A grassland monitoring method and system based on remote sensing image technology

Abstract

The present application relates to the technical field of grassland monitoring, and more particularly to a grassland monitoring method and system based on remote sensing image technology. The method comprises the following steps: collecting multispectral remote sensing image data of a target grassland area, and performing atmospheric correction processing to obtain each band surface reflectance data; calculating a grassland vegetation index based on the surface reflectance data, and comparing it with a preset healthy grassland reference index to generate a vegetation index change amount; extracting a short-wave infrared band reflectance according to the vegetation index change amount, and calculating a soil exposure change amount of the grassland area; identifying a surface reflectance abnormally enhanced area in the visible light band reflectance based on the vegetation index change amount and the soil exposure change amount as a salt representation area; the present application realizes accurate identification and distribution determination of a grassland salinization area and fine monitoring of the spatial distribution by fusing the vegetation index change amount and the soil exposure change amount for spatial correlation analysis.

Description

Technical Field

[0001] This invention relates to the field of grassland monitoring technology, and in particular to a grassland monitoring method and system based on remote sensing image technology. Background Technology

[0002] Grasslands are an important component of terrestrial ecosystems, and their growth status is closely related to ecological stability. Grasslands are prone to salinization, leading to vegetation degradation, soil structure damage, and a decline in ecological functions. Traditional monitoring methods rely heavily on manual field sampling and ground surveys. While these methods can obtain relatively accurate soil and vegetation information, they suffer from limitations such as limited monitoring range, long cycles, and high costs, making it difficult to meet the needs of continuous monitoring of large-scale grassland ecosystems. Multispectral remote sensing images can acquire surface reflectance information over a large spatial area, providing a new technical means for grassland status monitoring. Existing remote sensing-based grassland monitoring typically focuses on the analysis of single vegetation indices, underutilizing the spatial correlation between vegetation degradation, soil exposure, and salt accumulation, resulting in low accuracy in identifying grassland salinization and insufficient ability to determine its spread. Summary of the Invention

[0003] Therefore, it is necessary to provide a grassland monitoring method and system based on remote sensing image technology to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a grassland monitoring method based on remote sensing image technology includes the following steps:

[0005] Step S1: Collect multispectral remote sensing image data of the target grassland area and perform atmospheric correction processing to obtain surface reflectance data for each band;

[0006] Step S2: Calculate the grassland vegetation index based on the surface reflectance data and compare it with the preset healthy grassland reference index to generate the vegetation index change; extract the shortwave infrared reflectance based on the vegetation index change to calculate the soil exposure change in the grassland area.

[0007] Step S3: Identify areas of abnormally enhanced surface reflectance in the visible light band based on changes in vegetation index and soil exposure, and use these areas as salinity characterization regions; calculate the reflectance change direction between adjacent pixels based on the salinity characterization regions, determine the trend of surface reflectance change, and form information on the direction of salinity expansion.

[0008] Step S4: Based on the changes in soil exposure and the direction of salt expansion, construct the salinization classification results of grassland areas and generate a grassland salinization distribution map.

[0009] The present invention also provides a grassland monitoring system based on remote sensing image technology, for performing the grassland monitoring method based on remote sensing image technology as described above, the grassland monitoring system based on remote sensing image technology includes:

[0010] The remote sensing image acquisition module is used to acquire multispectral remote sensing image data of the target grassland area and perform atmospheric correction processing to obtain surface reflectance data for each band.

[0011] The vegetation index change determination module is used to calculate the grassland vegetation index based on surface reflectance data and compare it with the preset healthy grassland reference index to generate the vegetation index change amount; based on the vegetation index change amount, the short-wave infrared reflectance is extracted to calculate the soil exposure change amount in the grassland area.

[0012] The salinity expansion direction determination module is used to identify areas of abnormally enhanced surface reflectance in the visible light band based on changes in vegetation index and soil exposure, which serve as salinity characterization areas; and to calculate the reflectance change direction between adjacent pixels based on the salinity characterization areas, thereby determining the trend of surface reflectance change and forming salinity expansion direction information.

[0013] The grassland salinization distribution map generation module is used to construct the grassland salinization classification results based on the soil exposure changes and salt expansion direction information, and generate grassland salinization distribution maps.

[0014] The beneficial effects of this invention are as follows: By performing atmospheric correction processing on multispectral remote sensing image data of the target grassland area, surface reflectance data of each band are obtained. Based on a unified reflectance, a grassland vegetation index is calculated and compared with a preset healthy grassland reference index, thereby obtaining a vegetation index change that reflects the degree of change in grassland vegetation status, enabling quantitative spatial identification of grassland vegetation degradation. Furthermore, by combining shortwave infrared reflectance to calculate soil exposure changes, a correspondence is established between vegetation degradation and surface exposure, thereby improving the accuracy of grassland degradation identification.

[0015] By jointly identifying areas of abnormally enhanced surface reflectance in the visible light band using changes in vegetation index and soil exposure, these areas are used as salinity characterization areas. This allows for the effective extraction of surface salinity accumulation areas from remote sensing data. Furthermore, by calculating the reflectance change direction between adjacent pixels, the trend of surface reflectance variation is determined, providing information on the expansion direction of salinity within grassland areas. Combining soil exposure change and salinity expansion direction information, salinity levels in grassland areas are classified, and salinity distribution maps are drawn. This allows for a visual representation of the spatial distribution and expansion trend of grassland salinity, improving the accuracy and spatial discrimination capability of grassland salinity monitoring. Attached Figure Description

[0016] Figure 1 This is a schematic diagram illustrating the steps of a grassland monitoring method based on remote sensing image technology.

[0017] Figure 2 This is a schematic diagram of a grassland monitoring system based on remote sensing image technology.

[0018] Figure 3 Map showing the graded distribution of grassland salinization;

[0019] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0020] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0021] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0022] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0023] To achieve the above objectives, please refer to Figures 1 to 3 A grassland monitoring method based on remote sensing image technology includes the following steps:

[0024] All specific values ​​involved in this embodiment are exemplary parameters used to clearly illustrate the technical operation process and are not the only limitation of the present invention.

