A blockchain-based carbon sink transaction method

Abstract

This invention relates to the field of digital asset management technology and discloses a blockchain-based carbon sink trading method, comprising: acquiring environmental data and forest growth data of afforestation areas, performing edge processing to generate structured feature data, dynamically calculating carbon sink volume on the structured feature data, generating carbon sink calculation result data; uploading the carbon sink calculation result data to the blockchain and generating data micro-blocks; when data micro-blocks meet aggregation trigger conditions, aggregating several data micro-blocks to generate aggregated carbon sink data blocks; generating carbon sink asset entities based on the aggregated carbon sink data blocks and configuring lifecycle state machines for the carbon sink asset entities; constructing a ternary association graph in the blockchain; acquiring carbon sink asset trading requests, performing transaction matching, processing carbon sink asset status, and writing carbon sink asset trading information into the ternary association graph. This application forms a closed-loop management system for carbon sink trading.

Description

Technical Field

[0001] This invention relates to the field of digital asset management technology, and more specifically, to a blockchain-based carbon sink trading method. Background Technology

[0002] With global climate change becoming increasingly prominent, reducing greenhouse gas emissions and enhancing the carbon sequestration capacity of ecosystems have become crucial measures for addressing climate change. Carbon trading, as a market-based greenhouse gas emission reduction mechanism, transforms the carbon sequestration capacity of ecosystems into tradable assets, providing an effective path for promoting green and low-carbon development. Among these, afforestation carbon sequestration holds a significant position in both voluntary and mandatory emission reduction markets due to its long-term stable carbon sequestration capacity and comprehensive value in improving the ecological environment.

[0003] In recent years, blockchain technology, with its characteristics of distributed storage, immutability, and traceability, has been gradually applied to the field of carbon sink management. Existing blockchain-based carbon sink trading solutions typically only hash and store the verification results or key data summaries of carbon sink projects, using the blockchain to ensure the authenticity of data records. However, such solutions still have certain limitations: on the one hand, blockchain mainly undertakes data storage and verification functions, failing to fully utilize smart contracts to automate the management of business processes such as carbon sink asset verification, trading, and cancellation; on the other hand, existing solutions are insufficiently integrated with technologies such as IoT data collection and intelligent analysis models, making it difficult to achieve dynamic collection, real-time accounting, and continuous updating of carbon sink data; furthermore, existing solutions typically treat data storage, asset status management, and transaction tracking as independent functional modules, lacking a unified data association and status management mechanism. A closed-loop management system integrating trusted data collection, intelligent accounting, asset status management, and full-process traceability has not yet been formed.

[0004] Therefore, it is necessary to design a blockchain-based carbon sink trading method to solve the problems existing in the current technology. Summary of the Invention

[0005] In view of this, the present invention proposes a blockchain-based carbon sink trading method, which aims to solve the problem that existing carbon sink trading has not yet formed a closed-loop management system that integrates trusted data collection, intelligent accounting, asset status management and full-process traceability.

[0006] This invention proposes a blockchain-based carbon sink trading method, comprising: Environmental data and tree growth data of the afforestation area are acquired and edge processing is performed to generate structured feature data. The carbon sink amount is dynamically calculated on the structured feature data to generate carbon sink calculation result data. The carbon sequestration calculation results are uploaded to the blockchain and data micro-blocks are generated. When the data micro-blocks meet the aggregation triggering conditions, several data micro-blocks are aggregated to generate aggregated carbon sequestration data blocks, which are then stored in the blockchain. Carbon sequestration asset entities are generated based on the aggregated carbon sequestration data blocks, and a lifecycle state machine is configured for the carbon sequestration asset entities. A ternary association graph is constructed in the blockchain, which includes data entity nodes, model entity nodes, and asset entity nodes, as well as accounting association relationships, generation association relationships, and circulation association relationships; The system acquires carbon sink asset trading requests, performs transaction matching, processes the carbon sink asset status according to the smart contract, and writes the carbon sink asset trading information into the ternary correlation graph.

[0007] Furthermore, when performing dynamic carbon sequestration calculation on the structured feature data, the following steps are included: The environmental data in the structured feature data is input into the first branch network of the carbon sink accounting model, and the forest growth data in the structured feature data is input into the second branch network of the carbon sink accounting model. The output results of the first branch network and the output results of the second branch network are weighted and fused based on the attention mechanism fusion layer to obtain the fused feature vector. The fused feature vector is input into the fully connected regression layer to output the carbon sink accounting result data.

[0008] Furthermore, when generating data micro-blocks, the following are included: The carbon sequestration results data are constructed into data micro-blocks based on a preset time granularity. Each data micro-block includes the hash values ​​of multi-source data of environmental data and forest growth data, as well as the carbon sequestration increment, data batch identifier, and trusted timestamp of the carbon sequestration results data. The hash values ​​of the data micro-blocks are then written into the blockchain.

[0009] Furthermore, when aggregating several of the aforementioned data micro-blocks to generate an aggregated carbon sink data block, the process includes: When the aggregation triggering condition is met, the hash values ​​of all data micro-blocks generated since the previous aggregation time are extracted; the hash values ​​of the data micro-blocks, the version identifier hash value of the carbon sink accounting model in this aggregation, and the hash values ​​of the model's key hyperparameters are used as leaf nodes to construct a multi-branch Merkle tree and generate an aggregation root hash; the aggregation root hash, the time interval information of this aggregation, and the triggering condition identifier are packaged together to generate the aggregated carbon sink data block, and the aggregated carbon sink data block is written into the blockchain.

[0010] Furthermore, the state sequence of the lifecycle state machine includes: monitoring state, verification state, verified state, listing state, transaction locked state, partial delivery state, delivered state, partial cancellation state, full cancellation state, and invalid state.

[0011] Furthermore, the ternary correlation map includes: The accounting association relationship uses the accounting task identifier as a credential to record the input mapping relationship between the data entity node and the model entity node. The accounting association relationship includes the accounting task identifier, the hash list of data entity nodes, the version number of model entity nodes, and the accounting execution timestamp. The generated association relationship uses the model entity node version number and the verification agency identity identifier as credentials to record the verification generation mapping relationship between the model entity node and the asset entity node. The generated association relationship includes the model entity node version number, aggregated carbon sink data block identifier, verification timestamp and verification agency identity identifier. The circulation association uses the transaction contract address and transaction identifier as credentials to record the circulation mapping relationship between the asset entity node and the transaction event. The circulation association includes the asset entity node identifier, transaction identifier, transaction type, transaction timestamp, and transaction participant identity identifier.

