A wetland monitoring method, device and medium combining multi-source remote sensing and unmanned aerial vehicles

Abstract

The application discloses a kind of multi-source remote sensing combination wetland monitoring method, equipment and medium of unmanned plane, it is related to wetland monitoring technical field, the application has realized the efficient and accurate of wetland monitoring by the combination of satellite remote sensing and unmanned plane technology.First, satellite remote sensing image is carried out space-time continuity and noise suppression pretreatment, improves image quality.Then, segmentation fusion processing is used to optimize image, and the distribution range and community boundary of species are accurately extracted.Based on this data, the monitoring path of unmanned plane is planned, and the monitoring efficiency is improved.The high-resolution multispectral image obtained by unmanned plane is radiometrically corrected and geometrically corrected, and is fused with satellite remote sensing image to generate a high-precision wetland species distribution map.The scheme fully utilizes the advantages of satellite remote sensing in wide coverage and unmanned plane in high resolution, and solves the problems of data quality, information extraction and monitoring efficiency in traditional methods, providing strong support for wetland protection and management.

Description

Technical Field

[0001] This invention relates to the field of wetland monitoring technology, specifically to a wetland monitoring method, equipment, and medium using multi-source remote sensing combined with unmanned aerial vehicles (UAVs). Background Technology

[0002] As vital ecosystems with diverse and unique ecological functions, wetlands require careful monitoring for their protection, management, and sustainable development. Wetlands serve as habitats for numerous organisms, including fish, amphibians, aquatic plants, and various waterbirds. However, wetlands are often impacted by surrounding human activities, such as agricultural non-point source pollution and industrial wastewater discharge. Therefore, monitoring wetland vegetation growth and further monitoring water bodies based on species distribution are necessary to effectively implement preventative measures and safeguard the wetland's ecological environment.

[0003] Traditional wetland monitoring methods rely heavily on field surveys. While this approach can obtain relatively accurate local information, it has significant drawbacks such as low efficiency, high labor costs, and limited monitoring range, making it difficult to meet the needs of dynamic monitoring of large-area wetlands over long periods of time.

[0004] The emergence of satellite remote sensing technology has brought new opportunities for wetland monitoring. It can periodically acquire large-scale wetland images, providing macroscopic information such as wetland land cover types, vegetation distribution, and changes in water area. However, the spatial resolution of satellite remote sensing data is relatively limited, making it difficult to accurately depict the complex ecological structure, vegetation community details, and small-scale ecological processes within wetlands.

[0005] Unmanned aerial vehicle (UAV) remote sensing technology, with its high spatial resolution and flexible maneuverability, can acquire more detailed wetland image information in local areas, effectively compensating for the shortcomings of satellite remote sensing in monitoring at the microscopic level. However, UAVs are limited by factors such as endurance and flight range, and cannot provide comprehensive coverage monitoring of large areas of wetlands when used alone.

[0006] Furthermore, existing wetland monitoring often lacks in-depth integration and collaborative analysis of ground-based monitoring data and remote sensing data, resulting in insufficient, in-depth, and precise monitoring of wetland ecosystems. The rise of smart plot technology offers a solution to this problem. By setting up smart plots within wetlands, real-time in-situ monitoring data across multiple dimensions, such as meteorology, soil moisture, water quality, and biodiversity, can be acquired, providing richer and more accurate ground-based information for wetland monitoring and effectively complementing multi-source remote sensing data. Summary of the Invention

[0007] The technical problem this invention aims to solve is the low precision of existing remote sensing image processing and UAV flight monitoring path planning, which further leads to low precision in the collaborative analysis of data from both sources. The goal is to provide a wetland monitoring method, equipment, and medium that combines multi-source remote sensing with UAV technology. By integrating satellite remote sensing and UAV technology, this method achieves high efficiency and accuracy in wetland monitoring. First, the satellite remote sensing images undergo spatiotemporal continuity and noise suppression preprocessing to improve image quality. Next, segmentation and fusion processing is used to optimize the images, accurately extracting species distribution ranges and community boundaries. Based on this data, UAV monitoring paths are planned to improve monitoring efficiency. The high-resolution multispectral images acquired by the UAV, after radiometric and geometric correction, are fused with the satellite remote sensing images to generate a high-precision wetland species distribution map. This scheme fully leverages the wide coverage of satellite remote sensing and the high resolution of UAVs, overcoming the shortcomings of traditional methods in terms of data quality, information extraction, and monitoring efficiency, and providing strong support for wetland protection and management.

[0008] This invention is achieved through the following technical solution:

[0009] The first aspect of this invention provides a wetland monitoring method combining multi-source remote sensing and unmanned aerial vehicles (UAVs), comprising the following specific steps:

[0010] Multi-band remote sensing images of the target detection area are acquired based on satellite remote sensing.

[0011] Preprocessing is performed on the multi-band remote sensing images of the target detection area to obtain spatiotemporally continuous and noise-suppressed multi-band remote sensing images.

[0012] Multi-band remote sensing images that are spatiotemporally continuous and have been noise-suppressed are segmented and fused to obtain optimized multi-band remote sensing images.

[0013] Based on the band information of the optimized multi-band remote sensing image, spectral and spatial features are extracted.

[0014] Based on spectral and spatial features, the species distribution range and community boundaries are extracted to obtain community distribution data of the target detection area;

[0015] Based on the community distribution data of the target detection area, plan the monitoring path of the UAV;

[0016] The target area is monitored by a drone, and multispectral image data of the drone is acquired. The multispectral image data of the drone contains multiple bands with wavelength ranges similar to those of the remote sensing image data.

[0017] Radiometric and geometric corrections are performed on UAV multispectral image data to obtain UAV hyperspectral image data.

[0018] Band matching is performed between UAV hyperspectral image data and remote sensing image data, and the UAV hyperspectral image data and remote sensing image data are fused to identify wetland distribution areas and generate wetland species distribution maps.

[0019] Furthermore, the preprocessing of the multi-band remote sensing image of the target detection area specifically includes:

[0020] Interpolation and resampling are performed on the spatial features of coarse-resolution multi-band remote sensing images.

[0021] The time series data of the coarse-resolution image after interpolation and resampling are matched pixel-level with the cloudless high-resolution multi-band remote sensing image to obtain the corrected time series.

[0022] The missing values ​​in the original multi-band remote sensing images are filled by spatiotemporal modeling to restore the complete time series;

[0023] A weighted seasonality-trend decomposition filter is used to smooth the time series, outputting spatiotemporally continuous and noise-suppressed multi-band remote sensing images.

[0024] Furthermore, the segmentation and fusion processing of the spatiotemporally continuous and noise-suppressed multi-band remote sensing images specifically includes:

[0025] Randomly select a single pixel from a multi-band remote sensing image;

[0026] Within the target area, homogeneous and heterogeneous regions are segmented based on the pixels;

[0027] Merge homogeneous regions into a new object;

[0028] The heterogeneous region is segmented, and the homogeneous region within the heterogeneous region is merged into a new object. This process continues until the heterogeneous region is completely segmented, resulting in a segmented multi-band remote sensing image.

[0029] The segmented multi-band remote sensing images are integrated to obtain optimized multi-band remote sensing images.

[0030] Furthermore, based on the band information of the optimized multi-band remote sensing image, spectral and spatial features are extracted, specifically including:

[0031] Extract the spectral reflectance curve of each pixel to obtain the spectral characteristics reflecting the vegetation species;

[0032] The spatial resolution of multi-band remote sensing images is obtained, and texture features are calculated to obtain spatial features reflecting vegetation communities.

