An ecological restoration area identification method and system, and a computer device

Abstract

This invention relates to a method, system, and computer equipment for identifying ecological restoration zones. The method includes: acquiring land cover data and vegetation index time series data for each pixel within a study area; acquiring pixels that meet preset land cover change characteristics based on the land cover data to obtain a first type of ecological restoration zone based on land cover change; extracting the change characteristics of the vegetation index time series data for each pixel to acquire pixels that meet preset vegetation index change conditions to obtain a second type of ecological restoration zone based on vegetation index change; and making a comprehensive judgment based on preset ecological restoration zone comprehensive judgment rules, combining the first type of ecological restoration zone and the second type of ecological restoration zone to obtain the ecological restoration zone identification result. The method provided in this application can effectively overcome the problems of missed detection in gradual restoration and misjudgment of land use type change in single methods, thereby improving the accuracy, completeness, and reliability of the identification results.

Description

Technical Field

[0001] This invention relates to the technical field of ecological environment monitoring and remote sensing information processing, and in particular to a method, system and computer equipment for identifying ecological restoration areas. Background Technology

[0002] With global ecological degradation becoming increasingly prominent, ecological restoration has become a crucial measure for regional sustainable development and ecological security. Scientifically and accurately identifying the spatiotemporal locations of ecological restoration is a key prerequisite for objectively assessing the effectiveness of ecological protection and restoration and optimizing ecological governance strategies. Therefore, developing technical methods capable of stably and reliably identifying ecological restoration zones under large-scale, long-term time-series conditions has become an urgent need in this field.

[0003] Currently, remote sensing identification of ecological restoration areas mainly relies on two types of methods. The first type is based on land cover change, whose identification results heavily depend on the accuracy of land cover classification data. Accumulated classification errors over time directly affect the reliability of the results. Furthermore, it typically relies on explicit type transition thresholds and is insensitive to areas where vegetation is in a slow recovery phase but has not yet reached the land cover type transition criteria (i.e., "gradual recovery"), leading to an underestimation of the actual recovery area. The second type is based on the time-series evolution characteristics of vegetation indices. However, vegetation indices primarily reflect the growth status of green vegetation and cannot effectively distinguish the land use types they support. For example, changes in vegetation indices caused by human agricultural activities such as farmland reclamation and fallow may exhibit similar characteristics to the ecological restoration process, leading to misjudgments. In addition, vegetation indices are susceptible to interference from interannual climate fluctuations and short-term disasters, also introducing uncertainty into the results.

[0004] Therefore, existing methods that rely solely on land cover change or vegetation index evolution have limitations, making it difficult to simultaneously guarantee the accuracy and completeness of identification results in large-scale, long-term ecological restoration monitoring. Summary of the Invention

[0005] Based on this, the purpose of this invention is to provide a method, system, and computer device for identifying ecological restoration areas. By integrating land cover change determination and vegetation index time-series evolution characteristics, a comprehensive determination rule is constructed to accurately identify ecological restoration areas in large-scale regions. This effectively overcomes the problems of missed identification of gradual restoration and misjudgment of land use type changes that exist in single methods, thereby improving the accuracy, completeness, and reliability of the identification results.

[0006] Firstly, this application provides a method for identifying ecological restoration zones, including:

[0007] Acquire land cover data and vegetation index time series data for each pixel within the study area; wherein, the land cover data includes the land cover type of each pixel and the ecological land use type corresponding to the land cover type, and the ecological land use type includes non-ecological land and ecological land; Based on the land cover data, pixels that meet the preset land cover change characteristics are obtained to obtain the first type of ecological restoration area based on land cover change; Extract the variation characteristics of the vegetation index time series data of each pixel, obtain the pixels that meet the preset vegetation index variation conditions, and obtain the second type of ecological restoration area based on vegetation index variation. Based on the preset comprehensive judgment rules for ecological restoration areas, the first type of ecological restoration area and the second type of ecological restoration area are combined for comprehensive judgment to obtain the ecological restoration area identification result.

[0008] Secondly, this application provides an ecological restoration zone identification system, comprising: The data acquisition module is used to acquire land cover data and vegetation index time series data for each pixel within the study area; wherein, the land cover data includes the land cover type of each pixel and the ecological land use type corresponding to the land cover type, and the ecological land use type includes non-ecological land and ecological land; The first type of ecological restoration area analysis module is used to obtain pixels that meet the preset land cover change characteristics based on the land cover data, and obtain the first type of ecological restoration area based on land cover change. The second type of ecological restoration area analysis module is used to extract the change characteristics of the vegetation index time series data of each pixel, obtain the pixels that meet the preset vegetation index change conditions, and obtain the second type of ecological restoration area based on the vegetation index change. The ecological restoration area identification module is used to make a comprehensive judgment based on the preset ecological restoration area comprehensive judgment rules, combining the first type of ecological restoration area and the second type of ecological restoration area, to obtain the ecological restoration area identification result.

[0009] Thirdly, this application provides a computer device including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the ecological restoration zone identification method as described above.

[0010] This application provides a method, system, and computer equipment for identifying ecological restoration zones. By analyzing the land cover data and vegetation index time-series data characteristics of each pixel within the research area, it identifies two types of ecological restoration zones: a first type based on land cover change and a second type based on vegetation index change. Then, based on comprehensive ecological restoration zone judgment rules, it fuses and verifies the first and second types of ecological restoration zones, outputting high-precision ecological restoration zone identification results. Compared to existing technologies, the technical solution provided in this application, by fusing and judging the characteristics of land cover change and the time-series changes in vegetation indices, achieves more comprehensive, accurate, and stable automated identification of ecological restoration zones under large-scale, long-term time-series conditions. This significantly improves the objectivity and reliability of ecological restoration monitoring and assessment, providing strong technical support for the scientific evaluation and decision-making of ecological protection and restoration effectiveness. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A flowchart illustrating the steps of an ecological restoration zone identification method provided in this application embodiment; Figure 2 A flowchart illustrating the steps of a vegetation index-based identification method provided in this application embodiment; Figure 3 A flowchart illustrating the steps for assessing an ecological restoration zone is provided in this application embodiment; Figure 4 A flowchart illustrating the steps for analyzing the contribution of driving factors is provided in this application embodiment; Figure 5 This is a schematic diagram of the structure of an ecological restoration area identification system provided in an embodiment of this application; Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the protection scope of this application.