[0012] Furthermore, the ternary correlation graph also includes tamper-proof protection: When adding data entity nodes, model entity nodes, or asset entity nodes, or adding accounting relationships, generation relationships, or circulation relationships to the ternary association graph, the log of the ternary association graph change operation is recorded in the blockchain for evidence storage. The log record includes change type, change content hash, change timestamp, and operator identity identifier. Based on the snapshot triggering condition of the ternary association graph, the joint hash value of the current topology structure and entity attribute snapshot of the ternary association graph is determined, and the joint hash value is written into the blockchain to verify the integrity of the historical state of the ternary association graph.

[0013] Furthermore, when obtaining carbon asset trading requests and matching transactions, the process includes: Authentication of transaction participants is performed based on decentralized identifiers and verifiable credentials; the current lifecycle status of the carbon sink asset entity is read from the ternary association graph, and the carbon sink asset entities in the verified or listed status are selected to enter the transaction matching queue; intelligent matching is performed through multi-dimensional feature fusion based on the multi-dimensional attributes of the carbon sink asset entity and the historical transaction performance obtained from the circulation association.

[0014] Furthermore, when processing the status of carbon sink assets according to smart contracts, this includes: When the state transition conditions are detected, the carbon sink asset entity is driven to perform a state transition through the smart contract; each state transition event is recorded as a log in the blockchain, and the log includes the state before the transition, the state after the transition, the transition timestamp, the trigger condition identifier, and the decentralized identity identifier of the triggerer.

[0015] Furthermore, when acquiring environmental data and tree growth data for the afforestation area, this includes: The afforestation area is divided into several ecological units based on an ecological gradient zoning model. The sensor deployment density and acquisition frequency of each ecological unit are determined, and environmental data is acquired through an Internet of Things sensor node network. High-resolution images are acquired through periodic aerial photography using a drone equipped with a positioning system, and macroscopic vegetation data is obtained by connecting to a multi-source satellite remote sensing data interface. Based on the high-resolution images and the macroscopic vegetation data, the forest growth data is obtained.

[0016] Compared with existing technologies, the beneficial effects of this invention are as follows: By constructing a three-in-one on-chain governance architecture of dynamic aggregation of data micro-blocks, lifecycle state machine, and ternary association graph, a trusted closed loop for afforestation carbon sink trading is achieved, encompassing data collection, carbon sink verification, asset trading, and transparent supervision. The data root layer upgrades traditional static snapshot notarization to a two-layer temporal structure of "continuous micro-block on-chain + event-triggered aggregation," and solidifies the model version identifier and key parameter hashes into Merkle tree leaf nodes, enabling the cryptographic fingerprint of carbon sink assets to simultaneously bind the data source and accounting tools, eliminating the dual risks of data tampering and model manipulation. The state-driven layer binds a complete lifecycle state machine from "under monitoring" to "complete cancellation" to each carbon sink asset, ensuring the smooth execution of transactions. Constrained by rigid state transition conditions, the system enables atomic execution of asset transfer and fund clearing in complex financial scenarios such as installment delivery and secured transactions. The associated index layer weaves the originally isolated on-chain evidence records into a causal chain that can be penetrated with one click through a ternary graph containing three types of entities—data, models, and assets—and three types of explicit associated edges. The three form a tight causal closed loop of "data generation → model verification → asset formation → state transition → graph indexing → feedback matching." Each module works collaboratively on a unified on-chain data base rather than operating in isolation. This achieves comprehensive technical effects such as reduced carbon sink response latency, cryptographic verification and traceability at any stage of the entire life cycle, and significantly reduced transaction costs. It solves the fundamental problem of the triple separation of evidence storage, state, and traceability in existing solutions. Attached Figure Description

[0017] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A flowchart illustrating a blockchain-based carbon sink trading method provided in an embodiment of the present invention. Detailed Implementation

[0018] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0019] In some embodiments of this application, see Figure 1 As shown, a blockchain-based carbon sink trading method is proposed, including: S100: Acquire environmental data and tree growth data of the afforestation area, perform edge processing, generate structured feature data, dynamically calculate carbon sink from the structured feature data, and generate carbon sink calculation result data. S200: Upload the carbon sink accounting results data to the blockchain and generate data micro-blocks; when the data micro-blocks meet the aggregation trigger conditions, aggregate several data micro-blocks to generate aggregated carbon sink data blocks and store the aggregated carbon sink data blocks in the blockchain; generate carbon sink asset entities based on aggregated carbon sink data blocks and configure lifecycle state machines for carbon sink asset entities; S300: Construct a ternary association graph in the blockchain. The ternary association graph includes data entity nodes, model entity nodes, and asset entity nodes, as well as accounting associations, generation associations, and circulation associations. S400: Obtain carbon sink asset trading requests, perform transaction matching, process the status of carbon sink assets according to the smart contract, and write the carbon sink asset trading information into the ternary association graph.

[0020] In step S100, multi-source data collection is first conducted on the afforestation area to be monitored to obtain a data foundation that reflects changes in the regional ecological environment and the growth status of trees. Specifically, based on the topographic features, vegetation distribution characteristics, and ecological environment differences of the afforestation area, the area is divided into multiple ecological units using an ecological gradient zoning model. These ecological units may include different types of areas such as ridge areas, valley areas, and slope areas. Since different ecological units have different water conditions, light conditions, soil environment, and vegetation growth conditions, a corresponding sensor deployment strategy is determined based on the ecological characteristics of each ecological unit to improve the ability of the collected data to characterize regional carbon sink changes.

[0021] Specifically, after dividing the ecosystem into units, the deployment density and data acquisition frequency of sensor nodes are determined based on the area, degree of environmental change, and vegetation distribution of each unit. For example, in areas with significant environmental changes, the deployment density of sensor nodes is increased, and a shorter data acquisition cycle is used to obtain more continuous information on environmental changes; in areas with relatively stable environmental changes, the deployment density of sensor nodes is reduced, and a longer data acquisition cycle is used, thereby reducing data acquisition costs while ensuring data validity. Environmental data is collected through an IoT sensor node network deployed in different ecosystem units. This environmental data includes, but is not limited to, soil temperature, soil moisture, soil pH, soil nutrient content, light intensity, precipitation, wind speed, and regional carbon dioxide concentration changes.

