Water bird habitat ecological data analysis method and system, medium and electronic equipment

Abstract

The invention discloses a waterfowl habitat ecological data analysis method and system, a medium and electronic equipment. The waterfowl habitat ecological data analysis method comprises the steps that S1, ecological data of a waterfowl habitat is acquired and preprocessed; s2, constructing a waterfowl habitat ecological model based on the preprocessed data; s3, judging whether the ecological condition of the habitat meets a preset suitability standard or not according to an output result of the ecological model; the system comprises a multi-source data acquisition module based on a system architecture, an intelligent preprocessing module, an ecological model construction module, a habitat suitability evaluation module and a visualization and report module. Compared with the prior art, the method has the advantages of comprehensiveness (2) accuracy; and practicability.

Description

Technical Field

[0001] The present invention relates to the field of big data analysis, and specifically refers to a method, system, medium and electronic device for ecological data analysis of waterbird habitats. Background Art

[0002] As an important part of the natural wetland ecosystem, waterbirds play an important role in the material cycle, energy flow and information transmission of the wetland ecosystem. The health status of waterbird habitats is directly related to the effectiveness of biodiversity conservation and water resource management. Currently, the ecological assessment of waterbird habitats mainly relies on traditional on-site surveys and monitoring. This method is not only time-consuming and laborious, but also greatly affected by human factors, making it difficult to comprehensively and accurately reflect the true situation of the habitats.

[0003] With the rapid development of remote sensing technology, GIS and big data technology, new means have been provided for the ecological assessment of waterbird habitats. However, there are still many deficiencies in the existing technologies in practical applications. For example, the patent with the publication number CN109359411A proposes a method for estimating the vegetation coverage of marsh wetlands under the influence of climate change. Although it can quickly obtain vegetation information, it ignores the comprehensive consideration of key ecological factors such as water quality, soil physical and chemical properties, and waterbird population distribution, resulting in one-sided evaluation results.

[0004] In addition, the patent with the publication number CN119026801A provides a waterbird habitat suitability evaluation system, but there are deficiencies in data preprocessing, model construction and result evaluation. Specifically, the data cleaning and standardization processes are cumbersome and inefficient; the spatial non-stationarity influence of ecological factors is not fully considered during model construction; the result evaluation lacks dynamic threshold management and spatial continuity correction, resulting in limited accuracy and practicality of the evaluation results. Summary of the Invention

[0005] To solve the above technical problems, the technical solution provided by the present invention is: A method for ecological data analysis of waterbird habitats, including:

[0006] S1. Obtain the ecological data of the waterbird habitat and preprocess it;

[0007] S2. Based on the preprocessed data, construct an ecological model of the waterbird habitat;

[0008] S3. According to the output result of the ecological model, judge whether the ecological status of the habitat meets the preset suitability standard.

[0009] Preferably, S1 further includes:

[0010] S1.1. Obtain data on vegetation coverage, underwater terrain (water depth distribution at different water level elevations), water quality parameters, soil physical and chemical properties, waterbird population quantity and distribution information;

[0011] S1.2. Clean the data through statistical methods and interpolation correction, unify the data format standard after cleaning, fuse the standard data, correct the time series, and optimize the data quality.

[0012] Preferably, S1.1 further includes:

[0013] S1.1.1. Obtain Landsat-8 / 9 images with a resolution of 30m, Sentinel-2 images with a resolution of 10m, or unmanned aerial vehicle hyperspectral images with a resolution of 0.1 - 1m through remote sensing images. Take photos of the vegetation canopy within the quadrat using a digital camera, extract the NDVI value through ENVI and calculate the vegetation coverage percentage, establish a regression model, and invert the vegetation coverage; for the vegetation part, add the coverage of different types of vegetation such as trees, shrubs, herbs, and aquatic (wetland) vegetation, especially the coverage of submerged plants, and investigate the specific distribution of different vegetation.

[0014] S1.1.2. Set sampling points, use a multi-parameter water quality meter to measure water temperature, pH, dissolved oxygen, nutrients, chemical oxygen demand, and conductivity (salinity) on-site and upload the data to the terminal;

[0015] S1.1.3. Set sampling plots at different water depth terrains (shoals, exposed beaches, deep waters). Drill and sample at the 0 - 20cm surface soil and 20 - 50cm deep soil layers, record the longitude, latitude, and soil type, send the samples for inspection, measure the pH, organic matter content, particle size distribution, and pollutant concentration of the samples, and upload the data to the terminal; specifically, the soil is divided into onshore, water-land interface, and underwater.

[0016] S1.1.4. Set up infrared cameras in the habitat to monitor the activity rhythm information of waterbirds and upload the data to the terminal.

