Arbor load estimation method and system based on different growth coefficients

Abstract

The invention relates to the technical field of arbor load evaluation, and discloses an arbor load estimation method and system based on different growth coefficients, and the method comprises the steps: generating an arbor growth database, collecting arbor growth parameters, extracting arbor planting point coordinates, generating a development tree, and analyzing the growth difference between species. Dividing the arbor growth data set into a training set, a verification set and a test set; establishing an arbor load estimation model based on the arbor growth database, and training the arbor load estimation model by using the training set; estimating biomass accumulation of the arbor in a preset growth period, and calculating a total load in combination with the weight of the planting soil and a moisture saturation state; and evaluating the influence of the growth coefficient and the environmental variable on load prediction. The system comprises an arbor data acquisition module, an arbor load evaluation module and an evaluation result output module. The system can reflect the growth change of the arbor, so as to adjust the monitoring and maintenance strategy in time.

Description

Technical Field

[0001] The present invention relates to the technical field of tree load assessment, and in particular to a tree load estimation method and system based on different growth coefficients. Background Art

[0002] Tree load estimation relies on the growth characteristics of trees, indirectly estimating their weight by measuring and analyzing parameters such as diameter at breast height, tree height, and crown width. This method is widely used in urban green space planning, landscape design and construction, construction engineering, forestry engineering, and ecological research. Current techniques often rely on empirical evidence, without relevant calculation methods or software. While this method is simple and easy to implement, its accuracy is significantly affected by data quality and experience. Tree load estimation is a crucial component of construction engineering, forestry engineering, and ecological research, involving issues such as tree weight and wind resistance. Traditional methods often rely on empirical or measured data, which have several significant drawbacks. First, empirical theories are limited and subjective, lacking objective data support and calculation methods. Second, different tree species vary in trunk bulk density, crown morphology, and density, directly impacting the accuracy of load estimates. Third, trees of the same species in different growth stages (juvenile, mature, and aging) can experience significant variations in loads.

[0003] Prior art one, a Chinese patent with patent number: 202411131941.9, provides a water resource scheduling optimization method based on an empirical model, relating to the field of irrigation water resource allocation technology. By collecting and analyzing historical irrigation data for each growth stage, combined with farmland spectral reflectance data and environmental data, a crop growth coefficient model is constructed to determine the adjustment coefficients of crops at different growth stages. By collecting the type and growth stage of the crop currently to be irrigated, the adjustment coefficient is determined using historical irrigation data, and the optimal adjustment coefficient corresponding to the crop growth stage and type is matched. By collecting farmland spectral reflectance data and environmental data in real time, the current crop adjustment coefficient is constructed, and the crop water requirement is generated by combining the optimal adjustment coefficient. Finally, by setting irrigation priorities and combining the optimal irrigation amount, irrigation water resources are allocated to each farmland within the irrigation area. Although this ensures the rational allocation of water resources among different farmlands and optimizes the overall water resource allocation of the irrigation area, relying on historical irrigation data to calculate the adjustment coefficient leads to large load estimation errors due to environmental variables.

[0004] Prior art 2, a Chinese patent, patent number: 202410023505.3 discloses a method for identifying key factors in forest biomass estimation based on multimodal data fusion, which relates to the field of biomass estimation technology. If the biomass change exceeds the expected change area and the proportion of the sub-area exceeds a proportion threshold, a biomass prediction model is trained and obtained. If the biomass prediction model is reliable, a K-fold cross-validation is performed on the biomass prediction model, and an evaluation index set is constructed based on the verification results. The model performance coefficient is generated from the evaluation index set. If the model performance coefficient exceeds the quality threshold, the target feature is screened from the biomass prediction model. After analyzing and processing the target feature, the corresponding eigenvalue sequence is obtained. The principal component is determined from the eigenvalue sequence and the dimension is reduced. The corresponding part is screened from the principal component after the dimension reduction according to the contribution as the key factor. Although the scope of key factor identification is gradually narrowed, the efficiency of identification is improved, and the selected key factors are more reliable, the dimension of biomass estimation is single when the key factors are screened through a posteriori cross-validation.

[0005] Prior art three, a Chinese patent with patent number 202410873044.9, relates to a method and system for identifying seedling species. The seedling species identification method involves first collecting odor samples and environmental parameter data from seedlings of different plant species at various growth stages; then separating and analyzing the odor samples to generate an odor data stream. Key compound signatures are extracted to form odor signatures for each growth stage; then, combining the odor signatures with the environmental parameter data, a dynamic prediction model is established and trained to predict the seedling growth stage. Simultaneously, a plant species identification model is established and trained based on the odor signatures and environmental parameter data for each growth stage; finally, plant morphological characteristics are analyzed using computer vision data, and the odor signatures are integrated with the visual data to optimize and adjust the plant species identification results. While the accuracy and efficiency of plant species identification are improved by collecting odor and environmental data and combining advanced analytical methods with odor analysis and visual recognition, the reliance on indirect features such as odor samples to correct unstructured data results in significant errors.

[0006] Currently, existing technologies 1, 2, and 3 have the problem that traditional load estimation relies on static coefficient tables, resulting in poor accuracy in tree load estimation results. To solve the above problem, the present invention provides a tree load estimation method and system based on different growth coefficients. Summary of the Invention

[0007] The main purpose of the present invention is to provide a tree load estimation method and system based on different growth coefficients to solve the problem in the prior art that traditional load estimation relies on static coefficient tables and the tree load estimation results are less accurate.

