Artificial forest aboveground biomass monitoring method based on model fusion

Abstract

The invention discloses a man-made forest overground biomass monitoring method based on model fusion, and relates to the technical field of biomass monitoring, and the method comprises the steps: collecting airborne hyperspectral data and airborne LiDAR data in an observation region, and the real value of the biomass of any part of the observation region; the method comprises the following steps: respectively preprocessing airborne hyperspectral data and airborne LiDAR data, and constructing an aboveground biomass model based on the airborne LiDAR data and an aboveground biomass model based on the airborne hyperspectral data; and respectively calculating a prediction error of the overground biomass model based on the airborne LiDAR data to the monitored area and a prediction error of the overground biomass model based on the airborne hyperspectral data to the monitored area, thereby setting an airborne LiDAR data weight and an airborne hyperspectral data weight, and generating a man-made forest land biomass monitoring model. According to the method, airborne hyperspectral and LiDAR data are fused, accurate forest resource data are provided for forestry departments, and more scientific resource evaluation and ecological planning are assisted.

Description

Technical Field

[0001] The present invention relates to the technical field of biomass monitoring, and in particular to a method for monitoring aboveground biomass of artificial forests based on model fusion. Background Art

[0002] When it comes to monitoring planted forests, traditional single-source data monitoring of aboveground biomass relies on a single data source, resulting in partial information. For example, while satellite remote sensing data alone can cover large areas, its resolution is low, making it difficult to accurately identify the biochemical composition of vegetation, leading to large errors in biomass estimates. Forestry resource assessments require extremely high data accuracy. Inaccurate data can mislead ecological planning. For example, when formulating forest carbon sink accounting, inaccurate biomass data can bias carbon sink estimates and affect the correct assessment of the forest's ecological value. Manual field measurements of biomass were the primary method in the early days. Workers had to measure each tree in large areas of planted forests, a tedious and time-consuming process. In some vast forests, completing a comprehensive measurement could take months or even longer. With the development of forestry, the need for timely monitoring of forest dynamics is becoming increasingly stringent. Emergencies such as forest pests and diseases, fire hazards, and other emergencies require rapid responses. Traditional manual monitoring methods are unable to provide timely data on forest resource dynamics and are unable to meet the forestry department's needs for real-time forest supervision. Therefore, there is an urgent need to leverage advanced technologies to improve the efficiency of biomass monitoring. Traditional manual biomass monitoring not only requires a large amount of manpower but also specialized measurement tools, which is costly. Furthermore, to ensure measurement accuracy, multiple repeated measurements are often required, further increasing costs. For some forestry projects with limited funds, such high monitoring costs are unaffordable. Furthermore, the limited measurement range also restricts comprehensive monitoring of large forest areas. In the pursuit of sustainable development, a new method that can reduce costs while achieving high-precision, large-scale biomass monitoring is urgently needed to promote the sustainable development of forest resource monitoring.

[0003] Therefore, people need a method for monitoring aboveground biomass of artificial forests based on model fusion to solve the above problems. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for monitoring aboveground biomass of artificial forests based on model fusion, so as to solve the problems raised in the above-mentioned background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solution: a method for monitoring aboveground biomass of artificial forests based on model fusion, comprising the following steps: S1. Collect airborne hyperspectral data, airborne LiDAR data, and the true value of biomass in any part of the observation area; pre-process the airborne hyperspectral data and airborne LiDAR data respectively; S2. Constructing an aboveground biomass model based on airborne LiDAR data and an aboveground biomass model based on airborne hyperspectral data, respectively; S3. Calculate the prediction error of the aboveground biomass model based on the airborne LiDAR data and the prediction error of the aboveground biomass model based on the airborne hyperspectral data for the monitoring area, thereby setting the weight of the airborne LiDAR data and the weight of the airborne hyperspectral data; S4. Integrate the aboveground biomass model based on airborne LiDAR data and the aboveground biomass model based on airborne hyperspectral data to generate a plantation biomass monitoring model, and use the plantation biomass monitoring model to monitor biomass.