[0017] Preferably, S2 further includes:

[0018] S2.1. Input the preprocessed raster data and vector data. The raster data includes vegetation coverage, water quality parameters, and soil physical and chemical property data, and the vector data is waterbird population quantity and distribution information data. Use ArcGIS to unify all the data to a resolution of 10m and a coordinate system;

[0019] S2.2. Construct a judgment matrix based on vegetation coverage, water quality parameters, soil physical and chemical property data, and human interference (measurement of sound and light, distance from the road). Calculate the eigenvector corresponding to the maximum eigenvalue of the matrix, normalize it to obtain the weight value and perform a consistency test. If the consistency ratio ≥ 0.1, the judgment matrix needs to be readjusted; otherwise, accept the weight assignment;

[0020] S2.3. Calculate the habitat suitability index. If the habitat suitability index ≥ 0.7, it is marked as "highly suitable habitat"; if 0.4 ≤ habitat suitability index < 0.7, it is marked as "moderately suitable"; otherwise, it is marked as "unsuitable".

[0021] Preferably, S3 further includes:

[0022] S3.1. Input the habitat suitability index raster map, single-factor compliance status matrix, and spatial connectivity index;

[0023] S3.2. Single-factor veto, formula:

[0024]

[0025] where x ik is the value of the kth factor of the ith pixel. If any core factor exceeds the standard, it is directly marked as "unsuitable";

[0026] S3.2. If the single-factor veto is not triggered, enter the habitat suitability index grading, formula:

[0027]

[0028] where level 3 = highly suitable, 2 = moderately suitable, 1 = unsuitable;

[0029] S3.3. Modify the spatial continuity, formula:

[0030]

[0031] where N adj represents the number of similar pixels in the 3×3 neighborhood of the current pixel, and N total represents the total number of neighborhood pixels. If the "highly suitable" patch is isolated, it is downgraded to "moderately suitable".

[0032] Preferably, a supporting ecological data analysis system for waterbird habitats is also provided, including a multi-source data acquisition module, an intelligent preprocessing module, an ecological model construction module, a habitat suitability assessment module, and a visualization and reporting module based on the system architecture:

[0033] The functions of the multi-source data acquisition module include full-channel access and metadata management; the architecture forms include hardware interfaces and software interfaces; the working methods include automated acquisition and manual upload; the interaction methods include dashboards and alarm prompts;

[0034] The functions of the intelligent preprocessing module include one - key cleaning and standardization engine; the architecture forms include pipeline design and GPU acceleration; the working modes include users selecting data cleaning strategies and the system automatically generating preprocessing logs; the interaction modes include visual verification and parameter adjustment;

[0035] The functions of the ecological model construction module include dual - model fusion and real - time sensitivity analysis; the architecture forms include microservice architecture and containerized deployment; the working modes include loading grid data, calling the AHP service to calculate factor weights, and starting model training to generate a spatially varying coefficient raster map; the interaction modes include judgment scoring and model monitoring;

[0036] The functions of the habitat suitability assessment module include multi - level decision trees and dynamic threshold management; the architecture forms include rule engines and parallel computing; the working modes include inputting raster and matrix data, applying veto rules pixel - by - pixel, and outputting a suitability grading map with confidence annotations; the interaction modes include scenario simulation and conflict warning;

[0037] The functions of the visualization and reporting module include multi - dimensional display and intelligent interpretation; the architecture forms include WebGL rendering and template engines; the working modes include extracting statistical indicators and generating reports; the interaction modes include VR viewing and collaborative annotation.

[0038] Preferably, a computer - readable storage medium is also provided, storing a computer program, which implements the steps of the above - mentioned method when executed by a processor.

[0039] Preferably, an electronic device is also provided, including a memory, a processor, and a computer program stored on the memory. The processor implements the steps of the above - mentioned method when executing the program.

[0040] The advantages of the present invention compared with the prior art are as follows: (1) Comprehensiveness: The present invention comprehensively considers multiple ecological factors such as vegetation coverage, water quality parameters, soil physical and chemical properties, and the number and distribution information of waterbird populations, realizing a comprehensive assessment of waterbird habitats; (2) Accuracy: Through the optimization of links such as data cleaning, standardization, and model construction, the present invention improves the accuracy of the assessment results. At the same time, considering the impact of the spatial non - stationarity of ecological factors, the prediction ability of the model is further enhanced; (3) Practicality: The present invention introduces a single - factor one - vote veto rule and a spatial continuity correction mechanism, ensuring the practicality and operability of the assessment results. At the same time, the visualization and reporting module provides users with intuitive and easy - to - understand assessment results, facilitating decision - making for ecological protection and planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a schematic diagram of the steps of the waterbird habitat ecological data analysis method.

[0042] Figure 2 It is a schematic diagram of the architecture of the ecological data analysis system for waterbird habitats. Specific implementation manners

[0043] The present invention will be further described in detail below with reference to the accompanying drawings.

