A method and system for analyzing the dynamic evolution and driving forces of wetlands

Abstract

This invention belongs to the field of wetland evolution and driving force analysis technology, specifically involving a method and system for analyzing wetland dynamic evolution and driving forces. The method includes acquiring multi-temporal remote sensing images and corresponding measured water level data of the target lake's shoreline area; extracting the measured wetland boundaries and establishing a spatial distribution dataset of shoreline slope; determining the annual representative water level; calculating the horizontal displacement of the wetland boundaries segment by segment based on the shoreline slope spatial distribution dataset and water level deviation values; correcting the horizontal displacement of the measured wetland boundaries to obtain corrected wetland boundaries for each period under the annual representative water level; calculating the wetland area sequence based on the corrected wetland boundaries; analyzing the dynamic evolution patterns of the wetlands; extracting the annual average water level change; calculating the corresponding wetland area change segment by segment; and outputting the contribution of climate and hydrology to the dynamic evolution of the wetlands. This invention achieves accurate analysis of wetland evolution and effective decomposition of climate and hydrology driving forces in plateau rift lakes.

Description

Technical Field

[0001] This invention belongs to the field of wetland evolution and driving force analysis technology, specifically relating to a method and system for analyzing wetland dynamic evolution and driving forces. Background Technology

[0002] As an important ecosystem type, wetlands possess multiple functions, including hydrological regulation, biodiversity maintenance, and carbon sequestration. Accurate analysis of their dynamic evolution and driving mechanisms is fundamental to the ecological protection and management of lake basins. Due to their unique geological structure and climatic conditions, plateau rift lakes have developed typical wetland ecosystems in their riparian areas. Clarifying their evolutionary patterns and driving forces is of great significance for regional ecological assessment.

[0003] However, plateau rift lakes are characterized by steep shorelines. Under these topographical conditions, seasonal fluctuations in lake levels and water level rises and falls caused by interannual climate change result in significant horizontal shifts in wetland boundaries along the lakeshore. Even with limited water level changes, the gentle slopes along the shoreline can induce substantial horizontal displacements in the lateral direction. Furthermore, significant differences in slope across different shoreline sections lead to vastly different responses to the same water level changes, resulting in uneven spatial distribution of wetland boundaries extracted from remote sensing images across different periods. Moreover, because remote sensing images are acquired at different times of year and season, the lake levels corresponding to each image period are not identical. Therefore, the extracted wetland boundaries reflect the inundation state of the lakeshore under specific instantaneous water level conditions, rather than the inherent spatial extent of the wetland itself. This mechanism leads to discrepancies between wetland area observations from different periods, making it impossible to directly compare data from different periods under a unified water level benchmark. Consequently, a large number of false signals caused by water level fluctuations are mixed into the temporal changes in wetland area, making it difficult to obtain accurate results in the analysis of wetland dynamic evolution patterns and the quantitative decomposition of climate and hydrological driving forces based on observational data. Summary of the Invention

[0004] (1) Technical problems to be solved The purpose of this invention is to provide a method and system for analyzing the dynamic evolution and driving forces of wetlands, in order to solve the problems of uneven spatial displacement response of wetland boundaries caused by significant differences in the slope of the shoreline in plateau rift lakes, and inconsistent water levels at different times when remote sensing images are acquired, resulting in systematic deviations in the wetland area observed by remote sensing at different times, and thus inaccurate analysis of wetland dynamic evolution and driving forces.

[0005] (2) Technical solution To achieve the above objectives, on the one hand, the present invention provides a method for analyzing the dynamic evolution and driving forces of wetlands, the method comprising: Acquire multi-temporal remote sensing images of the target lake shore area and obtain measured water level data corresponding to the acquisition time of the multi-temporal remote sensing images; preprocess the multi-temporal remote sensing images and extract wetland boundaries to obtain measured wetland boundaries; obtain the surface slope of each shore section of the target lake shore area and establish a spatial distribution dataset of lake shore slope.

[0006] Determine the annual representative water level; calculate the water level deviation between the measured water level at the time of acquisition of multi-temporal remote sensing images and the corresponding annual representative water level based on the measured water level data; calculate the horizontal displacement of the wetland boundary caused by the water level deviation for each shore segment based on the spatial distribution dataset of lake shore slope and the water level deviation value.

[0007] Based on the measured horizontal displacement of the wetland boundary, the horizontal displacement of each bank segment of the measured wetland boundary is corrected to bring the measured wetland boundary down to the annual representative water level, thus obtaining the corrected wetland boundary.

[0008] The wetland area is calculated based on the corrected wetland boundary to obtain a wetland area sequence; the wetland area change rate for each time period is calculated based on the wetland area sequence to analyze the dynamic evolution pattern of the wetland.

[0009] The annual average water level for each time period is extracted from the measured water level data, and the annual average water level change for adjacent time periods is calculated. Based on the spatial distribution dataset of lakeside slope and the annual average water level change, the corresponding wetland area change is calculated for each shore segment. The wetland area change is compared with the wetland area change in the wetland area sequence to output the contribution of climate and hydrology to the dynamic evolution of wetlands.

[0010] Furthermore, the method for determining the annual representative water level includes: The arithmetic mean of the daily water levels for each year is calculated based on the measured water level data to obtain the annual average water level. The year is determined based on the imaging date of the remote sensing image, and the annual average water level of the corresponding year is used as the annual representative water level of the remote sensing image.

[0011] Furthermore, the method for calculating the horizontal displacement of the wetland boundary caused by the water level deviation on a segment-by-segment basis, based on the spatial distribution dataset of the lakeshore slope and the water level deviation value, includes: The slope values ​​corresponding to the central data of the lakeside slope spatial distribution are extracted point by point along the lake shoreline to obtain the shoreline slope sequence; the global mean of the shoreline slope sequence is calculated, and the difference between each slope value and the global mean is accumulated point by point in the shoreline slope sequence to obtain the cumulative deviation sequence; the turning point where the absolute value of the cumulative deviation in the cumulative deviation sequence changes from increasing to decreasing or from decreasing to increasing is identified and determined as the shoreline segment boundary.

[0012] The lake shore is divided according to the defined shoreline boundaries to obtain shoreline segmentation results. The median slope value within the shoreline segmentation results is taken to obtain the representative slope value of the shoreline segment. Based on the representative slope value of the shoreline segment and the corresponding water level deviation value, the horizontal displacement of the wetland boundary of each shoreline segment is calculated according to the trigonometric geometric relationship that the horizontal displacement of the wetland boundary is equal to the water level deviation value divided by the tangent of the representative slope value of the shoreline segment.

[0013] Furthermore, the method for identifying the turning point where the absolute value of the cumulative deviation in the cumulative deviation sequence changes from increasing to decreasing or from decreasing to increasing, and determining it as the boundary for shoreline segmentation, includes: Calculate the difference between the absolute values ​​of adjacent cumulative deviations for each position in the cumulative deviation sequence; and take the position where the difference between the absolute values ​​of adjacent cumulative deviations changes from positive to negative or from negative to positive as the candidate turning point.

[0014] The sum of the absolute values ​​of the differences between the cumulative deviation absolute values ​​at the candidate turning point and the cumulative deviation absolute values ​​at the immediately preceding and following candidate turning points is used to obtain the change range of the candidate turning point. For the first and last candidate turning points, the change range is used as a local reference benchmark. For the remaining candidate turning points, the average change range of the immediately preceding and following candidate turning points is used as a local reference benchmark.

