Search

CN-121999380-A - Forest structure remote sensing parameter-based method and medium for evaluating post-fire recovery process

CN121999380ACN 121999380 ACN121999380 ACN 121999380ACN-121999380-A

Abstract

The invention relates to the technical field of remote sensing image processing, in particular to an evaluation method of a post-fire recovery process based on forest structure remote sensing parameters, which comprises the following steps of S1, setting a research area, selecting a fire area and a non-fire area in the research area, preprocessing Landsat time sequence data to synthesize Landsat time sequence stacks, calculating an NBR time sequence, acquiring NBR values before and after occurrence of forest disturbance pixels by using LANDTRENDR algorithm, dividing the severity of the fire plaque into three grades of low, medium and high by combining RdNBR indexes, S2, performing remote sensing inversion on the forest structure parameters, and S3, reconstructing the post-fire recovery process. According to the invention, dNBR is introduced as a prediction variable, and the defect of the existing remote sensing inversion model in the aspect of fire disturbance intensity characterization is effectively solved by enhancing the response capability of the forest structure parameter remote sensing inversion model to the fire disturbance intensity.

Inventors

  • FAN HUI
  • PAN YINGYING
  • LI SHUAI
  • XU XIAO
  • LI YATING

Assignees

  • 云南大学

Dates

Publication Date
20260508
Application Date
20260114

Claims (9)

  1. 1. The method for evaluating the post-fire recovery process based on the forest structure remote sensing parameters is characterized by comprising the following steps of: Step S1, identifying and severity of a burning plaque, namely setting a research area, selecting a burning area and a non-burning area in the research area, synthesizing Landsat time sequence stacks and calculating NBR time sequences after pretreatment by taking Landsat time sequence data as a source, acquiring NBR values before and after occurrence of forest disturbance pixels by using LANDTRENDR algorithm, and dividing the severity of the burning plaque into three grades of low grade, medium grade and high grade by combining RdNBR index; S2, carrying out remote sensing inversion on forest structure parameters, namely taking a microwave band, a spectrum band and derivative indexes of the Sentinel-1 SAR and the Sentinel-2 MSI, wherein the derivative indexes comprise a microwave index and a spectrum index, a topographic variable extracted from SRTM DEM data, a fire plaque attribute index as a prediction variable, and taking forest structure parameters of a footprint point scale after quality screening and abnormal footprint point removal as dependent variables, and constructing each fire disturbance damage degree by adopting a random forest regression model; s3, reconstructing a post-fire recovery process, namely extracting a fire and non-fire pixel point pair by adopting a trend score matching method, and respectively calculating the ratio of forest structure parameters of plaque scales after extracting the forest structure parameters of the fire and non-fire pixel point pair Finally reconstructing a curve after fire by a space-time substitution method.
  2. 2. The method for evaluating a post-fire recovery process based on forest structure remote sensing parameters according to claim 1, wherein the calculation formula of the NBR value is shown as formula one: (equation I) The calculation formula of RdNBR is shown as formula two: (equation II).
  3. 3. The method for evaluating post-fire recovery process based on forest structure remote sensing parameters according to claim 1, wherein the ratio of forest structure parameters is The calculation formula of (a) is shown as a formula III: (equation three).
  4. 4. The method for evaluating a post-fire recovery process based on forest structure remote sensing parameters of claim 1, wherein the forest structure parameters comprise at least one of m75, h95, TCC, PAI, and FHD.
  5. 5. The method for evaluating a post-fire recovery process based on forest structure remote sensing parameters according to claim 1, wherein the microwave band comprises VV, VH.
  6. 6. The method for evaluating a post-fire recovery process based on remote sensing parameters of a forest structure of claim 1, wherein the spectral index comprises NBR, NDVI, EVI, MSAVI, TCB, TCW, TCG, TCA.
  7. 7. The method for evaluating a post-fire recovery process based on forest structure remote sensing parameters according to claim 1, wherein the microwave index is RVI.
  8. 8. The method for evaluating a post-fire recovery process based on forest structure remote sensing parameters according to claim 1, wherein the method for selecting the fire area and the non-fire area comprises the following specific steps: (1) Each fire perturbation plaque detected by Landsat is regarded as a fire zone, a 120 m buffer zone is constructed by the fire perturbation plaque, a 30m buffer zone adjacent to the fire perturbation plaque is removed, and a reserved 90 m buffer zone not adjacent to the fire perturbation plaque is regarded as a non-fire zone; (2) The method comprises the steps of selecting gradients and slope directions with larger topographic effects as observation variables, taking a standard deviation of 0.25 times of sample point trend scores as a caliper size when the trend scores are matched, judging the matching effect according to the standard mean deviation of the fire and non-fire pixel point pairs before and after the trend scores, the standard mean deviation reducing amplitude and the saliency of independent sample t test, wherein the standard mean deviation of the fire and non-fire pixel point pairs before the matching is large, the saliency difference exists between the independent sample t test results, and the standard mean deviation of the reserved fire and non-fire pixel point pairs after the matching is reduced, and the saliency does not exist any more.
  9. 9. A computer-readable storage medium storing a computer program, which when executed by a processor implements the method for evaluating post-fire recovery procedures based on forest structure remote sensing parameters according to any one of claims 1 to 8.

