Saline-alkali soil carbon sink estimation system based on machine learning
Through a machine learning-based saline-alkali land carbon sink estimation system, integrating multiple data sources and building a recurrent neural network model, the problem of complexity of saline-alkali land carbon sink estimation is solved, and more accurate and efficient carbon sink assessment is achieved, supporting ecological restoration and land Management decision making.
Patent Information
- Application Number
- CN202411056046.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-02
- Publication Date
- 2024-11-12
AI Technical Summary
Due to its unique soil chemical and physical properties, traditional carbon storage estimation methods are not applicable, making it difficult to accurately evaluate its carbon sink potential and dynamic changes.
Using a saline-alkali land carbon sink estimation system based on machine learning, including data preparation, feature processing and model calculation modules, the physical and chemical properties of soil and vegetation, meteorological data, and hydrological data are integrated, and the recurrent neural network model is constructed to screen key features. And conduct carbon sink estimates.
It improves the accuracy and efficiency of carbon sink estimation, reduces manual intervention, provides visual analysis of dynamic changes in carbon sinks, and enhances the scientific nature of ecological value assessment and land management.
Smart Images

Figure CN118940960A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of carbon sink calculation in saline-alkali land, and in particular relates to a saline-alkali land carbon sink estimation system based on machine learning. Background Art
[0002] As global climate change intensifies, finding effective carbon sinks has become the key to mitigating the greenhouse effect. As a potential carbon sink, saline-alkali land helps reduce the concentration of greenhouse gases in the atmosphere and combat global warming by absorbing carbon dioxide in the atmosphere. Saline-alkali land is widely distributed around the world. Through accurate carbon sink estimation, we can better assess the ecological value and improvement potential of these lands, and provide a scientific basis for the sustainable management and utilization of land. Accurate carbon sink estimation helps to understand the role of saline-alkali land in the ecosystem, guide ecological restoration work, improve soil quality, enhance biodiversity, and provide data support for the formulation of relevant environmental protection policies.
[0003] Although saline-alkali land has great carbon sink potential, it also faces challenges in estimation. Due to its high salt content, saline-alkali land has unique soil chemical and physical properties, which affect its carbon storage and release process. The carbon cycle process of saline-alkali land is relatively complex, involving multiple interacting systems such as soil, plants, and microorganisms. Moreover, the carbon storage of saline-alkali land is not static, but changes dynamically with time and environmental conditions. For example, changes in hydrological conditions may affect the fixation and release of soil carbon.
[0004] Due to the particularity of saline-alkali land, traditional soil carbon stock estimation methods are not fully applicable, and more accurate measurement techniques and models need to be developed and applied. Summary of the invention
[0005] In order to solve the above technical problems, the present invention proposes a saline-alkali land carbon sink estimation system based on machine learning to solve the problems existing in the above-mentioned prior art.
[0006] To achieve the above object, the present invention provides a saline-alkali land carbon sink estimation system based on machine learning, comprising:
[0007] Data preparation module, feature processing module, model calculation module, and display module;
[0008] The data preparation module is used to obtain remote sensing image data, physical and chemical properties of soil and vegetation, meteorological data, hydrological data and human data of the target saline-alkali land area;
[0009] The feature processing module is used to extract key features;
[0010] The model calculation module is used to construct a saline-alkali land carbon sink estimation model, and obtain a carbon sink estimation value based on key characteristics and the saline-alkali land carbon sink estimation model;
[0011] The display module is used to display the carbon sink capacity and its changing trend in the form of charts.
[0012] Optionally, the data preparation module includes a remote sensing data preparation unit, a ground survey data acquisition unit, and an environmental data acquisition unit;
[0013] The remote sensing data preparation unit is used to obtain remote sensing image data and perform region division based on the remote sensing image data;
[0014] The ground survey data collection unit performs on-site sampling according to the regional division results, obtains saline-alkali land soil samples and plant samples, and determines the physical and chemical properties of the soil and vegetation;
[0015] The environmental data acquisition unit acquires meteorological data, hydrological data and human data of the target saline-alkali land area.
[0016] Optionally, the remote sensing data preparation unit includes a data acquisition subunit and a region division unit;
[0017] The data acquisition subunit is used to obtain high-resolution remote sensing image data of the target saline-alkali land area through a satellite, and pre-process the remote sensing data;
[0018] The region division unit is used to segment the pre-processed remote sensing image data into regions according to the spectral characteristics, texture characteristics and spatial distribution characteristics of the saline-alkali land using image segmentation technology; a classification algorithm is used to classify and mark the segmented regions and to calculate the distribution statistics.