[0015] Candidate turning points with a change magnitude not greater than that of the local reference benchmark are eliminated in one go to obtain the shoreline segmentation boundary.

[0016] Furthermore, the method for dividing the lake shore according to the shore segmentation boundary to obtain the shore segmentation result includes: Map the sequential positions along the shoreline corresponding to the shoreline segmentation boundary to two-dimensional lake shoreline geometric coordinates to obtain the spatial coordinates of the shoreline segmentation boundary; cut the lake shoreline into an initial shoreline segment set based on the spatial coordinates.

[0017] For each initial shore segment in the initial shore segment set, the spatial distance between adjacent points is calculated point by point along the shoreline direction to obtain a sequence of adjacent point distances; the positions in the adjacent point distance sequence that exceed a preset spatial continuity threshold are identified as spatial discontinuities; the initial shore segments containing spatial discontinuities are truncated at the spatial discontinuities to obtain independent continuous sub-shore segments; the independent continuous sub-shore segments are merged with the initial shore segments that do not contain spatial discontinuities to obtain the shore segment division results.

[0018] Furthermore, the method for truncating the initial shoreline containing the spatial discontinuity at the spatial discontinuity point to obtain independent continuous sub-shoreline segments includes: All spatial discontinuities within the initial shoreline containing spatial discontinuities are arranged in a shoreline order to obtain an ordered sequence of spatial discontinuities. Each spatial discontinuity in the ordered sequence is used as a dividing point, and the initial shoreline containing spatial discontinuities is sequentially divided in a shoreline order to obtain continuous boundary point segments.

[0019] For each continuous boundary point segment, calculate the tangent direction of the sub-shore segment end at the open endpoint; gradually extend the open endpoint along the tangent direction of the sub-shore segment end, and take the shoreline projection point when the extension point first enters the preset shoreline buffer range as the coordinate of the closed endpoint; replace the open endpoint of each continuous boundary point segment with the coordinate of the closed endpoint, and splice the line segment connecting the two closed endpoints of each continuous boundary point segment along the shoreline with each continuous boundary point segment to obtain an independent continuous sub-shore segment closed at both ends.

[0020] Furthermore, the method for correcting the horizontal displacement of the measured wetland boundary segment by segment based on the horizontal displacement of the wetland boundary, and correcting the measured wetland boundary to below the annual representative water level, to obtain the corrected wetland boundary includes: All measured wetland boundaries are discretized into an ordered sequence of boundary points; for each boundary point in the ordered sequence, the direction of the line connecting the preceding and following boundary points is taken as the local tangent direction at the boundary point; the direction perpendicular to the local tangent direction is taken as the displacement application direction of the boundary point, thus obtaining the displacement direction of each boundary point.

[0021] Based on the division results of each shoreline segment, the horizontal displacement of the wetland boundary corresponding to each shoreline segment is applied along the displacement application direction of each boundary point within each shoreline segment to obtain the corrected set of boundary point coordinates; the corrected set of boundary point coordinates is then connected sequentially according to the shoreline segment order to obtain the corrected wetland boundary corrected to the annual representative water level.

[0022] Furthermore, the method of obtaining the displacement direction of each boundary point by using the direction perpendicular to the local tangent direction as the displacement application direction of the boundary point includes: For each boundary point in the ordered boundary point sequence, calculate the radius of curvature at the boundary point based on the coordinates of the adjacent boundary points; use the radius of curvature as the search radius to extract the water body region centered on the boundary point; calculate the local centroid coordinates of the water body region to obtain the local pointing vector from the boundary point to the local centroid coordinates.

[0023] For the local tangent direction, two mutually opposite perpendicular directions are selected as candidate displacement directions; the candidate displacement directions whose dot product with the local pointing vector is negative are marked as land-side candidate directions, and the candidate displacement directions whose dot product is positive are marked as lake-inside candidate directions; based on the sign of the water level deviation value corresponding to the shore segment where the boundary point is located, the displacement application direction of each boundary point in the shore segment with a positive water level deviation value is determined as the lake-inside candidate direction, and the displacement application direction of each boundary point in the shore segment with a negative water level deviation value is determined as the land-side candidate direction, thus obtaining the displacement direction of each boundary point.

[0024] Furthermore, the method for calculating the wetland area change rate at each time period based on the wetland area sequence and analyzing the dynamic evolution pattern of wetlands includes: The wetland area sequence is subtracted pairwise from each adjacent period to obtain the wetland area net change sequence for each period. The wetland area net change sequence for each period is divided by the corresponding initial wetland area to obtain the wetland area change rate sequence for each period. Based on the temporal variation characteristics of the wetland area change rate sequence, the wetland dynamic evolution process is divided into stages of continuous shrinkage, continuous expansion and fluctuating alternation, thus obtaining the wetland dynamic evolution law.

[0025] Based on the same inventive concept, the present invention also provides a wetland dynamic evolution and driving force analysis system, the system comprising: The data acquisition and preprocessing module is used to acquire multi-temporal remote sensing images of the target lake shore area and acquire measured water level data corresponding to the acquisition time of the multi-temporal remote sensing images; preprocess the multi-temporal remote sensing images and extract wetland boundaries to obtain measured wetland boundaries; acquire the surface slope of each shore section of the target lake shore area and establish a spatial distribution dataset of lake shore slope.

[0026] The wetland boundary correction module is used to determine the annual representative water level; calculate the water level deviation between the measured water level at the time of acquisition of multi-temporal remote sensing images and the corresponding annual representative water level based on the measured water level data; calculate the horizontal displacement of the wetland boundary caused by the water level deviation for each shore segment based on the spatial distribution dataset of lake shore slope and the water level deviation value; and perform horizontal displacement correction on each shore segment of the measured wetland boundary based on the horizontal displacement of the wetland boundary to correct the measured wetland boundary to below the annual representative water level, thus obtaining the corrected wetland boundary.

[0027] The wetland evolution module is used to calculate the wetland area based on the corrected wetland boundary to obtain a wetland area sequence; based on the wetland area sequence, it calculates the wetland area change rate for each time period and analyzes the dynamic evolution pattern of the wetland.

[0028] The driving force analysis module is used to extract the annual average water level for each time period based on the measured water level data and calculate the annual average water level change between adjacent time periods. Based on the spatial distribution dataset of lake shore slope and the annual average water level change, the module calculates the corresponding wetland area change for each shore segment, compares it with the wetland area change in the wetland area sequence, and outputs the contribution of climate and hydrology to the dynamic evolution of wetlands.

[0029] (3) Beneficial effects Compared with the prior art, the beneficial effects of the present invention are: 1. By determining the annual representative water level, the horizontal displacement of the wetland boundary is calculated for each shore segment based on the spatial distribution dataset of lakeside slope and the water level deviation value. The horizontal displacement of the measured wetland boundary is then corrected for each shore segment. The measured wetland boundary for each period is then reduced to the annual representative water level to obtain the corrected wetland boundary, which effectively reduces the systematic deviation of the wetland boundary caused by water level fluctuations.

