Vegetation-environment coupling prediction method and system based on multi-source remote sensing data fusion

Abstract

The present invention discloses a vegetation environment coupling prediction method and system based on multi-source remote sensing data fusion, which relates to the field of vegetation environment monitoring. Specifically disclosed are vegetation canopy structure data, vegetation environment temperature detection data, and vegetation component characteristic data. Vegetation canopy structure parameters are calculated through a data fusion algorithm; vegetation canopy structure parameters are introduced to construct a vegetation transpiration intensity outlier detection model; local drought or waterlogged areas are identified based on the vegetation transpiration intensity outlier detection model; trunk insect pest infestation vibration signals and trunk wormhole scanning data are detected based on local drought or waterlogged areas; and pest spread paths are simulated based on trunk insect pest infestation vibration signals and trunk wormhole scanning data. The present invention realizes the analysis of the specific ecological structure of vegetation, as well as the in-depth research and application of the immediate response mechanism for disasters.

Description

Technical Field

[0001] The present invention relates to the field of vegetation environment monitoring, and in particular to a vegetation environment coupling prediction method and system based on multi-source remote sensing data fusion. Background Art

[0002] In the field of vegetation management, intelligence is becoming a major trend in crop cultivation and landscaping. Multimodal perception systems are a key technology for improving management efficiency and accuracy. For example, by integrating diverse sensor devices and remote sensing technologies to monitor the vegetation environment and building environment-plant response models based on big data analysis, the goal is to provide scientific and accurate support for planting decisions. However, these technologies are typically used for crop monitoring. For landscaping vegetation management, there is a lack of in-depth research and application of the specific ecological structure of vegetation and immediate response mechanisms to disasters. Summary of the Invention

[0003] The technical problem to be solved by the present invention is that the vegetation management in garden landscaping still lacks analysis of the specific ecological structure of vegetation, as well as in-depth research and application of the immediate response mechanism to disasters. The purpose is to provide a vegetation environment coupling prediction method and system based on multi-source remote sensing data fusion, which solves the above technical problems by analyzing the canopy structure and dynamically modeling the transpiration intensity field, and real-time monitoring and prediction of the spread path of pests.

[0004] The present invention is achieved through the following technical solutions:

[0005] A vegetation environment coupling prediction method based on multi-source remote sensing data fusion includes: using multiple sensors to collect vegetation canopy structure data, vegetation environment temperature detection data and vegetation component characteristic data; calculating the vegetation canopy structure parameters using the above-mentioned vegetation canopy structure data, the above-mentioned vegetation environment temperature detection data and the above-mentioned vegetation component characteristic data through a data fusion algorithm; introducing the above-mentioned vegetation canopy structure parameters to construct a vegetation transpiration intensity outlier detection model; identifying local drought or waterlogged areas based on the above-mentioned vegetation transpiration intensity outlier detection model; detecting tree trunk pest boring vibration signals and tree trunk worm tunnel scanning data based on the local drought or waterlogged areas; and simulating the pest spread path based on the above-mentioned tree trunk pest boring vibration signals and the above-mentioned tree trunk worm tunnel scanning data.

[0006] The above-mentioned multiple sensors are used to collect vegetation canopy structure data, vegetation ambient temperature detection data and vegetation component characteristic data, specifically including:

[0007] LiDAR is used to detect vegetation canopy structure data; thermal infrared imaging is used to detect vegetation ambient temperature data; and hyperspectral camera is used to detect vegetation cost characteristic data.

[0008] Before the vegetation canopy structure data, the vegetation ambient temperature detection data, and the vegetation component characteristic data are calculated by a data fusion algorithm to obtain vegetation canopy structure parameters, the method further includes:

[0009] The above vegetation canopy structure data are spatially aligned with the satellite elevation data; the above vegetation ambient temperature detection data are temporally aligned using a dynamic time warping algorithm; and the above vegetation cost feature data are spectrally aligned using TV regularization.

[0010] The above-mentioned vegetation canopy structure parameters are introduced to construct a vegetation transpiration intensity outlier detection model, which specifically includes:

[0011] The Penman-Monteith equation was established by introducing canopy structure parameters to calculate transpiration intensity.