[0044] Example 1

[0045] As Figure 1 shown, this example provides a specific implementation manner of step S1 of the ecological data analysis method for waterbird habitats. S1 is used to obtain and preprocess the ecological data of waterbird habitats, specifically as follows:

[0046] (1) Obtain data on vegetation coverage, water quality parameters, soil physical and chemical properties, waterbird population quantity and distribution information. Among them, for obtaining vegetation coverage, Landsat-8 / 9 with a resolution of 30m, Sentinel-2 with a resolution of 10m or unmanned aerial vehicle hyperspectral images with a resolution of 0.1 - 1m are obtained through remote sensing images. In the quadrat, digital cameras are used to photograph the vegetation canopy, the NDVI value is extracted through ENVI and the vegetation coverage percentage is calculated, and a regression model is established to invert the vegetation coverage. The NDVI calculation formula:

[0047]

[0048] For obtaining water quality parameters, sampling points are set in the habitat according to the grid method (100m×100m) or along the water body gradient (near shore / center / downstream), and a portable multi-parameter water quality meter (such as YSI Pro DSS) is used to measure pH, dissolved oxygen (DO), and conductivity (EC) on-site, and the data is uploaded to the terminal.

[0049] For obtaining soil physical and chemical properties, borehole sampling is carried out according to the soil profile (0 - 20cm surface soil, 20 - 50cm deep soil), the longitude and latitude and soil type (tidal flat / marsh / grassland) are recorded, the samples are sent for inspection, the pH is measured by the potentiometric method, the organic matter content is measured by the potassium dichromate oxidation method, the particle size distribution is measured by a laser particle size analyzer, the concentrations of pollutants such as cadmium and arsenic are analyzed by atomic absorption spectrometry (AAS), and then the soil attribute spatial distribution map is generated using the ArcGIS geostatistical tool (Kriging interpolation method), and the data is uploaded to the terminal.

[0050] For obtaining the waterbird population quantity and distribution, camera traps (Browning or Reconyx) are deployed in key areas of the habitat to automatically photograph and identify the activity rhythms of waterbirds, and the data is uploaded to the terminal.

[0051] (2) Clean the data through statistical methods and interpolation correction, unify the data format standard after cleaning, fuse the standard data, correct the time series, and optimize the data quality.

[0052] In terms of data cleaning, for water quality parameters (such as TN > 10 mg / L) or population numbers (single point > 1000), the 3σ principle or box plot method is used to identify anomalies, and combined with field records to judge whether it is a true value. For missing soil data, the inverse distance weighted method (IDW) is used to interpolate and fill from adjacent points.

[0053] In terms of format standardization, the unified coordinate system is WGS84 UTM, the resolution is resampled to 10 m (Sentinel-2 benchmark), the GDAL library is used to align the UAV data with satellite images, and the error is controlled within 1 pixel. The water quality and soil data are converted into CSV format, and the fields are standardized named, pH → Water_pH, and the unit is unified as mg / L or %.

[0054] In terms of data fusion, historical remote sensing images are registered to the current data using ground control points, and the root mean square error (RMSE) < 0.5 pixels. The vegetation coverage (30 m), water quality (point data), and waterbird distribution (vector surface) are used to generate a 100 m × 100 m grid data set through spatial overlay.

[0055] In terms of time series correction, for vegetation data in different seasons, harmonic analysis is used to eliminate the influence of phenological differences, and the daily rainfall data (from meteorological stations) is matched to the waterbird survey weekly scale through cumulative summation.

[0056] In terms of quality control, for the vegetation / water body classification results extracted by remote sensing, 200 verification points are randomly selected, the confusion matrix is calculated, and the Kappa coefficient ≥ 0.8 is acceptable. For soil heavy metal detection, standard substances (GBW07401) need to be added for calibration, and the relative error < 5%.

[0057] In addition, in this embodiment, conventional artificial on-site investigation technical means are added, and the data after investigation can be analyzed and modeled through big data. In the specific data part, the investigation of benthic animals (quantity and distribution), water depth and water level related data (proportion and distribution of different water depth areas), vegetation coverage (used to refine the distribution and area of specific vegetation species, specific vegetation includes aquatic (wet) vegetation (submerged plants, floating-leaved plants, emergent plants, etc., specific species can include Euryale ferox (habitat for water pheasants), Phragmites australis, Nelumbo nucifera, Nymphaea tetragona, etc., the height of aquatic plants and the distance from the safe bird island), trees, shrubs and herbs in terrestrial plants (crown amplitude, tree height, etc. of trees)); safe islands (including slope, area, substrate composition (silt, hard or soft), etc.), and the safe islands can also be read out by remote sensing and the corresponding regional elevation map is added. Since this part of the content belongs to common means in the prior art, it is not used as the core technical feature of this solution, but only as a supplementary basic technical means for the embodiment, in order to improve the operability and completeness of the embodiment.