[0006] Furthermore, in step S1, N feature data {x 1L ,x 2L ,…,x nL ,…,x NL}, and obtain the I characteristic parameters of hyperspectral data {x 1H ,x 2H ,…,x iL ,…,x IH}, the preprocessing includes removing data noise and correcting radiation errors. The preprocessing operation is the key foundation of the entire artificial forest aboveground biomass monitoring method based on model fusion. Removing data noise will eliminate the messy signals that interfere with the accuracy of the data, ensuring that the airborne LiDAR data and hyperspectral data can clearly and accurately reflect the actual characteristics of the forest. This greatly improves the stability and reliability of the model when these data are subsequently used to build a biomass model, reducing the risk of errors caused by data fluctuations; correcting radiation errors is an important measure to ensure the authenticity of the data. Through correction, data acquired under different conditions can be analyzed under the same standard, making the comparison and fusion of data more scientific, so that the extracted feature data can truthfully display the growth status, spatial structure and other key information of forest vegetation, laying a solid foundation for the accurate monitoring of artificial forest aboveground biomass, and strongly supporting subsequent model construction and biomass estimation work.

[0007] Furthermore, in step S2, the nth feature data of the airborne LiDAR data is analyzed and the random forest algorithm is used to calculate the nth feature data x nL The importance score P(x nL ): ;

[0008] Where T is the number of decision trees in the random forest, err(O L ) does not use feature data x nLThe overall error, err(t L ) is the use of feature data x nL The error of the t-th decision tree is used to obtain the importance scores of the N feature data of the airborne LiDAR data. K feature data are selected from high to low according to the importance scores as the important feature data of the aboveground biomass model based on the airborne LiDAR data. Similarly, the hyperspectral data is analyzed to obtain B feature data as the important feature data of the aboveground biomass model based on the airborne hyperspectral data. The aboveground biomass model based on the airborne LiDAR data is constructed: ;

[0009] where β kL is the kth regression coefficient of the aboveground biomass model based on airborne LiDAR data, k = 1, 2, …, K, x kL represents the kth important feature data of the aboveground biomass model based on airborne LiDAR data, represents the error term of the aboveground biomass model based on airborne LiDAR data, represents the predicted biomass of aboveground biomass based on airborne LiDAR data; The feature data x is not used. nL The method for obtaining the overall error when : When constructing any decision tree in the random forest, a part of the observation area is randomly extracted from the entire observation area to construct the training set of the tree. The set of some observation areas that are not extracted into the training set becomes the out-of-bag samples of any decision tree; After all decision trees are built, for any decision tree, calculate the error MSE of any decision tree: ;

[0010] where y i represents the true value of the jth observation area in the out-of-bag sample, Y i represents the predicted value of the jth observation region in the out-of-bag sample, α represents the number of observation regions in the out-of-bag sample, and the predicted value is obtained by not considering the feature data x when splitting the node in any decision tree. nL Method for prediction; Then calculate the average error of all decision trees as the average error without using feature data x nL The overall error err(O L ); The usage characteristic data x nL The error of the t-th decision tree is obtained by randomly extracting some observation areas with replacement from the observation area to form a training subset for constructing the t-th decision tree. In the process of node splitting of the t-th decision tree, the feature data x is used. nLAll feature data including , determine the node splitting method according to the Gini bad rate criterion, and thus complete the construction of the t-th decision tree; The set of observation areas that are not drawn to build the t-th decision tree constitutes the out-of-bag data of the t-th decision tree. For the J-th observation area in the out-of-bag data, traverse downward according to the t-th decision tree and finally reach a leaf node. The predicted value z represented by the leaf node is J It is the predicted value of the Jth observation area in the out-of-bag data; Then calculate the feature data x nL The error err(t L ): ;

[0011] where Z J represents the true value of the Jth observation region in the out-of-bag data, and μ represents the number of observation regions in the out-of-bag data; The kth regression coefficient of the aboveground biomass model based on airborne LiDAR data is obtained by the least squares method; The aboveground biomass model based on airborne hyperspectral data was constructed using the method of constructing an aboveground biomass model based on airborne LiDAR data: ;