[0012] The Penman-Monteith equation uses the Isolation Forest algorithm to construct a transpiration intensity outlier detection model.

[0013] The above-mentioned detection of tree trunk insect infestation vibration signals and tree trunk insect tunnel scanning data based on local drought or waterlogged areas specifically includes:

[0014] Multiple trunk vibration sensors are used to capture trunk insect infestation signals of preset frequencies; X-ray tomography equipment is used to detect insect tunnels inside the trunk.

[0015] The simulation of the pest spread path based on the above-mentioned tree trunk pest infestation vibration signal and the above-mentioned tree trunk pest tunnel scanning data specifically includes:

[0016] Based on the characteristics of pest infestation vibration signals and the infestation of tree trunks, a pest infestation feature library of various pest types was constructed.

[0017] Identify the type of pests in the worm tunnel to be detected based on the above-mentioned borer feature library;

[0018] Collecting multiple sets of changes in the location of the tree trunk insect pests' vibration signals and the spread speed of at least one insect tunnel, and predicting the spread speed of at least one insect tunnel in the direction of the to-be-detected insect tunnel based on a deep learning classification model;

[0019] Generate a three-dimensional model of the pest spread in the above-mentioned insect tunnel to be detected.

[0020] The above-mentioned identification of local drought or waterlogged areas based on the vegetation transpiration intensity outlier detection model specifically includes:

[0021] When there are more outliers in transpiration intensity, the corresponding area is a local waterlogging area; when there are fewer outliers in transpiration intensity, the corresponding area is a local waterlogging area.

[0022] The vegetation environment coupling prediction system based on multi-source remote sensing data fusion includes: a data acquisition module: using multiple sensors to collect vegetation canopy structure data, vegetation environment temperature detection data and vegetation component characteristic data; a data fusion module: using the above vegetation canopy structure data, the above vegetation environment temperature detection data and the above vegetation component characteristic data to calculate vegetation canopy structure parameters through a data fusion algorithm; a model construction module: introducing the above vegetation canopy structure parameters to construct a vegetation transpiration intensity outlier detection model; a region recognition module: identifying local drought or waterlogged areas based on the above vegetation transpiration intensity outlier detection model; a worm tunnel detection module: detecting tree trunk insect pest boring vibration signals and tree trunk worm tunnel scanning data based on local drought or waterlogged areas; and a worm tunnel simulation module: simulating the pest spread path based on the above tree trunk insect pest boring vibration signals and the above tree trunk worm tunnel scanning data.

[0023] An electronic device includes a memory, a processor, and a computer program running on the processor. When the processor executes the computer program, the steps of the vegetation environment coupling prediction method based on multi-source remote sensing data fusion are implemented.

[0024] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the vegetation environment coupling prediction method based on multi-source remote sensing data fusion.

[0025] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0026] This application proposes a vegetation environment coupling prediction method and system based on multi-source remote sensing data fusion, which uses vegetation canopy structure data, vegetation environmental temperature detection data and vegetation component characteristic data to calculate vegetation canopy structure parameters through a data fusion algorithm; introduces vegetation canopy structure parameters to construct a vegetation transpiration intensity outlier detection model, thereby using the characteristics of vegetation canopy structure, environmental temperature and vegetation composition combined with vegetation transpiration intensity to identify local drought or waterlogged areas; and detects trunk insect pest gnawing vibration signals and trunk worm tunnel scanning data based on local drought or waterlogged areas; and simulates the pest spread path in local drought or waterlogged areas based on the trunk insect pest gnawing vibration signals and trunk worm tunnel scanning data. The present invention monitors grassland ecological degradation by identifying drought or waterlogged areas, and can also monitor and predict multi-dimensional ecological behaviors such as pest spread paths, canopy structure, and transpiration intensity fields in real time. It realizes that vegetation management in garden landscaping still lacks analysis of the specific ecological structure of vegetation, as well as in-depth research and application of immediate response mechanisms for disasters, providing more comprehensive, accurate, and real-time intelligent decision-making support for vegetation management, thereby filling the gaps in existing technologies in multimodal ecological event monitoring and prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the examples. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be considered as limiting the scope. A person of ordinary skill in the art can also derive other relevant drawings based on these drawings without inventive effort. In the drawings:

[0028] Figure 1 This is a flowchart of a vegetation environment coupling prediction method based on multi-source remote sensing data fusion according to an embodiment of the present application. DETAILED DESCRIPTION

[0029] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with examples and drawings. The exemplary embodiments of the present invention and their descriptions are only used to explain the present invention and are not intended to limit the present invention.