[0019] Optionally, the ground survey data collection unit determines the sample collection points of soil and vegetation in each area according to the regional classification marks and distribution, and combines the data measured by the samples according to the regional classification marks to obtain multiple data sets.
[0020] Optionally, the feature processing module includes a data preprocessing unit and a feature extraction unit;
[0021] The preprocessing unit is used to perform data cleaning and normalization on the ground survey data and environmental data to form a new data set and store it;
[0022] The feature extraction unit is used to screen and obtain key features from the physical and chemical properties of soil and vegetation, meteorological data, hydrological data, human data and remote sensing image data.
[0023] Optionally, the feature extraction unit selects features that meet the correlation coefficient requirements as candidate features by analyzing the correlation between the physical and chemical properties of soil and vegetation and meteorological data, hydrological data, anthropogenic data and carbon storage; uses a statistical algorithm to evaluate the contribution of each candidate feature to carbon sink estimation, and screens out key features from the candidate features based on the contribution value, wherein the key features include but are not limited to vegetation coverage, soil type, soil moisture extracted from remote sensing image data, soil pH, conductivity, salinity, vegetation biomass, vegetation carbon content in ground survey data; temperature, precipitation, tillage method, and irrigation amount in environmental data.
[0024] Optionally, the model calculation module includes a biomass estimation unit, a model construction unit, and a carbon sink estimation unit;
[0025] The biomass estimation unit is used to obtain biomass estimation values based on key features of soil and vegetation physical and chemical properties and high-resolution remote sensing image data;
[0026] The model building unit is used to build a recurrent neural network model, and introduce layer normalization and residual connection to obtain a saline-alkali land carbon sink estimation model;
[0027] The carbon sink estimation unit is used to obtain a carbon sink estimation value through a saline-alkali land carbon sink estimation model.
[0028] Optionally, the biomass estimation unit is used to classify the vegetation in each area of the saline-alkali land based on high-resolution remote sensing image data, and determine the vegetation type of each block; extract the vegetation index from the high-resolution remote sensing image data based on the vegetation type; and obtain the biomass estimation value through a pre-trained biomass estimation model based on the vegetation index and ground survey data.
[0029] Optionally, the carbon sink estimation unit is used to extract data corresponding to key features, and input the extracted data, corresponding vegetation types, and biological estimates into a saline-alkali land carbon sink estimation model to obtain a carbon sink estimation value.
[0030] Compared with the prior art, the present invention has the following advantages and technical effects:
[0031] The present invention integrates remote sensing image data, physical and chemical properties of soil and vegetation, meteorological data, hydrological data and human data through the data preparation module, ensuring the comprehensiveness and reliability of the estimation results; the feature processing module screens out key features by analyzing the correlation between different data and carbon reserves, thereby enhancing the prediction accuracy and scientificity of the model; the model calculation module adopts a recurrent neural network model, and introduces layer normalization and residual connection to improve the generalization ability and stability of the model, ensuring the accuracy of the estimation results; the system uses machine learning technology to realize the automation of carbon sink estimation, reduce manual intervention, and improve estimation efficiency and consistency; the display module intuitively displays the carbon sink capacity and its changing trend in the form of charts, which is convenient for analysis and decision-making, and enhances the comprehensibility and practicality of the results. The present invention significantly improves the accuracy and efficiency of carbon sink estimation by integrating remote sensing technology, ground surveys, environmental monitoring and machine learning algorithms, which helps to better evaluate the ecological value and improvement potential of saline-alkali land, and provides a scientific basis for the sustainable management and utilization of land. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0033] Figure 1 Schematic diagram of the system structure of an embodiment of the present invention. DETAILED DESCRIPTION
[0034] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0035] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0036] Embodiment 1
[0037] like Figure 1 As shown, this embodiment provides a saline-alkali land carbon sink estimation system based on machine learning, including:
[0038] Data preparation module, feature processing module, model calculation module, and display module;
[0039] The data preparation module is used to obtain remote sensing image data, physical and chemical properties of soil and vegetation, meteorological data, hydrological data and human data of the target saline-alkali land area; the human data includes farming methods, artificial irrigation water volume, etc.
[0040] As a specific implementation, the data preparation module includes a remote sensing data preparation unit, a ground survey data collection unit, and an environmental data collection unit;
[0041] The remote sensing data preparation unit is used to obtain remote sensing image data and perform region division based on the remote sensing image data;
[0042] The ground survey data collection unit performs on-site sampling according to the regional division results, obtains saline-alkali land soil samples and plant samples, and determines the physical and chemical properties of the soil and vegetation;
[0043] The environmental data acquisition unit acquires meteorological data, hydrological data and human data of the target saline-alkali land area.
