Search

CN-121999224-A - Method for analyzing space-time evolution topological interference response by dynamic evolution of estuary wetland

CN121999224ACN 121999224 ACN121999224 ACN 121999224ACN-121999224-A

Abstract

The invention relates to the technical field of hyperspectral remote sensing ecological environment monitoring, in particular to a estuary wetland dynamic evolution analysis method integrating space-time evolution topology analysis and interference response mechanism, which comprises the steps of constructing a benchmark segmentation template, the multi-time phase change detection is combined to define the re-segmentation area, and the local re-segmentation is only carried out in the real change area, so that the time sequence consistency of the object boundary of the non-change area is ensured, and the pseudo-change interference is effectively eliminated. On the basis, landscape element change information is extracted, object-level space-time evolution track codes are constructed, landscape topology indexes and ecological restoration disturbance indexes are synthesized, natural and artificial interference events are identified by combining hydrologic observation data, and finally dynamic evolution analysis results which quantitatively represent the interference-response relationship are generated. The method realizes accurate identification and mechanism interpretation of the evolution process of the estuary wetland, and provides reliable basis for physiological management.

Inventors

  • YANG YUXIN
  • YANG HUAN
  • LIU JING

Assignees

  • 西安中科西光航天科技集团有限公司

Dates

Publication Date
20260508
Application Date
20260129

Claims (10)

