Method and device for identifying mangrove ecological anomaly, computer device and storage medium

Abstract

The present application relates to the technical field of mangrove habitat analysis, and particularly relates to a mangrove ecological anomaly identification method, multi-dimensional feature data constructed based on mangrove correlation multi-source data, feature reconstruction is performed by using a time series convolution auto-encoder network to represent the dynamic evolution baseline of mangroves under normal phenology and tidal cycle, reconstruction error is obtained by combining multi-dimensional feature data and feature reconstruction data, the reconstruction error data reflecting the degree of deviation of the current grid unit multi-dimensional feature data from the natural and healthy phenology evolution baseline of mangroves, and the ecological anomaly occurrence change point identification and ecological anomaly grade division are performed according to the reconstruction error data and the change point detection model, thereby improving the accuracy and efficiency of mangrove ecological anomaly identification.

Description

Technical Field

[0001] This invention relates to the field of mangrove habitat analysis technology, and in particular to a method, apparatus, computer equipment, and storage medium for identifying mangrove ecological anomalies. Background Technology

[0002] Mangroves are important coastal ecosystems, and accurate assessment of their ecological health is crucial for ecological protection and restoration. Currently, mangrove ecological assessment methods based on UAV remote sensing technology mostly use single-section (single time node) observation data, extracting indicators such as vegetation indices and landscape patterns, and employing entropy weighting or analytic hierarchy process (AHP) methods for static weighted assessment.

[0003] However, mangroves grow in a complex intertidal marine environment, which is greatly affected by tides and seasonal phenological changes. Single-section assessments can easily misjudge normal phenological leaf fall or brief tidal submersion as ecological degradation, failing to distinguish between natural succession and ecological degradation, making it difficult to accurately identify mangrove ecological anomalies and lacking early warning capabilities. Summary of the Invention

[0004] Based on this, the purpose of this invention is to provide a method, apparatus, computer device, and storage medium for identifying mangrove ecological anomalies. The method involves constructing multidimensional feature data based on multi-source data associated with mangroves, reconstructing features using a time-series convolutional autoencoder network to characterize the dynamic evolution baseline of mangroves under normal phenological and tidal cycles, and combining the multidimensional feature data and the reconstructed feature data to calculate the reconstruction error. This yields reconstruction error data reflecting the degree to which the multidimensional feature data of the current grid cell deviates from the baseline of the natural healthy phenological evolution of the mangroves. Based on the reconstruction error data and a change point detection model, the method identifies ecological anomaly change points and classifies ecological anomaly levels, thereby improving the accuracy and efficiency of mangrove ecological anomaly identification.

[0005] In a first aspect, embodiments of this application provide a method for identifying ecological anomalies in mangrove forests, comprising the following steps:

[0006] Obtain multi-source data of mangroves from several grid cells in the target area at various times within a preset time observation period; Multidimensional feature extraction is performed on the mangrove-related multi-source data of each grid unit at each time period to obtain multidimensional feature data of each grid unit at each time period. Feature reconstruction is performed based on the multidimensional feature data of each grid unit at each time period and the pre-trained time series convolutional autoencoder network to obtain the feature reconstruction data of each grid unit at each time period. Reconstruction error is calculated based on the multidimensional feature data and feature reconstruction data of each grid cell at each time period to obtain the reconstruction error data of each grid cell at each time period. Based on the reconstruction error data of each grid cell at each time period and the preset change point detection model, the change points of ecological anomalies are identified and the ecological anomaly levels are classified, so as to obtain the mangrove ecological anomaly identification results of each grid cell.

[0007] Secondly, embodiments of this application provide a mangrove ecological anomaly identification device, comprising: The data acquisition module is used to obtain multi-source data of mangroves collected at various times within a preset time observation period for several grid units in the target area. The feature extraction module is used to extract multidimensional features from the mangrove-related multi-source data of each grid cell at each time period, and obtain multidimensional feature data of each grid cell at each time period. The feature reconstruction module is used to reconstruct features based on the multidimensional feature data of each grid unit at each time period and the pre-trained time series convolutional autoencoder network, so as to obtain the feature reconstruction data of each grid unit at each time period. The reconstruction error calculation module is used to calculate the reconstruction error based on the multidimensional feature data and feature reconstruction data of each grid cell at each time period, and to obtain the reconstruction error data of each grid cell at each time period. The anomaly identification module is used to identify ecological anomaly occurrence points and classify ecological anomaly levels based on the reconstruction error data of each grid cell at each period and the preset change point detection model, so as to obtain the mangrove ecological anomaly identification results of each grid cell.

