Farming space evolution simulation method and system based on digital twinning

Abstract

The present application relates to the technical field of digital twin simulation, in particular to a cultivated forest space evolution simulation method and system based on digital twin; the method comprises the following steps: acquiring multi-period remote sensing images, digital elevation models, historical cultivated land parcel boundaries with land class time series attributes and current land class survey parcel of a target area, constructing a slope cultivated land digital twin unit and determining a ridge neighboring weight; calculating a forest canopy apparent closure state based on the multi-period remote sensing images and determining a forest canopy apparent closure node; taking the forest canopy apparent closure node as a starting point, combining the ridge neighboring weight, vegetation apparent increase and slope surface action normalization value to calculate a ridge root knot locking state, determine a ridge root knot locking node and an evolution lag time; and further generating a cultivated forest space evolution stage diagram and determining a corrected cultivated forest space evolution completion time. The present application can identify the evolution process of forest canopy apparent closure leading and ridge root knot locking lag, and improve the accuracy of cultivated forest space evolution simulation.

Description

Technical Field

[0001] This invention relates to the field of digital twin simulation technology, specifically to a method and system for simulating the spatial evolution of cultivated forests based on digital twins. Background Technology

[0002] In scenarios such as returning farmland to forest, ecological restoration, soil and water conservation, and land use control, it is usually necessary to continuously track the spatial evolution process of sloping farmland from farmland to forest. After sloping farmland is converted to forest, the surface vegetation cover, canopy closure, slope runoff, field ridge structure, drainage ditch morphology, and root soil stabilization will change over time. These changes not only affect the evaluation of regional soil and water conservation and ecological restoration, but also affect the difficulty of subsequent farmland restoration, land use determination, and evaluation of ecological restoration effectiveness. In existing research and practice, returning farmland to forest is an important scenario for vegetation restoration and land use change. Data such as remote sensing images, land use survey patches, and digital elevation models are also often used for land use change monitoring and ecological restoration simulation.

[0003] Existing methods for analyzing the spatial evolution of cultivated land and forest rely heavily on multi-period remote sensing imagery, vegetation indices, land use survey patches, or land use change models to identify whether a target area has transitioned from cultivated land to forest. The spatial evolution status is then determined based on results such as increased vegetation cover, forest patch formation, or apparent canopy closure. With the development of digital twins and remote sensing temporal modeling technologies, land use change processes can be dynamically expressed using continuously updated spatial data, and forest or land use changes can also be modeled and predicted using remote sensing temporal data. However, in the process of converting cultivated land to forest on mountainous slopes, the early appearance of apparent canopy closure in remote sensing images does not necessarily mean that the original ridges, contour strips, and drainage ditches within the original cultivated land have completed structural locking. Especially in areas with distinct slope micro-topography, continuous ridge remnants, and preferential root development along ridges and drainage ditches, the surface canopy may close first, while root consolidation, litter accumulation, and slope reshaping near the original cultivated boundaries are still in a state of continuous accumulation. Summary of the Invention

[0004] The purpose of this invention is to address the problem in the background art where the apparent closure of the forest canopy and the locking of the root knots on the field ridges are not synchronized during the process of converting sloping farmland in mountainous areas to forest, resulting in inaccurate judgment of the completion time of the spatial evolution of farmland and forest. The invention proposes a method and system for simulating the spatial evolution of farmland and forest based on digital twins.

[0005] The technical solution of this invention: a method for simulating the spatial evolution of cultivated forests based on digital twins, comprising: S1. Acquire multi-period remote sensing images, digital elevation models, historical cultivated land parcel boundaries with land type time-series attributes, and current land type survey parcels of the target area. Filter out target areas that were historically sloping cultivated land and are currently forest land, or whose vegetation apparent value determined by the multi-period remote sensing images is higher than the historical cultivated land status. Construct digital twin units of sloping cultivated land and determine the adjacent weight of field ridges. S2. Calculate the apparent canopy closure status of the digital twin unit of the sloping farmland based on the multi-period remote sensing images, and determine the apparent canopy closure node; S3. Starting from the apparent closure node of the canopy, the root knot locking state of the field ridge is calculated using the adjacent weight of the field ridge, the apparent increase of vegetation relative to the historical cultivated land state, and the normalized value of slope action determined by the digital elevation model. S4. Determine the root knot locking node of the field ridge according to the root knot locking state, and determine the evolutionary lag time according to the canopy apparent closure node and the root knot locking node of the field ridge; S5. Based on the canopy apparent closure node, the field ridge root knot locking node, and the evolutionary lag time, generate a stage map of cultivated forest spatial evolution and determine the corrected completion time of cultivated forest spatial evolution.

