A Dynamic Monitoring Method for Mangrove Carbon Sequestration Based on Tidal Disturbance Compensation

By constructing a nonlinear mapping relationship between the tidal observation perspective drift index and the remote sensing signal, the remote sensing observation perspective drift was corrected, thus solving the impact of tidal changes on mangrove carbon sink monitoring and realizing the stabilization treatment and high-stability monitoring of mangrove carbon storage.

CN123117553APending Publication Date: 2026-11-17SHENZHEN NATURAL RESOURCES & REAL ESTATE ASSESSMENT & DEV RES CENT (SHENZHEN GEOLOGICAL ENVIRONMENT MONITORING CENT) +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611397735.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-09
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing remote sensing monitoring methods are difficult to accurately describe the impact of tidal disturbances on mangrove carbon sequestration monitoring under tidal change environments, resulting in inconsistencies in remote sensing observation results and unstable carbon storage estimations. In particular, the drift of remote sensing observation perspective is difficult to correct in complex tidal environments.

Method used

By acquiring multi-temporal remote sensing images, tidal water level sequences, and UAV lidar data, a dynamic three-dimensional tidal observation field is constructed, the tidal observation perspective drift index is calculated, a nonlinear mapping relationship is established for signal correction, stability characteristic parameters are extracted, and mangrove carbon storage is inverted.

Benefits of technology

It enables consistency analysis of remote sensing data under different tidal conditions, reduces the interference of tidal changes on remote sensing carbon structure information, improves the reliability and stability of mangrove carbon sink monitoring, and adapts to long-term remote sensing monitoring under complex tidal environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123117553A_ABST
    Figure CN123117553A_ABST
Patent Text Reader

Abstract

This invention relates to the field of ecological and environmental protection technology, and in particular to a method for dynamic monitoring of mangrove carbon sinks based on tidal level disturbance compensation. The method includes: Step 1: Constructing a multi-source remote sensing data acquisition and fusion system and a three-dimensional tidal dynamic observation field; Step 2: Establishing a tidal observation perspective drift index and generating a tidal observation perspective drift index matrix; Step 3: Constructing a tidal perspective drift compensation model based on mapping relationships, and then correcting the original remote sensing observation signals; Step 4: Fusing features and filtering variables to obtain a set of carbon sink inversion feature parameters; Step 5: Generating a mangrove carbon sink spatial distribution map dataset based on tidal observation perspective drift correction. This invention quantitatively characterizes the changes in remote sensing observation perspective caused by changes in water body coverage and mangrove structure exposure during tidal level changes, thus enabling the description of changes in remote sensing observation conditions under different tidal levels.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of ecological and environmental protection technology, and in particular to a method for dynamic monitoring of mangrove carbon sequestration based on tidal disturbance compensation. Background Technology

[0002] Mangrove ecosystems are typical coastal wetland ecosystems, typically located in the intertidal zone where land meets the sea, and are constantly influenced by tidal fluctuations. The tidal state corresponding to images acquired by remote sensing sensors at different times is not entirely consistent, causing the remote sensing observation process to be affected by dynamic tidal changes. Mangrove carbon sequestration monitoring is usually based on multi-temporal remote sensing data, analyzing spectral characteristics, radar backscattering characteristics, and vegetation structure changes acquired at different time points to invert biomass and carbon storage. However, in the intertidal environment, because the objects observed by remote sensing sensors dynamically change with tidal levels, the effective vegetation structure information corresponding to the same spatial pixel under different tidal conditions varies, creating a coupling effect between changes in remote sensing signals and changes in actual vegetation carbon structure. Especially when using multi-temporal remote sensing data for long-term dynamic monitoring, tidal changes may cause systematic shifts in remote sensing observation results, affecting the consistency between remote sensing data at different time scales.

[0003] Furthermore, existing remote sensing monitoring methods typically use the entire mangrove forest area as the observation object, focusing primarily on the changing characteristics of the remote sensing signal itself, while lacking an effective description of the changes in the observed state of different vegetation structural units under tidal influence. Because mangroves possess distinct vertical structural features, different structural units such as roots, trunks, and canopies have varying degrees of exposure and remote sensing response capabilities during tidal changes. Tidal variations alter the contribution ratio of different structural information to the remote sensing observation results, causing the vegetation structural information represented by the remote sensing data to change with tidal conditions. Therefore, accurately describing the changes in remote sensing observation perspective caused by tidal disturbances and effectively distinguishing and correcting these changes from the remote sensing signal is a key issue in improving the reliability of carbon sequestration monitoring in intertidal mangroves.

