A grassland ecological system monitoring method and system based on remote sensing data, an electronic device, and a storage medium

Abstract

The present application discloses a kind of grassland ecological system monitoring method, system based on remote sensing data, electronic equipment and storage medium, method includes: using satellite sensor to collect the original remote sensing image covering grassland area, and the original remote sensing image is preprocessed, obtain multispectral image dataset;Spectral features in multispectral image dataset are extracted and each band reflectivity and vegetation index are calculated, to determine the spectral feature distribution map of grassland vegetation;Spectral feature distribution map is identified using convolutional neural network, and vegetation structure feature is obtained;Function characteristic parameters are extracted and analyzed from vegetation structure feature, and quantitative characterization result is obtained;Based on spectral feature distribution map, vegetation structure feature and quantitative characterization result, using time series analysis method obtains the spatio-temporal variation trend of grassland vegetation state;Through clustering analysis method, spatio-temporal variation trend is compared with preset ecological health threshold value, to judge the health state partition of grassland ecological system, and obtain health condition classification map.

Description

Technical Field

[0001] This invention relates to the field of ecological environment monitoring technology, specifically to a method, system, electronic device, and storage medium for monitoring grassland ecosystems based on remote sensing data. Background Technology

[0002] Monitoring and managing grassland ecosystems is a key issue in ecology and environmental science, directly related to biodiversity maintenance, carbon cycle regulation, and sustainable socio-economic development. As an important component of the Earth's terrestrial ecosystems, the health status and dynamic changes in grassland services have a significant impact on regional and even global ecological balance. However, traditional monitoring methods such as ground surveys and plot measurements have significant limitations in coverage, data update frequency, and long-term consistency. These limitations make real-time dynamic monitoring of large-scale grassland ecosystems difficult, especially in the context of climate change and intensified human activities, urgently requiring more efficient and precise technological means.

[0003] Existing solutions largely rely on low-resolution remote sensing data or single-indicator analysis, making it difficult to comprehensively capture the complex changes in grassland vegetation and their deep correlation with ecosystem function. Specifically, current methods perform poorly in terms of insufficient resolution, inaccurate spectral feature extraction, and limited spatiotemporal dynamic analysis capabilities, resulting in monitoring results that fail to accurately reflect the true state of grassland ecosystems. The core challenge lies in how to achieve comprehensive analysis of the spectral, structural, and functional characteristics of grassland vegetation using remote sensing data, and how to establish an effective correlation between the spatiotemporal changes of these characteristics and the health status of the ecosystem. In particular, the accuracy of data interpretation, the comprehensiveness of feature recognition, and the reliability of model construction have become bottlenecks for technological breakthroughs in the processing and application of high-resolution multispectral imagery. These unresolved technical factors directly limit the ability to conduct long-term dynamic monitoring of grassland ecosystems, thereby affecting the scientific nature of management decisions.

[0004] Therefore, how to accurately identify the spatiotemporal variations of the spectral, structural, and functional characteristics of grassland vegetation using high-resolution multispectral imaging technology, and on this basis, construct a correlation model between vegetation traits and ecosystem functions, has become a key issue in the field of grassland ecosystem monitoring. Solving this problem will provide solid technical support for the health assessment and sustainable management of grassland ecosystems. Summary of the Invention

[0005] To address the above technical problems, this invention provides a method for monitoring grassland ecosystems based on remote sensing data, the method comprising:

[0006] Raw remote sensing images covering grassland areas are acquired using satellite sensors, and the raw remote sensing images are preprocessed to obtain a multispectral image dataset.

[0007] The spectral features of the multispectral image dataset were extracted using spectral analysis, and the reflectance and vegetation index of each band were calculated to determine the spectral feature distribution map of grassland vegetation.

[0008] A convolutional neural network is constructed, and the convolutional neural network is used to identify the spectral feature distribution map to obtain vegetation structure features;

[0009] Functional feature parameters are extracted from the vegetation structure features, and the functional feature parameters are analyzed using the random forest algorithm to obtain quantitative characterization results;

[0010] Based on the spectral feature distribution map, the vegetation structure characteristics, and the quantitative characterization results, the spatiotemporal variation trend of grassland vegetation status is obtained using time series analysis.

