Aerial image-based steppe area ecological environment soil erosion monitoring system and method

Abstract

The invention relates to the technical field of aerial image recognition, in particular to a steppe area ecological environment soil erosion monitoring system and method based on aerial images. The method comprises the following steps: dividing a grassland area into a plurality of monitoring grids, and collecting spectral feature images and vegetation coverage parameters of a plurality of time nodes in each grid through aerial photography under different meteorological conditions; based on the data, soil exposure changes are tracked to generate a first type of erosion features, and a second type of erosion features are obtained according to spatial distribution thereof. And combining the two types of characteristics to identify the soil erosion situation of the grassland, and finally comparing according to the soil erosion characteristics of the monitoring units in different time periods so as to determine the soil erosion condition of the grassland ecological environment. According to the grassland soil erosion situation identification method, the grassland soil erosion situation is comprehensively identified by combining the analysis of the two erosion characteristics, the efficiency of grassland ecological protection is improved, and the scientificity and systematicness of ecological monitoring are improved.

Description

Technical Field

[0001] This invention relates to the field of aerial image recognition technology, and in particular to a system and method for monitoring soil erosion in grassland areas based on aerial images. Background Technology

[0002] Traditional methods for monitoring soil erosion in grassland areas often rely on ground observations, lacking a comprehensive understanding of vegetation changes and soil erosion dynamics within the grassland region. Existing technologies have limited spatial coverage, making it difficult to capture the impact of different meteorological conditions on grassland ecosystems in real time. This results in low data update frequency and an inability to reflect the evolution of soil erosion problems in a timely manner. This makes assessing the health status of grassland ecosystems a complex and costly task. Traditional monitoring methods are ill-suited to rapidly changing environmental conditions. Due to the geographical characteristics and complex vegetation cover of grassland areas, existing monitoring technologies face significant challenges in accurately identifying and quantifying grassland soil erosion trends. Single monitoring methods cannot fully reveal the interaction between soil exposure and vegetation cover, limiting in-depth analysis of soil erosion characteristics. The one-sidedness of information and the lack of data from multiple time points severely restrict the dynamic monitoring of soil erosion in grassland areas, hindering the formation of effective ecological protection measures and regulatory mechanisms. Summary of the Invention

[0003] Therefore, it is necessary to provide a monitoring system and method for soil erosion in grassland areas based on aerial imagery to solve at least one of the aforementioned technical problems.

[0004] To achieve the above objectives, a method for monitoring soil erosion in grassland areas based on aerial imagery includes the following steps:

[0005] Step S1: Divide the grassland area to be monitored into multiple monitoring unit grids; under different meteorological conditions, collect spectral feature images of at least three time points in each monitoring unit grid by aerial photography, and record the corresponding vegetation cover parameters during the collection period;

[0006] Step S2: Based on spectral feature images and vegetation cover parameters, track the changes in soil exposure of each monitoring unit at the cross-phase to generate the first type of erosion features;

[0007] Step S3: Based on the distribution of the first type of erosion features in terms of spatial location and variation range, the second type of erosion features are obtained in a targeted manner;

[0008] Step S4: Combine the first type of erosion characteristics with the second type of erosion characteristics to identify the characteristics of grassland soil erosion.

[0009] Step S5: Compare the characteristics of grassland soil erosion at different time periods in each monitoring unit grid, and determine the ecological environment erosion situation based on the comparison results.

[0010] This invention also provides a grassland ecological environment soil erosion monitoring system based on aerial imagery, used to execute the grassland ecological environment soil erosion monitoring method based on aerial imagery described above. The grassland ecological environment soil erosion monitoring system based on aerial imagery includes:

[0011] The image acquisition module is used to divide the grassland area to be monitored into multiple monitoring unit grids; under different meteorological conditions, it acquires spectral feature images of at least three time points in each monitoring unit grid through aerial photography, and records the corresponding vegetation cover parameters during the acquisition period.

[0012] The feature fuzzy recognition module is used to track the changes in soil exposure of each monitoring unit at the cross-phase based on spectral feature images and vegetation cover parameters, and generate the first type of erosion features;

[0013] The feature precise identification module is used to obtain the second type of erosion features based on the distribution of the first type of erosion features in terms of spatial location and variation range.

[0014] The identification module is used to combine the first type of erosion features and the second type of erosion features to identify the characteristics of grassland soil erosion.

[0015] The erosion analysis module is used to compare the characteristics of grassland soil erosion at different time periods in each monitoring unit grid, and to determine the ecological environment erosion situation based on the comparison results.

[0016] This invention divides the grassland area to be monitored into multiple monitoring unit grids, enabling detailed monitoring and data collection for each grid. Based on multi-temporal spectral feature images, it ensures a comprehensive understanding of vegetation changes and soil exposure. By collecting data under different meteorological conditions, it significantly improves the accuracy and timeliness of monitoring, laying a solid foundation for subsequent analysis. The recording of vegetation cover parameters further enhances the ability to assess ecological conditions.

[0017] During the analysis phase, by comprehensively applying spectral feature images and vegetation cover parameters, the changes in soil exposure at different time points were effectively tracked, thereby generating the first type of erosion features and achieving a deep understanding of the grassland soil erosion mechanism. Based on the distribution and variation range of the first type of features, the second type of erosion features were specifically acquired, enabling the monitoring system to more accurately identify potential erosion areas. Combining the analysis of the two types of erosion features, a comprehensive identification of the grassland soil erosion situation was formed.

[0018] The subsequent data comparison function, by comparing the erosion trend characteristics between different time periods in the monitoring unit grid, ensures the accurate judgment of the ecological environment erosion situation. The system can promptly detect the progress and changes of erosion, enabling managers to formulate corresponding protection measures based on the ecological characteristics of different areas. This improves the overall efficiency of grassland ecological protection, enhances the scientific and systematic nature of ecological monitoring, and provides strong support for subsequent ecological restoration and sustainable development. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the steps involved in a method for monitoring soil erosion in grassland areas based on aerial imagery.

[0020] Figure 2 A schematic diagram of aerial surveying of the grassland;

[0021] Figure 3 This is a comparison chart of multi-temporal characteristic changes.

[0022] 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

[0023] 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.

[0024] 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.

[0025] 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.