[0033] Furthermore, the extraction of species distribution range and community boundaries based on spectral and spatial features to obtain community distribution data of the target detection area specifically includes:

[0034] The vegetation index is calculated based on the band information of the optimized multi-band remote sensing image.

[0035] Based on spectral and spatial features, morphological methods were used to extract the boundary information of each vegetation community unit, thereby obtaining the species distribution range and community boundary.

[0036] Based on vegetation index, species distribution range and community boundary, a threshold is set to separate vegetated areas from non-vegetated areas;

[0037] The vegetation area is divided into multiple community units using a region growing algorithm to obtain community distribution data for the target detection area.

[0038] Furthermore, the planning of the UAV's monitoring path based on the community distribution data of the target detection area specifically includes:

[0039] Based on the community distribution data of the target detection area, the types of vegetation communities that need to be monitored are determined according to the needs of the target detection area, and priorities are assigned to different communities.

[0040] Obtain drone performance parameters and, in conjunction with community priority, divide the monitoring area;

[0041] Based on the terrain data of the monitored area, calculate the forward overlap and lateral overlap;

[0042] Based on the aforementioned overlap requirements, an initial waypoint distribution is automatically generated;

[0043] Data was collected from the area covered by the initial waypoint distribution, and the overlap and stitching quality of the data were analyzed.

[0044] If the overlap and stitching quality do not meet the set thresholds, adjust the waypoint density.

[0045] The monitoring path for the UAV is generated based on the adjusted waypoint density.

[0046] Furthermore, the radiometric and geometric corrections are performed on the UAV multispectral image data:

[0047] The radiation correction includes:

[0048] The pixel brightness values ​​of the original image are converted into atmospheric outer surface reflectance data, and the radiometric calibration parameters monitored by the sensor are extracted to obtain radiometric calibration data.

[0049] Using the radiation value of the calibrated tarpaulin as the standard value, the radiation-calibrated data is subjected to reflection correction, and the radiation value data is converted into reflectance data.

[0050] The geometric correction includes:

[0051] Imaging is performed using semiconductor photosensitive elements to obtain spectral information along a single line in space;

[0052] The acquisition of spatial images and spectral data is completed through mechanical push-broom scanning.

[0053] The image data of the entire area acquired according to the monitoring path of the UAV will be stitched together in multiple rows to obtain the UAV hyperspectral image data.

[0054] Furthermore, the step of performing band matching between UAV hyperspectral image data and remote sensing imagery, fusing the UAV hyperspectral image data and remote sensing imagery, identifying wetland distribution areas, and generating a wetland species distribution map specifically includes:

[0055] Based on the band information of multi-band remote sensing images, determine the band range corresponding to the hyperspectral data;

[0056] Spectral reflectance data corresponding to the bands of multi-band remote sensing images are extracted from hyperspectral data, and the extracted spectral reflectance data is consistent with the band information of the remote sensing images in terms of wavelength range and resolution.

[0057] Based on the spectral reflectance information of the extracted hyperspectral data, a vegetation index identical to the remote sensing vegetation index is calculated.

[0058] Obtain vegetation indices calculated from the band reflectance values ​​of multi-band remote sensing images;

[0059] The vegetation index of the calculated UAV imagery is fused with the vegetation index of the remote sensing data.

[0060] Extracting features from fused data;

[0061] A classification model is constructed to classify the features of the fused data, identify wetland distribution areas, and generate wetland species distribution maps.

[0062] A second aspect of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement a wetland monitoring method combining multi-source remote sensing and unmanned aerial vehicles.

[0063] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a wetland monitoring method combining multi-source remote sensing and unmanned aerial vehicles.

[0064] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0065] 1. By integrating satellite remote sensing with UAVs, richer and more accurate data can be obtained, thereby improving the precision of wetland monitoring. Satellite remote sensing can provide large-scale macroscopic information, while UAVs can acquire high-resolution local information. The combination of the two can compensate for the shortcomings of a single technology and achieve complementary advantages.

[0066] 2. Drones are fast and flexible, enabling them to acquire multispectral image data of target areas in a short time. Furthermore, well-planned drone monitoring paths can further improve monitoring efficiency and reduce redundant and ineffective monitoring.

[0067] 3. Preprocessing and segmentation / fusion of satellite remote sensing images can effectively restore missing values, smooth noise, and improve the spatiotemporal continuity of data;

[0068] 4. Extracting spectral and spatial features can provide more comprehensive information for wetland monitoring; spectral features can reflect the species and health status of vegetation, while spatial features can reveal the distribution pattern and structural characteristics of vegetation communities.

[0069] 5. Through data fusion and classification model construction, a high-precision wetland species distribution map can be generated; this distribution map can clearly show the distribution range and boundaries of different species in wetlands, providing strong support for the protection and management of wetland biodiversity. Attached Figure Description

[0070] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort. In the drawings:

[0071] Figure 1 A monitoring flowchart provided for embodiments of the present invention;

[0072] Figure 2 A land cover type distribution map provided for an embodiment of the present invention. Detailed Implementation

[0073] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0074] As one possible implementation method, such as Figure 1 As shown, this embodiment provides a wetland monitoring method combining multi-source remote sensing and unmanned aerial vehicles (UAVs), including the following specific steps: acquiring multi-band remote sensing images of the target detection area based on satellite remote sensing; preprocessing the multi-band remote sensing images of the target detection area to obtain spatiotemporally continuous and noise-suppressed multi-band remote sensing images; segmenting and fusing the spatiotemporally continuous and noise-suppressed multi-band remote sensing images to obtain optimized multi-band remote sensing images; extracting spectral and spatial features based on the band information of the optimized multi-band remote sensing images; and extracting species distribution range and community boundaries based on the spectral and spatial features. The process involves obtaining community distribution data for the target detection area; planning the monitoring path of the UAV based on the community distribution data; using the UAV to monitor the target area and acquiring UAV multispectral image data, which includes multiple bands with wavelength ranges similar to those of the remote sensing image data; performing radiometric and geometric corrections on the UAV multispectral image data to obtain UAV hyperspectral image data; performing band matching between the UAV hyperspectral image data and the remote sensing image data; fusing the UAV hyperspectral image data and the remote sensing image data to identify wetland distribution areas and generate a wetland species distribution map.

[0075] This embodiment achieves efficient and accurate wetland monitoring by combining satellite remote sensing and UAV technology. First, the satellite remote sensing imagery undergoes spatiotemporal continuity and noise suppression preprocessing to improve image quality. Next, segmentation and fusion processing is used to optimize the imagery, accurately extracting species distribution ranges and community boundaries. Based on this data, UAV monitoring paths are planned to improve monitoring efficiency. The high-resolution multispectral imagery acquired by the UAV, after radiometric and geometric correction, is fused with the satellite remote sensing imagery to generate a high-precision wetland species distribution map. This scheme fully leverages the wide coverage of satellite remote sensing and the high resolution of UAVs, overcoming the shortcomings of traditional methods in terms of data quality, information extraction, and monitoring efficiency, providing strong support for wetland protection and management.