[0011] By comparing the spatiotemporal change trends with preset ecological health thresholds using cluster analysis, the health status zones of the grassland ecosystem are determined, resulting in a health status classification map.

[0012] Preferably, the method for obtaining the multispectral image dataset includes:

[0013] The first image dataset was obtained by using satellite sensors to acquire raw remote sensing images covering grassland areas and extracting them from multispectral data.

[0014] The first image was denoised and cloud removal was performed using Gaussian filtering to obtain the second image dataset.

[0015] The second image dataset is filtered using mean filtering, and the spatial distribution characteristics of the multispectral data are determined. Spectral analysis is then used to determine the coverage area of ​​the grassland region, thus obtaining the multispectral image dataset.

[0016] Preferably, the method for obtaining the spectral feature distribution map includes:

[0017] The multispectral image dataset is acquired, and spectral features are extracted using spectral analysis to obtain initial spectral data.

[0018] The reflectance of each band in the initial spectral data is calculated to obtain the band reflectance value, and the vegetation index is calculated based on the band reflectance value to obtain the index distribution data;

[0019] The initial spectral data and the exponential distribution data are fused to obtain a fused feature set, and the fused feature set is smoothed using the mean filtering method to obtain the spectral feature distribution map.

[0020] Preferably, the method for obtaining the vegetation structure characteristics includes:

[0021] A convolutional neural network is constructed, and the convolutional neural network is used to identify the spectral feature distribution map to obtain a preliminary distribution map of vegetation structure;

[0022] Vegetation structure information is extracted from the preliminary distribution map of the vegetation structure, and the vegetation structure information includes: height characteristics, density characteristics, and coverage.

[0023] The height features are classified using a first preset threshold to obtain layered height distribution data;

[0024] The density of the vegetation structure is calculated using the density feature and the coverage. If the density exceeds a second preset threshold, the preliminary distribution map of the vegetation structure is adjusted to obtain a corrected feature matrix, thus obtaining the vegetation structure feature.

[0025] Preferably, the method for obtaining the quantitative characterization result includes:

[0026] The functional parameters are extracted from the vegetation structure features to obtain the initial dataset;

[0027] The initial data were analyzed and processed using the random forest algorithm to obtain the trend of photosynthetic efficiency.

[0028] Based on the trend of photosynthetic efficiency, water use characteristics are obtained, and photosynthetic efficiency distribution data are obtained.

[0029] If the photosynthetic efficiency distribution data exceeds the third preset threshold, the random forest algorithm is optimized by adjusting the parameters, and the vegetation data is analyzed through the optimized algorithm model to obtain the quantitative value of photosynthetic efficiency.

[0030] The quantitative values ​​of the water use characteristics and the photosynthetic efficiency are compared to determine the deviation range of the efficiency assessment, and the outliers within the deviation range are obtained to obtain the quantitative characterization results of the functional characteristics.

[0031] Preferably, the method for obtaining the spatiotemporal change trend includes:

[0032] Principal component analysis was used to reduce the dimensionality of the quantitative characterization results to obtain the dimensionality-reduced feature data.

[0033] The temporal and spatial changes of the dimensionality-reduced feature data are detected using time series analysis to obtain dynamic change features;

[0034] Based on the aforementioned dynamic change characteristics, the rate of change in the time and spatial dimensions is calculated to obtain quantitative indicators of vegetation status.

[0035] If the quantitative indicator exceeds the fourth preset threshold, the deep spatiotemporal features of the quantitative indicator are extracted, the abnormal trend of the vegetation status is determined, and the correlation between the abnormal trend and the dynamic change features is obtained to obtain the spatiotemporal change trend.

[0036] Preferably, the method for obtaining the health status classification map includes:

[0037] Based on the aforementioned spatiotemporal change trend, trend data is obtained, and K-means clustering is used to cluster the trend data to obtain clustering results;

[0038] Features of each partition are extracted from the clustering results and compared with the preset ecological health threshold to determine the health status;

[0039] If the health status is lower than the ecological health threshold, it is marked as an unhealthy area, and the unhealthy area is processed by the spatial difference method to generate preliminary health status classification data;

[0040] The preliminary health status classification data is visualized using GIS tools to obtain the health status classification map.