[0004] Meanwhile, existing carbon storage inversion methods often directly establish model relationships using remote sensing characteristic parameters and sample plot survey data, failing to fully consider the impact of dynamic changes in observation conditions during remote sensing data acquisition. This leads to uncertainties in remote sensing characteristic parameters acquired under different tidal conditions. Especially in complex tidal environments, the remote sensing observation perspective drifts with tidal changes, making it difficult for traditional methods to ensure a unified observational basis across multiple temporal remote sensing features, thus affecting the stability of mangrove carbon storage spatial distribution estimation. Therefore, addressing the problem of dynamic changes in remote sensing observation states caused by tidal changes during intertidal mangrove remote sensing monitoring, how to effectively characterize the observation perspective drift caused by tidal effects, establish the correlation between changes in remote sensing observation perspective and remote sensing signal response, and achieve stabilization of remote sensing data under the influence of tidal disturbances, thereby improving the reliability of dynamic monitoring of mangrove carbon sinks, remains a technical problem that needs to be solved. Summary of the Invention

[0005] This invention addresses the problems existing in the prior art by providing a dynamic monitoring method for mangrove carbon sequestration based on tidal disturbance compensation, mainly comprising:

[0006] A method for dynamic monitoring of mangrove carbon sequestration based on tidal disturbance compensation includes,

[0007] Step 1: Acquire multi-temporal optical remote sensing images, synthetic aperture radar images, tidal water level sequences, digital elevation models, and UAV lidar point cloud data, and perform preprocessing, time synchronization, and 3D reconstruction on the acquired data to construct a dynamic tidal 3D observation field;

[0008] Step 2: Based on the tidal dynamic three-dimensional observation field, analyze the changes in water cover and vegetation exposure at different tide levels, calculate the area of ​​the effective remote sensing observation area, and then establish the tidal observation perspective drift index to generate the tidal observation perspective drift index matrix.

[0009] Step 3: Spatial matching of the tidal observation view drift index matrix, radar backscattering coefficient data, and surface reflectivity data to establish a nonlinear mapping relationship, and then constructing a tidal view drift compensation model based on the mapping relationship to correct the original remote sensing observation signal.

[0010] Step 4: Based on the remote sensing feature data corrected by the tidal observation perspective, extract radar backscattering features, vegetation spectral features and tidal observation perspective stability features, and fuse and filter the features to obtain a set of carbon sink inversion feature parameters.

[0011] Step 5: Based on the set of carbon sink inversion characteristic parameters and mangrove sample plot survey data, establish a ground biomass inversion model, calculate the absolute carbon storage value of mangroves, and generate a dataset of mangrove carbon sink spatial distribution maps based on tidal observation perspective drift correction.

[0012] Furthermore, the data preprocessing and tidal dynamic three-dimensional observation field construction method in step one includes: performing point cloud filtering, ground point classification, and three-dimensional point cloud reconstruction on UAV lidar point cloud data to separate surface points and vegetation points, thereby generating a digital terrain model and a canopy height model; performing radiometric calibration, atmospheric correction, and geometric correction on multi-temporal optical remote sensing images to obtain surface reflectivity data; performing radiometric calibration, multi-view processing, filtering and noise reduction, and terrain correction on synthetic aperture radar images to obtain radar backscattering coefficient data; performing time synchronization interpolation on tidal water level sequences based on the absolute imaging time nodes corresponding to the multi-temporal optical remote sensing images and synthetic aperture radar images to obtain instantaneous tidal water level data corresponding to the remote sensing observation time; and constructing a tidal dynamic three-dimensional observation field dataset containing spatial pixel elevation information, vegetation vertical structure information, and tidal water body coverage status information based on the digital terrain model, canopy height model, and instantaneous tidal water level data.

[0013] Furthermore, the method for calculating the tidal observation perspective drift index and generating the matrix in step two includes: based on the tidal dynamic three-dimensional observation field dataset, analyzing the spatial variation relationships of water coverage, canopy exposure, trunk exposure, and root exposure under different tidal conditions, and determining the effective carbon observation area corresponding to each spatial pixel; constructing the tidal observation perspective drift index based on the area ratio relationship between the effective carbon observation area and the baseline carbon storage area; and generating the tidal observation perspective drift index matrix according to the calculation results of the tidal observation perspective drift index corresponding to each spatial pixel.

[0014] Furthermore, the formula for calculating the tidal observation perspective drift index is as follows:

[0015] ,

[0016] ,

[0017] in, express Each spatial pixel in The tidal observation perspective drift index corresponding to the given moment; Indicates the first Each spatial pixel in The first one that is always in an effective remote sensing observation state The area of ​​mangrove-like structural units, where structural units include root structure, trunk structure, and canopy structure; Indicates the first The contribution weight of mangrove-like structural units to remotely sensed carbon structure response signals; Indicates the preset reference tidal state time node; This indicates the number of mangrove structural unit categories.

[0018] Furthermore, the tidal observation viewpoint drift compensation and remote sensing signal correction method in step three includes: spatially matching the tidal observation viewpoint drift index matrix, radar backscattering coefficient data, and surface reflectivity data to establish a nonlinear mapping relationship between the tidal observation viewpoint drift index and the remote sensing observation signal; constructing a tidal viewpoint drift compensation model based on the nonlinear mapping relationship, and estimating the signal offset in the original remote sensing observation signal caused by changes in the tidal observation area to obtain the tidal viewpoint drift compensation amount; and correcting the original radar backscattering coefficient data and surface reflectivity data according to the tidal viewpoint drift compensation amount to obtain a standardized remote sensing carbon structure dataset after eliminating the influence of tidal observation viewpoint drift.

[0019] Furthermore, the tidal perspective drift compensation model is expressed as follows:

[0020]

[0021] in, Indicates remote sensing observation signals, This represents the signal variation term caused by the drift in the tidal observation perspective. This represents the response terms of the true carbon structure of mangroves. This represents the observation error term.