[0026] To achieve the above objectives, please refer to Figures 1 to 3 A method for monitoring soil erosion in grassland areas based on aerial imagery includes the following steps:

[0027] Step S1: Divide the grassland area to be monitored into multiple monitoring unit grids; under different meteorological conditions, collect spectral feature images of at least three time points in each monitoring unit grid by aerial photography, and record the corresponding vegetation cover parameters during the collection period;

[0028] Step S2: Based on spectral feature images and vegetation cover parameters, track the changes in soil exposure of each monitoring unit at the cross-phase to generate the first type of erosion features;

[0029] Step S3: Based on the distribution of the first type of erosion features in terms of spatial location and variation range, the second type of erosion features are obtained in a targeted manner;

[0030] Step S4: Combine the first type of erosion characteristics with the second type of erosion characteristics to identify the characteristics of grassland soil erosion.

[0031] Step S5: Compare the characteristics of grassland soil erosion at different time periods in each monitoring unit grid, and determine the ecological environment erosion situation based on the comparison results.

[0032] In this embodiment, please refer to Figure 2 By regularly dividing the grassland area to be monitored into multiple independent monitoring unit grids according to a uniform scale, the entire area is divided into multiple independent monitoring unit grids. During the actual monitoring process, using an aerial photography platform equipped with a multispectral sensor, spectral feature images of each monitoring unit grid are acquired at no less than three different times under different meteorological conditions such as sunny, cloudy and light wind disturbance. At the same time, vegetation cover parameter data corresponding to each aerial photography time are recorded synchronously during the aerial photography mission.

[0033] Furthermore, the grid side length is determined based on the overall area of ​​the grassland. For example, the grassland is divided into square grids with a side length of 50 meters. Fixed flight routes and overlap rates are set using aerial photography flight planning software to ensure that spectral feature images acquired at different time points completely cover the same monitoring unit in spatial range. At the same time, vegetation cover parameters for the corresponding time points are obtained through ground quadrat surveys or remote sensing inversion. For example, 400 monitoring unit grids are divided in a 100-hectare grassland area. An aerial photography mission is carried out in spring, summer, and early autumn, and the corresponding vegetation cover value is recorded for each grid.

[0034] By combining spectral feature images and vegetation cover parameters, comparative analysis is performed on images of the same monitoring unit at different time points to track changes in the extent and degree of soil exposure. The results of these changes are then summarized to form the first type of erosion characteristics. By setting initial scanning conditions based on vegetation cover, exposure change tracking is initiated only for monitoring units that meet the conditions. Furthermore, the erosion phenomenon is comprehensively characterized by the displacement of surface features in the spectral images and changes in the vegetation index.

[0035] Furthermore, the vegetation cover parameter of each monitoring unit is read. When the cover is lower than a threshold of, for example, 0.45, the key analysis process of that grid is automatically triggered. By matching the feature points of the spectral feature images of the grid at three time points, the displacement of the surface feature points in the time series is calculated. At the same time, combined with the decrease of the vegetation index over time, it is determined whether there is a trend of continuous expansion of soil exposure in the monitoring unit. For example, if the exposed surface area of ​​a certain grid increases from 5% to 12% in three images, accompanied by a continuous decrease in the vegetation index, then this change process is recorded as the first type of erosion feature.

[0036] By analyzing the spatial distribution and variation of the first type of erosion characteristics in each monitoring unit grid, areas with relatively concentrated erosion activity are screened out, and spectral acquisition strategies are adjusted for these areas to obtain the second type of erosion characteristics. The comprehensive erosion index of each monitoring unit is calculated, active erosion monitoring units are identified, and high-density, multi-band spectral acquisition is then carried out on these monitoring units.

[0037] Furthermore, the topsoil offset data is averaged within the grid, the vegetation attenuation gradient is spatially statistically analyzed, and the two are normalized and weighted to obtain the comprehensive erosion index. When the index exceeds a preset threshold, the corresponding grid is marked as an active erosion monitoring unit. For these grids, the parameters of the multispectral image acquisition device are adjusted, for example, increasing the number of bands from 10 to 20 and shortening the sampling interval from 30 days to 10 days, to obtain more detailed spectral response data. These fine-grained spectral data are then aligned and corrected spatially with the first type of erosion features to form the second type of erosion features. For example, in a certain active erosion grid, high-density acquisition reveals that a specific band is more sensitive to changes in exposed fine soil, thus supplementing the original erosion information.

[0038] By jointly analyzing the first and second types of erosion characteristics, and judging the spatial superposition and changing trends of different types of erosion information, the current erosion situation of each monitoring unit is identified. By simultaneously considering the topsoil displacement amplitude, vegetation attenuation degree, and abnormal features in the high-density spectral response, the strength and development direction of erosion are determined.

[0039] Furthermore, the two types of erosion features are resampled according to a unified spatial resolution, and a joint feature table is constructed. When a monitoring unit simultaneously exhibits large surface displacement, significant vegetation attenuation, and abnormal spectral reflectance changes, it is determined that the unit is in a state of intensified erosion. For example, if the displacement of a grid increases significantly within two months and the high-density spectrum shows an increase in the proportion of exposed fine sand, it is identified as an erosion development trend.

[0040] By comparing the characteristics of grassland soil erosion at different time periods in each monitoring unit grid with horizontal and vertical data, the changing patterns of erosion are analyzed, and the overall ecological environment erosion is determined accordingly. Time series comparisons are used to identify whether erosion continues to expand, slows down, or remains stable.

[0041] Furthermore, the erosion characteristics of the same monitoring unit are compared over consecutive time periods on a monthly or quarterly basis. When multiple adjacent time periods show increased erosion, it is determined that the ecological environment of the area is highly eroded. For example, if more than 30% of the monitoring units in a grassland are identified as showing increased erosion over three consecutive time periods within a year-long monitoring cycle, it can be determined that the grassland as a whole is in a state of obvious ecological erosion, thus completing the entire monitoring and judgment process.

[0042] Preferably, step S2 specifically includes:

[0043] The initial scanning conditions for tracking changes in soil bareness are determined based on vegetation cover parameters, and topsoil layer offset data are measured through spectral feature images.