[0030] 2. By accumulating slope deviations point by point along the lake shoreline and identifying turning points, the shoreline is automatically divided into sections. The horizontal displacement of the wetland boundary of each section is calculated using the trigonometric relationship between the slope value and the water level deviation value of the representative section. The direction of displacement application is determined by combining the vertical direction of the local tangent at the boundary point with the centroid of the water body. This achieves fine correction for each section and improves the accuracy of wetland boundary calculation. Attached Figure Description

[0031] Figure 1 This is a flowchart of a wetland dynamic evolution and driving force analysis method according to Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the module composition of a wetland dynamic evolution and driving force analysis system according to Embodiment 2 of the present invention. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0033] Before providing examples, it is necessary to describe the application scenarios of this invention. This invention is applicable to the dynamic monitoring of wetlands in plateau rift lakes.

[0034] Example 1: As Figure 1 As shown in the figure, this embodiment provides a method for analyzing the dynamic evolution and driving forces of wetlands, the method including: S1. Acquire multi-temporal remote sensing images of the target lake shore area and obtain measured water level data corresponding to the acquisition time of the multi-temporal remote sensing images; preprocess the multi-temporal remote sensing images and extract the wetland boundary to obtain the measured wetland boundary; obtain the surface slope of each shore section of the target lake shore area and establish a spatial distribution dataset of lake shore slope.

[0035] For example, a plateau rift lake is taken as the target lake. The shoreline of the target lake has a relatively steep slope, with an average surface slope of about 5° to 25°.

[0036] Thirty-eight Landsat multi-temporal remote sensing images of the target lake's shoreline area over 31 consecutive years were acquired, with a spatial resolution of 30m. The temporal focus was primarily on the dry season, with seven years also including images from the wet season. The acquisition times of the images were kept as consistent as possible across years, and each image was selected based on cloud cover below 5%, with imaging time accurate to the day. The measured water level data corresponding to the acquisition times of the multi-temporal remote sensing images were obtained from daily water level records at the target lake's shoreline hydrological stations, with a time accuracy of one day. The daily water level value closest to the imaging date of each remote sensing image was used as the corresponding measured water level data for each period. The measured water level data for the 38 images range from 1886.2m to 1887.8m.

[0037] Radiometric calibration, atmospheric correction, and geometric registration were performed sequentially on multi-temporal remote sensing images. Radiometric calibration converted the raw DN values ​​to top-atmosphere reflectance. Atmospheric correction used the FLAASH model to convert top-atmosphere reflectance to surface reflectance. Geometric registration used a 1:50,000 topographic map as a reference, with registration errors controlled within 0.5 pixels. On the preprocessed images from each period, the improved normalized difference water index M was used to extract the water body extent. The formula for calculating M is: ; in, Green band surface reflectance, The surface reflectance is represented by the first band of shortwave infrared. Using M≥0 as the water body discrimination threshold, images from each period are classified into water bodies and non-water bodies. The set of pixels at the boundary between the water and non-water body classification results is used as the initial wetland boundary. After sub-pixel edge refinement, the initial wetland boundary is smoothly interpolated to an equivalent accuracy of 2.5m spatial resolution to obtain the measured wetland boundary for each period. It should be noted that this embodiment uses the water-land boundary line of a plateau rift lake as a proxy indicator for the water level-sensitive boundary of the lakeside wetland, reflecting the dynamic boundary between open water and lakeside land, which moves with changes in water level.

[0038] The surface slope of each section of the shoreline in the target lake area was derived from 30m resolution SRTM DEM data. Using the DEM raster within a 300m buffer zone from the lake shoreline towards the land as the analysis unit, a third-order unweighted difference algorithm was used to calculate the surface slope pixel-by-pixel, with slope values ​​expressed in degrees (°). The pixel-by-pixel slope calculation formula is as follows: ; in, pixel coordinates Surface slope (°); and Elevation The slope gradients in the east-west and north-south directions are calculated using the above formula, and the results are converted from radians to degrees. The above pixel-by-pixel surface slope raster is overlaid with a 300m buffer zone mask over the lakeshore, and the slope raster within the target lakeshore area is cropped to form a spatial distribution dataset of lakeshore slope. The spatial distribution dataset of lakeshore slope covers approximately 41.6km of the lakeshore.

[0039] S2. Calculate the arithmetic mean of the daily water levels for each year based on the measured water level data to obtain the annual average water level; determine the year based on the imaging date of the remote sensing image, and use the annual average water level of the corresponding year as the annual representative water level of the remote sensing image; calculate the water level deviation between the measured water level at the time of acquisition of the multi-temporal remote sensing image and the corresponding annual representative water level based on the measured water level data.

[0040] For example, the arithmetic mean of the daily measured water level data for each year is taken to obtain 31 annual average water level values, which constitute an annual average water level series. , The year is represented by the number of years. The 31 values ​​of the annual average water level sequence of the target lake are distributed between 1886.5m and 1887.6m, showing an overall interannual variation pattern of first decreasing and then increasing.

[0041] For 38 multi-temporal remote sensing images, the measured water level corresponding to the imaging date of each remote sensing image was read. Determine the year to which the image belongs. Extract the representative water level for the corresponding year. Calculate the water level deviation value : ; in, The image is captured on the date of its formation. For the first The actual water level measured on the day the landscape image was captured. This represents the annual representative water level for the year in which this image is located. Water level deviation values ​​corresponding to 38 images of the target lake. It is distributed between -0.65m and 0.62m.

[0042] The slope values ​​corresponding to the central data of the lakeside slope spatial distribution are extracted point by point along the lake shoreline to obtain the shoreline slope sequence; the global mean of the shoreline slope sequence is calculated, and the difference between each slope value and the global mean is accumulated point by point in the shoreline slope sequence to obtain the cumulative deviation sequence; Calculate the difference between the absolute values ​​of adjacent cumulative deviations for each position in the cumulative deviation sequence; and take the position where the difference between the absolute values ​​of adjacent cumulative deviations changes from positive to negative or from negative to positive as the candidate turning point.

[0043] The sum of the absolute values ​​of the cumulative deviation at the candidate turning point and the difference between the absolute values ​​of the cumulative deviation at the immediately preceding and following candidate turning points is calculated to obtain the variation range of the candidate turning point. For the first and last candidate turning points, the variation range is used as a local reference benchmark. For the remaining candidate turning points, the average variation range of the immediately preceding and following candidate turning points is used as a local reference benchmark. Candidate turning points with variation ranges not greater than the local reference benchmark are eliminated in one go to obtain the shoreline segmentation boundary.

[0044] Map the sequential positions along the shoreline corresponding to the shoreline segmentation boundary to two-dimensional lake shoreline geometric coordinates to obtain the spatial coordinates of the shoreline segmentation boundary; cut the lake shoreline into an initial shoreline segment set based on the spatial coordinates.

[0045] For example, sampling was conducted point-by-point at 30m intervals along the shoreline of the target lake in a clockwise direction, resulting in a total of 1387 shoreline sampling points (corresponding to a total shoreline length of approximately 41.6km). For each sampling point, the slope value of the nearest neighbor pixel in the lakeshore slope spatial distribution dataset was extracted, centered on the sampling point coordinates, to obtain the shoreline slope sequence. ( ), the unit is degrees (°).

[0046] Calculate the coastal slope sequence global mean , obtained the target lake .

[0047] By summing the differences between each slope value and the global mean at each point, a cumulative deviation sequence is obtained. : ; Cumulative Deviation Series initial value Termination value The absolute value reflects the total accumulated amount of global slope deviation.