[0012] where β bH is the b-th regression coefficient of the aboveground biomass model based on airborne hyperspectral data, b=1,2,…,B, x bH represents the bth important feature data of B aboveground biomass models based on airborne hyperspectral data, represents the error term of the aboveground biomass model based on airborne hyperspectral data, The data represents the predicted aboveground biomass based on airborne hyperspectral data. From a data processing and analysis perspective, the use of a random forest algorithm to calculate the importance score of feature data is a highlight. By comparing the error of the decision tree when using and not using specific feature data, the value of each feature is accurately assessed. This process can deeply mine the effective information in airborne LiDAR and hyperspectral data, avoiding model redundancy caused by blindly using all data, and making the selected K (for LiDAR data) and B (for hyperspectral data) important feature data more representative. For example, in complex forest environments, key features such as vegetation height and spectral reflectance that contribute most to biomass estimation can be quickly identified, laying a solid foundation for subsequent model construction.

[0013] From a model construction perspective, aboveground biomass models were constructed based on the selected key feature data, improving the model's relevance and accuracy. Least squares regression coefficients were obtained, enabling the model to optimally fit the data and minimize errors. Using the airborne LiDAR data model as an example, by assigning appropriate weights to different feature data, the relationship between forest three-dimensional structure and biomass can be more accurately reflected. Furthermore, this model construction method is universal and applicable to airborne hyperspectral data as well, ensuring consistent model construction logic for both data sources, facilitating subsequent model fusion and results analysis.

[0014] From the perspective of error evaluation and optimization, the reliability of the model is effectively verified by calculating the decision tree error through out-of-bag samples and then finding the average value of the overall error. nL When using feature data, the overall error is calculated using out-of-bag samples, which can clearly understand the impact of the missing feature on the model; when using feature data, the error of each decision tree is calculated, and the evaluation is further refined, which helps to adjust the model parameters in time and optimize the model performance, thereby significantly improving the prediction accuracy of the aboveground biomass of artificial forests and providing more reliable data support for forestry resource monitoring.

[0015] Furthermore, in step S3, the initial biomass M is obtained by drone for any monitoring area, and the aboveground biomass prediction M based on the aboveground biomass is calculated based on the aboveground biomass model based on the airborne LiDAR data. L The aboveground biomass model of airborne hyperspectral data is used to calculate the predicted biomass M based on the aboveground biomass of airborne hyperspectral data. H , thereby setting the airborne LiDAR data weight (|M H -M|) / (|M L -M|+|M H -M|) and airborne hyperspectral data weight (|M L -M|) / (|M L -M|+|M H -M|), obtain the initial biomass through drones, combine the two models to set weights, and dynamically adjust the role of airborne LiDAR and hyperspectral data in biomass calculation according to the actual conditions of different monitoring areas, making data fusion more in line with reality and improving monitoring accuracy.

[0016] Furthermore, in step S4, a plantation biomass monitoring model is generated: ;

[0017] in It represents the predicted amount of biomass in artificial forests, effectively integrating the advantages of the two models, giving full play to the ability of LiDAR data to capture forest structure information and hyperspectral data to capture vegetation biochemical information, achieving a more accurate and comprehensive prediction of the biomass of artificial forests, and providing strong data support for the scientific management of forestry resources.

