or measured data, etc. Finally, the fine digital elevation model is fused to generate the ground object white model.

[0091] (4) Three-dimensional modeling method based on fitting algorithm

[0092] Based on point cloud data, combined with model library, automatic three-dimensional modeling of model is completed. First, a forest tree model library is constructed; then, taking the model library as the basic unit, the conversion parameters of spatial transformation (translation, rotation or scaling) between models are obtained by automatically extracting or manually collecting the feature points of the three-dimensional model construction method of forest trees, and finally the modeling is completed.

[0093] (5) Three-dimensional modeling method based on artificial intelligence

[0094] Based on three-dimensional tree point cloud, dense point cloud data covering all elements is formed, and based on this, tree trunk, branch and leaf point clouds are obtained. Through artificial intelligence (AI) technology, combined with field photos, the bark material corresponding to the field is given to generate tree trunks, and after the corresponding leaves are pasted, a 1:1 restored three-dimensional tree species model is made, and single tree reconstruction is completed.

[0095] The forest simulation model not only contains the spatial position distribution information of each forest tree, but also includes the basic information of the forest trees, and these basic information can be integrated into the digital twin platform.

[0096] S2. Obtain stand structure parameters from the digital twin platform according to the basic information of the forest trees.

[0097] Based on the basic information of the forest trees, the data processing and analysis software (stand spatial structure parameter calculation formula) of the digital twin platform is used to obtain various stand structure parameters related to forest dynamic growth, including diameter at breast height, tree height, crown width, crown length, first branch height, stand density, openness, competition index, layer index, spatial density index, angular scale, diameter size ratio, mixed degree, breast height cross-sectional area, stand density, herb coverage, total crown area, main crown layer crown area, and lower crown layer crown area.

[0098] Further, by consulting materials, the stand age information is obtained, combined with the tree height parameter, and the platform is used for comprehensive analysis to obtain the site index level of different plots as the correction coefficient of the dynamic growth model. The correction coefficient of the site index level of 14 is 1, and the correction coefficient increases (decreases) by 0.2 for each increase (decrease) of 1 level.

[0099] S3. Use the relationship between productivity growth and stand structure parameters to establish a growth dynamic model of the arbor layer (forest trees with a diameter at breast height greater than or equal to 5 cm) at the stand scale.

[0100] In the target stand, 8-12 standard trees of different tree species are selected respectively, and the tree core samples are taken by growth cone, pretreated, and cross-dated using dendrochronology instrument to obtain the radial growth information of the trees in the past 5 years; the productivity growth amount of the trees in the past 5 years is calculated by using a unary collection equation; the productivity growth amount of the stand is calculated; the relationship model of the parameters and the productivity growth amount of the stand is established in combination with the parameter information in S2; based on the relationship model, the productivity increment of the stand in the next 5 years under the conditions of other site index levels is calculated in combination with the current site index level; the growth amount of each tree and the number of stand regeneration are inversely calculated based on the productivity increment in the next 5 years; the growth information of the stand regeneration seedlings is obtained based on S1, and the trees with a diameter at breast height of less than 5 cm after 5 years are randomly simulated into the simulation model.

[0101] S4. Calculate the stand comprehensive structure index and the shrub layer diversity comprehensive index based on the output of the arbor layer growth dynamic model, and simulate the species diversity of the shrub layer based on the stand comprehensive structure index and the shrub layer diversity comprehensive index.

[0102] S41. Investigate all shrub species growing naturally in the local area, establish a local shrub species database, and complete real scene modeling to form a shrub species model library.

[0103] S42. Based on S2 and S3, obtain the stand structure parameters after 5 years, and construct a stand comprehensive structure index factor (S) = 0.81*diameter at breast height + 0.91*crown length + 0.74*forest layer index - 0.21*competition index + 0.37*space density index + 0.1*angle size - 0.23*stand density.

[0104] Construct a shrub layer diversity comprehensive index (R) = 0.70*Shannon-Wiener index + 0.35*Margalef index + 0.94*Pielou index + 0.93*Simpson index.

[0105] S43. Establish the relationship between S and R: R = 0.82*S, calculate R from S; then calculate the Margalef index from R to obtain the number of shrub species simulated after 5 years.

[0106] S44. After obtaining the number of shrub species, randomly select the corresponding number of shrubs from the shrub species model library obtained in S41 for simulation to complete the diversity simulation of the shrub layer.

[0107] S5. Dynamic simulation of natural regeneration density based on stand structure parameters, including:

[0108] S51. Investigate all arbor species growing naturally in the local area, establish a database of local arbor regeneration species, and complete real scene modeling to form a regeneration seedling model library.

[0109] S52. Natural regeneration density (Y) in the next 5 years = 3600 + 3982.99 * size ratio of dominant forest tree diameter - 1160.60 / ratio of main canopy layer area to lower canopy layer area - 493.696.