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

[0015] In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims. In the description of this application, it should be understood that the terms "first," "second," "third," etc., are used only to distinguish similar objects and are not necessarily used to describe a specific order or sequence, nor should they be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0016] Furthermore, in the description of this application, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0017] Currently, remote sensing identification of ecological restoration areas mainly relies on two types of methods: The first category is identification methods based on land cover change. This method analyzes multiple land cover classification maps to directly identify areas where non-ecological land such as arable land and construction land has transformed into ecological land such as forest, shrubland, and grassland, and classifies these areas as ecological restoration zones. This method is logically intuitive and relatively efficient. However, its identification results heavily depend on the accuracy of the land cover classification data; the accumulation of classification errors over time directly affects the reliability of the results. More importantly, this method typically relies on explicit type transition thresholds and is insensitive to areas where vegetation is in a slow recovery phase but has not yet met the land cover type transition criteria (i.e., "gradual recovery"), leading to an underestimation of the actual recovery area.

[0018] The second category is identification methods based on the time-series evolution characteristics of vegetation indices. This method analyzes the changing trends or abrupt changes of indicators such as the Normalized Difference Vegetation Index (NDVI) over long time series to detect the continuous improvement of vegetation growth, thereby identifying potential ecological restoration areas. This method is relatively effective for gradual restoration processes. However, vegetation indices primarily reflect the growth status of green vegetation and cannot effectively distinguish the land use types they support. For example, changes in vegetation indices caused by human agricultural activities such as farmland reclamation and fallow may exhibit similar characteristics to ecological restoration processes, leading to misjudgments. Furthermore, vegetation indices are susceptible to interference from interannual climate fluctuations and short-term disasters, also introducing uncertainty into the results.

[0019] To this end, this application provides a method for identifying ecological restoration areas. By acquiring land cover data and vegetation index time series data of each pixel, analyzing their change characteristics, and fusing and verifying the preliminary analysis results of the two, the final ecological restoration area identification result is obtained, which can significantly improve the accuracy and reliability of identification.

[0020] Please see Figures 1 to 4 This application provides a method for identifying ecological restoration areas, including: S101, acquire land cover data and vegetation index time series data for each pixel within the study area; wherein, the land cover data includes the land cover type of each pixel and the ecological land use type corresponding to the land cover type, and the ecological land use type includes non-ecological land and ecological land; S102, Based on the land cover data, obtain pixels that meet the preset land cover change characteristics to obtain the first type of ecological restoration area based on land cover change; S103, extract the change characteristics of the vegetation index time series data of each pixel, obtain the pixels that meet the preset vegetation index change conditions, and obtain the second type of ecological restoration area based on vegetation index change. S104. Based on the preset comprehensive judgment rules for ecological restoration areas, a comprehensive judgment is made by combining the first type of ecological restoration area and the second type of ecological restoration area to obtain the ecological restoration area identification result.

[0021] This application provides a method for identifying ecological restoration zones. By analyzing the characteristics of land cover data and vegetation index time-series data of each pixel within the research area, it identifies two types of ecological restoration zones: a first type based on land cover change and a second type based on vegetation index change. Then, based on a comprehensive ecological restoration zone determination rule, the method combines and verifies the first and second types of ecological restoration zones to output a high-precision ecological restoration zone identification result. The technical solution provided in this application, by fusing and judging the characteristics of land cover change and the time-series changes in vegetation indices, achieves more comprehensive, accurate, and stable automated identification of ecological restoration zones under large-scale, long-term time-series conditions. This significantly improves the objectivity and reliability of ecological restoration monitoring and assessment, providing strong technical support for the scientific evaluation and decision-making of ecological protection and restoration effectiveness.

[0022] For step S101, land cover data and vegetation index time series data for each pixel within the study area are obtained.

[0023] The pixel is the smallest unit and basic sampling point that constitutes a remote sensing digital image or other digital image. Each pixel corresponds to a certain area of ​​the ground, and its data value determines the size of the ground area it represents. For example, when the resolution of a remote sensing image is 30 meters, one pixel represents an area of ​​30 meters × 30 meters on the ground.

[0024] The land cover data is time-series data based on time changes, for example, using years as the unit, combined with annual land cover data of the study area for several consecutive years to obtain land cover data.

[0025] The land cover data includes the land cover type of each pixel and the corresponding ecological land use type. Specifically, the land cover data includes the time series data of the land cover type of each pixel in each year and the corresponding time series data of the ecological land use type.

[0026] In one embodiment, the land cover type includes arable land, impermeable surface, forest, shrubland, grassland, etc., and the ecological land type is determined according to the land cover type.

[0027] The types of ecological land use include ecological land and non-ecological land. In one embodiment, arable land and impermeable surfaces correspond to non-ecological land, while forests, shrublands, and grasslands correspond to ecological land.

[0028] The vegetation index time series data is a set of several vegetation index values ​​collected at fixed time intervals and arranged chronologically. The collection interval of the vegetation index can be set to the same interval as that of the land cover data, and the collection starts at the same start time, thereby aligning and synchronizing the time data. The vegetation index is an indicator used to qualitatively or quantitatively reflect the growth status of surface vegetation, obtained by combining and calculating the reflectance of different bands using remote sensing sensors. For example, it can be the normalized vegetation index (NVDI) of a region.