[0022] To obtain data reflecting the spatial growth status of trees, this embodiment utilizes a drone equipped with a positioning system to periodically photograph the afforestation area. The positioning system acquires spatial location information corresponding to the images collected by the drone, enabling spatial matching between images collected at different times. The drone, equipped with multispectral or hyperspectral imaging equipment, acquires high-resolution images of the afforestation area, including vegetation texture, canopy structure, and spectral response characteristics. Simultaneously, macroscopic vegetation data covering a large area is obtained through a multi-source satellite remote sensing data interface. This macroscopic vegetation data includes information such as vegetation cover, leaf area index, and regional biomass. By performing spatiotemporal fusion processing on the high-resolution drone images and satellite remote sensing data, tree growth data characterizing the tree growth status is obtained. This tree growth data includes tree height, crown width, diameter at breast height (DBH), vegetation cover, and biomass information.

[0023] In a preferred embodiment, the ecological characteristics of the afforestation area are analyzed by satellite remote sensing images. Based on the topographic relief, vegetation distribution and environmental differences, the afforestation area is divided into three ecological units: ridge area, valley area and slope area. Differentiated sensor deployment strategies are formulated based on the ecological environment change characteristics of different ecological units. Specifically, for ridge areas with drastic environmental changes and significant influence from sunlight and water conditions, a high-density deployment of one sensor node per 25 hectares is adopted, with a node spacing of approximately 50 meters. Soil sensors and miniature weather stations are also included, acquiring soil and meteorological data with a data collection cycle of 10–30 minutes. For river valley areas with relatively stable water conditions, a low-density deployment of one sensor node per 100 hectares is adopted, with a node spacing of approximately 200 meters. Soil sensors are the primary component, and soil environmental data is acquired with a data collection cycle of 45–60 minutes. For slope areas with ecological characteristics between ridge and valley areas, a medium-density deployment of one sensor node per 50 hectares is adopted, with a node spacing of approximately 100 meters. Soil sensors and simplified weather stations are also included, and environmental data is acquired with a data collection cycle of 30–45 minutes. To obtain data on the growth status of trees in afforestation areas, a combination of UAV remote sensing and satellite remote sensing is used for multi-scale data acquisition. The process involves using intelligent drones equipped with RTK high-precision positioning systems to periodically patrol the afforestation area along pre-planned dynamic routes. High-resolution remote sensing images with a resolution of 0.2–0.6 meters are acquired using multispectral or hyperspectral sensors. Based on these high-resolution images, information on tree canopy structure, vegetation spectral characteristics, and spatial distribution is extracted to estimate tree growth parameters such as tree height, crown width, and diameter at breast height (DBH), and to identify abnormal growth conditions such as pests and diseases. Simultaneously, multi-source satellite remote sensing data from Landsat-8 and Sentinel-2 are accessed via a multi-source satellite remote sensing data interface. A spatiotemporal fusion algorithm is used to fuse remote sensing data from different times and spatial resolutions to obtain macroscopic vegetation characteristic data of the afforestation area, including vegetation cover, leaf area index, and biomass. By integrating high-precision local observation data from drones and macroscopic monitoring data from satellite remote sensing, multi-scale tree growth data comprehensively reflecting the growth status of trees in the afforestation area is generated.

[0024] After acquiring environmental and forest growth data, the data is sent to an edge computing unit for preprocessing to reduce the amount of raw data transmission and improve subsequent calculation efficiency. Specifically, the edge computing unit performs anomaly detection, data cleaning, and data format standardization on the sensor-collected data, eliminating invalid data caused by sensor anomalies or communication interference. Simultaneously, it performs lightweight feature extraction on the high-resolution imagery collected by the UAV, converting the raw imagery into structured feature information suitable for carbon sequestration calculations. After the above processing, data from different sources and in different formats are converted into structured feature data in a unified format and stored in association according to their corresponding data source identifiers, collection time identifiers, and spatial location information.

[0025] Dynamic carbon sequestration is performed based on structured feature data. A carbon sequestration model that integrates environmental temporal information and forest spatial information is used to process the structured feature data. Specifically, the carbon sequestration model includes a first branch network, a second branch network, an attention mechanism fusion layer, and a fully connected regression layer. The first branch network is used to process environmental data, and the second branch network is used to process forest growth data.

[0026] The data input to the first branch network is a continuous environmental data stream after edge preprocessing, including environmental parameters such as soil moisture, soil temperature, pH, nitrogen, phosphorus, and potassium content, light intensity, precipitation, wind speed, and carbon dioxide concentration gradient. The environmental data is compiled into time series data according to the acquisition frequency of the corresponding sensors and organized according to a preset time window, such as a sliding time window based on data from the past 30 or 90 days, inputting continuously changing environmental parameters into the first branch network. The first branch network uses a Long Short-Term Memory (LSTM) network or a Transformer network to analyze the correlation between environmental parameters at different time points, extracting the temporal features of how environmental factors affect forest growth and carbon sequestration capacity over time, and outputting a fixed-dimensional temporal feature vector.

[0027] The input data to the second branch network consists of remote sensing image feature data, including forest growth parameters such as tree height, crown width, diameter at breast height (DBH), vegetation cover, leaf area index, and biomass. This remote sensing image feature data is obtained from UAV multispectral / hyperspectral imagery and satellite remote sensing imagery through edge artificial intelligence processing. Specifically, the MobileNet convolutional neural network extracts features from the original remote sensing images, generating spatial feature tensors containing multi-band and multi-resolution information. The second branch network employs a convolutional neural network (CNN) to extract spatial structural features from the remote sensing images through convolution operations at different scales, including information such as crown density, vertical structure of the forest canopy, and spatial distribution differences in biomass. The output is a spatial feature vector characterizing the current spatial carbon sink status of the forest stand.

[0028] The temporal feature vector output from the first branch network and the spatial feature vector output from the second branch network are input into the attention mechanism fusion layer for fusion processing. The attention mechanism fusion layer does not use fixed weights for simple fusion, but dynamically adjusts the contribution weights of temporal and spatial features based on the quality of the current input data, environmental conditions, and the degree of influence of different features on the carbon sink estimation results. For example, when the afforestation area experiences recent environmental fluctuations such as continuous drought or heavy rainfall, the attention weight corresponding to the temporal feature vector is increased, making the model more attentive to the immediate impact of environmental changes on carbon sink capacity; when the quality of UAV remote sensing imagery is high and accurately reflects the vegetation status, the attention weight corresponding to the spatial feature vector is increased, making the model more accurately reflect the current stand biomass and carbon storage status.

[0029] After calculating the corresponding attention weights using an attention mechanism, the temporal and spatial feature vectors are weighted and fused to generate a fused feature vector. This fused feature vector simultaneously incorporates information about environmental dynamics and forest spatial growth status. Subsequently, the fused feature vector is input into a fully connected regression layer, and the initial carbon sink prediction result is output through feature mapping relationships.

