CN-121999399-A - Method and system for measuring blue carbon reserves of coastal wetland based on habitat perception and missing codes

CN121999399ACN 121999399 ACN121999399 ACN 121999399ACN-121999399-A

Abstract

The invention discloses a method for measuring blue carbon reserves of a coastal wetland based on habitat perception and missing codes, which comprises the steps of obtaining habitat type labels and microclimate parameters of a target area, generating sensor dispatching instructions and dispatching state codes according to the habitat type labels and the microclimate parameters, executing active compensation according to pod posture change, generating data confidence marks, constructing feature vectors of grid units, constructing structural missing indication codes, inputting the structural missing indication codes and the feature vectors into an inversion model in a combined mode, outputting blue carbon reserves predicted values and corresponding uncertainty indexes of the target area, generating a recovery strategy and dynamically correcting subsequent flight planning parameters when the uncertainty indexes exceed a preset threshold, and repeatedly executing feature construction, inversion and uncertainty assessment processes after recovery data are updated until global uncertainty meets preset requirements. The method and the system improve the stability and the reliability of the blue carbon reserve space inversion result under the complex coastal wetland environment.

Inventors

XIAO XI
CHEN LEI
ZHU YINGJIE
CHEN MEIHAO
HE JUNYU
FAN DONGYANG
HUA XIAOJING
FENG YIXIONG

Assignees

浙江大学

Dates

Publication Date: 20260508
Application Date: 20260409

Claims (10)

1. The method for measuring the blue carbon reserves of the coastal wetland based on the habitat perception and missing codes is characterized by comprising the following steps of: (1) Acquiring a habitat type label of a target area and synchronously acquiring microclimate parameters; (2) Generating a sensor scheduling instruction according to the habitat category label and the microclimate parameter by combining a preset threshold value table, and generating a scheduling state code representing the working state of the sensor; (3) Performing space-time registration and gridding treatment on the acquired multi-source heterogeneous remote sensing data to construct a feature vector of a grid unit; (4) According to the sensor scheduling instruction and the data confidence level mark, a structural deletion indication code is constructed, and is input into an inversion model in a combined mode with the feature vector, and a blue carbon reserve predicted value and a corresponding uncertainty index of a target area are output; (5) And after the recovery data is updated, repeatedly executing the characteristic construction, inversion and uncertainty evaluation processes until the global uncertainty meets the preset requirement.
2. The method for measuring the blue carbon reserves of the coastal wetland based on the perception and missing codes of the habitat according to claim 1, wherein the habitat categories comprise mangrove habitat, salt marsh habitat and photo beach habitat, and the microclimate parameters comprise wind speed, humidity and air pressure.
3. The method for measuring the blue carbon reserves of the coastal wetland based on the habitat perception and missing codes according to claim 1 is characterized in that in the step (1), the method for obtaining the habitat categories of the target area comprises the steps of (a) classifying the habitat categories in real time by utilizing preview images acquired by a vision sensor, (b) matching a priori habitat distribution map of the target area with positioning information, and (c) fusing the results of the steps (a) and (b) and selecting a high confidence result.
4. The method for measuring the blue carbon reserves of the coastal wetland based on the habitat awareness and missing codes according to claim 1, wherein the generating the sensor scheduling instruction according to the habitat type and the microclimate parameters and combining with a preset threshold table comprises the following steps: When the habitat class is mangrove habitat, starting a laser radar, starting a hyperspectral imager and/or a multispectral camera to observe all elements; when the habitat type is a salt-pond habitat or a photo-beach habitat, turning off the laser radar or reducing the sampling frequency to a preset lower limit, and turning on a hyperspectral imager and/or a multispectral camera; Triggering a nacelle protection mode when the local instantaneous wind speed exceeds a preset threshold value, and closing the laser radar or reducing the working load of the laser radar; when the ambient humidity exceeds a preset threshold, the corresponding optical image data is marked as low confidence data.
5. The method for measuring the blue carbon reserves of the coastal wetland based on the ecological environment perception and the missing codes according to claim 1, wherein the steps of executing the active compensation according to the change of the attitude of the nacelle and generating the data confidence level mark comprise the steps of monitoring the attitude angular speed omega of the nacelle in real time, judging that the physical active compensation cannot completely eliminate disturbance when the attitude angular speed omega exceeds a preset stability limit threshold value, and marking the observation data acquired in the time window as low confidence level data.
6. The method for measuring blue carbon reserves of the coastal wetland based on the perception and missing codes according to claim 1, wherein in the step (3), the feature vector comprises: (1) The spectrum characteristics are that the normalized vegetation index NDVI, the enhanced vegetation index EVI and the soil conditioning vegetation index SAVI are calculated based on multispectral or hyperspectral data; (2) The vertical structure characteristic is that 95% quantiles of the height distribution are extracted from the laser radar point cloud to serve as canopy height characteristics H 95 , and when the laser radar is in a closed state, the vertical structure characteristic value is marked as missing; (3) The terrain features comprise a grid unit digital elevation E, a gradient S and a terrain-hydrodynamic coupling index T, wherein the calculation model of the T is as follows: E avg and S avg are respectively the average elevation and the average gradient of the target area; (4) The environmental characteristics include real-time wind speed, humidity and air pressure.
7. The method for measuring blue carbon reserves of the coastal wetland based on the environmental awareness and the missing codes according to claim 1, wherein in the step (4), the structural missing indication codes comprise missing masks constructed for the data arrival states of the sensor channels and scheduling state codes generated by the mapping of the scheduling instructions.
8. The method for measuring blue carbon reserves of the coastal wetland based on the perception of the habitat and the missing codes according to claim 1, wherein in the step (4), the structural missing indication codes and the feature vectors are input into an inversion model in a joint way, and the method comprises the following steps: Splicing the structural deletion indication code with the feature vector to construct an extended feature vector containing habitat prior information; And inputting the extended feature vector into an inversion model.
9. The method for measuring the blue carbon reserves of the coastal wetland based on the habitat perception and missing codes, which is characterized in that in the step (4), the inversion model is an integrated inversion model and comprises a base learner layer and a meta learner layer, wherein the base learner layer comprises at least two of a gradient lifting tree model, a random forest model and a support vector regression model, and the meta learner layer fuses the output of the base learner by adopting a robust regression algorithm.
10. A littoral wetland blue carbon reserve measurement system based on habitat perception and missing coding, comprising: The flight platform is controlled by the control unit; the self-balancing nacelle is mounted on the flying platform and is used for bearing the multi-mode sensing array and executing attitude active compensation; The multi-mode sensing array comprises a laser radar, a hyperspectral imager, a multispectral camera and a microclimate sensor, and is controlled by the control unit; a control unit comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the method for measuring blue carbon reserves of a coastal wetland based on habitat awareness and missing codes as claimed in any one of claims 1 to 9 when the computer program is executed.