[0041] The present invention also provides a grassland ecosystem monitoring system based on remote sensing data. The monitoring system applies the monitoring method described above and includes: a spectral data acquisition module, a feature distribution calculation module, a structural feature extraction module, a quantitative characterization analysis module, a spatiotemporal trend analysis module, and a health status monitoring module.

[0042] The spectral data acquisition module uses satellite sensors to acquire raw remote sensing images covering the grassland area and preprocesses the raw remote sensing images to obtain a multispectral image dataset.

[0043] The feature distribution calculation module uses spectral analysis to extract spectral features from the multispectral image dataset and calculates reflectance and vegetation index for each band to determine the spectral feature distribution map of grassland vegetation.

[0044] The structural feature extraction module is used to construct a convolutional neural network, and the convolutional neural network is used to identify the spectral feature distribution map to obtain vegetation structure features;

[0045] The quantitative characterization analysis module is used to extract functional feature parameters from the vegetation structure features, and to analyze the functional feature parameters using the random forest algorithm to obtain the quantitative characterization results.

[0046] The spatiotemporal trend analysis module obtains the spatiotemporal change trend of grassland vegetation status using time series analysis based on the spectral feature distribution map, the vegetation structure characteristics, and the quantitative characterization results.

[0047] The health status monitoring module compares the spatiotemporal change trend with the preset ecological health threshold using cluster analysis to determine the health status zoning of the grassland ecosystem and obtain a health status classification map.

[0048] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the above-described method for monitoring grassland ecosystems based on remote sensing data.

[0049] The present invention also provides a computer-readable storage medium storing a computer program that, when executed, implements the above-described method for monitoring grassland ecosystems based on remote sensing data.

[0050] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0051] This invention acquires high-resolution multispectral satellite imagery, extracts vegetation spectral features using spectral analysis algorithms, and identifies vegetation structural features using convolutional neural networks. Subsequently, a random forest algorithm is applied to analyze vegetation functional characteristics, such as photosynthetic efficiency and water use efficiency. This invention innovatively combines spectral, structural, and functional features, detects dynamic changes through time-series analysis, and finally uses cluster analysis to compare the changing data with preset ecological health thresholds, achieving a zonal assessment of the grassland ecosystem's health status. This multidimensional and dynamic assessment method provides a scientific basis for the monitoring and management of grassland ecosystems, helping to promptly identify ecological problems and formulate corresponding protection measures. Attached Figure Description

[0052] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention;

[0054] Figure 2 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention.

[0055] Explanation of reference numerals in the attached figures:

[0056] 1010, Processor; 1020, Memory; 1030, Input / Output Interface; 1040, Communication Interface; 1050, Bus. Detailed Implementation

[0057] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this disclosure should have the ordinary meaning understood by one of ordinary skill in the art to which this disclosure pertains. The terms "first," "second," and similar terms used in the embodiments of this disclosure do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0059] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0060] Example 1

[0061] In this embodiment, as Figure 1 As shown, a method for monitoring grassland ecosystems based on remote sensing data is characterized in that the method includes:

[0062] S1. Use satellite sensors to acquire raw remote sensing images covering grassland areas, and preprocess the raw remote sensing images to obtain a multispectral image dataset.

[0063] The method for obtaining a multispectral image dataset includes: acquiring original remote sensing images covering the grassland area using satellite sensors, extracting a first image dataset using multispectral data; denoising and cloud removal of the first image using Gaussian filtering to obtain a second image dataset; filtering the second image dataset using mean filtering, determining the spatial distribution characteristics of the multispectral data, and using spectral analysis to determine the coverage area of ​​the grassland area to obtain a multispectral image dataset.