[0008] Thirdly, embodiments of this application provide a computer device, including: a processor, a memory, and a computer program stored in the memory and executable on the processor; when the computer program is executed by the processor, it implements the steps of the mangrove ecological anomaly identification method as described in the first aspect.

[0009] Fourthly, embodiments of this application provide a storage medium storing a computer program, which, when executed by a processor, implements the steps of the mangrove ecological anomaly identification method as described in the first aspect.

[0010] This application provides a method, apparatus, computer device, and storage medium for identifying mangrove ecological anomalies. Based on multi-dimensional feature data constructed from multi-source data associated with mangroves, a time-series convolutional autoencoder network is used for feature reconstruction to characterize the dynamic evolution baseline of mangroves under normal phenological and tidal cycles. The reconstruction error is calculated by combining the multi-dimensional feature data and the reconstructed feature data to obtain reconstruction error data reflecting the degree to which the multi-dimensional feature data of the current grid cell deviates from the baseline of the natural healthy phenological evolution of the mangrove. Based on the reconstruction error data and a change point detection model, ecological anomaly change points are identified, and ecological anomaly levels are classified, improving the accuracy and efficiency of mangrove ecological anomaly identification.

[0011] To better understand and implement this invention, the following detailed description is provided in conjunction with the accompanying drawings. Attached Figure Description

[0012] Figure 1 This is a flowchart illustrating a method for identifying mangrove ecological anomalies according to an embodiment of this application. Figure 2 This is a flowchart illustrating step S2 of the mangrove ecological anomaly identification method provided in one embodiment of this application; Figure 3 This is a flowchart illustrating step S3 of a mangrove ecological anomaly identification method provided in one embodiment of this application; Figure 4 This is a flowchart illustrating step S4 of a mangrove ecological anomaly identification method provided in one embodiment of this application. Figure 5 This is a flowchart illustrating step S5 of a mangrove ecological anomaly identification method provided in one embodiment of this application. Figure 6 A flowchart illustrating a method for identifying mangrove ecological anomalies provided in another embodiment of this application; Figure 7 This is a schematic diagram of the structure of a mangrove ecological anomaly identification device provided in one embodiment of this application; Figure 8 This is a schematic diagram of the structure of a computer device provided in one embodiment of this application. Detailed Implementation

[0013] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0014] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0015] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0016] Please see Figure 1 , Figure 1 This is a flowchart illustrating a method for identifying mangrove ecological anomalies according to an embodiment of this application. The method includes the following steps: S1: Obtain multi-source data of mangrove forests at various times collected from several grid cells in the target area within a preset time observation period.

[0017] The execution subject of the mangrove ecological anomaly identification method is the identification device of the mangrove ecological anomaly identification method (hereinafter referred to as the identification device). In an optional embodiment, the identification device can be a computer device, a server, or a server cluster composed of multiple computer devices.

[0018] In this embodiment, the identification device obtains mangrove-related multi-source data collected at various times within a preset time observation period for several grid units of the target area. The mangrove-related multi-source data includes multispectral data, water body data, and mangrove patch data.

[0019] S2: Multidimensional feature extraction is performed on the mangrove-related multi-source data of each grid cell at each time period to obtain multidimensional feature data of each grid cell at each time period.

[0020] In this embodiment, the identification device performs multi-dimensional feature extraction based on the mangrove-related multi-source data of each grid unit at each period to obtain multi-dimensional feature data of each grid unit at each period. The multi-dimensional feature data includes pest and disease stress dimension feature data, vegetation function dimension feature data, disturbance evaluation dimension feature data, environmental quality dimension feature data, and landscape pattern dimension feature data.

[0021] Please see Figure 2 , Figure 2 The flowchart of step S2 in the mangrove ecological anomaly identification method provided in one embodiment of this application is as follows: S21: Based on the near-infrared band reflectance, red-edge band reflectance and the preset red-edge normalized vegetation index calculation algorithm in the multispectral data, the red-edge normalized vegetation index is obtained as the feature data of the pest and disease stress dimension.

[0022] In this embodiment, the identification device obtains the red-edge normalized vegetation index (RBVI) ​​based on the near-infrared reflectance, red-edge reflectance, and a preset red-edge normalized vegetation index calculation algorithm in the multispectral data. This RBVI serves as the characteristic data for the pest and disease stress dimension, used to highly sensitively characterize abnormal decline in canopy chlorophyll content and early stress. The red-edge normalized vegetation index calculation algorithm is as follows:

[0023] In the formula, The red-edged normalized vegetation index, For near-infrared reflectivity, This refers to the reflectivity of the red-edge band.

