Mangrove forest and tidal flat dynamic monitoring method integrating tide level control and multi-source characteristics

Abstract

The invention discloses a mangrove forest and tidal flat dynamic monitoring method fusing tide level control and multi-source features, and relates to the technical field of image processing, and the method comprises the steps: defining a mangrove forest suitable growth region in a research region; generating an annual low-tide-level synthetic image according to the available image of the suitable growth area of the mangrove forest; according to the original tide level data, tide level matching images are screened from all the available images; features of mangrove forest, tidal flat and spartina alterniflora are extracted from the annual low-tide-level synthetic image and the tide level matching image respectively; performing training and hyper-parameter adjustment and optimization on the random forest model by utilizing the characteristics of the mangrove forest, the tidal flat and the spartina alterniflora; using the trained random forest model to classify the research area to obtain an initial distribution diagram of mangrove forest, tidal flat and spartina alterniflora; and post-processing the initial distribution diagram to obtain an annual distribution diagram of the mangrove forest, the tidal flat and the spartina alterniflora. The accuracy, reliability and efficiency of dynamic monitoring of the coastal wetland ecosystem can be remarkably improved.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a method for dynamic monitoring of mangroves and tidal flats that integrates tidal level control and multi-source features. Background Technology

[0002] Coastal wetlands, as sensitive zones where land and sea interact, are ecosystems of global importance. Mangroves, as the ecological cornerstone of these areas, play an irreplaceable role in mitigating storm surges, stabilizing shorelines, purifying water, and providing crucial habitats for numerous marine organisms. Crucially, the vast intertidal mudflats provide indispensable space and growth substrate for mangrove seed dispersal, establishment, and seaward expansion; their width, elevation, and stability directly constrain the distribution pattern and succession process of mangroves. Therefore, comprehensive monitoring of mangrove ecosystems requires precise identification and dynamic tracking of tidal flats as a key element. However, in recent decades, under the dual pressures of human activities (such as land reclamation and coastal engineering) and natural changes, the global mangrove area has been declining. Simultaneously, invasive plants such as Spartina alterniflora, with their strong adaptability and reproductive capabilities, have rapidly expanded on coastal tidal flats, competing with native mangroves for living space and resources, seriously threatening coastal biodiversity and ecological security. Against this backdrop, the simultaneous and accurate identification of mangroves, Spartina alterniflora, and tidal flats, and the clarification of the spatiotemporal evolution relationships among the three, have become urgent needs for carrying out ecological restoration, assessing ecosystem health, and implementing scientific management.

[0003] Remote sensing technology, with its advantages of macroscopic, rapid, and periodic observation, has become a core means of large-scale, long-term ecological environment monitoring. However, in the complex coastal wetland environment, the use of remote sensing technology for accurate land cover classification and dynamic monitoring has long faced several key technical bottlenecks. The most prominent is the problem of spectral confusion: mangroves, Spartina alterniflora, and exposed tidal flats exhibit very similar spectral response characteristics in the visible and near-infrared bands, exhibiting typical phenomena of "same species, different spectra" and "same spectrum, different species," making effective differentiation difficult using only a few spectral bands. Secondly, there is the severe impact of tidal interference: the periodic rise and fall of tides causes tidal flats and vegetation to be periodically submerged and exposed by seawater, resulting in constant and drastic dynamic changes in their visible range, water content, and spectral characteristics in remote sensing images. This interference not only affects the classification accuracy of a single time phase but also makes image results acquired at different times (such as tidal flat area in different years) incomparable due to differences in tide levels at the time of imaging, producing a large number of "false changes" that greatly mislead the analysis of the true evolutionary patterns. In addition, seasonal phenological changes in vegetation and noise from cloud cover further increase the uncertainty and complexity of monitoring. Summary of the Invention

[0004] In view of this, embodiments of this application provide a method and related equipment for dynamic monitoring of mangroves and tidal flats that integrates tidal level control and multi-source characteristics, so as to accurately monitor mangroves and tidal flats.

[0005] One aspect of this application provides a method for dynamic monitoring of mangroves and tidal flats that integrates tidal level control and multi-source characteristics. The method includes the following steps:

[0006] Delineate suitable mangrove growth areas within the study area;

[0007] Generate annual low tide composite images based on available images of the mangrove suitable growth area;

[0008] Tide-matching images are obtained by filtering all available images based on the original tide data;

[0009] Features of mangroves, tidal flats, and Spartina alterniflora were extracted from the annual low tide composite image and the tide matching image, respectively.