[0048] For cumulative deviation series Calculate the difference in absolute value of adjacent cumulative deviations for each position. : ; in, Therefore 0. Will From positive to negative (i.e.) and ) or from negative to positive (i.e. and ) position Marked as a candidate turning point. A sign shift means... At this location, the cumulative deviation sequence either changes from increasing to decreasing or from decreasing to increasing, meaning it reaches a local maximum or local minimum. A total of 94 candidate turning points were identified for the target lake, with an average spacing of approximately 14 sampling points between adjacent candidate turning points, corresponding to a shoreline length of approximately 420 m.

[0049] For each candidate turning point ( ), range of change Calculated according to the following rules: ; in, and The first The candidate turning point is the one immediately preceding and following the candidate turning point. The total absolute value of the cumulative deviation jump between candidate inflection points and their adjacent candidate inflection points was quantified. The larger the value, the more significant the slope change at the candidate inflection point. (First and last candidate inflection points) and If there are only one-sided adjacent terms, the formula for calculating the magnitude of change degenerates into the difference of one-sided absolute values.

[0050] Local reference datum Determined according to the following rules: ; The first and last candidate turning points are used as local reference benchmarks based on their own change range, rather than on the change range of adjacent positions.

[0051] The magnitude of change Candidate turning points are considered pseudo-turns caused by slight local slope disturbances and are removed from the candidate turning point set in one go. After removal, the remaining candidate turning points are renumbered, and the change magnitude and local reference benchmark are not recalculated. The reason for performing the removal only once is that if it is performed iteratively, the adjacency relationship of the candidate turning point set changes after each round of removal. and The calculation results change accordingly, which may lead to significant turning points that should have been retained being mistakenly removed in subsequent rounds, thus compromising the stability of shoreline segmentation. After a one-time removal process, the target lake retained 37 shoreline segmentation boundaries, corresponding to locations where the slope characteristics along the shoreline changed significantly. The average distance between adjacent shoreline segmentation boundaries was approximately 37 sampling points, corresponding to a shoreline length of approximately 1.1 km.

[0052] The serial numbers of the sampling points along the 37 shoreline sections dividing the boundary were determined. Mapping to two-dimensional lake shoreline geometric coordinates, the mapping rule is to map along the shoreline in sequence. Planar coordinates of each sampling point in the WGS 84 / UTM projection coordinate system The spatial coordinates serve as the boundaries for dividing the shoreline. Based on the spatial coordinates of the 37 shoreline dividing boundaries, the target lake shoreline vector line elements are truncated at each shoreline dividing boundary to obtain 38 initial shoreline segments, forming an initial shoreline set.

[0053] For each initial shore segment in the initial shore segment set, the spatial distance between adjacent points is calculated point by point along the lake shoreline to obtain a sequence of adjacent point distances; positions in the adjacent point distance sequence that exceed a preset spatial continuity threshold are identified as spatial discontinuities. All spatial discontinuities within the initial shoreline containing spatial discontinuities are arranged in a shoreline order to obtain an ordered sequence of spatial discontinuities. Each spatial discontinuity in the ordered sequence is used as a dividing point, and the initial shoreline containing spatial discontinuities is sequentially divided in a shoreline order to obtain continuous boundary point segments.

[0054] For the open endpoints of each continuous boundary point segment, calculate the tangent direction of the sub-shore segment end at the open endpoint; extend the open endpoints step by step along the tangent direction of the sub-shore segment end, and use the shoreline projection point when the extended point first enters the preset shoreline buffer range as the coordinates of the closed endpoint; replace the open endpoints of each continuous boundary point segment with the coordinates of the closed endpoints, and splice the line segment connecting the two closed endpoints of each continuous boundary point segment along the shoreline with each continuous boundary point segment to obtain an independent continuous sub-shore segment closed at both ends; merge the independent continuous sub-shore segment with the initial shore segment without spatial discontinuities to obtain the shore segment division results.

[0055] For example, the 38 initial shore segments in the initial shore segment set are processed segment by segment. Each initial shore segment consists of several ordered boundary points, and the spatial distance between adjacent boundary points is... Calculated using Euclidean distance: ; in, and The coordinates of two sequentially adjacent points along the initial shoreline in the WGS 84 / UTM projection coordinate system are given. The point number is calculated for each pair of adjacent points. This forms a sequence of distances between adjacent points.

[0056] Preset spatial continuity threshold The value is taken as 3 times the spatial resolution of the remote sensing image, that is... The rationale for using 3 times the pixel size is that the spatial accuracy of the wetland boundary extraction results is limited by the 30m resolution image. The distance between adjacent boundary points in a normal continuous shoreline does not exceed one pixel diagonal length (approximately 42m). When the distance between adjacent points exceeds 90m, it can be determined that there is an actual spatial break in the wetland boundary at that location, which is due to a discontinuous jump rather than normal geometric curvature. In the adjacent point distance sequence... The point corresponding to the position of m The points were identified as spatial discontinuities. Of the 38 initial shoreline segments in the target lake, 5 segments contained spatial discontinuities, while 33 initial shoreline segments without spatial discontinuities were directly retained.

[0057] Five initial shoreline segments containing spatial discontinuities were processed separately. All spatial discontinuities within each initial shoreline segment were arranged in coastal order, resulting in an ordered sequence of spatial discontinuities. Using each spatial discontinuity in the ordered sequence as a dividing point, the initial shoreline segments containing spatial discontinuities were sequentially divided: each spatial discontinuity breaks the shoreline at that point; the spatial discontinuity itself is not included in any continuous boundary point segment. Among the endpoints of each segmented continuous boundary point segment, the endpoint connecting to other continuous boundary point segments is the closed endpoint, i.e., the endpoint of the original shoreline segment; the endpoint facing the spatial discontinuity is the open endpoint. After initial shoreline segmentation at spatial discontinuities, we obtain... Each of the five initial shoreline segments containing spatial discontinuities in the target lake has one to three spatial discontinuities, resulting in a total of 13 continuous boundary point segments and 16 open endpoints.

[0058] The 16 open endpoints of the 13 consecutive boundary point segments are processed one by one. The tangent direction of the sub-bank segment end at each open endpoint is determined by the direction of the line connecting the open endpoint and the second adjacent inner boundary point: Let the coordinates of the open endpoint be... The coordinates of the second adjacent inner boundary point are The tangent vector at the end of the sub-bank section is After normalization, the unit tangent direction vector is obtained. The purpose of choosing the second adjacent inner point instead of the first immediately adjacent point is to reduce the impact of local orientation jitter caused by boundary discretization on tangent estimation.

[0059] The selection principle for the preset shoreline buffer range is: it should be greater than the extension step size to ensure no missed detections occur during step-by-step judgment, and less than the wetland boundary discretization spacing to avoid misjudging shorelines at non-target locations as termination points. In this embodiment, the extension step size is 1m and the wetland boundary discretization spacing is 2.5m. Considering the above constraints, the preset shoreline buffer range is selected as follows: The endpoint will be opened along the tangent direction at the end of the sub-bank section. Extend the line step by step in increments of 1 meter, and calculate the coordinates of each extension point. ( The termination condition is that the shortest distance between the extension point and the shoreline is less than the preset shoreline buffer zone for the first time. The perpendicular coordinates of the extension points that meet the termination condition are projected onto the shoreline as the coordinates of the closed endpoints. The maximum extension step limit is set at 500 steps (corresponding to 500m). This limit is based on the fact that discontinuities in wetland boundary space are usually caused by local breaks in the extracted boundary at complex terrain features such as estuaries and islands. The actual break distance is usually on the order of tens of meters, and 500m can cover extreme terrain conditions. If the extension step limit is exceeded but the termination condition is not met, the shoreline projection point of the extension point with the smallest distance from the shoreline during the extension process is taken as the coordinate of the closed endpoint. The actual extension distances of the 16 open endpoints of the target lake are all between 8m and 63m, and the limit is not triggered.