[0024] S22: Based on the near-infrared band reflectance, infrared band reflectance in the multispectral data and the preset normalized vegetation index calculation algorithm, the normalized vegetation index is obtained as the feature data of the vegetation functional dimension.

[0025] In this embodiment, the identification device obtains the Normalized Difference Vegetation Index (NDV) based on the near-infrared reflectance, infrared reflectance, and a preset normalized vegetation index calculation algorithm in the multispectral data. This NV serves as the vegetation functional dimension feature data, used to characterize the growth density of the mangrove canopy structure. The NV calculation algorithm is as follows:

[0026] In the formula, Normalized Difference Vegetation Index (NDVI) This refers to the reflectivity in the infrared band.

[0027] S23: Perform multispectral classification based on the multispectral data to obtain the total area of ​​invasive alien species; obtain the invasive alien species ratio based on the total area of ​​invasive alien species, the total area of ​​grid cells, and a preset invasive alien species ratio calculation algorithm, and use it as the feature data of the interference evaluation dimension.

[0028] In this embodiment, the identification device performs multispectral classification based on the multispectral data to obtain the total area of ​​invasive alien species. The identification device then uses the total area of ​​invasive alien species, the total area of ​​the raster cells, and a preset algorithm for calculating the invasive alien species ratio to obtain the invasive alien species ratio. This ratio serves as the feature data for the interference evaluation dimension, characterizing the extent to which invasive alien species such as Spartina alterniflora encroach upon the native mangrove habitat. The algorithm for calculating the invasive alien species ratio is as follows:

[0029] In the formula, The proportion of invasive alien species, This represents the total area affected by invasive alien species. This represents the total area of ​​the grid cells.

[0030] S24: Based on the area of ​​spectrally abnormal water bodies, the total area of ​​the water network, and the preset water body apparent anomaly index calculation algorithm in the water body data, obtain the water body apparent anomaly index as the feature data of the environmental quality dimension.

[0031] In this embodiment, the identification device obtains a water body apparent anomaly index based on the area of ​​spectrally anomalous water bodies, the total area of ​​the water network, and a preset water body apparent anomaly index calculation algorithm in the water body data. This index serves as the environmental quality dimension feature data, used to quantify the apparent anomaly degree of pollution or eutrophication in the tidal channels and water networks within the mangrove forest. The water body apparent anomaly index calculation algorithm is as follows:

[0032] In the formula, This refers to the apparent anomaly index of water bodies. The area of ​​water bodies with spectral anomalies. This represents the total area of ​​the water network.

[0033] S25: Based on the total number of mangrove patches, the total area of ​​grid cells, and the preset mangrove patch density calculation algorithm in the mangrove patch data, the mangrove patch density is obtained as the landscape pattern dimension feature data.

[0034] In this embodiment, the identification device obtains the mangrove patch density based on the total number of mangrove patches, the total area of ​​grid cells, and a preset mangrove patch density calculation algorithm in the mangrove patch data. This density is used as the landscape pattern dimension feature data to measure the degree of spatial fragmentation of mangrove habitats after natural or human-induced division. The mangrove patch density calculation algorithm is as follows:

[0035] In the formula, For mangrove patch density, This represents the total number of mangrove patches.

[0036] S3: Based on the multidimensional feature data of each grid unit at each time period and the pre-trained time series convolutional autoencoder network, feature reconstruction is performed to obtain the feature reconstruction data of each grid unit at each time period.

[0037] The time-series convolutional autoencoder network is trained using historical data from periods without significant degradation records. The training set includes multidimensional feature data of several sample grid cells for each historical period. The network parameters are optimized by minimizing the mean square error (MSE) as the objective function. This network represents the baseline of dynamic evolution of mangroves under normal phenology and tidal cycles.

[0038] In this embodiment, the identification device reconstructs features based on the multidimensional feature data of each grid cell at each time period and the pre-trained time-series convolutional autoencoder network, learns the dynamic baseline of multidimensional indicators of mangroves, and obtains the feature reconstruction data of each grid cell at each time period. This effectively separates the background noise caused by normal leaf fall or temporary flooding, and improves the accuracy of ecological assessment.

[0039] The time-series convolutional autoencoder network includes one-dimensional convolutional layers and transposed convolutional layers; please refer to... Figure 3 , Figure 3 The flowchart of step S3 in the mangrove ecological anomaly identification method provided in one embodiment of this application is as follows: S31: Stack the multidimensional feature data of each grid cell at each time period to construct the temporal feature tensor of each grid cell.