[0110] S53. After obtaining the stand regeneration density, a regeneration seedling model is randomly selected from the regeneration seedling model library according to the number of seedlings for simulation.

[0111] S6. Optimize the dynamic simulation of the arbor layer, shrub layer and natural regeneration density.

[0112] The optimization point of the arbor layer dynamic growth simulation is the number of regeneration seedlings entering the arbor layer and growth information.

[0113] By dynamically adjusting the proportion of regeneration seedlings entering the arbor layer (such as optimizing the survival rate of seedlings with a diameter of less than 5 cm from 40% to a dynamic threshold based on site conditions), introducing a growth model (such as correcting the diameter and height growth curve of regeneration seedlings based on environmental factors), combining monitoring data (such as unmanned aerial vehicle remote sensing and manual investigation) fed back in real time by the digital twin platform, and using the random forest algorithm to optimize the site index correction coefficient, the adaptability of arbor layer growth prediction is improved.

[0114] The optimization point of the dynamic simulation of shrub layer species diversity is the height and ground diameter of shrubs.

[0115] In the simulation of shrub layer diversity, environmental factors such as light intensity and soil moisture are integrated, the growth parameters of shrub height and ground diameter are dynamically corrected, the spatial distribution of shrub species is optimized through a species competition model (such as adjusting the competition weight in the S index), and the coefficient of the diversity comprehensive index (R = 0.82 * S) is calibrated combined with field investigation data, to improve the ecological rationality of species diversity prediction.

[0116] The optimization point of the dynamic simulation of natural regeneration density is the growth information of regeneration seedlings.

[0117] For the natural regeneration density formula, the regeneration coefficient is analyzed by regression analysis of historical data, and a dynamic model of seedling survival rate is introduced (the survival rate threshold is adjusted according to climate data), combined with the difference between simulation results and actual observation, the gradient boosting tree algorithm is used to iteratively optimize the formula parameters, to enhance the accuracy and environmental adaptability of regeneration density prediction.

[0118] On the other hand, the present application also discloses a forest dynamic growth model construction system, comprising:

[0119] The forest simulation construction module is configured to acquire forest multi-source data, construct a real scene three-dimensional model based on the forest multi-source data, and obtain a forest simulation model by performing fine modeling on single trees based on the real scene three-dimensional model; and the forest simulation model and forest basic information are integrated into a digital twin platform.

[0120] The parameter extraction module is configured to acquire stand structure parameters from the digital twin platform according to the forest basic information.

[0121] The arbor layer simulation module is configured to establish a stand-scale arbor layer growth dynamic model by using the relationship between productivity growth and stand structure parameters.

[0122] The shrub layer simulation module is configured to calculate a stand comprehensive structure index and a shrub layer diversity comprehensive index based on the output of the arbor layer growth dynamic model, and perform dynamic simulation of shrub layer species diversity according to the stand comprehensive structure index and the shrub layer diversity comprehensive index.

[0123] The natural regeneration density simulation module is configured to perform dynamic simulation of natural regeneration density based on the stand structure parameters.

[0124] The simulation improvement module is configured to optimize the dynamic simulation of the arbor layer, the shrub layer, and the natural regeneration density.

[0125] The embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments, and the same or similar parts between the embodiments can be referred to each other. For the device disclosed by the embodiments, since it corresponds to the method disclosed by the embodiments, the description is relatively simple, and the related parts can be referred to the method part.