[0029] In this embodiment, by acquiring land cover data and vegetation index time series data for each pixel within the study area, analytical reference data is provided for the dual-parameter fusion identification of the ecological restoration area, thus providing effective data support.

[0030] For step S102, based on the land cover data, pixels that meet the preset land cover change characteristics are obtained to obtain the first type of ecological restoration area based on land cover change.

[0031] In one embodiment, step S102 includes: Pixels whose ecological land use type changes from non-ecological land use to ecological land use within a first time window of a preset length, and remain as ecological land use for the remaining time of the first time window after the change, are identified as the first type of ecological restoration area, and the first change time of the ecological land use type change is recorded.

[0032] To prevent misjudgments due to accidental errors or short-term disturbances in the land cover data (such as the impact of natural disasters or human intervention), a first time window of preset length is introduced. The changes in the ecological land use type of the pixel within the first time window are analyzed to determine whether they meet the characteristics of land cover change, thereby identifying whether the pixel belongs to an ecological restoration zone. In one embodiment, the length of the first time window is in years, and can be set to 5 years, 10 years, etc.

[0033] In one embodiment, a land cover state function is defined as follows:

[0034] Define the length of the first time window as W. If the pixel is in a certain year within the first time window... Non-ecological land, i.e. In the second year Transformed into ecological land, that is Then, if the ecological land use type remains ecological land use for the remaining years of the first time window, it is determined that the pixel has experienced an ecological restoration event and belongs to the first type of ecological restoration area.

[0035] In this embodiment, by introducing a time window to detect changes in regional land cover types over a certain period of time, the problem of inaccurate identification caused by remote sensing classification errors and land use fluctuations can be effectively filtered out.

[0036] For step S103, the change characteristics of the vegetation index time series data of each pixel are extracted, and pixels that meet the preset vegetation index change conditions are obtained to obtain the second ecological restoration area based on the vegetation index change.

[0037] In one embodiment, step S103 includes: S301, perform segmented fitting on the vegetation index time series data of each pixel to obtain a vegetation index time series of several consecutive time periods.

[0038] By using a segmented fitting method, the vegetation index time series data, which consists of several consecutive years, is divided into several consecutive time periods. The length of each time period can be the same preset value or different lengths.

[0039] Vegetation growth is affected by a variety of factors such as climate and human intervention. The vegetation index of a region does not rise or fall linearly, but fluctuates. In one embodiment, the vegetation index time series can be segmented according to the time of fluctuation, thereby dividing it into several time periods of different durations.

[0040] For each time period, the vegetation index values ​​within that time period are fitted using a linear fitting method to obtain the vegetation index change characteristics for each time period.

[0041] In one embodiment, the vegetation index time series data is represented as follows:

[0042] in, Let be the vegetation index value of the i-th pixel at time t.

[0043] S302, calculate the slope of vegetation index change for each time period.

[0044] A linear fit is performed on the vegetation index within the time period, and the slope of the fitted line is calculated to obtain the slope of the vegetation index change corresponding to each time period. The slope of the vegetation index change is used to quantify the rate and direction of the vegetation index change over time within the time period, which can determine whether the ecology is recovering and the speed of recovery within that time period.

[0045] In one embodiment, the time interval fitted piecewise to a straight line can be represented by a piecewise linear function:

[0046] in, Let be the slope of the vegetation index change in the j-th time period. This represents the intercept for the corresponding time period. Let j be the time range of the j-th time period.

[0047] S303, obtain the pixels where the slope of the vegetation index change is greater than a preset first slope threshold and the time period is greater than a preset first time length, obtain the second type of ecological restoration area, and record the start time of the time period as the second transition time.

[0048] The slope of the vegetation index change corresponding to each time period is compared with a preset first slope threshold. The first slope threshold can be set according to actual conditions and can be used as a criterion for judging the ecological restoration status. When the slope of the vegetation index change is greater than the first slope threshold, the ecological restoration status of the area can be considered good, and the judgment is not caused by small-scale fluctuations. The duration of the time period must also be greater than a preset first time length to further confirm that the pixel is undergoing continuous ecological restoration and avoid data interference caused by short-term fluctuations.

[0049] Pixels that simultaneously satisfy the conditions of a vegetation index change slope greater than a preset slope threshold and a time period duration greater than a preset time length are identified as the second type of ecological restoration area. The start time of the time period is recorded as the start time of ecological restoration and is used as the second transition time.

[0050] By further identifying the temporal variation characteristics of vegetation indices, we can identify areas where the ecological land use type has not changed but is in the process of restoration, thus obtaining the identification status of the second type of ecological restoration area.

[0051] In this embodiment, by analyzing the time series data of vegetation index, the changing characteristics of vegetation index are obtained, which can identify areas where vegetation biomass, coverage, or vitality is continuously and significantly improving, but may not yet meet the land cover type conversion standard. This effectively makes up for the defect that the first type of ecological restoration area obtained based on land cover data in step S102 above may be missed.

[0052] For step S104, based on the preset comprehensive judgment rules for ecological restoration areas, a comprehensive judgment is made by combining the first type of ecological restoration area and the second type of ecological restoration area to obtain the ecological restoration area identification result.

[0053] The comprehensive determination rule for ecological restoration areas is used to integrate the ecological restoration area identification results obtained by two methods: land cover data characteristics and vegetation index change characteristics, to obtain the final ecological restoration area identification result.

[0054] In one embodiment, step S104 includes: S401, obtain a pixel that belongs to both the first type of ecological restoration area and the second type of ecological restoration area as the first pixel, determine the first pixel as an ecological restoration area, and classify the ecological restoration area type according to the land cover type of the pixel.

[0055] The first pixel that is simultaneously identified as a first-class ecological restoration area and a second-class ecological restoration area in terms of spatial location is obtained. The first pixel is judged as an ecological restoration area in the comprehensive judgment based on the change characteristics of land cover data and the change characteristics of vegetation index. Therefore, the area of ​​the first pixel is directly judged as an ecological restoration area.