[0076] In this embodiment, satellite remote sensing imagery is subject to various noise interferences during acquisition, such as sensor noise and atmospheric noise. These noises degrade image quality and affect subsequent analysis and applications. Preprocessing with spatiotemporal continuity and noise suppression can effectively reduce these noises and improve image quality. Meanwhile, UAV multispectral imagery data is also affected by radiometric and geometric distortions during acquisition. Radiometric and geometric corrections can eliminate these distortions and improve image accuracy. Traditional remote sensing image analysis methods may be limited by image resolution and noise when extracting species distribution ranges and community boundaries. Segmentation and fusion processing and feature extraction can more accurately identify vegetation distribution and community boundaries, improving the accuracy of analysis. When using satellite remote sensing imagery or UAV imagery alone for wetland distribution area identification, insufficient data resolution or spectral information may affect identification accuracy. By fusing UAV hyperspectral imagery data with satellite remote sensing imagery, the advantages of both can be fully utilized, improving the accuracy of wetland distribution area identification. Satellite remote sensing imagery and UAV imagery each have different advantages: satellite imagery has a wide coverage but relatively low resolution; UAV imagery has high resolution but limited coverage. Combining these two methods enables the integration of large-scale and high-resolution monitoring, improving monitoring efficiency. Planning UAV monitoring paths requires accurate community distribution data. Extracting community distribution data from satellite remote sensing imagery provides a scientific basis for UAV monitoring path planning, enhancing the targeting and efficiency of monitoring. Therefore, this embodiment employs preprocessing for spatiotemporal continuity and noise suppression, resulting in multi-band remote sensing images that are consistent in time and space with significantly reduced noise levels, providing a high-quality data foundation for subsequent analysis. Furthermore, after radiometric and geometric correction, the radiometric and geometric accuracy of the UAV multispectral imagery data is significantly improved, more accurately reflecting the spectral and spatial characteristics of ground features. Through segmentation and fusion processing and feature extraction, the distribution range of species and community boundaries can be extracted more precisely, providing more accurate data support for ecological monitoring and protection. By fusing UAV hyperspectral imagery data with satellite remote sensing imagery, wetland distribution areas can be identified more accurately, generating high-precision wetland species distribution maps. Planning UAV monitoring paths based on community distribution data extracted from satellite remote sensing imagery improves the efficiency and targeting of UAV monitoring, reducing unnecessary monitoring areas. By combining satellite remote sensing imagery and UAV imagery through band matching and fusion, a combination of large-scale and high-resolution monitoring has been achieved, improving the overall efficiency of monitoring and data integration capabilities.

[0077] In some possible implementations, multi-band remote sensing imagery of the target detection area is acquired based on satellite remote sensing. Multi-band remote sensing imagery provides rich spectral information, which helps distinguish different types of vegetation and ground features. Preprocessing of the multi-band remote sensing imagery of the target detection area specifically includes:

[0078] Interpolation and resampling are performed on the spatial features of coarse-resolution multi-band remote sensing images.

[0079] The time series data of the coarse-resolution image after interpolation and resampling are matched pixel-level with the cloudless high-resolution multi-band remote sensing image to form a corrected time series.

[0080] And based on similarity-weighted fusion, a corrected time series is formed;

[0081] Missing values ​​in the original multi-band remote sensing imagery are filled in using spatiotemporal modeling to reconstruct a complete time series. A spatiotemporal model is constructed using known spatiotemporal data based on the selected modeling method. The missing values ​​in the original multi-band remote sensing imagery are then estimated using this constructed model. For each missing pixel, its spectral value is calculated using the spatiotemporal information of its surrounding known pixels, thereby reconstructing the complete time series imagery.

[0082] The Weighted Seasonal-Trend Decomposition Filter (WTDF) is a filter used for time series analysis. It decomposes a time series into trend, seasonal, and random components. This decomposition helps in understanding the structure of the time series and can be used for tasks such as prediction and anomaly detection. Therefore, this embodiment uses the WTDF to smooth the time series, outputting spatiotemporally continuous and noise-suppressed multi-band remote sensing images. Based on the characteristics of the time series data and the analysis requirements, the weight allocation principle of the weighted filter is determined. According to the determined weight allocation principle, the weighted filter is designed. The designed weighted filter is applied to the decomposed trend component, seasonal component, and residual. The trend decomposition first decomposes the time series data of each multi-band remote sensing image. The trend decomposition decomposes the time series into trend component, seasonal component, and residual component. The trend component reflects the long-term trend of the time series, the seasonal component represents the periodic variation pattern of the time series, and the residual component is the random fluctuation after removing the trend and seasonality. Through this decomposition, the different variation characteristics in the time series can be clearly separated.

[0083] In some possible implementations, a weighted seasonality-trend decomposition filter is used to smooth the time series. Specifically, this involves: First, using moving averages or other smoothing techniques to extract the trend component of the time series. The moving average can be a simple moving average (SMA) or a weighted moving average (WMA), where weights are used to emphasize data at certain points in time. The trend component is subtracted from the original time series to obtain a detrended time series. Then, the seasonal component is extracted by calculating the average of the detrended time series over each seasonal period. For example, for monthly data, the average for each month can be calculated. The seasonal component is then subtracted from the detrended time series to obtain the stochastic component. The stochastic component is the portion of the time series that cannot be explained by trend and seasonality. When extracting trend and seasonal components, weighting methods can be used to emphasize data at certain points in time. For example, a weighted moving average can be used to extract the trend component, where weights are used to emphasize more recent data points.

[0084] In some possible implementations, the time-series data of the coarse-resolution image after interpolation and resampling are pixel-level matched with a cloud-free high-resolution multi-band remote sensing image to form a corrected time-series. Specifically, this involves: using the cloud-free high-resolution multi-band remote sensing image as the registration reference image, as it has high spatial resolution and clear texture features, providing an accurate spatial reference for the coarse-resolution image; extracting feature points from the reference image and the coarse-resolution image after interpolation and resampling; calculating the geometric transformation parameters of the coarse-resolution image relative to the reference image based on the feature point matching results; then performing a geometric transformation on the coarse-resolution image and resampling it again to spatially align it with the reference image, achieving pixel-level matching. During the pixel-level matching process, for each pixel, its similarity in the coarse-resolution and high-resolution images is calculated. The higher the similarity, the closer the features of the pixel are in the two images, and the greater its corresponding weight. Based on the calculated similarity weights, the pixel values ​​of the coarse-resolution and high-resolution images are weighted and fused. For each pixel, the fused value is a weighted sum of the values ​​of the corresponding pixels in the coarse-resolution image and the high-resolution image, each multiplied by its similarity weight. The fused image combines the advantages of both images, preserving the spatial details of the high-resolution image while utilizing the spectral information of the coarse-resolution image, thus forming a corrected time-series image.

[0085] In some possible implementations, a segmentation and fusion process is performed on a spatiotemporally continuous and noise-suppressed multi-band remote sensing image to obtain an optimized multi-band remote sensing image. The segmentation and fusion process is as follows: A single pixel is randomly selected from the multi-band remote sensing image, and its position (row and column number) and spectral characteristics (reflectance values ​​of each band) are recorded. Within the target area, with the randomly selected starting pixel as the center, pixels in its surrounding neighborhood (usually 4-neighborhood or 8-neighborhood) are searched. The spectral difference between each pixel in the neighborhood and the starting pixel is calculated. For each band, the difference in reflectance between the two pixels is calculated. Then, the differences of all bands are weighted and summed. If the spectral difference between a pixel in the neighborhood and the starting pixel is less than a set threshold, the pixel is classified as a homogeneous region; otherwise, it is classified as a heterogeneous region. Homogeneous regions are merged into a new object, heterogeneous regions are segmented, and homogeneous regions within heterogeneous regions are merged into the new object. This process continues until all heterogeneous regions are segmented, resulting in a segmented multi-band remote sensing image. Finally, the segmented multi-band remote sensing images are integrated to obtain an optimized multi-band remote sensing image. Based on the above segmentation and merging operations, a new image is generated, where the value of each pixel corresponds to its corresponding object number. In this way, the image is divided into multiple object regions with similar spectral characteristics. Through the above segmentation and fusion processing, homogeneous regions in the image are effectively merged, and heterogeneous regions are reasonably segmented, thereby improving the readability and analytical value of the image while retaining important information. This method is particularly suitable for remote sensing applications that require land cover classification, change detection, or landscape analysis.