[0022] Furthermore, the carbon sink inversion feature parameter extraction and optimization method in step four includes: extracting radar polarization feature parameters, scattering structure parameters, and texture feature parameters from remote sensing feature data corrected for tidal observation perspective, and extracting normalized vegetation index, enhanced vegetation index, and vegetation spectral feature parameters from surface reflectance data; calculating the tidal perspective stability feature parameters of each spatial pixel within a continuous tidal cycle based on the tidal observation perspective drift index matrix to obtain tidal perspective stability features; fusing radar polarization feature parameters, scattering structure parameters, texture feature parameters, vegetation spectral feature parameters, and tidal perspective stability features to construct a mangrove carbon structure feature matrix corrected for tidal observation perspective; and using a feature contribution evaluation method to screen variables in the mangrove carbon structure feature matrix to obtain a set of target feature parameters for mangrove carbon storage inversion.

[0023] Furthermore, the method for mangrove carbon storage inversion and spatial distribution generation in step five includes: obtaining mangrove sample plot survey data for the corresponding region based on the carbon sink inversion feature parameter set, the mangrove sample plot survey data including aboveground biomass data, belowground biomass data, and carbon content per unit biomass data; establishing a nonlinear mapping model between the carbon sink inversion feature parameter set and mangrove aboveground biomass, inputting the carbon sink inversion feature parameter set into the nonlinear mapping model to obtain the mangrove aboveground biomass spatial distribution matrix; performing carbon storage conversion calculation on the mangrove aboveground biomass spatial distribution matrix according to the preset ratio coefficient of belowground biomass to aboveground biomass and the carbon content per unit biomass conversion coefficient to obtain the absolute carbon storage value of mangroves corresponding to each spatial pixel; and generating a mangrove carbon sink spatial distribution map dataset based on tidal observation perspective drift correction based on the absolute carbon storage value of mangroves.

[0024] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0025] This invention establishes a mapping relationship between the tidal observation perspective drift index and the radar backscattering coefficient and surface reflectivity changes. This allows for the acquisition of remote sensing response offsets caused by tidal level changes, and the use of these offsets to compensate and correct the original remote sensing observation signals. This enables remote sensing data acquired under different tidal conditions to be analyzed under a unified observation benchmark, improving the consistency and comparability between multi-temporal remote sensing data, reducing the interference of tidal level changes on the expression of remote sensing carbon structure information, and making the changes in remote sensing signals closer to the real responses generated by the structural changes of mangroves themselves.

[0026] By constructing a tidal observation perspective drift index, we can quantitatively characterize the changes in remote sensing observation perspective caused by changes in water coverage and mangrove structure exposure during tidal level changes. This transforms the impact of tidal disturbances on the remote sensing observation process from an uncertain environmental factor into a calculable drift parameter, enabling a description of changes in remote sensing observation conditions under different tidal levels and providing a foundation for subsequent remote sensing data consistency processing.

[0027] By considering the exposure differences and remote sensing response characteristics of different structural units such as mangrove roots, trunks, and canopies during tidal processes, the dynamic changes of effective remote sensing observation information in the intertidal environment can be described more accurately, improving the ability of remote sensing feature parameters to express the vertical structural information of mangroves. Simultaneously, carbon sink inversion based on remote sensing feature data corrected for tidal observation perspective drift can improve the stability of remote sensing input data in carbon storage estimation and enhance the reliability of monitoring results for the spatial distribution of mangrove carbon sinks.

[0028] By linking the tidal observation perspective drift index, perspective compensation process, and carbon sink inversion process, continuous processing from tidal disturbance identification and remote sensing observation perspective correction to dynamic carbon sink monitoring is achieved. This enables remote sensing monitoring of mangroves to adapt to the dynamic changes in observation conditions under complex tidal environments, providing effective technical support for long-term, continuous, and highly stable remote sensing monitoring of intertidal blue carbon ecosystems. Attached Figure Description

[0029] Figure 1 This is a flowchart of the steps of the present invention. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0031] Please refer to Figure 1 This embodiment of a dynamic monitoring method for mangrove carbon sequestration based on tidal disturbance compensation specifically includes:

[0032] Step 1: Acquire multi-temporal optical remote sensing images, synthetic aperture radar images, tidal water level sequences, digital elevation models, and UAV lidar point cloud data, and perform preprocessing, time synchronization, and 3D reconstruction on the acquired data to construct a dynamic tidal 3D observation field.

[0033] Specifically, the study focused on the mangrove monitoring area along the Shenzhen coast, selecting a 5 km² mangrove distribution area as the monitoring scope. Multi-temporal remote sensing data from January to December 2025 were acquired. This included 36 multi-temporal optical remote sensing images with a spatial resolution of 10 m, 48 synthetic aperture radar images with a spatial resolution of 10 m, 30 consecutive days of tidal level data with hourly intervals, and digital elevation model data with a spatial resolution of 12.5 m. Furthermore, a UAV equipped with a lidar device was used to conduct low-altitude scanning of the monitoring area, acquiring lidar point cloud data with an average point density of 150 points / m², resulting in approximately 750 million three-dimensional point cloud data points.