[0058] Embodiment 2

[0059] As shown Figure 1 in the figure, this embodiment provides a specific implementation manner of step S2 of the ecological data analysis method for waterbird habitats. S1 constructs an ecological model of waterbird habitats based on the preprocessed data, which is specifically as follows:

[0060] (1) Input the preprocessed raster data and vector data. The raster data includes vegetation coverage, water quality parameters, and soil physical and chemical property data, and the vector data is the data of waterbird population quantity and distribution information. Use ArcGIS to unify all data to a resolution of 10m and a coordinate system;

[0061] (2) Construct a judgment matrix based on vegetation coverage, water quality parameters, soil physical and chemical property data, and human disturbance. Calculate the eigenvector corresponding to the maximum eigenvalue (λ_max) of the matrix, normalize it to obtain the weight value and perform a consistency test. If the consistency ratio ≥ 0.1, then the judgment matrix needs to be readjusted; otherwise, accept the weight assignment;

[0062] (3) Quantify the spatially non-stationary impact of environmental factors on waterbird distribution. Use the Gaussian kernel function to determine the spatial weight. If the water quality coefficient > 0 and passes the significance test, then the waterbird density in this area is sensitive to water quality; otherwise, water quality is not the main limiting factor;

[0063] (4) Calculate the habitat suitability index. If the habitat suitability index ≥ 0.7, then mark it as "highly suitable habitat"; if 0.4 ≤ habitat suitability index < 0.7, then mark it as "moderately suitable"; otherwise, mark it as "unsuitable".

[0064] Among them, the ecological model of waterbird habitats adopts an analytic hierarchy process (AHP) + geographically weighted regression (GWR) combined model, taking into account expert experience and spatial heterogeneity. The model aims to establish a habitat suitability index (HSI) to quantify the influence of environmental factors on the survival of waterbirds.

[0065] First, the analytic hierarchy process (AHP) is used to determine the factor weights, and the importance of factors such as vegetation coverage (V), water quality (W), soil quality (S), and human disturbance (H) is compared pairwise. The formula:

[0066]

[0067] Code example of the analytic hierarchy process (AHP):

[0068] from pyDecision.algorithm import ahp

[0069] matrix = [[1,3,5,7],[1 / 3,1,3,5],[1 / 5,1 / 3,1,3],[1 / 7,1 / 5,1 / 3,1]]

[0070] weights, cr = ahp(matrix)

[0071] Geographically Weighted Regression (GWR) code example:

[0072] library(spgwr)

[0073] gwr_model<-gwr(y~x1+x2, data = spdf, bandwidth = 0.1)

[0074] summary(gwr_model$SDF$coefficients)

[0075] Then, find the eigenvector corresponding to the maximum eigenvalue (λ_max) of the matrix. After normalization, the weight values are obtained. The formula is:

[0076] W = [w V , w W , w S , w H T

[0077] Consistency test: If the consistency ratio (CR) < 0.1, the test is passed. The formula is:

[0078]

[0079] Among them, RI is the random consistency index. When n = 4, RI = 0.89. If CR ≥ 0.1, the judgment matrix needs to be re-adjusted; otherwise, accept the weight assignment.

[0080] Secondly, Geographically Weighted Regression (GWR) is used to quantify the spatially non-stationary impact of environmental factors on the distribution of waterbirds. The formula is:

[0081] y i = β0(u i , v i ) + β1(u i , v i )x 1i + β2(u i , v i )x 2i +…+ ∈ i Here, an algorithm can be added, that is, the appearance of specific characteristic waterbirds and the appearance of protected-level species can be given additional weighted scores to increase the hsi of this area or the overall habitat.

[0082] ​Among them, y i represents the waterbird density at the i-th location (birds / km 2 ), (u i , v i ) represents the spatial coordinates; β k (u i , v i ) represents the regression coefficient varying with spatial location, and x ki represents the environmental variables (such as NDVI, TN concentration, etc.).

[0083] Calculation formula for the habitat suitability index (HSI):

[0084]

[0085] Among them, w k represents the factor weight obtained by AHP, and f k (x k ) represents the single-factor suitability function (normalized to 0 - 1).

[0086] If HSI ≥ 0.7, it is marked as "highly suitable habitat"; if 0.4 ≤ HSI < 0.7, it is marked as "moderately suitable"; otherwise, it is marked as "unsuitable".

[0087] To improve the model accuracy, model validation is also introduced in this embodiment. The AUC value is calculated through the ROC curve (>0.8 indicates an excellent model). If AUC < 0.7, variables need to be reselected or weights adjusted; otherwise, the model can be used.

[0088] The following is the specific implementation method of this embodiment:

[0089] Input: Vegetation coverage in a certain area = 65%, TN concentration = 1.2 mg / L, human disturbance intensity = 0.3 (normalized value).