[0064] In this embodiment, satellite sensors capture spectral information in multiple bands, including visible and near-infrared, covering the grassland area to be monitored, generating a raw remote sensing image with a resolution of 30 meters. A first image dataset is then extracted using the multispectral data. Data from different bands in the first image dataset are stored in layers, for example, the red, green, and near-infrared bands are extracted as independent layers for easier subsequent analysis. Gaussian filtering is used to remove noise from the first image dataset. In this embodiment, radiometric correction techniques can also be used to eliminate sensor bias, or spatial smoothing methods can be used to reduce the impact of random noise. The second image dataset is obtained. Mean filtering is then applied to the second image dataset. Mean filtering calculates the average value of each pixel by sliding a 3×3 window. For example, if the original value of a pixel is 150, the values ​​of its eight surrounding pixels fluctuate between 130 and 160. After filtering, the value approaches 145, resulting in a smooth transition. Then, the spatial distribution characteristics of the multispectral data are determined. Spectral analysis is used to determine the coverage area of ​​the grassland region. Combined with the spectral reflectance curve, typical characteristics of grassland vegetation are identified, such as high reflectance in the near-infrared band. If the reflectance of a certain area is above 0.4, it can be determined to be a grassland-covered area, thus obtaining the multispectral image dataset.

[0065] S2. Use spectral analysis to extract spectral features from multispectral image datasets, calculate reflectance and vegetation index for each band, and determine the spectral feature distribution map of grassland vegetation.

[0066] The method for obtaining the spectral feature distribution map includes: acquiring a multispectral image dataset, extracting spectral features through spectral analysis to obtain initial spectral data; calculating the reflectance of each band in the initial spectral data to obtain band reflectance values, and calculating the vegetation index based on the band reflectance values ​​to obtain index distribution data; fusing the initial spectral data and the index distribution data to obtain a fused feature set, and smoothing the fused feature set using the mean filtering method to obtain the spectral feature distribution map.

[0067] In this embodiment, for images of grassland areas, common near-infrared and red bands are selected, and their spectral reflectance characteristics are analyzed to generate initial spectral data. For the initial spectral data, the reflectance of each band is calculated. The reflectance of the near-infrared band is 0.4, while that of the red band is 0.15. Based on the above reflectance values, the difference between the near-infrared band and the red band, divided by their sum, yields a vegetation index between 0 and 1, i.e., exponential distribution data. The spatial distribution of the vegetation exponential distribution data is superimposed with the multi-band features of the initial spectral data to form a fused feature set. Then, the fused feature set is smoothed using the mean filtering method to obtain a spectral feature distribution map.

[0068] S3. Construct a convolutional neural network and use it to identify the spectral feature distribution map to obtain vegetation structure features.

[0069] The method for obtaining vegetation structure features includes: constructing a convolutional neural network, using the convolutional neural network to identify the spectral feature distribution map to obtain a preliminary vegetation structure distribution map; extracting vegetation structure information based on the preliminary vegetation structure distribution map, including height features, density features, and coverage; classifying the height features using a first preset threshold to obtain layered height distribution data; calculating the density of the vegetation structure using density features and coverage, and if the density exceeds a second preset threshold, adjusting the preliminary vegetation structure distribution map to obtain a corrected feature matrix, thus obtaining the vegetation structure features.

[0070] In this embodiment, the constructed convolutional neural network can adopt a three-layer convolutional structure, with each layer using a 3×3 convolutional kernel, to progressively extract spatial and spectral features of the image, ultimately outputting a preliminary distribution map. When extracting vegetation structure information from the preliminary distribution map, edge detection technology can be used to identify vegetation boundaries. Specifically, for a 1000m... 2 In the grassland area, vegetation height features were detected, showing an average height of 0.8m in some areas and 0.3m in others. Density features were obtained through pixel statistics, with vegetation coverage accounting for 60% of pixels per square meter. Coverage can be calculated based on the proportion of green areas in the image; for example, a certain area may have a coverage of 75%. When classifying height features using a preset threshold, 0.5m can be set as the dividing point. Areas above this value are classified as high-vegetation areas, and those below are classified as low-vegetation areas. When calculating density using density features and coverage, a simple ratio relationship is introduced. When the ratio of density to coverage is greater than 1.2, it is judged as a high-density area. If the density exceeds the preset range, for example, the ratio reaches 1.5, it indicates that the image processing may be affected by lighting or shadow interference. In this case, the feature matrix is ​​corrected by adjusting the contrast or brightness parameters to obtain the vegetation structure features.