[0040] In this embodiment, the recognition device stacks the multidimensional feature data of each grid cell at each time period in a time sequence to construct the temporal feature tensor of each grid cell.

[0041] S32: Based on the temporal feature tensors of each grid cell at each time period and the one-dimensional convolutional layer, perform temporal feature extraction and dimensionality reduction to obtain the latent feature tensors of each grid cell at each time period.

[0042] In this embodiment, the recognition device performs temporal feature extraction and dimensionality reduction based on the temporal feature tensors of each grid unit at each time period and the one-dimensional convolutional layer to obtain the latent feature tensors of each grid unit at each time period, wherein the latent feature tensors are:

[0043] In the formula, For the latent feature tensor, It is a non-linear activation function. Here is the encoder weight matrix. For time-series feature tensors, This is the first bias term.

[0044] S33: Based on the latent feature tensors of each grid cell at each time period and the transposed convolutional layer, feature reconstruction is performed to obtain the feature reconstruction tensor of each grid cell.

[0045] In this embodiment, the recognition device performs feature reconstruction based on the latent feature tensors of each grid cell at each time period and the transposed convolutional layer to obtain the feature reconstruction tensor of each grid cell. The feature reconstruction tensor includes feature reconstruction data from each time period, and the feature reconstruction tensor is:

[0046] In the formula, Reconstruct the tensor for features. This is the decoder weight matrix. This is the second bias term.

[0047] S4: Calculate the reconstruction error based on the multidimensional feature data and feature reconstruction data of each grid cell at each time period to obtain the reconstruction error data of each grid cell at each time period.

[0048] In this embodiment, the identification device calculates the reconstruction error based on the multidimensional feature data and feature reconstruction data of each grid cell at each period, and obtains the reconstruction error data of each grid cell at each period. The reconstruction error data reflects the degree to which the multidimensional feature data of the current grid cell deviates from the baseline of the natural healthy phenological evolution of the mangrove. If the reconstruction error data is close to zero, it indicates that the mangrove is in normal tidal inundation or seasonal leaf fall. If the reconstruction error data is large, it indicates that the evolution trajectory of the mangrove has deviated from the normal baseline and may be due to early disease stress or environmental pollution.

[0049] Please see Figure 4 , Figure 4 The flowchart of step S4 in the mangrove ecological anomaly identification method provided in one embodiment of this application is as follows: S41: Based on the multidimensional feature data, feature reconstruction data, and preset reconstruction error calculation algorithm of each grid cell at each time period, obtain the reconstruction error data of each grid cell at each time period.

[0050] In this embodiment, the recognition device obtains the reconstruction error data of each grid cell at each time period based on the multidimensional feature data, feature reconstruction data, and a preset reconstruction error calculation algorithm of each grid cell at each time period. The reconstruction error calculation algorithm is as follows:

[0051] In the formula, Let be the reconstruction error data for the t-th period. For the multidimensional feature data of the t-th period, Reconstruct the data for the features of the t-th period.

[0052] S5: Based on the reconstruction error data of each grid cell at each period and the preset change point detection model, identify the change points of ecological anomalies and classify the ecological anomaly levels to obtain the mangrove ecological anomaly identification results for each grid cell.

[0053] In this embodiment, the identification device identifies ecological anomaly occurrence change points and classifies ecological anomaly levels based on the reconstruction error data of each grid unit at each period and the preset change point detection model, thereby obtaining the mangrove ecological anomaly identification results for each grid unit. It can identify the time node and spatial location of anomaly occurrence in the early stage of reconstruction error change, reducing the lag of traditional post-event assessment and realizing early location and classification of ecological anomalies.

[0054] Please see Figure 5 , Figure 5 The flowchart of step S5 in the mangrove ecological anomaly identification method provided in one embodiment of this application includes steps S51 to S52, as follows: S51: Based on the reconstruction error data of each grid cell at each time period and the preset joint posterior probability calculation algorithm for the occurrence of change points, obtain the joint posterior probability of the occurrence of change points of each grid cell at each time period.