[0030] Example

[0031] like Figure 1 As shown, an embodiment of the present application provides a vegetation environment coupling prediction method based on multi-source remote sensing data fusion, including: using multiple sensors to collect vegetation canopy structure data, vegetation environment temperature detection data and vegetation component characteristic data; the above-mentioned vegetation canopy structure data, the above-mentioned vegetation environment temperature detection data and the above-mentioned vegetation component characteristic data are calculated through a data fusion algorithm to obtain vegetation canopy structure parameters; the above-mentioned vegetation canopy structure parameters are introduced to construct a vegetation transpiration intensity outlier detection model; local drought or waterlogged areas are identified based on the above-mentioned vegetation transpiration intensity outlier detection model; tree trunk pest boring vibration signals and tree trunk worm tunnel scanning data are detected based on the local drought or waterlogged areas; and pest spread paths are simulated based on the above-mentioned tree trunk pest boring vibration signals and the above-mentioned tree trunk worm tunnel scanning data.

[0032] The above-mentioned multiple sensors are used to collect vegetation canopy structure data, vegetation ambient temperature detection data and vegetation component characteristic data, specifically including:

[0033] LiDAR is used to detect vegetation canopy structure data; thermal infrared imaging is used to detect vegetation ambient temperature data; and hyperspectral camera is used to detect vegetation cost characteristic data.

[0034] The LiDAR system uses a 1550nm solid-state radar, with a point cloud density increased to 500 points / m², enabling penetrating vegetation scanning. The thermal infrared imaging system has been upgraded to a cooled indium antimonide detector (1280×1024 resolution) with a temperature sensitivity of 0.03°C. The hyperspectral camera covers the 400-2500nm wavelength band and subdivides multiple vegetation component characteristic channels, such as the 650nm chlorophyll a and 1450nm water absorption bands.

[0035] Before the vegetation canopy structure data, the vegetation ambient temperature detection data, and the vegetation component characteristic data are calculated by a data fusion algorithm to obtain vegetation canopy structure parameters, the method further includes:

[0036] The above vegetation canopy structure data are spatially aligned with the satellite elevation data; the above vegetation ambient temperature detection data are temporally aligned using a dynamic time warping algorithm; and the above vegetation cost feature data are spectrally aligned using TV regularization.

[0037] Spatial domain registration specifically includes: point cloud rasterization (matching satellite resolution), which can be expressed as:

[0038] grid = voxelize(lidar, resolution=dem.resolution);

[0039] lidar represents point cloud data of vegetation canopy structure data; dem represents satellite elevation data; voxelize represents a voxelization function that divides the lidar data into a three-dimensional grid (voxels); lidar,resolution=dem.resolution means setting the voxel grid resolution of the lidar data to be consistent with the dem data.

[0040] Fine registration based on the iterative closest point (ICP) algorithm can be expressed as:

[0041] transform = ICP_optimize(grid, dem, max_iter=100);

[0042] ICP_optimize indicates the iterative closest point (ICP) algorithm; grid indicates the lidar data after point cloud rasterization; dem indicates the satellite elevation data; grid, dem, max_iter=100 indicates that the grid data uses the dem data as the reference benchmark for spatial resolution registration, and the maximum number of iterations of the ICP algorithm is 100.

[0043] Generate a canopy height model, which can be expressed as:

[0044] CHM = grid - dem .

[0045] Time domain registration (thermal infrared data synchronization) uses the dynamic time warping (DTW) algorithm, which can be expressed as:

[0046] C(i,j) = D(i,j) + min{C(i-1,j), C(i,j-1), C(i-1,j-1)};

[0047] C(i,j) represents the cumulative distance matrix; where D(i,j) represents the distance between the current time series and the standard time series; C(i-1,j) means that a point in the reference sequence one step before D(i,j) matches only one point in the target sequence; C(i,j-1) means that a point in the reference sequence one step before D(i,j) matches multiple consecutive points in the target sequence; C(i-1,j-1) means that a point in the target sequence one step before D(i,j) matches multiple consecutive points in the reference sequence.