Description

Forest structure remote sensing parameter-based method and medium for evaluating post-fire recovery process Technical Field The invention relates to the technical field of remote sensing image processing, in particular to an evaluation method and medium of a post-fire recovery process based on forest structure remote sensing parameters. Background The forest fire can change the landscape dynamics of a forest ecological system and influence ecological processes such as forest succession, nutrient circulation, energy circulation and the like. Accurately identifying the recovery process of the forest ecological system after fire is helpful for comprehensively recognizing the influence of forest fire on the forest ecological system. In the related research of early post-fire forest recovery detection, researchers mostly adopt a traditional field investigation method to obtain parameters such as canopy height, tree species, age, canopy density, damage degree, soil seed bank and the like of the forest in the sample plot at a certain moment after a forest fire so as to estimate vegetation recovery speed, thereby analyzing the forest recovery process after fire disturbance. With the development of remote sensing technology, related researches on forest recovery detection after fire based on optical and microwave remote sensing images are largely emerging, but the inversion precision is poor, and the improvement is needed. Disclosure of Invention Features and advantages of the invention will be set forth in part in the description which follows, or may be obvious from the description, or may be learned by practice of the invention. In order to overcome the problems in the prior art, the invention provides an evaluation method of a post-fire recovery process based on forest structure remote sensing parameters, which specifically comprises the following steps: Step S1, identifying and severity of a burning plaque, namely setting a research area, selecting a burning area and a non-burning area in the research area, synthesizing Landsat time sequence stacks and calculating NBR time sequences after pretreatment by taking Landsat time sequence data as a source, acquiring NBR values before and after occurrence of forest disturbance pixels by using LANDTRENDR algorithm, and dividing the severity of the burning plaque into three grades of low grade, medium grade and high grade by combining RdNBR index; S2, carrying out remote sensing inversion on forest structure parameters, namely taking a microwave band, a spectrum band and derivative indexes of the Sentinel-1 SAR and the Sentinel-2 MSI, wherein the derivative indexes comprise a microwave index and a spectrum index, a topographic variable extracted from SRTM DEM data, a fire plaque attribute index as a prediction variable, and taking forest structure parameters of a footprint point scale after quality screening and abnormal footprint point removal as dependent variables, and constructing each fire disturbance damage degree by adopting a random forest regression model; s3, reconstructing a post-fire recovery process, namely extracting a fire and non-fire pixel point pair by adopting a trend score matching method, and respectively calculating the ratio of forest structure parameters of plaque scales after extracting the forest structure parameters of the fire and non-fire pixel point pair Finally reconstructing a curve after fire by a space-time substitution method. Preferably, the formula of calculation of the NBR value is shown as formula one: (equation I) The calculation formula of RdNBR is shown as formula two: (equation II). Preferably, the forest structure parameter ratioThe calculation formula of (a) is shown as a formula III: (equation three). Preferably, the forest structure parameters include at least one of m75, h95, TCC, PAI, and FHD. Preferably, the microwave band includes VV, VH. Preferably, the spectral index comprises NBR, NDVI, EVI, MSAVI, TCB, TCW, TCG, TCA. Preferably, the microwave index is RVI. Preferably, the method for selecting the fire zone and the non-fire zone comprises the following specific steps: (1) Each fire perturbation plaque detected by Landsat is regarded as a fire zone, a 120 m buffer zone is constructed by the fire perturbation plaque, a 30m buffer zone adjacent to the fire perturbation plaque is removed, and a reserved 90 m buffer zone not adjacent to the fire perturbation plaque is regarded as a non-fire zone; (2) The method comprises the steps of selecting gradients and slope directions with larger topographic effects as observation variables, taking a standard deviation of 0.25 times of sample point trend scores as a caliper size when the trend scores are matched, judging the matching effect according to the standard mean deviation of the fire and non-fire pixel point pairs before and after the trend scores, the standard mean deviation reducing amplitude and the saliency of independent sample t test, wherein the standard mean deviati