[0055] In this embodiment, the identification device obtains the joint posterior probability of change point occurrence for each grid cell at each time period based on the reconstruction error data of each grid cell at each time period and a preset joint posterior probability calculation algorithm for change point occurrence. The joint posterior probability calculation algorithm for change point occurrence is as follows:

[0056] In the formula, Let be the joint posterior probability of the change point occurring at time t. Let be the length of the journey in period t, representing the time span since the last ecological anomaly occurred. This represents the cumulative value of the reconstruction error data up to the t-th period. Let be the length of the variable point run in the (t-1)th period. To predict the likelihood probability, let be the length of the previous run. Under the given conditions, the probability density of the reconstructed error data observed in the t-th period is: The prior transition probability for the run length is determined by a predefined conditional risk function and is used to characterize the prior probability distribution of ecological anomaly change points. Let be the normalization constant for the total probability, and let be the numerator term for all possible... The boundary likelihood value is obtained by summing.

[0057] By using reconstructed error data as the observation input data for the change point detection model, the originally complex, high-dimensional multidimensional features can be effectively reduced in dimensionality and transformed into a scalar time series that intuitively represents the "health deviation". This process abandons the traditional approach of relying on experience to set static index thresholds, allowing the change point detection model to focus on analyzing the dynamic accumulation process of deviation. This achieves the technical effect of highly sensitive capture and adaptive dynamic detection of subtle ecological anomalies under complex intertidal habitats and natural background noise interference.

[0058] S52: Determine the occurrence period of ecological anomalies based on the joint posterior probability of change points in each period of each grid cell and the preset probability threshold, and obtain several occurrence periods of ecological anomalies in each grid cell; classify the ecological anomaly level based on the reconstruction error data of each occurrence period of ecological anomalies in each grid cell, and obtain the ecological anomaly level data of each occurrence period of ecological anomalies in each grid cell.

[0059] In this embodiment, the identification device determines the occurrence period of ecological anomalies based on the joint posterior probability of change points in each period of each grid cell and a preset probability threshold, thereby obtaining several occurrence periods of ecological anomalies in each grid cell. Specifically, if the joint posterior probability of change points is greater than the probability threshold, the identification device takes the corresponding period as the occurrence period of ecological anomalies.

[0060] The identification device classifies ecological anomalies into different levels based on the reconstruction error data of each ecological anomaly occurrence period in each grid cell, thereby obtaining ecological anomaly level data for each ecological anomaly occurrence period in each grid cell. Specifically, based on the reconstruction error data of each ecological anomaly occurrence period in each grid cell, the identification device can use the natural breakpoint method or a preset threshold to classify each grid cell under each ecological anomaly occurrence period into different ecological anomaly levels, wherein the ecological anomaly levels include mild stress, moderate anomaly, and severe degradation.

[0061] Please see Figure 6 , Figure 6 The flowchart of another embodiment of the mangrove ecological anomaly identification method provided in this application also includes steps S6-S7, as follows: S6: Obtain the location coordinate data of each grid cell, the standard orthophotos of each ecological anomaly occurrence period in each grid cell, and the canopy height model.

[0062] In this embodiment, the identification device obtains the location coordinate data of each grid cell, the standard orthophotos of each ecological anomaly occurrence period of each grid cell, and the canopy height model.

[0063] S7: Based on the location coordinate data of each grid cell, the standard orthophotos of each ecological anomaly occurrence period of each grid cell, the canopy height model, and the ecological anomaly level data, construct a spatial distribution map of mangrove ecological anomalies to obtain the spatial distribution map of mangrove ecological anomalies of each ecological anomaly occurrence period in the target area.

[0064] In this embodiment, the identification device constructs a spatial distribution map of mangrove ecological anomalies based on the location coordinate data of each grid cell, the standard orthophotos of each ecological anomaly occurrence change point of each grid cell, the canopy height model, and the ecological anomaly level data. This obtains a spatial distribution map of mangrove ecological anomalies in the target area for each ecological anomaly occurrence change point. The spatial distribution map of mangrove ecological anomalies is used to visually display the actual geographical coverage, aggregation center, and dynamic evolution trend of ecological anomalies within the mangrove community (including the distribution areas of major dominant tree species, tidal creeks, and forest gaps, as well as complex micro-topographic areas) under the current ecological anomaly occurrence change point.

[0065] Specifically, the identification device inputs the location coordinate data of each grid cell, the standard orthophotos of each ecological anomaly occurrence period of each grid cell, the canopy height model, and the ecological anomaly level data into a preset geographic information system (GIS). The data is then mapped to the standard spatial coordinate system of the target mangrove area for spatial overlap and attribute association. Spatial kernel density analysis, spatial clustering, and vectorized boundary extraction are performed on the grid cells of each ecological anomaly occurrence period to generate a spatial distribution map of mangrove ecological anomalies for each ecological anomaly occurrence period.