[0048] Spectral domain registration (hyperspectral feature extraction): The spectral features of chlorophyll / water / lignin are collected using an end-member spectral library and optimized by maintaining spatial continuity through TV regularization.

[0049] Comprehensive data fusion algorithm for calculating canopy structural parameters , which can be expressed as:

[0050] ;

[0051] Where n represents the number of spatial sampling points in the same canopy area when multiple sensors synchronously collect vegetation canopy structure data, vegetation ambient temperature detection data, and vegetation component characteristic data; is the canopy structure data; is the ambient temperature detection data; T is the affine transformation matrix; λ is the hyperspectral gradient regularization term; It is the vegetation cost characteristic data.

[0052] The above-mentioned vegetation canopy structure parameters are introduced to construct a vegetation transpiration intensity outlier detection model, which specifically includes:

[0053] The Penman-Monteith equation was established by introducing canopy structure parameters to calculate transpiration intensity.

[0054] The Penman-Monteith equation uses the Isolation Forest algorithm to construct a transpiration intensity outlier detection model.

[0055] Specifically, the canopy structure parameter is introduced into the Penman-Monteith equation , which can be expressed as:

[0056] ;

[0057] Indicates the slope of the saturated water vapor pressure-temperature curve; represents the net radiation from the surface; represents the soil heat flux, which is usually negligible; represents the psychrometer constant; represents the average temperature; Indicates the average wind speed; represents the saturated water vapor pressure; Indicates the actual water vapor pressure.

[0058] The saturated water vapor pressure can be calculated as follows:

[0059] ;

[0060] in, The temperature is in degrees Celsius.

[0061] The actual water vapor pressure can be calculated from the relative humidity:

[0062] ;

[0063] in, is the saturated water vapor pressure; is the relative humidity.

[0064] Net surface radiation:

[0065] ;

[0066] in, Represents the incident shortwave radiation (direct solar radiation + diffuse radiation); Represents the shortwave radiation reflected by the surface (depending on the surface albedo ); Represents the downward long-wave radiation of the atmosphere (greenhouse effect); Represents the upward longwave radiation from the surface (surface thermal radiation).

[0067] The above-mentioned detection of tree trunk insect infestation vibration signals and tree trunk insect tunnel scanning data based on local drought or waterlogged areas specifically includes:

[0068] Multiple trunk vibration sensors are used to capture trunk insect infestation signals of preset frequencies; X-ray tomography equipment is used to detect insect tunnels inside the trunk.

[0069] The simulation of the pest spread path based on the above-mentioned tree trunk pest infestation vibration signal and the above-mentioned tree trunk pest tunnel scanning data specifically includes:

[0070] Based on the characteristics of pest infestation vibration signals and the infestation of tree trunks, a pest infestation feature library of various pest types was constructed.

[0071] Identify the type of pests in the worm tunnel to be detected based on the above-mentioned borer feature library;

[0072] Collecting multiple sets of changes in the location of the tree trunk insect pests' vibration signals and the spread speed of at least one insect tunnel, and predicting the spread speed of at least one insect tunnel in the direction of the to-be-detected insect tunnel based on a deep learning classification model;

[0073] Generate a three-dimensional model of the pest spread in the above-mentioned insect tunnel to be detected.