[0126] The above description of the disclosed embodiments enables a person skilled in the art to implement or use the present application. Various modifications to the embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for constructing a forest dynamic growth model, characterized in that, Includes the following steps: Acquire multi-source forest data; construct a realistic 3D model based on the multi-source forest data; refine the model of individual trees based on the realistic 3D model to obtain a forest simulation model; integrate the forest simulation model and basic forest information into a digital twin platform; the multi-source forest data includes airborne lidar data, hyperspectral data, orthophotos, ground-based lidar data, and manual survey data; Based on the basic forest information, forest stand structure parameters are obtained from the digital twin platform; the forest stand structure parameters include diameter at breast height (DBH), tree height, crown width, crown length, height below the first branch, as well as stand density, openness, competition index, forest layer index, spatial density index, angular scale, DBH ratio, mixed degree, DBH sectional area, stand density, herbaceous cover, total crown area, main crown area, and lower crown area; By utilizing the relationship between productivity growth and the aforementioned stand structure parameters, a dynamic model of tree layer growth at the stand scale is established, including: Select a standard log and measure its historical radial growth information; Based on the historical radial growth information, the historical productivity growth of forest trees was calculated using a univariate acquisition equation, and the productivity growth of forest stands was extrapolated. A relationship model is constructed between the forest stand productivity growth and the forest stand structure parameters, and the relationship model is corrected using the current site index. Based on the current site index level, estimate the stand productivity growth under other site index levels; Based on the increase in forest stand productivity, the growth of diameter at breast height (DBH) and tree height of each tree is calculated in reverse. Combined with the basic information of the trees, the number and distribution of new seedlings entering the canopy layer are simulated to complete the dynamic growth modeling of the canopy layer at the stand scale. Based on the output of the tree layer growth dynamic model, the comprehensive stand structure index and the comprehensive shrub layer diversity index are calculated. Then, based on these indices, a dynamic simulation of shrub layer species diversity is performed, including: Construct a database of shrub species; Based on the tree layer growth dynamic model, the predicted stand structure parameters are output, and the comprehensive stand structure index S and the comprehensive shrub layer diversity index R are calculated using the following formulas: S = 0.81 * Diameter at breast height (DBH) + 0.91 * Crown length + 0.74 * Forest layer index - 0.21 * Competition index + 0.37 * Spatial density index + 0.1 * Angle scale - 0.23 * Stand density; R = 0.70 * Shannon-Wiener index + 0.35 * Margalef index + 0.94 * Pielou index + 0.93 * Simpson index; The simulated number of shrub species was obtained based on the relationship between the stand comprehensive structure index S and the shrub layer diversity comprehensive index R. Shrubs with the same number of species as the simulated shrub layer are randomly selected from the shrub species database for simulation to complete the shrub layer diversity simulation. Dynamic simulation of natural regeneration density was performed based on the aforementioned forest stand structure parameters.

2. The method for constructing a forest dynamic growth model according to claim 1, characterized in that, Constructing a realistic 3D model based on the aforementioned multi-source forest data includes: By acquiring multi-angle images through oblique photogrammetry, and combining them with ground control point data, aerial triangulation and point-pointing were performed to complete the control network adjustment and aerial triangulation accuracy check. Based on the verified control point data, modeling software is used to reconstruct the model and generate an initial three-dimensional model of the forest stand that includes topography, vegetation, and building features. The geometric structure of the initial three-dimensional model of the forest stand is corrected to obtain a real-world three-dimensional model.

3. The method for constructing a forest dynamic growth model according to claim 1, characterized in that, Based on the aforementioned real-world 3D model, different 3D modeling methods are used to refine the modeling of individual trees for different scenarios; The different 3D modeling methods include: 3D modeling methods based on the construction results of real-scene 3D models, 3D modeling methods based on point cloud data, 3D modeling methods based on digital photogrammetry systems, 3D modeling methods based on fitting algorithms, and 3D modeling methods based on artificial intelligence.

4. The method for constructing a forest dynamic growth model according to claim 1, characterized in that, Dynamic simulation of natural regeneration density based on the aforementioned stand structure parameters includes: Establish a database of updated tree species and complete real-world modeling to create a database of updated seedling models; The formula for calculating forest stand regeneration density Y is as follows: Y = 3600 + 3982.99 * ratio of diameter at breast height (DBH) of dominant tree species - 1160.60 / ratio of main canopy area to lower canopy area - 493.696; After obtaining the forest stand regeneration density, a regeneration seedling model is randomly selected from the seedling model library according to the number of seedlings for simulation.

5. The method for constructing a forest dynamic growth model according to claim 1, characterized in that, It also includes optimizing the dynamic simulation of tree layer, shrub layer, and natural regeneration density.

6. A forest dynamic growth model construction system, used to implement the forest dynamic growth model construction method as described in any one of claims 1-5, characterized in that, include: The forest simulation building module is used to acquire multi-source forest data; Based on the multi-source forest data, a real-scene 3D model is constructed. On the basis of the real-scene 3D model, a detailed model of individual trees is performed to obtain a forest simulation model. The forest simulation model and basic forest information are then integrated into a digital twin platform. The parameter extraction module is used to obtain forest stand structure parameters from the digital twin platform based on the basic forest information; The tree layer simulation module is used to establish a dynamic model of tree layer growth at the stand scale by utilizing the relationship between productivity growth and the stand structure parameters. The shrub layer simulation module is used to calculate the comprehensive stand structure index and the comprehensive shrub layer diversity index based on the output of the tree layer growth dynamic model, and to perform dynamic simulation of shrub layer species diversity based on the comprehensive stand structure index and the comprehensive shrub layer diversity index. The natural regeneration density simulation module is used to perform dynamic simulation of natural regeneration density based on the forest stand structure parameters.

7. A forest dynamic growth model construction system according to claim 6, characterized in that, It also includes a simulation improvement module for optimizing the dynamic simulation of tree layer, shrub layer, and natural regeneration density.

Images