[0018] Compared with the existing technology, the beneficial effects achieved by the present invention are: on the one hand, the traditional single data monitoring method has large errors in biomass estimation. The present invention integrates airborne hyperspectral and LiDAR data. The former can accurately identify the biochemical components of vegetation and judge the growth status of plants by virtue of its continuous spectral characteristics; the latter uses laser detection to obtain the three-dimensional structure of the forest and clarify parameters such as tree height and density. The combination of the two fully covers the information required for biomass estimation. In actual monitoring, the prediction results of the fusion model are closer to the actual biomass, reducing prediction errors, providing the forestry department with accurate forest resource data, and facilitating more scientific resource assessment and ecological planning; On the one hand, in the past, when measuring biomass manually on the ground, workers had to measure each tree one by one. In large-scale artificial forest areas, this process was time-consuming and difficult. The monitoring method based on model fusion, with the help of remote sensing technology, can quickly collect data on large areas of forests. By using equipment mounted on aircraft, large-area scans can be completed in a short period of time to obtain massive amounts of data. The data is quickly analyzed and processed by the model, and the biomass monitoring results can be output. Compared with traditional methods, the monitoring cycle is shortened, allowing forestry departments to keep abreast of forest dynamics and changes, and gain valuable time to respond to emergencies such as forest pests and diseases, fire hazards, etc. On the other hand, traditional manual monitoring requires extensive manpower and specialized measurement tools. Due to the limited measurement range, multiple measurements are often required to ensure accuracy. The new method reduces field measurement work and manpower requirements. A single remote sensing measurement can capture multi-dimensional data, enabling diverse analyses. The efficient data collection and precise results reduce overall monitoring costs, enabling forestry projects to achieve high-precision, large-scale biomass monitoring, thereby promoting the sustainable development of forest resource monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 This is a flow chart of the aboveground biomass monitoring method of artificial forests based on model fusion 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. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0021] See also Figure 1 The present invention provides a technical solution: a method for monitoring aboveground biomass of artificial forests based on model fusion, comprising the following steps: S1. Collect airborne hyperspectral data, airborne LiDAR data, and the true value of biomass in any part of the observation area; pre-process the airborne hyperspectral data and airborne LiDAR data respectively; S2. Constructing an aboveground biomass model based on airborne LiDAR data and an aboveground biomass model based on airborne hyperspectral data, respectively; S3. Calculate the prediction error of the aboveground biomass model based on the airborne LiDAR data and the prediction error of the aboveground biomass model based on the airborne hyperspectral data for the monitoring area, thereby setting the weight of the airborne LiDAR data and the weight of the airborne hyperspectral data; S4. Integrate the aboveground biomass model based on airborne LiDAR data and the aboveground biomass model based on airborne hyperspectral data to generate a plantation biomass monitoring model, and use the plantation biomass monitoring model to monitor biomass.

[0022] In step S1, N feature data {x 1L ,x 2L ,…,x nL ,…,x NL}, and obtain the I characteristic parameters of hyperspectral data {x 1H ,x 2H ,…,x iL ,…,x IH}, the preprocessing includes removing data noise and correcting radiation errors. The preprocessing operation is the key foundation of the entire artificial forest aboveground biomass monitoring method based on model fusion. Removing data noise will eliminate the messy signals that interfere with the accuracy of the data, ensuring that the airborne LiDAR data and hyperspectral data can clearly and accurately reflect the actual characteristics of the forest. This greatly improves the stability and reliability of the model when these data are subsequently used to build a biomass model, reducing the risk of errors caused by data fluctuations; correcting radiation errors is an important measure to ensure the authenticity of the data. Through correction, data acquired under different conditions can be analyzed under the same standard, making the comparison and fusion of data more scientific, so that the extracted feature data can truthfully display the growth status, spatial structure and other key information of forest vegetation, laying a solid foundation for the accurate monitoring of artificial forest aboveground biomass, and strongly supporting subsequent model construction and biomass estimation work.

[0023] In step S2, the nth feature data of the airborne LiDAR data is analyzed and the random forest algorithm is used to calculate the nth feature data x nL The importance score P(x nL ): ;

[0024] Where T is the number of decision trees in the random forest, err(O L ) does not use feature data x nL The overall error, err(t L ) is the use of feature data x nL The error of the t-th decision tree is used to obtain the importance scores of the N feature data of the airborne LiDAR data. K feature data are selected from high to low according to the importance scores as the important feature data of the aboveground biomass model based on the airborne LiDAR data. Similarly, the hyperspectral data is analyzed to obtain B feature data as the important feature data of the aboveground biomass model based on the airborne hyperspectral data. The aboveground biomass model based on the airborne LiDAR data is constructed: ; where β kL is the kth regression coefficient of the aboveground biomass model based on airborne LiDAR data, k = 1, 2, …, K, x kL represents the kth important feature data of the aboveground biomass model based on airborne LiDAR data, represents the error term of the aboveground biomass model based on airborne LiDAR data, represents the predicted biomass of aboveground biomass based on airborne LiDAR data; The feature data x is not used. nLThe method for obtaining the overall error when : When constructing any decision tree in the random forest, a part of the observation area is randomly extracted from the entire observation area to construct the training set of the tree. The set of some observation areas that are not extracted into the training set becomes the out-of-bag samples of any decision tree; After all decision trees are built, for any decision tree, calculate the error MSE of any decision tree: ;