[0090] Calculation:

[0091] f V =(65 - 30) / 40 = 0.875

[0092] f W =(1.2 < 2.0) → 1.0 (assuming that the TN threshold ≤ 2.0 mg / L is suitable)

[0093] f H =1 - 0.3 = 0.7

[0094] HSI = 0.55×0.875 + 0.26×1.0 + 0.13×0.8 + 0.06×0.7 = 0.85

[0095] Output: If HSI = 0.85 ≥ 0.7, it is determined as "highly suitable habitat", and it is recommended to be included in the protected area.

[0096] Example 3

[0097] As Figure 1 shown, this example provides a specific implementation of step S3 of the ecological data analysis method for waterbird habitats. S3 determines whether the ecological status of the habitat meets the preset suitability criteria according to the output results of the ecological model, as follows:

[0098] (1) Input data, including the habitat suitability index (HSI) raster map (0-1 standardized value); single-factor compliance status matrix (binary raster, 1 = compliant, 0 = non-compliant); spatial connectivity index (proportion of adjacent pixels in patches). The preset criteria for this model are as follows: ① Core factor threshold: vegetation coverage ≥ 30% (S V = 0.3); water body TN ≤ 2.0 mg / L (S w = 0.2); soil lead ≤ 300 mg / kg (S p = 300). ② HSI classification threshold: highly suitable: T1 = 0.7; moderately suitable: T2 = 0.5; unsuitable: T3 < 0.5. ③ Spatial continuity threshold: proportion of suitable patches in the neighborhood ≥ 70% (C min = 70%).

[0099] (2) Single-factor veto, formula:

[0100]

[0101] where x ik is the value of the kth factor for the ith pixel. If any core factor exceeds the standard, it is directly marked as "unsuitable";

[0102] Code example:

[0103]

[0104] (3) If the single-factor veto is not triggered, then enter the habitat suitability index classification, formula:

[0105]

[0106] where level 3 = highly suitable, 2 = moderately, 1 = unsuitable.

[0107] Habitat suitability index classification code example:

[0108]

[0109] (3) Correct spatial continuity, formula:

[0110]

[0111] Among them, N adj represents the number of homogeneous pixels within the 3×3 neighborhood of the current pixel, and N total represents the total number of neighborhood pixels. If the "highly suitable" patch is isolated, it is downgraded to "moderately suitable".

[0112] Example code for correcting spatial continuity:

[0113]

[0114] The full - process judgment steps of S3 are as follows: ①Data pre - processing: Unify the HSI, single - factor data, and spatial connectivity index to the same resolution and coordinate system. ②Pixel - by - pixel judgment: Traverse each pixel, check whether the single - factor veto is triggered. For non - vetoed pixels, divide the preliminary level according to the HSI value. For initially highly suitable pixels, perform neighborhood connectivity correction. ③Result aggregation: Statistically analyze the area and spatial distribution of pixels at each level; if the proportion of the area of the highly suitable area ≥60% and the area of the largest patch > 1 km 2 , output "overall suitable"; otherwise, output "locally suitable for repair".

[0115] Example code for the full - process judgment steps of S3:

[0116]

[0117]

[0118]

[0119] The following is the specific implementation method of this embodiment:

[0120] For a certain pixel, HSI = 0.75, vegetation coverage = 25% (not up to standard), TN = 1.8 mg / L (up to standard), soil lead = 200 mg / kg (up to standard). Start single - factor veto:

[0121] Veto = 1 (because vegetation < 30%) → Class = 1

[0122] Final output: Marked as "unsuitable".

[0123] Example 4

[0124] As Figure 1 and Figure 2As shown in the figure, this embodiment provides a specific implementation of the data acquisition and preprocessing process. In the data collection stage, Landsat-8 / 9 satellite remote sensing images are used to obtain vegetation cover data with a resolution of 30 meters. High-resolution images with a resolution of 10 meters are obtained through Sentinel-2 satellites for fine vegetation analysis. UAVs are flown to take hyperspectral images with a resolution of 0.1-1 meter, covering the entire study area. Multiple sampling points are set up in the study area, and a multi-parameter water quality meter is used to measure on-site data such as pH, dissolved oxygen, and conductivity. Borehole sampling is carried out in different soil layers (0-20 cm and 20-50 cm) to measure soil pH, organic matter content, particle size distribution, and pollutant concentration. The activity rhythm of waterbirds is monitored by using infrared cameras + integrated audio and video detection + on-site human monitoring, and the population quantity and distribution information of waterbirds are recorded.

[0125] In the data preprocessing stage, the ENVI software is used to extract the NDVI value from the remote sensing images and calculate the vegetation cover percentage. Statistical analysis and interpolation correction are carried out on the water quality and soil data to unify the data format. The vegetation, water quality, soil, and waterbird data are fused, and the time series is corrected to ensure data quality.