[0044] Calculate vegetation attenuation gradient using topsoil layer offset data;

[0045] The short-term meteorological stability observation window is identified by the correlation analysis node between topsoil migration data and vegetation attenuation gradient;

[0046] The first type of erosion characteristics are identified by vegetation attenuation gradient within a short-term meteorological stability observation window.

[0047] In this embodiment, by reading the vegetation coverage parameters corresponding to each monitoring unit, it is first determined whether it is necessary to initiate a tracking scan of soil exposure changes. If the scanning conditions are met, the positional changes of the topsoil layer are measured using spectral feature images acquired at different times to obtain topsoil layer offset data. Then, the vegetation attenuation gradient is calculated based on the topsoil layer offset data, and a correlation analysis relationship is established between the topsoil layer offset data and the vegetation attenuation gradient to identify the short-term meteorological stability observation window. Finally, only within the short-term meteorological stability observation window, the first type of erosion features are extracted based on the changes in the vegetation attenuation gradient. The entire process is executed sequentially in the order of first screening the objects, then calculating the changes, and then limiting the time window.

[0048] Furthermore, when determining the starting scanning conditions for tracking changes in soil exposure, the existing vegetation cover parameters of each monitoring unit can be read and compared. The vegetation cover value at the current time point can be compared with a pre-set coverage threshold. When the value is lower than the threshold, it is considered that the monitoring unit has a potential risk of soil exposure. By generating a starting scanning identifier, the subsequent spectral image analysis process of the monitoring unit can be initiated, thereby avoiding complex calculations for all monitoring units at the same time.

[0049] It should be explained that when measuring topsoil layer migration data through spectral feature images, spectral feature images acquired by the same monitoring unit at different time points are selected as input data. By comparing fixed surface textures or brightness abrupt change areas in the images, several feature points are selected in the first time point image, and the corresponding positions are found in subsequent time point images. By calculating the positional changes of these feature points between different time points, the migration of the topsoil layer in the time series is obtained, and the migration results of multiple feature points are integrated into the topsoil layer migration data of the monitoring unit.

[0050] Specifically, the topsoil migration data is combined with the vegetation index changes at the corresponding time points. This involves arranging the topsoil migration data in chronological order and synchronously reading the vegetation index values ​​under the same time series. By comparing the magnitude of the decline in vegetation index between adjacent time points, a gradient value reflecting the rate of vegetation decay is obtained, thus forming vegetation decay gradient data that changes over time.

[0051] Furthermore, when identifying short-term weather stability observation windows, the judgment can be based on the synchronous change relationship between topsoil shift data and vegetation attenuation gradient. First, find the time period on the time axis where the rate of change of topsoil shift decreases significantly, and then check whether the vegetation attenuation gradient in that time period shows a change from violent fluctuations to a relatively stable state. When both types of data show a slowing trend in change within the same time range, that time range is marked as a short-term weather stability observation window.

[0052] Furthermore, when identifying the first type of erosion characteristics through the vegetation attenuation gradient within a short-term meteorological stability observation window, only the vegetation attenuation gradient data within that observation window can be extracted for analysis. The overall direction and magnitude of the vegetation attenuation gradient change within that time window can be statistically analyzed. When the vegetation attenuation gradient continues to change negatively or slowly rises but does not return to the initial level within the window, this change state is recorded as the first type of erosion characteristic and associated with the corresponding monitoring unit for storage.

[0053] Preferably, the initial scanning conditions for tracking changes in soil bareness are determined based on vegetation cover parameters, and the topsoil layer shift data are measured using spectral feature images, specifically as follows:

[0054] The initial vegetation coverage values ​​of each monitoring unit grid are extracted from the vegetation coverage parameters;

[0055] When the initial vegetation coverage value is lower than the preset coverage threshold, a start scan trigger signal for the corresponding monitoring unit grid is generated; the spectral feature images of the corresponding monitoring unit grid at different time nodes are selected based on the start scan trigger signal.

[0056] Feature point matching is performed on spectral feature images at different time points to identify the corresponding positions of the same feature points in images of different time phases; the spatial position offset of the feature points between different time points is calculated by the corresponding positions to obtain the topsoil layer offset data;

[0057] By integrating the surface displacement vectors of all feature points within each monitoring unit grid, continuous topsoil layer migration data is generated.

[0058] In this embodiment, please refer to Figure 3 By reading and judging existing vegetation coverage parameters, the monitoring unit grids that need to be focused on are selected, and under certain conditions, the spectral feature images of the corresponding time nodes are called to measure the spatial position changes of the topsoil layer in the time dimension, thereby forming topsoil layer offset data.

[0059] It should be explained that the monitoring unit grid is used as the index unit, and the vegetation coverage parameter records bound to the grid are read one by one. Specifically, the earliest vegetation coverage data obtained within the same monitoring period is selected as the initial vegetation coverage value, and this value is associated with the corresponding monitoring unit grid number and stored.

[0060] Furthermore, the initial vegetation coverage value of each monitoring unit grid is compared with the threshold. When the value is less than, for example, the coverage threshold of 0.45, a start scan trigger identifier is generated for the monitoring unit in the system. Then, the spectral feature image data corresponding to the monitoring unit at different time points are filtered out in the data management module through the identifier.

[0061] It should be explained that within the same monitoring unit grid, at least two spectral feature images acquired at different time points are selected as comparison objects. Specifically, several surface feature points with obvious texture or brightness changes are manually or automatically selected in the spectral feature image of the first time point, and the corresponding feature point positions are found in the spectral feature images of subsequent time points. By recording the changes in the pixel position of the same feature point in different images, the pixel displacement is converted into the actual spatial displacement, thereby obtaining the spatial position offset of the feature point in the time series, and the offset is recorded as part of the topsoil layer offset data.

[0062] Furthermore, after calculating the spatial offset of a single feature point, the displacement results of all feature points within the same monitoring unit grid within the same time period are summarized. This is done by organizing the displacement direction and magnitude of each feature point into a set of surface displacement vectors according to a unified spatial coordinate system. This set is then processed to be continuous so that it can reflect the overall displacement trend of the topsoil layer in the time dimension, thereby generating continuous topsoil layer offset data.

[0063] Preferably, the short-term meteorological stability observation window is identified at the correlation analysis node between topsoil migration data and vegetation attenuation gradient as follows:

[0064] The time point for potential meteorological stability triggering moments is determined based on the rate of change of topsoil migration data.