[0010] The features of the mangroves, tidal flats, and Spartina alterniflora were used to train and optimize the hyperparameters of the random forest model.

[0011] The trained random forest model was used to classify the study area to obtain the initial distribution maps of the mangroves, tidal flats and Spartina alterniflora;

[0012] The initial distribution map is post-processed to obtain the annual distribution maps of the mangroves, tidal flats, and Spartina alterniflora.

[0013] In some embodiments, defining a suitable mangrove growth area within the study area includes the following steps:

[0014] Based on the digital elevation model and shoreline data, the suitable growth area of ​​the mangroves is defined in the study area; wherein the definition rules include: extending a preset distance from the shoreline boundary; and areas where the superimposed elevation is lower than a preset height and the slope is less than a preset angle.

[0015] In some embodiments, generating composite low tide images for different years based on available images of the suitable mangrove growth area includes the following steps:

[0016] For the target year, the cloud cover percentage within the suitable mangrove growth area was recalculated and all available images were selected.

[0017] Calculate the improved normalized water index of the available imagery;

[0018] The available images with the lowest 50% of the improved normalized water index were selected as candidate images.

[0019] The candidate images are then combined using median values ​​to generate the annual low tide composite image.

[0020] In some embodiments, the step of filtering the available images based on the original tide data to obtain tide-matching images includes the following steps:

[0021] The original tidal data is processed into a continuous sequence at the minute level using an interpolation method;

[0022] Based on the transit timestamp of the available images, the corresponding instantaneous tide level is queried and extracted from the minute-level continuous sequence; the target tide level is obtained through iterative search; the maximum value of the lowest instantaneous tide level in each annual low tide composite image is taken as the starting point, and the iteration is performed upwards by step size; wherein, the goal of the iterative search is that in each year, the difference between the instantaneous tide level of at least one available image and the target tide level is less than a preset threshold.

[0023] For each year, the available images from all available images that have a difference between the instantaneous tide level and the target tide level within a preset tolerance range are selected as the tide level matching images.

[0024] In some embodiments, the method further includes the following steps:

[0025] If the number of available images that meet the preset tolerance range in any year exceeds the preset number, then the standard deviation of the difference between the instantaneous tide level and the target tide level of the available images that meet the preset tolerance range is calculated.

[0026] The available image that minimizes the overall standard deviation for all years is selected as the tide-matching image.

[0027] In some embodiments, the extraction of features of mangroves, tidal flats, and Spartina alterniflora from the annual low tide composite image and the tide-matched image respectively includes the following steps:

[0028] The spectral bands, spectral indices, and texture features of mangroves, tidal flats, and Spartina alterniflora were extracted from the annual low tide composite image and the tide matching image, respectively.

[0029] The spectral bands include blue light band, green light band, red light band, near-infrared band, short-wave infrared 1 band and short-wave infrared 2 band;

[0030] The spectral indices include the normalized vegetation index, the land surface water index, the modified normalized differential water body index, the enhanced vegetation index, the normalized water body index, the near-infrared vegetation index, the shadowless automatic water body extraction index, and the shadowed automatic water body extraction index.

[0031] The texture features include contrast, correlation, energy, and homogeneity calculated based on the gray-level co-occurrence matrix.

[0032] In some embodiments, the post-processing of the initial distribution map to obtain the annual distribution maps of mangroves, tidal flats, and Spartina alterniflora includes the following steps:

[0033] Mathematical morphology methods were used to remove noise from small patches in the initial distribution map, and the mangrove boundaries in the initial distribution map were corrected based on the reference image to obtain the annual distribution maps of mangroves, tidal flats and Spartina alterniflora.

[0034] Another aspect of this application embodiment provides a dynamic monitoring device for mangroves and tidal flats that integrates tidal level control and multi-source characteristics, the device comprising:

[0035] Regional delineation units are used to define suitable areas for mangrove growth within the study area;

[0036] An image synthesis unit is used to generate an annual low tide composite image based on available images of the suitable mangrove growth area.