[0044] As a specific implementation, the remote sensing data preparation unit includes a data acquisition subunit and a region division unit;
[0045] The data acquisition subunit is used to obtain high-resolution remote sensing image data of the target saline-alkali land area through a satellite, and pre-process the remote sensing data;
[0046] The region division unit is used to segment the pre-processed remote sensing image data into regions according to the spectral characteristics, texture characteristics and spatial distribution characteristics of the saline-alkali land using image segmentation technology; a classification algorithm is used to classify and mark the segmented regions and to calculate the distribution statistics.
[0047] As a specific implementation, the ground survey data collection unit determines the sample collection points of soil and vegetation in each region according to the regional classification marks and distribution conditions, and combines the data measured by the samples according to the regional classification marks to obtain multiple data sets.
[0048] The feature processing module is used to extract key features;
[0049] As a specific implementation, the feature processing module includes a data preprocessing unit and a feature extraction unit;
[0050] The preprocessing unit is used to perform data cleaning and normalization on the ground survey data and environmental data to form a new data set and store it;
[0051] The feature extraction unit is used to screen and obtain key features from the physical and chemical properties of soil and vegetation, meteorological data, hydrological data, human data and remote sensing image data.
[0052] As a specific implementation method, the feature extraction unit selects features that meet the correlation coefficient requirements as candidate features by analyzing the correlation between the physical and chemical properties of soil and vegetation and meteorological data, hydrological data, anthropogenic data and carbon storage; uses a statistical algorithm to evaluate the contribution of each candidate feature to carbon sink estimation, and selects key features from the candidate features according to the contribution value, wherein the key features include but are not limited to vegetation coverage, soil type, soil moisture extracted from remote sensing image data, soil pH, conductivity, salinity, vegetation biomass, vegetation carbon content in ground survey data; temperature, precipitation, tillage method, and irrigation amount in environmental data.
[0053] The model calculation module is used to construct a saline-alkali land carbon sink estimation model, and obtain a carbon sink estimation value based on key characteristics and the saline-alkali land carbon sink estimation model;
[0054] Specifically, there are multiple servers storing the soil organic carbon estimation model, and the business load status of the multiple servers is obtained; if the server with the lowest business load status is the current server, based on the soil organic carbon estimation model stored in the current server that matches the identification information of the target saline-alkali land area, the extracted data is processed, and the carbon sink estimation value is output through the model.
[0055] As a specific implementation, the model calculation module includes a biomass estimation unit, a model construction unit, and a carbon sink estimation unit;
[0056] The biomass estimation unit is used to obtain biomass estimation values based on key features of soil and vegetation physical and chemical properties and high-resolution remote sensing image data;
[0057] The model building unit is used to build a recurrent neural network model, and introduce layer normalization and residual connection to obtain a saline-alkali land carbon sink estimation model;
[0058] The carbon sink estimation unit is used to obtain a carbon sink estimation value through a saline-alkali land carbon sink estimation model.
[0059] As a specific implementation, the biomass estimation unit is used to classify the vegetation in each area of the saline-alkali land based on high-resolution remote sensing image data, and determine the vegetation type of each block; extract the vegetation index from the high-resolution remote sensing image data based on the vegetation type; based on the vegetation index and ground survey data, obtain the biomass estimation value through a pre-trained biomass estimation model.
[0060] As a specific implementation, the carbon sink estimation unit is used to extract data corresponding to key features, and input the extracted data, corresponding vegetation types, and biological estimates into a saline-alkali land carbon sink estimation model to obtain a carbon sink estimation value.
[0061] Specifically, the time series and spatial distribution analysis of carbon storage data can be performed to evaluate the dynamic changes of saline-alkali land carbon sinks and predict future carbon sink potential. Time series analysis includes seasonal change analysis and interannual change analysis;
[0062] Seasonal variation analysis: Analyze the variation patterns of soil organic carbon and carbonate in different seasons (such as spring, summer, autumn, and winter). For example, the soil organic carbon in saline-alkali paddy fields in western Jilin Province shows a seasonal variation pattern of "one decrease, one increase, and one decrease".
[0063] Interannual variation analysis: Compare the changes in saline-alkali land carbon storage in different years, especially the impact of development years on carbon storage. For example, the longer the paddy field soil in western Jilin has been developed, the higher the total carbon and organic carbon content of the soil, and the lower the carbonate content.
[0064] The display module is used to display the carbon sink capacity and its changing trend in the form of charts.