[0008] To achieve the above object, the present invention provides the following technical solutions: A method for estimating tree loads based on different growth coefficients includes: generating a tree growth database, collecting tree growth parameters, extracting tree planting point coordinates, estimating the biomass accumulation of the tree within a preset growth cycle, and calculating the total load in combination with the weight of the planting soil and the water saturation state; and evaluating the impact of the growth coefficient and environmental variables on load prediction.

[0009] As a further improvement of the present invention, the tree growth parameters are preprocessed to obtain a tree growth data set, and the tree growth data set is divided into a training set, a validation set, and a test set.

[0010] As a further improvement of the present invention, the process of analyzing growth differences between species includes the following steps: Generate a tree growth database and collect growth parameters of different tree species; tree species growth parameters include biomass, physiological indicators, and environmental impact data; clean the collected heterogeneous data and perform normalization to eliminate dimensional differences; Perform differential tests on growth parameters of different species to identify species groups with differences; determine whether there are differences in leaf area index between different species under the same environment; extract features of tree growth parameters and group species according to growth parameter characteristics; Gene sequence data were collected, multiple sequence comparisons were performed, and developmental trees were constructed to analyze the correlation between evolutionary differences and phenotypic differences. The tree growth dataset was formed and divided into training set, validation set and test set.

[0011] As a further improvement of the present invention, the process of grouping species according to growth parameter characteristics comprises the following steps: The observational data comes from multiple unobserved tree data sets, each with different dynamic tree parameters. The tree data parameters for each set are inferred, the optimal number of groupings is automatically inferred, the probability of each species belonging to a different group is calculated, and the calculated species are assigned to the group with the highest probability. Extract features from the dynamic tree parameters corresponding to each group to obtain differences in species growth patterns; test whether there are differences in growth performance between different groups under the same environment. If differences are found, regroup them; If it is judged that there is no difference, the grouping results will be integrated to form the tree growth data set.

[0012] As a further improvement of the present invention, a tree load estimation model is established based on the tree growth database, and the tree load estimation model is trained using a training set; the biomass accumulation of the tree within a preset growth cycle is estimated, and the total load is calculated in combination with the planting soil weight and water saturation state.

[0013] As a further improvement of the present invention, the process of calculating the total load in combination with the weight of the planting soil and the water saturation state includes the following steps: A tree load estimation model is established based on the tree growth database. The training set of the tree growth data set is divided into two parts. The tree growth data is divided into k folds, and the tree load estimation model is trained with k-1 folds each time. Among them, each tree sample has a corresponding prediction result; The tree growth patterns are learned in different subsets and features to generate preliminary prediction results. After the first training is completed, the tree load estimation model generates predictions for the validation set and the retained subset data set. The predicted results are used as input features for the second training, and the predicted values ​​are used to form a new feature matrix; the predicted results and true labels of the first trained tree load estimation model are used to train the second tree load estimation model, and the output of the first trained tree load estimation model is learned to be weighted or combined.

[0014] As a further improvement of the present invention, the process of dividing the training set of the tree growth data set into two parts includes the following steps: The training set in the tree growth dataset is equally divided into k subsets. If the total number of training samples is N, each fold contains approximately N / k subsets. K cycles of training are performed, and k01 subsets are selected and merged into the training data for the current iteration in each cycle. The first cycle uses the 2nd to kth folds as the training set, the second cycle uses the 1st, 3rd to kth folds, and so on; the remaining subset is used as the validation set, and the tree load estimation model is trained on the k-1th fold training data; After each training is completed, the current tree load estimation model is used to predict the retained validation set subset to obtain the predicted value of the subset sample; the prediction results of all k validation sets are integrated to form a set of predicted values ​​corresponding to the original training set samples.

[0015] As a further improvement of the present invention, the process of generating the preliminary prediction result includes the following steps: After training is completed, the validation set is used to predict the 1-fold validation set retained in the current round; predictions are made on the independently retained test set and data that did not participate in the current round of training; The prediction results are recorded as preliminary prediction values, and all samples are used as validation sets for prediction. For the independently retained sub-datasets, the tree load estimation model makes unified predictions after each round of training and integrates multiple prediction results. Adjust the input data in different training rounds, learn the rules under different data distributions and generate diverse prediction values; integrate the prediction values ​​into a new feature matrix as the input features of the second training.

[0016] As a further improvement of the present invention, the influence of the growth coefficient and environmental variables on the load prediction is evaluated. If the evaluation result is greater than a preset threshold, the tree load estimation model is adjusted; if the evaluation result is less than the threshold, the tree load result is output to the cloud.

[0017] To achieve the above object, the present invention also provides the following technical solutions: A tree load estimation system based on different growth coefficients is applied to the tree load estimation method based on different growth coefficients. The tree load estimation system based on different growth coefficients includes: The tree data acquisition module is used to generate a tree growth database, collect tree growth parameters, extract tree planting point coordinates, generate a developmental tree, and analyze growth differences between species; pre-process the tree growth parameters to obtain a tree growth dataset, and divide the tree growth dataset into a training set, a validation set, and a test set; The tree load assessment module is used to establish a tree load estimation model based on the tree growth database and train the tree load estimation model using a training set; it estimates the biomass accumulation of trees within a preset growth cycle and calculates the total load based on the weight of the planting soil and the water saturation state; The evaluation result output module is used to evaluate the impact of growth coefficients and environmental variables on load prediction. If the evaluation result is greater than the preset threshold, the tree load estimation model is adjusted; if the evaluation result is less than the threshold, the tree load result is output to the cloud.