[0065] The vegetation stress decline trend in the vegetation attenuation gradient is detected based on the time node of the potential meteorological stability trigger moment.

[0066] When the vegetation stress decline trend matches the preset ecological restoration response data with the preset similarity coefficient, the regional convergence stability of the vegetation attenuation gradient is detected, and the monitoring unit is confirmed to have entered the short-term meteorological stability observation window based on the regional convergence stability.

[0067] In this embodiment, by synchronously analyzing the topsoil shift data and vegetation attenuation gradient data of the same monitoring unit in the time series, the time position where the rate of change of the topsoil shift data slows down significantly is first located from the topsoil shift data as a potential meteorological stability trigger moment. Then, the vegetation attenuation gradient is checked around the time position to see if there is a corresponding stress drop. When the two form a consistent stability feature in time, it is confirmed that the monitoring unit has entered the short-term meteorological stability observation window.

[0068] It should be explained that within the same monitoring unit, the topsoil shift data at consecutive time points are arranged in chronological order, and the shift change between adjacent time points is calculated. By comparing the topsoil shift amplitude in two consecutive aerial photography cycles, when the shift change in multiple consecutive time periods is significantly less than the previous change level, for example, from an average of 2 cm per week to less than 0.5 cm, the time point when the rate of change begins to decrease is marked as the potential meteorological stability trigger moment, and its corresponding time label is recorded.

[0069] Furthermore, taking the marked potential meteorological stability trigger time as the center, the gradient data of several consecutive time points before and after the vegetation attenuation gradient data are extracted from the vegetation attenuation gradient data. Specifically, the vegetation attenuation gradient values ​​of 3 to 5 observation cycles before and after the time node are selected, and their direction and magnitude of change are analyzed. When it is found that the vegetation attenuation gradient changes from continuous intensification to slowing down, or from rapid decline to slow recovery, it is determined that there is a vegetation stress decline trend during this period.

[0070] It should be explained that after confirming the vegetation stress decline trend, the vegetation attenuation gradient change curve corresponding to this trend is compared with the pre-stored ecological restoration response data. By comparing the consistency of the change direction and the relative similarity of the change magnitude, a matching degree value is obtained. When the matching degree value reaches a similarity coefficient of, for example, 0.8 or higher, it is further checked whether the vegetation attenuation gradient in the same monitoring unit and its adjacent grids shows synchronous slowing or convergence change characteristics. When multiple adjacent areas show a decrease in fluctuation and a consistent change direction in the same time period, it is confirmed that the monitoring unit has regional convergence stability.

[0071] Furthermore, after regional convergence stability is confirmed, the corresponding time range is marked as a short-term meteorological stability observation window, and this mark is bound and stored with the monitoring unit number as an effective time window for subsequent erosion feature extraction. For example, in a certain monitoring unit, if the topsoil migration rate decreases significantly in the middle of a certain month, and the vegetation attenuation gradient changes from a rapid decrease to a stable fluctuation in the same time period, and the matching degree with the preset ecological restoration response data reaches 0.85, and the surrounding adjacent grids show similar change trends, then this time period is confirmed and recorded as the short-term meteorological stability observation window of that monitoring unit.

[0072] Preferably, detecting the vegetation stress decline trend in the vegetation attenuation gradient based on the time node of the potential meteorological stability triggering moment specifically involves:

[0073] Based on the time node of the potential meteorological stability trigger moment, extract the data sequence of continuous vegetation index change points before and after the corresponding moment from the vegetation decay gradient.

[0074] Analyze the directionality of vegetation index changes in the data series and identify the recovery period when the vegetation index changes from a decline to a slow recovery.

[0075] The vegetation stress decline trend during the vegetation stress decline process can be judged based on the persistence characteristics of vegetation index changes during the recovery period.

[0076] In this embodiment, the potential meteorological stability trigger moment is used as the time reference point. Vegetation index change data within a certain range before and after the moment is extracted from the time series corresponding to the vegetation attenuation gradient. By analyzing the direction and duration of vegetation index change within this time range, it is determined whether there is a vegetation stress decline trend that is changing from a stressed state to a recovering state, thereby completing the identification of the vegetation stress decline process.

[0077] It needs to be explained that the time label corresponding to the potential meteorological stability trigger moment is located. Centered on this time label, several consecutive time points of vegetation index values ​​are selected from the vegetation decay gradient data, both forward and backward. For example, data points from 5 observation cycles before and after this moment are selected and arranged into a continuous data sequence in chronological order.

[0078] Furthermore, point-by-point comparison of the extracted data sequence is used to determine the direction of change of vegetation index between adjacent time points. When the vegetation index gradually changes from a continuous decline to a decrease in the rate of decline and finally shows a slight rebound in multiple consecutive time points, the time range from negative change to positive change is marked as the vegetation index rebound period, and the start and end times of this period are recorded.

[0079] It should be explained that during the confirmed recovery period, the changes in the vegetation index are observed and statistically analyzed to determine whether the vegetation index has maintained a recovery or remained basically stable without a significant decline again in multiple consecutive observation periods. When the recovery or stability lasts for a preset time period, such as 3 to 4 consecutive observation periods, the change process is judged as a vegetation stress decline trend.

[0080] Preferably, when the matching degree between the vegetation stress decline trend and the preset ecological restoration response data reaches a preset similarity coefficient, the regional convergence stability of the vegetation attenuation gradient is detected, and the monitoring unit is confirmed to have entered the short-term meteorological stability observation window based on the regional convergence stability. Specifically:

[0081] When the vegetation stress decline trend matches the preset ecological restoration response data with the preset similarity coefficient, the continuous change trend of the vegetation attenuation gradient is monitored. When the vegetation attenuation gradient shows a trend of slow convergence from fluctuation, the corresponding convergence and stabilization trend is generated.

[0082] The short-term stable range of vegetation decline is determined based on the duration and change pattern corresponding to the convergent stable trend.

[0083] The regional convergence stability is confirmed by the direction of the vegetation attenuation gradient fluctuation in the short-term stable interval, and the monitoring unit is confirmed to have entered the short-term meteorological stability observation window by the regional convergence stability.