[0025] where y i represents the true value of the jth observation area in the out-of-bag sample, Y i represents the predicted value of the jth observation region in the out-of-bag sample, α represents the number of observation regions in the out-of-bag sample, and the predicted value is obtained by not considering the feature data x when splitting the node in any decision tree. nL Method for prediction; Then calculate the average error of all decision trees as the average error without using feature data x nL The overall error err(O L ); The usage characteristic data x nL The error of the t-th decision tree is obtained by randomly extracting some observation areas with replacement from the observation area to form a training subset for constructing the t-th decision tree. In the process of node splitting of the t-th decision tree, the feature data x is used. nL All feature data including , determine the node splitting method according to the Gini bad rate criterion, and thus complete the construction of the t-th decision tree; The set of observation areas that are not drawn to build the t-th decision tree constitutes the out-of-bag data of the t-th decision tree. For the J-th observation area in the out-of-bag data, traverse downward according to the t-th decision tree and finally reach a leaf node. The predicted value z represented by the leaf node is J It is the predicted value of the Jth observation area in the out-of-bag data; Then calculate the feature data x nL The error err(t L ): ; where Z J represents the true value of the Jth observation region in the out-of-bag data, and μ represents the number of observation regions in the out-of-bag data; The kth regression coefficient of the aboveground biomass model based on airborne LiDAR data is obtained by the least squares method; The aboveground biomass model based on airborne hyperspectral data was constructed using the method of constructing an aboveground biomass model based on airborne LiDAR data: ; where β bH is the b-th regression coefficient of the aboveground biomass model based on airborne hyperspectral data, b=1,2,…,B, x bH represents the bth important feature data of B aboveground biomass models based on airborne hyperspectral data, represents the error term of the aboveground biomass model based on airborne hyperspectral data, The data represents the predicted aboveground biomass based on airborne hyperspectral data. From a data processing and analysis perspective, the use of a random forest algorithm to calculate the importance score of feature data is a highlight. By comparing the error of the decision tree when using and not using specific feature data, the value of each feature is accurately assessed. This process can deeply mine the effective information in airborne LiDAR and hyperspectral data, avoiding model redundancy caused by blindly using all data, and making the selected K (for LiDAR data) and B (for hyperspectral data) important feature data more representative. For example, in complex forest environments, key features such as vegetation height and spectral reflectance that contribute most to biomass estimation can be quickly identified, laying a solid foundation for subsequent model construction.

[0026] From a model construction perspective, aboveground biomass models were constructed based on the selected key feature data, improving the model's relevance and accuracy. Least squares regression coefficients were obtained, enabling the model to optimally fit the data and minimize errors. Using the airborne LiDAR data model as an example, by assigning appropriate weights to different feature data, the relationship between forest three-dimensional structure and biomass can be more accurately reflected. Furthermore, this model construction method is universal and applicable to airborne hyperspectral data as well, ensuring consistent model construction logic for both data sources, facilitating subsequent model fusion and results analysis.

[0027] From the perspective of error evaluation and optimization, the reliability of the model is effectively verified by calculating the decision tree error through out-of-bag samples and then finding the average value of the overall error. nL When using feature data, the overall error is calculated using out-of-bag samples, which can clearly understand the impact of the missing feature on the model; when using feature data, the error of each decision tree is calculated, and the evaluation is further refined, which helps to adjust the model parameters in time and optimize the model performance, thereby significantly improving the prediction accuracy of the aboveground biomass of artificial forests and providing more reliable data support for forestry resource monitoring.