[0056] In one embodiment, the ecological restoration zone type is determined based on the land cover type of the first pixel. For example, referring to the above embodiment, the land cover type in ecological land may include forest, shrubland, and grassland, and the ecological restoration zone type may correspond to forest restoration zone, shrubland restoration zone, and grassland restoration zone.

[0057] S402, if the pixel belongs to the first type of ecological restoration area but not the second type of ecological restoration area, obtain the net primary productivity time series of the pixel.

[0058] Vegetation indices may respond poorly to certain restoration types (such as slow-growing tree species and early-stage vegetation), or there may be noise in the vegetation index time series. To distinguish between "true restoration but undetected vegetation indices" and "land cover misclassification or non-ecological restoration transitions," a net primary productivity time series is introduced for verification.

[0059] Net primary productivity (NPP) refers to the total organic matter produced by green plants through photosynthesis per unit time and per unit area (i.e., gross primary productivity GPP), minus the respiration (R) consumed by the plants themselves. It directly reflects the productivity and carbon sequestration capacity of a vegetation community under natural conditions and is a key indicator for measuring the health of ecosystem structure and function.

[0060] S403, based on the net primary productivity time series, obtain the slope of net primary productivity change within a second time window of a preset length, retain the pixels whose net primary productivity change slope is greater than a preset second slope threshold as second pixels, and determine the second pixels as ecological restoration areas.

[0061] Obtain the net primary productivity data of the cell within a second time window of a preset length, and calculate its change slope. If the change slope of net primary productivity is greater than the preset second slope threshold, it indicates that the ecosystem productivity of the region has changed significantly after the land cover type change, and the second cell is retained as an ecological restoration area.

[0062] S404, if the pixel belongs to the second type of ecological restoration area but not to the first type of ecological restoration area, obtain the pixel whose ecological land use type is ecological land as the third pixel, and determine the third pixel as an ecological restoration area.

[0063] For pixels belonging to the second type of ecological restoration zone but not the first type, it indicates that their land cover type did not change within the first time window, or changed again due to various factors after the initial change. For slow or gradual restoration, the plants are undergoing restoration. The ecological land use type of the pixel is identified. If its ecological land use type is already ecological land, it indicates that the vegetation improvement is occurring on the ecological baseline and is a true ecological restoration. The pixel is retained as a third pixel. If it is non-ecological land, the increase in the vegetation index may be caused by human agricultural activities, and such pixels are removed.

[0064] S405, if the pixel does not belong to either the first type of ecological restoration area or the second type of ecological restoration area, then the pixel is determined to be a non-ecological restoration area.

[0065] If the pixel does not belong to either the first type of ecological restoration area or the second type of ecological restoration area, it means that the pixel does not meet the criteria for ecological restoration area judgment, whether it is identified based on land cover data or vegetation index. Such pixels are then judged as non-ecological restoration areas.

[0066] In one embodiment, the ecological restoration zone identification results include ecological restoration zone location information, ecological restoration zone type, start time, etc., providing an accurate data foundation for scientific research.

[0067] This embodiment combines the identification results of the first type of ecological restoration area and the second type of ecological restoration area to construct a fusion identification system for ecological restoration area identification. Multiple logical cross-validation is performed on the two types of results, which effectively reduces the misjudgment rate of ecological restoration area identification and improves the identification accuracy and precision.

[0068] The embodiments of this application can accurately locate ecological restoration areas through the above-described method for identifying ecological restoration areas. Based on the inventive concept of this application, this application further analyzes the restoration status of ecological restoration areas.

[0069] To avoid assessment biases introduced by differences in the scale and spatial structure of ecological land use bases when comparing different spatial units solely based on ecological restoration area or the number of restored pixels, this application introduces an ecological restoration density index to quantitatively characterize the intensity of ecological restoration. In one embodiment, after obtaining the final ecological restoration area identification result, the following step is also included: S501, Based on the ecological restoration zone identification results, the ecological restoration density of the ecological restoration zone is calculated using the ecological restoration density formula.

[0070] Ecological restoration density (ERD) is an indicator used to quantify the intensity of ecological restoration in a specific area. It is characterized by comparing the relative proportion of the number of pixels in the ecological restoration area to the total number of pixels in the ecological land area.

[0071] In one embodiment, the ecological restoration density formula is:

[0072] in, Let be the ecological restoration density of subregion j. This represents the area of ​​a single grid cell within the ecological restoration area. This refers to the area of ​​a single grid cell within ecological land use. and These represent the number of pixels in the ecological restoration area and the number of pixels in the ecological land use area for year i within sub-region j.

[0073] The sub-regions are obtained by dividing the region according to preset regional division rules, which can be formulated according to research objectives and actual needs. In some embodiments, the division of the sub-regions may include: administrative unit division (e.g., province, city, county, etc.), natural geographical unit division (e.g., watershed, ecological functional zone, geomorphic unit, etc.), and land cover type, etc.

[0074] The higher the ecological restoration density value, the stronger the ecological restoration capacity of the sub-region or the wider the restoration range.

[0075] S502, based on the ecological restoration area identification results, at least one landscape index is obtained using a preset landscape index calculation formula.

[0076] To further characterize the ecological restoration process from a spatial structure perspective, a landscape index is introduced to quantitatively represent the morphological and spatial distribution characteristics of the ecological restoration area, supplementing spatial organization information that is difficult to reflect based solely on quantity or density indicators.

[0077] In one embodiment, the landscape index includes at least the area-weighted patch fractal dimension (FRAC) and the mean Euclidean nearest neighbor distance (ENN).

[0078] The formula for calculating the fractal dimension of the area-weighted patch is:

[0079] The area-weighted patch fractal dimension is used to describe the complexity of the boundaries of ecological restoration patches. Let fractal dimension be the area-weighted patch dimension of subregion j. Let be the area of ​​the i-th ecological restoration patch within sub-region j. Let be the perimeter of the i-th ecological restoration patch within sub-region j, and n be the number of ecological restoration patches within sub-region j.