[0126] Example Five

[0127] As Figure 1 and Figure 2 shown, this embodiment provides a specific implementation of the ecological model construction and habitat suitability assessment process. In the model construction stage, the preprocessed raster data (vegetation coverage, water quality parameters, soil physical and chemical properties) and vector data (waterbird population quantity and distribution) are input. ArcGIS is used to unify all data to a resolution of 10 meters and the same coordinate system. A judgment matrix is constructed, including factors such as vegetation coverage, water quality, soil physical and chemical properties, and human interference, and the weight value is calculated and consistency test is carried out. The Gaussian kernel function is used to quantify the impact of environmental factors on the spatial non-stationarity of waterbird distribution.

[0128] In the habitat suitability assessment stage, the habitat suitability index is calculated and classified according to the preset threshold (such as ≥0.7 is highly suitable, 0.4-0.7 is moderately suitable, <0.4 is unsuitable). The single-factor one-vote veto rule is applied. If any core factor (such as severe water quality exceeding the standard) does not meet the standard, it is directly marked as unsuitable. The spatial continuity is corrected to ensure that the "highly suitable" patches are not isolated. If they are isolated, they are downgraded to "moderately suitable".

[0129] Example Five

[0130] As Figure 1 and Figure 2As shown, this embodiment presents the specific implementation of the construction and application of a data analysis system, and designs a system architecture based on a multi-source data collection module, an intelligent preprocessing module, an ecological model construction module, a habitat suitability assessment module, and a visualization and reporting module. The multi-source data collection module supports full-channel access and metadata management, and realizes automated collection and manual upload through hardware interfaces and software interfaces. The intelligent preprocessing module includes a one-key cleaning and standardization engine, adopts a pipeline design and GPU acceleration, and provides visualization verification and parameter adjustment functions. The ecological model construction module integrates dual-model fusion and real-time sensitivity analysis functions, adopts a microservices architecture and containerized deployment, and supports grid data loading, AHP service invocation, and model training. The habitat suitability assessment module includes a multi-level decision tree and dynamic threshold management, adopts a rule engine and parallel computing, supports raster and matrix data input, and outputs a suitability grading map with confidence annotations. The visualization and reporting module provides multi-dimensional display and intelligent interpretation functions, adopts WebGL rendering and a template engine, and supports statistical index extraction and report generation.

[0131] The overall workflow of the system includes:

[0132] (1) After the system starts, the multi-source data collection module begins to collect data from various data sources (such as remote sensing devices, water quality monitoring stations, soil testing laboratories, infrared cameras, etc.). Receive data from different data sources, including vegetation coverage, water quality parameters, soil physical and chemical properties, waterbird population quantity and distribution information, etc. Record the metadata of each data source, such as data time, source, accuracy, etc., for subsequent data management and traceability.

[0133] (2) The intelligent preprocessing module cleans the received data, removes incorrect, duplicate, and abnormal data to ensure the accuracy and reliability of the data. Unify the format and standard of the cleaned data, such as resolution, coordinate system, etc., for subsequent data analysis and processing. Integrate the standardized data into a unified dataset to provide a basis for subsequent model construction.

[0134] (3) Input the preprocessed data into the ecological model construction module, including raster data and vector data. Use tools such as ArcGIS to unify the data to the same resolution and coordinate system, construct a judgment matrix, calculate the factor weights, and perform a consistency test. Then, quantify the spatial non-stationary impact of environmental factors on waterbird distribution, and use the Gaussian kernel function to determine the spatial weights. Based on the input data and the constructed model, perform model training to generate a spatially varying coefficient raster map for subsequent habitat suitability assessment.

[0135] (4) Input the spatially varying coefficient raster map, single-factor compliance status matrix, and spatial connectivity index output by the ecological model construction module into the habitat suitability assessment module. According to the preset criteria, conduct a one-vote veto judgment on any core factor. If it exceeds the standard, directly mark it as "unsuitable". If the single-factor one-vote veto is not triggered, classify the habitat according to the habitat suitability index, such as "highly suitable", "moderately suitable", and "unsuitable". Perform spatial continuity correction on the classified habitat to ensure the accuracy and rationality of the evaluation results.

[0136] (5) The visualization and reporting module extracts the evaluation results and relevant data from the habitat suitability assessment module. Based on the extracted data, generate a report containing content such as data overview, analysis results, and suggestions. Use visualization tools such as charts and maps to intuitively display the evaluation results and data distribution. This module provides a user interaction interface, allowing users to view the report, analysis results, and data visualization display, and perform operations such as collaborative annotation and VR viewing.

[0137] Example Six

[0138] This example aims to further refine and expand the method in the above-mentioned invention patent claim book, especially for more detailed evaluation of specific bird species (wading birds, swimming birds, ducks, plovers, etc.) and different regions. By introducing considerations of the ecological needs of specific bird species and regional differences, this example can more accurately evaluate the suitability of waterbird habitats and provide a scientific basis for ecological protection and restoration.