[0028] In step S3, the initial biomass M is obtained by drone for any monitoring area, and the aboveground biomass prediction M based on the aboveground biomass is calculated by the aboveground biomass model based on the airborne LiDAR data. LThe aboveground biomass model of airborne hyperspectral data is used to calculate the predicted biomass M based on the aboveground biomass of airborne hyperspectral data. H , thereby setting the airborne LiDAR data weight (|M H -M|) / (|M L -M|+|M H -M|) and airborne hyperspectral data weight (|M L -M|) / (|M L -M|+|M H -M|), obtain the initial biomass through drones, combine the two models to set weights, and dynamically adjust the role of airborne LiDAR and hyperspectral data in biomass calculation according to the actual conditions of different monitoring areas, making data fusion more in line with reality and improving monitoring accuracy.

[0029] Furthermore, in step S4, a plantation biomass monitoring model is generated: ; in It represents the predicted amount of biomass in artificial forests, effectively integrating the advantages of the two models, giving full play to the ability of LiDAR data to capture forest structure information and hyperspectral data to capture vegetation biochemical information, achieving a more accurate and comprehensive prediction of the biomass of artificial forests, and providing strong data support for the scientific management of forestry resources.

[0030] Example 1: In the S1 stage, professional aerial remote sensing equipment is first used to collect airborne hyperspectral data and airborne LiDAR data within the observation area of ​​the eucalyptus forest. At the same time, a small, representative area is selected within the eucalyptus forest to conduct field measurements of the tree diameter at breast height and tree height. The acquired airborne hyperspectral and airborne LiDAR data are then preprocessed to remove data noise caused by factors such as equipment noise and atmospheric interference, and to correct for radiation errors. Ultimately, N feature data for the airborne LiDAR data, such as canopy height and canopy width, and I feature parameters for the hyperspectral data, such as spectral reflectance at different wavelengths, are obtained.

[0031] Entering the S2 stage, for each feature data of the airborne LiDAR data, the random forest algorithm is used to calculate its importance score. For example, for the feature data "canopy height", the overall error err(0) of the random forest model when it is not used is calculated. L ), and the error err(t L ), thus obtaining its importance score P(x nLBased on the scores, K important feature data were selected, and the regression coefficients were determined using the least squares method to construct an aboveground biomass model based on airborne LiDAR data. Similarly, the hyperspectral data were analyzed to construct an aboveground biomass model based on airborne hyperspectral data.

[0032] In the S3 stage, the initial biomass M is obtained in the forest using drones. Then, the first predicted biomass M is calculated using the aboveground biomass model based on airborne LiDAR data and airborne hyperspectral data. L and the second prediction quantity M H , set the weights of the two types of data.

[0033] Finally, in the S4 stage, the two models are integrated according to the determined weights to generate a plantation biomass monitoring model, which accurately predicts the aboveground biomass of the entire eucalyptus forest and provides data support for subsequent forest resource management and ecological assessment.

[0034] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A method for monitoring aboveground biomass of artificial forests based on model fusion, characterized by: The method comprises the following steps: S1. Collect airborne hyperspectral data, airborne LiDAR data and the real value of biomass in any part of the observation area; pre-process the airborne hyperspectral data and airborne LiDAR data respectively; S2, constructing an aboveground biomass model based on airborne LiDAR data and an aboveground biomass model based on airborne hyperspectral data based on airborne LiDAR data and airborne hyperspectral data respectively; S3, respectively calculating the prediction error of the aboveground biomass model based on the onboard LiDAR data for the monitoring area and the prediction error of the aboveground biomass model based on the airborne hyperspectral data for the monitoring area, thereby setting the weight of the airborne LiDAR data and the weight of the airborne hyperspectral data; S4. Integrate the aboveground biomass model based on airborne LiDAR data and the aboveground biomass model based on airborne hyperspectral data to generate an artificial forest biomass monitoring model, and use the artificial forest biomass monitoring model to monitor the biomass.

