Artificial Forest Carbon Sink Monitoring Report Automatic Generation System and Method

Abstract

This application relates to the field of artificial forest carbon sinks, and specifically discloses an automatic generation system and method for artificial forest carbon sink monitoring reports. First, it obtains carbon dioxide values at different environmental heights, coniferous artificial forest growth environment data, and coniferous artificial forest ecological productivity text data collected by carbon dioxide monitors. Then, using deep learning technology, it performs feature extraction and correlation analysis on the three. Finally, through a generator, it generates an artificial forest carbon sink monitoring report, thereby providing more accurate and effective support for artificial forest management and environmental protection.

Description

Technical Field

[0001] This application relates to the field of carbon sinks in planted forests, and more specifically, to a system and method for automatically generating a monitoring report on the carbon sink of planted forests. Background Art

[0002] A planted forest is a forest area that is artificially planted and managed. Its main purpose is to achieve the expected forest land effect under specific spatial layouts and management measures through artificially selected and cultivated tree species. These forest lands are usually established through technical means such as sowing, planting, or cutting to ensure that the individual trees have good genetic qualities and adaptability. The individual trees in a planted forest are usually of the same age and are evenly distributed throughout the forest land, making the overall stand structure uniform and reasonable.

[0003] However, current technologies pose some challenges in the carbon sink monitoring of planted forests. Existing methods mainly measure the carbon dioxide content directly in the air of planted forests. Based on single monitoring data, it is impossible to accurately evaluate the carbon dioxide level in a specific space, thus limiting the accuracy and comprehensiveness of the assessment of the carbon sink effect of planted forests.

[0004] Therefore, a system and method for automatically generating a monitoring report on the carbon sink of planted forests are desired. Summary of the Invention

[0005] To solve the above technical problems, this application is proposed. Embodiments of this application provide a system and method for automatically generating a monitoring report on the carbon sink of planted forests. First, it obtains carbon dioxide values at different environmental heights, growth environment data of coniferous planted forests, and text data on the ecological productivity of coniferous planted forests collected by a carbon dioxide monitor. Then, using deep learning technology, it extracts features and conducts correlation analysis on the three. Finally, through a generator, it generates a monitoring report on the carbon sink of planted forests, thereby providing more accurate and effective support for the management of planted forests and environmental protection.

[0006] According to one aspect of this application, a system for automatically generating a monitoring report on the carbon sink of planted forests is provided, which includes:

[0007] A carbon sink data acquisition module for planted forests, which is used to obtain carbon dioxide values at different environmental heights, growth environment data of coniferous planted forests, and text data on the ecological productivity of coniferous planted forests collected by a carbon dioxide monitor;

[0008] A carbon sink data extraction module for planted forests, which is used to extract a carbon dioxide feature vector of coniferous planted forests and a multi-modal association feature vector of the growth state of planted forests from the carbon dioxide values at different environmental heights, the growth environment data of coniferous planted forests, and the text data on the ecological productivity of coniferous planted forests collected by the carbon dioxide monitor;

[0009] A carbon sink monitoring report generation module, which is used to generate an artificial forest carbon sink monitoring report based on the coniferous plantation carbon dioxide eigenvector and the multi-modal correlation eigenvector of the artificial forest growth state.

[0010] According to another aspect of the present application, there is provided a method for automatically generating an artificial forest carbon sink monitoring report, which includes:

[0011] Obtain the carbon dioxide values at different environmental heights, the coniferous plantation growth environment data, and the coniferous plantation ecological productivity text data collected by a carbon dioxide monitor;

[0012] Extract the coniferous plantation carbon dioxide eigenvector and the multi-modal correlation eigenvector of the artificial forest growth state from the carbon dioxide values at different environmental heights, the coniferous plantation growth environment data, and the coniferous plantation ecological productivity text data collected by the carbon dioxide monitor;

[0013] Generate an artificial forest carbon sink monitoring report based on the coniferous plantation carbon dioxide eigenvector and the multi-modal correlation eigenvector of the artificial forest growth state.

[0014] Compared with the prior art, the automatic generation system and method for an artificial forest carbon sink monitoring report provided by the present application first obtain the carbon dioxide values at different environmental heights, the coniferous plantation growth environment data, and the coniferous plantation ecological productivity text data collected by a carbon dioxide monitor, then use deep learning technology to perform feature extraction and correlation analysis on the three, and finally generate an artificial forest carbon sink monitoring report through a generator, so as to provide more accurate and effective support for artificial forest management and environmental protection. Description of the Drawings

[0015] By describing the embodiments of the present application in more detail in conjunction with the drawings, the above and other objects, features, and advantages of the present application will become more obvious. The drawings are used to provide a further understanding of the embodiments of the present application, and constitute a part of the specification, and are used to explain the present application together with the embodiments of the present application, and do not constitute a limitation to the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0016] Figure 1 It is a block diagram of an automatic generation system for an artificial forest carbon sink monitoring report according to an embodiment of the present application.

[0017] Figure 2 It is a block diagram of an artificial forest carbon sink data extraction module in the automatic generation system for an artificial forest carbon sink monitoring report according to an embodiment of the present application.

[0018] Figure 3 It is a block diagram of an artificial forest growth environment feature extraction unit in the automatic generation system for an artificial forest carbon sink monitoring report according to an embodiment of the present application.

[0019] Figure 4 It is a block diagram of a carbon sink monitoring report generation module in an artificial forest carbon sink monitoring report automatic generation system according to an embodiment of the present application.

[0020] Figure 5 It is a flowchart of an artificial forest carbon sink monitoring report automatic generation method according to an embodiment of the present application.

[0021] Figure 6 It is a block diagram of an electronic device according to an embodiment of the present application. Detailed implementation manners

[0022] Next, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the example embodiments described herein.

[0023] Figure 1 It is a block diagram of an artificial forest carbon sink monitoring report automatic generation system according to an embodiment of the present application. As Figure 1 shown, the artificial forest carbon sink monitoring report automatic generation system 100 according to an embodiment of the present application includes: an artificial forest carbon sink data acquisition module 110, configured to acquire carbon dioxide values at different environmental heights, coniferous artificial forest growth environment data, and coniferous artificial forest ecological productivity text data collected by a carbon dioxide monitor; an artificial forest carbon sink data extraction module 120, configured to extract a coniferous artificial forest carbon dioxide feature vector and an artificial forest growth state multi-modal association feature vector from the carbon dioxide values at different environmental heights, the coniferous artificial forest growth environment data, and the coniferous artificial forest ecological productivity text data collected by the carbon dioxide monitor; and a carbon sink monitoring report generation module 130, configured to generate an artificial forest carbon sink monitoring report based on the coniferous artificial forest carbon dioxide feature vector and the artificial forest growth state multi-modal association feature vector.