[0025] Step S1: Collect multispectral remote sensing image data of the target grassland area and perform atmospheric correction processing to obtain surface reflectance data for each band;

[0026] Step S2: Calculate the grassland vegetation index based on the surface reflectance data and compare it with the preset healthy grassland reference index to generate the vegetation index change; extract the shortwave infrared reflectance based on the vegetation index change to calculate the soil exposure change in the grassland area.

[0027] Step S3: Identify areas of abnormally enhanced surface reflectance in the visible light band based on changes in vegetation index and soil exposure, and use these areas as salinity characterization regions; calculate the reflectance change direction between adjacent pixels based on the salinity characterization regions, determine the trend of surface reflectance change, and form information on the direction of salinity expansion.

[0028] Step S4: Based on the changes in soil exposure and the direction of salt expansion, construct the salinization classification results of grassland areas and generate a grassland salinization distribution map.

[0029] In one embodiment, a multispectral remote sensing satellite or UAV remote sensing platform is used to acquire images of the target grassland area, obtaining multispectral remote sensing image data including visible light, near-infrared, and short-wave infrared bands. The multispectral image includes information from multiple bands such as red, green, blue, near-infrared, and short-wave infrared. After acquisition, atmospheric correction processing is performed on the multispectral remote sensing image data to eliminate the influence of atmospheric scattering and absorption on the remote sensing image, converting the original radiance data into surface reflectance data. After atmospheric correction processing, surface reflectance data of the target grassland area in various spectral bands can be obtained, providing a unified data foundation for subsequent vegetation index calculations and surface change analysis.

[0030] Based on the surface reflectance data obtained in step S1, red light band reflectance and near-infrared band reflectance are extracted, and the grassland vegetation index is calculated based on the relationship between them. For example, the Normalized Difference Vegetation Index (NDVI) model is used to calculate the current grassland vegetation index value to reflect the growth status of the grassland vegetation. Subsequently, the calculated current grassland vegetation index value is compared with a pre-set healthy grassland reference index, and the difference between the two is calculated to obtain the vegetation index change. This vegetation index change can reflect the degree of decline of the current grassland vegetation relative to its healthy state.

[0031] After obtaining the vegetation index changes, shortwave infrared reflectance was extracted in areas with significant vegetation decline, and the band ratio changes were calculated by combining it with red light reflectance to reflect the degree of surface soil exposure. When vegetation cover decreases, soil exposure increases, and shortwave infrared reflectance typically shows an increasing trend. By performing spatial correspondence analysis between shortwave infrared reflectance changes and vegetation index changes, areas where vegetation decline and soil exposure increase synchronously were identified, and the soil exposure changes in grassland areas were calculated accordingly.

[0032] Based on the vegetation index changes and soil exposure changes obtained in step S2, the two types of data are spatially overlaid to identify areas exhibiting both vegetation decline and soil exposure enhancement. Visible light reflectance data, such as red, green, and blue light reflectance, are further extracted from these areas. When the visible light reflectance of certain areas is significantly higher than that of surrounding areas, it indicates salt accumulation; saline-alkali surfaces typically exhibit high reflectance characteristics. By spatially connecting these areas of abnormally enhanced reflectance, continuous areas of abnormally enhanced surface reflectance are formed and used as salinity characterization areas.

[0033] Within the salinity characterization region, the reflectance differences between adjacent pixels are calculated, and the spatial direction of reflectance variation is analyzed. For example, by calculating the reflectance gradient between pixels, the direction of reflectance variation from low-value areas to high-value areas is determined. This direction of variation reflects the expansion trend of salinity on the land surface. By comprehensively analyzing the reflectance variation direction of the entire salinity characterization region, information on the direction of salinity expansion is obtained.

[0034] Based on the soil exposure change data obtained in step S2 and the salt spread direction information obtained in step S3, the salinization level of the grassland area is analyzed. The soil exposure change data for each spatial unit within the grassland area is statistically analyzed, and the spatial diffusion trend of salt is analyzed in conjunction with the salt spread direction information. Subsequently, based on the magnitude of the soil exposure change data and the salt spread trend, each spatial unit is classified into different salinization levels, such as mild, moderate, and severe salinization. A salinization distribution map of the grassland area is then drawn based on the spatial distribution of each level, thus visually displaying the degree of grassland salinization and its spatial distribution.

[0035] In another embodiment, taking a grassland monitoring area as an example, image data is acquired using a multispectral remote sensing satellite with a spatial resolution of 10m. The image coverage area is approximately 20km², containing spectral data in the red band (center wavelength approximately 665nm), near-infrared band (center wavelength approximately 842nm), and short-wave infrared band (center wavelength approximately 1610nm). The acquired raw image resolution is 2000×2000 pixels. After image correction using an atmospheric correction algorithm (such as the 6S model), the radiance values ​​are converted into surface reflectance data between 0 and 1. For example, the red reflectance of a certain pixel is 0.18, the near-infrared reflectance is 0.46, and the short-wave infrared reflectance is 0.32. The corrected surface reflectance data for each band will serve as the basic input data for subsequent vegetation index calculations and soil exposure analysis.

[0036] Suppose a pixel in a grassland area has a red reflectance of 0.20 and a near-infrared reflectance of 0.50. The vegetation index (VRI) for this pixel, calculated using the NDVI formula, is 0.43. If the preset healthy grassland reference index is 0.65, then the change in the VRI for this pixel is −0.22, indicating a decrease in vegetation cover compared to a healthy state. Further extraction of the short-wave infrared reflectance at this pixel location is performed; assuming a value of 0.36, the ratio to the calculated red reflectance is 1.80. Through statistical analysis of the entire area, if multiple adjacent pixels simultaneously exhibit a decrease in the VRI and an increase in the short-wave infrared reflectance ratio, these areas are identified as areas of increased soil exposure, and the corresponding change in soil exposure is calculated. For example, if approximately 320 pixels in a certain area show this type of change, then the corresponding area is considered a region of increased soil exposure.