[0030] To improve the reliability of carbon sequestration results and avoid deviations from actual growth patterns due to sensor noise, extreme weather changes, and other factors, this embodiment introduces the Richards growth equation as a physical constraint into the training and inference process of the carbon sequestration model. Specifically, during the model training phase, the error between the predicted results and the actual measurement results is used as the basic optimization objective. Simultaneously, a deviation constraint between the predicted results and the Richards growth law is added, ensuring that the model learns the patterns of data change while adhering to the objective law that trees gradually transition from a rapid growth phase to a stable growth phase and approach their growth limit. During the model inference phase, physical constraints limit the impact of abnormal input data on the predicted results, keeping the output results within a reasonable range consistent with the growth patterns of trees, thereby improving the model's stability and robustness under extreme environmental conditions.

[0031] After the aforementioned temporal feature extraction, spatial feature extraction, attention fusion, and physical constraint regression processing, the model outputs the final carbon sink estimate, which represents the carbon sink increment of the corresponding afforestation area in the current period. Simultaneously, the central cloud platform generates a unique carbon sink accounting task identifier (Task ID) for this carbon sink accounting task, and encapsulates the accounting task identifier, the data batch identifier participating in the accounting, the carbon sink accounting model version identifier used, and the final carbon sink estimate result into a carbon sink accounting result data package, which is then sent to the blockchain module.

[0032] In step S200, the blockchain module performs trusted storage and asset management of the carbon sink accounting results data. In this embodiment, the blockchain module adopts a progressive on-chain data structure of "data root layer - state-driven layer - association index layer" to realize continuous storage of carbon sink accounting data, carbon sink asset lifecycle management, and traceability of subsequent transaction processes.

[0033] First, at the root data level, the carbon sequestration results are processed into micro-blocks. Specifically, the continuously generated carbon sequestration results are divided according to a preset time granularity, and the carbon sequestration results generated within the same time period are constructed into corresponding data micro-blocks. The preset time granularity can be set according to actual application needs, such as dividing according to daily, weekly, or other fixed periods. When generating data micro-blocks, hash calculations are performed on the multi-source data participating in carbon sequestration to obtain corresponding data fingerprint information. The multi-source data includes environmental data collected by IoT sensors and forest growth data obtained by drones and satellite remote sensing equipment. Further, the hash values ​​of the multi-source data, the carbon sequestration increment values ​​in the carbon sequestration results, the data batch identifier, and the trusted timestamp information are written into the corresponding data micro-blocks.

[0034] The trusted timestamp is used to prove the generation time of the corresponding carbon sink accounting result data, avoiding subsequent data backtracking or time sequence disorder. After each data micro-block is generated, a hash calculation is performed on the data micro-block, and the resulting hash value is written into the blockchain distributed ledger, so that each data micro-block formed in each time period has a unique on-chain identifier, realizing continuous evidence storage of carbon sink accounting data from the generation stage. Compared with the traditional method of only storing the final verification result, this embodiment continuously uploads data micro-blocks to the chain, making carbon sink data a continuous and tamper-proof time series record.

[0035] To avoid rapid growth in on-chain data volume due to the continuous storage of a large number of data micro-blocks, and to ensure data integrity during long-term carbon asset management, this embodiment uses a dynamic aggregation mechanism to aggregate multiple data micro-blocks. Specifically, a dynamic aggregation trigger contract is set up in the blockchain module. This contract monitors the status of data micro-blocks in real time and determines whether to execute the data aggregation operation based on preset aggregation trigger conditions.

[0036] The aggregation triggering conditions include at least one of the following: the cumulative carbon sequestration increase since the last aggregation has reached a preset threshold; a preset time period has elapsed since the last aggregation; or a preset ecological event has occurred in the afforestation area, as detected by a blockchain oracle. Ecological events can include abnormal events affecting tree growth, such as fires, rainstorms, and droughts. When any aggregation triggering condition is met, the dynamic aggregation triggering contract automatically retrieves all data micro-blocks generated since the last aggregation moment and performs aggregation processing.

[0037] Specifically, during the aggregation process, the hash values ​​corresponding to each data micro-block are extracted. These micro-block hash values, the version identifier hash value of the current carbon sequestration accounting model, and the hash values ​​of key model parameters are used together as leaf nodes of a Merkle tree. A multi-branch Merkle tree structure is then used to calculate the corresponding aggregate root hash. The model version identifier hash value indicates the specific model version used in this carbon sequestration calculation, while the key model parameter hash value indicates the important parameter configurations during model operation. By incorporating data source information and accounting model information into the Merkle tree calculation, the generated aggregate root hash simultaneously contains evidence of carbon sequestration data and accounting methodology.

[0038] Subsequently, the aggregated root hash, the time interval information corresponding to the current aggregation process, and the trigger condition identifier are encapsulated to generate an aggregated carbon sink data block, which is then written into the blockchain. The time interval information records the data period covered by this aggregation, including the start and end times; the trigger condition identifier records whether the aggregation was triggered by a cumulative carbon sink threshold, a time period, or an ecological event. The aggregated carbon sink data block formed in this way serves as the basic data unit for subsequent carbon sink asset generation.

[0039] Because the Merkle tree is constructed using data hashes, model version hashes, and model parameter hashes during the aggregation process, any change in the content of any data micro-block will result in a change in its corresponding hash value, which in turn will cause a change in the Merkle tree root hash. Similarly, changes in the carbon sequestration model version or key model parameters will also lead to inconsistencies in the aggregation root hash. Therefore, this dynamic aggregation mechanism can simultaneously ensure the authenticity of both the source of carbon sequestration data and the source of the calculation method, avoiding the problem of artificially increasing carbon sequestration through data replacement or model parameter adjustments.

[0040] In the state-driven layer, corresponding carbon sink asset entities are generated based on aggregated carbon sink data blocks, and a lifecycle state machine is configured for each carbon sink asset entity in the blockchain to achieve state management of the entire process of carbon sink assets from formation, verification, trading to cancellation. Specifically, each aggregated carbon sink data block corresponds to an independent carbon sink asset entity, which includes at least an asset identifier, the corresponding aggregated carbon sink data block identifier, the amount of carbon sink, the generation time, and the current lifecycle state information.