[0060] With closed endpoint coordinates Replace the open endpoint coordinates of each consecutive boundary point segment After replacement, both ends of each continuous boundary point segment are closed endpoints. A connecting line segment is extracted along the shoreline between the closed endpoint and the original open endpoint: on the shoreline vector line elements, the shoreline arc segment between the shoreline projection point corresponding to the open endpoint and the coordinates of the closed endpoint is extracted. This shoreline arc segment is then joined end-to-end with the continuous boundary point segment at the closed endpoint to obtain an independent continuous sub-shoreline segment with closed ends. The initial shoreline segments containing spatial discontinuities in the target lake, after the above processing, yielded 13 independent continuous sub-shoreline segments with closed ends. These 13 independent continuous sub-shoreline segments with closed ends are merged with 33 initial shoreline segments without spatial discontinuities, resulting in 46 shoreline segment divisions.

[0061] The median slope value within the segment division results is taken to obtain the representative slope value of the segment. Based on the representative slope value of the segment and the corresponding water level deviation value, the horizontal displacement of the wetland boundary of each segment is calculated according to the trigonometric geometric relationship that the horizontal displacement of the wetland boundary is equal to the water level deviation value divided by the tangent of the representative slope value of the segment.

[0062] All measured wetland boundaries are discretized into an ordered sequence of boundary points. For each boundary point in the ordered sequence, the direction of the line connecting the preceding and following boundary points is taken as the local tangent direction at the boundary point. For each boundary point in the ordered sequence, the radius of curvature at the boundary point is calculated based on the coordinates of the preceding and following boundary points. Using the radius of curvature as the search radius, the water body region centered on the boundary point is extracted. The local centroid coordinates of the water body region are calculated to obtain the local pointing vector from the boundary point to the local centroid coordinates of the water body.

[0063] For the local tangent direction, two mutually opposite perpendicular directions are selected as candidate displacement directions; the candidate displacement directions whose dot product with the local pointing vector is negative are marked as land-side candidate directions, and the candidate displacement directions whose dot product is positive are marked as lake-inside candidate directions; based on the sign of the water level deviation value corresponding to the shore segment where the boundary point is located, the displacement application direction of each boundary point in the shore segment with a positive water level deviation value is determined as the lake-inside candidate direction, and the displacement application direction of each boundary point in the shore segment with a negative water level deviation value is determined as the land-side candidate direction, thus obtaining the displacement direction of each boundary point.

[0064] Based on the division results of each shoreline segment, the horizontal displacement of the wetland boundary corresponding to each shoreline segment is applied along the displacement application direction of each boundary point within each shoreline segment to obtain the corrected set of boundary point coordinates; the corrected set of boundary point coordinates is then connected sequentially according to the shoreline segment order to obtain the corrected wetland boundary corrected to the annual representative water level.

[0065] For example, the slope values ​​of all pixels falling within the shoreline range in the spatial distribution dataset of lakeside slopes for each of the 46 shoreline segments are extracted, and the median is taken as the representative slope value of the shoreline segment. ( The reason for using the median instead of the mean is that there are steep cliffs and alluvial platforms at the mouth of the target lake in some areas, and the corresponding slope values ​​deviate significantly from the main topography of the shoreline. The median is more resistant to such outliers than the mean.

[0066] Horizontal displacement of wetland boundaries in each section The slope value represented by the shoreline With the Water level deviation value corresponding to the landscape image Calculated using trigonometric geometric relationships: ; in, The unit is meters. The unit is degrees (°). The unit is meters (m). Positive values ​​indicate that the wetland boundary should be corrected towards the inland side of the lake, while negative values ​​indicate that the wetland boundary should be corrected towards the land side. Taking a section of the northwest shore of the target lake as an example, the representative slope value is... The water level deviation value corresponding to the 8th remote sensing image (high water season) m, the horizontal displacement of the wetland boundary is calculated. m; Representative slope value of a certain section of the southeast coast. The same scene remote sensing image corresponds m. A total of 1748 horizontal displacements of wetland boundaries were calculated from 38 remote sensing images and 46 shoreline segments. The absolute values ​​range from 0.0m to 5.53m.

[0067] The measured wetland boundaries of 38 periods were discretized into an ordered sequence of boundary points. The discretization interval was 2.5m, consistent with the equivalent accuracy of refining the wetland boundaries to 2.5m in the first step. Interpolation along the equal arc length of the measured wetland boundary vector profile resulted in a number of boundary points between 11,200 and 14,700 after discretization for each period. These points were arranged clockwise to form an ordered sequence of boundary points. ( , (This varies depending on the period). For each boundary point in the ordered sequence of boundary points. Pi is assigned to the nearest perpendicular segment on the shoreline vector element, thereby establishing the correspondence between each boundary point and the shoreline segment.

[0068] For ordered boundary point sequences Each boundary point in Local tangent direction vector From adjacent boundary points and The direction of the connecting line is determined: ; The first and last boundary points of an ordered sequence of boundary points ( and Processed in a ring shape. Pick , Pick .

[0069] Boundary points radius of curvature at Calculated using the coordinates of adjacent boundary points according to the three-point circle fitting formula: ; When the three points are collinear, the denominator approaches zero, at which point the radius of curvature... Theoretically, it tends towards infinity; in practice, it will... The upper limit is set at 500m to avoid the search radius being too large, causing the water area extraction range to cross the shoreline boundary. The lower limit is not truncated; the actual calculated value of the target lake is used. It is distributed between 4.1m and 500m (including truncation).

[0070] With radius of curvature Using the search radius, boundary points are extracted from the preprocessed remote sensing images of each period. Centered on, with radius The set of pixels with M≥0 within a circular region is used as the boundary point. The centroid coordinates of the local water body are determined by calculating the arithmetic mean of the coordinates of the centers of all pixels within the water body region. Boundary points Local pointing vector to the local water body centroid coordinates for: ; Local pointing vector No normalization is applied; it is used only for direction determination. When there are no pixels with M≥0 within the circular region (the boundary point falls on the edge of an extremely narrow body of water), the search radius is expanded to... If no results are found after re-extraction, the average coordinates of the center of the pixels in the entire lake are taken as the local centroid coordinates of the water body. This situation occurred 7 times in 38 periods of data for the target lake, all of which were located in the extremely shallow shoals at the river mouth where the lake flows into the lake.

[0071] unit vector of local tangent direction By taking two mutually opposite perpendicular directions, we can obtain the candidate displacement directions. and Calculate the local pointing vector. respectively with and dot product: ; Candidate displacements with a negative dot product point away from the local water body's centroid, i.e., towards the land side, and are marked as land-side candidate directions; candidate displacements with a positive dot product point towards the local water body's centroid, i.e., towards the inner side of the lake, and are marked as inner-lake-side candidate directions.