[0066] Please refer to Figure 7 , Figure 7 This is a schematic diagram of the structure of a mangrove ecological anomaly identification device provided in one embodiment of this application. The device can be implemented entirely or partially through software, hardware, or a combination of both. The device 7 includes: Data acquisition module 71 is used to obtain multi-source data of mangroves collected at various times within a preset time observation period for several grid units in the target area; The feature extraction module 72 is used to extract multi-dimensional features based on the mangrove-related multi-source data of each grid unit at each time period, and obtain multi-dimensional feature data of each grid unit at each time period. The feature reconstruction module 73 is used to reconstruct features based on the multidimensional feature data of each grid unit at each time period and the pre-trained time series convolutional autoencoder network, so as to obtain the feature reconstruction data of each grid unit at each time period. The reconstruction error calculation module 74 is used to calculate the reconstruction error based on the multidimensional feature data and feature reconstruction data of each grid cell at each time period, and to obtain the reconstruction error data of each grid cell at each time period. The anomaly identification module 75 is used to identify ecological anomaly occurrence points and classify ecological anomaly levels based on the reconstruction error data of each grid cell at each period and the preset change point detection model, so as to obtain the mangrove ecological anomaly identification results of each grid cell.

[0067] In this embodiment, a data acquisition module obtains multi-source data of mangroves associated with several grid units in the target area at various times within a preset time observation period; a feature extraction module performs multi-dimensional feature extraction based on the multi-source data of mangroves associated with each grid unit at various times to obtain multi-dimensional feature data of each grid unit at various times; a feature reconstruction module performs feature reconstruction based on the multi-dimensional feature data of each grid unit at various times and a pre-trained time-series convolutional autoencoder network to obtain feature reconstruction data of each grid unit at various times; a reconstruction error calculation module calculates the reconstruction error based on the multi-dimensional feature data of each grid unit at various times and the feature reconstruction data to obtain reconstruction error data of each grid unit at various times; and an anomaly identification module identifies ecological anomaly occurrence points and classifies ecological anomaly levels based on the reconstruction error data of each grid unit at various times and a preset change point detection model to obtain the mangrove ecological anomaly identification results of each grid unit. Multidimensional feature data constructed based on multi-source data associated with mangroves is used to reconstruct features using a time-series convolutional autoencoder network to characterize the dynamic evolution baseline of mangroves under normal phenology and tidal cycles. The reconstruction error is obtained by combining the multidimensional feature data and the reconstructed feature data to obtain reconstruction error data reflecting the degree to which the multidimensional feature data of the current grid cell deviates from the natural healthy phenological evolution baseline of the mangroves. Based on the reconstruction error data and the change point detection model, the occurrence of ecological anomalies is identified and the ecological anomaly level is classified, which improves the accuracy and efficiency of mangrove ecological anomaly identification.

[0068] Please refer to Figure 8 , Figure 8This is a schematic diagram of the structure of a computer device provided in one embodiment of this application. The computer device 8 includes: a processor 81, a memory 82, and a computer program 83 stored in the memory 82 and executable on the processor 81. The computer device can store multiple instructions, which are adapted to be loaded and executed by the processor 81 as described above. Figures 1 to 6 The method steps of the illustrated embodiment, and the specific execution process, can be found in the illustration. Figures 1 to 6 The specific details of the illustrated embodiments will not be elaborated here.

[0069] The processor 81 may include one or more processing cores. The processor 81 connects to various parts of the server using various interfaces and lines, and executes various functions and processes data of the mangrove ecological anomaly identification device 7 by running or executing instructions, programs, code sets, or instruction sets stored in the memory 82, and by calling data stored in the memory 82. Optionally, the processor 81 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 81 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU mainly handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required to be displayed on the touch screen; and the modem is used for wireless communication. It is understood that the modem may also not be integrated into the processor 81 and may be implemented as a separate chip.

[0070] The memory 82 may include random access memory (RAM) or read-only memory. Optionally, the memory 82 may include a non-transitory computer-readable storage medium. The memory 82 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 82 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch instructions), instructions for implementing the various method embodiments described above, etc.; the data storage area may store data involved in the various method embodiments described above, etc. Optionally, the memory 82 may also be at least one storage device located remotely from the aforementioned processor 81.

[0071] This application embodiment also provides a storage medium that can store multiple instructions, which are adapted to be loaded and executed by a processor as described above. Figures 1 to 6 The method steps of the illustrated embodiment, and the specific execution process, can be found in the illustration. Figures 1 to 6 The specific details of the illustrated embodiments will not be elaborated here.