[0034] First, the acquired UAV LiDAR point cloud data was preprocessed. A progressive morphological filtering method was used to remove non-ground noise points, and ground points were classified according to elevation differences, resulting in two categories: surface points and vegetation points. Approximately 210 million surface points and 540 million vegetation points were ultimately retained. Based on the surface point data, a digital terrain model with a spatial resolution of 0.5 m was generated using spatial interpolation to characterize the micro-topographic undulations of the mangrove area. Vegetation height information was calculated based on the elevation difference between vegetation points and surface points, generating a canopy height model with a spatial resolution of 0.5 m to characterize the vertical structure of the mangrove forest.

[0035] Subsequently, radiometric calibration was performed on the acquired multi-temporal optical remote sensing images, converting the raw digital values ​​into sensor radiance values. An atmospheric correction model was then used to eliminate the effects of atmospheric scattering and absorption. At the same time, geometric correction was performed using ground control points to keep the spatial positioning error of the images within one pixel, thus obtaining surface reflectance data for the monitored area. The synthetic aperture radar images were then subjected to radiometric calibration, multi-view processing, filtering and noise reduction, and terrain correction. An adaptive filtering method with a 5×5 window was used to reduce the influence of speckle noise, and finally, radar backscattering coefficient data consistent with the spatial location of the optical remote sensing images were obtained.

[0036] Subsequently, based on the absolute imaging times recorded by optical remote sensing images and synthetic aperture radar images, the tidal level sequence is time-matched. For example, if an optical remote sensing image was acquired at 10:30 AM on June 15, 2025, and the tidal level monitoring data recording time interval is 1 hour, then the instantaneous tide level at that moment is calculated to be 1.26 m using a linear interpolation method. The same method is used to obtain the instantaneous tide level data for all remote sensing images at the corresponding times, thereby achieving synchronization between the remote sensing observation time and the tidal change process.

[0037] Finally, the digital terrain model with a spatial resolution of 0.5 m, the canopy height model with a spatial resolution of 0.5 m, and the instantaneous tidal level data at the corresponding remote sensing observation time were spatially fused. Using each spatial pixel as the basic unit, the tidal water cover status was determined based on the relationship between pixel elevation and tidal level. At the same time, the exposure status of the mangrove canopy, trunk, and root system was determined by combining the canopy height model. A dynamic three-dimensional tidal observation field dataset containing spatial pixel elevation information, vegetation vertical structure information, and tidal water cover status information was constructed, providing basic data for subsequent calculation of tidal observation perspective drift index and remote sensing monitoring of mangrove carbon sinks.

[0038] Step 2: Based on the tidal dynamic three-dimensional observation field, analyze the changes in water cover and vegetation exposure at different tide levels, calculate the area of ​​the effective remote sensing observation area, and then establish the tidal observation perspective drift index to generate the tidal observation perspective drift index matrix.

[0039] Specifically, based on the tidal dynamic three-dimensional observation field dataset constructed in step one, the tidal observation perspective drift index is calculated according to the exposure changes of mangrove vertical structural units under different tidal conditions, and a tidal observation perspective drift index matrix is ​​generated.

[0040] A spatial pixel numbered P (i=1250) within the monitoring area was selected as an example. This spatial pixel has a spatial resolution of 10 m × 10 m and a pixel area of ​​100 m². Based on the spatial elevation information, tidal level information, and vegetation vertical structure information in the tidal dynamic three-dimensional observation field, the mangrove structure within this spatial pixel was divided into three categories: root structure, trunk structure, and canopy structure, i.e., the number of structural unit categories. =3.

[0041] The changes in SAR backscattering coefficient and optical reflectivity, as well as structural information acquired by UAV lidar, were obtained from thirty mangrove sample plots. The remote sensing response contributions of different structural units were statistically analyzed. The remote sensing response contribution of the root system structure was calculated as follows: =15, contribution of remote sensing response to tree trunk structure: =45, contribution of remote sensing response of canopy structure: =90.

[0042] The contribution weights of the root system structure are calculated as follows:

[0043] =

[0044] Similarly, the contribution weights of the tree trunk structure are calculated. =0.3; Contribution weight of canopy structure =0.6. Therefore, this embodiment obtains ( , , = (0.1, 0.3, 0.6).

[0045] Determine the preset reference tidal state time node based on the dynamic three-dimensional tidal observation field. In this embodiment, the lowest tide state is selected as the baseline state, at which the tide level is 0.20 m. Spatial analysis is used to obtain the effective remote sensing observation area of ​​different structural units of this pixel under the baseline tidal state, including the root system area. =30m², =50m², the effective area of ​​the canopy structure is =100m². Therefore, the effective remote sensing structure contribution under the baseline tidal state is:

[0046] ,

[0047] Right now 8.

[0048] Furthermore, the remote sensing image acquisition time t was selected, corresponding to a tide level of 1.50 m. Based on the analysis of the tidal dynamic three-dimensional observation field, tidal cover prevents some structural units from being effectively observed by the remote sensing sensor. Among these, the effective area of ​​the root system structure... Effective observation area of ​​tree trunk structure Effective observation area of ​​canopy structure .