[0024] In the above-mentioned automatic generation system 100 for the artificial forest carbon sink monitoring report, the artificial forest carbon sink data acquisition module 110 is used to acquire carbon dioxide values at different environmental heights, coniferous artificial forest growth environment data, and coniferous artificial forest ecological productivity text data collected by a carbon dioxide monitor. It should be understood that an artificial forest refers to a forest area designed and managed by humans, whose main purpose is to achieve the expected ecological effects under specific layouts and management measures through carefully selected and cultivated tree species. These forest lands are generally established by methods such as sowing, planting, or cuttings to ensure that the trees have excellent genetic characteristics and the ability to adapt to the environment. The trees in an artificial forest are usually of the same age and are evenly distributed throughout the area, making the overall structure of the forest land uniform and reasonable. However, in the aspect of artificial forest carbon sink monitoring, there are currently some technical challenges. The current monitoring methods usually evaluate the carbon sink effect by directly measuring the carbon dioxide content in the air. However, this method relies on a single data source and is difficult to accurately reflect the carbon dioxide concentration in a specific area. Therefore, there are limitations in the accuracy and comprehensiveness of the evaluation of the artificial forest carbon sink benefits. Therefore, in the technical solution of this application, by acquiring carbon dioxide values at different environmental heights, coniferous artificial forest growth environment data, and coniferous artificial forest ecological productivity text data collected by a carbon dioxide monitor, and combining deep learning technology to generate an artificial forest carbon sink monitoring report, it can provide more accurate and efficient support for artificial forest management and environmental protection.

[0025] Specifically, obtaining the carbon dioxide values at different environmental heights, the growth environment data of coniferous plantations, and the text data on the ecological productivity of coniferous plantations collected by carbon dioxide monitors is for comprehensively evaluating the carbon sink capacity and ecological health of plantations. Among them, the carbon dioxide monitoring data provides information on the carbon dioxide concentration at different heights in the plantation area. This is the basis for evaluating the forest carbon sink function because the change in carbon dioxide concentration directly reflects the absorption and release of carbon by plants. By monitoring the carbon dioxide values at different heights, the carbon exchange conditions at different levels inside the forest can be understood, thus more comprehensively evaluating the forest's carbon sequestration capacity. The growth environment data of coniferous plantations covers ecological environment factors such as soil quality, climate conditions, and water supply. These data are crucial for understanding the growth status and ecological functions of the forest. The text data on the ecological productivity of coniferous plantations includes detailed information on tree growth, yield, stand productivity, and biomass. These data provide productivity indicators for the forest ecosystem and can illustrate the growth efficiency of the forest under different conditions. By analyzing these text data, the carbon storage potential of the forest can be quantified, and the productivity changes under different environmental conditions can be evaluated. The comprehensive utilization of these three types of data can provide comprehensive basic information for monitoring the carbon sink capacity of coniferous plantations. This not only helps to accurately evaluate the role of forests in carbon emission reduction but also provides a scientific basis for forest management and protection strategies, supporting the implementation of sustainable forestry practices. In a specific embodiment of this application, the growth environment data of coniferous plantations includes climate data and soil data, and the text data on the ecological productivity of coniferous plantations includes stand productivity data and biomass data.

[0026] In a specific embodiment of the application, taking low - efficiency coniferous forests (such as aerial - seeding masson pine forests and Chinese fir plantations) as the research objects, analyzing the impact of transformation measures on their stand characteristics, using models to simulate the succession process of stands before and after transformation, and comprehensively considering factors such as stand characteristics, environmental factors, and human disturbances, analyzing the change characteristics of carbon sinks at different succession stages and their impact on the carbon cycle process, and clarifying the driving mechanism of the spatio - temporal changes of carbon sinks in two types of low - quality coniferous forests after transformation. Specifically, by conducting investigations and calibrations of model tree species parameters, collecting environmental data such as climate and soil required by the models, simulating the growth and carbon sink dynamics of typical plantations in the study area within a specific time period, then, collecting verification data such as stand productivity and biomass (remote sensing or field surveys) in the study area to verify the accuracy of the simulation results, and finally, providing an annual value dataset of the simulation results (stand productivity, biomass, ecosystem carbon budget) and forming a report text.

[0027] In the above-mentioned automatic generation system 100 for the artificial forest carbon sink monitoring report, the artificial forest carbon sink data extraction module 120 is used to extract the coniferous artificial forest carbon dioxide feature vector and the multi-modal association feature vector of the artificial forest growth state from the carbon dioxide values at different environmental heights collected by the carbon dioxide monitor, the growth environment data of the coniferous artificial forest, and the ecological productivity text data of the coniferous artificial forest. In this way, it can help researchers and managers better understand the carbon sink capacity of the artificial forest and the complexity of its growth state. This multi-dimensional information integration can provide more accurate carbon sink assessment and more effective forest management strategies, thus supporting the realization of sustainable forest management and climate change mitigation goals.

[0028] Figure 2 It is a block diagram of the artificial forest carbon sink data extraction module in the automatic generation system for the artificial forest carbon sink monitoring report according to the embodiment of the present application. As Figure 2 shown, in a specific embodiment of the present application, the artificial forest carbon sink data extraction module 120 includes: an artificial forest carbon dioxide feature extraction unit 121, which is used to perform feature extraction on the carbon dioxide values at different environmental heights collected by the carbon dioxide monitor to obtain the coniferous artificial forest carbon dioxide feature vector; an artificial forest growth environment feature extraction unit 122, which is used to perform feature extraction on the growth environment data of the coniferous artificial forest to obtain the artificial forest growth environment association feature vector; an ecological productivity text feature extraction unit 123, which is used to perform feature extraction on the ecological productivity text data of the coniferous artificial forest to obtain the artificial forest ecological productivity semantic association feature vector; and a growth state multi-modal feature association unit 124, which is used to associate the artificial forest growth environment association feature vector and the artificial forest ecological productivity semantic association feature vector to obtain the artificial forest growth state multi-modal association feature vector.