[0080] This index is used to quantify the complexity and stability of the shape of ecological restoration patches. Theoretically, its value is between 1 and 2. The larger the value, the more tortuous the patch boundary and the more complex the shape, which usually means less regular human interference and a more natural ecological process.

[0081] The formula for calculating the average Euclidean nearest neighbor distance is:

[0082] in, Let be the average Euclidean nearest neighbor distance for ecological restoration patches in sub-region j. Let be the total area of ​​the ecological restoration patches within sub-region j. Let be the area of ​​the i-th ecological restoration patch within sub-region j. Let be the Euclidean distance between the geometric center of the i-th ecological restoration patch within sub-region j and the geometric center of the nearest ecological restoration patch.

[0083] The average Euclidean nearest neighbor distance is used to characterize the spatial clustering or connection features between ecological restoration patches.

[0084] This index is used to quantify the spatial clustering and dispersion of ecological restoration patches. It is the area-weighted average of the distances from each patch to its nearest neighbor. A smaller ENN value indicates a more compact distribution of patches and potentially better connectivity; a larger value indicates more isolated and dispersed patches.

[0085] S503, using a preset formula for calculating the comprehensive ecological restoration index, and combining the ecological restoration density and the landscape index, the comprehensive ecological restoration index is obtained.

[0086] The comprehensive ecological restoration index integrates the above-mentioned multi-dimensional indicators into a single comprehensive score, realizing a single quantitative representation and ranking of the regional ecological restoration level, and providing a clear priority basis for management decisions.

[0087] In one embodiment, the formula for calculating the comprehensive ecological restoration index is:

[0088] in, Let be the comprehensive ecological restoration index for sub-region j. Let be the ecological restoration density of subregion j. Let fractal dimension be the area-weighted patch dimension of subregion j. Let be the average Euclidean nearest neighbor distance for ecological restoration patches in sub-region j. and These are the preset weighting coefficients.

[0089] This embodiment is used to assess the restoration status of ecological restoration areas. By using ERD, the spatial scope of restoration is transformed into standardized intensity indicators, making restoration efforts between areas with different background conditions comparable. By using FRAC and ENN, the spatial pattern dimension is introduced into regional assessment for the first time to further assess the restoration status of ecological restoration areas and effectively improve the efficiency of ecological restoration planning by relevant personnel.

[0090] According to the inventive concept of this application, after step S603, the following step is also included: S601, acquire at least one type of driving factor data related to ecological restoration; wherein each type of driving factor data includes at least one driving factor.

[0091] Among them, driving factors refer to natural environmental conditions and human activities that may influence the occurrence, rate, or spatial pattern of ecological restoration.

[0092] In one embodiment, the types of driving factor data may include four major categories: meteorological, soil, topographic, and socioeconomic.

[0093] In one embodiment, the meteorological driving factor data includes at least one or more of the following driving factors: 1. Rainfall (PPT) Precipitation is a critical water source limiting vegetation growth and ecological restoration, directly affecting soil moisture content, vegetation transpiration, and primary productivity. In arid or semi-arid regions, precipitation variation is often the decisive factor in the success or failure of ecological restoration. PPT is a cumulative meteorological variable, using the summation of monthly precipitation to obtain the annual total precipitation.

[0094]

[0095] Where y represents the year and m represents the month.

[0096] 2. Potential evaporation (PET) PET (Peak Evapotranspiration Requirement) characterizes the maximum evapotranspiration demand under unconstrained water conditions, reflecting the potential drought level of a region. A higher PET generally indicates greater water stress on vegetation, hindering the sustainable progress of ecological restoration. PET is also a cumulative variable, and is statistically analyzed using an annual-scale summation method.

[0097] Where y represents the year and m represents the month.

[0098] 3. Vapor pressure deficit (VPD) VPD reflects the atmosphere's evaporative demand for water and is an important indicator for measuring vegetation transpiration stress. Higher VPD inhibits stomatal conductance, thus limiting photosynthetic efficiency. VPD is a state variable, and its statistics are based on annual averages.

[0099] Where y represents the year and m represents the month.

[0100] 4. Wind speed (WS) Wind speed indirectly affects ecological restoration processes by enhancing evapotranspiration and influencing vegetation structure stability, particularly in coastal and arid regions. WS is a state variable, and its statistical value is calculated using annual averages.

[0101] Where y represents the year and m represents the month.

[0102] 5. Average temperature (TEMP) Temperature is an important factor controlling the growth cycle and physiological activities of vegetation, and has a significant impact on photosynthetic rate and phenological processes.

[0103]

[0104] Where y represents the year and m represents the month.

[0105] In one embodiment, the driving factor data for soil types includes at least one or more of the following driving factors: 1. Soil texture Soil texture determines the soil's water retention, aeration, and nutrient retention capacity, thus affecting vegetation establishment and growth conditions. Soil texture is expressed as a percentage, and each type meets the following criteria:

[0106] Among them, Sand represents sand content, Silt represents silt content, and Clay represents clay content.

[0107] 2. Soil Moisture (SM) Soil moisture directly reflects the availability of water resources for vegetation and serves as an important bridge connecting climate conditions and vegetation responses.

[0108]

[0109] Where y represents the year and m represents the month.

[0110] In one embodiment, the driving factor data for the terrain class includes at least one or more of the following driving factors: 1. Elevation Elevation exerts fundamental constraints on the spatial pattern of ecological restoration by regulating temperature gradients, precipitation distribution, and the intensity of human activities.