[0049] Therefore, the effective remote sensing structure contribution under the current tidal condition is:

[0050] ,

[0051] Right now .

[0052] The drift index is calculated based on the tidal observation perspective using the following formula:

[0053] .

[0054] Therefore, the tidal observation viewpoint drift index corresponding to this spatial pixel at a tide level of 1.50 m is 0.846. This result indicates that under the current tidal condition, due to water cover causing the root structure to be completely invisible and part of the trunk structure to be invisible, the effective remote sensing carbon structure contribution of this spatial pixel is reduced to 84.6% of the baseline state. Following the above method, pixel-by-pixel calculations were performed on all spatial pixels in the tidal dynamic three-dimensional observation field and at different remote sensing observation time points to obtain the corresponding values ​​for each spatial location. Numerical values ​​were generated and arranged according to spatial coordinates to form a tidal observation view drift index matrix. This matrix is ​​used to describe the variation characteristics of the effective remote sensing contribution of mangrove vertical structure under different tidal states, and serves as the input parameter for subsequent tidal view drift compensation model construction and remote sensing signal correction.

[0055] Step 3: Spatially match the tidal observation view drift index matrix, radar backscattering coefficient data, and surface reflectivity data to establish a nonlinear mapping relationship. Based on the mapping relationship, construct a tidal view drift compensation model to correct the original remote sensing observation signal.

[0056] Specifically, based on the tidal observation view drift index matrix obtained in step two, a typical spatial pixel numbered P (i=1250) is selected as the analysis object. The tidal observation view drift index corresponding to this spatial pixel is spatially matched with the synchronously acquired synthetic aperture radar image data and optical remote sensing image data. In this embodiment, by analyzing the tidal observation view drift index, radar backscattering coefficient, and surface reflectivity data corresponding to multiple consecutive remote sensing observation time nodes, a sample dataset describing the relationship between tidal observation view drift and remote sensing signal changes is constructed.

[0057] Drift index from the perspective of tidal observation Using the changes in radar backscattering coefficient and surface reflectivity as input variables, a nonlinear mapping relationship is established. This embodiment employs a nonlinear regression model to fit the relationship between the two variables, obtaining a mapping function between the tidal observation viewpoint drift index and the remote sensing signal variation term. Through model fitting, it was obtained that when the tidal observation perspective drift index corresponding to spatial pixel P (i=1250) changes from the baseline state... 1 reduced to At that time, due to the reduction in the effective contribution of remote sensing observation structures caused by tidal cover, the radar backscattering coefficient produced a systematic shift of about −1.1 dB, and the surface reflectivity produced a change of about −0.03.

[0058] Based on the established tidal perspective drift compensation model: ,in: This represents the original remote sensing observation signal; This represents the change in remote sensing signal caused by the drift of the tidal observation perspective. This represents the true carbon structure response of mangroves after eliminating the influence of tidal observation perspective drift; This represents random observation error.

[0059] Based on remote sensing observations of spatial pixel P at a tide level of 1.50 m. The original radar backscattering coefficient observation value at that moment was The following calculations were performed based on the tidal perspective drift compensation model: This means that changes in the tidal observation area cause a systematic decrease in the remote sensing signal relative to the baseline state of −1.1 dB. This is due to the observation error term. This represents random error, estimated through the residual term during model training. It is not directly deducted as part of the tidal system offset compensation. Therefore, this embodiment corrects for the systematic offset caused by tidal observation viewpoint drift. The original radar backscattering coefficient is corrected based on the tidal viewpoint drift compensation: -9.6 (-1.1) = -8.5 dB. Therefore, the radar backscattering coefficient of this spatial pixel after tidal perspective drift compensation is corrected from -9.6 dB to -8.5 dB. Simultaneously, for the surface reflectance data, the reflectance offset caused by tidal perspective drift is calculated to be -0.03 based on the same nonlinear mapping relationship. The original surface reflectance of this pixel is 0.32, so the corrected surface reflectance is 0.35. This allows for the acquisition of standardized remote sensing carbon structure data after eliminating the influence of tidal observation perspective drift.

[0060] Following the above method, pixel-by-pixel calculations are performed on all spatial pixels within the study area, based on the corresponding spatial locations. The exponential matrix is ​​used to estimate the tidal perspective drift compensation amount, and the radar backscattering coefficient data and surface reflectivity data are corrected respectively. Finally, a standardized remote sensing carbon structure dataset after tidal perspective drift correction is generated, which provides a data foundation for the subsequent extraction of characteristic parameters for mangrove carbon sink inversion.

[0061] Step 4: Based on the remote sensing feature data corrected by the tidal observation perspective, extract radar backscattering features, vegetation spectral features and tidal observation perspective stability features, and then fuse and filter the features to obtain a set of carbon sink inversion feature parameters.