[0037] The image matching unit is used to filter out tide-matching images from all available images based on the original tide data;

[0038] The feature extraction unit is used to extract features of mangroves, tidal flats and Spartina alterniflora from the annual low tide composite image and the tide matching image, respectively.

[0039] The model training unit is used to train the random forest model and tune its hyperparameters using the features of the mangroves, tidal flats and Spartina alterniflora;

[0040] The initial distribution determination unit is used to classify the study area using the trained random forest model to obtain the initial distribution map of the mangroves, tidal flats and Spartina alterniflora;

[0041] The annual distribution determination unit is used to post-process the initial distribution map to obtain the annual distribution maps of the mangroves, tidal flats and Spartina alterniflora.

[0042] Another aspect of this application embodiment provides an electronic device, including a processor and a memory;

[0043] The memory is used to store programs;

[0044] The processor executes the program to implement any of the methods described above.

[0045] Another aspect of this application provides a computer-readable storage medium storing a program that is executed by a processor to implement the method described in any of the above embodiments.

[0046] This application includes at least the following beneficial effects:

[0047] This application defines suitable mangrove growth areas within the study area; generates annual low-tide composite images based on available images of suitable mangrove growth areas; selects tidal level matching images from all available images based on raw tidal data; extracts features of mangroves, tidal flats, and Spartina alterniflora from the annual low-tide composite images and tidal level matching images, respectively; trains and hyperparameter-tunes a random forest model using the features of mangroves, tidal flats, and Spartina alterniflora; classifies the study area using the trained random forest model to obtain initial distribution maps of mangroves, tidal flats, and Spartina alterniflora; and post-processes the initial distribution maps to obtain annual distribution maps of mangroves, tidal flats, and Spartina alterniflora. This application fundamentally solves the problem of distorted dynamic analysis of tidal flats caused by inconsistent tides by performing hierarchical preprocessing and selection of available images, combined with rich feature extraction from the trained random forest model. It also achieves high-precision, automated differentiation of mangroves, tidal flats, and Spartina alterniflora, thus forming a complete technical solution from data preprocessing to high-precision boundary identification, significantly improving the accuracy, reliability, and efficiency of dynamic monitoring of coastal wetland ecosystems. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 A flowchart illustrating the dynamic monitoring method for mangroves and tidal flats that integrates tidal level control and multi-source characteristics, provided in an embodiment of this application;

[0050] Figure 2 An example flowchart of a dynamic monitoring method for mangroves and tidal flats that integrates tidal level control and multi-source features, provided in an embodiment of this application;

[0051] Figure 3 The structural block diagram of the dynamic monitoring device for mangroves and tidal flats that integrates tidal level control and multi-source characteristics provided in the embodiments of this application is shown. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0053] Before providing a detailed description of the embodiments of this application, some related technologies involved in the embodiments of this application will be described first, as follows:

[0054] To address these challenges, existing technologies have developed various solutions, one of which is a "machine learning classification method based on multi-temporal image synthesis and multi-feature fusion." This approach typically involves two core steps: The first step is multi-temporal image synthesis. To suppress noise from clouds, shadows, and transient tides, this technique collects all available optical satellite imagery (such as the Landsat series) within a monitoring year (e.g., one year). After removing cloud-covered images using cloud detection algorithms, all remaining qualified images are combined using median composition to generate a composite image representing the "normal" state of that year. This method effectively suppresses transient noise and provides a more stable classification basis. The second step is multi-feature fusion and classification. Based on the generated annual composite image, this method not only uses the original spectral bands but also introduces multiple vegetation indices (such as NDVI), water indices (such as MNDWI), and texture features based on the gray-level co-occurrence matrix (GLCM) to construct a multi-dimensional feature space. Subsequently, advanced ensemble learning algorithms such as Random Forest are used for classification. Random forest models have been widely proven to perform well in classifying complex scenarios due to their efficient handling of high-dimensional features, insensitivity to outliers, and natural resistance to overfitting. This "median synthesis + random forest" approach is currently a high-performing and widely used advanced scheme in coastal wetland remote sensing classification, constituting the closest existing technology to be improved in this embodiment. However, this scheme still has inherent limitations in addressing the consistency of tidal levels in cross-period dynamic monitoring, making it difficult to support precise studies requiring highly comparable analysis of tidal flat evolution and its interaction with mangroves. Therefore, effectively eliminating the impact of tidal fluctuations on remote sensing monitoring and achieving accurate and comparable dynamic extraction of tidal flat and vegetation information from different periods has become a pressing technical challenge in this field.