[0029] It should be understood that the spatial distribution of carbon dioxide concentration is a key factor in evaluating the forest carbon sink capacity. Since there may be significant differences in the carbon dioxide concentration at different heights within the forest, these differences directly affect the carbon storage and exchange processes. By performing feature extraction on these data, the carbon dioxide values at different heights can be transformed into meaningful feature vectors, capturing the carbon concentration characteristics at each height level. This helps to understand the specific situation of the carbon cycle within the forest, such as the differences in carbon absorption and release between the canopy layer and the ground layer. The feature extraction process can simplify the complex and raw carbon dioxide measurement values into an easily analyzable form by transforming the original data into feature vectors. Through feature extraction methods, such as the combination of convolutional layers and fully connected layers, key features representative of the forest carbon sink function can be extracted. This simplification not only helps to reduce the complexity of data processing but also improves the accuracy and efficiency of analysis.

[0030] Furthermore, the extraction of features from the growth environment data of coniferous plantations is to transform complex environmental information into a form that is convenient for analysis and application, in order to support in-depth research and management of forest ecosystems. This process helps to reveal the impact of environmental factors on the growth and health of plantations, thereby optimizing forest management strategies. The complexity of environmental data usually involves multiple variables, such as soil type, climate conditions, water content, etc. The interaction of these factors and their combined impact on forest growth are complex. Through feature extraction, these raw data can be simplified into representative feature vectors. These feature vectors condense the key information in the environmental data, making it easier to analyze and model, thus better understanding the impact of environmental factors on coniferous plantations.

[0031] Furthermore, the extraction of features from the text data of the ecological productivity of coniferous plantations is to transform text information into a numerical form with high information density and operability, so as to more effectively analyze and utilize ecological productivity data. This process is crucial for in-depth understanding of forest productivity, optimizing management strategies, and promoting scientific research. The complexity of text data usually contains a large amount of descriptive and quantitative information, such as tree growth rate, biomass, productivity indicators, etc. Since these data exist in the form of natural language or unstructured text, it may be very difficult to directly analyze and extract useful information. Through feature extraction, these raw text data can be transformed into structured feature vectors, from which key information can be refined and the hidden ecological productivity information in the text can be captured. Through natural language processing (NLP) techniques, such as word embedding models (Word Embeddings) or transformers, the semantic information in the text data can be transformed into vectors in a high-dimensional feature space. These vectors capture the semantic relationships and context information in the text, thus providing profound insights for analysis.

[0032] In particular, the association of the growth environment association feature vectors of plantations with the semantic association feature vectors of the ecological productivity of plantations is to comprehensively consider environmental factors and productivity data, so as to more comprehensively and accurately evaluate and understand the growth status of plantations. This multi-modal feature vector integrates information from different data sources, helping to provide more in-depth ecological analysis and management decision support. It should be understood that the relationship between the growth environment and ecological productivity is complex and multi-dimensional. Analyzing environmental data or productivity data alone often cannot fully reveal their mutual influence and combined effects. Among them, the growth environment association feature vectors include the external conditions for forest growth, such as climate, soil, and water, while the semantic association feature vectors of ecological productivity cover the results and performances of forest growth, such as biomass, stand productivity, tree growth rate, etc. By associating these two types of feature vectors, environmental conditions and growth results can be considered comprehensively, obtaining a more comprehensive description of the growth status, and further understanding the actual performances of forests under different environmental conditions.

[0033] In a specific embodiment of the present application, the artificial forest carbon dioxide feature extraction unit 121 includes: arranging the carbon dioxide values at different environmental heights collected by the carbon dioxide monitor into a coniferous artificial forest carbon dioxide input vector; passing the coniferous artificial forest carbon dioxide input vector through a coniferous artificial forest carbon dioxide feature encoder including a one-dimensional convolutional layer and a fully connected layer to obtain the coniferous artificial forest carbon dioxide feature vector.

[0034] It should be understood that the spatial distribution of carbon dioxide is of great significance in the forest ecosystem. The carbon dioxide concentrations at different heights reflect the carbon exchange conditions at different levels within the forest. There may be significant differences in the carbon dioxide concentrations in the canopy layer, sub-canopy layer, and surface layer, and these differences affect the carbon absorption and release processes of the forest. In practical applications, the original carbon dioxide measurement data is usually unstructured. Arranging it into an input vector can integrate the concentration data at different heights into a unified format. This arrangement form not only facilitates data storage and management but also lays a foundation for applying machine learning algorithms and data analysis models.

[0035] Furthermore, passing the coniferous artificial forest carbon dioxide input vector through a feature encoder including a one-dimensional convolutional layer and a fully connected layer is to effectively extract and represent the key features in the carbon dioxide data, thereby improving the performance of data analysis and model prediction. Among them, the role of the one-dimensional convolutional layer is to extract local features in the input vector. The one-dimensional convolutional layer slides through different parts of the data through convolutional operations and filters the data to capture the spatial patterns and structural features in the input vector. In the carbon dioxide monitoring data, such local features may include the concentration change trends at different height levels or the short-term fluctuations of the carbon dioxide concentration in the time series. The role of the fully connected layer is to further integrate and map the features extracted by the convolutional layer into a lower-dimensional feature space. The fully connected layer flattens the feature map output by the convolutional layer and performs a linear combination through a weight matrix and an activation function to finally obtain a compact feature vector. By using a deep learning model, such as a convolutional neural network, the feature encoder can automatically learn and extract the features most useful for analysis and decision-making, improving the prediction ability and accuracy of the model. Specifically, using the fully connected layer of the coniferous artificial forest carbon dioxide feature encoder including a one-dimensional convolutional layer and a fully connected layer to perform a fully connected encoding on the coniferous artificial forest carbon dioxide input vector to extract the high-dimensional implicit features of the feature values at each position in the coniferous artificial forest carbon dioxide input vector; and using the one-dimensional convolutional layer of the coniferous artificial forest carbon dioxide feature encoder including a one-dimensional convolutional layer and a fully connected layer to perform a one-dimensional encoding on the coniferous artificial forest carbon dioxide input vector to extract the high-dimensional implicit correlation features of the correlations between the feature values at each position in the coniferous artificial forest carbon dioxide input vector.

[0036] Figure 3 This is a block diagram of the artificial forest growth environment feature extraction unit in the artificial forest carbon sink monitoring report automatic generation system according to an embodiment of the present application. As Figure 3 shown, in a specific embodiment of the present application, the artificial forest growth environment feature extraction unit 122 includes: a growth environment data preprocessing subunit 1221, configured to perform data preprocessing on the coniferous artificial forest growth environment data to obtain an artificial forest growth environment data word feature matrix; a growth environment data convolutional coding subunit 1222, configured to pass the artificial forest growth environment data word feature matrix through an artificial forest growth environment data feature encoder based on a convolutional neural network to obtain the artificial forest growth environment association feature vector.