[0041] The lifecycle state machine includes the following states: Monitoring, Verification, Verified, Listing, Transaction Locked, Partial Delivery, Delivery, Partial Cancellation, Full Cancellation, and Expiration. Specifically, Monitoring indicates that the data for the carbon asset is still being collected and third-party verification has not yet been completed; Verification indicates that aggregated carbon asset data blocks have been generated and are in the data integrity verification and inspection phase; Verified indicates that the carbon asset has passed inspection and meets the conditions for entering the trading market; Listing indicates that the carbon asset has been published on the trading platform and is awaiting trading; Transaction Locked indicates that the buyer and seller have reached a trading agreement, and the asset and trading funds are locked; Partial Delivery and Delivery indicate that the carbon asset has completed partial or full ownership transfer, respectively; Partial Cancellation and Full Cancellation indicate that the corresponding carbon credits have been used and cancelled; and Expiration indicates that the carbon asset has lost its trading value due to exceeding its validity period or ecological damage.

[0042] The transitions between different lifecycle states are controlled by pre-defined state transition rules in the smart contract. For example, when the dynamic aggregation trigger contract completes the aggregation process and the continuously collected data meets the time window requirements for verification, the carbon sink asset entity transitions from the monitoring state to the verification state; when the data trust storage contract completes data integrity verification and receives a pre-defined number of digital signatures from verification agencies, the carbon sink asset entity transitions from the verification state to the verified state; when the asset holder submits a listing request through decentralized identity authentication, the carbon sink asset entity transitions from the verified state to the listing state.

[0043] Once the trading platform completes the transaction matching and the smart contract verifies that the buyer's funds lock-up conditions are met, the carbon sink asset entity will be transferred from the listing status to the transaction lock-up status. When the delivery conditions are met, the automated transaction performance contract will execute the transfer of asset ownership and fund settlement, and the asset status will be transferred to the partial delivery status or the delivered status according to the delivery quantity. When the buyer submits proof of carbon sink usage and passes the smart contract verification, the asset status will be further transferred to the partial cancellation status or the full cancellation status.

[0044] During each lifecycle state transition, the smart contract generates a corresponding state transition log and writes it to the blockchain. The state transition log includes at least the state before the transition, the state after the transition, the transition time, triggering conditions, and the identity information of the triggering entity. By combining the lifecycle state machine with the blockchain's notarization mechanism, the formation, verification, trading, and cancellation processes of carbon sink assets are all subject to on-chain rules, achieving trusted management of carbon sink assets throughout their entire lifecycle.

[0045] In step S300, an association index layer is set up in the blockchain to construct a ternary association graph. This graph, in the form of a directed graph, associates the data generated during the carbon sink accounting process, the accounting model used, the formed carbon sink assets, and subsequent transaction events. Unlike conventional knowledge graphs used only for information retrieval, the entity nodes in the ternary association graph in this embodiment all correspond to specific objects that have been notarized in the blockchain. Verifiable associations between entities are established through on-chain credentials, enabling carbon sink assets to be traced along the data source, accounting model, asset formation, and transaction flow path.

[0046] Specifically, the ternary association graph includes data entity nodes, model entity nodes, and asset entity nodes. Among them, the data entity nodes originate from data micro-blocks and aggregated carbon sink data blocks generated in S200 and stored in the blockchain. Each data entity node corresponds to a specific data micro-block or aggregated carbon sink data block and records its data block identifier, hash value, generation time, and data batch identifier, so that the on-chain data entities can correspond one-to-one with the actual stored data objects.

[0047] The model entity nodes originate from the carbon sequestration calculation model used in S100 for dynamic carbon sequestration calculation. Corresponding model entity nodes are established for different versions of the carbon sequestration calculation model, recording information such as the model version number, release date, model file hash, training dataset version, and difference summary hash generated by model version updates. By registering the model version as an independent entity on the blockchain, the carbon sequestration calculation model used at different times can be clearly distinguished, avoiding the problem of not being able to identify the specific calculation tool during subsequent verification.

[0048] The asset entity node originates from carbon sink asset entities generated in S200 based on aggregated carbon sink data blocks. Each aggregated carbon sink data block corresponds to one carbon sink asset entity node. The asset entity node records information such as asset identifier, corresponding aggregated carbon sink data block identifier, current lifecycle status, and state migration history, enabling the asset entity node to establish a correspondence with specific carbon sink assets in the blockchain and to be updated synchronously as the asset status changes.

[0049] After constructing the three types of entity nodes mentioned above, explicit relationships are established between these entity nodes. These relationships include accounting relationships, generation relationships, and circulation relationships, which respectively describe the input relationship between carbon sink data and the accounting model, the generation relationship between the accounting model and carbon sink assets, and the circulation relationship between carbon sink assets and transaction events. Through these relationships, the originally independent data storage, model management, asset management, and transaction records are organized into an on-chain relational structure with clear causal relationships.

[0050] Specifically, the accounting association is used to record the input mapping relationship between data entity nodes and model entity nodes. When S100 completes a carbon sink accounting, the central cloud platform generates a corresponding accounting task identifier (Task ID) and associates the data batch identifier, the model version identifier used, and the accounting result with it. After receiving the accounting result data, the blockchain module determines the data entity node and model entity node corresponding to this accounting based on the Task ID and establishes an accounting association pointing from the data entity node to the model entity node. The accounting association records at least the accounting task identifier, the hash list of data entity nodes participating in the accounting, the model entity node version number, and the accounting execution timestamp, thereby enabling the specific data and model version used to be determined from a carbon sink accounting result.

[0051] The generation of associations is used to record the verification generation mapping relationship between model entity nodes and asset entity nodes. After S200 generates an aggregated carbon sink data block and further forms a carbon sink asset entity, it determines the corresponding model entity node based on the model version identifier recorded in the aggregated carbon sink data block, and establishes a generation association between the model entity node and the asset entity node. The generation association records at least the model entity node version number, the aggregated carbon sink data block identifier, the verification timestamp, and the verification body's identity identifier, thus indicating which model version the corresponding carbon sink asset was based on and which verification body confirmed it.

[0052] The circulation association is used to record the circulation mapping relationship between asset entity nodes and transaction events. When carbon sink assets enter the trading process such as listing, trading lock-in, delivery, or cancellation, the corresponding smart contract generates a transaction event and determines the target asset entity node based on the asset identifier corresponding to the transaction event, establishing a circulation association between the asset entity node and the transaction event. The circulation association records at least the asset entity node identifier, transaction identifier, transaction type, transaction timestamp, and the identity identifiers of the transaction participants. The transaction contract address and transaction identifier serve as verification credentials for the corresponding transaction event, enabling regulatory authorities to query the historical transaction and circulation process of the asset based on the asset identifier.