2. The method for monitoring aboveground biomass of artificial forests based on model fusion according to claim 1, characterized in that: In step S1, N feature data {x 1L ,x 2L ,…,x nL ,…,x NL }, and obtain the I characteristic parameters of hyperspectral data through preprocessing {x 1H ,x 2H ,…,x iL ,…,x IH }, the preprocessing includes removing data noise and correcting radiation errors.

3. The method for monitoring aboveground biomass of artificial forests based on model fusion according to claim 2, characterized in that: In step S2, the nth feature data of the airborne LiDAR data is analyzed and the random forest algorithm is used to calculate the nth feature data x nL The importance score P(x nL ): ； Where T is the number of decision trees in the random forest, err(O L ) does not use feature data x nL The overall error, err(t L ) is the use of feature data x nL The error of the t-th decision tree is obtained, and then the importance scores of the N feature data of the airborne LiDAR data are obtained. K feature data are selected from high to low according to the importance scores as the important feature data of the aboveground biomass model based on the airborne LiDAR data. Similarly, the hyperspectral data are analyzed to obtain B feature data as important feature data of the aboveground biomass model based on airborne hyperspectral data, and the aboveground biomass model based on airborne LiDAR data is constructed.

4. The method for monitoring aboveground biomass of artificial forests based on model fusion according to claim 3 is characterized in that: The feature data x is not used. nL The method for obtaining the overall error when : When constructing any decision tree in the random forest, randomly select some observation areas from the entire observation area to construct the training set of the tree. The set of some observation areas that are not selected to enter the training set becomes the out-of-bag samples of any decision tree; When all decision trees are built, the error MSE of any decision tree is calculated, and then the average error of all decision trees is calculated as the average error of the decision tree without using feature data x. nL The overall error err(O L ).

5. The method for monitoring aboveground biomass of artificial forests based on model fusion according to claim 4 is characterized in that: The use characteristic data x nL The error of the t-th decision tree is obtained by randomly selecting part of the observation area with replacement from the observation area to form a training subset for constructing the t-th decision tree. In the process of node splitting of the t-th decision tree, the feature data x is used. nL All feature data including , determine the node splitting method according to the Gini bad rate criterion, and thus complete the construction of the t-th decision tree; The set of observation areas that are not drawn to build the t-th decision tree constitutes the out-of-bag data of the t-th decision tree. For the J-th observation area in the out-of-bag data, traverse downward according to the t-th decision tree and finally reach a leaf node. The predicted value z represented by the leaf node J It is the predicted value of the Jth observation area in the out-of-bag data; Then, based on the true value of the Jth observation area in the out-of-bag data and the number of observation areas in the out-of-bag data, the feature data x is calculated. nL The error err(t L ).

6. The method for monitoring aboveground biomass of artificial forests based on model fusion according to claim 5, characterized in that: The kth regression coefficient of the aboveground biomass model based on airborne LiDAR data is obtained by the least squares method.

7. The method for monitoring aboveground biomass of artificial forests based on model fusion according to claim 6, characterized in that: The method of constructing an aboveground biomass model based on airborne LiDAR data was adopted. The error term of the aboveground biomass model based on airborne hyperspectral data and the predicted biomass of aboveground biomass based on airborne hyperspectral data were considered to construct an aboveground biomass model based on airborne hyperspectral data.

8. The method for monitoring aboveground biomass of artificial forests based on model fusion according to claim 7, characterized in that: In step S3, the initial biomass M is obtained by drone for any monitoring area, and the predicted biomass M of the aboveground biomass based on the airborne LiDAR data is calculated by the aboveground biomass model based on the airborne LiDAR data. L The aboveground biomass model based on airborne hyperspectral data is used to calculate the predicted biomass M of the aboveground biomass based on airborne hyperspectral data. H , thereby setting the weight of the airborne LiDAR data (|M H -M|) / (|M L -M|+|M H -M|) and airborne hyperspectral data weight (|M L -M|) / (|M L -M|+|M H -M|).

9. The method for monitoring aboveground biomass of artificial forests based on model fusion according to claim 8, characterized in that: In step S4, a plantation biomass monitoring model is generated to obtain the predicted biomass of the plantation biomass. .

Images