  1. 1. A method for analyzing space-time evolution topological interference response of dynamic evolution of estuary wetland is characterized by comprising the following steps of; step 1, acquiring remote sensing images of a target estuary wetland in at least two different time phases; Step 2, performing object-oriented segmentation on the remote sensing image of the reference time phase, and outputting a reference segmentation template; Step 3, limiting a change area range based on a multi-time phase change detection result, inheriting the object boundary of the reference segmentation template in a non-change area, executing local re-segmentation only in the change area, expanding and mapping the reference segmentation template to other time-phase remote sensing images, and generating a segmentation result with a time sequence consistent object boundary; Step 4, comparing the reference segmentation template with the segmentation result of the corresponding time phase in the segmentation results with the time sequence consistent object boundary, and extracting the change information of the estuary wetland landscape elements; step 5, constructing object-level space-time evolution track codes for each landscape object, and calculating landscape topology indexes based on the track codes; And 6, introducing an ecological restoration disturbance index, combining hydrologic observation data and human activity data, identifying the type of disturbance and generating a dynamic evolution analysis result of the relation between the disturbance and the response.
  2. 2. The method for analyzing space-time evolution topological interference response of dynamic evolution of estuary wetland according to claim 1, wherein the performing object-oriented segmentation on the remote sensing image of the reference time phase in step 2 comprises: step 2.1, extracting spectral features, texture features and spatial features of pixels in the reference time phase remote sensing image; Step 2.2, based on the spectral features, the texture features and the spatial features extracted in the step 2.1, adopting a multi-scale segmentation algorithm, taking the minimization of spectral heterogeneity and the minimization of shape complexity as constraint targets, iteratively combining adjacent pixels with similar features, and outputting an initial segmentation object; And 2.3, calculating the average area and the shape index of the initial segmentation object output in the step 2.2, merging the initial segmentation objects with the areas lower than a preset area threshold or the shape indexes lower than a preset shape threshold, and outputting a reference segmentation template.
  3. 3. The method for analyzing space-time evolution topological interference response by dynamic evolution of estuary wetland according to claim 2, wherein mapping the reference segmentation template to the registered remaining time-phase remote sensing images in step 3 comprises: step 3.1, using the remote sensing image of the reference time phase as a geometric reference, and adopting a polynomial correction model to perform geometric fine correction on the remote sensing images of the other time phases obtained in the step 1; Step 3.2, performing spatial superposition on the vector boundary image layer of the reference segmentation template output in the step 2 and the rest of the time phase remote sensing images subjected to geometric fine correction in the step 3.1; and 3.3, in the boundary range of the spatial superposition in the step 3.2, performing re-segmentation based on the spectral features of the rest time-phase remote sensing images after the geometric fine correction, and outputting a segmentation result with a time sequence consistent object boundary.
  4. 4. The method for analyzing the spatiotemporal evolution topological interference response of the dynamic evolution of the estuary wetland according to claim 1, wherein the step 4 of extracting the variation information of the estuary wetland landscape elements comprises the following steps: Step 4.1, constructing a time sequence data set based on the multi-temporal remote sensing image obtained in the step 1 and the segmentation result with the time sequence consistent object boundary generated in the step 3, and extracting the time sequence spectrum characteristic of each segmentation object in the segmentation result, wherein the time sequence spectrum characteristic comprises the standard deviation, the ratio of the maximum value to the minimum value of a multi-temporal normalized vegetation index (NDVI) and the normalized water body index and the area of the segmentation object; Step 4.2, constructing a vegetation weather identification rule based on the time sequence spectrum characteristics extracted in the step 4.1, identifying a change candidate object by utilizing a weather difference threshold value of different vegetation in a growth period, subdividing the vegetation object in the change candidate object into mangrove forest, spartina alterniflora and other vegetation types, and judging the specific conversion type of the landscape element by combining the reference time phase attribute; and 4.3, performing morphological open operation processing on the change candidate object marked in the step 4.2, removing isolated plaques with the area smaller than the noise threshold value, and outputting a change area.
  5. 5. The method for analyzing space-time evolution topological interference response of dynamic evolution of estuary wetland according to claim 4, wherein calculating topological index based on the change information extracted in step 4 comprises: Step 5.1, abstracting the change area output in the step 4.3 into a landscape patch, and constructing a patch vector layer; Step 5.2, calculating a landscape connectivity index by adopting a graph theory algorithm based on the plaque vector layer constructed in the step 5.1; step 5.3, constructing a space-time evolution track code of the landscape object, distributing a unique identification number to each geographic object in the benchmark segmentation template, tracking the state change of the unique identification number in all time-phase images, and constructing a state chain of a state like T1 state- > T2 state- > Tn state; and 5.4, carrying out pattern recognition based on the state chain, and automatically classifying the evolution track into an artificial repair type, a natural siltation type or a city erosion type according to the logical characteristics of state evolution, thereby being used as a classification basis for calculating the topological index.
  6. 6. The method for dynamically evolving and analyzing the spatiotemporal evolution topology disturbance response according to claim 5, wherein the step 6 of identifying the disturbance event includes: Step 6.1, acquiring hydrologic observation data which are similar to the remote sensing image acquired in the step 1, wherein the hydrologic observation data comprise runoff quantity, sand conveying quantity and flood pulse frequency; step 6.2, when the change trend of the topological index is obviously related to the hydrologic observation data obtained in the step 6.1 in time sequence, identifying the hydrologic event as a natural interference event; And 6.3, calculating an ecological restoration disturbance index, wherein the calculation comprises the steps of comprehensively evaluating the landscape fragmentation index and the landscape connectivity index, identifying the event as positive artificial restoration disturbance when the landscape fragmentation index of the change area is monitored to rise but the landscape connectivity index simultaneously shows a rising trend, and identifying the event as negative development encroachment disturbance when the landscape fragmentation index rises and the landscape connectivity index falls.
  7. 7. The method for dynamically evolving and analyzing the spatiotemporal evolution topology interference response of the estuary wetland according to claim 6, wherein the step 6 of generating the dynamic evolution analysis result representing the relation between the interference and the response comprises the steps of: Step 7.1, establishing a quantitative response model of the type of the interference event and the change of the topological index, wherein the model takes the characteristic value in the hydrographic observation data obtained in the step 6.1 as an input variable and the change rate of the output landscape connectivity index as an output variable; Step 7.2, outputting a strong response conclusion of the estuary wetland to the hydrologic pulse when the landscape connectivity index change rate corresponding to the identified natural interference event exceeds a preset sensitivity threshold; and 7.3, generating an evaluation conclusion based on the identification result in the step 6.3, namely outputting an evaluation result of ecological function improvement for positive artificial restoration interference, and outputting a wetland ecological risk early warning conclusion for negative development encroaching interference when the fragmentation index increase slope exceeds a preset sustainable threshold value.
  8. 8. The method for dynamically evolving and analyzing space-time evolving topological disturbance response of estuary wetland according to claim 4, further comprising, after step 4.3, performing confidence optimization on the variation region: Step 8.1, clustering pixel characteristics in the variation area output in the step 4.3 by adopting a Gaussian mixture model, and dividing a clustering result into three real variation types of vegetation succession, sediment accumulation and erosion and tidal channel transition and pseudo variation types caused by noise; Step 8.2, calculating posterior probability of each pixel belonging to the corresponding class in the three true change classes in the clustering result of the step 8.1; And 8.3, removing pixels with posterior probability values lower than a first confidence threshold based on the posterior probability values calculated in the step 8.2, performing morphological closing operation on pixels with posterior probability values lower than a second confidence threshold but higher than the first confidence threshold, and outputting optimized change information.
  9. 9. The method for analyzing the spatio-temporal evolution topological interference response of the dynamic evolution of the estuary wetland according to claim 8, wherein the determining the clustering number of the gaussian mixture model in the step 8.1 comprises the following steps: Step 9.1, setting a candidate cluster quantity set to be evaluated; Step 9.2, training a Gaussian mixture model on the pixel characteristic data set of the change area in the step 8.1 for each candidate quantity in the candidate cluster quantity set in the step 9.1, and calculating Bayesian information criterion scores and contour coefficients of the Gaussian mixture model; step 9.3, carrying out normalization processing on the Bayesian information criterion score corresponding to each candidate number obtained in the step 9.2, and carrying out weighted summation on the normalized Bayesian information criterion score and the contour coefficient to obtain a comprehensive evaluation score of each candidate number; And 9.4, selecting the candidate number with the optimal comprehensive evaluation score obtained in the step 9.3 as the clustering number of the Gaussian mixture model.
  10. 10. The method for analyzing the spatio-temporal evolution topological interference response of the dynamic evolution of the estuary wetland according to claim 1, wherein the preprocessing after the remote sensing image is acquired in the step 1 comprises the following steps: Step 10.1, performing radiation calibration and atmospheric correction on the remote sensing image obtained in the step 1, and calculating a normalized vegetation index and a normalized water body index, wherein the remote sensing image comprises a multispectral image and a panchromatic image; Step 10.2, adopting Gram-Schmidt spectrum sharpening, and injecting the space details of the full-color image processed in the step 10.1 into the multispectral image processed in the step 10.1 to generate a fusion image; And 10.3, removing the image with cloud cover rate higher than 5% or solar altitude angle lower than 30 degrees from the fused image generated in the step 10.2 from the remote sensing image obtained in the step 1.