[0037] Suppose that in a grassland area, step S2 yields approximately 500 pixels exhibiting both decreased vegetation index and increased soil exposure. Visible light reflectance data is further extracted from these pixels. For example, some pixels show a red light reflectance of 0.32, while the surrounding area's average is 0.21, demonstrating a significant increase in reflectance. Through spatial connectivity analysis, 180 adjacent pixels are connected to form three continuous regions, which are then identified as salinity characterization areas. Reflectance gradients are then calculated for the pixels within these regions. For instance, within a certain region, the reflectance difference between adjacent pixels gradually increases from 0.02 to 0.08, increasing along a northeastward direction. This indicates that the salinity expansion in this region is from southwest to northeast.

[0038] Assume a 20 km² grassland area is divided into 4000 spatial units. Analysis revealed significant changes in soil exposure in 1200 units, with approximately 700 units exhibiting low changes, 400 in the moderate range, and 100 in the high range. Further analysis of the salinity expansion direction showed that the high-value range was primarily distributed continuously from southwest to northeast. Based on this information, salinization was classified into three levels: areas with soil exposure changes less than 0.10 were classified as slightly salinized; areas with changes between 0.10 and 0.25 were classified as moderately salinized; and areas with changes greater than 0.25 were classified as severely salinized. A salinization distribution map was drawn based on the spatial location of each level of area. The severely salinized area is mainly concentrated within an area of ​​approximately 1.2 km², extending in a band towards the northeast.

[0039] Please refer to [link / reference needed] for further information. Figure 3 The spatial differentiation results of salinization degree obtained by multispectral remote sensing data analysis are shown in the figure. Different colored areas correspond to mild, moderate and severe salinization levels, respectively. The visualization results are formed by combining information such as vegetation decline, soil exposure changes and salt expansion trends, and classifying according to preset grading standards.

[0040] Of particular importance, step S1 includes:

[0041] Acquire multispectral remote sensing image data of the target grassland area;

[0042] Radiometric correction is performed on multispectral remote sensing image data to eliminate differences in sensor response;

[0043] Atmospheric correction is performed on the radiometrically corrected remote sensing image data to eliminate the effects of atmospheric scattering and absorption;

[0044] The surface reflectance data of the grassland area were determined based on the corrected data of each band.

[0045] In one embodiment, images of the target grassland area are acquired using a remote sensing satellite or UAV platform, obtaining multispectral remote sensing image data containing multiple spectral bands. The multispectral remote sensing image data includes at least the blue, green, red, near-infrared, and short-wave infrared bands. After acquiring the multispectral remote sensing image data, radiometric correction is performed on each band to eliminate the influence of differences in response sensitivity between different sensors during imaging, ensuring a unified reference standard for radiance across different bands. After radiometric correction, atmospheric correction is performed on the remote sensing image data. An atmospheric transmission model is established to compensate for factors such as atmospheric scattering and aerosol absorption, thereby reducing interference from the atmospheric environment on the reflectance information of the remote sensing image. After atmospheric correction, the true reflectance data of each band at the land surface is obtained, and a surface reflectance dataset of the target grassland area is constructed based on this reflectance data.

[0046] Preferably, step S2 includes:

[0047] Extract the red and near-infrared reflectance data from the surface reflectance data to determine the distribution of grassland vegetation index and calculate the current grassland vegetation index value.

[0048] Calculate the difference between the current grassland vegetation index value and the preset healthy grassland reference index to form the vegetation index change, and construct a vegetation decline distribution map based on the vegetation index change.

[0049] Shortwave infrared reflectance is extracted within the area covered by the vegetation attenuation distribution map, and the change in the ratio of shortwave infrared reflectance to red light reflectance is calculated to form the surface exposure response value.

[0050] Spatial correspondence analysis was performed between surface exposure response values ​​and vegetation index changes to identify areas where vegetation index changes and surface exposure response values ​​increased synchronously, thus generating soil exposure change data.

[0051] In one embodiment, red light band reflectance and near-infrared band reflectance are extracted from the surface reflectance data obtained in step S1, and a correspondence between the reflectance data is established according to the spatial location of the pixels. The grassland vegetation index is calculated using the red light band and near-infrared band reflectance to obtain the vegetation index distribution of the grassland area. Statistical analysis is then performed on the vegetation index of each pixel within the grassland area to obtain the current grassland vegetation index value. Subsequently, this current grassland vegetation index value is compared with a pre-established healthy grassland reference index. The difference between the two is calculated to obtain the vegetation index change, and the spatial distribution of this change identifies areas with significant vegetation decay, thus obtaining a vegetation decay distribution map.

[0052] Within the area covered by the vegetation attenuation distribution map, shortwave infrared reflectance is further extracted, and the ratio between the two is calculated in conjunction with red light reflectance to reflect the changing characteristics of surface exposure. This yields the surface exposure response value, which is then analyzed in relation to the change in vegetation index according to spatial location. This identifies areas where both show simultaneous enhancement in space, thus determining the location where significant changes in soil exposure occur within the grassland area, and obtaining the change in soil exposure accordingly.

[0053] In another embodiment, taking a remote sensing image of a grassland monitoring area as an example, pixel-level analysis is performed on a grassland area of ​​approximately 20 km². The remote sensing image has a spatial resolution of 10 m and contains approximately 200,000 pixels. Red light and near-infrared reflectance are extracted from the surface reflectance data. For example, if a pixel has a red light reflectance of 0.18 and a near-infrared reflectance of 0.52, the vegetation index for that pixel is calculated to be 0.49 based on their relationship. After repeating the above calculation for all pixels, a vegetation index distribution map of the entire grassland area is obtained, and the current average grassland vegetation index is statistically calculated to be 0.46. Assuming a healthy grassland reference index of 0.60, the difference between the two is −0.14, thus determining that there is a certain degree of vegetation attenuation in the area. The degree of vegetation attenuation is then classified according to the magnitude of the difference between different pixels, forming a vegetation attenuation distribution map.