[0086] Homogeneous regions refer to areas that are similar in spectral characteristics (such as reflectance in different bands) and spatial location. The similarity between pixels can be determined by setting a threshold (such as a spectral difference threshold). If the difference in reflectance between two pixels in different bands is less than this threshold, they are considered to belong to homogeneous regions. Heterogeneous regions refer to areas that differ significantly in spectral characteristics or spatial location, i.e., regions that do not meet the homogeneity criteria.

[0087] In some possible implementations, spectral and spatial features are extracted based on the band information of the optimized multi-band remote sensing image, specifically including:

[0088] Information on all bands is obtained from multi-band remote sensing images, including the center wavelength and band range. For example, a common multi-band image may contain blue light (0.45-0.52 μm), green light (0.52-0.60 μm), red light (0.60-0.70 μm), and near-infrared light (0.70-0.90 μm).

[0089] Extract the spectral reflectance curve of each pixel to obtain the spectral characteristics that reflect the vegetation species. The spectral reflectance curve is a set of reflectance values ​​of each pixel in different bands, which can reflect the spectral characteristics of the vegetation species.

[0090] By acquiring the spatial resolution of multi-band remote sensing imagery and calculating texture features, spatial characteristics reflecting vegetation communities can be obtained. Spatial resolution refers to the actual ground area represented by each pixel in the remote sensing image, which determines the smallest detail of ground features that the image can resolve. Texture features reflect the spatial distribution characteristics of vegetation communities.

[0091] In some possible implementations, based on spectral and spatial features, the species distribution range and community boundaries are extracted to obtain community distribution data of the target detection area, specifically including:

[0092] The vegetation index is calculated based on the band information of the optimized multi-band remote sensing image.

[0093] Based on spectral and spatial features, morphological extraction is used to extract the boundary information of each vegetation community unit, obtaining the species distribution range and community boundary. Specifically, morphological dilation is used to fill small holes within the vegetation area, enhancing the connectivity of the vegetation area; morphological erosion is used to refine the boundary of the vegetation area and remove noise on the boundary; the boundary of the vegetation area is further optimized through a combination of dilation and erosion; and edge detection algorithm is used to extract the boundary of the vegetation area, obtaining the boundary information of the vegetation community unit.

[0094] Based on vegetation index, species distribution range and community boundary, a threshold is set to separate vegetated areas from non-vegetated areas;

[0095] A region growing algorithm is used to divide the vegetation region into multiple community units, obtaining community distribution data for the target detection area. Specifically, multiple seed points are randomly selected within the vegetation region, and growth conditions are set, such as spectral similarity (difference in vegetation indices less than a certain threshold). Starting from a seed point, each pixel in its neighborhood is checked to see if it meets the growth conditions. If it does, the pixel is merged into the current growing region, and this pixel is used as a new seed point to continue growing. This process is repeated until no new pixels can be merged into the current region. The generated regions are then merged and optimized, removing excessively small regions or merging adjacent similar regions. The divided vegetation community units are labeled as different categories, generating a community distribution map.

[0096] In some possible implementations, the monitoring path of the UAV is planned based on the community distribution data of the target detection area, specifically including:

[0097] Based on the community distribution data of the target detection area, the types of vegetation communities that need to be monitored are determined according to the needs of the target detection area, and priorities are assigned to different communities.

[0098] Acquire the drone's performance parameters, including flight speed, flight time, camera resolution, sensor type, and maximum flight altitude. These parameters can be obtained from the drone's technical manual or data provided by the manufacturer. Combined with community priority, divide the monitoring area. Based on the distribution range and priority of vegetation communities, divide the target detection area into multiple monitoring zones. High-priority community areas can be divided into smaller monitoring units to ensure more detailed monitoring; low-priority areas can be divided into larger monitoring units to improve monitoring efficiency.

[0099] Based on the topographic data of the monitored area, forward overlap and lateral overlap are calculated. Forward overlap refers to the proportion of overlapping portions of images on adjacent flight paths relative to a single image, while lateral overlap refers to the proportion of overlapping portions of images on adjacent flight paths relative to a single image.

[0100] Based on the calculated forward and lateral overlap, an initial waypoint distribution is automatically generated. Waypoints should be evenly distributed within the monitoring area to ensure that the distance between adjacent waypoints meets the overlap requirements.

[0101] Based on the generated initial waypoint distribution, execute the UAV flight mission and collect data from the monitored area. Ensure the UAV flies along the predetermined waypoints and captures images or videos at each waypoint. Collect data from the range area based on the initial waypoint distribution and analyze the overlap and stitching quality of the data.

[0102] If the overlap and stitching quality do not meet the set thresholds, adjust the waypoint density.

[0103] The monitoring path for the UAV is generated based on the adjusted waypoint density.

[0104] In some possible implementations, radiometric and geometric corrections are performed on the UAV multispectral image data:

[0105] Radiation correction includes:

[0106] The pixel brightness values ​​of the original image are converted into atmospheric outer surface reflectance data, and the radiometric calibration parameters monitored by the sensor are extracted to obtain radiometric calibration data.

[0107] Using the radiation value of the calibrated tarpaulin as the standard value, the radiation-calibrated data is subjected to reflection correction, and the radiation value data is converted into reflectance data.

[0108] Geometric correction includes:

[0109] Imaging is performed using semiconductor photosensitive elements to obtain spectral information along a single line in space;

[0110] The acquisition of spatial images and spectral data is completed through mechanical push-broom scanning.

[0111] The image data of the entire area acquired according to the monitoring path of the UAV will be stitched together in multiple rows to obtain the UAV hyperspectral image data.

[0112] Radiometric and geometric corrections can significantly improve the quality and usability of UAV multispectral imagery data. Radiometric correction eliminates the influence of sensor and atmospheric conditions, yielding accurate reflectance data; geometric correction corrects geometric distortions in the imagery, ensuring spatial consistency between images. This corrected data can provide reliable support for applications such as vegetation monitoring and land use analysis.

[0113] In some possible implementations, the corrected image data is saved in a standard image format, and a quality check is performed on the corrected image data, including the accuracy of reflectance data and the precision of geometric correction. The quality of the correction results can be evaluated by comparing them with reference data (such as ground-based measured data or high-resolution imagery).

[0114] In some possible implementations, band matching is performed between UAV hyperspectral imagery data and remote sensing imagery, and the UAV hyperspectral imagery data and remote sensing imagery are fused to identify wetland distribution areas and generate wetland species distribution maps, specifically including:

[0115] Based on the band information of multi-band remote sensing images, the center wavelength and band range of the bands are obtained from the metadata of multi-band remote sensing images, and the center wavelength and band range of the bands are obtained from the metadata of UAV hyperspectral images. By comparing the band ranges of hyperspectral images and multi-band remote sensing images, the band range in hyperspectral images that is closest to the bands of remote sensing images is determined, and the band range in UAV hyperspectral image data corresponding to the band information of multi-band remote sensing images is obtained.