[0062] Specifically, based on the standardized remote sensing carbon structure dataset obtained in step three, spatial pixel P (i=1250) was selected as an example to extract features from the remote sensing feature data after tidal observation viewpoint drift correction. First, radar polarization feature parameters, scattering structure parameters, and texture feature parameters were extracted from the corrected synthetic aperture radar data. The radar polarization feature parameters include the backscattering coefficients of the horizontal transmission and horizontal reception polarization channels, the backscattering coefficients of the horizontal transmission and vertical reception polarization channels, and the polarization ratio parameter. The extracted horizontal transmission and horizontal reception polarization backscattering coefficients for this pixel were -8.5 dB, -13.2 dB, and the polarization ratio parameter was 0.64. Simultaneously, scattering structure parameters and texture feature parameters were calculated based on the spatial grayscale variation characteristics of the radar image. The scattering structure parameter value was 0.72, the grayscale co-occurrence matrix contrast texture parameter was 18.6, and the mean texture parameter was 0.43.

[0063] Vegetation spectral characteristic parameters, including the Normalized Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI), were extracted from the corrected surface reflectance data. For each spatial pixel, the NDVI was calculated to be 0.78 and the EVI to be 0.52. Simultaneously, near-infrared and red band reflectance parameters were extracted, with a near-infrared reflectance of 0.46 and a red band reflectance of 0.09, used to characterize the mangrove vegetation cover and growth status.

[0064] Based on the tidal observation perspective drift index matrix, the variation of the tidal observation perspective of this spatial pixel within consecutive tidal cycles is analyzed. Specifically, tidal observation perspective drift index data for 10 consecutive tidal cycles are selected, with values ​​of 1.08, 1.12, 1.15, 1.20, 1.18, 1.25, 1.21, 1.17, 1.14, and 1.10, respectively. Tidal observation perspective stability characteristic parameters are calculated based on the data. The calculated mean of the tidal observation perspective drift index is 1.16, and the standard deviation is 0.055. The standard deviation characterizes the stability of the tidal observation perspective; a smaller standard deviation indicates greater stability under tidal influence in the region. Finally, the tidal perspective stability characteristic parameters corresponding to this spatial pixel are obtained.

[0065] Following the above method, features were extracted from all spatial pixels within the study area to obtain a multi-source remote sensing feature set, including radar polarization feature parameters, scattering structure parameters, texture feature parameters, vegetation spectral feature parameters, and tidal perspective stability features. These different types of features were normalized to convert each feature parameter to the range of 0–1, and then fused according to spatial pixel location to construct a mangrove carbon structure feature matrix corrected for tidal observation perspective. For example, for a monitoring area of ​​5 km², after being divided at a spatial resolution of 10 m × 10 m, a total of 50,000 spatial pixels were obtained. Each spatial pixel corresponds to a carbon structure feature vector containing 15 remote sensing feature parameters, ultimately forming a mangrove carbon structure feature matrix with a size of 50,000 × 15.

[0066] A feature contribution evaluation method was used to screen variables in the mangrove carbon structure feature matrix. This embodiment employed the random forest feature importance evaluation method, calculating the contribution of 15 candidate feature parameters to obtain an importance score for each parameter. For example, the radar polarization ratio parameter had a contribution of 0.18, the canopy scattering structure parameter had a contribution of 0.16, the normalized difference vegetation index (NDVI) had a contribution of 0.15, the tidal perspective stability parameter had a contribution of 0.12, and the texture contrast parameter had a contribution of 0.08. Features were then screened based on a preset contribution threshold of 0.05, retaining valid feature parameters with contributions higher than the threshold, ultimately obtaining a set of carbon storage inversion feature parameters containing eight target feature parameters.

[0067] The target feature parameter set includes radar polarization ratio parameters, scattering structure parameters, main texture feature parameters, normalized vegetation index, enhanced vegetation index, vegetation spectral feature parameters, and tidal perspective stability feature parameters, etc., which are used to establish a nonlinear mapping model between remote sensing feature parameters and mangrove aboveground biomass, and realize the spatial inversion of mangrove carbon storage.

[0068] Step 5: Based on the set of carbon sink inversion characteristic parameters and mangrove sample plot survey data, establish a ground biomass inversion model, calculate the absolute carbon storage value of mangroves, and generate a dataset of mangrove carbon sink spatial distribution maps based on tidal observation perspective drift correction.

[0069] Specifically, based on the set of target characteristic parameters for carbon storage inversion obtained in step four, field surveys were conducted on mangrove plots corresponding to the monitoring area to obtain plot survey data for model construction. In this embodiment, 30 mangrove plot survey quadrats were set up within the study area, each with an area of ​​20 m × 20 m. By measuring parameters such as tree height, diameter at breast height (DBH), and number of mangrove plants within the quadrats, and combining this with the mangrove biomass calculation model, aboveground biomass data corresponding to each plot were obtained. Simultaneously, underground biomass data were obtained through root sampling and proportion estimation methods, and carbon content per unit biomass was obtained through experimental measurement. For example, one typical plot, numbered S15, was found to have an aboveground biomass of 135 t / hm², an underground biomass of 27 t / hm², and a carbon content per unit biomass of 0.45, based on the survey calculations.

[0070] Based on the survey data of the aforementioned 30 sample plots, the target feature parameters for carbon storage inversion at the corresponding spatial locations of each sample plot obtained in step four are used as input variables for the model, and the measured aboveground biomass of the sample plots is used as the output variable for the model. A nonlinear mapping model between the set of target feature parameters for carbon storage inversion and the aboveground biomass of mangroves is established. In this embodiment, a random forest regression model is used for modeling. Eight target feature parameters corresponding to each sample plot are used as input variables, including radar polarization feature parameters, scattering structure parameters, texture feature parameters, vegetation spectral feature parameters, and tidal perspective stability feature parameters, etc. The measured aboveground biomass of the corresponding sample plot is used as the target variable. The nonlinear response relationship between remote sensing feature parameters and aboveground biomass is obtained through model training.