Description

Method for analyzing space-time evolution topological interference response by dynamic evolution of estuary wetland Technical Field The invention relates to the technical field, in particular to a method for analyzing space-time evolution topology interference response of dynamic evolution of estuary wetland. Background The dynamic evolution analysis in the estuary wetland ecological monitoring technology is systematic research on the change of a wetland ecosystem along with time, and is characterized in that the analysis of a space-time evolution process, namely the distribution and evolution rule of the wetland morphology and function in time and space dimensions, the topology concept is used for describing the spatial structure and the connection relation of wetland landscape elements, the interference refers to the external pressure of the ecosystem caused by natural or artificial factors, the response represents the ecological change and adaptation behavior of the wetland to the interference, and the monitoring method can logically deduce the evolution mechanism of the estuary wetland under the action of multiple environments by integrating various technical characteristics, so that scientific basis is provided for ecological assessment and sustainable management. The existing river mouth wetland ecological monitoring technology based on object-oriented image analysis has the technical pain point that when dynamic monitoring is carried out by utilizing multi-phase remote sensing images, the object boundaries generated by the same geographic entity in different time phases cannot be matched accurately due to the fact that the parameters or image features adopted by each independent segmentation are different. Such spatio-temporal inconsistencies of segmentation results introduce a large amount of spurious variation information that does not originate from the real surface variations. For example, a continuous mangrove patch, due to the different spectral responses of rainy and drought images, may be segmented into objects of different shapes and numbers in two time phases, and thus misjudged as broken or expanded, masking the real evolution track of the wetland, such as the gradual change process caused by natural succession or artificial interference. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a method for analyzing space-time evolution topological interference response of dynamic evolution of a estuary wetland, which solves the technical problems that a large amount of pseudo-change information interference is caused and the real evolution track of the wetland is difficult to restore due to inconsistent object-oriented segmentation boundaries of remote sensing images in different time phases. In order to solve the technical problems, the invention comprises the following specific contents: The invention provides a method for analyzing space-time evolution topological interference response of dynamic evolution of estuary wetland, which comprises the following steps of; step 1, acquiring remote sensing images of a target estuary wetland in at least two different time phases; Step 2, performing object-oriented segmentation on the remote sensing image of the reference time phase, and outputting a reference segmentation template; Step 3, limiting a change area range based on a multi-time phase change detection result, inheriting the object boundary of the reference segmentation template in a non-change area, executing local re-segmentation only in the change area, expanding and mapping the reference segmentation template to other time-phase remote sensing images, and generating a segmentation result with a time sequence consistent object boundary; Step 4, comparing the reference segmentation template with the segmentation result of the corresponding time phase in the segmentation results with the time sequence consistent object boundary, and extracting the change information of the estuary wetland landscape elements; step 5, constructing object-level space-time evolution track codes for each landscape object, and calculating landscape topology indexes based on the track codes; And 6, introducing an ecological restoration disturbance index, combining hydrologic observation data and human activity data, identifying the type of disturbance and generating a dynamic evolution analysis result of the relation between the disturbance and the response. Further, in the method for analyzing the space-time evolution topological interference response of the dynamic evolution of the estuary wetland, in the step 2, the remote sensing image of the reference time phase is subjected to object-oriented segmentation, which comprises the following steps: step 2.1, extracting spectral features, texture features and spatial features of pixels in the reference time phase remote sensing image; Step 2.2, based on the spectral features, the texture features and the spatial features extract