[0071] S4. Extract functional feature parameters from vegetation structure characteristics, and analyze the functional feature parameters using the random forest algorithm to obtain quantitative characterization results.

[0072] The method for obtaining quantitative characterization results includes: extracting functional parameters from vegetation structure features to obtain an initial dataset; analyzing and processing the initial data using a random forest algorithm to obtain the trend of photosynthetic efficiency; obtaining water use characteristics based on the trend of photosynthetic efficiency to obtain photosynthetic efficiency distribution data; if the photosynthetic efficiency distribution data exceeds a third preset threshold, optimizing the random forest algorithm by adjusting parameters, and analyzing the vegetation data through the optimized algorithm model to obtain a quantitative value of photosynthetic efficiency; comparing the quantitative values ​​of water use characteristics and photosynthetic efficiency to determine the deviation range of efficiency assessment, obtaining outliers within the deviation range, and obtaining the quantitative characterization results of functional features.

[0073] In this embodiment, parameters such as height and density are extracted from vegetation structure features and combined with spectral data to initially form an initial dataset containing multidimensional features. Then, based on statistical analysis, the initial dataset is dimensionality-reduced to identify feature terms strongly correlated with functional parameters. A random forest algorithm is then used to process the initial dataset to determine the trend of photosynthetic efficiency. Based on this trend, water use-related features are obtained to acquire efficiency distribution data. If the efficiency distribution data exceeds a preset threshold, the random forest model is optimized by adjusting parameters to obtain an updated model. In this embodiment, the preset threshold can be set to an efficiency fluctuation range of ±1. 0%; If the efficiency distribution data of a certain area reaches 90%, exceeding the upper limit of the threshold, the tree depth or feature weight of the random forest is adjusted, the model is re-optimized, and the vegetation data is analyzed through the optimized algorithm model to obtain the quantitative value of photosynthetic efficiency; The quantitative values ​​of water use characteristics and photosynthetic efficiency are compared to determine the deviation range of efficiency assessment. The deviation range can be determined by the difference between the quantitative value and the measured water use rate; If the outliers within the deviation range are concentrated in a certain area, such as if the deviation value is consistently higher than 0.1, it may be due to differences in vegetation species or external environmental interference. The outliers within the deviation range are obtained to obtain the quantitative characterization results of functional characteristics.

[0074] S5. Based on the spectral feature distribution map, vegetation structure characteristics and quantitative characterization results, the spatiotemporal variation trend of grassland vegetation status is obtained by using time series analysis.

[0075] The methods for obtaining spatiotemporal change trends include: using principal component analysis to reduce the dimensionality of the quantitative characterization results to obtain dimensionality-reduced feature data; using time series analysis to detect the temporal and spatial changes of the dimensionality-reduced feature data to obtain dynamic change features; based on the dynamic change features, calculating the rate of change of the time and spatial dimensions to obtain quantitative indicators of vegetation status; if the quantitative indicators exceed a fourth preset threshold, extracting the deep spatiotemporal features of the quantitative indicators, judging the abnormal trend of vegetation status, and obtaining the correlation between the abnormal trend and the dynamic change features to obtain the spatiotemporal change trend.

[0076] In this embodiment, principal component analysis (PCA) is used to reduce the dimensionality of the quantitative characterization results. If the infrared band of the spectral features is highly correlated with the photosynthetic rate of the functional features, PCA will highlight this correlation, resulting in dimensionality-reduced feature data. For the dimensionality-reduced feature data, time series analysis is applied to detect temporal and spatial changes and determine dynamic change features. Based on the dynamic change features, the rate of change in the time and spatial dimensions is calculated to obtain the quantitative indicators of vegetation status. Specifically, the rate of change in the time dimension can be defined as the relative percentage change in photosynthetic rate within a certain time period, and the rate of change in the spatial dimension can be defined as the relative percentage change in spatial changes. If the quantitative indicators of vegetation status exceed a fourth preset threshold, deep spatiotemporal features are extracted to determine the abnormal trend of vegetation status. The correlation between the abnormal trend of vegetation status and the dynamic change features is obtained, and the spatiotemporal change trend is output.