[0037] It should be understood that the necessity of data preprocessing lies in that the original growth environment data usually contains noise, missing values, and inconsistencies, which will affect the accuracy of analysis. Through data preprocessing, incorrect data can be cleaned, missing values can be filled, and the data can be standardized. This process makes the data more consistent and reliable, thus providing a clear basis for further analysis. Here, the growth environment data may include multiple features, such as climate conditions, soil types, and moisture content, etc. Converting these features into a word feature matrix can structure the complex environmental data into an easy-to-process matrix form. By encoding and vectorizing the environmental features, unstructured or semi-structured data can be converted into a numerical form, making it suitable for calculation and analysis.

[0038] Furthermore, processing the artificial forest growth environment data word feature matrix through a feature encoder based on a convolutional neural network is to automatically extract and represent the key features in the data through deep learning technology, thereby enhancing the ability and accuracy of data analysis. Convolutional neural networks are particularly suitable for processing and analyzing data with spatial or temporal structures, such as local patterns and features in growth environment data. Among them, the advantage of convolutional neural networks lies in their powerful feature extraction ability. Convolutional neural networks can capture local dependencies and spatial features in the data through one-dimensional or two-dimensional convolutional layers. When processing artificial forest growth environment data, convolutional neural networks can effectively identify feature patterns under different environmental conditions, such as temperature changes, humidity distributions, etc. The convolutional layer filters the feature matrix through a sliding convolutional kernel to extract representative features, which help to reveal the key patterns and relationships in the data. The automated nature of the convolutional operation makes the feature extraction process without manual feature design. In traditional data analysis, manual selection and construction of features are required, which may miss important information. Using a convolutional neural network can automatically learn and extract the most relevant features, thereby improving the comprehensiveness and accuracy of the analysis. Specifically, each layer of the artificial forest growth environment data feature encoder based on the convolutional neural network performs convolutional processing, mean pooling processing based on the local feature matrix, and non-linear activation processing on the input data respectively during the forward pass of the layer to output the artificial forest growth environment associated feature vector by the last layer of the artificial forest growth environment data feature encoder based on the convolutional neural network, where the input of the artificial forest growth environment data feature encoder based on the convolutional neural network is the artificial forest growth environment data word feature matrix.

[0039] In a specific embodiment of the present application, the growth environment data preprocessing subunit 1221 includes: passing the coniferous artificial forest growth environment data through an artificial forest growth environment word embedding module to obtain an artificial forest growth environment numerical word vector sequence; arranging the artificial forest growth environment numerical word vector sequence two-dimensionally to obtain the artificial forest growth environment data word feature matrix.

[0040] It should be understood that the basic concept of word embedding is to transform discrete and non-numerical data into continuous and dense vector representations. In the growth environment data, there may be multiple different features, such as temperature, humidity, soil type, etc. The original data of these features are usually categorical or have different dimensions. Through the word embedding module, these features can be transformed into low-dimensional and dense numerical vectors. The numerical vector of each feature can capture the similarities and relationships between features, making the data representation more meaningful and easy to process. Specifically, the coniferous plantation growth environment data is tokenized to obtain a growth environment word sequence; the embedding layer of the plantation growth environment word embedding module is used to map each growth environment word in the growth environment word sequence into a word embedding vector to obtain a plantation growth environment numerical word vector sequence.

[0041] Furthermore, the basic idea of two-dimensional arrangement is to reorganize the numerical information in a one-dimensional vector sequence into a two-dimensional matrix. This process usually includes partitioning or reshaping the vector sequence into a matrix form with fixed number of rows and columns. In this way, the structured information of the data can be expressed more clearly. Each row or column may correspond to specific environmental features or time points, thus retaining the local relationships and structural features of the data. The advantage of the two-dimensional feature matrix is its ability to capture spatial or temporal patterns in the data. For example, when processing coniferous plantation growth environment data, the two-dimensional matrix can arrange the values of different environmental variables (such as temperature, humidity, etc.) at different time points or different spatial positions in different positions of the matrix. In this way, the rows and columns of the matrix can represent different features and time steps, enabling the model to identify and learn the time series features and spatial distribution patterns in the environmental data.

[0042] In a specific embodiment of the present application, the ecological productivity text feature extraction unit 123 includes: obtaining multiple coniferous plantation ecological productivity text understanding feature vectors by passing the coniferous plantation ecological productivity text data through a coniferous plantation ecological productivity text editor; obtaining the coniferous plantation ecological productivity semantic association feature vectors by passing the multiple coniferous plantation ecological productivity text understanding feature vectors through a coniferous plantation ecological productivity multi-scale neighborhood feature extraction module based on sample dimensions.

[0043] It should be understood that the role of the text editor is to convert unstructured text data into structured feature representations. In the ecological productivity text, the descriptions usually involve different ecological indicators, environmental conditions, and productivity data. Through the text editor, these descriptions are parsed and encoded into feature vectors. Each feature vector captures the key information and patterns in the text, such as the productivity level, the impact of environmental factors, etc. By generating multiple feature vectors, the multi-level information and complex relationships in the text can be captured, thus providing richer context information for subsequent analysis. Specifically, the embedding layer of the plantation ecological productivity text editor is used to convert the features of the coniferous plantation ecological productivity text data into embedding vectors to obtain a sequence of ecological productivity embedding vectors; the Transformer-based Bert model of the plantation ecological productivity text editor is used to perform global context semantic encoding on the sequence of ecological productivity embedding vectors to obtain multiple plantation ecological productivity text understanding feature vectors.

[0044] Furthermore, the feature vectors of the text understanding of multiple plantation ecological productivity are processed by a multi-scale neighborhood feature extraction module based on the sample dimension in order to deeply explore and extract the potential semantic relationships and multi-level features in the text data. This process can capture the complex patterns and relationships in the text at different scales and ranges through the multi-scale neighborhood feature extraction technology, thereby improving the understanding and prediction capabilities of ecological productivity. Among them, the core of the multi-scale neighborhood feature extraction module is to analyze data from different scales and perspectives. This module can identify local and global patterns in different ranges by performing multi-scale analysis on feature vectors. For example, for the feature vectors in the ecological productivity text, multi-scale neighborhood analysis can capture the short-term and long-term relationships between features and identify the micro and macro factors that affect productivity. Such a method can more comprehensively understand the information in the text data and discover the potential correlation patterns therein. Feature extraction in the sample dimension makes the processing of feature vectors more detailed. By performing feature extraction on the sample dimension, the features of each sample can be deeply analyzed and the specific factors affecting productivity in each sample can be identified. In this way, not only can the global correlation features be extracted, but also the local information of each sample can be refined to obtain a more accurate feature representation. Specifically, each layer of the multi-scale neighborhood feature extraction module of artificial forest ecological productivity based on sample dimension performs the following operations on the input data in the forward transmission of the layer: performing convolution processing on the input data based on the first convolution kernel to obtain a first convolution feature map; performing convolution processing on the input data based on the second convolution kernel to obtain a second convolution feature map; performing convolution processing on the input data based on the third convolution kernel to obtain a third convolution feature map; performing convolution processing on the input data based on the fourth convolution kernel to obtain a fourth convolution feature map, wherein the first convolution kernel, the second convolution kernel, and the fourth convolution kernel are respectively used as the input data. The convolution kernel, the third convolution kernel and the fourth convolution kernel have different sizes; the first convolution feature map, the second convolution feature map, the third convolution feature map and the fourth convolution feature map are cascaded to obtain a multi-scale convolution feature map; the multi-scale convolution feature map is subjected to mean pooling processing along the channel dimension to obtain a pooled feature map; and the pooled feature map is subjected to nonlinear activation processing to obtain an activated feature map; wherein the output of the last layer of the convolution neural network model with a multi-scale convolution structure is the semantic association feature vector of the artificial forest ecological productivity.