[0084] In this embodiment, when the matching degree between the vegetation stress decline trend and the preset ecological restoration response data reaches the preset similarity coefficient, the regional convergence stability of the vegetation attenuation gradient is detected, and the monitoring unit is confirmed to have entered the short-term meteorological stability observation window. Under the premise that the vegetation stress decline trend has been identified, the trend is compared with the ecological restoration response data pre-stored in the system. When the similarity of the change characteristics of the two meets the set conditions, the change state of the vegetation attenuation gradient in time and space is tracked. By identifying whether the change has changed from fluctuation to convergence, it is determined whether there is a stable observation window.

[0085] It should be explained that by calling the vegetation attenuation gradient time series of the corresponding monitoring unit, the confirmed vegetation stress decline trend segment is aligned and compared with the ecological restoration response data point by point. When the matching degree reaches a similarity coefficient threshold of, for example, 0.8, the vegetation attenuation gradient change after that time period is continuously monitored to observe whether its fluctuation amplitude gradually decreases over time. When a significant narrowing of the fluctuation range is detected, the change state is marked as a trend of slow convergence from fluctuation, and a corresponding convergence stabilization trend indicator is generated.

[0086] Furthermore, after generating the convergence and stabilization trend indicator, the duration of the trend is statistically analyzed, and the change pattern of the vegetation decay gradient within this time period is used to make a judgment. When the vegetation decay gradient maintains small fluctuations and does not show obvious reverse changes within multiple consecutive observation periods, such as the change amplitude within 5 consecutive observation periods being lower than the previous average level, this continuous time period is defined as the short-term stable interval of vegetation decay, and its start and end time range is recorded.

[0087] It should be explained that, within the established short-term stable range, statistical analysis is performed on the fluctuation direction of the vegetation attenuation gradient within this range to determine whether the vegetation attenuation gradient within this range mainly exhibits a small change in a single direction, or whether it alternates between positive and negative directions but with limited amplitude. When this fluctuation pattern occurs simultaneously in the same monitoring unit and its adjacent areas, it is confirmed that the monitoring unit has regional convergence stability.

[0088] Furthermore, after regional convergence stability is confirmed, the corresponding short-term stable interval is directly marked as the short-term meteorological stability observation window, and this window is bound and stored with the monitoring unit number. For example, in a certain monitoring unit, the matching degree between the vegetation stress decline trend and the ecological restoration response data reaches 0.85. Subsequently, the vegetation attenuation gradient changes from obvious fluctuations to small convergence changes in 6 consecutive observation periods, and adjacent monitoring units show similar change directions. Then, the time period corresponding to these 6 observation periods is confirmed as the short-term meteorological stability observation window.

[0089] Preferably, the regional convergence stability is confirmed based on the fluctuation direction of the vegetation attenuation gradient within the short-term stable interval, and the monitoring unit is confirmed to have entered the short-term meteorological stability observation window based on the regional convergence stability.

[0090] The positive and negative rate fluctuations of the convergence rate can be identified by the direction of the vegetation decay gradient fluctuation in the short-term stable interval.

[0091] The convergence stability of the region is confirmed based on the changing frequencies of positive and negative velocity fluctuations.

[0092] The convergence trend of regional convergence stability is used to determine the convergence indicator of change to determine whether the change tends to stagnate.

[0093] By using change convergence indicators to match meteorological stability trigger conditions, it can be confirmed whether the monitoring unit has entered the short-term meteorological stability observation window.

[0094] In this embodiment, within the established short-term stable range, the directionality of vegetation attenuation gradient changes over time is analyzed in detail. By distinguishing different rate fluctuation types in the change process, their frequency of occurrence is counted, and it is determined whether the change shows a convergence trend. When the convergence trend reaches the preset judgment condition, a change convergence indicator is generated to trigger the meteorological stability judgment.

[0095] It should be explained that the vegetation decay gradient values ​​corresponding to consecutive time points are extracted within the short-term stable interval. The changes in vegetation decay gradient between adjacent time points are compared. When the gradient change direction at a certain time point is consistent with the previous time point, the change is marked as a positive rate fluctuation. When the gradient change direction is opposite to the previous time point, the change is marked as a negative rate fluctuation.

[0096] Furthermore, after marking the positive and negative rate fluctuations, the frequency and distribution of their occurrence within the short-term stable interval are statistically analyzed. The frequency values ​​of the positive and negative rate fluctuations are calculated. When the frequencies of the two types of fluctuations are close and the overall fluctuation amplitude remains within a small range, it is determined that the vegetation attenuation change of the monitoring unit in this time interval presents a regional convergent stable state, and this state is taken as the determination result of regional convergent stability.

[0097] It should be explained that after the regional convergence stability is confirmed, the overall trend of vegetation decay gradient is further observed. When the vegetation decay gradient fluctuates slightly around a certain value in multiple consecutive time points without significant deviation, for example, the change amplitude is less than the set range of 0.02 in 4 to 6 consecutive observation periods, a change convergence indicator is generated.

[0098] Furthermore, the change convergence flag is compared with the preset meteorological stability triggering conditions in the system. When a monitoring unit simultaneously meets the conditions that the change convergence flag exists and the previous potential meteorological stability triggering time has been recorded within the same time period, the time period is confirmed and marked as a short-term meteorological stability observation window. For example, in a certain monitoring unit, the short-term stability interval is 7 days. Within these 7 days, the frequency of positive rate fluctuations and negative rate fluctuations is close, and the change amplitude of vegetation attenuation gradient remains within a small range. Therefore, a change convergence flag is generated.

[0099] Preferably, identifying the positive and negative velocity fluctuations of the convergence rate based on the fluctuation direction of the vegetation attenuation gradient within the short-term stable interval specifically involves:

[0100] Numerical sequences of vegetation decay gradients at continuous time points are extracted within a short-term stable interval.

[0101] By performing a difference operation on the convergence rates of adjacent time points in the numerical sequence, a sequence of convergence rate changes is obtained.

[0102] Based on the sign of the values ​​in the convergence rate change sequence, positive changes are marked as positive rate fluctuations, and negative changes are marked as negative rate fluctuations.