[0071] After the model is established, the set of target feature parameters corresponding to all spatial pixels in the study area is input into the nonlinear mapping model, and the spatial distribution matrix of mangrove aboveground biomass is calculated pixel by pixel. For example, for spatial pixel numbered P (i=1250), its input target feature parameters include: radar polarization ratio parameter 0.64, scattering structure parameter 0.72, gray-level co-occurrence matrix contrast texture parameter 18.6, mean texture parameter 0.43, normalized vegetation index 0.78, enhanced vegetation index 0.52, near-infrared band reflectivity 0.46, and tidal view stability parameter 0.055. After calculation by the nonlinear mapping model, the predicted aboveground biomass value corresponding to this spatial pixel is 150 t / hm². Using the same method, the calculation is performed on all 50,000 spatial pixels in the monitoring area to obtain the spatial distribution matrix of mangrove aboveground biomass for the entire study area.

[0072] Based on the ratio between underground and aboveground biomass in mangrove ecosystems and the carbon content per unit biomass, carbon storage conversion was performed on the spatial distribution matrix of aboveground biomass. In this embodiment, the ratio of underground to aboveground biomass was determined to be 0.20 based on the results of the sample plot survey, meaning that underground biomass is approximately 20% of aboveground biomass; the carbon content per unit biomass conversion factor was taken as 0.45. Taking spatial pixel P as an example, its predicted aboveground biomass value is 150 t / hm², then the corresponding underground biomass is:

[0073]

[0074] Therefore, the total biomass corresponding to this pixel is:

[0075]

[0076] Further calculations yielded the following absolute carbon storage value for the mangrove forest corresponding to this spatial pixel:

[0077]

[0078] That is, the carbon storage per unit area of ​​mangroves within this spatial pixel is 81 tC / hm².

[0079] Following the above method, carbon storage conversion calculations were performed on all spatial pixels within the study area to obtain the absolute carbon storage value of mangroves corresponding to each spatial pixel. This value was then rasterized based on spatial location relationships to generate a mangrove carbon sink spatial distribution map dataset based on tidal observation perspective drift correction. For example, within a monitoring area of ​​5 km², a total of 50,000 spatial pixel carbon storage data points were generated, with each spatial pixel corresponding to an absolute carbon storage value, thus forming a carbon sink distribution map that reflects the spatial differences in mangrove carbon storage across different regions.

[0080] Therefore, this scheme, by introducing remote sensing characteristic parameters corrected for tidal observation perspective drift during the carbon storage inversion process, reduces the impact of water cover changes under different tidal conditions on remote sensing signals and improves the stability and accuracy of remote sensing monitoring results for mangrove carbon sinks.

[0081] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the concept of this application. For example, technical solutions formed by substituting the above-described features with technical features having similar functions disclosed in this application.

Claims

1. A mangrove carbon sink dynamic monitoring method based on tide level disturbance compensation, characterized by: include, Step 1: Acquire multi-temporal optical remote sensing images, synthetic aperture radar images, tidal water level sequences, digital elevation models, and UAV lidar point cloud data, and perform preprocessing, time synchronization, and 3D reconstruction on the acquired data to construct a dynamic tidal 3D observation field; Step 2: Based on the tidal dynamic three-dimensional observation field, analyze the changes in water cover and vegetation exposure at different tide levels, calculate the area of ​​the effective remote sensing observation area, and then establish the tidal observation perspective drift index to generate the tidal observation perspective drift index matrix. Step 3: Spatial matching of the tidal observation view drift index matrix, radar backscattering coefficient data, and surface reflectivity data to establish a nonlinear mapping relationship, and then constructing a tidal view drift compensation model based on the mapping relationship to correct the original remote sensing observation signal. Step 4: Based on the remote sensing feature data corrected by the tidal observation perspective, extract radar backscattering features, vegetation spectral features and tidal observation perspective stability features, and fuse and filter the features to obtain a set of carbon sink inversion feature parameters. Step 5: Based on the set of carbon sink inversion characteristic parameters and mangrove sample plot survey data, establish a ground biomass inversion model, calculate the absolute carbon storage value of mangroves, and generate a dataset of mangrove carbon sink spatial distribution maps based on tidal observation perspective drift correction.

2. The method according to claim 1, wherein: The data preprocessing and tidal dynamic 3D observation field construction method in step one includes: performing point cloud filtering, ground point classification, and 3D point cloud reconstruction on UAV lidar point cloud data to separate surface points and vegetation points, thereby generating a digital terrain model and a canopy height model; performing radiometric calibration, atmospheric correction, and geometric correction on multi-temporal optical remote sensing images to obtain surface reflectivity data; performing radiometric calibration, multi-view processing, filtering and noise reduction, and terrain correction on synthetic aperture radar images to obtain radar backscattering coefficient data; performing time synchronization interpolation on tidal water level sequences based on the absolute imaging time nodes corresponding to the multi-temporal optical remote sensing images and synthetic aperture radar images to obtain instantaneous tidal water level data corresponding to the remote sensing observation time; and constructing a tidal dynamic 3D observation field dataset containing spatial pixel elevation information, vegetation vertical structure information, and tidal water body coverage status information based on the digital terrain model, canopy height model, and instantaneous tidal water level data.