[0054] Subsequently, shortwave infrared reflectance is extracted within the vegetation attenuation area. For example, if the shortwave infrared reflectance of a pixel is 0.36, its ratio to the red reflectance (0.18) is calculated, yielding a ratio of approximately 2.0, which characterizes the surface exposure level at that location. After calculating the ratio for all relevant pixels, the spatial distribution of surface exposure response values ​​is obtained. Further spatial overlay analysis is performed on the surface exposure response values ​​and vegetation index changes. For instance, when the vegetation index of certain pixels decreases by more than 0.12 and the corresponding surface exposure response value increases by more than 30%, it is determined that there is a significant change in soil exposure in that area. By performing spatial statistics on these pixels, the distribution of soil exposure changes within the grassland area is obtained.

[0055] Preferably, the red and near-infrared reflectance of the surface reflectance data are extracted to determine the distribution of the grassland vegetation index and calculate the current grassland vegetation index value, including:

[0056] From the surface reflectance data, the red and near-infrared reflectance data of the grassland area were extracted, and the relative difference between the near-infrared reflectance and the red reflectance was calculated.

[0057] The relative difference data are arranged according to the spatial location of the grassland area to form the grassland vegetation index distribution.

[0058] The distribution of vegetation indices within the grassland area is statistically analyzed to generate the current grassland vegetation index value.

[0059] In one embodiment, the red and near-infrared reflectance of the grassland area is extracted from the surface reflectance data obtained in step S1, and a correspondence is established according to the spatial location of each pixel in the remote sensing image. For each pixel, the relative difference between the near-infrared and red reflectance is calculated to characterize the vegetation growth status at that pixel location. Subsequently, the obtained relative difference data is arranged according to the spatial location of each pixel within the grassland area, ensuring that the vegetation index data corresponding to each pixel maintains a consistent spatial distribution, thereby obtaining the spatial distribution of the grassland vegetation index throughout the entire grassland area. Further, statistical analysis is performed on the above vegetation index distribution within the grassland area, for example, by summarizing the vegetation index of each pixel according to the number of pixels, and calculating its average value or overall statistical value, thereby obtaining the current grassland vegetation index value reflecting the overall vegetation status of the current grassland. Through the above process, the distribution characteristics of vegetation growth status within the grassland area are reflected at the spatial level.

[0060] In another embodiment, taking a grassland area of ​​approximately 15 km² as an example, a remote sensing image with a resolution of 10 m is processed to obtain approximately 150,000 pixel data points. The red-band reflectance and near-infrared reflectance of each pixel are extracted. For example, if a pixel has a red-band reflectance of 0.21 and a near-infrared reflectance of 0.55, the relative difference of that pixel is calculated to be approximately 0.45. This calculation is repeated for all pixels to obtain a set of vegetation index data for the grassland area. These data points are then arranged according to the spatial location of the pixels in the image, thus obtaining a complete distribution of grassland vegetation indices. Subsequently, the vegetation indices of all pixels within the grassland area are statistically analyzed. For example, the overall average value is calculated to be 0.47, and this value is used as the current grassland vegetation index value to characterize the overall vegetation status of the area at the current time.

[0061] Preferably, the vegetation index data are arranged according to the spatial location of the grassland area to form the grassland vegetation index distribution, including:

[0062] Starting from the boundary of the grassland area, the relative difference data are arranged in a direction from the boundary inward, forming a sequence of arrangement extending from the boundary inward.

[0063] Based on the changes in the relative differences between adjacent positions in the permutation sequence, determine the permutation direction from the boundary to the interior;

[0064] Based on the arrangement direction, adjust the relative difference data that are not distributed in that direction so that the relative difference data form a continuous distribution structure within the grassland area;

[0065] The distribution of grassland vegetation indices was determined based on the continuous distribution structure.

[0066] In one embodiment, after obtaining the relative difference data corresponding to each pixel, the starting position of the arrangement is determined based on the boundary range of the grassland area in the remote sensing image. Starting from the outer boundary of the grassland area, pixels at the boundary are taken as initial pixels, and the relative difference data of adjacent pixels are read layer by layer in a direction extending from the boundary into the grassland, thus obtaining an arrangement sequence that gradually extends from the boundary inward. In this arrangement sequence, the relative difference data between adjacent pixels are compared, and their numerical change trends are analyzed, such as determining whether they show a gradual increase, a gradual decrease, or local fluctuation, thereby determining the main spatial arrangement direction of the relative difference data.

[0067] With a defined arrangement direction as a reference, relative difference data that do not unfold according to this direction are re-incorporated, so that related pixels are connected to the existing sequence in the same spatial arrangement, thereby forming a continuous distribution structure of relative difference data within the grassland area. Based on this continuous distribution structure, the spatial distribution of grassland vegetation index within the grassland area is established, so that the vegetation index of each pixel maintains the spatial position correspondence and can reflect the trend of change from the boundary to the interior.

[0068] In another embodiment, assuming the remote sensing image of a grassland area has a size of 1000×800 pixels, after calculating the relative difference data, the outer pixels of the image are used as the starting point of the boundary, and the relative difference data are arranged in a layer-by-layer manner expanding from the boundary to the center. For example, if the relative difference data at a certain boundary pixel is 0.32, and its neighboring pixels inward are 0.36, 0.41, and 0.44 respectively, by comparing the changes between adjacent pixels, it is determined that the relative difference data as a whole shows a trend of gradually increasing from the boundary to the inward. For pixel data not arranged in this direction, they are re-incorporated into their corresponding positions according to their spatial adjacency, so that the relative difference data of each pixel forms a continuous distribution within the grassland area. In this way, the complete spatial distribution of grassland vegetation index is obtained, representing the spatial variation characteristics of vegetation growth status in the grassland area.

[0069] Preferably, the calculation of the difference between the current grassland vegetation index value and the preset healthy grassland reference index to form the vegetation index change, and the construction of a vegetation decline distribution map based on the vegetation index change, includes:

[0070] Within the grassland area, the current grassland vegetation index value is matched with the preset healthy grassland reference index in the execution space.

[0071] Calculate the difference between the current grassland vegetation index value and the healthy grassland reference index at the corresponding spatial location to determine the change in vegetation index;

[0072] Vegetation change levels are classified based on the magnitude of changes in vegetation indices.

[0073] A grassland vegetation attenuation distribution map was constructed based on the spatial distribution of vegetation change levels.