[0116] If the spectral resolution of the hyperspectral image is higher than that of the remote sensing image, the hyperspectral data needs to be spectrally resampled to make its spectral resolution consistent with that of the remote sensing image. That is, the spectral reflectance data corresponding to the band of the multi-band remote sensing image is extracted from the hyperspectral data. The extracted spectral reflectance data is consistent with the band information of the remote sensing image in terms of wavelength range and resolution.

[0117] Based on the spectral reflectance information of the extracted hyperspectral data, a vegetation index identical to the remote sensing vegetation index is calculated.

[0118] Obtain vegetation indices calculated from the band reflectance values ​​of multi-band remote sensing images;

[0119] The vegetation index of the calculated UAV imagery is fused with the vegetation index of the remote sensing data.

[0120] Extracting features from fused data;

[0121] By constructing a classification model to classify the features of the fused data, and through steps such as band matching, vegetation index calculation and fusion, feature extraction and classification, UAV hyperspectral image data and remote sensing images can be effectively fused, wetland distribution areas can be identified, and wetland species distribution maps can be generated.

[0122] In some possible implementations, a wetland monitoring method combining multi-source remote sensing and unmanned aerial vehicles (UAVs) can be directly applied to various wetland monitoring projects, whether freshwater wetlands, coastal wetlands, or inland marsh wetlands. As long as there is a need to monitor multiple aspects of the wetland ecosystem (land cover, vegetation, water quality, biodiversity, etc.), the multi-source remote sensing data and smart plot collaborative monitoring and assessment method of this embodiment can be adopted. For example, in the monitoring of large-scale river wetlands, satellite remote sensing data can be used to monitor the distribution and changes of wetlands throughout the entire watershed. UAVs can conduct detailed surveys of key areas such as riverbanks and river islands in the river wetlands. Smart plots are set up in different ecological functional zones of the river wetlands (such as estuary wetlands and floodplain wetlands) to monitor changes in water quality, soil, and biodiversity in real time. Through multi-source data fusion analysis, the health status and dynamic changes of the river wetland ecosystem can be comprehensively understood.

[0123] This technology can be further expanded to areas such as wetland ecological restoration effect assessment and wetland ecological compensation mechanism research. In wetland ecological restoration projects, through long-term multi-source data monitoring and analysis, changes in various indicators of the wetland ecosystem before and after restoration, such as increased vegetation coverage, improved water quality, and restored biodiversity, can be compared to accurately assess the effectiveness of ecological restoration projects and provide a scientific basis for optimizing restoration plans and making funding decisions. In wetland ecological compensation mechanism research, the monitoring method in this embodiment can be used to obtain data related to the ecosystem service value of wetlands (such as water conservation, climate regulation, and biodiversity maintenance), providing data support for rationally determining ecological compensation standards and scope.

[0124] Specifically, this embodiment can be directly applied to wetland water quality ecological monitoring and evaluation. By combining satellite remote sensing with UAV technology, wetland water quality monitoring is not only highly efficient and accurate, but also extends the monitoring scope to the field of water quality ecological monitoring, forming a comprehensive and efficient integrated wetland and water quality ecological monitoring scheme, providing stronger support for wetland protection and management.

[0125] In wetland monitoring, the first step is to preprocess the satellite remote sensing images to ensure spatiotemporal continuity and suppress noise. This process, using advanced algorithms and technologies, effectively improves image quality, making wetland information in the images clearer and more accurate. Subsequently, segmentation and fusion processing is used to optimize the preprocessed images, accurately extracting the distribution range and boundaries of wetland communities. This enables monitoring personnel to accurately locate key areas of the wetlands and conduct targeted monitoring work.

[0126] Based on the accurately extracted data on the distribution range and boundaries of wetland species and water sources, the monitoring path of the UAV is further planned. The UAV can quickly and efficiently monitor the wetland according to the planned path, greatly improving monitoring efficiency. The high-resolution multispectral imagery acquired by the UAV during the monitoring process is fused with satellite remote sensing imagery after radiometric and geometric correction. This fusion fully leverages the advantages of satellite remote sensing's wide coverage and the high resolution of the UAV, ultimately generating a high-precision wetland species and water source distribution map. This map clearly shows the distribution of different species and water sources in the wetland.

[0127] In water quality ecological monitoring, satellite remote sensing can analyze the spectral characteristics of water bodies to make preliminary judgments about water quality conditions, such as identifying the distribution of suspended solids, algae, and other substances, thereby roughly understanding the eutrophication level of the water body. Based on the eutrophication level, further analysis of the water's ecological environment (such as the state of water pollution and the growth of flora and fauna) can be conducted. Drones, carrying specialized water quality monitoring equipment, can penetrate into different areas of wetlands to obtain more precise water quality data, including key water quality indicators such as water temperature, pH, dissolved oxygen, chemical oxygen demand (COD), and ammonia nitrogen. Through the flexible flight and precise positioning of drones, comprehensive and multi-level monitoring of wetland water bodies can be achieved, compensating for the shortcomings of traditional water quality monitoring methods in terms of spatial coverage and monitoring frequency. Furthermore, based on the spectral images of remote sensing, after determining the distribution range and boundaries of water sources, further monitoring of water source distribution data can be conducted, analyzing the distribution data of objects on the water surface, thereby further determining the distribution of species and garbage in the water source.

[0128] Specifically, specific band combinations of multispectral UAV imagery can be used to extract water quality parameters. For example, by analyzing the spectral characteristics of near-infrared and red light bands, chlorophyll content and suspended solids concentration in water bodies can be estimated, thereby accurately assessing the nutrient status and pollution level of wetland water bodies. For small wetland patches or special landform features (such as oxbow lakes and river islands), multispectral UAVs, with their high resolution advantage, can perform detailed boundary extraction and morphological feature analysis, obtaining detailed information such as area, perimeter, and shape index. This microscopic information is of great significance for a deeper understanding of the ecological structure and processes within wetlands. Combining wetland monitoring with water quality ecological monitoring forms a complete integrated monitoring system for wetlands and water quality ecology.

[0129] In practical applications, preprocessing and segmentation / fusion of satellite remote sensing imagery helps identify key wetland areas and species distribution, providing key monitoring targets and regions for water quality ecological monitoring. UAVs then use this information to conduct targeted water quality monitoring, acquiring detailed water quality data. Furthermore, combining water quality monitoring results with wetland species distribution maps allows for a more comprehensive analysis of the wetland ecosystem's health. For example, comparing water quality data and species distribution in different areas reveals the impact of water quality changes on wetland biodiversity, providing a scientific basis for wetland ecological restoration and protection measures. This integrated monitoring scheme fully leverages the advantages of satellite remote sensing's wide coverage and UAVs' high resolution and maneuverability, effectively solving problems such as low data quality, inaccurate information extraction, and low monitoring efficiency in traditional monitoring methods. It not only provides precise data support for wetland protection and management but also enables timely detection of vegetation and water quality conditions within wetland ecosystems, providing strong support for the sustainable development and protection of wetland ecosystems.