[0072] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0073] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0074] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the algorithm. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0075] In the embodiments provided by this invention, it should be understood that the disclosed apparatus / terminal devices and methods can be implemented in other ways. For example, the apparatus / terminal device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0076] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0077] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0078] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms.

[0079] This invention is not limited to the above-described embodiments. If any modifications or variations to this invention do not depart from the spirit and scope of this invention, and if such modifications and variations fall within the scope of the claims and equivalent technologies of this invention, then this invention also intends to include such modifications and variations.

Claims

1. A method for identifying ecological anomalies in mangrove forests, characterized by, Includes the following steps: Obtain multi-source data of mangroves from several grid cells in the target area at various times within a preset time observation period; Multidimensional feature extraction is performed on the mangrove-related multi-source data of each grid unit at each time period to obtain multidimensional feature data of each grid unit at each time period. Feature reconstruction is performed based on the multidimensional feature data of each grid unit at each time period and the pre-trained time series convolutional autoencoder network to obtain the feature reconstruction data of each grid unit at each time period. Reconstruction error is calculated based on the multidimensional feature data and feature reconstruction data of each grid cell at each time period to obtain the reconstruction error data of each grid cell at each time period. Based on the reconstruction error data of each grid cell at each time period and the preset change point detection model, the change points of ecological anomalies are identified and the ecological anomaly levels are classified, so as to obtain the mangrove ecological anomaly identification results of each grid cell.

2. The method for identifying ecological anomalies in mangroves according to claim 1, characterized in that: The mangrove-related multi-source data includes multispectral data, water body data, and mangrove patch data; the multidimensional feature data includes pest and disease stress dimension feature data, vegetation function dimension feature data, disturbance evaluation dimension feature data, environmental quality dimension feature data, and landscape pattern dimension feature data. The step of extracting multidimensional features from mangrove-related multi-source data at various times for each raster unit to obtain multidimensional feature data for each raster unit at various times includes the following steps: Based on the near-infrared reflectance and red-edge reflectance in the multispectral data, and using a preset red-edge normalized vegetation index (RBVI) ​​calculation algorithm, the RBVI is obtained as the feature data for the pest and disease stress dimension. The RBVI calculation algorithm is as follows: In the formula, The red-edged normalized vegetation index, For near-infrared reflectivity, Reflectivity in the red-edge band; Based on the near-infrared reflectance and infrared reflectance in the multispectral data, and a preset normalized vegetation index (NVI) calculation algorithm, a normalized vegetation index is obtained as the vegetation functional dimension feature data. The NVI calculation algorithm is as follows: In the formula, Normalized Difference Vegetation Index (NDVI) For infrared reflectivity; Multispectral classification is performed based on the multispectral data to obtain the total area of ​​invasive alien species; the invasion ratio of alien species is obtained based on the total area of ​​invasive alien species, the total area of ​​raster cells, and a preset algorithm for calculating the invasion ratio of alien species, and is used as the feature data of the interference evaluation dimension, wherein the algorithm for calculating the invasion ratio of alien species is as follows: In the formula, The proportion of invasive alien species, This represents the total area affected by invasive alien species. This represents the total area of ​​the grid cells; Based on the area of ​​spectrally anomalous water bodies, the total area of ​​the water network, and a preset water body apparent anomaly index calculation algorithm from the water body data, a water body apparent anomaly index is obtained as the environmental quality dimension feature data. The water body apparent anomaly index calculation algorithm is as follows: In the formula, This refers to the apparent anomaly index of water bodies. The area of ​​water bodies with spectral anomalies. The total area of ​​the water network; Based on the total number of mangrove patches, the total area of ​​raster cells, and a preset mangrove patch density calculation algorithm in the mangrove patch data, the mangrove patch density is obtained as the landscape pattern dimension feature data. The mangrove patch density calculation algorithm is as follows: In the formula, For mangrove patch density, This represents the total number of mangrove patches.

3. The method for identifying ecological anomalies in mangroves according to claim 1, characterized in that: The time-series convolutional autoencoder network includes a one-dimensional convolutional layer and a transposed convolutional layer; The step of reconstructing features based on the multidimensional feature data of each grid unit at each time period and a pre-trained time-series convolutional autoencoder network to obtain the feature reconstruction data of each grid unit at each time period includes the following steps: The multidimensional feature data of each grid cell at different times are stacked to construct the temporal feature tensor of each grid cell. Temporal feature extraction and dimensionality reduction are performed based on the temporal feature tensors of each grid unit at each time period and the one-dimensional convolutional layer to obtain the latent feature tensors of each grid unit at each time period. Based on the latent feature tensors of each grid cell at each time period and the transposed convolutional layer, feature reconstruction is performed to obtain the feature reconstruction tensor of each grid cell, wherein the feature reconstruction tensor includes the feature reconstruction data of each time period.