[0074] In one embodiment, after calculating the grassland vegetation index, the current grassland vegetation index value corresponding to each pixel within the grassland area is determined, and pre-established healthy grassland reference index data is retrieved. The healthy grassland reference index is derived from historical monitoring data, healthy grassland sample data from the same area, or long-term average vegetation index data. Subsequently, a one-to-one correspondence between the current grassland vegetation index value and the healthy grassland reference index is established within the grassland area according to the spatial location of the pixels, so that each pixel location has a corresponding current vegetation index value and a reference vegetation index value.

[0075] After spatial mapping, the difference between the current grassland vegetation index value and the healthy grassland reference index is calculated for each pixel location. This difference is used as the vegetation index change at that pixel location. When the current vegetation index is lower than the healthy grassland reference index, the difference reflects the degree of decline in vegetation growth at that location. When the difference is small or close to zero, it indicates that the vegetation status in that area is close to a healthy level. Based on the magnitude of the vegetation index change, each pixel within the grassland area is classified into different levels, such as slight change, moderate change, and significant change, thus obtaining the corresponding vegetation change level information. The vegetation change levels of each pixel are combined and expressed according to their spatial location in the remote sensing image, thereby obtaining a vegetation attenuation distribution map within the grassland area.

[0076] In another embodiment, assuming a remote sensing image of a grassland area contains 2000×1500 pixels, after calculating the current vegetation index, the average vegetation index of historically healthy grassland is selected as a reference index, for example, a reference value of 0.72. For a given pixel location, if the current vegetation index is 0.61, the calculated change in vegetation index is -0.11; for another pixel location, if the current vegetation index is 0.45, the corresponding change is -0.27. Based on the magnitude of the change, the change is divided into multiple level intervals, for example, within -0.10 is a slight change, -0.10 to -0.20 is a moderate change, and below -0.20 is a significant decay. The different levels are then combined according to the spatial location of the pixels to obtain the vegetation decay distribution map of the grassland area.

[0077] Preferably, constructing a grassland vegetation attenuation distribution map based on the spatial distribution of vegetation change levels includes:

[0078] Starting from the boundary of the grassland area, the vegetation change levels are traversed according to spatial units;

[0079] When the vegetation change levels of adjacent spatial units show a continuous increasing or decreasing relationship, the adjacent spatial units are connected to form a decay region.

[0080] The unconnected spatial units are compared again, and spatial units whose grade difference with the connected areas is within a preset range are connected to the attenuation area.

[0081] Determine the distribution map of grassland vegetation attenuation based on the attenuation areas.

[0082] In one embodiment, the boundary of the grassland area is used as the starting point for spatial traversal. The grassland area is divided into spatial units according to a preset grid scale, and the vegetation change level is read unit by unit according to spatial position. Subsequently, the vegetation change level is compared and judged between the current spatial unit and its adjacent spatial units. When the vegetation change level of adjacent spatial units shows a continuous increasing relationship or a continuous decreasing relationship, it is determined that the two have a consistent vegetation decay trend in space, and a connection relationship is established between the adjacent spatial unit and the current spatial unit to form an initial decay area.

[0083] After the initial attenuation region is formed, the level difference comparison is performed again on the unconnected spatial units in the grassland region. The level difference is calculated between the unconnected spatial units and the boundary spatial units in the adjacent attenuation regions. When the level difference is within the preset difference range, it is determined that the unconnected spatial unit has the same attenuation characteristics as the attenuation region, and the spatial unit is connected to the corresponding attenuation region, thereby gradually expanding the spatial range of the attenuation region.

[0084] Once the connection relationship of all spatial units has been determined, the spatial location of each attenuation region and the distribution of its internal vegetation change level are summarized and identified. The distribution relationship of each attenuation region in the grassland region is then graphically represented to obtain a grassland vegetation attenuation distribution map that reflects the spatial distribution status of grassland vegetation attenuation.

[0085] In another implementation, spatial neighborhood weights are introduced in the process of determining the connection between spatial units. The vegetation change level of multiple adjacent spatial units in the neighborhood is used as a comprehensive judgment basis. When the difference between the spatial unit to be connected and the average level of the neighborhood is less than a preset threshold, the spatial unit is merged into the corresponding attenuation region to further improve the stability of the attenuation region division and obtain the grassland vegetation attenuation distribution map accordingly.

[0086] Preferably, the shortwave infrared reflectance is extracted within the area covered by the vegetation attenuation distribution map, and the change in the ratio of shortwave infrared reflectance to red light reflectance is calculated to form the surface exposure response value, including:

[0087] Within the area covered by the vegetation attenuation distribution map, shortwave infrared reflectance and red light reflectance are extracted according to spatial location.

[0088] Calculate the ratio of shortwave infrared reflectance to red light reflectance at the same spatial location;

[0089] Based on the spatial variation trend of the ratio at each spatial location in the vegetation attenuation distribution map, the change path of the ratio of shortwave infrared band reflectance to red light band reflectance is determined.

[0090] Spatially integrate the comparison value change paths to generate surface exposure response values.

[0091] In one embodiment, within the area covered by the grassland vegetation attenuation distribution map, spatial units are used as the basic analysis units. The corresponding remote sensing band reflectance data are extracted unit by unit according to the spatial location order, including shortwave infrared band reflectance and red light band reflectance. A one-to-one correspondence is established between the two types of band reflectance and the corresponding spatial units.

[0092] At each spatial unit location, the ratio of the corresponding shortwave infrared band reflectance to the red band reflectance is calculated to obtain the band reflectance ratio for that spatial unit location. This ratio characterizes the difference between vegetation cover and bare surface characteristics at that location. After obtaining the band reflectance ratios for each spatial unit, the continuity of the ratio changes between adjacent spatial units is determined by combining the spatial structure of each attenuation region in the vegetation attenuation distribution map. Furthermore, the trend of the ratio changing from low to high or from high to low is identified according to spatial adjacency relationships to determine the path of change in the ratio of shortwave infrared band reflectance to red band reflectance, thereby reflecting the spatial direction of the transition from vegetation cover to bare soil.