[0053] To ensure the integrity of the ternary association graph itself, graph change notarization is performed when new entity nodes or associations are added. Specifically, when new data entity nodes, model entity nodes, or asset entity nodes are added, or when new accounting associations, generated associations, or transferred associations are added, a corresponding graph change log is generated and written to the blockchain for notarization. The graph change log includes at least the change type, change content hash, change timestamp, and operator identification. The change content hash represents the digital fingerprint of the newly added or modified content, enabling subsequent consistency verification of the graph content before and after the change.

[0054] Set the snapshot trigger conditions for the ternary association graph, and perform an integrity check on the current ternary association graph when the snapshot trigger conditions are met. The snapshot trigger conditions can be triggered according to a preset time period, or triggered when the number of graph entity nodes or the number of association relationships reaches a preset value. For example, a graph snapshot can be generated once a day, or a snapshot can be triggered every time the number of graph entity nodes increases by 100. After triggering, the entity nodes and their attribute information in the current ternary association graph are obtained, and the topological connection relationships between each entity node are obtained. A joint hash calculation is performed on the topological structure and entity attribute snapshots to obtain the graph joint hash value, and the joint hash value is written to the blockchain.

[0055] When it is necessary to verify the historical state of a ternary association graph, the graph entities, entity attributes, and relationships corresponding to the time to be verified are obtained. The current joint hash value is then recalculated in the same way as the generated graph joint hash value. This recalculated joint hash value is compared with the corresponding historical snapshot joint hash value stored in the blockchain. If they match, it is determined that the ternary association graph at the corresponding time point has not undergone any tampering inconsistent with the on-chain records. If they do not match, it is determined that the entity attributes or topological relationships of the graph have undergone abnormal changes, thus achieving the integrity verification of the historical state of the ternary association graph.

[0056] To establish a correspondence between on-chain data and off-chain raw data, this embodiment employs a combined on-chain and off-chain data storage approach. For large volumes of raw data, such as sensor readings and UAV remote sensing images, off-chain databases or distributed file storage systems can be stored, with corresponding content identifiers or hash values ​​stored on the blockchain. Core information used to verify the credibility of carbon sink assets, such as data fingerprints, aggregate root hashes, entity relationships, state transition records, and transaction records, is stored on the blockchain. Thus, without significantly increasing the storage pressure on the blockchain, on-chain records can be linked to corresponding off-chain raw data via hash values.

[0057] For data requiring access control, attribute-based encryption can be used. Data owners pre-set access policies, using the access subject's role, geographic attributes, and time attributes as access conditions. For example, access to the corresponding raw data can be restricted to subjects with an inspector role located within a pre-defined geographic area of ​​the target afforestation project and during a specified monitoring period. When an access subject submits a data access request, the access conditions are verified based on their attribute credentials. Access to the corresponding data is only granted if the access subject's attributes meet the pre-defined access policy, thus ensuring the traceability of carbon sequestration data while restricting unauthorized access to the raw data.

[0058] Through the above methods, the ternary association graph organizes data entities, model entities, asset entities, and transaction events according to their accounting, generation, and circulation relationships. It establishes verifiable association paths using credentials such as accounting task identifiers, model versions and verification agency identities, and transaction contract addresses and transaction identifiers. This allows any carbon sink asset to be queried along the path of "data entity → model entity → asset entity → transaction event," realizing full-process association and traceability from the source of carbon sink data, accounting model, asset formation to transaction circulation.

[0059] In step S400, the trading platform module, as the front-end interaction layer for carbon sink asset trading, is used to receive carbon sink asset trading requests submitted by trading participants and interact with the carbon sink asset entities, life cycle status, and ternary association graph in the blockchain to complete the identity verification of trading participants, carbon sink asset screening, transaction matching, transaction performance, and on-chain transaction information, thereby realizing the linkage between the carbon sink asset trading process and the status of on-chain assets.

[0060] First, when a participant submits a carbon asset trading request through the trading platform, the platform obtains the participant's decentralized identifier (DID) and the corresponding verifiable credential (VC). The platform then verifies the participant's identity based on the VC. The VC is issued by a trusted authentication authority that has pre-certified the participant's identity. The platform determines the VC's validity by verifying its digital signature, without needing to store the participant's complete identity information on the platform. After successful verification, the participant's decentralized identity is used as the credential for subsequent trading operations and to record the participant's relevant actions on the blockchain.

[0061] The trading platform determines trading demand based on carbon sink asset trading requests and obtains candidate carbon sink asset entities from the ternary association graph in the blockchain. For each candidate carbon sink asset entity, its corresponding lifecycle status is queried based on its asset identifier, and the candidate assets are screened according to preset trading access conditions. Specifically, only carbon sink asset entities in the certified or listed state are included in the trading matching scope. Carbon sink assets in the certified state indicate that carbon sink certification has been completed and they meet the trading conditions, while carbon sink assets in the listed state indicate that the asset holder has submitted a listing request and entered the trading market. Carbon sink assets that do not meet the trading access conditions, such as those in the monitoring, certification, trading lock, partial cancellation, full cancellation, or invalidation states, are not included in the trading matching queue, thereby avoiding duplicate trading of carbon sink assets that have not been certified, have been locked, or have been cancelled.

[0062] After completing the transaction access screening, the trading platform's intelligent matching engine matches transactions based on the demand information in the transaction request and the attribute information of the candidate carbon sink assets. The attribute information of carbon sink assets includes basic attributes such as carbon sink volume, transaction price, and vintage year; ecological attributes such as tree species and carbon sequestration efficiency; risk attributes such as regional risk coefficient and certification standards; and credit attributes such as the project owner's historical compliance rate. The trading platform performs feature processing on these different types of attributes to form multi-dimensional features for transaction matching, and comprehensively evaluates different attributes according to the needs of the transaction participants to determine candidate carbon sink assets with a high degree of matching with the transaction request.

[0063] During transaction matching, the trading platform also obtains historical transaction information of candidate carbon sink assets from the circulation relationships in the ternary correlation graph. This historical transaction information includes historical transaction prices, historical circulation frequency, and historical transaction performance. This historical transaction information is used as an auxiliary feature for transaction matching, integrated with the fundamental, ecological, risk, and credit attributes of the carbon sink assets to obtain matching results between candidate carbon sink assets and transaction requests. By using both the current state of the asset and historical circulation information as matching criteria, the platform avoids matching transactions solely based on carbon sink quantity and price, improving the alignment between transaction matching results and actual transaction demand.