[0072] Based on boundary points Water level deviation value corresponding to the assigned shore section The sign of the sign determines the direction of displacement at each boundary point: The displacement direction of each boundary point within the shoreline is determined as the candidate direction on the inner side of the lake, indicating that the water level is higher than the annual representative water level during imaging, and the wetland boundary needs to be corrected by shrinking towards the inner side of the lake; The displacement direction of each boundary point within the shoreline is determined as the candidate direction on the land side, indicating that the water level is lower than the annual representative water level during imaging, and the wetland boundary needs to be expanded and corrected towards the land side. No displacement is applied to the shoreline, and the coordinates of the boundary points remain unchanged.

[0073] Boundary points Horizontal displacement of wetland boundary along the direction of displacement application ( Boundary point (The corresponding shoreline section number) is used to obtain the corrected boundary point coordinates. : ; in, Boundary point The displacement is applied in a unit vector direction (candidate direction on the land side or candidate direction on the lake side). The absolute value is used in coordinate calculations, and the direction is controlled by the displacement applied direction unit vector. The corrected set of boundary point coordinates. According to the original order of the ordered boundary point sequence ( By connecting these points sequentially, the corrected wetland boundaries are obtained at the annual representative water level. After correcting the measured wetland boundaries of the target lake for 38 periods, 38 corrected wetland boundaries are obtained. The absolute value of the difference between the closed polygon area of ​​the corrected wetland boundary and the measured wetland boundary area for each period is distributed between 0.01 km² and 0.12 km².

[0074] S3. Calculate the wetland area based on the corrected wetland boundary to obtain a wetland area sequence; calculate the difference between each pair of adjacent periods in the wetland area sequence to obtain a sequence of net changes in wetland area for each period; divide the net change in wetland area for each period by the corresponding initial wetland area to obtain a sequence of wetland area change rates for each period; based on the temporal variation characteristics of the wetland area change rate sequence, divide the wetland dynamic evolution process into stages of continuous shrinkage, continuous expansion, and alternating fluctuations to obtain the dynamic evolution law of wetlands.

[0075] For example, the area of ​​the closed polygon of the 38-period corrected wetland boundary is calculated period by period, using the Gaussian area formula according to the ordered boundary point sequence: ; in, For the first The first phase of wetland boundary correction The corrected boundary point coordinates For the first The total number of boundary points of the wetland boundary is corrected periodically, and the subscripts are treated as rings. Time to take ), Units are The target lake's corrected wetland boundary area (Phase 38). Distributed in to Between these, arranged by imaging year, a wetland area sequence is formed. ( When multiple periods of corrected wetland boundaries exist in the same year, the arithmetic mean of the areas of the corrected wetland boundaries in each period is taken as the wetland area sequence value for that year. The target lake has two images in each of the seven years. After taking the average, the wetland area sequence still maintains 31 annual values, which are aligned with the time axis of the annual average water level sequence.

[0076] wetland area sequence By taking the difference between adjacent periods, the sequence of net changes in wetland area for each period can be obtained. : ; Indicates the wetland area in the th order. To the Net increase during the yearly period, This indicates a net decrease. The time series shows the net change in wetland area of ​​the target lake over various periods. The 30 values ​​are distributed in to between.

[0077] Net change in wetland area at different time periods Divide by the corresponding initial wetland area The sequence of wetland area change rate for each time period was obtained. : ; Indicates the first The net increase in wetland area during the period This indicates a net decrease. The time series of wetland area change rates for the target lake. The 30 values ​​are distributed in to between.

[0078] Based on the wetland area change rate series The dynamic evolution stages of wetlands are divided based on their temporal variation characteristics, and the division rules are as follows: If continuous One or more time periods All values ​​are negative, indicating a continuous shrinkage phase; if continuous One or more time periods All values ​​are positive, indicating a continuous expansion phase; the remaining alternating positive and negative intervals are identified as fluctuation and alternation phases. This is a preset threshold for the number of consecutive time periods, and its value is a positive integer. The selection criteria for the values ​​are as follows: Annual images in the wetland area sequence may be distributed across different seasons. Although annual representative water level corrections have been completed for wetland boundaries, residual seasonal disturbances may still be introduced into the annual average area sequence when the temporal distribution of image acquisition is uneven. The probability of consecutive unidirectional changes over one or two consecutive annual periods falling into residual disturbances is relatively high, while consecutive unidirectional changes over three or more consecutive annual periods can be identified as trend evolution signals after water level correction has been completed. When the temporal distribution of the images is relatively uniform, a value can be selected. When the phase distribution is uneven, it is recommended to take to In this embodiment, .

[0079] target lake wetland area change rate series The analysis presents a three-segment structure: the analysis period consists of 12 consecutive periods from year 1 to year 13. All values ​​were negative, indicating a continuous shrinkage phase, with the wetland area decreasing from 18.76 km² to 13.41 km². From year 13 to 22, there were nine alternating periods of positive and negative values, indicating a fluctuating phase, with the wetland area oscillating between 12.8 km² and 14.6 km². From year 22 to 31, there were nine consecutive periods... All values ​​were positive, indicating a continuous expansion phase, with the corresponding wetland area increasing from 13.05 km² to 16.38 km². The division into three phases represents the dynamic evolution pattern of the wetland.

[0080] S4. Extract the annual average water level for each time period from the measured water level data, and calculate the annual average water level change for adjacent time periods; calculate the corresponding wetland area change for each shore segment based on the spatial distribution dataset of lakeside slope and the annual average water level change, compare it with the wetland area change in the wetland area sequence, and output the contribution of climate and hydrology to the dynamic evolution of wetlands.

[0081] For example, the arithmetic mean of daily water levels is extracted from measured water level data over the years to obtain 31 annual average water level values, which are then compared with the aforementioned annual average water level sequence. Completely identical, with no double counting. The wetland area sequence... Annual index and average annual water level sequence After confirming the one-to-one correspondence between the years and the timelines, the annual average water level change sequence for adjacent periods is obtained by subtracting adjacent years. : ; Annual average water level change series of the target lake in adjacent time periods The 30 values ​​are distributed in m to Between m, and the wetland area sequence The general direction of change is consistent across the corresponding time periods, but the magnitude of the change varies depending on the slope of the shoreline.

[0082] The changes in wetland area corresponding to the annual average water level changes were calculated segment by segment for each of the 46 shoreline sections. In the Changes in wetland area over time period Approximate by the product of the horizontal displacement of the wetland boundary caused by water level changes and the effective arc length of the shoreline: in, For the first The slope value (°) represents the slope of the shoreline. For the first Arc length of the shoreline (m). For the first Average annual water level change (m) between adjacent time periods. The unit is m², which can be converted to km². This indicates that the annual rise in water level has led to the first The area of ​​wetlands along the shoreline has increased (the wetland boundary has expanded towards the land side). This indicates that the decrease in average annual water level has led to a reduction in wetland area. The arc lengths of the 46 shoreline segments of the target lake are shown. Distributed between 0.3km and 3.2km, with a total arc length of 41.6km, consistent with the aforementioned shoreline division results.