[0093] After determining the paths of each ratio change, the paths of ratio change within the same attenuation region are spatially integrated. The magnitude and direction of the ratio change on the path are uniformly identified, and the integrated results are mapped to the corresponding spatial unit locations to obtain the surface exposure response value that characterizes the degree of surface soil exposure.

[0094] In another embodiment, after calculating the ratio of shortwave infrared band reflectivity to red band reflectivity, the ratio changes of adjacent spatial units are smoothed to reduce the impact of outliers in a single spatial unit on the overall change path. Based on this, the ratio change path is determined, and the surface exposure response value is obtained by spatially summarizing the path results.

[0095] Preferably, step S3 includes:

[0096] Within the grassland area, the changes in vegetation index and soil exposure are spatially superimposed to form a degradation superimposed sequence;

[0097] Based on the degraded superposition sequence, the reflectance in the visible light band was extracted, and spatial units with synchronous reflectance enhancement were screened and identified as units with abnormally enhanced surface reflectance.

[0098] Connect adjacent units with anomaly enhancement of surface reflectance to form anomaly enhancement region of surface reflectance, which is then identified as a salinity characterization region.

[0099] Calculate the magnitude pattern of reflectance between adjacent pixels within the salinity characterization region to determine the direction of reflectance change;

[0100] Information on the direction of salt expansion is generated based on the spatial extension trend of the reflectivity change direction.

[0101] In one embodiment, within the grassland area, vegetation index change data and soil exposure change data corresponding to the surface exposure response values ​​generated in the previous steps are acquired. The two types of data are then matched pixel-by-pixel according to spatial location, establishing a correspondence between vegetation index change and soil exposure change under the same spatial coordinates. Based on this, the two types of changes are spatially overlaid, and a joint characterization information including the degree of vegetation attenuation and soil exposure is established based on each spatial unit, thereby forming a degradation overlay sequence covering the entire grassland area to reflect the synergistic relationship between grassland vegetation degradation and surface exposure changes. After obtaining the degradation overlay sequence, visible light reflectance data from remote sensing images at the corresponding spatial locations are further retrieved, including blue, green, and red light reflectance, and analyzed unit-by-unit according to the spatial unit order in the degradation overlay sequence. By comparing the changes in reflectance in the visible light bands of each spatial unit, when a spatial unit simultaneously exhibits a significant increase in reflectance in multiple visible light bands, and this increase trend is consistent with the vegetation attenuation and soil exposure increase trends characterized in the degradation superposition sequence, the spatial unit is selected as a unit with abnormally enhanced surface reflectance.

[0102] Based on spatial adjacency, the selected surface reflectance anomalous enhancement units are connected. When multiple anomalous enhancement units are spatially adjacent and have similar reflectance variation characteristics, they are merged into the same continuous region to form a surface reflectance anomalous enhancement region. Since saline-alkali soils typically exhibit high visible light reflectance, this anomalous enhancement region is identified as a salinity characterization region to indicate the spatial location of grassland salt accumulation or salinization.

[0103] After determining the salinity characterization area, the pixels within that area are used as the analysis objects. The magnitude of the visible light reflectance values ​​between adjacent pixels is analyzed, that is, the magnitude relationship and the change range of the reflectance of adjacent pixels are compared, so as to determine the spatial increasing or decreasing trend of reflectance, and thus determine the direction of reflectance change.

[0104] Based on the reflectance variation direction formed between each pixel, the continuous extension of the reflectance in space is analyzed. Pixel sequences with consistent variation directions are connected, and their overall extension trend is extracted to generate information on the expansion direction of salt in the grassland area.

[0105] In another embodiment, after obtaining the degraded superimposed sequence, spatial units with synchronously enhanced reflectance are directly identified based on the spatial variation of reflectance in the visible light band, and these units are connected into continuous regions through spatial adjacency to determine the salinity characterization region. Within this region, the reflectance magnitude patterns of adjacent pixels are further compared to extract the direction of reflectance change from low to high or from high to low, and the salinity expansion direction information is determined based on the continuous spatial extension of this direction of change.

[0106] Preferably, connecting adjacent surface reflectance anomaly enhancement units to form a surface reflectance anomaly enhancement region, and identifying it as a salinity characterization region includes:

[0107] The visible light band reflectance is extracted from the spatial location corresponding to the degraded superimposed sequence to form a visible light band reflectance sequence.

[0108] Spatial units that are synchronously enhanced with the degradation superimposed sequence are screened from the visible light reflectance sequence and identified as surface reflectance anomalous enhancement units.

[0109] Spatially connect adjacent surface reflectance anomaly enhancement units to form a continuous surface reflectance anomaly enhancement region;

[0110] The areas with abnormally enhanced surface reflectance were identified as salinity characterization areas.

[0111] In one embodiment, after a degraded overlay sequence is formed in the grassland area, visible light band reflectance data from the remote sensing image is retrieved according to the spatial location corresponding to the degraded overlay sequence. The visible light band reflectance data includes reflectance information for the blue, green, and red bands, and is organized according to the arrangement order of spatial units, so that each spatial unit corresponds to a set of visible light band reflectance values, forming a visible light band reflectance sequence consistent with the spatial structure of the degraded overlay sequence.

[0112] After obtaining the visible light reflectance sequence, this sequence is compared with the degradation overlay sequence. Specifically, the changes in visible light reflectance of each spatial unit are compared with the trends in the degradation overlay sequence. When a spatial unit shows a simultaneous increase in both vegetation index and soil exposure in the degradation overlay sequence, and a synchronous increase in reflectance in the visible light reflectance sequence, this spatial unit is identified as a unit with abnormally increased surface reflectance. This method allows for the selection of spatial locations that exhibit both vegetation degradation and high surface reflectance characteristics.

[0113] Based on spatial adjacency, spatial connections are performed on the selected surface reflectance anomaly enhancement units. Specifically, the boundary adjacency or corner adjacency between spatial units is used as the connection basis. Adjacent anomaly enhancement units are traversed unit by unit. When two or more anomaly enhancement units are spatially continuous, they are merged into the same connection region, thereby gradually forming a continuously distributed surface reflectance anomaly enhancement region.