[0103] By statistically analyzing the frequency and distribution patterns of positive and negative rate fluctuations within the short-term stable interval, we can identify the positive and negative rate fluctuations of the convergence rate.

[0104] In this embodiment, the positive and negative rate fluctuations of the convergence rate are identified based on the fluctuation direction of the vegetation decay gradient in the short-term stable interval. Within the determined short-term stable interval, the data on the change of the vegetation decay gradient over time are continuously read. By analyzing the differences in the direction of change between adjacent time points, the change process is divided into different types of rate fluctuations.

[0105] It needs to be explained that extracting the numerical sequence of vegetation decay gradients at consecutive time points within a short-term stable interval specifically involves locating the start and end times corresponding to the short-term stable interval. This is done by extracting the gradient values ​​corresponding to each observation time point from the vegetation decay gradient database in chronological order within that time range. For example, within a 7-day short-term stable interval, the vegetation decay gradient values ​​are obtained once a day, thus forming a numerical sequence containing 7 consecutive time points.

[0106] Furthermore, in the established numerical sequence of vegetation attenuation gradient, the numerical changes of two adjacent time points are compared one by one in chronological order. The direction and magnitude of change between the later time point and the previous time point are recorded, and the results are organized into a continuous sequence of changes to reflect the subtle fluctuations of the vegetation attenuation gradient within the short-term stable range.

[0107] It needs to be explained that after obtaining the sequence of changes, the direction of each change is determined. When a change shows that the direction of change is consistent with the previous stage or continues to converge in the same direction, the change is marked as a positive rate fluctuation. When a change shows that the direction of change is opposite to the previous stage or reverses, the change is marked as a negative rate fluctuation.

[0108] Furthermore, after marking the fluctuation types of all changes, all positive and negative rate fluctuations within the short-term stable interval are summarized and statistically analyzed. The number of occurrences of the two types of fluctuations in the entire interval is calculated, and their distribution on the time axis is analyzed to determine whether they exhibit alternating or concentrated patterns, thereby completing the identification of positive and negative rate fluctuations in the convergence rate.

[0109] Preferably, the method of confirming the convergence stability of the region based on the changing frequencies of the forward and reverse velocity fluctuations is as follows:

[0110] Based on the frequency of occurrence of positive and negative rate fluctuations within the short-term stable range, the frequency values ​​of positive and negative fluctuations are obtained.

[0111] The ratio of the positive fluctuation frequency value to the negative fluctuation frequency value is used to obtain the fluctuation frequency balance coefficient.

[0112] When the fluctuation frequency balance coefficient is within the preset balance range, the convergence region of the monitoring unit is determined to be bidirectional balanced fluctuation.

[0113] Extract the characteristics of bidirectional equilibrium fluctuations and confirm the regional convergence stability based on the duration of the bidirectional equilibrium fluctuation characteristics.

[0114] In this embodiment, after marking the positive and negative rate fluctuations, all time points within the short-term stable interval are traversed and counted. The number of occurrences of the positive rate fluctuation marker and the number of occurrences of the negative rate fluctuation marker are accumulated respectively. The corresponding number is divided by the total number of time points within the short-term stable interval to form the positive fluctuation frequency value and the negative fluctuation frequency value. For example, if the short-term stable interval is 20 time points in length, with 9 occurrences of the positive rate fluctuation and 11 occurrences of the negative rate fluctuation, the frequency values ​​of 0.45 and 0.55 are obtained respectively.

[0115] A unified numerical calculation module is used to calculate the ratio between the two. The quotient between the positive and negative fluctuation frequency values ​​is used as the fluctuation frequency balance coefficient. Before the calculation, the frequency values ​​are checked for minimum thresholds. For example, if any frequency value is less than 0.05, smoothing is performed first.

[0116] First, set the equilibrium range of the fluctuation frequency equilibrium coefficient, for example, between 0.8 and 1.25. After the fluctuation frequency equilibrium coefficient is calculated in real time, compare it with the equilibrium range. If the calculation result falls within the range, mark the current convergence region as a bidirectional equilibrium fluctuation state in the monitoring unit status record. It should be explained that the equilibrium range can be calibrated based on historical monitoring data, but remains unchanged during a single analysis.

[0117] After determining that it is a bidirectional equilibrium fluctuation, the duration of this state on the time axis is continuously recorded, and the characteristic parameters of the alternating distribution of positive and negative rate fluctuations in the corresponding time period are extracted simultaneously. Specifically, these parameters include the time interval range of the alternation and the frequency deviation amplitude. Furthermore, when the duration of the bidirectional equilibrium fluctuation state exceeds the preset minimum stable duration threshold, such as when it is maintained for more than 6 consecutive time points, it is confirmed that the corresponding region has reached a convergent stable state.

[0118] Most importantly, step S3 is as follows:

[0119] The distribution of topsoil migration data and vegetation attenuation gradient data in each monitoring unit grid is extracted from the first type of erosion characteristics. The comprehensive erosion index of each monitoring unit grid is calculated, and active erosion monitoring units are selected based on the comprehensive erosion index.

[0120] Based on the differences in the variation amplitude of topsoil layer migration data and vegetation attenuation gradient data in the active erosion monitoring unit, the acquisition parameters of the depth spectral analysis path of the multispectral image acquisition device are determined. Among them, the monitoring unit with the larger variation amplitude adopts a higher density of spectral bands and a shorter sampling interval.

[0121] Based on the acquired parameters, targeted high-density spectral acquisition is performed on the erosion-active monitoring unit to obtain fine-grained spectral responses;

[0122] The fine-grained spectral response and the first type of erosion features are spatially registered and aligned, and the spatial position of the fine-grained spectral response data is corrected according to the displacement of the topsoil layer offset data to obtain the second type of erosion features.

[0123] In this embodiment, based on the constructed monitoring unit grid system, the first type of erosion characteristic data is mapped to the corresponding grid units according to spatial coordinates, forming a raster map of topsoil layer offset data and a raster map of vegetation attenuation gradient data respectively. Then, within each monitoring unit grid, the topsoil layer offset amplitude and vegetation attenuation gradient values ​​are normalized and weighted and superimposed to obtain the comprehensive erosion index of the corresponding grid. For example, the average topsoil layer offset in a certain grid is 3.2 cm and the average vegetation attenuation gradient is 0.18. After uniform scale conversion, a single comprehensive erosion index is generated.