[0018] The present invention facilitates long-term monitoring and maintenance. The estimation method based on different growth coefficients can reflect changes in tree growth, allowing for timely adjustments to monitoring and maintenance strategies. This helps ensure the healthy growth and safety of trees while extending their service life. The tree load estimation method based on different growth coefficients offers significant benefits in improving estimation precision and accuracy, optimizing resource utilization efficiency, and enabling long-term monitoring and maintenance. These advantages make this method promising for broad application in urban green space planning, landscape design and construction, construction engineering, forestry engineering, and other fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 1. A schematic flow chart of steps of an embodiment of a method for estimating tree loads based on different growth coefficients according to the present invention; Figure 2 This is a schematic flow chart of steps for analyzing growth differences between species in an embodiment of a method for estimating tree loads based on different growth coefficients according to the present invention; Figure 3 This is a schematic flow chart of the steps of grouping species according to growth parameter characteristics in one embodiment of a method for estimating tree loads based on different growth coefficients of the present invention; Figure 4This is a flowchart of an embodiment of a method for estimating tree loads based on different growth coefficients of the present invention, and a step of calculating the total load in combination with the weight of the planting soil and the water saturation state; Figure 5 This is a flow chart of steps for dividing a training set of a tree growth data set into two parts according to an embodiment of the tree load estimation method based on different growth coefficients of the present invention; Figure 6 A schematic flow chart of steps for generating preliminary prediction results according to an embodiment of a method for estimating tree loads based on different growth coefficients of the present invention; Figure 7 This is a flow chart of steps for integrating predicted values ​​into a new feature matrix according to an embodiment of a method for estimating tree loads based on different growth coefficients of the present invention; Figure 8 1. A schematic flow chart of steps for learning to weight or combine outputs of a first trained tree load estimation model according to an embodiment of a tree load estimation method based on different growth coefficients of the present invention; Figure 9 This is a schematic flow chart of steps for evaluating the influence of growth coefficients and environmental variables on load prediction in an embodiment of a method for estimating tree loads based on different growth coefficients of the present invention; Figure 10 This is a functional module diagram of an embodiment of a tree load estimation system based on different growth coefficients of the present invention; Figure 11 This is a schematic structural diagram of an embodiment of an electronic device of the present invention; Figure 12 This is a schematic structural diagram of an embodiment of a storage medium of the present invention. DETAILED DESCRIPTION

[0020] 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. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0021] The terms "first," "second," and "third" in this disclosure are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features indicated. Therefore, features specified as "first," "second," or "third" may explicitly or implicitly include at least one of such features. In the description of this disclosure, "plurality" means at least two, for example, two, three, etc., unless otherwise specifically defined. All directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of this disclosure are intended only to illustrate the relative positional relationships and movement of components in a specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indications will also change accordingly. Furthermore, the terms "including," "having," and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or elements is not limited to the listed steps or elements and may optionally include steps or elements not listed, or may optionally include other steps or elements inherent to such process, method, product, or apparatus.

[0022] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0023] like Figure 1 As shown, this embodiment provides an embodiment of a method for estimating tree loads based on different growth coefficients. In this embodiment, the method for estimating tree loads based on different growth coefficients specifically includes the following steps: Step S1: Generate a tree growth database, collect tree growth parameters, extract tree planting point coordinates, generate a developmental tree, and analyze growth differences between species; preprocess the tree growth parameters to obtain a tree growth dataset, and divide the tree growth dataset into a training set, a validation set, and a test set; Step S2: Establishing a tree load estimation model based on the tree growth database, and using the training set to train the tree load estimation model; estimating the biomass accumulation of the tree within a preset growth period, and calculating the total load in combination with the weight of the planting soil and the water saturation state; Step S3: Evaluate the impact of growth coefficients and environmental variables on load prediction. If the evaluation result is greater than a preset threshold, adjust the tree load estimation model; if the evaluation result is less than the threshold, output the tree load result to the cloud.

[0024] Preferably, this embodiment generates a tree growth database and collects growth parameters, combines coordinate extraction with developmental tree generation, analyzes growth differences between species, and provides comprehensive basic data support for subsequent model establishment. At the same time, the data is preprocessed and divided into training sets, validation sets, and test sets to ensure the scientific nature and effectiveness of model training; a load estimation model is constructed based on the tree growth database, and the model parameters are optimized through the training set to achieve accurate estimation of tree biomass accumulation and total load, thereby providing a reliable basis for agricultural planning and resource management; by evaluating the impact of growth coefficients and environmental variables on load prediction, the model is dynamically adjusted to improve prediction accuracy. If the evaluation result does not reach the threshold, the load result is output to the cloud for further analysis and application, reflecting the flexibility and practicality of the model. Tree load estimation is usually based on certain mathematical models and assumptions, such as classifying tree species, treating tree trunks as cylinders for volume calculation, and estimating crown weight through branches and leaves and gaps, etc., to improve estimation accuracy. Factors such as the tree's growth environment, soil conditions, and water availability all affect its growth rate and size. Load estimation based on different regions, climate conditions, and growth coefficients can fully account for these environmental factors, ensuring the accuracy and reliability of the estimated results. In landscape design and construction, appropriate tree load estimation can optimize resource efficiency and avoid over- or under-design. Over-design wastes resources, while under-design can lead to safety hazards. Load estimation based on different growth coefficients allows for rational design based on the tree's actual growth conditions, thereby improving resource efficiency and reducing construction and maintenance costs. It also facilitates long-term monitoring and maintenance. Estimation methods based on different growth coefficients can reflect changes in tree growth, allowing for timely adjustments to monitoring and maintenance strategies. This helps ensure the healthy growth and safety of trees while extending their service life. Tree load estimation methods based on different growth coefficients offer significant benefits in terms of improved estimation precision and accuracy, optimized resource efficiency, and long-term monitoring and maintenance. These advantages make this method promising for broad application in urban green space planning, landscape design and construction, construction engineering, forestry engineering, and other fields.