[0139] Step 1.1: On the basis of the original data, add ecological need data for specific bird species, such as the need of wading birds for shallow water areas, the need of swimming birds for open water areas, the need of ducks and geese for food abundance, and the need of plovers for tidal flats and wetlands. This data can be obtained through literature research, expert consultation, and on-site observation.

[0140] Step 1.2: According to the regional characteristics, divide different ecological regions, such as rivers, lakes, tidal flats, wetlands, etc., and collect ecological data for each region separately. At the same time, consider the ecological connections between regions, such as the interaction between the upstream and downstream of rivers and the interaction between wetlands and surrounding farmlands.

[0141] Step 1.3: Preprocess the collected data, including data cleaning, format unification, time series correction, etc., to ensure the accuracy and consistency of the data.

[0142] Step 2.1: In ArcGIS, unify the preprocessed raster data and vector data to the same resolution and coordinate system, and input the ecological need data and regional difference data of specific bird species respectively.

[0143] Step 2.2: When constructing the judgment matrix, in addition to considering the original vegetation coverage, water quality parameters, soil physical and chemical properties, and human disturbances, it is also necessary to add the ecological demand factors of specific bird species and the regional difference factors. Calculate the weights of each factor by the AHP method and conduct a consistency test.

[0144] Step 2.3: When quantifying the spatially non-stationary impact of environmental factors on the distribution of specific bird species, use the Gaussian kernel function to determine the spatial weights, and pay special attention to the environmental factors sensitive to specific bird species, such as the sensitivity of wading birds to water depth and substrate, and the sensitivity of swimming birds to water area and food resources.

[0145] Step 2.4: When calculating the habitat suitability index, set different thresholds and weights according to the ecological needs of specific bird species and regional differences to obtain a more accurate habitat suitability classification result.

[0146] Step 3.1: Input the habitat suitability index raster map, single-factor compliance status matrix, and spatial connectivity index, while considering the ecological needs of specific bird species and regional differences.

[0147] Step 3.2: In the single-factor veto rule, add the core ecological demand factors of specific bird species, such as the water depth factor for wading birds and the water area factor for swimming birds. If any core factor exceeds the standard, it is directly marked as "unsuitable".

[0148] Step 3.3: In the habitat suitability index classification, adjust the classification criteria and weights according to the ecological needs of specific bird species and regional differences to obtain a more accurate suitability classification result.

[0149] Step 3.4: When correcting the spatial continuity, consider the migration paths and ecological connections of specific bird species, such as the migration needs of wading birds between the upper and lower reaches of the river and the movement needs of swimming birds between different waters. If the "highly suitable" patches are isolated and do not meet the ecological needs of specific bird species, they are downgraded to "moderately suitable".

[0150] Step 4.1: Use WebGL rendering and template engine to present the habitat suitability classification results in a multi-dimensional display manner, including maps, charts, animations, etc.

[0151] Step 4.2: In the report, in addition to including statistical indicators and suitability classification results, it is also necessary to specifically analyze the impact of the ecological needs of specific bird species and regional differences, and put forward targeted ecological protection and restoration suggestions.

[0152] In this embodiment, by introducing considerations of the ecological needs of specific bird species and regional differences, the original ecological data analysis method for waterbird habitats is further refined and expanded. Through more accurate data acquisition, model construction, and suitability assessment, this embodiment can more accurately evaluate the suitability of waterbird habitats and provide a more scientific basis for ecological protection and restoration.

[0153] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of the present invention claimed is defined by the appended claims and their equivalents.

Claims

1. Method for analyzing ecological data of waterbird habitats, characterized in that Including: S1. Obtain and preprocess the ecological data of waterbird habitats; S2. Construct an ecological model of waterbird habitats based on the preprocessed data; S3. According to the output results of the ecological model, determine whether the ecological status of the habitat meets the preset suitability criteria.

2. The ecological data analysis method for waterbird habitats according to claim 1, wherein The S1 further includes: S1.

1. Obtain data on vegetation coverage, underwater topography (water depth distribution at different water level elevations), water quality parameters, soil physical and chemical properties, waterbird population quantity and distribution information; S1.

2. Clean the data through statistical methods and interpolation correction, unify the data format standards after cleaning, fuse the standard data, correct the time series, and optimize the data quality.

3. The ecological data analysis method for waterbird habitats according to claim 1, wherein S1.1 further includes: S1.1.

1. Obtain Landsat-8 / 9 images with a resolution of 30m, Sentinel-2 images with a resolution of 10m, or UAV hyperspectral images with a resolution of 0.1 - 1m through remote sensing images. Take photos of the vegetation canopy within the quadrats using a digital camera. Extract the NDVI value through ENVI and calculate the vegetation coverage percentage. Establish a regression model to invert the vegetation coverage; the vegetation includes the coverage of trees, shrubs, grasses, aquatic (wet) vegetation, and submerged plants, and investigate the specific distribution of different vegetation; S1.1.