Description

Method and system for measuring blue carbon reserves of coastal wetland based on habitat perception and missing codes Technical Field The invention relates to the technical field of ecological environment remote sensing monitoring, in particular to a method and a system for measuring blue carbon reserves of a coastal wetland based on habitat perception and missing codes. Background The blue carbon ecosystem is an important component of the global carbon cycle, wherein mangroves, coastal salt-water and intertidal light beaches have higher carbon capture and storage capacity. The coastal wetland habitat is affected by the action of sea Liu Jiaohu and the tidal process, has obvious space breaking characteristics, and has obvious differences among different habitats in vegetation types, vertical structures and submerged frequencies. Therefore, establishing a blue carbon reserve monitoring means which gives consideration to spatial resolution, operation efficiency and habitat adaptability is a real requirement in the field of carbon sink metering. The existing blue carbon reserve monitoring method mainly comprises artificial sample area investigation and satellite remote sensing monitoring. The manual investigation has the advantages of high precision, high operation cost, limited space coverage and difficult access to muddy beach areas, and satellite remote sensing has the advantages of large-scale coverage, but is limited by factors such as space resolution, revisit period, frequent cloud and fog in coastal areas, and the like, so that the fine structure information of the crushed wetland is difficult to accurately obtain, and inversion deviation is easy to generate in weak texture areas such as light beach. In recent years, a low-altitude unmanned aerial vehicle remote sensing technology with laser radar, hyperspectral or multispectral imaging equipment is gradually applied to the field of blue carbon monitoring, and a new technical path is provided for acquiring high-resolution and multisource remote sensing data. However, aiming at the complex habitat condition of the coastal wetland, the existing unmanned aerial vehicle multi-mode remote sensing technology still has the following defects: First, there is a discrepancy between full element observations and load energy efficiency. The demand difference of different coastal habitats on remote sensing information is obvious, the mangrove area has strong dependence on the vertical structure information, and the light beach or salt marsh area mainly depends on the spectrum information. In the prior art, a collection mode that the sensor is opened in the whole course is adopted, a large amount of redundant data is generated in part of habitats, and the energy consumption and the data storage processing burden are increased. Second, the reliability of data under complex weather conditions is insufficient. Coastal areas often accompany gusts, high humidity and salt fog environments, which are prone to causing severe disturbance of nacelle attitude and degradation of imaging quality. The prior art lacks a real-time sensing and response mechanism for microclimate conditions, and is difficult to effectively mark or compensate disturbed data in a data acquisition stage. Third, structural deletions induced by sensor scheduling are not effectively exploited. In order to reduce energy consumption, a part of technologies implement start-stop control on the sensor, so that the generated non-random missing data is generally regarded as abnormal data and is subjected to interpolation or rejection processing, the prior information of the habitat contained in the scheduling behavior of 'sensor off' is ignored, and systematic inversion deviation is easily introduced. Fourth, there is a lack of closed loop measurement mechanisms based on uncertainty assessment. Most of the existing operation flows are in an open loop mode, when the uncertainty of inversion results of local areas is high, the flight and acquisition strategies are difficult to adjust in time to carry out complement measurement, and the reliability of the final result in a key area is affected. Therefore, it is necessary to provide a blue carbon reserve measurement method and system which can adapt to the difference of the coastal wetland habitat, support the on-demand scheduling of sensors, utilize the structural missing information caused by the scheduling and have uncertainty-driven recovery capability. Disclosure of Invention The invention provides a method and a system for measuring the blue carbon reserves of a coastal wetland based on habitat perception and missing codes, which improve the stability and reliability of the space inversion result of the blue carbon reserves in a complex coastal wetland environment. The technical scheme of the invention is as follows: a method for measuring blue carbon reserves of a coastal wetland based on habitat perception and missing codes comprises the following steps: (1)