[0045] In the above-mentioned automatic artificial forest carbon sink monitoring report generation system 100, the carbon sink monitoring report generation module 130 is used to generate an artificial forest carbon sink monitoring report based on the coniferous artificial forest carbon dioxide feature vector and the artificial forest growth state multi-modal association feature vector. It should be understood that the carbon dioxide feature vector reflects the concentration and absorption capacity of carbon dioxide in the artificial forest. Coniferous artificial forests absorb carbon dioxide through photosynthesis, and this process is crucial for mitigating climate change. The carbon dioxide feature vector provides data on the distribution of carbon dioxide at different times and in different spaces, helping to evaluate the carbon absorption and emission of the forest. By analyzing these features, the effectiveness of the forest carbon sink and potential improvement points can be identified. The artificial forest growth state multi-modal association feature vector integrates different growth information of the forest, including the growth rate, health status, coverage, etc. of the trees. These features can provide a comprehensive perspective on the health and productivity of the forest ecosystem. The association analysis of multi-modal data allows for a comprehensive consideration of the impact of the growth state on the carbon sink capacity, such as how tree density and growth rate affect the absorption and storage of carbon dioxide. Combining these two types of feature vectors can generate a comprehensive artificial forest carbon sink monitoring report. The report synthesizes various aspects of information on carbon dioxide absorption capacity and growth state, providing a comprehensive assessment of the forest carbon sink function.

[0046] Figure 4 It is a block diagram of the carbon sink monitoring report generation module in the automatic artificial forest carbon sink monitoring report generation system according to the embodiment of the present application. As Figure 4 shown, in a specific embodiment of the present application, the carbon sink monitoring report generation module 130 includes: a coniferous artificial forest carbon sink feature fusion unit 131, which is used to fuse the coniferous artificial forest carbon dioxide feature vector and the artificial forest growth state multi-modal association feature vector to obtain a coniferous artificial forest carbon sink monitoring generation feature vector; a coniferous artificial forest carbon sink feature optimization unit 132, which is used to perform topological space constraint guided by the target parameter space on the coniferous artificial forest carbon sink monitoring generation feature vector to obtain an optimized coniferous artificial forest carbon sink monitoring generation feature vector; and an artificial forest carbon sink monitoring report generation unit 133, which is used to generate an artificial forest carbon sink monitoring report by passing the optimized coniferous artificial forest carbon sink monitoring generation feature vector through a generator.

[0047] It should be understood that by fusing the coniferous artificial forest carbon dioxide feature vector and the artificial forest growth state multi-modal association feature vector, the carbon absorption capacity and growth state of the forest can be combined to form a more comprehensive feature representation. This comprehensive information can more accurately reflect the actual situation of the forest carbon sink, rather than relying solely on a single data source.

[0048] Specifically, in the technical solution of this application, the feature vector for coniferous plantation carbon sink monitoring is obtained by integrating information from multiple aspects such as carbon dioxide data, growth environment data, and ecological productivity data. Among them, the information from these multiple aspects is very high-dimensional and complex, and may contain many different and intertwined information. When the weights of the generator are adapted to the high-dimensional feature vector obtained from the information of these multiple aspects, it may not be able to accurately capture the uniqueness of each feature, resulting in overfitting some features or ignoring other features, thereby introducing interference in the generation process, and leading to class coherence interference in the generated coniferous plantation carbon sink monitoring report. Therefore, in the technical solution of this application, topological space constraints guided by the target parameter space are imposed on the feature vector for coniferous plantation carbon sink monitoring to obtain an optimized feature vector for coniferous plantation carbon sink monitoring.

[0049] Among them, imposing topological space constraints guided by the target parameter space on the feature vector for coniferous plantation carbon sink monitoring to obtain an optimized feature vector for coniferous plantation carbon sink monitoring includes: extracting the target parameter matrix for coniferous plantation carbon sink monitoring generated by the generator; performing nodal decomposition on the target parameter matrix for coniferous plantation carbon sink monitoring generated by taking row vectors as units to obtain a set of target parameter node encoding vectors for coniferous plantation carbon sink monitoring; using each target parameter node encoding vector in the set of target parameter node encoding vectors for coniferous plantation carbon sink monitoring as a wandering topological space, respectively imposing topological space constraints on the feature vector for coniferous plantation carbon sink monitoring to obtain a set of constrained feature vectors for coniferous plantation carbon sink monitoring; calculating the position-wise mean vector of the set of constrained feature vectors for coniferous plantation carbon sink monitoring to obtain the optimized feature vector for coniferous plantation carbon sink monitoring.

[0050] Among them, each target parameter node encoding vector in the set of target parameter node encoding vectors generated by the coniferous plantation carbon sink monitoring is used as a wandering topological space, and the topological space constraint is respectively performed on the coniferous plantation carbon sink monitoring generated feature vector to obtain a set of coniferous plantation carbon sink monitoring generated feature vectors after constraint, including: multiplying the coniferous plantation carbon sink monitoring generated feature vector by the transposed vector of the coniferous plantation carbon sink monitoring generated target parameter node encoding vector, and then calculating the natural exponential function value of the multiplication result to obtain the coniferous plantation carbon sink monitoring generated weighted exponential response weight; calculating the Euclidean distance between the coniferous plantation carbon sink monitoring generated feature vector and the coniferous plantation carbon sink monitoring generated target parameter node encoding vector to obtain the coniferous plantation carbon sink monitoring generated node encoding distance value; performing a dot product on the coniferous plantation carbon sink monitoring generated node encoding distance value and the coniferous plantation carbon sink monitoring generated feature vector, and calculating the natural exponential function value for each eigenvalue of the vector after the dot product to obtain the coniferous plantation carbon sink monitoring generated distance-guided exponential feature vector; performing a dot product on the coniferous plantation carbon sink monitoring generated weighted exponential response weight and the coniferous plantation carbon sink monitoring generated distance-guided exponential feature vector to obtain the coniferous plantation carbon sink monitoring generated feature vector after constraint.