[0111] 2. Slope Slope affects the intensity of soil erosion and land use patterns, indirectly restricting the sustainability of ecological restoration, as calculated from elevation data:

[0112] In one embodiment, the socioeconomic driving factor data includes at least one or more of the following driving factors: 1. Ecological Engineering (EP) Ecological engineering accelerates vegetation restoration through human intervention and is an important manifestation of policy-driven ecological restoration. Its forms include:

[0113] 2. Gross Domestic Product (GDP) GDP reflects the level of regional economic development and indirectly affects land use intensity and the capacity for ecological restoration. Its spatial distribution expression is as follows:

[0114] Where k represents an administrative unit.

[0115] 3. Point of Impermeability (POI) The proportion of impermeable surfaces reflects the degree of urbanization and is usually positively correlated with ecological degradation pressure. The calculation formula is as follows:

[0116] S602, Standardize the driving factor data.

[0117] Since the driving factors have different dimensions and numerical ranges (e.g., precipitation is measured in millimeters, GDP in billions of yuan, and slope in degrees), directly inputting them into the model would cause the model training to be affected by the dimensions. In this embodiment, each driving factor is standardized. The purpose of the standardization process is to eliminate the differences in dimensions and scales, transforming all driving factors into a unified, dimensionless numerical range.

[0118] In one embodiment, each driving factor is standardized using an extreme value normalization method, and the calculation formula is as follows:

[0119] in, The original observation value of the driving factor on the i-th sample. This represents the minimum sample value of the driving factor in the entire training dataset. This represents the maximum sample value of the driving factor in the entire training dataset. This is the value after standardization.

[0120] This embodiment standardizes the driving factor data to eliminate dimensional differences, providing a reliable data foundation for subsequent model analysis.

[0121] S603, The standardized driving factor data and the ecological restoration comprehensive index are input into a preset machine learning model for training to construct a driving factor analysis model.

[0122] In one embodiment, the machine learning model is an XGBoost model, which can process multi-source nonlinear data such as meteorological, soil, topographic, and socioeconomic data, quantitatively characterize the impact of each driving factor on the comprehensive ecological restoration index, and identify key limiting and promoting factors.

[0123] First, the eXtreme Gradient Boosting (XGBoost) algorithm is used to establish a nonlinear mapping relationship between the comprehensive ecological restoration index and multidimensional driving factors. XGBoost is an ensemble learning method based on the gradient boosting framework. It iteratively constructs multiple regression decision trees and integrates them in an additive model to achieve high-precision fitting of complex nonlinear relationships. During model training, the model parameters are optimized by minimizing the objective function, which consists of a loss term and a regularization term, to improve prediction accuracy while suppressing model complexity and reducing the risk of overfitting. By introducing the first and second gradient information of the loss function, the model's convergence efficiency and computational stability are improved.

[0124] In this application, the XGBoost algorithm is used to construct a regression model between the Ecological Restoration Index (ERI) and multidimensional driving factors. XGBoost trains and optimizes the model by minimizing the overall objective function, which is defined as:

[0125] in, This represents the true normal recovery composite index value of the i-th spatial unit; This represents the comprehensive ecological restoration index value of the prediction model; This is the loss function, used to measure the error between the predicted and actual values. Indicates the first A regression tree; This is a regularization term used to constrain model complexity.

[0126] In regression problems, the loss function The squared error form is usually used, and its expression is:

[0127] Regularization term Defined as:

[0128] in, Indicates the first The number of leaf nodes in the regression tree; Indicates the first The weight values ​​of each leaf node. and This is the regularization coefficient, used to balance the model's fitting ability and complexity.

[0129] During model training, XGBoost uses an additive model approach to progressively construct the prediction function:

[0130] in, Indicates the first The regression tree generated in rounds of iterations is used to analyze the samples. The prediction results.

[0131] To improve optimization efficiency, this invention performs a second-order Taylor expansion of the objective function in each iteration, utilizing the first-order gradient of the loss function. With second gradient The objective function approximates the current round:

[0132] in:

[0133] In this way, XGBoost can improve the model's ability to fit the complex nonlinear relationship between ecological restoration driving factors and the comprehensive ecological restoration index while ensuring computational efficiency.

[0134] S604. Based on the driving factor analysis model, a preset machine learning model interpretation method is used to analyze the contribution of the driving factor data to the comprehensive ecological restoration index.

[0135] In one embodiment, the machine learning model interpretation method is the SHAP method. The SHAP interpretation method can quantify the positive or negative impact of each factor (such as annual precipitation, slope, GDP) on the recovery process in each specific region, and rank their importance.

[0136] After completing model training, to improve the interpretability of the model results, this invention further introduces the Shapley Additive exPlanations (SHAP) method to interpret and analyze the XGBoost model. Based on Shapley value theory in game theory, SHAP quantifies the marginal contribution of each input variable to the model data results, thus enabling interpretability analysis of complex machine learning models. By calculating the SHAP value distribution of each driving factor, its positive or negative impact on the comprehensive ecological restoration index can be identified, and the relative importance of different factors in the overall prediction can be assessed. The different driving factors can be denoted as... Explanation Model It can be represented as:

[0137] Where T is the total number of driving factor types, This represents the Shapley value of the i-th driving factor. This represents a constant term when all input variables take null values.

[0138] Shapley value The calculation method is as follows: in all subsets of bad driving factors i In this process, the marginal contribution of driving factor i (i.e., the difference in model output with and without driving factor i) is calculated and averaged.

[0139]

[0140] Where f is the original prediction model, S is the feature subset, N is the set of all features, i represents a certain feature driving factor, and n is the total number of driving factors. This represents the set of driving factors that does not contain driving factor i.

[0141] By calculating the average absolute value of the SHAP value of each driving factor, the global importance ranking of the influence of each driving factor on the ERI value is obtained, and the key driving factors are identified.

[0142] In this embodiment, the XGBoost model can effectively analyze the complex, nonlinear, and potentially interactive mapping relationships between dozens of factors, such as meteorology, soil, topography, and socioeconomic factors, and the level of ecological restoration. The SHAP method is used to quantify the contribution and direction of each driving factor to the comprehensive ecological restoration index in each specific region and year, thereby identifying key driving factors and improving the scientific nature and effectiveness of ecological restoration work.