[0114] After a continuous region is formed, the overall reflectance characteristics of the region are comprehensively determined. Since salt accumulation usually leads to a significant increase in surface reflectance, and this characteristic is more pronounced in the visible light band, the region with abnormally enhanced surface reflectance is identified as a salinity characterization region.

[0115] In another embodiment, the visible light reflectance is directly extracted from the spatial location corresponding to the degraded superimposed sequence, and spatial units whose reflectance changes with the degraded superimposed sequence are selected as surface reflectance anomalous enhancement units. Subsequently, these units are connected according to spatial adjacency. When a continuous region is formed, the continuous region is identified as a surface reflectance anomalous enhancement region and used as a salinity characterization region.

[0116] Of particular importance, step S4 includes:

[0117] The soil exposure variation of each spatial unit within the grassland area was extracted, and its distribution was determined according to spatial location.

[0118] Introduce salt expansion direction information into each spatial unit to determine the salt expansion direction relationship between spatial units;

[0119] Based on the magnitude of soil exposure changes and the direction of salt expansion, the salinity levels of each spatial unit were classified to obtain the salinity classification results of grassland areas.

[0120] Based on the spatial distribution of salinization levels within the grassland area, a grassland salinization distribution map was drawn.

[0121] In one embodiment, after determining the direction of salinity expansion, the soil exposure change corresponding to each spatial unit within the grassland area is extracted. Specifically, the grassland area is divided into spatial units according to the spatial pixel structure of the remote sensing image, and the soil exposure change data obtained in the aforementioned steps is read unit by unit, so that each spatial unit corresponds to a soil exposure change value, thereby obtaining a soil exposure change distribution consistent with the spatial structure of the grassland area. In this way, the intensity of changes in soil exposure at various locations in the grassland area is intuitively reflected, providing basic data for determining the degree of salinization. After obtaining the distribution of soil exposure change, the salinity expansion direction information determined in the aforementioned steps is introduced into each spatial unit. Specifically, the salinity expansion direction information is spatially matched with the corresponding spatial unit, so that each spatial unit has a corresponding salinity expansion direction, and the expansion path of salinity in the grassland area is determined by the directional relationship between adjacent spatial units. When the salinity expansion direction of a certain spatial unit points to an adjacent spatial unit, the adjacent spatial unit is considered to be the downstream location of salinity expansion, thereby establishing the salinity expansion direction relationship between spatial units, thus describing the expansion trend of salinity in the grassland area.

[0122] After determining the direction of salinity expansion, a comprehensive assessment is made based on changes in soil exposure and this relationship. Specifically, the magnitude of changes in soil exposure is used as the fundamental criterion for salinity intensity, adjusted for the degree of influence of the salinity expansion direction on each spatial unit. A spatial unit is considered to have a high salinity level if it not only has a high change in soil exposure but is also downstream of a salinity expansion path or at a convergence point. Conversely, a unit is considered to have a low salinity level if its soil exposure change is low or it is not located on a salinity expansion path. Using this assessment method, each spatial unit within the grassland area is classified into salinity levels, thus obtaining the salinity classification results for the grassland area.

[0123] After obtaining the salinization classification results, the spatial distribution of each salinization level in the grassland area is graphically represented. Specifically, the spatial units corresponding to different salinization levels are marked within the grassland area, and different display methods are set according to the differences in levels, thereby presenting the distribution of salinization degree in the overall spatial structure. A grassland salinization distribution map is drawn based on the spatial distribution relationship of salinization levels to intuitively reflect the location and distribution range of salinization in the grassland area.

[0124] The present invention also provides a grassland monitoring system based on remote sensing image technology, for performing the grassland monitoring method based on remote sensing image technology as described above, the grassland monitoring system based on remote sensing image technology includes:

[0125] The remote sensing image acquisition module 101 is used to acquire multispectral remote sensing image data of the target grassland area and perform atmospheric correction processing to obtain surface reflectance data of each band.

[0126] The vegetation index change determination module 102 is used to calculate the grassland vegetation index based on surface reflectance data, compare it with the preset healthy grassland reference index, and generate the vegetation index change amount; extract the short-wave infrared reflectance based on the vegetation index change amount to calculate the soil exposure change amount in the grassland area.

[0127] The salinity expansion direction determination module 103 is used to identify areas of abnormally enhanced surface reflectance in the visible light band reflectance based on the changes in vegetation index and soil exposure, and to use these areas as salinity characterization areas; based on the salinity characterization areas, it calculates the reflectance change direction between adjacent pixels, determines the trend of surface reflectance change, and forms salinity expansion direction information.

[0128] The grassland salinization distribution map generation module 104 is used to construct the grassland salinization classification results based on the soil exposure change and salt expansion direction information, and generate a grassland salinization distribution map.

[0129] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A grassland monitoring method based on remote sensing image technology, characterized in that, Includes the following steps: Step S1: Collect multispectral remote sensing image data of the target grassland area and perform atmospheric correction processing to obtain surface reflectance data for each band; Step S2: Calculate the grassland vegetation index based on the surface reflectance data and compare it with the preset healthy grassland reference index to generate the vegetation index change; extract the shortwave infrared reflectance based on the vegetation index change to calculate the soil exposure change in the grassland area. Step S3: Identify areas of abnormally enhanced surface reflectance in the visible light band based on changes in vegetation index and soil exposure, and use these areas as salinity characterization regions; calculate the reflectance change direction between adjacent pixels based on the salinity characterization regions, determine the trend of surface reflectance change, and form information on the direction of salinity expansion. Step S4: Based on the changes in soil exposure and the direction of salt expansion, construct the salinization classification results of grassland areas and generate a grassland salinization distribution map.

2. The grassland monitoring method based on remote sensing image technology according to claim 1, characterized in that, Step S2 includes: Extract the red and near-infrared reflectance data from the surface reflectance data to determine the distribution of grassland vegetation index and calculate the current grassland vegetation index value. Calculate the difference between the current grassland vegetation index value and the preset healthy grassland reference index to form the vegetation index change, and construct a vegetation decline distribution map based on the vegetation index change. Shortwave infrared reflectance is extracted within the area covered by the vegetation attenuation distribution map, and the change in the ratio of shortwave infrared reflectance to red light reflectance is calculated to form the surface exposure response value. Spatial correspondence analysis was performed between surface exposure response values ​​and vegetation index changes to identify areas where vegetation index changes and surface exposure response values ​​increased synchronously, thus generating soil exposure change data.