[0055] Despite advancements in existing technologies, the following key limitations remain in complex coastal wetland environments:

[0056] a. Tidal interference cannot be completely eliminated: Although existing multi-temporal composite methods (such as annual median composite) can average the tidal influence, they cannot guarantee the consistency of tidal height in composite images across different years. This leads to "pseudo-changes" in tidal flat area due to tidal level differences between years, making the results of cross-year tidal flat dynamic analysis unreliable and seriously affecting the scientific nature of evolutionary pattern research.

[0057] b. Feature utilization is not optimized: Many methods fail to systematically combine spectral, index and texture features, or feature selection is not targeted, resulting in limited ability to distinguish easily confused land cover such as mangroves and Spartina alterniflora.

[0058] c. Insufficient automation and intelligence in the process: From image selection and preprocessing to classification and post-processing, the existing solutions rely on human experience and judgment in many aspects. There is a lack of a complete, automated monitoring workflow that is tailored to the special environment of the intertidal zone, resulting in low work efficiency and poor repeatability of results.

[0059] To address the shortcomings of existing technologies in high-precision dynamic monitoring, this embodiment aims to provide a method for dynamic monitoring of mangroves and tidal flats that integrates tidal level control and multi-source features. This method aims to construct a high-dimensional feature space integrating spectral, index, and texture data through an innovative, hierarchical image preprocessing and selection workflow. Combined with a high-performance random forest model, it fundamentally solves the distortion problem in tidal flat dynamic analysis caused by tidal inconsistencies and achieves high-precision, automated differentiation of mangroves, Spartina alterniflora, and tidal flats. This forms a complete technical solution from data preprocessing to high-precision boundary identification, significantly improving the accuracy, reliability, and efficiency of dynamic monitoring of coastal wetland ecosystems.

[0060] Reference Figure 1 This application provides a method for dynamic monitoring of mangroves and tidal flats that integrates tidal level control and multi-source characteristics, specifically including the following steps S100~S160:

[0061] S100: Define suitable areas for mangrove growth within the study area;

[0062] S110: Generate an annual low tide composite image based on available images of the suitable mangrove growth area;

[0063] S120: Based on the original tide level data, obtain tide level matching images from all available images;

[0064] S130: Extract features of mangroves, tidal flats, and Spartina alterniflora from the annual low tide composite image and the tide matching image, respectively;

[0065] S140: Use the features of the mangroves, tidal flats and Spartina alterniflora to train and optimize the hyperparameters of the random forest model;

[0066] S150: The trained random forest model is used to classify the study area to obtain the initial distribution map of the mangroves, tidal flats and Spartina alterniflora;

[0067] S160: Post-process the initial distribution map to obtain the annual distribution maps of the mangroves, tidal flats and Spartina alterniflora.

[0068] The following section will provide a detailed introduction and explanation of the solutions in the embodiments of this application, using specific application examples.

[0069] like Figure 2 As shown in the figure, this embodiment provides a method for dynamic monitoring of mangroves and tidal flats that integrates tidal level control and multi-source characteristics. The specific steps are as follows:

[0070] S1: Delineation of suitable mangrove growth areas in the study area.

[0071] Based on the Digital Elevation Model (DEM) and shoreline data, precise suitable growth zones for mangroves are automatically delineated. Specific rules are: areas extending 2 kilometers from the shoreline boundary; areas with an overlay elevation of less than 20 meters and a slope of less than 15 degrees. This effectively focuses on the core areas where mangroves are likely to distribute and expand, eliminating interference from irrelevant landforms.

[0072] S2: Screen and synthesize high-quality annual low-tide images to generate annual baseline images that are least affected by tides and clouds for mangrove / Spartina alterniflora species classification.

[0073] S21: Initial screening based on cloud cover: For the target year, the cloud cover percentage in the S1 area is recalculated and initial screening is performed to obtain all available images. While ensuring image quality, this effectively expands the candidate pool of available images and provides a richer data foundation for subsequent low tide screening.

[0074] S22: Initial screening based on tide level: Calculate the modified normalized water index (MNDWI). The lower the MNDWI value, the less water cover and the lower the tide level. We select the 50% of images with the lowest MNDWI values ​​as the candidate set to initially ensure that the selected images are in a low tide state.