[0025] Furthermore, if Figure 2 As shown, the process of analyzing growth differences between species in step S1 specifically includes the following steps: Step S11: Generate a tree growth database and collect growth parameters of different tree species; wherein the growth parameters of tree species include biomass, physiological indicators, and environmental impact data; clean the collected heterogeneous data and perform normalization processing to eliminate dimensional differences; Step S12: performing a difference test on the growth parameters of different species to identify species groups with differences; determining whether there are differences in the leaf area index of different species under the same environment; performing feature extraction on the tree growth parameters and grouping the species according to the growth parameter characteristics; Step S13: Gather gene sequence data, perform multiple sequence comparison, construct a developmental tree to analyze the correlation between evolutionary differences and phenotypic differences, form an arbor growth dataset, and divide the arbor growth dataset into a training set, a validation set, and a test set.

[0026] Preferably, this embodiment generates a tree growth database and collects growth parameters of different tree species, including biomass, physiological indicators and environmental impact data. Through data cleaning and normalization, the dimensional differences of heterogeneous data are eliminated to ensure data quality and consistency. A difference test is performed on the growth parameters of different species to identify species groups with significant differences; at the same time, it is determined whether there are differences in the leaf area index of different species under the same environment, and feature extraction is performed on the tree growth parameters, and species are grouped according to the characteristics of growth parameters. Gene sequence data are collected, multiple sequence comparisons are performed, and a developmental tree is constructed to analyze the correlation between evolutionary differences and phenotypic differences to form a tree growth data set, which is divided into a training set, a validation set and a test set. Through difference testing and feature extraction, the differences in growth characteristics between different species are revealed, providing a basis for classification and evolutionary analysis; the evolutionary relationship of species is revealed through phylogenetic tree analysis, and a comprehensive tree growth data set is constructed in combination with phenotypic data to provide reliable data support for model training and prediction.

[0027] Furthermore, if Figure 3 As shown, the process of grouping species according to growth parameter characteristics in step S12 specifically includes the following steps: Step S121: Observation data comes from multiple unobserved tree data groups, each group having different dynamic tree parameters; each group of tree data parameters is inferred, the optimal number of groupings is automatically inferred, the probability of each group of species belonging to different groups is calculated, and the calculated species are assigned to the group with the highest probability; Step S122: extracting features from the dynamic tree parameters corresponding to each group to obtain differences in growth patterns that characterize the species; testing whether there are differences in growth performance between different groups under the same environment; and regrouping if differences are found. Step S123: If it is determined that there is no difference, the grouping results are integrated to form a tree growth data set.

[0028] Optimally, this embodiment achieves automated grouping optimization by estimating the optimal number of groups for different dynamic tree parameters, calculating the likelihood of species belonging to different groups, and assigning species to the group with the highest probability. This improves data processing efficiency and accuracy. By extracting features from each group of dynamic tree parameters, the differences in species growth patterns are revealed. The system then examines whether different groups perform differently under the same environment. If no differences are found, the grouping results are integrated to form a complete tree growth dataset.

[0029] Furthermore, if Figure 4 As shown, the process of calculating the total load in step S2 in combination with the weight of the planting soil and the water saturation state specifically includes the following steps: Step S21: establishing a tree load estimation model based on the tree growth database, dividing the training set of the tree growth data set into two parts, dividing the tree growth data into k folds, and using k-1 folds to train the tree load estimation model each time; Among them, each tree sample has a corresponding prediction result; Step S22: Learning tree growth patterns in different subsets and features to generate preliminary prediction results; after the first training is completed, the tree load estimation model generates prediction values ​​for the validation set and the retained sub-dataset; Step S23: Use the prediction results as input features for the second training, and construct a new feature matrix with the prediction values; use the prediction results and true labels of the first training tree load estimation model to perform the second training tree load estimation model, and learn to weight or combine the output of the first training tree load estimation model.

[0030] Preferably, this embodiment achieves efficient use of data and cross-validation, thereby improving the generalization ability of the model and reducing the risk of overfitting. By learning the growth patterns of trees of different subsets and features, generating preliminary prediction results, and making predictions on the validation set and the retained subset, the accuracy of the model is further improved. The predicted values ​​are used as new feature inputs for secondary training to further optimize the model performance, reflecting the strategy of feature fusion and model combination. By weighting or combining the prediction results of the first stage, the model can more accurately capture complex nonlinear relationships, thereby improving the accuracy and stability of the final prediction results.

[0031] Furthermore, if Figure 5 As shown, the process of dividing the training set of the tree growth data set into two parts in step S21 specifically includes the following steps: Step S211: Split the training set in the tree growth dataset into k subsets equally. If the total number of training samples is N, each fold contains approximately N / k subsets. Perform k cycles of training, selecting k01 subsets in each cycle and merging them into the training data for the current iteration. Step S212: The first cycle uses the 2nd to kth folds as the training set, the second cycle uses the 1st, 3rd to kth folds, and so on; the remaining subset is used as the validation set, and the tree load estimation model is trained on the k-1th fold training data; Step S213: After each training is completed, the current tree load estimation model is used to predict the retained validation set subset to obtain the predicted value of the subset sample; the prediction results of all k validation sets are integrated to form a set of predicted values ​​corresponding to the original training set samples.