[0083] The results of the 46 shoreline segment divisions were presented in the [number]th [year]. Changes in wetland area over time period Summing, we get the first... Changes in the total wetland area of ​​the entire lake caused by climate and hydrology over a period of time : ; Target lake in 30 time periods Distributed in km² to Between km²

[0084] Will Changes in wetland area in the wetland area series To make a comparison, define the first The proportion of the contribution of climate and hydrology to wetland dynamic evolution over time period : ; when When =0 (i.e., there is no net change in wetland area between adjacent years), Meaningless, recorded as null, and not included in subsequent mean calculations. The range of values ​​is not limited to ,when This indicates that the direction of climate and hydrological driving forces is consistent with the actual change in wetland area; when This indicates that the direction of climate and hydrology driving change is opposite to the actual direction of change, and non-climate and hydrology driving factors (including land use changes such as reclamation and conversion of farmland to wetlands) dominate wetland evolution during this period.

[0085] Changes in wetland area not driven by climate and hydrology for: ; Climate and hydrological driving contribution ratios of the target lake over 30 time periods Distributed in to Between. During the period of sustained contraction (years 1 to 13, spanning 12 years), The mean is This indicates that climate and hydrology drive (the annual average water level continues to decline, and the annual average water level change) The mean is m) contributed approximately 7% to the wetland area shrinkage, with the remaining 93% caused by non-climate and hydrological driving factors such as reclamation; during the continuous expansion phase (years 22 to 31, spanning 9 years), The mean is This indicates that the average annual water level has risen ( The mean is (m) is the main driver of wetland area recovery, with non-climate and hydrological drivers contributing approximately 93%. The fluctuation and alternation phase (years 13 to 22, spanning 9 years) is significant. The fact that the value exceeds 100% or is negative during certain periods indicates that there is a mutual offsetting effect between climate-driven and non-climate-driven hydrological factors.

[0086] 30 time periods , and The data is organized and output in chronological order to form a table showing the contribution of climate and hydrology to the dynamic evolution of wetlands. The table provides the wetland area change caused by climate and hydrology, the wetland area change caused by non-climate and hydrology, and the contribution ratio of climate and hydrology for each time period.

[0087] Example 2: Based on the same inventive concept, such as Figure 2 As shown, this embodiment also provides a wetland dynamic evolution and driving force analysis system, the system comprising: The data acquisition and preprocessing module is used to acquire multi-temporal remote sensing images of the target lake shore area and acquire measured water level data corresponding to the acquisition time of the multi-temporal remote sensing images; preprocess the multi-temporal remote sensing images and extract wetland boundaries to obtain measured wetland boundaries; acquire the surface slope of each shore section of the target lake shore area and establish a spatial distribution dataset of lake shore slope.

[0088] The wetland boundary correction module is used to determine the annual representative water level; calculate the water level deviation between the measured water level at the time of acquisition of multi-temporal remote sensing images and the corresponding annual representative water level based on the measured water level data; calculate the horizontal displacement of the wetland boundary caused by the water level deviation for each shore segment based on the spatial distribution dataset of lake shore slope and the water level deviation value; and perform horizontal displacement correction on each shore segment of the measured wetland boundary based on the horizontal displacement of the wetland boundary to correct the measured wetland boundary to below the annual representative water level, thus obtaining the corrected wetland boundary.

[0089] The wetland evolution module is used to calculate the wetland area based on the corrected wetland boundary to obtain a wetland area sequence; based on the wetland area sequence, it calculates the wetland area change rate for each time period and analyzes the dynamic evolution pattern of the wetland.

[0090] The driving force analysis module is used to extract the annual average water level for each time period based on the measured water level data and calculate the annual average water level change between adjacent time periods. Based on the spatial distribution dataset of lake shore slope and the annual average water level change, the module calculates the corresponding wetland area change for each shore segment, compares it with the wetland area change in the wetland area sequence, and outputs the contribution of climate and hydrology to the dynamic evolution of wetlands.

[0091] It should be noted that the specific ways in which each module performs operations in the system described in the above embodiments have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0092] Finally, it should be noted that although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for analyzing the dynamic evolution and driving forces of wetlands, characterized in that, The method includes: Acquire multi-temporal remote sensing images of the target lake shore area and obtain measured water level data corresponding to the acquisition time of the multi-temporal remote sensing images; preprocess the multi-temporal remote sensing images and extract wetland boundaries to obtain measured wetland boundaries; obtain the surface slope of each shore section of the target lake shore area and establish a spatial distribution dataset of lake shore slope. Determine the annual representative water level; calculate the water level deviation between the measured water level at the time of acquisition of multi-temporal remote sensing images and the corresponding annual representative water level based on the measured water level data; calculate the horizontal displacement of the wetland boundary caused by the water level deviation for each shore segment based on the spatial distribution dataset of lake shore slope and the water level deviation value; perform horizontal displacement correction on each shore segment of the measured wetland boundary based on the horizontal displacement of the wetland boundary, and correct the measured wetland boundary to below the annual representative water level to obtain the corrected wetland boundary; The wetland area is calculated based on the corrected wetland boundary to obtain a wetland area sequence; the wetland area change rate is calculated for each time period based on the wetland area sequence to analyze the dynamic evolution pattern of the wetland. The annual average water level for each time period is extracted from the measured water level data, and the annual average water level change for adjacent time periods is calculated. Based on the spatial distribution dataset of lakeside slope and the annual average water level change, the corresponding wetland area change is calculated for each shore segment. The wetland area change is compared with the wetland area change in the wetland area sequence to output the contribution of climate and hydrology to the dynamic evolution of wetlands.

2. The method for analyzing the dynamic evolution and driving forces of wetlands according to claim 1, characterized in that, The method for determining the annual representative water level includes: The arithmetic mean of the daily water levels for each year is calculated based on the measured water level data to obtain the annual average water level. The year is determined based on the imaging date of the remote sensing image, and the annual average water level of the corresponding year is used as the annual representative water level of the remote sensing image.

3. The method for analyzing the dynamic evolution and driving forces of wetlands according to claim 2, characterized in that, The method for calculating the horizontal displacement of the wetland boundary caused by the water level deviation on a segment-by-segment basis, based on the spatial distribution dataset of the lake shore slope and the water level deviation value, includes: The slope values ​​corresponding to the central data of the lakeside slope spatial distribution are extracted point by point along the lake shoreline to obtain the shoreline slope sequence; the global mean of the shoreline slope sequence is calculated, and the difference between each slope value and the global mean is accumulated point by point in the shoreline slope sequence to obtain the cumulative deviation sequence; the turning points in the cumulative deviation sequence where the absolute value of the cumulative deviation changes from increasing to decreasing or from decreasing to increasing are identified as the shoreline segmentation boundaries; The lake shore is divided according to the defined shoreline boundaries to obtain shoreline segmentation results. The median slope value within the shoreline segmentation results is taken to obtain the representative slope value of the shoreline segment. Based on the representative slope value of the shoreline segment and the corresponding water level deviation value, the horizontal displacement of the wetland boundary of each shoreline segment is calculated according to the trigonometric geometric relationship that the horizontal displacement of the wetland boundary is equal to the water level deviation value divided by the tangent of the representative slope value of the shoreline segment.