[0130] As one possible implementation method, satellite remote sensing data, with its macroscopic and periodic characteristics, enables continuous monitoring of large areas of wetlands. For example, by regularly acquiring multispectral and hyperspectral images from the Beijing-3 series or ZY-1 series satellites, it is possible to clearly grasp the distribution patterns of wetland land cover types on a large spatial scale and their dynamic changes over time. For instance, comparing satellite images from the past few years allows for a direct observation of the transformation between marshland and forest areas within wetlands, providing long-term basic data support for wetland land use planning and ecological protection. Multispectral UAV data, with its high spatial resolution and rich spectral information, enables detailed exploration of key areas within wetlands. In terms of wetland vegetation community structure analysis, multiple specific spectral bands (such as blue, green, red, and near-infrared bands) of multispectral UAV imagery can capture subtle differences in spectral reflectance among different vegetation species. Through object-oriented classification methods, the distribution range and community boundaries of various wetland vegetation types can be accurately identified. For example, reeds and cattails, two common wetland plants, have different reflectivities in the near-infrared band. Multispectral UAV imagery can be used to clearly delineate their growth areas, and then their distribution patterns and ecological functions in wetland ecosystems can be studied in detail.

[0131] In some possible implementations, regarding the visualization of monitoring results, a Geographic Information System (GIS) platform can be used to overlay information such as wetland land cover type distribution, vegetation growth status, water level changes, and ecosystem health assessment results onto a map in the form of thematic layers, forming an intuitive wetland monitoring thematic map. This allows users such as wetland management departments and researchers to clearly understand the overall spatial pattern of the wetland and the distribution characteristics of various ecological elements. For example, by using land cover type layers marked with different colors, the distribution range and interrelationships of different land use types in the wetland can be quickly identified; based on the NDVI rendering effect of the vegetation growth status layer, the distribution of wetland vegetation with varying growth rates and its relationship with the surrounding environment can be intuitively judged; and based on the water level change layer, the fluctuations in wetland water levels in different seasons and years and their impact on the wetland ecosystem can be clearly seen. Meanwhile, various charts such as line graphs, bar charts, and scatter plots are used to display the changing trends of various wetland monitoring indicators over time. For example, the annual change curve of wetland area clearly shows the rate of increase or decrease and time nodes of wetland area; the bar chart of seasonal change of vegetation index reflects the seasonal pattern of vegetation growth; and the scatter plot of water quality parameters over time helps to analyze the relationship between water quality changes and time, season, and external disturbance factors (such as precipitation and surrounding human activities). These charts provide powerful tools for in-depth analysis of the dynamic changes of wetland ecosystems. For key areas or typical wetland landforms, three-dimensional models are constructed using UAV imagery for three-dimensional display. For example, the three-dimensional model of an oxbow lake in a wetland allows for detailed observation of its topographic features, surrounding vegetation distribution details, and water morphology changes. This provides an immersive visualization tool for in-depth research on the microstructure and ecological functions of wetlands, greatly improving the readability and usability of wetland monitoring data and facilitating efficient data analysis, decision-making, and scientific research by relevant personnel.

[0132] As one possible implementation method, this embodiment provides a wetland monitoring system for a national nature reserve, such as... Figure 2 As shown, in a wetland within a national nature reserve, five smart plots were selected based on the wetland's ecological functional zoning and topographic features. These plots, labeled A, B, C, D, and E, were evenly distributed across the core area, buffer zone, and experimental area of ​​the wetland. Comprehensive monitoring equipment was installed in each smart plot for monitoring purposes.

[0133] Based on the monitoring needs and geographical location of this wetland, the ZY1E and BJ3N satellites were selected as the primary satellite remote sensing data sources. The ZY1E satellite's orbital characteristics allow it to cover the target wetland every 16 days. Its multispectral spatial resolution is 10 meters, and its panchromatic resolution reaches 2.5 meters. It also possesses rich spectral information, covering visible and near-infrared bands, making it suitable for monitoring various wetland characteristics. The ZY1E satellite has a revisit period of 2 days and a multispectral spatial resolution of 10 meters, providing high accuracy for vegetation monitoring. Image data from these two satellites over the past 5 years will be acquired quarterly to ensure long-term dynamic monitoring of the wetland.

[0134] After acquiring satellite remote sensing image data, preprocessing was performed using the professional remote sensing image processing software ENVI. First, radiometric calibration was performed. Based on the radiometric calibration parameter files provided by the ZY1E and BJ3N satellites, the raw DN values ​​of the images were converted into radiance or reflectance values, thereby eliminating the influence of sensor-related factors and atmospheric scattering and absorption on data accuracy. In the geometric correction stage, ground control points evenly distributed within the wetland area were collected (the coordinates of these control points were obtained through high-precision GPS measurements). A polynomial fitting method was used to correct the images to a unified CGCS2000 geographic coordinate system, ensuring that the geometric accuracy of the images met the requirements of subsequent analysis. Finally, the images were cropped based on the wetland boundary vector data to extract the image of the target wetland area, resulting in preprocessed satellite remote sensing image data, ready for subsequent wetland information extraction.

[0135] Based on the macroscopic wetland characteristics and smart plot distribution reflected in the ZY1E satellite remote sensing data, a detailed UAV flight plan was formulated. The Zongheng CW-15 UAV platform was selected, as it possesses excellent stability and payload capacity, and can carry various high-precision remote sensing devices. The onboard cameras included a 6-channel multispectral resolution 0.08-meter Changguang Yuchen MS600 optical camera to acquire multispectral image data, and a Ruibo DG6M full-frame five-lens oblique photography camera to acquire five-lens RGB images of the wetland for creating a realistic 3D model and digital orthophoto.

[0136] Data collection using drones was conducted in spring and autumn when meteorological conditions were suitable (wind speed less than 5 m / s, no precipitation, and cloud cover less than 10%). Flight routes were planned based on the location of the smart sample plots and key monitoring areas within the wetlands identified by satellite remote sensing (such as areas with complex changes in wetland vegetation communities and areas with dynamic changes in water boundaries) to ensure comprehensive coverage of these key areas. The flight altitude was set between 100-150 meters according to different monitoring needs to obtain image data with sufficient resolution. During the flight, the drone's flight control system automatically recorded key parameters such as flight trajectory, flight altitude, shooting angle, and shooting time.

[0137] After the UAV data acquisition was completed, PhotoScan and Reconstruction Master software were used for preprocessing. The software first dehazed and denoised the large amount of raw UAV imagery, performed radiometric correction, and then conducted aerial triangulation. Utilizing the UAV's built-in high-precision positioning information and a small number of ground control points selected on the imagery (these control points can share some with the control points used in satellite remote sensing geometric correction to ensure coordinate system consistency), the imagery was corrected to the same CGCS2000 geographic coordinate system as the satellite remote sensing imagery. Finally, radiometric correction was performed, converting the imagery's grayscale values ​​into actual radiance or reflectance values ​​based on the camera's radiometric calibration parameters. This resulted in preprocessed UAV imagery data that could be fused and analyzed with satellite remote sensing data and smart plot data.

[0138] Wetland land cover type classification was performed using the spectral characteristics of preprocessed ZY1E satellite remote sensing imagery. Spectral feature analysis was conducted on a large number of sample areas with known land cover types. For example, wetland water bodies have low reflectance in the near-infrared band, while vegetation has high reflectance in the near-infrared band. By setting threshold conditions for reflectance in different bands, different land cover types such as wetland water bodies, marsh vegetation, mudflats, woodland, and grassland were gradually distinguished. After classification, the land cover type classification results for wetlands were obtained, and the area of ​​each type and its changes over time were statistically analyzed.