[0124] After screening the active erosion monitoring units, the maximum variation of topsoil migration data and the fluctuation range of vegetation attenuation gradient for each active erosion monitoring unit within the continuous observation period are statistically analyzed. The two types of variation amplitudes are then input as parameters into the acquisition parameter configuration module. Specifically, when the topsoil migration variation of a monitoring unit exceeds 2 cm and the vegetation attenuation gradient variation exceeds 0.15, a higher density of spectral bands is assigned to the monitoring unit, for example, from the original 12 bands to 24 bands, while the sampling time interval is shortened.

[0125] During the operation of the multispectral image acquisition device, the control system schedules the acquisition parameters set for each active erosion monitoring unit as described above, so that the acquisition device automatically switches to high-density spectral acquisition mode when it enters the spatial range of the corresponding monitoring unit, and completes continuous scanning acquisition according to the set number of bands and sampling interval, thereby obtaining fine-grained spectral response data covering the monitoring unit, such as completing a complete spectral scan of 24 bands within 5 seconds.

[0126] After completing the high-density spectral acquisition, the fine-grained spectral response data and the first type of erosion characteristic data are first spatially registered according to the spatial coordinate system of the monitoring unit grid. This ensures that the two types of data correspond to the same monitoring unit position under the same spatial reference frame. Then, the topsoil layer offset data within the monitoring unit is retrieved and used as the basis for spatial correction. The pixel positions of the fine-grained spectral response are translated or resampled and adjusted. When the topsoil layer as a whole shifts by 1.5 cm in a certain direction, the spectral response data is simultaneously spatially corrected in the corresponding direction. Furthermore, the corrected fine-grained spectral response and the original erosion characteristics together constitute the second type of erosion characteristic data.

[0127] Of particular importance is the extraction of topsoil migration data and vegetation attenuation gradient data from the first type of erosion characteristics, their distribution across each monitoring unit grid, the calculation of the comprehensive erosion index for each monitoring unit grid, and the selection of active erosion monitoring units based on the comprehensive erosion index:

[0128] The topsoil layer migration data is spatially rasterized, the average displacement within each monitoring unit grid is statistically analyzed, and a displacement intensity distribution map is generated.

[0129] Spatial interpolation is performed on the vegetation attenuation gradient data to calculate the average attenuation gradient within each monitoring unit grid and generate a vegetation degradation distribution map.

[0130] The average displacement and average attenuation gradient of each monitoring unit grid are normalized and then weighted and summed to obtain the comprehensive erosion index of each monitoring unit grid.

[0131] The comprehensive erosion index is compared with the preset active threshold, and monitoring unit grids with a comprehensive erosion index greater than or equal to the preset active threshold are selected and marked as active erosion monitoring units.

[0132] In this embodiment, the collected topsoil displacement data is spatially rasterized according to the monitoring unit grid, the average displacement of all feature points in each grid is calculated, and a displacement intensity distribution map is generated. For example, the average displacement in a certain grid is 3.5 cm. The displacement intensity distribution map is used to visually display the topsoil movement intensity of each grid. At the same time, spatial interpolation is performed on the vegetation attenuation gradient data. The average attenuation gradient in each monitoring unit grid is calculated by linear or surface interpolation methods between sampling points, and a vegetation degradation distribution map is generated.

[0133] The average displacement and average attenuation gradient values ​​of each monitoring unit grid are normalized to unify the two types of data to a standard scale of 0 to 1. The maximum-minimum normalization method is used for topsoil migration data, and the same normalization method is used for vegetation attenuation gradient. Then, the normalized displacement and attenuation gradient are weighted and summed according to preset weights, for example, the displacement weight is 0.6 and the attenuation gradient weight is 0.4, to obtain a comprehensive index value between 0 and 1.

[0134] For each monitoring unit grid, the comprehensive index value is compared with a pre-set activity threshold. For example, if the activity threshold is 0.65, and the comprehensive index of a certain grid is 0.72, the grid is identified as an active erosion monitoring unit. Active monitoring units can be directly used for subsequent high-density spectral acquisition and generation of second-type erosion features. The selected active erosion monitoring units are marked on the map, and their grid number, location coordinates, and comprehensive index value are recorded.

[0135] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0136] 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 method for monitoring soil erosion in grassland areas based on aerial imagery, characterized in that, Includes the following steps: Step S1: Divide the grassland area to be monitored into multiple monitoring unit grids; Under different meteorological conditions, aerial photography was used to collect spectral feature images of at least three time points in each monitoring unit grid, while the corresponding vegetation cover parameters were recorded during the collection period. Step S2: Based on spectral feature images and vegetation cover parameters, track the changes in soil exposure of each monitoring unit at the cross-phase to generate the first type of erosion features; Step S3: Based on the distribution of the first type of erosion features in terms of spatial location and variation range, the second type of erosion features are obtained in a targeted manner; Step S4: Combine the first type of erosion characteristics with the second type of erosion characteristics to identify the characteristics of grassland soil erosion. Step S5: Compare the characteristics of grassland soil erosion at different time periods in each monitoring unit grid, and determine the ecological environment erosion situation based on the comparison results.

2. The method for monitoring soil erosion in grassland areas based on aerial imagery according to claim 1, characterized in that, Step S2 is as follows: The initial scanning conditions for tracking changes in soil bareness are determined based on vegetation cover parameters, and topsoil layer offset data are measured through spectral feature images. Calculate vegetation attenuation gradient using topsoil layer offset data; The short-term meteorological stability observation window is identified by the correlation analysis node between topsoil migration data and vegetation attenuation gradient; The first type of erosion characteristics are identified by vegetation attenuation gradient within a short-term meteorological stability observation window.

3. The method for monitoring soil erosion in grassland areas based on aerial imagery according to claim 2, characterized in that, The initial scanning conditions for tracking changes in soil bareness are determined based on vegetation cover parameters. Specifically, topsoil layer migration data are measured using spectral feature images. The initial vegetation coverage values ​​of each monitoring unit grid are extracted from the vegetation coverage parameters; When the initial vegetation coverage value is lower than the preset coverage threshold, a start scan trigger signal for the corresponding monitoring unit grid is generated; the spectral feature images of the corresponding monitoring unit grid at different time nodes are selected based on the start scan trigger signal. Feature point matching is performed on spectral feature images at different time points to identify the corresponding positions of the same feature points in images of different time phases; the spatial position offset of the feature points between different time points is calculated by the corresponding positions to obtain the topsoil layer offset data; By integrating the surface displacement vectors of all feature points within each monitoring unit grid, continuous topsoil layer migration data is generated.