[0065] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A saline-alkali land carbon sink estimation system based on machine learning, characterized in that: The following steps are involved: Data preparation module, feature processing module, model calculation module, and display module; The data preparation module is used to obtain remote sensing image data, physical and chemical properties of soil and vegetation, meteorological data, hydrological data and human data of the target saline-alkali land area; The feature processing module is used to extract key features; The model calculation module is used to construct a saline-alkali land carbon sink estimation model, and obtain a carbon sink estimation value based on key characteristics and the saline-alkali land carbon sink estimation model; The display module is used to display the carbon sink capacity and its changing trend in the form of charts.
2. The saline-alkali land carbon sink estimation system based on machine learning according to claim 1 is characterized in that: The data preparation module includes a remote sensing data preparation unit, a ground survey data acquisition unit, and an environmental data acquisition unit; The remote sensing data preparation unit is used to obtain remote sensing image data and perform region division based on the remote sensing image data; The ground survey data collection unit performs on-site sampling according to the regional division results, obtains saline-alkali land soil samples and plant samples, and determines the physical and chemical properties of the soil and vegetation; The environmental data acquisition unit acquires meteorological data, hydrological data and human data of the target saline-alkali land area.
3. The saline-alkali land carbon sink estimation system based on machine learning according to claim 2 is characterized in that: The remote sensing data preparation unit includes a data acquisition subunit and a region division unit; The data acquisition subunit is used to obtain high-resolution remote sensing image data of the target saline-alkali land area through a satellite, and pre-process the remote sensing data; The region division unit is used to segment the pre-processed remote sensing image data into regions according to the spectral characteristics, texture characteristics and spatial distribution characteristics of the saline-alkali land using image segmentation technology; a classification algorithm is used to classify and mark the segmented regions and to calculate the distribution statistics.
4. The saline-alkali land carbon sink estimation system based on machine learning according to claim 3 is characterized in that: The ground survey data collection unit determines the sample collection points of soil and vegetation in each region according to the regional classification marks and distribution conditions, and combines the data measured by the samples according to the regional classification marks to obtain multiple data sets.
5. The saline-alkali land carbon sink estimation system based on machine learning according to claim 1, characterized in that: The feature processing module includes a data preprocessing unit and a feature extraction unit; The preprocessing unit is used to perform data cleaning and normalization on the ground survey data and environmental data to form a new data set and store it; The feature extraction unit is used to screen and obtain key features from the physical and chemical properties of soil and vegetation, meteorological data, hydrological data, human data and remote sensing image data.
6. The saline-alkali land carbon sink estimation system based on machine learning according to claim 5 is characterized in that: The feature extraction unit selects features that meet the correlation coefficient requirements as candidate features by analyzing the correlation between the physical and chemical properties of soil and vegetation and meteorological data, hydrological data, anthropogenic data and carbon storage; Statistical algorithms are used to evaluate the contribution of each candidate feature to carbon sink estimation, and key features are screened out from the candidate features based on the contribution value. Key features include but are not limited to vegetation coverage, soil type, soil moisture extracted from remote sensing image data, soil pH, conductivity, salinity, vegetation biomass, and vegetation carbon content in ground survey data; and temperature, precipitation, tillage methods, and irrigation amount in environmental data.
7. The saline-alkali land carbon sink estimation system based on machine learning according to claim 1, characterized in that: The model calculation module includes a biomass estimation unit, a model construction unit, and a carbon sink estimation unit; The biomass estimation unit is used to obtain biomass estimation values based on key features of soil and vegetation physical and chemical properties and high-resolution remote sensing image data; The model building unit is used to build a recurrent neural network model, and introduce layer normalization and residual connection to obtain a saline-alkali land carbon sink estimation model; The carbon sink estimation unit is used to obtain a carbon sink estimation value through a saline-alkali land carbon sink estimation model.
8. The saline-alkali land carbon sink estimation system based on machine learning according to claim 7, characterized in that: The biomass estimation unit is used to classify the vegetation in each area of the saline-alkali land based on high-resolution remote sensing image data and determine the vegetation type of each block; extract the vegetation index from the high-resolution remote sensing image data based on the vegetation type; and obtain the biomass estimation value through a pre-trained biomass estimation model based on the vegetation index and ground survey data.
9. The saline-alkali land carbon sink estimation system based on machine learning according to claim 8, characterized in that: The carbon sink estimation unit is used to extract data corresponding to key features, and input the extracted data, corresponding vegetation types, and biological estimation quantities into a saline-alkali land carbon sink estimation model to obtain a carbon sink estimation value.
Citation Information
Cited By
Ecological product carbon sink value accounting method
CN119226722A
A method for calculating the carbon sink value of ecological products
CN119226722B
Carbon sink data analysis method and system for coal mining subsidence land
CN119849766A
Crop carbon sink calculation model, crop carbon credit calculation method and storage medium
CN120509906A