[0139] The Normalized Difference Vegetation Index (NDVI) is calculated to monitor wetland vegetation growth. By analyzing the changes in NDVI values ​​from satellite remote sensing images at different times, NDVI time-series curves are plotted to intuitively understand the seasonal, interannual, and long-term growth trends of wetland vegetation. For example, if the NDVI value continuously increases over a certain period, it indicates that the vegetation in the area is growing well and the coverage is increasing; conversely, it may indicate vegetation degradation.

[0140] This study utilizes the texture and shape features of high-resolution UAV imagery to analyze the structure of wetland vegetation communities using an object-oriented classification method. First, the imagery is segmented at multiple scales. Based on the actual distribution of the wetland vegetation communities, appropriate segmentation scale parameters are selected to divide the imagery into objects of different sizes and shapes. Then, based on the objects' spectral characteristics (such as color differences between different vegetation types in multispectral imagery), texture features (such as the texture patterns formed by the arrangement of plant leaves), and shape features (such as the outline shape of the vegetation community), combined with vegetation species information obtained from field surveys, the distribution range and community boundaries of different vegetation species are identified. For example, the reed community, common in wetlands, exhibits specific textures and relatively high vegetation height in the imagery. These features allow for accurate differentiation from other vegetation and precise mapping of its distribution boundaries.

[0141] For small wetland patches or unique landforms (such as oxbow lakes and river islands), fine-grained boundary extraction is performed using UAV imagery. Edge detection algorithms (such as the Canny edge detection algorithm) are employed to identify the boundary contours of these unique landforms by analyzing the grayscale variations of image pixels, and then calculating detailed information such as their area, perimeter, and shape index. This microscopic information is of great significance for in-depth research into the ecological structure and function of wetlands.

[0142] This study correlates soil moisture data collected from smart plots with vegetation growth data from satellite remote sensing and UAV imagery. For example, in smart plot A, long-term monitoring revealed that when soil moisture was between 25% and 35%, satellite remote sensing and UAV imagery showed the most vigorous wetland vegetation growth, with high vegetation cover and NDVI values. Further analysis showed that when soil moisture content was too high (above 40%) or too low (below 20%), vegetation growth was inhibited, manifested as a decrease in NDVI values ​​and a reduction in vegetation cover. This fusion analysis allows for a deeper exploration of the mechanisms by which soil moisture affects wetland vegetation growth, providing a scientific basis for wetland water resource management and vegetation protection.

[0143] By combining water quality data from smart sample plots with satellite remote sensing monitoring of water area changes and UAV imagery of water morphology, the pollution diffusion pathways and ecological self-purification capacity of wetland water bodies are analyzed. For example, smart sample plot C is located near the estuary of a major river in the wetland. Water quality monitoring data from this plot show that chemical oxygen demand (COD) and ammonia nitrogen concentrations increase during the rainy season. Simultaneously, satellite remote sensing monitoring revealed a significant increase in the water area of ​​the region, while UAV imagery clearly showed the direction of water flow and the diffusion pattern at the river estuary. Through comprehensive analysis of these data, it can be determined that the main diffusion pathway of pollutants in the wetland is from the estuary into the wetland interior along with the water flow. Furthermore, based on the differences in vegetation growth and water self-purification capacity in different areas, the degradation and purification effects of the wetland on pollutants can be assessed.

[0144] Based on biodiversity monitoring data from smart sample plots and vegetation community distribution information from satellite remote sensing and UAV imagery, the stability and health status of wetland ecosystems are assessed. For example, biodiversity monitoring data in the area of ​​smart sample plot E shows a decrease in the types and numbers of birds and insects in recent years. Analysis combining satellite remote sensing and UAV imagery reveals changes in the vegetation community structure surrounding the area, with some wetland vegetation being converted into farmland, leading to habitat loss and disruption of ecological connectivity. This multi-source data fusion analysis allows for a comprehensive assessment of the health status of wetland ecosystems, timely identification of ecological problems, and the implementation of corresponding protection measures.

[0145] A wetland ecosystem health assessment index system was established, which comprehensively considers wetland ecological information extracted from multi-source data. This includes land cover diversity index, vegetation cover and health index (combining satellite remote sensing and UAV imagery vegetation information, and comprehensively assessing indicators such as vegetation cover area ratio and mean NDVI), comprehensive water quality index (based on smart plot water quality data, using the comprehensive water pollution index method to integrate multiple water quality parameters into a dimensionless index), and biodiversity index.

[0146] The weights of each indicator were determined using the analytic hierarchy process (AHP). Through expert scoring and the construction of a judgment matrix, the relative importance weight of each indicator in the evaluation system was calculated. For example, the weights were determined as follows: land cover diversity index 0.25, vegetation cover and health index 0.3, comprehensive water quality index 0.2, and biodiversity index 0.25. Then, the fuzzy comprehensive evaluation method was used to quantitatively assess the health status of the wetland ecosystem. First, an evaluation set was determined, divided into five levels: healthy, sub-healthy, slightly degraded, moderately degraded, and severely degraded. Based on the actual monitoring values ​​and weights of each indicator, the membership degree of the wetland ecosystem to each evaluation level was calculated using fuzzy transformation, ultimately determining the health level of the wetland ecosystem. For example, the calculated health status assessment result for this wetland ecosystem was sub-healthy, with relatively low vegetation cover and health index and biodiversity index, indicating certain pressure on wetland vegetation and biodiversity, requiring strengthened protection and management.

[0147] Using the ArcGIS geographic information system platform, information such as land cover type distribution, vegetation growth status, water level changes, and ecosystem health assessment results of wetlands are overlaid on a map as thematic layers to create a wetland monitoring thematic map. On the map, different land cover types are represented by different colors, such as blue for wetland water bodies, green for marsh vegetation, and brown for woodland, visually displaying the overall land use pattern of the wetland. Vegetation growth status is displayed through hierarchical rendering using NDVI values; darker colors indicate higher NDVI values ​​and more vigorous vegetation growth. Water level changes are displayed by overlaying contour lines of water levels from different periods, clearly showing the dynamic changes in wetland water levels. Ecosystem health assessment results are marked on the map with different symbols or colors, facilitating quick identification of the health status of different areas of the wetland.

[0148] By creating various charts and graphs such as line graphs, bar charts, and scatter plots, the changing trends of various wetland monitoring indicators over time are displayed. For example, a line graph showing the annual change in wetland area visually reflects the increase or decrease in wetland area over the past few years; a bar chart showing the seasonal change in vegetation index displays the changes in the growth vitality of wetland vegetation in different seasons; and a scatter plot of water quality parameters over time analyzes the fluctuation patterns of water quality indicators and their relationship with factors such as season and precipitation. These charts and graphs help to deeply analyze the dynamic changes in wetland ecosystems and provide data support for wetland management decisions.

[0149] For key areas or typical wetland landforms, 3D models are constructed using drone imagery for three-dimensional visualization. For example, for a large oxbow lake area in the wetland, five-lens drone imagery data is used to construct a 3D model of the oxbow lake using 3D modeling software (Reconstruct Master). This allows for intuitive observation of the oxbow lake's topographic features, surrounding vegetation distribution, and water morphology, providing a powerful tool for in-depth research on the ecological functions and evolution of the oxbow lake. Simultaneously, monitoring data from smart sample plots is displayed on the monitoring platform in the form of real-time data charts. For instance, the smart sample plot monitoring interface displays real-time changes in meteorological data (temperature, humidity, precipitation, etc.) using dynamic line graphs, and real-time monitoring values ​​of soil moisture and water quality parameters using bar charts. This allows wetland managers to easily view the dynamic changes in in-situ wetland monitoring information, promptly identify anomalies, and take appropriate measures.