[0077] S6. By comparing the spatiotemporal change trends with the preset ecological health thresholds using cluster analysis, the health status zones of the grassland ecosystem are determined, and a health status classification map is obtained.

[0078] The method for obtaining the health status classification map includes: obtaining trend data based on spatiotemporal change trends; clustering the trend data using the K-means clustering method to obtain clustering results; extracting the features of each partition from the clustering results and comparing them with a preset ecological health threshold to determine the health status; if the health status is lower than the ecological health threshold, it is marked as an unhealthy area, and the unhealthy area is processed using the spatial difference method to generate preliminary health status classification data; and using GIS tools to visualize the preliminary health status classification data to obtain the health status classification map.

[0079] It should be noted that the method of this disclosure embodiment can be executed by a single device, such as a computer or server. The method of this embodiment can also be applied to a distributed scenario, where multiple devices cooperate to complete the task. In such a distributed scenario, one of these devices may execute only one or more steps of the method of this disclosure embodiment, and the multiple devices will interact with each other to complete the method described.

[0080] It should be noted that the above description describes some embodiments of this disclosure. Other embodiments are within the scope of the appended claims. In some cases, it should be understood that the sequence number of each step in the above embodiments does not imply the order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention. The actions or steps recorded in the claims can be performed in a different order than that in the above embodiments and can still achieve the desired result. In addition, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0081] Example 2

[0082] In this embodiment, a grassland ecosystem monitoring system based on remote sensing data includes: a spectral data acquisition module, a feature distribution calculation module, a structural feature extraction module, a quantitative characterization analysis module, a spatiotemporal trend analysis module, and a health status monitoring module.

[0083] The spectral data acquisition module uses satellite sensors to acquire raw remote sensing images covering the grassland area and preprocesses the raw remote sensing images to obtain a multispectral image dataset.

[0084] The feature distribution calculation module uses spectral analysis to extract spectral features from the multispectral image dataset and calculates reflectance and vegetation index for each band to determine the spectral feature distribution map of grassland vegetation.

[0085] The structural feature extraction module is used to construct a convolutional neural network, which is then used to identify the spectral feature distribution map to obtain vegetation structural features.

[0086] The quantitative characterization analysis module is used to extract functional feature parameters from vegetation structure features and analyze the functional feature parameters using the random forest algorithm to obtain quantitative characterization results.

[0087] The spatiotemporal trend analysis module uses time series analysis to obtain the spatiotemporal change trend of grassland vegetation status based on spectral feature distribution maps, vegetation structure characteristics, and quantitative characterization results.

[0088] The health status monitoring module compares the spatiotemporal change trends with preset ecological health thresholds using cluster analysis to determine the health status zoning of the grassland ecosystem and obtain a health status classification map.

[0089] The system described in the above embodiments is used to implement the corresponding grassland ecosystem monitoring method based on remote sensing data in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0090] It should be noted that the aforementioned grassland ecosystem monitoring system based on remote sensing data is presented in the form of functional units. The term "module" here can be implemented in software and / or hardware, without specific limitations.

[0091] For example, a "module" can be a software program, hardware circuit, or a combination of both that implements the above functions. Hardware circuits may include application-specific integrated circuits (ASICs), electronic circuits, processors (e.g., shared processors, proprietary processors, or group processors) and memory for executing one or more software or firmware programs, combined logic circuits, and / or other suitable components that support the described functions.

[0092] Example 3

[0093] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the grassland ecosystem monitoring method based on remote sensing data as described in any of the above embodiments.

[0094] Figure 2 This embodiment illustrates a more specific hardware structure of an electronic device, which may include a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are interconnected internally via the bus 1050.

[0095] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0096] The memory 1020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 1020 and is called and executed by the processor 1010.