[0051] In particular, considering that when constructing a coniferous plantation carbon sink monitoring system, it deals with multi-source heterogeneous data such as data from carbon dioxide monitors and growth environment data. Each data source has its unique information and structure. For example, the variation law of carbon dioxide concentration with time and height, and the historical records of the growth conditions of coniferous forests (such as soil type, moisture content). When extracting and fusing these different data features into a coniferous plantation carbon sink monitoring generated feature vector, the goal of this application is to create a feature vector that can not only reflect the current state of the forest ecosystem but also capture its long-term evolution pattern. This means that the fused data should not only accurately reflect the growth status of coniferous plantations under existing environmental conditions (such as instantaneous carbon dioxide concentration, soil humidity, and temperature), but also be able to reveal the stable features formed based on historical data accumulation. To ensure this, after the process of generating the coniferous plantation carbon sink monitoring generated feature vector, attention also needs to be paid to how to keep the main characteristics of the original data from being blurred. Based on this, in the technical solution of this application, the topological space constraint guided by the target parameter space is performed on the coniferous plantation carbon sink monitoring generated feature vector to obtain an optimized coniferous plantation carbon sink monitoring generated feature vector.

[0052] In the embodiments of the present application, specifically, the feature vectors generated for coniferous plantation carbon sink monitoring are subjected to topological space constraints guided by the target parameter space to obtain optimized feature vectors generated for coniferous plantation carbon sink monitoring, which are used for: processing the feature vectors generated for coniferous plantation carbon sink monitoring with the following optimization formula to obtain the optimized feature vectors generated for coniferous plantation carbon sink monitoring; wherein, the optimization formula is:

[0053]

[0054]

[0055]

[0056] Wherein, represents the target parameter matrix generated for coniferous plantation carbon sink monitoring, represents the 1st, 2nd, th, th target parameter node encoding vectors of the set of target parameter node encoding vectors generated for coniferous plantation carbon sink monitoring, represents the transpose of a vector, represents the natural exponential function, represents the feature vectors generated for coniferous plantation carbon sink monitoring, represents matrix multiplication, represents element-wise multiplication, represents calculating the vector and the vector, represents the th constrained feature vector generated for coniferous plantation carbon sink monitoring in the set of constrained feature vectors generated for coniferous plantation carbon sink monitoring, represents the total number of the set of constrained feature vectors generated for coniferous plantation carbon sink monitoring, represents the optimized feature vectors generated for coniferous plantation carbon sink monitoring.

[0057] That is, to address the above technical problems, in the technical solution of the present application, the feature vectors generated for coniferous plantation carbon sink monitoring are subjected to topological space constraints guided by the target parameter space. This process first extracts the key parameters for decision-making from the trained generator, and these parameters form a matrix in a high-dimensional space, where each row represents the weights or influencing factors in different dimensions. Through the target parameter matrix generated for coniferous plantation carbon sink monitoring, the position and shape of the model decision boundary can be understood, and thus it can be inferred which input features are most critical to the prediction results.

[0058] Next, the target parameter matrix generated for the coniferous plantation carbon sink monitoring is node - decomposed in units of row vectors to obtain a set of target parameter node - encoding vectors for the coniferous plantation carbon sink monitoring. Here, each row vector serves as a node in graph theory, meaning that each set of parameters is now regarded as an entity with potential connectivity. This transformation allows the application of methods from graph theory and network science to explore the interactions between features. The node - encoding vectors not only carry the information of the original parameters but also imply knowledge about the topology of the entire system. The node - decomposition further reveals the inherent connection patterns or structures in the data, enabling the optimized feature vectors to better adapt to new task requirements.

[0059] Then, using each target parameter node - encoding vector in the set of target parameter node - encoding vectors for the coniferous plantation carbon sink monitoring as a walking topological space, the topological space constraints are respectively imposed on the feature vectors generated for the coniferous plantation carbon sink monitoring to obtain a set of feature vectors generated for the coniferous plantation carbon sink monitoring after constraint. Using the topological space defined by the node - encoding vectors for "walking" is actually simulating an exploratory process aimed at finding those feature transformations that can best preserve the characteristics of the original data structure. Each step is determined by the probability distribution of the current state for the next position. The topological space constraints ensure that even in different contexts, the feature representation still retains certain invariance. At the same time, it can also promote cross - domain transfer learning because it emphasizes the general relationships between features rather than the details of a specific domain. In this way, the reconstruction of the feature space is achieved, making the optimized feature vectors more compact and having better generalization ability.

[0060] Finally, the position - wise mean vector of the set of feature vectors generated for the coniferous plantation carbon sink monitoring after constraint is calculated to obtain the optimized feature vectors generated for the coniferous plantation carbon sink monitoring. Calculating the mean vector is a statistical aggregation method for integrating the optimal solutions from multiple perspectives. The idea behind this step is to reduce the bias caused by a single estimate by fusing the information provided by different sample points. The averaging process is equivalent to performing a Soft Voting, enhancing the expressiveness of the common features and making the optimized feature vectors more stable and reliable.

[0061] Furthermore, the role of the generator is to convert the optimized feature vectors generated for coniferous plantation carbon sink monitoring into a structured report. The optimized feature vectors generated for coniferous plantation carbon sink monitoring usually contain a large amount of numerical data and complex pattern information. For non-professional users, it may be difficult to directly interpret these data. By mapping the optimized feature vectors generated for coniferous plantation carbon sink monitoring into a specific report template, the generator can present the data in the form of charts, images, and text descriptions, making the report content more intuitive and understandable. In a specific embodiment of the present application, the coniferous plantation carbon sink monitoring report includes stand productivity, biomass, and ecosystem carbon budget.

[0062] In summary, the embodiment of the present application first obtains the carbon dioxide values at different environmental heights, the growth environment data of coniferous plantations, and the text data of the ecological productivity of coniferous plantations collected by a carbon dioxide monitor, then uses deep learning technology to perform feature extraction and correlation analysis on the three, and finally generates a coniferous plantation carbon sink monitoring report through a generator, so as to provide more accurate and effective support for coniferous plantation management and environmental protection.