3. The grassland monitoring method based on remote sensing image technology according to claim 2, characterized in that, Extracting red and near-infrared reflectance data from land surface reflectance data, determining the distribution of grassland vegetation indices, and calculating the current grassland vegetation index value includes: From the surface reflectance data, the red and near-infrared reflectance data of the grassland area were extracted, and the relative difference between the near-infrared reflectance and the red reflectance was calculated. The relative difference data are arranged according to the spatial location of the grassland area to form the grassland vegetation index distribution. The distribution of vegetation indices within the grassland area is statistically analyzed to generate the current grassland vegetation index value.

4. The grassland monitoring method based on remote sensing image technology according to claim 3, characterized in that, The vegetation index data are arranged according to the spatial location of grassland areas to form the grassland vegetation index distribution, including: Starting from the boundary of the grassland area, the relative difference data are arranged in a direction from the boundary inward, forming a sequence of arrangement extending from the boundary inward. Based on the changes in the relative differences between adjacent positions in the permutation sequence, determine the permutation direction from the boundary to the interior; Based on the arrangement direction, adjust the relative difference data that are not distributed in that direction so that the relative difference data form a continuous distribution structure within the grassland area; The distribution of grassland vegetation indices was determined based on the continuous distribution structure.

5. The grassland monitoring method based on remote sensing image technology according to claim 2, characterized in that, The difference between the current grassland vegetation index value and the preset healthy grassland reference index is calculated to form the vegetation index change, and a vegetation decline distribution map is constructed based on the vegetation index change, including: Within the grassland area, the current grassland vegetation index value is matched with the preset healthy grassland reference index in the execution space. Calculate the difference between the current grassland vegetation index value and the healthy grassland reference index at the corresponding spatial location to determine the change in vegetation index; Vegetation change levels are classified based on the magnitude of changes in vegetation indices. A grassland vegetation attenuation distribution map was constructed based on the spatial distribution of vegetation change levels.

6. The grassland monitoring method based on remote sensing image technology according to claim 5, characterized in that, The grassland vegetation attenuation distribution map constructed based on the spatial distribution of vegetation change levels includes: Starting from the boundary of the grassland area, the vegetation change levels are traversed according to spatial units; When the vegetation change levels of adjacent spatial units show a continuous increasing or decreasing relationship, the adjacent spatial units are connected to form a decay region. The unconnected spatial units are compared again, and spatial units whose grade difference with the connected areas is within a preset range are connected to the attenuation area. Determine the distribution map of grassland vegetation attenuation based on the attenuation areas.

7. The grassland monitoring method based on remote sensing image technology according to claim 2, characterized in that, Shortwave infrared reflectance was extracted within the vegetation attenuation distribution map coverage area, and the change in the ratio of shortwave infrared reflectance to red light reflectance was calculated to form the surface exposure response value, including: Within the area covered by the vegetation attenuation distribution map, shortwave infrared reflectance and red light reflectance are extracted according to spatial location. Calculate the ratio of shortwave infrared reflectance to red light reflectance at the same spatial location; Based on the spatial variation trend of the ratio at each spatial location in the vegetation attenuation distribution map, the change path of the ratio of shortwave infrared band reflectance to red light band reflectance is determined. Spatially integrate the comparison value change paths to generate surface exposure response values.

8. The grassland monitoring method based on remote sensing image technology according to claim 1, characterized in that, Step S3 includes: Within the grassland area, the changes in vegetation index and soil exposure are spatially superimposed to form a degradation superimposed sequence; Based on the degraded superposition sequence, the reflectance in the visible light band was extracted, and spatial units with synchronous reflectance enhancement were screened and identified as units with abnormally enhanced surface reflectance. Connect adjacent units with anomaly enhancement of surface reflectance to form anomaly enhancement region of surface reflectance, which is then identified as a salinity characterization region. Calculate the magnitude pattern of reflectance between adjacent pixels within the salinity characterization region to determine the direction of reflectance change; Information on the direction of salt expansion is generated based on the spatial extension trend of the reflectivity change direction.

9. The grassland monitoring method based on remote sensing image technology according to claim 8, characterized in that, Connecting adjacent units with anomaly enhancement in surface reflectance forms a region with anomaly enhancement in surface reflectance, which is then identified as a salinity characterization region, including: The reflectance of the visible light band is extracted from the corresponding spatial position of the degraded superimposed sequence to form a visible light band reflectance sequence; Spatial units that are synchronously enhanced with the degradation superimposed sequence are selected from the visible light reflectance sequence and identified as surface reflectance anomalous enhancement units. Spatially connect adjacent surface reflectance anomaly enhancement units to form a continuous surface reflectance anomaly enhancement region; The areas with abnormally enhanced surface reflectance were identified as salinity characterization areas.

10. A grassland monitoring system based on remote sensing image technology, characterized in that, For performing the grassland monitoring method based on remote sensing image technology as described in claim 1, the grassland monitoring system based on remote sensing image technology includes: The remote sensing image acquisition module is used to acquire multispectral remote sensing image data of the target grassland area and perform atmospheric correction processing to obtain surface reflectance data for each band. The vegetation index change determination module is used to calculate the grassland vegetation index based on surface reflectance data and compare it with the preset healthy grassland reference index to generate the vegetation index change amount; based on the vegetation index change amount, the short-wave infrared reflectance is extracted to calculate the soil exposure change amount in the grassland area. The salinity expansion direction determination module is used to identify areas of abnormally enhanced surface reflectance in the visible light band based on changes in vegetation index and soil exposure, which serve as salinity characterization areas; and to calculate the reflectance change direction between adjacent pixels based on the salinity characterization areas, thereby determining the trend of surface reflectance change and forming salinity expansion direction information. The grassland salinization distribution map generation module is used to construct the grassland salinization classification results based on the soil exposure changes and salt expansion direction information, and generate grassland salinization distribution maps.

Images