[0032] Preferably, this embodiment divides the training set equally into k subsets to ensure that the size of each subset is close, thereby ensuring the stability of model training and verification. This process helps to reduce the deviation caused by uneven data distribution and improve the generalization ability of the model. Through k-fold training, k-1 subsets are selected as training sets each time, and the remaining 1 subset is used as the verification set, thereby avoiding the deviation problem that may be caused by a single verification set. The prediction results of the k verification sets are integrated into a complete set of prediction values ​​for the final performance evaluation. This method can comprehensively reflect the performance of the model on different data subsets, avoid the randomness of a single verification result, and thus more accurately evaluate the overall performance of the model. Through multiple iterative training and verification, the model can learn and adjust on multiple different data subsets, thereby reducing the risk of overfitting and improving the model's adaptability to unknown data.

[0033] Furthermore, if Figure 6 As shown, the process of generating the preliminary prediction result in step S22 specifically includes the following steps: Step S221: After training is completed, the validation set is used to predict the 1-fold validation set retained in the current round; predictions are made on the independently retained test set and data that did not participate in the current round of training; Step S222: The prediction results are recorded as preliminary prediction values, and all samples are used as validation sets for prediction. For the independently retained sub-datasets, the tree load estimation model performs unified predictions after each round of training, and integrates multiple prediction results. Step S223: Adjust the input data in different training rounds, learn the rules under different data distributions and generate diversified prediction values; integrate the prediction values ​​into a new feature matrix as the input features of the second training.

[0034] Preferably, after training is completed, this embodiment uses the validation set to predict the 1-fold validation set retained in the current round, and also predicts the independently retained test set and data that did not participate in the current round of training. The prediction results are recorded as preliminary prediction values, and all samples are used as the validation set for prediction. For the independently retained sub-dataset, the tree load estimation model performs unified predictions after each round of training and integrates multiple prediction results. The input data is adjusted in different training rounds, and the rules are learned under different data distributions to generate diverse prediction values. The prediction values ​​are integrated into a new feature matrix as the input features for the second training.

[0035] Furthermore, if Figure 7 As shown, the process of integrating the predicted values ​​into a new feature matrix in step S223 specifically includes the following steps: Step S2231: retain a different validation subset in each round, and use the remaining as a training subset; after each round of training, use the currently retained validation set to calculate the performance of the tree load estimation model under the current distribution; an independent test set is used by default, which does not participate in the current round of training data, and predictions are performed after training is completed; Step S2232: Iterate sequentially, obtaining multiple prediction results for each sample from different distribution training to form an initial prediction value set; each sample is predicted a preset number of times, and each result is used as a new feature; the prediction results of each round of validation set are spliced ​​into a new feature matrix; Step S2233: perform multiple predictions on the independent test set and finally integrate them into new features of the test set; input the integrated new feature matrix into the tree load estimation model for secondary training and fuse the predicted values.

[0036] Preferably, this embodiment divides the training set and validation set multiple times, and uses different subsets for model training and validation to ensure the generalization ability of the model under different data distributions. Through multiple iterations, multiple prediction results are obtained for each sample to form a new feature matrix. This technical feature utilizes the ensemble learning idea of ​​the model to generate new features by fusing the prediction results of multiple models, thereby improving the prediction accuracy and robustness of the model. The multiple prediction results of the independent test set are integrated into new features for secondary training of the model. This technical feature further enhances the prediction ability of the model through feature engineering, while fusing the prediction results to reduce the risk of overfitting and improve the performance of the model on new data.

[0037] Furthermore, if Figure 8 As shown, the process of learning to weight or combine the output of the first training tree load estimation model in step S23 specifically includes the following steps: Step S231: combining the new feature matrix with the true labels of the original data set to form training data for a second training tree load estimation model; the new feature matrix is ​​used as input features, and the true labels are used as target variables; Step S232: Learn the weight coefficients of the predicted values ​​and capture the complex interactions in the prediction process; set the prediction error value and train with the goal of minimizing the prediction error value; evaluate the effectiveness of the weighted / combined strategy on the validation set; Step S233: After the second training is completed, the predicted value of the first training for the new data is used as the input of the second training, and the final prediction result after optimized weighting or combination is output; the tree data collected in real time is input into the tree load assessment model for evaluation.

[0038] Preferably, this embodiment combines the new feature matrix with the true labels of the original dataset to form training data for the second training tree load estimation model. The weight coefficients of the predicted values ​​are learned to capture the complex interactions during the prediction process. A prediction error value is set and training is performed with the goal of minimizing the prediction error. After the second training is completed, the predicted values ​​for the new data from the first training are used as input for the second training, and the final prediction results, after optimized weighting or combination, are output. This iterative optimization method not only improves the model's prediction accuracy but also enhances its adaptability to dynamic data, reflecting the advantages of model fusion and multi-stage optimization.