[0143] The ecological restoration zone identification method provided in this application acquires long-term land cover data and vegetation index time series data, and identifies the ecological restoration status of each pixel within the study area based on the change characteristics of the two sets of data, thus obtaining a first-class and a second-class ecological restoration zone. Next, according to a comprehensive judgment rule, the preliminary identification results of the two types of ecological restoration zones are logically fused to obtain the ecological restoration zone identification result. Based on this, a comprehensive ecological restoration index integrating restoration intensity and spatial pattern quality is further constructed to achieve a standardized quantitative assessment of the restoration level. Finally, multi-dimensional driving factors are integrated, and the complex nonlinear relationship between them and the restoration level is analyzed using the XGBoost model. The contribution and direction of each factor are quantified using the SHAP method.

[0144] The method provided in this application effectively overcomes the inherent defects of traditional single methods, such as missed detections in gradual restoration and misjudgments of land use type changes, and significantly improves the identification accuracy and reliability. At the same time, the comprehensive index and the quantification results of driving contributions provided offer strong data and theoretical support for the scientific evaluation of ecological restoration effectiveness, the precise formulation of management decisions, and the differentiated configuration of restoration measures, greatly improving the objectivity, automation level, and decision support capabilities of ecological restoration monitoring and evaluation work.

[0145] Secondly, please refer to Figure 5 , Figure 5 This is a schematic diagram of the structure of an ecological restoration area identification system provided in an embodiment of this application.

[0146] This application provides an ecological restoration zone identification system, including: The data acquisition module 11 is used to acquire land cover data and vegetation index time series data of each pixel in the study area; wherein, the land cover data includes the land cover type of each pixel and the ecological land type corresponding to the land cover type, and the ecological land type includes non-ecological land and ecological land; The first type of ecological restoration area analysis module 12 is used to obtain pixels that meet the preset land cover change characteristics based on the land cover data, and obtain the first type of ecological restoration area based on land cover change. The second type of ecological restoration area analysis module 13 is used to extract the change characteristics of the vegetation index time series data of each pixel, obtain the pixels that meet the preset vegetation index change conditions, and obtain the second type of ecological restoration area based on the vegetation index change. The ecological restoration area identification module 14 is used to make a comprehensive judgment based on the preset ecological restoration area comprehensive judgment rules, combining the first type of ecological restoration area and the second type of ecological restoration area, to obtain the ecological restoration area identification result.

[0147] It should be noted that the above-described ecological restoration area identification system, when implementing an ecological restoration area identification method, is only illustrated by the division of the above-described functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. The ecological restoration area identification system provided in the above-described embodiments is used to implement the ecological restoration area identification method described in the above-described embodiments. Its operating method and principle are the same as the ecological restoration area identification method described above. That is, the ecological restoration area identification system and the ecological restoration area identification method provided in the above-described embodiments belong to the same concept. The implementation process is detailed in the above-described method embodiments and will not be repeated here.

[0148] Thirdly, this embodiment provides a computer device. Please refer to [link / reference needed]. Figure 6 , Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Figure 6 As shown, the computer device 21 includes: a processor 210, a memory 211, and a computer program 212 stored in the memory 211 and executable on the processor 210, such as an ecological restoration zone identification program; the processor 210 executes the computer program 212 to implement the methods described in the above embodiments.

[0149] The processor 210 may include one or more processing cores. The processor 210 connects to various parts within the computer device 21 using various interfaces and lines. It executes various functions of the computer device 21 and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 211, and by accessing data in the memory 211. Optionally, the processor 210 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 210 may integrate one or more of the following: a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), and a modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required for the touch screen; and the modem handles wireless communication. It is understood that the modem may also be implemented as a separate chip, rather than integrated into the processor 210.

[0150] The memory 211 may include random access memory (RAM) or read-only memory. Optionally, the memory 211 may include a non-transitory computer-readable storage medium. The memory 211 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 211 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch instructions), instructions for implementing the various method embodiments described above, etc.; the data storage area may store data involved in the various method embodiments described above, etc. Optionally, the memory 211 may also be at least one storage device located remotely from the aforementioned processor 210.

[0151] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0152] Fourthly, embodiments of this application also provide a computer-readable storage medium that can store multiple instructions. These instructions are applicable to being loaded by a processor and executing the method steps of the above embodiments. For details of the execution process, please refer to the specific description of the above embodiments, which will not be repeated here.

[0153] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0154] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. A method for identifying ecological restoration areas, characterized in that, include: Acquire land cover data and vegetation index time series data for each pixel within the study area; wherein, the land cover data includes the land cover type of each pixel and the ecological land use type corresponding to the land cover type, and the ecological land use type includes non-ecological land and ecological land; Based on the land cover data, pixels that meet the preset land cover change characteristics are obtained to obtain the first type of ecological restoration area based on land cover change; Extract the variation characteristics of the vegetation index time series data of each pixel, obtain the pixels that meet the preset vegetation index variation conditions, and obtain the second type of ecological restoration area based on vegetation index variation. Based on the preset comprehensive judgment rules for ecological restoration areas, the first type of ecological restoration area and the second type of ecological restoration area are combined for comprehensive judgment to obtain the ecological restoration area identification result.

2. The method for identifying ecological restoration areas according to claim 1, characterized in that, The step of obtaining pixels that meet preset land cover change characteristics based on the land cover data to obtain the first type of ecological restoration zone based on land cover change includes: Pixels whose ecological land use type changes from non-ecological land use to ecological land use within a first time window of a preset length, and remain as ecological land use for the remaining time of the first time window after the change, are identified as the first type of ecological restoration area, and the first change time of the ecological land use type change is recorded.