[0150] The above specific implementation cases demonstrate in detail the practical application process of wetland monitoring methods based on multi-source remote sensing data and smart plots. From data acquisition, processing, and analysis to the final result evaluation and visualization, each link is closely connected, fully demonstrating the effectiveness and practicality of the method of this invention, and providing comprehensive and accurate technical support for wetland monitoring and protection.

[0151] As one possible implementation, this embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements a wetland monitoring method combining multi-source remote sensing and unmanned aerial vehicles.

[0152] As one possible implementation, this embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements a wetland monitoring method combining multi-source remote sensing and unmanned aerial vehicles (UAVs).

[0153] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment 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 within the scope of protection of the present invention.

Claims

1. A method for monitoring wetlands using multi-source remote sensing combined with unmanned aerial vehicles, characterized by, The method comprises the following specific steps: acquiring multi-band remote sensing images of a target detection area based on satellite remote sensing; preprocessing the multi-band remote sensing images of the target detection area to obtain multi-band remote sensing images that are continuous in time and space and have noise suppressed; performing segmentation fusion processing on the multi-band remote sensing images that are continuous in time and space and have noise suppressed to obtain optimized multi-band remote sensing images; extracting spectral features and spatial features according to band information of the optimized multi-band remote sensing images; extracting species distribution ranges and community boundaries based on the spectral features and the spatial features to obtain community distribution data of the target detection area; planning a monitoring path of a UAV based on the community distribution data of the target detection area; wherein planning the monitoring path of the UAV based on the community distribution data of the target detection area specifically comprises: determining vegetation community types that need to be monitored in priority based on the community distribution data of the target detection area and the requirements of the target detection area, and assigning priorities to different communities; acquiring UAV performance parameters, combining the community priorities, and dividing the monitoring area; calculating heading overlap and lateral overlap based on terrain data of the monitoring area; automatically generating an initial flight point distribution based on the overlap requirements; collecting data of a range area according to the initial flight point distribution and analyzing overlap and splicing quality of the data; if the overlap and the splicing quality do not meet a set threshold, adjusting the flight point density; and generating the monitoring path of the UAV according to the adjusted flight point density; monitoring the target area using a UAV to acquire UAV multi-spectral image data, wherein the UAV multi-spectral image data contains multiple bands similar to the wavelength range of the remote sensing image data; performing radiation correction and geometric correction on the UAV multi-spectral image data to obtain UAV hyperspectral image data; performing band matching on the UAV hyperspectral image data and the remote sensing image, fusing the UAV hyperspectral image data and the remote sensing image, identifying a wetland distribution area, and generating a wetland species distribution map.

2. The method of claim 1, wherein, The preprocessing of the multi-band remote sensing images of the target detection area specifically comprises: performing interpolation operation and resampling processing on spatial features of the coarse-resolution multi-band remote sensing images; performing pixel-level matching on time series data of the coarse-resolution images after the interpolation operation and the resampling processing and cloud-free high-resolution multi-band remote sensing images to obtain corrected time series; performing time-space modeling filling on missing values in the original multi-band remote sensing images to restore complete time series; performing smoothing processing on the time series using a weighted seasonal-trend decomposition filter to output the multi-band remote sensing images that are continuous in time and space and have noise suppressed.

3. The method for monitoring wetlands using multi-source remote sensing integrated with UAV as claimed in claim 1, wherein, The segmentation fusion processing on the multi-band remote sensing images that are continuous in time and space and have noise suppressed specifically comprises: randomly selecting a single pixel in the multi-band remote sensing images; segmenting homogeneous regions and heterogeneous regions in the target area based on the pixel; merging the homogeneous regions into a new object; segmenting the heterogeneous regions and merging homogeneous regions in the heterogeneous regions into the new object until the segmentation of the heterogeneous regions is completed to obtain segmented multi-band remote sensing images; The segmented multi-band remote sensing image is integrated to obtain an optimized multi-band remote sensing image.

4. The method for monitoring wetlands using multi-source remote sensing integrated with UAV as claimed in claim 1, wherein, According to the band information of the optimized multi-band remote sensing image, spectral features and spatial features are extracted, specifically including: A spectral reflectance curve of each pixel is extracted to obtain spectral features reflecting plant species; The spatial resolution of the multi-band remote sensing image is obtained, and texture features are calculated to obtain spatial features reflecting plant communities. 5.The method of claim 1, wherein, Based on the spectral features and the spatial features, the distribution range of the species and the community boundary are extracted to obtain community distribution data of the target detection area, specifically including: According to the band information of the optimized multi-band remote sensing image, a vegetation index is calculated; Based on the spectral features and the spatial features, the boundary information of each plant community unit is extracted using morphological methods to obtain the distribution range of the species and the community boundary; According to the vegetation index, the distribution range of the species and the community boundary, a threshold is set to separate the vegetation area from the non-vegetation area; The vegetation area is divided into multiple community units using a region growing algorithm to obtain the community distribution data of the target detection area.

6. The method for monitoring wetlands using multi-source remote sensing integrated with UAV as claimed in claim 1, wherein, The unmanned aerial vehicle multi-spectral image data is subjected to radiation correction and geometric correction. The radiation correction includes: The pixel brightness value data of the original image is converted into atmospheric outer surface reflectivity data, and the radiation calibration parameters monitored by the sensor are extracted to obtain radiation calibration data; The radiation value data of the calibrated tarpaulin is taken as a standard value, and the radiation calibration data is subjected to reflection correction to convert the radiation value data into reflectivity data. The geometric correction includes: A semiconductor photosensitive element is used for imaging to obtain spectral information on a spatial line; Mechanical push scanning is used to complete the collection of the entire spatial image and spectral data; The image data of the entire region obtained according to the monitoring path of the unmanned aerial vehicle is subjected to multi-row band splicing to obtain unmanned aerial vehicle hyperspectral image data.

7. The method for monitoring wetlands using multi-source remote sensing integrated with UAV as claimed in claim 1, wherein, The unmanned aerial vehicle hyperspectral image data is subjected to band matching with the remote sensing image, the unmanned aerial vehicle hyperspectral image data is fused with the remote sensing image, the wetland distribution area is identified, and a wetland species distribution map is generated, specifically including: Based on the band information of the multi-band remote sensing image, the corresponding band range of the hyperspectral data is determined; Spectral reflectivity data corresponding to the bands of the multi-band remote sensing image is extracted from the hyperspectral data, and the extracted spectral reflectivity data is consistent with the band information of the remote sensing image in terms of wavelength range and resolution; According to the extracted spectral reflectivity information of the hyperspectral data, the same vegetation index as the remote sensing vegetation index is calculated; The vegetation index calculated based on the band reflectivity value of the multi-band remote sensing image is obtained; The calculated vegetation index of the unmanned aerial vehicle image is fused with the vegetation index of the remote sensing data; Based on the fused data, fused data features are extracted; A classification model is constructed to classify the fused data features, identify the wetland distribution area, and generate a wetland species distribution map.

8. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to implement the wetland monitoring method of multi-source remote sensing combined with an unmanned aerial vehicle according to any one of claims 1 to 7.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the wetland monitoring method of multi-source remote sensing combined with an unmanned aerial vehicle according to any one of claims 1 to 7.

Images