[0039] Furthermore, if Figure 9 As shown, the process of evaluating the growth coefficient and the impact of environmental variables on load prediction in step S3 specifically includes the following steps: Step S31: Filter the prediction results for environmental variables and growth coefficients that affect load prediction; determine the contribution of parameters such as crown width and tree height to biomass through variable screening; and select the most unfavorable stress factor as the key variable by comparing the temperature stress coefficient and the water stress coefficient. Step S32: Calculate the difference in the mean values ​​of different genotypes to determine the influence of the variables; if the error index exceeds the preset threshold or the contribution of the key variable deviates from the expectation, trigger the adjustment instruction of the tree load assessment model; Step S33: Re-evaluate the adjusted model on the independent validation set; output the adjusted load results as formatted data through the cloud interface and record the adjustment log.

[0040] Preferably, this embodiment determines the contribution of parameters such as crown width and tree height to biomass by screening environmental variables and growth coefficients, and selects the most unfavorable stress factor as the key variable. Optimize the model input variables, improve the accuracy and reliability of model predictions, and reduce the impact of irrelevant variables on the model. Calculate the difference in mean values ​​of different genotypes to determine the impact of variables; if the error index exceeds the preset threshold or the contribution of the key variable deviates from expectations, the adjustment instruction is triggered. Dynamically adjust the model parameters to ensure that the model can be adaptively optimized when facing different genotypes, and improve the robustness and generalization ability of the model. Re-evaluate the adjusted model on an independent validation set, output formatted data through the cloud interface, and record the adjustment log. Verify the performance of the adjusted model to ensure the stability and traceability of the model in practical applications, while achieving efficient management and sharing of data.

[0041] like Figure 10 As shown, this embodiment further provides an embodiment of a tree load estimation system based on different growth coefficients. In this embodiment, the tree load estimation system based on different growth coefficients is applied to the tree load estimation method based on different growth coefficients in the above embodiment. The tree load estimation system based on different growth coefficients includes: Tree data collection module 1 is used to generate a tree growth database, collect tree growth parameters, extract tree planting point coordinates, generate a developmental tree, and analyze growth differences between species; pre-process the tree growth parameters to obtain a tree growth dataset, and divide the tree growth dataset into a training set, a validation set, and a test set; Tree load assessment module 2 is used to establish a tree load estimation model based on the tree growth database and train the tree load estimation model using the training set; it estimates the biomass accumulation of the tree within the preset growth cycle and calculates the total load based on the planting soil weight and water saturation state; The evaluation result output module 3 is used to evaluate the impact of the growth coefficient and environmental variables on load prediction. If the evaluation result is greater than the preset threshold, the tree load estimation model is adjusted; if the evaluation result is less than the threshold, the tree load result is output to the cloud.

[0042] Among them, the biomass-soil-water coupled load formula when the tree load results are output to the cloud is:

[0043] Where, Indicates the output value of the tree comprehensive load; represents the species growth morphology correction coefficient; represents the maximum biomass accumulation rate; represents the xylem density gradient parameter; represents the soil compression effect factor; Indicates the wet density of soil on site; represents the critical liquefaction density threshold; Indicates the i Grade branch dry matter mass; represents the canopy height attenuation coefficient; Indicates the i Relative height of grade branches; represents the water form conversion constant; represents the saturated cross-sectional area of ​​the root system; represents the pore water pressure angle; Cloud transmission verification formula:

[0044] Where, Indicates the cloud standardized load value; represents the atmospheric transmission attenuation factor; Indicates the real-time ambient temperature; Indicates the reference temperature; Indicates the annual temperature fluctuation range; Indicates the k The weight coefficient of each meteorological station; Indicates the k the atmospheric pressure gradient at the site; Dynamic correction feedback formula:

[0045] Where, represents the model correction amount; represents the learning rate decay coefficient; Indicates the actual measured value of the field sensor; represents the model prediction value; represents the temporal forgetting factor; represents the spatial discreteness weight; Represents the Laplace norm of the soil displacement tensor.

[0046] The core functions of the formulation system include: biomechanical-fluid dynamics coupling, which quantitatively models the interaction between biomass and soil and water by tensor-integrating xylem density gradients, soil compressibility, and pore water pressure. Cloud-based data fidelity conversion incorporates atmospheric transmission attenuation factors and pressure gradient weights to correct for the effects of ambient temperature and humidity on load transmission. An adaptive correction system dynamically calibrates the deviation between the model and sensor data, combining temporal forgetting and spatial discreteness. Parameter fitting and computational stability are optimized using the Levenberg-Marquardt nonlinear least squares algorithm. Scale adaptability ensures numerical stability within the model's scale by constraining parameter ranges and performing normalization operations. The formulation system utilizes third-order coupling to achieve: tensor fusion of biomechanical and fluid dynamic parameters; cloud-based data fidelity conversion that accounts for atmospheric transmission attenuation; and an adaptive correction system that incorporates a temporal forgetting mechanism. All parameters are nonlinearly fitted using the Levenberg-Marquardt algorithm to ensure computational stability within the scale.

[0047] Preferably, this embodiment generates a tree growth database, collects and analyzes tree growth parameters, extracts planting site coordinates, and generates a developmental tree, thereby achieving comprehensive monitoring and recording of tree growth status. Based on the tree growth database, a load estimation model is established to estimate tree biomass accumulation and calculate the total load based on soil weight and moisture status, providing a scientific basis for agricultural management. The impact of growth coefficients and environmental variables on load prediction is assessed, and the model is adjusted based on the assessment results or load results are output to the cloud, enabling efficient data sharing and management.

[0048] In this embodiment, the following calculation process is involved: Trunk Volume: Using the Formula ; in is the shape factor (the difference between the actual shape of the trunk and the cylinder), usually ranging from 0.4 to 0.7.