2. Set sampling points, use a multi-parameter water quality meter to measure water temperature, pH, dissolved oxygen, nutrient salts, chemical oxygen demand, conductivity (salinity) on-site and upload the data to the terminal; S1.1.

3. Set sampling in plots with different water depth topographies (shoals, exposed beaches, deep waters). Drill for samples at the 0 - 20cm surface soil and 20 - 50cm deep soil layers, record the longitude and latitude and soil type, send the samples for inspection, measure the pH, organic matter content, particle size distribution, and pollutant concentration of the samples and upload the data to the terminal; specifically, the soil is divided into onshore, land-water interface, and in-water; S1.1.

4. Set up infrared cameras in the habitat to monitor the activity rhythm information of waterbirds and upload the data to the terminal.

4. The ecological data analysis method for waterbird habitats according to claim 1, wherein The S2 further includes: S2.

1. Input the preprocessed raster data and vector data. The raster data includes vegetation coverage, water quality parameters, and soil physical and chemical property data, and the vector data is the waterbird population quantity and distribution information data. Use ArcGIS to unify all the data to a resolution of 10m and the coordinate system; S2.

2. Construct a judgment matrix based on vegetation coverage, water quality parameters, soil physical and chemical property data, and human interference. Calculate the eigenvector corresponding to the maximum eigenvalue (λ_max) of the matrix, normalize it to obtain the weight value and perform a consistency test. If the consistency ratio ≥ 0.1, the judgment matrix needs to be readjusted; otherwise, accept the weight distribution; S2.

3. Quantify the spatially non-stationary influence of environmental factors on waterbird distribution, use the Gaussian kernel function to determine the spatial weight. If the water quality coefficient > 0 and passes the significance test, the waterbird density in this area is sensitive to water quality; otherwise, water quality is not the main limiting factor; S2.

4. Calculate the habitat suitability index. If the habitat suitability index ≥ 0.7, mark it as "highly suitable habitat"; if 0.4 ≤ habitat suitability index < 0.7, mark it as "moderately suitable"; otherwise, mark it as "unsuitable".

5. The ecological data analysis method for waterbird habitats according to claim 1, characterized in that The S3 further includes: S3.

1. Input the habitat suitability index grid map, single-factor compliance status matrix, and spatial connectivity index; S3.

2. Single-factor veto, formula: where x ik is the k-th factor value of the i-th pixel. If any core factor exceeds the standard, it is directly marked as "unsuitable". S3.

2. If the single-factor veto is not triggered, enter the habitat suitability index grading, formula: Among them, level 3 = highly suitable, 2 = medium, 1 = unsuitable; S3.

3. Modify the spatial continuity, formula: Among them, N adj represents the number of homogeneous pixels within the 3×3 neighborhood of the current pixel, and N total represents the total number of neighborhood pixels. If the "highly suitable" patch is isolated, it is downgraded to "moderately suitable".

6. The waterbird habitat ecological data analysis system includes a multi-source data acquisition module, an intelligent preprocessing module, an ecological model construction module, a habitat suitability assessment module, and a visualization and reporting module based on the system architecture. It is characterized in that: The functions of the multi-source data acquisition module include full-channel access and metadata management; the architecture form includes hardware interfaces and software interfaces; the working mode includes automated acquisition and manual upload; the interaction mode includes dashboards and alarm prompts; The functions of the intelligent preprocessing module include one-key cleaning and a standardization engine; The architecture form includes pipeline design and GPU acceleration; the working mode includes the user selecting a data cleaning strategy and the system automatically generating preprocessing logs; the interaction mode includes visual verification and parameter adjustment; The functions of the ecological model construction module include dual-model fusion and real-time sensitivity analysis; The architecture form includes a microservices architecture and containerized deployment; the working mode includes loading grid data, calling the AHP service to calculate factor weights, and starting model training to generate a spatially varying coefficient grid map; the interaction mode includes judgment scoring and model monitoring; The functions of the habitat suitability assessment module include multi-level decision trees and dynamic threshold management; the architecture form includes a rule engine and parallel computing; the working mode includes inputting grid and matrix data, applying veto rules pixel by pixel, and outputting a suitability grading map with confidence annotations; the interaction mode includes scenario simulation and conflict warning; The functions of the visualization and reporting module include multi-dimensional display and intelligent interpretation; the architecture form includes WebGL rendering and a template engine; the working mode includes extracting statistical indicators and generating reports; the interaction mode includes VR viewing and collaborative annotation.

7. A computer-readable storage medium storing a computer program, characterized in that, When the program is executed by the processor, the steps of the method described in any one of claims 1-5 are implemented.

8. An electronic device, comprising a memory, a processor, and a computer program stored on the memory, characterized in that, When the processor executes the program, the steps of the method described in any one of claims 1-5 are implemented.

Images