3. The method for identifying ecological restoration areas according to claim 1, characterized in that, The step of extracting the change characteristics of the vegetation index time series data of each pixel, obtaining pixels that meet the preset vegetation index change conditions, and obtaining the second type of ecological restoration zone based on vegetation index change includes: The vegetation index time series data of each pixel are segmented and fitted to obtain a vegetation index time series of several consecutive time periods. Calculate the slope of vegetation index change for each of the aforementioned time periods; Pixels whose vegetation index change slope is greater than a preset first slope threshold and whose time period is greater than a preset first time length are obtained to identify the second type of ecological restoration zone, and the start time of the time period is recorded as the second transition time.

4. The method for identifying ecological restoration areas according to claim 1, characterized in that, The ecological restoration zone identification result is obtained by combining the first type of ecological restoration zone and the second type of ecological restoration zone based on the preset comprehensive judgment rules, including: A pixel belonging to both the first type of ecological restoration area and the second type of ecological restoration area is identified as the first pixel. The first pixel is determined to be an ecological restoration area, and the ecological restoration area type is classified according to the land cover type of the pixel. If the pixel belongs to the first type of ecological restoration area but not the second type of ecological restoration area, obtain the net primary productivity time series of the pixel; Based on the net primary productivity time series, the slope of net primary productivity change within a second time window of a preset length is obtained, and pixels with a net primary productivity change slope greater than a preset second slope threshold are retained as second pixels, and the second pixels are determined as ecological restoration areas. If the pixel belongs to the second type of ecological restoration area but not to the first type of ecological restoration area, the pixel whose ecological land use type is ecological land is obtained as the third pixel, and the third pixel is determined to be an ecological restoration area. If the pixel does not belong to either the first type of ecological restoration area or the second type of ecological restoration area, then the pixel is determined to be a non-ecological restoration area.

5. The method for identifying ecological restoration areas according to claim 1, characterized in that, The method of comprehensively determining ecological restoration areas based on preset ecological restoration area rules, combining the first type of ecological restoration area and the second type of ecological restoration area to obtain the ecological restoration area identification result, further includes: Based on the identification results of the ecological restoration area, the ecological restoration density of the ecological restoration area is calculated using the ecological restoration density formula; Based on the ecological restoration area identification results, at least one landscape index is obtained using a preset landscape index calculation formula; The ecological restoration comprehensive index is obtained by using a preset formula for calculating the ecological restoration comprehensive index, combined with the ecological restoration density and the landscape index.

6. The method for identifying ecological restoration areas according to claim 5, characterized in that, The landscape index includes at least the area-weighted patch fractal dimension and the average Euclidean nearest neighbor distance; The formula for ecological restoration density is: in, Let be the ecological restoration density of subregion j. This represents the area of ​​a single grid cell within the ecological restoration area. This refers to the area of ​​a single grid cell within ecological land use. and These represent the number of pixels in the ecological restoration area and the number of pixels in the ecological land use area for year i, respectively, within sub-region j. The formula for calculating the fractal dimension of the area-weighted patch is: in, Let fractal dimension be the area-weighted patch dimension of subregion j. Let be the area of ​​the i-th ecological restoration patch within sub-region j. Let be the perimeter of the i-th ecological restoration patch within sub-region j, and n be the number of ecological restoration patches within sub-region j. The formula for calculating the average Euclidean nearest neighbor distance is: in, Let be the average Euclidean nearest neighbor distance for ecological restoration patches in sub-region j. Let be the total area of ​​the ecological restoration patches within sub-region j. Let be the area of ​​the i-th ecological restoration patch within sub-region j. Let be the Euclidean distance between the geometric center of the i-th ecological restoration patch within sub-region j and the geometric center of the nearest ecological restoration patch; The formula for calculating the comprehensive ecological restoration index is as follows: in, Let be the comprehensive ecological restoration index for sub-region j. Let be the ecological restoration density of subregion j. Let fractal dimension be the area-weighted patch dimension of subregion j. Let be the average Euclidean nearest neighbor distance for ecological restoration patches in sub-region j. and These are the preset weighting coefficients.

7. The method for identifying ecological restoration areas according to claim 5 or 6, characterized in that, After obtaining the comprehensive ecological restoration index by using a preset formula for calculating the comprehensive ecological restoration index, combined with the ecological restoration density and the landscape index, the method further includes: Acquire data on at least one type of driving factor related to ecological restoration; wherein each type of driving factor data includes at least one driving factor; The driving factor data is standardized. The standardized driving factor data and the ecological restoration comprehensive index are input into a preset machine learning model for training to construct a driving factor analysis model; Based on the driving factor analysis model, a preset machine learning model interpretation method is used to analyze the contribution of the driving factor data to the comprehensive ecological restoration index.

8. The method for identifying ecological restoration areas according to claim 7, characterized in that, The machine learning model is the XGBoost model, and the machine learning model interpretation method is the SHAP method.

9. An ecological restoration zone identification system, characterized in that, include: The data acquisition module is used to acquire land cover data and vegetation index time series data for each pixel within the study area; wherein, the land cover data includes the land cover type of each pixel and the ecological land use type corresponding to the land cover type, and the ecological land use type includes non-ecological land and ecological land; The first type of ecological restoration area analysis module is used to obtain pixels that meet the preset land cover change characteristics based on the land cover data, and obtain the first type of ecological restoration area based on land cover change. The second type of ecological restoration area analysis module is used to extract the change characteristics of the vegetation index time series data of each pixel, obtain the pixels that meet the preset vegetation index change conditions, and obtain the second type of ecological restoration area based on the vegetation index change. The ecological restoration area identification module is used to make a comprehensive judgment based on the preset ecological restoration area comprehensive judgment rules, combining the first type of ecological restoration area and the second type of ecological restoration area, to obtain the ecological restoration area identification result.

10. A computer device comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the ecological restoration zone identification method as described in any one of claims 1 to 8.

Images