[0049] Crown volume; assuming the crown is conical, use the formula

[0050] in For the crown width, is the crown height (if the crown height is unknown, it can be roughly taken as 1 / 3 to 1 / 2 of the tree height).

[0051] Determine density: Trunk density ( ): The density of different tree species is different. The density of common trees ranges from 0.4 to 0.8g / cm 3 , and there are large differences between different tree species (such as pine about 0.5, oak about 0.7).

[0052] Crown density ( ): The density of branches and leaves is significantly lower than that of the trunk, usually ranging from 0.1 to 0.3 g / cm 3 (Varies depending on water content, leaf volume, and tree species).

[0053] The interior of a tree canopy includes leaves, twigs, and branches, which are affected by many factors such as tree species, age, growing environment, and season. The following is a typical volume ratio based on research and experience: Table 1 Typical volume distribution

[0054] Fill factor of the crown: Fill factor of the crown ( ) is typically between 0.2 and 0.5, depending on the tree species, age, and growing environment. For example, broadleaf trees (such as oak) have a fill factor of 0.25-0.35, while coniferous trees (such as pine) have a fill factor of 0.35-0.45.

[0055] Calculate weight: Total weight = trunk weight + crown weight, ,in V is the total volume, is the density.

[0056] The load considered in tree design = G × RGR, where G represents the actual load of the tree and RGR is the growth coefficient.

[0057] Table 2 Growth coefficients of different tree species

[0058] Furthermore, the tree load estimation method based on different growth coefficients of this embodiment can be used in an embodiment of tree load estimation that deeply integrates low-carbon (carbon sink), AI, and smart management and maintenance technologies: The AI-driven precise management and maintenance system for carbon sink forests uses the biomass data output by the tree load model (step S2) to link with the carbon sink measurement algorithm to estimate the carbon sequestration of individual trees in real time. In the smart management and maintenance platform, AI automatically marks tree species with high carbon sink potential (such as nanmu and camphor) and prioritizes the allocation of maintenance resources.

[0059] Low-carbon value: The system dynamically optimizes planting density. When the load model predicts that the total load of mature trees in a certain area is close to the soil bearing capacity threshold, AI will recommend a thinning plan and process the thinned wood into long-lasting wood products to achieve carbon sequestration, forming a "measurement-management-carbon lock-in" closed loop.

[0060] Smart flood detention tree networks in sponge cities: Intelligent linkage: Integrating the load model's water saturation calculation function (step S2), IoT soil sensors transmit real-time data during heavy rain warnings. AI automatically activates drainage systems around high-load-tolerant trees (such as banyan trees), while simultaneously calculating flood retention benefits using a canopy rainfall interception model. Load prediction results are integrated with the BIM system to simulate the long-term impact of different tree species combinations on underground garage roof loads during the design phase, ensuring no overloads during the 50-year growth cycle, achieving both ecological benefits and structural safety.

[0061] Digital twin protection project for ancient and famous trees: LiDAR scanning of ancient trees creates a three-dimensional growth model, which is then input into a load estimation system (step S1) to generate temporal and spatial biomass variation curves. AI compares this with historical disaster data (e.g., typhoon-induced lodging cases). When the predicted risk of a sudden load change exceeds 25%, reinforcement measures are triggered. Blockchain stores carbon sequestration data for ancient trees, generating tradable carbon certificates. Maintenance personnel view real-time load and stress cloud maps using AR glasses, enabling millimeter-level precision support operations.

[0062] These cases demonstrate how load estimation technology has been upgraded from a simple structural safety tool to a systematic solution that integrates carbon management, disaster warning, and digital assets.

[0063] like Figure 11 As shown, this embodiment provides an embodiment of an electronic device. In this embodiment, the electronic device 4 includes a processor 41 and a memory 42 coupled to the processor 41.

[0064] The memory 42 stores program instructions for implementing the tree load estimation method based on different growth coefficients according to any one of the above embodiments.

[0065] The processor 41 is configured to execute program instructions stored in the memory 42 to perform layout of tree load estimation methods based on different growth coefficients.

[0066] The processor 41 may also be referred to as a CPU (Central Processing Unit). The processor 41 may be an integrated circuit chip with signal processing capabilities. The processor 41 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component. The general-purpose processor may be a microprocessor or any conventional processor.

[0067] Further, Figure 12This is a schematic diagram of the structure of a storage medium in an embodiment of the present application. The storage medium 5 in the embodiment of the present application stores program instructions 51 that can implement all of the above methods. The program instructions 51 can be stored in the above storage medium in the form of a software product, including a number of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or terminal devices such as a computer, server, mobile phone, and tablet.

[0068] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interface, device or unit, which can be electrical, mechanical or other forms.

[0069] In addition, the functional units in the various embodiments of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated units may be implemented in the form of hardware or in the form of software functional units. The above is only an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

[0070] The above detailed description of the specific embodiments of the invention is intended to be illustrative only, and the present invention is not limited to the specific embodiments described above. For those skilled in the art, any equivalent modifications or substitutions to the invention are also within the scope of the present invention. Therefore, equivalent changes, modifications, and improvements made without departing from the spirit and scope of the present invention are also encompassed within the scope of the present invention.

Claims

1. A method for estimating tree loads based on different growth coefficients, characterized in that: The tree load estimation method based on different growth coefficients includes: generating a tree growth database, collecting tree growth parameters, extracting tree planting point coordinates, estimating the biomass accumulation of the tree within a preset growth cycle, and calculating the total load in combination with the weight of the planting soil and the water saturation state; and evaluating the impact of the growth coefficient and environmental variables on load prediction.