[0075] S23: Image Compositing: The candidate images selected in S22 are combined using median compositing to generate a high-quality composite image of the annual low tide level. This image effectively suppresses noise from clouds, shadows, and seasonal vegetation phenological changes, forming stable classification baseline data.

[0076] S3: Filter comparable tidal flat images across years based on precise matching of tide levels.

[0077] The core innovation of using raw tide data to filter images that are precisely matched to the tide level is to meet the comparability problem of cross-year tidal flats.

[0078] S31: Tide Level Data Correlation: The downloaded raw tide level data is processed into a minute-level continuous sequence using interpolation methods and uploaded to the Google Earth Engine (GEE) cloud platform. In GEE, based on the precise transit timestamps of each remote sensing image initially selected in S21, the corresponding instantaneous tide level values ​​are queried and extracted from the minute-level tide level sequence, thereby achieving precise correlation between tide level data and remote sensing images.

[0079] S32: Iterative Search for the Optimal Target Tide Level: First, calculate the lowest tide level for all available images in each year. If the lowest tide level of a particular year deviates significantly from the normal range of other years due to data anomalies (such as being affected by typhoons, sensor malfunctions, etc.), the data for that abnormal year is removed and not included in subsequent calculations. The highest value of the "annual lowest tide level" among all years is used as the initial value for the iterative search. Set the upper limit of the tide level search (1.5 meters in this embodiment) and the search step size (0.1 meters in this embodiment). Through iterative calculation, find an optimal "target tide level" among all years, such that in each year, at least one image has an instantaneous tide level value close to the target tide level.

[0080] S33: Construction of a Comparable Image Set Across Years: For each year, select images from all images whose tide levels differ from the optimal target tide level determined in S32 within a preset tolerance range (e.g., ±0.3 meters). If more than the preset number of images meet the criteria for a given year, further calculate the standard deviation of the difference between the tide levels of these images and the target tide level, and select the image combination that minimizes the overall standard deviation for all years to ensure a high degree of consistency in tide levels across year-on-year images.

[0081] S4: Construction of high-dimensional feature space and random forest classification.

[0082] S41: Feature Extraction: From the images generated in S2 and S3, three types of features are extracted to form an 18-dimensional high-dimensional feature space (including 6 spectral bands, 8 spectral indices, and 4 texture features), as shown in Table 1 below:

[0083] Table 1

[0084]

[0085] The characteristics are explained below:

[0086] The spectral bands include Blue, Green, Red, Near Infrared (NIR), Shortwave Infrared 1 (SWIR1), and Shortwave Infrared 2 (SWIR2).

[0087] Spectral indices include the Normalized Difference Vegetation Index (NDVI), Land Surface Water Index (LSWI), Modified Normalized Difference Water Index (mNDWI), Enhanced Vegetation Index (EVI), Normalized Difference Water Index (NDWI), Near-Infrared Reflectance of Vegetation (NIRv), Automated Water Extraction Index - no shadows (AWEInsh), and Automated Water Extraction Index - shadows (AWEIsh).

[0088] Texture features include contrast, correlation, energy, and homogeneity, which are calculated based on the gray-level co-occurrence matrix.

[0089] S42: Model Training and Classification: The collected sample data was randomly divided into training and test sets at a ratio of 80% and 20%, respectively. The random forest model was trained and its hyperparameters were tuned using the training set, and then its accuracy was verified on the test set. Finally, the trained model was used to classify the entire study area, obtaining initial distribution maps of mangroves, Spartina alterniflora, and tidal flats.

[0090] S43: Post-processing optimization: The initial distribution map is optimized, including using mathematical morphology methods (such as opening and closing operations) to remove noise from small patches, and performing necessary fine corrections on the mangrove boundaries based on high-resolution reference images to obtain the final annual distribution map.

[0091] In summary, the embodiments of this application include the following key technical solutions:

[0092] a) Layered image preprocessing and classification process: The overall monitoring process is divided into two independent branches with different purposes: "generating high-quality annual low tide images" (S2) and "generating cross-year comparable tidal flat images with accurate tide level matching" (S3), which are finally integrated in the land cover classification stage (S4).