[0074] Optionally, a high-sensitivity accelerometer is used to collect vibration signals from various pests (such as longhorn beetles and bark beetles) as they bore into tree trunks. This allows for the collection of various vibration signal characteristics, such as frequency spectrum analysis (FFT), time-domain energy, and pulse intervals. Combined with the collected data on the types of tree trunks commonly infested by various pests, deep learning is used to identify the corresponding pest types, thereby constructing a library of pest infestation signatures for various pest types. Optionally, based on the location of the pest infestation vibration signal acquisition, i.e., the change in the infestation area of ​​the tunnels obtained through CT scanning or ultrasonic detection, three-dimensional tunnel structural data, such as diameter, branch angle, depth, and tortuosity, is collected to determine the spread velocity in at least one branch direction, i.e., the change in tunnel cross-sectional shape and area. This solves the problem of low tunnel prediction accuracy due to differences in tree hardness and pest type. A three-dimensional model of the pest spread is generated by simulating the changes in the location of the detected tunnel vibration signal and the spread velocity in the corresponding tunnel direction. Among them, the MarchingCubes algorithm can be used for three-dimensional reconstruction, and the predicted three-dimensional model of pest spread can be generated with the help of three-dimensional software simulation.

[0075] The above-mentioned identification of local drought or waterlogged areas based on the vegetation transpiration intensity outlier detection model specifically includes:

[0076] When there are more outliers in transpiration intensity, the corresponding area is a local waterlogging area; when there are fewer outliers in transpiration intensity, the corresponding area is a local waterlogging area.

[0077] When transpiration intensity outliers are clustered and their number is high according to the preset anomaly threshold, it indicates that transpiration is suppressed (e.g., due to root hypoxia), and the area is classified as waterlogged. When transpiration intensity outliers are sparsely distributed and their number is significantly low according to the preset anomaly threshold, it indicates that transpiration is intensified, possibly due to forced water loss, and the area is classified as drought.

[0078] In summary, the embodiments of the present application provide a vegetation environment coupling prediction method and system based on multi-source remote sensing data fusion: vegetation canopy structure data, vegetation environmental temperature detection data and vegetation component characteristic data are used to calculate vegetation canopy structure parameters through a data fusion algorithm; vegetation canopy structure parameters are introduced to construct a vegetation transpiration intensity outlier detection model, so as to use the characteristics of vegetation canopy structure, environmental temperature and vegetation composition combined with vegetation transpiration intensity to identify local drought or waterlogged areas; and based on the local drought or waterlogged areas, trunk insect pest gnawing vibration signals and trunk worm tunnel scanning data are detected; and the pest spread path of the local drought or waterlogged areas is simulated according to the trunk insect pest gnawing vibration signals and trunk worm tunnel scanning data. The present invention realizes the monitoring of grassland ecological degradation by identifying drought or waterlogged areas, and can also monitor and predict multi-dimensional ecological behaviors such as pest spread path, canopy structure, transpiration intensity field in real time, realizing that the vegetation management of garden landscaping still lacks the analysis of vegetation-specific ecological structure, as well as the in-depth research and application of the immediate response mechanism for disasters, providing more comprehensive, accurate and real-time intelligent decision-making support for vegetation management, thereby filling the gap in the existing technology in multimodal ecological event monitoring and prediction. The embodiment of the present application combines multimodal sensing technologies such as lidar, thermal infrared imaging and hyperspectral camera, as well as data fusion algorithms and cellular automaton models, which can not only realize canopy structure analysis and transpiration intensity field dynamic modeling, but also can monitor and predict pest spread path in real time, providing comprehensive, accurate and real-time intelligent decision-making support for vegetation management. By integrating multiple ecological data and intelligent algorithms, the present application aims to achieve more refined and dynamic monitoring and prediction of vegetation environment, and further enhance the scientificity and effectiveness of ecological decision-making.

[0079] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. The vegetation environment coupling prediction method based on multi-source remote sensing data fusion is characterized by: include: Use multiple sensors to collect vegetation canopy structure data, vegetation ambient temperature detection data, and vegetation component characteristic data; The vegetation canopy structure data, the vegetation environment temperature detection data and the vegetation component characteristic data are calculated by a data fusion algorithm to obtain vegetation canopy structure parameters; The vegetation canopy structure parameters are introduced to construct a vegetation transpiration intensity outlier detection model, specifically including: introducing the canopy structure parameters to establish the Penman-Monteith equation to calculate the transpiration intensity; using the IsolationForest algorithm in the Penman-Monteith equation to construct a transpiration intensity outlier detection model; Identifying local drought or waterlogging areas based on the vegetation transpiration intensity outlier detection model, specifically including: when there are many transpiration intensity outliers, the corresponding area is a local waterlogging area; when there are few transpiration intensity outliers, the corresponding area is a local waterlogging area; Detect tree trunk insect infestation vibration signals and trunk insect tunnel scanning data based on local drought or waterlogged areas; The pest spreading path is simulated according to the vibration signal of the tree trunk pest infestation and the tree trunk pest tunnel scanning data.