2. The method for estimating tree loads based on different growth coefficients according to claim 1, characterized in that: The tree growth parameters are preprocessed to obtain the tree growth dataset, which is then divided into a training set, a validation set, and a test set.

3. The method for estimating tree loads based on different growth coefficients according to claim 2, characterized in that: The process of analyzing growth differences between species involves the following steps: Generate a tree growth database and collect growth parameters of different tree species; tree species growth parameters include biomass, physiological indicators, and environmental impact data; clean the collected heterogeneous data and perform normalization to eliminate dimensional differences; Perform differential tests on growth parameters of different species to identify species groups with differences; determine whether there are differences in leaf area index between different species under the same environment; extract features of tree growth parameters and group species according to growth parameter characteristics; Gene sequence data were collected, multiple sequence comparisons were performed, and developmental trees were constructed to analyze the correlation between evolutionary differences and phenotypic differences. The tree growth dataset was formed and divided into training set, validation set and test set.

4. The method for estimating tree loads based on different growth coefficients according to claim 3, characterized in that: The process of grouping species based on growth parameter characteristics includes the following steps: The observational data comes from multiple unobserved tree data sets, each with different dynamic tree parameters. The tree data parameters for each set are inferred, the optimal number of groupings is automatically inferred, the probability of each species belonging to a different group is calculated, and the calculated species are assigned to the group with the highest probability. Extract features from the dynamic tree parameters corresponding to each group to obtain differences in species growth patterns; test whether there are differences in growth performance between different groups under the same environment; if there are differences, regroup them; If it is judged that there is no difference, the grouping results will be integrated to form the tree growth data set.

5. The method for estimating tree loads based on different growth coefficients according to claim 2, characterized in that: A tree load estimation model was established based on the tree growth database, and the training set was used to train the tree load estimation model. The biomass accumulation of the tree within the preset growth cycle was estimated, and the total load was calculated in combination with the planting soil weight and water saturation state.

6. The method for estimating tree loads based on different growth coefficients according to claim 5, characterized in that: The process of calculating the total load based on the weight of the planting soil and the water saturation state includes the following steps: A tree load estimation model is established based on the tree growth database. The training set of the tree growth data set is divided into two parts. The tree growth data is divided into k folds, and the tree load estimation model is trained with k-1 folds each time. Among them, each tree sample has a corresponding prediction result; The tree growth patterns are learned in different subsets and features to generate preliminary prediction results. After the first training is completed, the tree load estimation model generates predictions for the validation set and the retained subset data set. The predicted results are used as input features for the second training, and the predicted values ​​are used to form a new feature matrix; the predicted results and true labels of the first trained tree load estimation model are used to train the second tree load estimation model, and the output of the first trained tree load estimation model is learned to be weighted or combined.

7. The method for estimating tree loads based on different growth coefficients according to claim 6, characterized in that: The process of dividing the training set of the tree growth dataset into two parts includes the following steps: The training set in the tree growth dataset is equally divided into k subsets. If the total number of training samples is N, each fold contains approximately N / k subsets. K cycles of training are performed, and k01 subsets are selected and merged into the training data for the current iteration in each cycle. The first cycle uses the 2nd to kth folds as the training set, the second cycle uses the 1st, 3rd to kth folds, and so on; the remaining subset is used as the validation set, and the tree load estimation model is trained on the k-1th fold training data; After each training is completed, the current tree load estimation model is used to predict the retained validation set subset to obtain the predicted value of the subset sample; the prediction results of all k validation sets are integrated to form a set of predicted values ​​corresponding to the original training set samples.

8. The method for estimating tree loads based on different growth coefficients according to claim 6, characterized in that: The process of generating preliminary forecast results includes the following steps: After training is completed, the validation set is used to predict the 1-fold validation set retained in the current round; predictions are made on the independently retained test set and data that did not participate in the current round of training; The prediction results are recorded as preliminary prediction values, and all samples are used as validation sets for prediction. For the independently retained sub-datasets, the tree load estimation model makes unified predictions after each round of training and integrates multiple prediction results. Adjust the input data in different training rounds, learn the rules under different data distributions and generate diverse prediction values; integrate the prediction values ​​into a new feature matrix as the input features of the second training.

9. The method for estimating tree loads based on different growth coefficients according to claim 4, characterized in that: Evaluate the impact of growth coefficients and environmental variables on load prediction. If the evaluation result is greater than the preset threshold, adjust the tree load estimation model; if the evaluation result is less than the threshold, output the tree load result to the cloud.

10. A tree load estimation system based on different growth coefficients, which is applied to the tree load estimation method based on different growth coefficients according to any one of claims 1 to 9, characterized in that: The tree load estimation system based on different growth coefficients includes: The tree data acquisition module is used to generate a tree growth database, collect tree growth parameters, extract tree planting point coordinates, generate a developmental tree, and analyze growth differences between species; pre-process the tree growth parameters to obtain a tree growth dataset, and divide the tree growth dataset into a training set, a validation set, and a test set; The tree load assessment module is used to establish a tree load estimation model based on the tree growth database and train the tree load estimation model using a training set; it estimates the biomass accumulation of trees within a preset growth cycle and calculates the total load based on the weight of the planting soil and the water saturation state; The evaluation result output module is used to evaluate the impact of growth coefficients and environmental variables on load prediction. If the evaluation result is greater than the preset threshold, the tree load estimation model is adjusted; if the evaluation result is less than the threshold, the tree load result is output to the cloud.

Images