[0093] b) Precise tide matching methods for cross-year comparable analysis: including the "optimal target tide iterative search algorithm" (S32) and the "image selection method with the goal of minimizing the standard deviation of tide levels in cross-year candidate images" (S33).

[0094] c) Multidimensional feature combination for high-precision land cover classification: The specific combination of the feature space used for protection, consisting of six spectral bands, eight spectral indices and four texture features.

[0095] Beneficial effects:

[0096] More scientific: The original "precise matching of tide levels" workflow fundamentally solves the industry problem of "pseudo-changes" in tidal flat area caused by inconsistent tide levels, making the conclusions of cross-year dynamic analysis true and reliable.

[0097] Higher classification accuracy: Through a three-level preprocessing process of "buffer zone definition -> low tide screening -> median synthesis", combined with a high-dimensional feature space of "spectral + index + texture" and a random forest model, the ability to distinguish easily confused land cover types such as mangroves and Spartina alterniflora is greatly improved.

[0098] High degree of automation and systematization: It provides a complete and automated solution from data preprocessing and feature engineering to intelligent classification, which greatly reduces human intervention, ensures the consistency and repeatability of monitoring results, and improves work efficiency.

[0099] Reference Figure 3 This application provides a dynamic monitoring device for mangroves and tidal flats that integrates tidal level control and multi-source characteristics, including:

[0100] Regional delineation units are used to define suitable areas for mangrove growth within the study area;

[0101] An image synthesis unit is used to generate an annual low tide composite image based on available images of the suitable mangrove growth area.

[0102] The image matching unit is used to filter out tide-matching images from all available images based on the original tide data;

[0103] The feature extraction unit is used to extract features of mangroves, tidal flats and Spartina alterniflora from the annual low tide composite image and the tide matching image, respectively.

[0104] The model training unit is used to train the random forest model and tune its hyperparameters using the features of the mangroves, tidal flats and Spartina alterniflora;

[0105] The initial distribution determination unit is used to classify the study area using the trained random forest model to obtain the initial distribution map of the mangroves, tidal flats and Spartina alterniflora;

[0106] The annual distribution determination unit is used to post-process the initial distribution map to obtain the annual distribution maps of the mangroves, tidal flats and Spartina alterniflora.

[0107] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0108] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this application are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is changed and sub-operations described as part of a larger operation are executed independently.

[0109] Furthermore, although this application is described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the described functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding this application. Rather, given the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed herein, the actual implementation of the module will be understood within the scope of conventional technology for an engineer. Therefore, those skilled in the art can implement the application set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and not intended to limit the scope of this application, which is determined by the full scope of the appended claims and their equivalents.

[0110] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0111] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0112] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0113] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0114] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0115] Although embodiments of this application have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the claims and their equivalents.

[0116] The above is a detailed description of the preferred embodiments of this application, but this application is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of this application, and these equivalent modifications or substitutions are all included within the scope defined by the claims of this application.

Claims

1. A method for dynamic monitoring of mangroves and tidal flats integrating tidal level control and multi-source characteristics, characterized in that, The method includes the following steps: Delineate suitable mangrove growth areas within the study area; Generate annual low tide composite images based on available images of the mangrove suitable growth area; Tide-matching images are obtained by filtering all available images based on the original tide data; Features of mangroves, tidal flats, and Spartina alterniflora were extracted from the annual low tide composite image and the tide matching image, respectively. The features of the mangroves, tidal flats, and Spartina alterniflora were used to train and optimize the hyperparameters of the random forest model. The trained random forest model was used to classify the study area to obtain the initial distribution maps of the mangroves, tidal flats and Spartina alterniflora; The initial distribution map is post-processed to obtain the annual distribution maps of the mangroves, tidal flats, and Spartina alterniflora.

2. The method for dynamic monitoring of mangroves and tidal flats integrating tidal level control and multi-source characteristics according to claim 1, characterized in that, Defining suitable mangrove growth areas within the study area includes the following steps: Based on the digital elevation model and shoreline data, the suitable growth area of ​​the mangroves is defined in the study area; wherein the definition rules include: extending a preset distance from the shoreline boundary; and areas where the superimposed elevation is lower than a preset height and the slope is less than a preset angle.