4. The method for monitoring soil erosion in grassland areas based on aerial imagery according to claim 2, characterized in that, The specific steps for identifying short-term meteorological stability observation windows in the correlation analysis of topsoil migration data and vegetation attenuation gradient are as follows: The time point for potential meteorological stability triggering moments is determined based on the rate of change of topsoil migration data. The vegetation stress decline trend in the vegetation attenuation gradient is detected based on the time node of the potential meteorological stability trigger moment. When the vegetation stress decline trend matches the preset ecological restoration response data with a preset similarity coefficient, the regional convergence stability of the vegetation attenuation gradient is detected, and the monitoring unit is confirmed to have entered the short-term meteorological stability observation window based on the regional convergence stability.

5. The method for monitoring soil erosion in grassland areas based on aerial imagery according to claim 4, characterized in that, Based on the time node of the potential meteorological stability triggering moment, the vegetation stress decline trend in the vegetation attenuation gradient is specifically detected as follows: Based on the time node of the potential meteorological stability trigger moment, extract the data sequence of continuous vegetation index change points before and after the corresponding moment from the vegetation decay gradient. Analyze the directionality of vegetation index changes in the data series and identify the recovery period when the vegetation index changes from a decline to a slow recovery. The vegetation stress decline trend during the vegetation stress decline process can be judged based on the persistence characteristics of vegetation index changes during the recovery period.

6. The method for monitoring soil erosion in grassland areas based on aerial imagery according to claim 4, characterized in that, When the matching degree between the vegetation stress decline trend and the preset ecological restoration response data reaches the preset similarity coefficient, the regional convergence stability of the vegetation attenuation gradient is detected, and the monitoring unit is confirmed to have entered the short-term meteorological stability observation window based on the regional convergence stability. Specifically: When the vegetation stress decline trend matches the preset ecological restoration response data with the preset similarity coefficient, the continuous change trend of the vegetation attenuation gradient is monitored. When the vegetation attenuation gradient shows a trend of slow convergence from fluctuation, the corresponding convergence and stabilization trend is generated. The short-term stable range of vegetation decline is determined based on the duration and change pattern corresponding to the convergent stable trend. The regional convergence stability is confirmed by the direction of the vegetation attenuation gradient fluctuation in the short-term stable interval, and the monitoring unit is confirmed to have entered the short-term meteorological stability observation window by the regional convergence stability.

7. The method for monitoring soil erosion in grassland areas based on aerial imagery according to claim 6, characterized in that, The regional convergence stability is determined by the direction of vegetation attenuation gradient fluctuations within the short-term stable interval, and the regional convergence stability is used to confirm whether the monitoring unit has entered the short-term meteorological stability observation window. Specifically: The positive and negative rate fluctuations of the convergence rate can be identified by the direction of the vegetation decay gradient fluctuation in the short-term stable interval. The convergence stability of the region is confirmed based on the changing frequencies of positive and negative velocity fluctuations. The convergence trend of regional convergence stability is used to determine the convergence indicator of change to determine whether the change tends to stagnate. By using change convergence indicators to match meteorological stability trigger conditions, it can be confirmed whether the monitoring unit has entered the short-term meteorological stability observation window.

8. The method for monitoring soil erosion in grassland areas based on aerial imagery according to claim 7, characterized in that, The positive and negative velocity fluctuations of the convergence rate are identified based on the direction of the vegetation attenuation gradient fluctuation in the short-term stable interval. Numerical sequences of vegetation decay gradients at continuous time points are extracted within a short-term stable interval. By performing a difference operation on the convergence rates of adjacent time points in the numerical sequence, a sequence of convergence rate changes is obtained. Based on the sign of the values ​​in the convergence rate change sequence, positive changes are marked as positive rate fluctuations, and negative changes are marked as negative rate fluctuations. By statistically analyzing the frequency and distribution patterns of positive and negative rate fluctuations within the short-term stable interval, we can identify the positive and negative rate fluctuations of the convergence rate.

9. The method for monitoring soil erosion in grassland areas based on aerial imagery according to claim 7, characterized in that, The convergence stability of the region is confirmed based on the changing frequencies of positive and negative velocity fluctuations, specifically as follows: Based on the frequency of occurrence of positive and negative rate fluctuations within the short-term stable range, the frequency values ​​of positive and negative fluctuations are obtained. The ratio of the positive fluctuation frequency value to the negative fluctuation frequency value is used to obtain the fluctuation frequency balance coefficient. When the fluctuation frequency balance coefficient is within the preset balance range, the convergence region of the monitoring unit is determined to be bidirectional balanced fluctuation. Extract the characteristics of bidirectional equilibrium fluctuations and confirm the regional convergence stability based on the duration of the bidirectional equilibrium fluctuation characteristics.

10. A soil erosion monitoring system for grassland areas based on aerial imagery, characterized in that, For performing the method for monitoring soil erosion in grassland areas based on aerial images as described in claim 1, the system for monitoring soil erosion in grassland areas based on aerial images comprises: The image acquisition module is used to divide the grassland area to be monitored into multiple monitoring unit grids; under different meteorological conditions, it acquires spectral feature images of at least three time points in each monitoring unit grid through aerial photography, and records the corresponding vegetation cover parameters during the acquisition period. The feature fuzzy recognition module is used to track the changes in soil exposure of each monitoring unit at the cross-phase based on spectral feature images and vegetation cover parameters, and generate the first type of erosion features; The feature precise identification module is used to obtain the second type of erosion features based on the distribution of the first type of erosion features in terms of spatial location and variation range. The identification module is used to combine the first type of erosion features and the second type of erosion features to identify the characteristics of grassland soil erosion. The erosion analysis module is used to compare the characteristics of grassland soil erosion at different time periods in each monitoring unit grid, and to determine the ecological environment erosion situation based on the comparison results.

Images