[0063] As described above, the coniferous plantation carbon sink monitoring report automatic generation system 100 according to the embodiment of the present application can be implemented in various terminal devices. In one example, the coniferous plantation carbon sink monitoring report automatic generation system 100 can be integrated into the terminal device as a software module and / or a hardware module. For example, the coniferous plantation carbon sink monitoring report automatic generation system 100 can be a software module in the operating system of the terminal device, or can be an application program developed for the terminal device; of course, the coniferous plantation carbon sink monitoring report automatic generation system 100 can also be one of the many hardware modules of the terminal device.

[0064] Alternatively, in another example, the coniferous plantation carbon sink monitoring report automatic generation system 100 and the terminal device can also be separate devices, and the coniferous plantation carbon sink monitoring report automatic generation system 100 can be connected to the terminal device through a wired and / or wireless network and transmit interaction information according to a predefined data format.

[0065] Figure 5 It is a flowchart of the coniferous plantation carbon sink monitoring report automatic generation method according to the embodiment of the present application. As Figure 5As shown, a method for automatically generating an artificial forest carbon sink monitoring report according to an embodiment of the present application includes: S110, obtaining carbon dioxide values at different environmental heights, coniferous artificial forest growth environment data, and coniferous artificial forest ecological productivity text data collected by a carbon dioxide monitor; S120, extracting a coniferous artificial forest carbon dioxide feature vector and an artificial forest growth state multimodal association feature vector from the carbon dioxide values at different environmental heights, the coniferous artificial forest growth environment data, and the coniferous artificial forest ecological productivity text data collected by the carbon dioxide monitor; S130, generating an artificial forest carbon sink monitoring report based on the coniferous artificial forest carbon dioxide feature vector and the artificial forest growth state multimodal association feature vector.

[0066] Here, those skilled in the art can understand that the specific operations of each step in the above method for automatically generating an artificial forest carbon sink monitoring report have been described in detail in the description of the artificial forest carbon sink monitoring report automatic generation system above with reference to Figures 1 to 4 and therefore, the repeated description thereof will be omitted.

[0067] Next, reference is made to Figure 6 to describe an electronic device according to an embodiment of the present application.

[0068] As Figure 6 shown, the electronic device 10 includes an input device 11, an input interface 12, a central processing unit 13, a memory 14, an output interface 15, an output device 16, and a bus 17. Among them, the input interface 12, the central processing unit 13, the memory 14, and the output interface 15 are connected to each other through the bus 17, and the input device 11 and the output device 16 are respectively connected to the bus 17 through the input interface 14 and the output interface 15, and then connected to other components of the electronic device 10.

[0069] Specifically, the input device 11 receives input information from the outside and transmits the input information to the central processing unit 13 through the input interface 12; the central processing unit 13 processes the input information based on computer-executable instructions stored in the memory 14 to generate output information, temporarily or permanently stores the output information in the memory 14, and then transmits the output information to the output device 16 through the output interface 15; the output device 16 outputs the output information to the outside of the electronic device 10 for user use.

[0070] In one embodiment, Figure 6 the electronic device 10 shown can be implemented as a network device, and the network device may include: a memory configured to store a program; a processor configured to run the program stored in the memory to execute any one of the methods for automatically generating an artificial forest carbon sink monitoring report described in the above embodiments.

[0071] According to an embodiment of the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, an embodiment of the present application includes a computer program product that includes a computer program tangibly embodied on a machine-readable medium, the computer program including program code for performing the methods shown in the flowcharts. In such an embodiment, the computer program can be downloaded and installed from a network, and / or installed from a removable storage medium.

[0072] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, and appropriate combinations thereof. In a hardware implementation, the division of the functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, one physical component can have multiple functions, or one function or step can be executed by several physical components in cooperation. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include a computer storage medium (or non-transitory medium) and a communication medium (or transitory medium). As is well known to those of ordinary skill in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassette, tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer.

[0073] It can be understood that the above embodiments are merely exemplary embodiments adopted to illustrate the principles of the present application, and the present application is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present application, and these modifications and improvements are also considered within the protection scope of the present application.

Claims

1. A system for automatically generating plantation carbon sink monitoring reports, characterized in that: include: The plantation carbon sink data acquisition module is used to obtain the carbon dioxide values ​​at different environmental heights collected by the carbon dioxide monitor, the coniferous plantation growth environment data and the coniferous plantation ecological productivity text data; A plantation carbon sink data extraction module is used to extract a coniferous plantation carbon dioxide feature vector and a plantation growth state multimodal correlation feature vector from the carbon dioxide values ​​at different environmental heights collected by the carbon dioxide monitor, the coniferous plantation growth environment data, and the coniferous plantation ecological productivity text data; A carbon sink monitoring report generating module, used for generating a plantation carbon sink monitoring report based on the coniferous plantation carbon dioxide feature vector and the plantation growth state multimodal correlation feature vector; Wherein, the carbon sink monitoring report generation module includes: A coniferous plantation carbon sink feature fusion unit, used for fusing the coniferous plantation carbon dioxide feature vector and the plantation growth state multimodal correlation feature vector to obtain a coniferous plantation carbon sink monitoring generation feature vector; A coniferous plantation carbon sink characteristic optimization unit, used for performing a target parameter space-oriented topological space constraint on the coniferous plantation carbon sink monitoring and generating characteristic vector to obtain an optimized coniferous plantation carbon sink monitoring and generating characteristic vector; A plantation forest carbon sink monitoring report generating unit, used for passing the optimized coniferous plantation forest carbon sink monitoring generating feature vector through a generator to generate a plantation forest carbon sink monitoring report; Wherein, the coniferous plantation carbon sink characteristic optimization unit is used for: Extracting a coniferous plantation carbon sink monitoring generation target parameter matrix of the generator; Decomposing the coniferous plantation carbon sink monitoring generation target parameter matrix into nodes in units of row vectors to obtain a set of coniferous plantation carbon sink monitoring generation target parameter node encoding vectors; Taking each coniferous plantation carbon sink monitoring target parameter node encoding vector in the set of coniferous plantation carbon sink monitoring target parameter node encoding vectors as the walking topological space, topological space constraints are respectively performed on the coniferous plantation carbon sink monitoring feature vectors to obtain a set of coniferous plantation carbon sink monitoring feature vectors after constraints; The positional mean vector of the set of the constrained coniferous plantation carbon sink monitoring generated feature vectors is calculated to obtain the optimized coniferous plantation carbon sink monitoring generated feature vectors.