[0064] Once the intelligent matching engine identifies the target carbon sink asset and the corresponding trading participants, the trading platform sends a transaction execution request to the automated transaction fulfillment contract in the blockchain module. The automated transaction fulfillment contract first reads the current lifecycle state of the target carbon sink asset in the ternary correlation graph and verifies the transaction execution conditions for that lifecycle state. If the target carbon sink asset is listed and meets the transaction execution conditions, the transaction is allowed to proceed; if the target carbon sink asset is not listed, or the asset corresponding to the transaction request has undergone a state change, the transaction is rejected. This avoids erroneous transactions caused by inconsistencies between the asset state on the trading platform and the actual asset state on the blockchain.

[0065] Once the transaction execution conditions are verified, the automated transaction fulfillment contract will transfer the lifecycle status of the target carbon sink asset from the listed state to the transaction-locked state. Simultaneously, it will lock the buyer's prepayment according to the transaction rules and freeze the corresponding carbon sink asset for the seller. During the transaction lock-up period, the locked carbon sink asset cannot enter other transaction matching processes, thus preventing the same carbon sink asset from being sold to multiple transaction participants simultaneously.

[0066] During the transaction fulfillment process, the system determines whether the carbon asset delivery requirements are met based on pre-set delivery conditions. Delivery conditions may include the passing of the agreed verification results, the arrival of the agreed delivery time, or the fulfillment of verification conditions for the corresponding stage of phased delivery. When only part of the delivery requirements are met, the automated transaction fulfillment contract migrates the lifecycle status of the target carbon asset from the transaction-locked state to the partially delivered state and records the quantity of carbon assets that have been delivered. When all delivery conditions are met, the lifecycle status of the target carbon asset is migrated to the delivered state, and the ownership transfer information of all carbon assets is recorded.

[0067] When executing carbon asset delivery, the automated transaction fulfillment contract simultaneously performs the transfer of ownership and settlement of transaction funds according to preset trading rules. This involves updating the asset ownership record in the blockchain and transferring transaction funds that meet the settlement conditions to the seller's corresponding account. By using smart contracts to link asset status, asset ownership, and fund settlement, carbon asset delivery and fund payment are executed in the same transaction fulfillment process, reducing inconsistencies in transaction status caused by human intervention.

[0068] After the transaction state migration is completed, the automated transaction fulfillment contract establishes a corresponding circulation relationship in the ternary association graph. Specifically, starting from the asset entity node corresponding to the target carbon sink asset, a circulation relationship is established with this transaction event as the associated object, and information such as the transaction contract address, transaction identifier, transaction type, transaction timestamp, transaction participant identity identifiers, and delivery quantity are written into the circulation relationship. At the same time, a graph change log is generated for this new operation in the graph, and the change log is stored on-chain in accordance with the S300 method, so that this transaction can establish a continuous association with the historical accounting, verification, and other transaction records of the target carbon sink asset.

[0069] After a transaction is fulfilled, the trading platform generates a transaction certificate based on the transaction results recorded on the blockchain. The transaction certificate includes at least the transaction contract address, transaction identifier, transaction timestamp, and the corresponding state transition sequence for this transaction, and is sent to both parties. Both parties can use the transaction certificate to query the corresponding on-chain transaction records and verify the state changes of the carbon asset before and after the transaction.

[0070] When a carbon asset is actually used and cancelled after a transaction is completed, the trading platform receives the corresponding carbon usage certificate and submits it to a smart contract for verification. Upon successful verification, the carbon asset's status is transitioned from a delivered state to a partially cancelled or fully cancelled state based on the actual amount of carbon credits cancelled. When only a portion of the carbon credits are cancelled, the platform records the amount cancelled and the remaining available carbon credits. When all carbon credits of the corresponding asset are cancelled, its status is transitioned to a fully cancelled state, and the asset is prevented from re-entering the trading matching queue.

[0071] In the aforementioned transaction process, each lifecycle state transition is executed by a smart contract according to preset state transition rules, generating a corresponding on-chain state transition log. The state transition log includes at least the state before the transition, the state after the transition, the transition timestamp, the trigger condition identifier, and the decentralized identity identifier of the triggerer. By embedding the state transition rules into the smart contract, carbon sink assets in a state where trading or delivery is prohibited are unable to perform corresponding operations, thereby achieving state constraints in the carbon sink asset trading process.

[0072] Through the above method, the carbon sink accounting results generated by the data acquisition module are first continuously stored in the form of data micro-blocks, and then formed into aggregated carbon sink data blocks after the aggregation conditions are met. These aggregated carbon sink data blocks then generate carbon sink asset entities. Subsequently, a ternary association graph is used to establish the association between data entities, model entities, and asset entities. A lifecycle state machine controls the assets to gradually move from the monitoring and verification stages to the verified and listed stages. After a transaction is matched, a smart contract drives the assets to sequentially undergo transaction locking, partial delivery, or delivery, and the transaction events are written into the ternary association graph. After the carbon sink asset is used and cancelled, it is migrated to a partially cancelled or fully cancelled state based on the number of cancellations. Thus, a carbon sink asset can form a continuous on-chain association record from the source of accounting data, accounting model, asset formation, transaction flow, to final cancellation, providing regulatory bodies or transaction participants with complete and transparent traceability evidence.

[0073] In summary, by constructing a three-in-one on-chain governance architecture of dynamic aggregation of data micro-blocks, lifecycle state machine, and ternary association graph, a trusted closed loop for afforestation carbon sink trading is achieved, encompassing data collection, carbon sink verification, asset trading, and transparent supervision. The data root layer upgrades traditional static snapshot notarization to a two-layer temporal structure of "continuous micro-block on-chaining + event-triggered aggregation," and solidifies model version identifiers and key parameter hashes into Merkle tree leaf nodes. This ensures that the cryptographic fingerprint of carbon sink assets is simultaneously bound to the data source and accounting tools, eliminating the dual risks of data tampering and model manipulation. The state-driven layer binds each carbon sink asset to a complete lifecycle state machine from "under monitoring" to "complete cancellation," making transaction execution subject to rigid state transitions. Conditional constraints enable atomic execution of asset transfer and fund clearing in complex financial scenarios such as installment delivery and secured transactions. The associated index layer weaves the originally isolated on-chain evidence records into a causal chain that can be penetrated with one click through a ternary graph containing three types of entities—data, models, and assets—and three types of explicit associated edges. The three form a tight causal closed loop of "data generation → model verification → asset formation → state transition → graph indexing → feedback matching". Each module works collaboratively on a unified on-chain data base instead of operating in isolation. This achieves comprehensive technical effects such as reduced carbon sink response latency, cryptographic verification and traceability at any stage of the entire life cycle, and significantly reduced transaction costs. It solves the fundamental problem of the triple separation of evidence storage, state, and traceability in existing solutions.