[0097] The input / output interface 1030 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components within the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touchscreens, microphones, various sensors, etc., while output devices may include displays, speakers, vibrators, indicator lights, etc.

[0098] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB (Universal Serial Bus), network cable, etc.) or wireless means (such as mobile network, WIFI (Wireless Fidelity), Bluetooth, etc.).

[0099] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.

[0100] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0101] The system described in the above embodiments is used to implement the corresponding grassland ecosystem monitoring method based on remote sensing data in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0102] Example 4

[0103] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, the present invention also provides a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the grassland ecosystem monitoring method based on remote sensing data as described in any of the above embodiments.

[0104] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0105] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to execute the grassland ecosystem monitoring method based on remote sensing data as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0106] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of this disclosure (including the claims) is limited to these examples; within the framework of this disclosure, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of this disclosure as described above, which are not provided in detail for the sake of brevity.

[0107] Additionally, to simplify the description and discussion, and to avoid obscuring the embodiments of this disclosure, the provided drawings may or may not show well-known power / ground connections to integrated circuit (IC) chips and other components. Furthermore, the apparatus may be shown in block diagram form to avoid obscuring the embodiments of this disclosure, and this also takes into account the fact that the details of implementation of these block diagram apparatuses are highly dependent on the platform on which the embodiments of this disclosure will be implemented (i.e., these details should be fully understood by those skilled in the art). While specific details (e.g., circuitry) have been set forth to describe exemplary embodiments of this disclosure, it will be apparent to those skilled in the art that the embodiments of this disclosure may be implemented without these specific details or with variations thereof. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0108] Although this disclosure has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0109] Therefore, the units of the various examples described in the embodiments of this application 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 technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0110] This disclosure is intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for monitoring a grassland ecosystem based on remote sensing data, characterized in that, The method comprises: acquiring original remote sensing images covering a grassland area by using a satellite sensor, and preprocessing the original remote sensing images to obtain a multispectral image dataset; extracting spectral features in the multispectral image dataset by using a spectral analysis method, and calculating band reflectivity and vegetation indexes to determine a spectral feature distribution map of grassland vegetation; constructing a convolutional neural network, and identifying the spectral feature distribution map by using the convolutional neural network to obtain vegetation structure features; extracting functional feature parameters from the vegetation structure features, and analyzing the functional feature parameters by using a random forest algorithm to obtain quantification results; obtaining a spatiotemporal variation trend of a grassland vegetation state by using a time series analysis method based on the spectral feature distribution map, the vegetation structure features, and the quantification results; comparing the spatiotemporal variation trend with a preset ecological health threshold by using a cluster analysis method to determine a health state partition of a grassland ecosystem, and obtaining a health condition classification map; the method for obtaining the quantification results comprises: extracting the functional feature parameters from the vegetation structure features to obtain an initial dataset; analyzing and processing the initial data by using a random forest algorithm to obtain a variation trend of photosynthetic efficiency; obtaining a water use feature according to the variation trend of the photosynthetic efficiency to obtain photosynthetic efficiency distribution data; if the photosynthetic efficiency distribution data exceeds a third preset threshold, adjusting parameters to optimize the random forest algorithm, and analyzing vegetation data by using an optimized algorithm model to obtain a quantification value of the photosynthetic efficiency; comparing the water use feature and the quantification value of the photosynthetic efficiency to determine an efficiency evaluation deviation range, obtaining abnormal points in the deviation range, and obtaining the quantification results of the functional features. 2.The method of claim 1, wherein, the method for obtaining the multispectral image dataset comprises: acquiring original remote sensing images covering a grassland area by using a satellite sensor, and extracting a first image dataset by using multispectral data; performing denoising and cloud layer elimination on the first image by using a Gaussian filtering method to obtain a second image dataset; filtering the second image dataset by using mean filtering, determining spatial distribution features of multispectral data, determining a coverage range of the grassland area by using spectral analysis, and obtaining the multispectral image dataset. 3.The method of claim 1, wherein, the method for obtaining the spectral feature distribution map comprises: obtaining the multispectral image dataset, extracting spectral features by using a spectral analysis method, and obtaining initial spectral data; calculating reflectivity of each band in the initial spectral data to obtain band reflectance, and calculating vegetation indexes based on the band reflectance to obtain index distribution data; performing feature fusion on the initial spectral data and the index distribution data to obtain a fusion feature set, and performing smoothing processing on the fusion feature set by using a mean filtering method to obtain the spectral feature distribution map. 4.The method of claim 1, wherein, the method for obtaining the vegetation structure features comprises: constructing a convolutional neural network, identifying the spectral feature distribution map by using the convolutional neural network, and obtaining a preliminary vegetation structure distribution map; According to the vegetation structure preliminary distribution map, vegetation structure information is extracted, and the vegetation structure information includes height characteristics, density characteristics, and coverage; The height characteristics are classified by using a first preset threshold, and layered height distribution data is obtained; The density characteristics and the coverage are used to calculate the density of the vegetation structure, and if the density exceeds a second preset threshold, the vegetation structure preliminary distribution map is adjusted to obtain a corrected feature matrix, and the vegetation structure characteristics are obtained. 5.The method of claim 1, wherein, The method for obtaining the spatiotemporal change trend includes: The principal component analysis method is used to reduce the dimension of the quantitative representation result, and reduced dimension feature data is obtained; The time variation and the space variation of the reduced dimension feature data are detected by using the time series analysis method, and a dynamic change feature is obtained; Based on the dynamic change feature, the change rates of the time dimension and the space dimension are calculated, and a quantitative index of the vegetation state is obtained; If the quantitative index exceeds a fourth preset threshold, deep spatiotemporal features of the quantitative index are extracted, an abnormal trend of the vegetation state is judged, and the correlation between the abnormal trend and the dynamic change feature is obtained, and the spatiotemporal change trend is obtained. 6.The method of claim 1, wherein, The method for obtaining the health condition classification map includes: Trend data is obtained based on the spatiotemporal change trend, the K-means clustering method is used to cluster the trend data, and clustering results are obtained; Each partition feature is extracted from the clustering results, and is compared with a preset ecological health threshold to judge a health state; If the health state is lower than the ecological health threshold, the unhealthy area is marked, and the unhealthy area is processed by using the spatial difference method to generate preliminary health condition classification data; The preliminary health condition classification data is visualized by using a GIS tool, and the health condition classification map is obtained.