3. The mangrove carbon sequestration monitoring method based on tidal disturbance compensation according to claim 1, characterized in that: The method for calculating the tidal observation perspective drift index and generating the matrix in step two includes: based on the tidal dynamic three-dimensional observation field dataset, spatial analysis is performed on the water coverage, root exposure, trunk exposure, and canopy exposure states within the mangrove area under different tidal conditions, thereby determining the effective observation range of remote sensing sensors for different vertical structural units of the mangrove under different tidal conditions; according to the differences in the contribution of different structural units such as mangrove roots, trunks, and canopies to the remote sensing observation signal, corresponding structural contribution weights are set, and based on the effective observation structural contribution under the current tidal state and the contribution of the complete mangrove structure under the baseline tidal state, a tidal observation perspective drift index is constructed; based on the calculation results of the tidal observation perspective drift index for each spatial pixel at different time nodes, a tidal observation perspective drift index matrix is ​​generated to characterize the degree of drift in remote sensing observation contributions caused by tidal changes.

4. The mangrove carbon sequestration monitoring method based on tidal disturbance compensation according to claim 3, characterized in that: The formula for calculating the tidal observation perspective drift index is: , , in, express Each spatial pixel in The tidal observation perspective drift index corresponding to the given moment; Indicates the first Each spatial pixel in The first one that is always in an effective remote sensing observation state The area of ​​mangrove-like structural units, where structural units include root structure, trunk structure, and canopy structure; Indicates the first The contribution weight of mangrove-like structural units to remotely sensed carbon structure response signals; Indicates the preset reference tidal state time node; This indicates the number of mangrove structural unit categories.

5. The mangrove carbon sequestration monitoring method based on tidal disturbance compensation according to claim 1, characterized in that: The method for compensating for tidal observation viewpoint drift and correcting remote sensing signals in step three includes: spatially matching the tidal observation viewpoint drift index matrix, radar backscattering coefficient data, and surface reflectivity data to establish a nonlinear mapping relationship between the tidal observation viewpoint drift index and the remote sensing signal; constructing a tidal viewpoint drift compensation model based on the nonlinear mapping relationship, and estimating the signal offset in the original remote sensing signal caused by changes in the tidal observation area to obtain the tidal viewpoint drift compensation amount; and correcting the original radar backscattering coefficient data and surface reflectivity data according to the tidal viewpoint drift compensation amount to obtain a standardized remote sensing carbon structure dataset after eliminating the influence of tidal observation viewpoint drift.

6. The mangrove carbon sequestration monitoring method based on tidal disturbance compensation according to claim 5, characterized in that: The tidal perspective drift compensation model is expressed as follows: in, Indicates remote sensing observation signals, This represents the signal variation term caused by the drift in the tidal observation perspective. This represents the response terms of the true carbon structure of mangroves. This represents the observation error term.

7. The mangrove carbon sequestration monitoring method based on tidal disturbance compensation according to claim 1, characterized in that: The carbon sink inversion feature parameter extraction and optimization method in step four includes: extracting radar polarization feature parameters, scattering structure parameters, and texture feature parameters from remote sensing feature data corrected for tidal observation perspective, and extracting normalized vegetation index, enhanced vegetation index, and vegetation spectral feature parameters from surface reflectance data; calculating the tidal perspective stability feature parameters of each spatial pixel within a continuous tidal cycle based on the tidal observation perspective drift index matrix to obtain tidal perspective stability features; fusing radar polarization feature parameters, scattering structure parameters, texture feature parameters, vegetation spectral feature parameters, and tidal perspective stability features to construct a mangrove carbon structure feature matrix corrected for tidal observation perspective; and using a feature contribution evaluation method to screen variables in the mangrove carbon structure feature matrix to obtain a set of target feature parameters for mangrove carbon storage inversion.

8. The mangrove carbon sequestration monitoring method based on tidal disturbance compensation according to claim 1, characterized in that: The method for mangrove carbon storage inversion and spatial distribution generation in step five includes: obtaining mangrove sample plot survey data for the corresponding region based on the carbon sink inversion feature parameter set, including aboveground biomass data, belowground biomass data, and carbon content per unit biomass data; establishing a nonlinear mapping model between the carbon sink inversion feature parameter set and mangrove aboveground biomass, inputting the carbon sink inversion feature parameter set into the nonlinear mapping model to obtain the mangrove aboveground biomass spatial distribution matrix; performing carbon storage conversion calculation on the mangrove aboveground biomass spatial distribution matrix according to the preset ratio coefficient of belowground biomass to aboveground biomass and the carbon content per unit biomass conversion coefficient to obtain the absolute carbon storage value of mangroves corresponding to each spatial pixel; and generating a mangrove carbon sink spatial distribution map dataset based on tidal observation perspective drift correction based on the absolute carbon storage value of mangroves.