[0074] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the present invention.

Claims

1. A blockchain-based carbon sink trading method, characterized in that, include: Environmental data and tree growth data of the afforestation area are acquired and edge processing is performed to generate structured feature data. The carbon sink amount is dynamically calculated on the structured feature data to generate carbon sink calculation result data. The carbon sequestration calculation results are uploaded to the blockchain and data micro-blocks are generated. When the data micro-blocks meet the aggregation triggering conditions, several data micro-blocks are aggregated to generate aggregated carbon sequestration data blocks, and the aggregated carbon sequestration data blocks are stored in the blockchain. Carbon sink asset entities are generated based on the aggregated carbon sink data blocks, and a lifecycle state machine is configured for the carbon sink asset entities; wherein, the aggregation triggering conditions include: the cumulative carbon sink increment reaches a preset increment threshold after the last aggregation is completed, a preset time period has been reached since the last aggregation is completed, and a preset ecological event has occurred in the afforestation area through a blockchain oracle; A ternary association graph is constructed in the blockchain, which includes data entity nodes, model entity nodes, and asset entity nodes, as well as accounting association relationships, generation association relationships, and circulation association relationships; The system acquires carbon sink asset trading requests, performs transaction matching, processes the carbon sink asset status according to the smart contract, and writes the carbon sink asset trading information into the ternary correlation graph.

2. The blockchain-based carbon sink trading method according to claim 1, characterized in that, When performing dynamic carbon sequestration calculation on the structured feature data, the following steps are included: The environmental data in the structured feature data is input into the first branch network of the carbon sink accounting model, and the forest growth data in the structured feature data is input into the second branch network of the carbon sink accounting model. The output results of the first branch network and the output results of the second branch network are weighted and fused based on the attention mechanism fusion layer to obtain the fused feature vector. The fused feature vector is input into the fully connected regression layer to output the carbon sink accounting result data.

3. The blockchain-based carbon sink trading method according to claim 2, characterized in that, When generating data micro-blocks, the following are included: The carbon sequestration results data are constructed into data micro-blocks based on a preset time granularity. Each data micro-block includes the hash values ​​of multi-source data of environmental data and forest growth data, as well as the carbon sequestration increment, data batch identifier, and trusted timestamp of the carbon sequestration results data. The hash values ​​of the data micro-blocks are then written into the blockchain.

4. The blockchain-based carbon sink trading method according to claim 3, characterized in that, When aggregating several of the aforementioned data micro-blocks to generate an aggregated carbon sink data block, the process includes: When the aggregation triggering condition is met, the hash values ​​of all data micro-blocks generated since the previous aggregation time are extracted; the hash values ​​of the data micro-blocks, the version identifier hash value of the carbon sink accounting model in this aggregation, and the hash values ​​of the model's key hyperparameters are used as leaf nodes to construct a multi-branch Merkle tree and generate an aggregation root hash; the aggregation root hash, the time interval information of this aggregation, and the triggering condition identifier are packaged together to generate the aggregated carbon sink data block, and the aggregated carbon sink data block is written into the blockchain.

5. The blockchain-based carbon sink trading method according to claim 4, characterized in that, The state sequence of the lifecycle state machine includes: monitoring, verification, verified, listing, transaction locked, partial delivery, delivered, partial cancellation, full cancellation, and invalidation.

6. The blockchain-based carbon sink trading method according to claim 5, characterized in that, The ternary correlation map includes: The accounting association relationship uses the accounting task identifier as a credential to record the input mapping relationship between the data entity node and the model entity node. The accounting association relationship includes the accounting task identifier, the hash list of data entity nodes, the version number of model entity nodes, and the accounting execution timestamp. The generated association relationship uses the model entity node version number and the verification agency identity identifier as credentials to record the verification generation mapping relationship between the model entity node and the asset entity node. The generated association relationship includes the model entity node version number, aggregated carbon sink data block identifier, verification timestamp and verification agency identity identifier. The circulation association uses the transaction contract address and transaction identifier as credentials to record the circulation mapping relationship between the asset entity node and the transaction event. The circulation association includes the asset entity node identifier, transaction identifier, transaction type, transaction timestamp, and transaction participant identity identifier.

7. The blockchain-based carbon sink trading method according to claim 6, characterized in that, The ternary correlation graph also includes tamper-proof protection: When adding the data entity node, the model entity node, or the asset entity node, or adding the accounting relationship, the generation relationship, or the circulation relationship in the ternary association graph, the log of the ternary association graph change operation is recorded in the blockchain for evidence storage. The log record includes the change type, change content hash, change timestamp, and operator identity identifier. Based on the snapshot triggering condition of the ternary association graph, the joint hash value of the current topological structure and entity attribute snapshot of the ternary association graph is determined, and the joint hash value is written into the blockchain to verify the integrity of the historical state of the ternary association graph.

8. The blockchain-based carbon sink trading method according to claim 7, characterized in that, When receiving carbon asset trading requests and matching transactions, the process includes: Authentication of transaction participants is performed based on decentralized identifiers and verifiable credentials; the current lifecycle status of the carbon sink asset entity is read from the ternary association graph, and the carbon sink asset entities in the verified or listed status are selected to enter the transaction matching queue; intelligent matching is performed through multi-dimensional feature fusion based on the multi-dimensional attributes of the carbon sink asset entity and the historical transaction performance obtained from the circulation association.

9. The blockchain-based carbon sink trading method according to claim 8, characterized in that, When processing the status of carbon sink assets according to smart contracts, the following are included: When the state transition conditions are detected, the carbon sink asset entity is driven to perform a state transition through the smart contract; each state transition event is recorded as a log in the blockchain, and the log includes the state before the transition, the state after the transition, the transition timestamp, the trigger condition identifier, and the decentralized identity identifier of the triggerer.

10. The blockchain-based carbon sink trading method according to claim 1, characterized in that, When obtaining environmental data and tree growth data for afforestation areas, the following should be included: The afforestation area is divided into several ecological units based on an ecological gradient zoning model. The sensor deployment density and acquisition frequency of each ecological unit are determined, and environmental data is acquired through an Internet of Things sensor node network. High-resolution images are acquired through periodic aerial photography using a drone equipped with a positioning system, and macroscopic vegetation data is obtained by connecting to a multi-source satellite remote sensing data interface. Based on the high-resolution images and the macroscopic vegetation data, the forest growth data is obtained.

Images