4. The method for analyzing the dynamic evolution and driving forces of wetlands according to claim 3, characterized in that, The method for identifying the turning points in the cumulative deviation sequence where the absolute value of the cumulative deviation changes from increasing to decreasing or from decreasing to increasing, and determining these as the boundaries for shoreline segmentation, includes: Calculate the difference between the absolute values ​​of adjacent cumulative deviations for each position in the cumulative deviation sequence; and identify the positions where the difference between the absolute values ​​of adjacent cumulative deviations changes from positive to negative or from negative to positive as candidate turning points. The sum of the absolute values ​​of the differences between the cumulative deviation absolute values ​​at the candidate turning point and the cumulative deviation absolute values ​​at the immediately preceding and following candidate turning points is used to obtain the change range of the candidate turning point. For the first and last candidate turning points, the change range is used as a local reference benchmark. For the remaining candidate turning points, the average change range of the immediately preceding and following candidate turning points is used as a local reference benchmark. Candidate turning points with a change magnitude not greater than that of the local reference benchmark are eliminated in one go to obtain the shoreline segmentation boundary.

5. The method for analyzing the dynamic evolution and driving forces of wetlands according to claim 4, characterized in that, The method for dividing the lake shore according to the shore segment division boundary to obtain the shore segment division result includes: Mapping the sequential positions along the shoreline corresponding to the shoreline segmentation boundary to two-dimensional lake shoreline geometric coordinates yields the spatial coordinates of the shoreline segmentation boundary; the lake shoreline is then cut into an initial set of shorelines based on these spatial coordinates. For each initial shore segment in the initial shore segment set, the spatial distance between adjacent points is calculated point by point along the shoreline direction to obtain a sequence of adjacent point distances; the positions in the adjacent point distance sequence that exceed a preset spatial continuity threshold are identified as spatial discontinuities; the initial shore segments containing spatial discontinuities are truncated at the spatial discontinuities to obtain independent continuous sub-shore segments; the independent continuous sub-shore segments are merged with the initial shore segments that do not contain spatial discontinuities to obtain the shore segment division results.

6. The method for analyzing the dynamic evolution and driving forces of wetlands according to claim 5, characterized in that, The method for truncating the initial shoreline containing the spatial discontinuity at the spatial discontinuity point to obtain independent continuous sub-shoreline segments includes: Arrange all spatial discontinuities in the initial shoreline containing spatial discontinuities in the order along the shoreline to obtain an ordered sequence of spatial discontinuities; use each spatial discontinuity in the ordered sequence of spatial discontinuities as a dividing point, and divide the initial shoreline containing spatial discontinuities in the order along the shoreline to obtain each continuous boundary point segment. For each continuous boundary point segment, calculate the tangent direction of the sub-shore segment end at the open endpoint; gradually extend the open endpoint along the tangent direction of the sub-shore segment end, and take the shoreline projection point when the extension point first enters the preset shoreline buffer range as the coordinate of the closed endpoint; replace the open endpoint of each continuous boundary point segment with the coordinate of the closed endpoint, and splice the line segment connecting the two closed endpoints of each continuous boundary point segment along the shoreline with each continuous boundary point segment to obtain an independent continuous sub-shore segment closed at both ends.

7. The method for analyzing the dynamic evolution and driving forces of wetlands according to claim 6, characterized in that, The method for correcting the horizontal displacement of the measured wetland boundary segment by segment based on the measured horizontal displacement of the wetland boundary, and correcting the measured wetland boundary to below the annual representative water level, includes: All measured wetland boundaries are discretized into an ordered sequence of boundary points; for each boundary point in the ordered sequence, the direction of the line connecting the preceding and following boundary points is taken as the local tangent direction at the boundary point; the direction perpendicular to the local tangent direction is taken as the displacement application direction of the boundary point, thus obtaining the displacement direction of each boundary point; Based on the division results of each shoreline segment, the horizontal displacement of the wetland boundary corresponding to each shoreline segment is applied along the displacement application direction of each boundary point within each shoreline segment to obtain the corrected set of boundary point coordinates; the corrected set of boundary point coordinates is then connected sequentially according to the shoreline segment order to obtain the corrected wetland boundary corrected to the annual representative water level.

8. The method for analyzing the dynamic evolution and driving forces of wetlands according to claim 7, characterized in that, The method of obtaining the displacement direction of each boundary point by taking the direction perpendicular to the local tangent direction as the displacement application direction of the boundary point includes: For each boundary point in the ordered boundary point sequence, calculate the radius of curvature at the boundary point based on the coordinates of the adjacent boundary points; use the radius of curvature as the search radius to extract the water body region centered on the boundary point; calculate the local centroid coordinates of the water body region to obtain the local pointing vector from the boundary point to the local centroid coordinates; For the local tangent direction, two mutually opposite perpendicular directions are selected as candidate displacement directions; the candidate displacement directions whose dot product with the local pointing vector is negative are marked as land-side candidate directions, and the candidate displacement directions whose dot product is positive are marked as lake-inside candidate directions; based on the sign of the water level deviation value corresponding to the shore segment where the boundary point is located, the displacement application direction of each boundary point in the shore segment with a positive water level deviation value is determined as the lake-inside candidate direction, and the displacement application direction of each boundary point in the shore segment with a negative water level deviation value is determined as the land-side candidate direction, thus obtaining the displacement direction of each boundary point.

9. The method for analyzing the dynamic evolution and driving forces of wetlands according to claim 1, characterized in that, The method for calculating the wetland area change rate at each time period based on the wetland area sequence and analyzing the dynamic evolution of wetlands includes: The wetland area sequence is subtracted pairwise from each adjacent period to obtain the wetland area net change sequence for each period. The wetland area net change sequence for each period is divided by the corresponding initial wetland area to obtain the wetland area change rate sequence for each period. Based on the temporal variation characteristics of the wetland area change rate sequence, the wetland dynamic evolution process is divided into stages of continuous shrinkage, continuous expansion and fluctuating alternation, thus obtaining the wetland dynamic evolution law.

10. A wetland dynamic evolution and driving force analysis system, used to perform the method according to any one of claims 1 to 9, characterized in that, The system includes: The data acquisition and preprocessing module is used to acquire multi-temporal remote sensing images of the target lake shore area and acquire measured water level data corresponding to the acquisition time of the multi-temporal remote sensing images; preprocess the multi-temporal remote sensing images and extract wetland boundaries to obtain measured wetland boundaries; acquire the surface slope of each shore section of the target lake shore area and establish a spatial distribution dataset of lake shore slope. The wetland boundary correction module is used to determine the annual representative water level; calculate the water level deviation between the measured water level at the time of acquisition of multi-temporal remote sensing images and the corresponding annual representative water level based on the measured water level data; calculate the horizontal displacement of the wetland boundary caused by the water level deviation for each shore segment based on the spatial distribution dataset of lake shore slope and the water level deviation value; and perform horizontal displacement correction on each shore segment of the measured wetland boundary based on the horizontal displacement of the wetland boundary to correct the measured wetland boundary to below the annual representative water level, thus obtaining the corrected wetland boundary. The wetland evolution module is used to calculate the wetland area based on the corrected wetland boundary to obtain a wetland area sequence; based on the wetland area sequence, it calculates the wetland area change rate for each time period and analyzes the dynamic evolution pattern of the wetland. The driving force analysis module is used to extract the annual average water level for each time period based on the measured water level data and calculate the annual average water level change between adjacent time periods. Based on the spatial distribution dataset of lake shore slope and the annual average water level change, the module calculates the corresponding wetland area change for each shore segment, compares it with the wetland area change in the wetland area sequence, and outputs the contribution of climate and hydrology to the dynamic evolution of wetlands.

Images