2. The vegetation environment coupling prediction method based on multi-source remote sensing data fusion according to claim 1 is characterized in that: The use of multiple sensors to collect vegetation canopy structure data, vegetation ambient temperature detection data, and vegetation component characteristic data specifically includes: LiDAR is used to detect vegetation canopy structure data; thermal infrared imaging is used to detect vegetation ambient temperature data; and hyperspectral camera is used to detect vegetation cost characteristic data.

3. The vegetation environment coupling prediction method based on multi-source remote sensing data fusion according to claim 2 is characterized in that: Before the vegetation canopy structure data, the vegetation environment temperature detection data and the vegetation component characteristic data are calculated by a data fusion algorithm to obtain vegetation canopy structure parameters, the method further includes: The vegetation canopy structure data and the satellite elevation data are spatially aligned; the vegetation ambient temperature detection data are temporally aligned using a dynamic time warping algorithm; and the vegetation cost feature data are spectrally aligned using TV regularization.

4. The vegetation environment coupling prediction method based on multi-source remote sensing data fusion according to claim 1 is characterized in that: The detection of tree trunk insect infestation vibration signals and tree trunk insect tunnel scanning data based on local drought or waterlogged areas specifically includes: Multiple trunk vibration sensors are used to capture trunk insect infestation signals of preset frequencies; X-ray tomography equipment is used to detect insect tunnels inside the trunk.

5. The vegetation environment coupling prediction method based on multi-source remote sensing data fusion according to claim 1 is characterized in that: The simulating of the pest spreading path according to the tree trunk pest boring vibration signal and the tree trunk pest tunnel scanning data specifically includes: Based on the characteristics of pest infestation vibration signals and the infestation of tree trunks, a pest infestation feature library of various pest types was constructed. Identify the type of pests in the wormhole to be detected based on the borer feature library; Collecting changes in the collection positions of multiple groups of vibration signals of tree trunk borer pests and the spread speed in at least one direction of an insect tunnel, and predicting the spread speed in at least one direction of the insect tunnel to be detected based on a deep learning classification model; A three-dimensional model of pest spread in the insect tunnel to be detected is generated.

6. The vegetation environment coupling prediction system based on multi-source remote sensing data fusion is characterized by: include: Data acquisition module: uses multiple sensors to collect vegetation canopy structure data, vegetation environment temperature detection data and vegetation component characteristic data; Data fusion module: the vegetation canopy structure data, the vegetation environment temperature detection data and the vegetation component characteristic data are calculated by a data fusion algorithm to obtain vegetation canopy structure parameters; Model construction module: introducing the vegetation canopy structure parameters to construct a vegetation transpiration intensity outlier detection model, specifically including: introducing the canopy structure parameters to establish the Penman-Monteith equation to calculate the transpiration intensity; using the Penman-Monteith equation to use the Isolation Forest algorithm to construct a transpiration intensity outlier detection model; Region identification module: Identifying local drought or waterlogging areas based on the vegetation transpiration intensity outlier detection model, specifically including: when there are many transpiration intensity outliers, the corresponding area is a local waterlogging area; when there are few transpiration intensity outliers, the corresponding area is a local waterlogging area; Insect tunnel detection module: Detects vibration signals of tree trunk infestations and scans of tree trunk insect tunnels based on local drought or waterlogged areas; The insect tunnel simulation module simulates the insect pest spreading path according to the vibration signal of the tree trunk insect pest and the tree trunk insect tunnel scanning data.

7. An electronic device comprising a memory, a processor, and a computer program running on the processor, characterized in that: When the processor executes the computer program, the steps of the vegetation environment coupling prediction method based on multi-source remote sensing data fusion as described in any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the vegetation environment coupling prediction method based on multi-source remote sensing data fusion as claimed in any one of claims 1 to 5 are implemented.

Images