4. The method for identifying ecological anomalies in mangroves according to claim 1, characterized in that, The step of calculating the reconstruction error based on the multidimensional feature data and feature reconstruction data of each grid cell at each time period to obtain the reconstruction error data of each grid cell at each time period includes the following steps: Based on the multidimensional feature data and feature reconstruction data of each grid cell at each time period, and the preset reconstruction error calculation algorithm, the reconstruction error data of each grid cell at each time period is obtained, wherein the reconstruction error calculation algorithm is as follows: In the formula, Let be the reconstruction error data for the t-th period. For the multidimensional feature data of the t-th period, Reconstruct the data for the features of the t-th period.

5. The method for identifying ecological anomalies in mangroves according to claim 1, characterized in that: The results of identifying ecological anomaly change points include anomaly confidence and the period label corresponding to the ecological anomaly change point. The process of calculating anomaly confidence and identifying ecological anomaly change points based on the reconstruction error data of each grid cell at different times and a preset change point detection model, to obtain the ecological anomaly change point identification results for each grid cell, includes the following steps: Based on the reconstruction error data of each grid cell at each time period and the preset joint posterior probability calculation algorithm for change point occurrence, the joint posterior probability of change point occurrence for each grid cell at each time period is obtained, wherein the joint posterior probability calculation algorithm for change point occurrence is as follows: In the formula, Let be the joint posterior probability of the change point occurring at time t. Let be the length of the journey in period t, representing the time span since the last ecological anomaly occurred. This represents the cumulative value of the reconstruction error data up to the t-th period. Let be the length of the variable point run in the (t-1)th period. To predict the likelihood probability, let be the length of the previous run. Under the given conditions, the probability density of the reconstructed error data observed in the t-th period is: The prior transition probability for the run length is determined by a predefined conditional risk function and is used to characterize the prior probability distribution of ecological anomaly change points. Let be the normalization constant for the total probability, and let be the numerator term for all possible... The boundary likelihood value obtained by summing; Based on the joint posterior probability of change points occurring in each period of each grid cell and a preset probability threshold, the occurrence period of ecological anomalies is determined, and several occurrence periods of ecological anomalies in each grid cell are obtained. Based on the reconstruction error data of each occurrence period of ecological anomalies in each grid cell, the ecological anomaly level is classified, and the ecological anomaly level data of each occurrence period of ecological anomalies in each grid cell are obtained.

6. The method for identifying ecological anomalies in mangroves according to claim 1, characterized in that, It also includes the following steps: Obtain the location coordinate data of each grid cell, the standard orthophotos of each ecological anomaly occurrence period in each grid cell, and the canopy height model; Based on the location coordinate data of each grid cell, the standard orthophotos of each ecological anomaly occurrence period of each grid cell, the canopy height model, and the ecological anomaly level data, a spatial distribution map of mangrove ecological anomalies is constructed to obtain the spatial distribution map of mangrove ecological anomalies of each ecological anomaly occurrence period in the target area.

7. A device for identifying ecological anomalies in mangrove forests, characterized in that, include: The data acquisition module is used to obtain multi-source data of mangroves collected at various times within a preset time observation period for several grid units in the target area. The feature extraction module is used to extract multidimensional features from the mangrove-related multi-source data of each grid cell at each time period, and obtain multidimensional feature data of each grid cell at each time period. The feature reconstruction module is used to reconstruct features based on the multidimensional feature data of each grid unit at each time period and the pre-trained time series convolutional autoencoder network, so as to obtain the feature reconstruction data of each grid unit at each time period. The reconstruction error calculation module is used to calculate the reconstruction error based on the multidimensional feature data and feature reconstruction data of each grid cell at each time period, and to obtain the reconstruction error data of each grid cell at each time period. The anomaly identification module is used to identify ecological anomaly occurrence points and classify ecological anomaly levels based on the reconstruction error data of each grid cell at each period and the preset change point detection model, so as to obtain the mangrove ecological anomaly identification results of each grid cell.

8. A computer device, characterized in that, include: A processor, a memory, and a computer program stored in the memory and executable on the processor; the computer program, when executed by the processor, implements the steps of the mangrove ecological anomaly identification method as described in any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium stores a computer program, which, when executed by a processor, implements the steps of the mangrove ecological anomaly identification method as described in any one of claims 1 to 6.

Images