7. A monitoring system for grassland ecosystem based on remote sensing data, the monitoring system applying the monitoring method according to any one of claims 1-6, characterized in that, It includes: A spectrum data acquisition module, a feature distribution calculation module, a structure feature extraction module, a quantitative representation analysis module, a spatiotemporal trend analysis module, and a health condition monitoring module; The spectrum data acquisition module uses a satellite sensor to collect original remote sensing images covering a grassland area, and pre-processes the original remote sensing images to obtain a multispectral image data set; The feature distribution calculation module extracts spectral features in the multispectral image data set by using a spectral analysis method, calculates band reflectance and vegetation indices, and determines a spectral feature distribution map of grassland vegetation; The structure feature extraction module is used to construct a convolutional neural network, and the convolutional neural network is used to identify the spectral feature distribution map to obtain vegetation structure characteristics; The quantitative representation analysis module is used to extract functional feature parameters from the vegetation structure characteristics, and a random forest algorithm is used to analyze the functional feature parameters to obtain a quantitative representation result; The spatiotemporal trend analysis module uses a time series analysis method to obtain a spatiotemporal change trend of a grassland vegetation state based on the spectral feature distribution map, the vegetation structure characteristics, and the quantitative representation result; The health condition monitoring module compares the spatiotemporal change trend with a preset ecological health threshold by using a clustering analysis method to judge a health state partition of a grassland ecosystem, and obtains a health condition classification map.

8. An electronic device, comprising: The computer readable storage medium stores a computer program, and when the computer program is executed, the method for monitoring grassland ecological system based on remote sensing data according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and when the computer program is executed, the method for monitoring grassland ecological system based on remote sensing data according to any one of claims 1 to 6 is implemented.

Images