3. The method for dynamic monitoring of mangroves and tidal flats integrating tidal level control and multi-source characteristics according to claim 1, characterized in that, The process of generating composite low-tide images for different years based on available images of the suitable mangrove growth area includes the following steps: For the target year, the cloud cover percentage within the suitable mangrove growth area was recalculated and all available images were selected. Calculate the improved normalized water index of the available imagery; The available images with the lowest 50% of the improved normalized water index were selected as candidate images. The candidate images are then combined using median values ​​to generate the annual low tide composite image.

4. The method for dynamic monitoring of mangroves and tidal flats integrating tidal level control and multi-source characteristics according to claim 1, characterized in that, The step of filtering out tide-matching images from all available images based on the original tide data includes the following steps: The original tidal data is processed into a continuous sequence at the minute level using an interpolation method; Based on the transit timestamp of the available images, the corresponding instantaneous tide level is queried and extracted from the minute-level continuous sequence; the target tide level is obtained through iterative search; the maximum value of the lowest instantaneous tide level in each annual low tide composite image is taken as the starting point, and the iteration is performed upwards by step size; wherein, the goal of the iterative search is that in each year, the difference between the instantaneous tide level of at least one available image and the target tide level is less than a preset threshold. For each year, the available images from all available images that have a difference between the instantaneous tide level and the target tide level within a preset tolerance range are selected as the tide level matching images.

5. The method for dynamic monitoring of mangroves and tidal flats integrating tidal level control and multi-source characteristics according to claim 4, characterized in that, The method further includes the following steps: If the number of available images that meet the preset tolerance range in any year exceeds the preset number, then the standard deviation of the difference between the instantaneous tide level and the target tide level of the available images that meet the preset tolerance range is calculated. The available image that minimizes the overall standard deviation for all years is selected as the tide-matching image.

6. The method for dynamic monitoring of mangroves and tidal flats integrating tidal level control and multi-source characteristics according to claim 1, characterized in that, The extraction of features of mangroves, tidal flats, and Spartina alterniflora from the annual low tide composite image and the tide matching image includes the following steps: The spectral bands, spectral indices, and texture features of mangroves, tidal flats, and Spartina alterniflora were extracted from the annual low tide composite image and the tide matching image, respectively. The spectral bands include blue light band, green light band, red light band, near-infrared band, short-wave infrared 1 band and short-wave infrared 2 band; The spectral indices include the normalized vegetation index, the land surface water index, the modified normalized differential water body index, the enhanced vegetation index, the normalized water body index, the near-infrared vegetation index, the shadowless automatic water body extraction index, and the shadowed automatic water body extraction index. The texture features include contrast, correlation, energy, and homogeneity calculated based on the gray-level co-occurrence matrix.

7. The method for dynamic monitoring of mangroves and tidal flats integrating tidal level control and multi-source characteristics according to any one of claims 1 to 6, characterized in that, The post-processing of the initial distribution map to obtain the annual distribution maps of mangroves, tidal flats, and Spartina alterniflora includes the following steps: Mathematical morphology methods were used to remove noise from small patches in the initial distribution map, and the mangrove boundaries in the initial distribution map were corrected based on the reference image to obtain the annual distribution maps of mangroves, tidal flats and Spartina alterniflora.

8. A dynamic monitoring device for mangroves and tidal flats that integrates tidal level control and multi-source characteristics, characterized in that: The device includes: Regional delineation units are used to define suitable areas for mangrove growth within the study area; An image synthesis unit is used to generate an annual low tide composite image based on available images of the suitable mangrove growth area. The image matching unit is used to filter out tide-matching images from all available images based on the original tide data; The feature extraction unit is used to extract features of mangroves, tidal flats and Spartina alterniflora from the annual low tide composite image and the tide matching image, respectively. The model training unit is used to train the random forest model and tune its hyperparameters using the features of the mangroves, tidal flats and Spartina alterniflora; The initial distribution determination unit is used to classify the study area using the trained random forest model to obtain the initial distribution map of the mangroves, tidal flats and Spartina alterniflora; The annual distribution determination unit is used to post-process the initial distribution map to obtain the annual distribution maps of the mangroves, tidal flats and Spartina alterniflora.

9. An electronic device, characterized in that, The electronic device includes a processor and a memory; The memory is used to store programs; The processor executes the program to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The storage medium stores a program that is executed by a processor to implement the method as described in any one of claims 1 to 7.

Images