2. The automatic generation system of plantation carbon sink monitoring report according to claim 1 is characterized in that: The plantation carbon sink data extraction module comprises: A plantation carbon dioxide feature extraction unit, used for extracting features from the carbon dioxide values ​​at different environmental heights collected by the carbon dioxide monitor to obtain a carbon dioxide feature vector for the coniferous plantation; A plantation forest growth environment feature extraction unit, used for extracting features from the coniferous plantation forest growth environment data to obtain a plantation forest growth environment-related feature vector; An ecological productivity text feature extraction unit, used for extracting features from the coniferous plantation ecological productivity text data to obtain a semantically associated feature vector of the plantation ecological productivity; The growth state multimodal feature association unit is used to associate the artificial forest growth environment association feature vector with the artificial forest ecological productivity semantic association feature vector to obtain the artificial forest growth state multimodal association feature vector.

3. The automatic generation system of plantation carbon sink monitoring report according to claim 2 is characterized in that: The plantation carbon dioxide feature extraction unit comprises: Arranging the carbon dioxide values ​​at different environmental heights collected by the carbon dioxide monitor as a coniferous plantation carbon dioxide input vector; The coniferous forest carbon dioxide input vector is passed through a coniferous forest carbon dioxide feature encoder comprising a one-dimensional convolutional layer and a fully connected layer to obtain the coniferous forest carbon dioxide feature vector.

4. The automatic generation system of plantation carbon sink monitoring report according to claim 3 is characterized in that: The artificial forest growth environment feature extraction unit comprises: A growth environment data preprocessing subunit, used for preprocessing the coniferous plantation growth environment data to obtain a plantation growth environment data word feature matrix; The growth environment data convolution encoding subunit is used to obtain the artificial forest growth environment associated feature vector by passing the artificial forest growth environment data word feature matrix through an artificial forest growth environment data feature encoder based on a convolutional neural network.

5. The automatic generation system of plantation carbon sink monitoring report according to claim 4 is characterized in that: The growth environment data preprocessing subunit includes: The coniferous plantation growth environment data is passed through a plantation growth environment word embedding module to obtain a plantation growth environment word vector sequence; The artificial forest growth environment numeral word vector sequence is arranged in two dimensions to obtain the artificial forest growth environment data word feature matrix.

6. The automatic generation system of plantation carbon sink monitoring report according to claim 5 is characterized in that: The ecological productivity text feature extraction unit comprises: The coniferous plantation ecological productivity text data is passed through a plantation ecological productivity text editor to obtain a plurality of plantation ecological productivity text understanding feature vectors; The multiple plantation forest ecological productivity text comprehension feature vectors are passed through a sample dimension-based plantation forest ecological productivity multi-scale neighborhood feature extraction module to obtain the plantation forest ecological productivity semantic association feature vector.

7. The automatic generation system of plantation carbon sink monitoring report according to claim 6 is characterized in that: Taking each coniferous plantation carbon sink monitoring target parameter node encoding vector in the set of the coniferous plantation carbon sink monitoring target parameter node encoding vector as the walking topological space, topological space constraints are respectively performed on the coniferous plantation carbon sink monitoring feature vector to obtain a set of constrained coniferous plantation carbon sink monitoring feature vectors, including: After multiplying the coniferous plantation carbon sink monitoring generation feature vector and the transposed vector of the coniferous plantation carbon sink monitoring generation target parameter node encoding vector, calculating the natural exponential function value of the multiplication result to obtain the coniferous plantation carbon sink monitoring generation weighted index response weight; Calculating the Euclidean distance between the coniferous plantation carbon sink monitoring generation feature vector and the coniferous plantation carbon sink monitoring generation target parameter node coding vector to obtain the coniferous plantation carbon sink monitoring generation node coding distance value; Performing a dot multiplication on the coniferous plantation carbon sink monitoring generation node encoding distance value and the coniferous plantation carbon sink monitoring generation characteristic vector, and calculating a natural exponential function value for each eigenvalue of the vector after the dot multiplication to obtain a coniferous plantation carbon sink monitoring generation distance guidance index characteristic vector; The response weight of the weighted index generated by the coniferous plantation carbon sink monitoring and the characteristic vector of the distance guidance index generated by the coniferous plantation carbon sink monitoring are point-multiplied to obtain the constrained coniferous plantation carbon sink monitoring and characteristic vector.

8. A method for automatically generating a plantation carbon sink monitoring report, characterized in that: include: Obtain the carbon dioxide values ​​at different environmental altitudes collected by the carbon dioxide monitor, the growth environment data of coniferous plantations, and the text data of the ecological productivity of coniferous plantations; Extracting a coniferous plantation carbon dioxide feature vector and a multimodal correlation feature vector of the plantation growth state from the carbon dioxide values ​​at different environmental heights collected by the carbon dioxide monitor, the coniferous plantation growth environment data, and the coniferous plantation ecological productivity text data; Generate a plantation carbon sink monitoring report based on the coniferous plantation carbon dioxide feature vector and the multimodal correlation feature vector of the plantation growth state; Wherein, based on the coniferous plantation carbon dioxide feature vector and the multimodal correlation feature vector of the plantation growth state, a plantation carbon sink monitoring report is generated, including: The coniferous plantation carbon dioxide feature vector and the plantation growth state multimodal correlation feature vector are merged to obtain a coniferous plantation carbon sink monitoring generation feature vector; Performing target parameter space-oriented topological space constraints on the coniferous plantation carbon sink monitoring generated feature vector to obtain an optimized coniferous plantation carbon sink monitoring generated feature vector; Passing the optimized coniferous plantation carbon sink monitoring generated feature vector through a generator to generate a plantation carbon sink monitoring report; The method of performing a target parameter space-oriented topological space constraint on the coniferous plantation carbon sink monitoring generated feature vector to obtain an optimized coniferous plantation carbon sink monitoring generated feature vector includes: Extracting a coniferous plantation carbon sink monitoring generation target parameter matrix of the generator; Decomposing the coniferous plantation carbon sink monitoring generation target parameter matrix into nodes in units of row vectors to obtain a set of coniferous plantation carbon sink monitoring generation target parameter node encoding vectors; Taking each coniferous plantation carbon sink monitoring target parameter node encoding vector in the set of coniferous plantation carbon sink monitoring target parameter node encoding vectors as the walking topological space, topological space constraints are respectively performed on the coniferous plantation carbon sink monitoring feature vectors to obtain a set of coniferous plantation carbon sink monitoring feature vectors after constraints; The positional mean vector of the set of the constrained coniferous plantation carbon sink monitoring generated feature vectors is calculated to obtain the optimized coniferous plantation carbon sink monitoring generated feature vectors.

Images