[0006] Preferably, a digital twin unit for sloping farmland is constructed, and the proximity weights of field ridges are determined, including: Based on the boundaries of the historical cultivated land parcels with land type time-series attributes and the current land type survey parcels, target parcels that were historically sloping cultivated land and are currently forest land, or whose vegetation apparent value determined by the multi-period remote sensing images is higher than the historical cultivated land status, are selected, and each target parcel is treated as a digital twin unit of sloping cultivated land. The slope of each pixel within the digital twin unit of the sloping farmland is calculated based on the digital elevation model. Based on the historical cultivated land patch boundaries, contour line directions, and local slope break lines with land type time-series attributes, the spatial line sets of original field ridges, contour cultivated zones, and drainage ditches are extracted. Calculate the shortest distance from each pixel within the digital twin unit of the sloping farmland to the spatial line set; The proximity weights of field ridges are determined based on the shortest distance and the median of the shortest distances from each pixel in the digital twin unit of the sloping farmland to the spatial line set.

[0007] Preferably, the apparent canopy closure state of the digital twin unit of the sloping farmland is calculated based on the multi-period remote sensing images, and the apparent canopy closure nodes are determined, including: The multi-period remote sensing images were registered in the same seasonal phase. The normalized vegetation index of each pixel is calculated based on the near-infrared and red reflectance of each remote sensing image. The median of the normalized vegetation index of each pixel in the digital twin unit of the sloping farmland is taken to obtain the apparent vegetation value of the digital twin unit of the sloping farmland at each time. Based on the period when the digital twin unit of the sloping farmland was in the historical farmland state, the vegetation baseline value under the historical farmland state is determined. Based on the reference sample set of stable forest land, the apparent reference value of vegetation in stable forest land was determined; Based on the vegetation apparent value, the vegetation baseline value, and the stable forest vegetation apparent reference value, the canopy apparent closure state is calculated, and the moment when the canopy apparent closure state first reaches the normalized upper limit state corresponding to the stable forest vegetation apparent reference value is determined as the canopy apparent closure node.

[0008] Preferably, the stable forest land reference sample set consists of stable forest land patches located in the same target area as the sloping farmland digital twin unit, and obtained by matching according to slope difference, aspect difference and image temporal phase difference; the apparent vegetation reference value of stable forest land is the median of the apparent vegetation value of each stable forest land patch in the stable forest land reference sample set.

[0009] Preferably, starting from the apparent canopy closure node, the root-knot locking state of the field ridge is calculated using the proximity weights of the field ridge, the apparent increase in vegetation relative to the historical cultivated land state, and the normalized value of slope action determined by the digital elevation model, including: Based on the slope of each pixel within the digital twin unit of the sloping farmland, the normalized value of the slope effect is determined; Using the aforementioned canopy apparent closure node as the starting point, the increase in vegetation apparent value relative to the vegetation baseline value is calculated for each pixel in each remote sensing time phase. Multiply the adjacent weight of the field ridge, the increased apparent vegetation and the normalized value of the slope effect, and take the median of the multiplication results of each pixel in the digital twin unit of the sloping farmland to obtain the intensity of continuous vegetation occupation near the field ridge in each remote sensing time phase. The intensity of continuous vegetation occupation near the field ridge is accumulated according to the time interval between adjacent remote sensing phases to obtain the cumulative amount of root knot locking of the field ridge; The locking state of the field ridge root knot is calculated based on the cumulative locking amount and the stable locking reference amount.

[0010] Preferably, the normalized value of the slope effect is determined based on the sine value of the pixel slope and the median of the sine values ​​of the slope of each pixel in the digital twin unit of the sloping farmland, and is used to characterize the influence of slope on slope runoff, deposition and root soil stabilization.

[0011] Preferably, the stable locking reference value is determined based on a stable converted farmland to forest sample set, which consists of stable converted farmland to forest plots matched with the digital twin unit of the sloping farmland in terms of slope, aspect, and conversion period. The conversion period is determined by the boundary of the historical farmland plot with land type time-series attributes and the current land type survey plot. The field ridge root knot locking state is the normalized result of the cumulative field ridge root knot locking value relative to the stable locking reference value.

[0012] Preferably, determining the root-knot locking node of the field ridge based on the root-knot locking state, and determining the evolutionary lag time based on the canopy apparent closure node and the root-knot locking node of the field ridge, includes: The moment when the locking state of the field ridge root knot first reaches the normalized upper limit state is determined as the locking node of the field ridge root knot. The time difference between the root knot locking node of the field ridge and the apparent closure node of the forest canopy is determined as the evolutionary lag time.

[0013] Preferably, based on the canopy apparent closure nodes, the field ridge root knot locking nodes, and the evolutionary lag time, a stage map of cultivated forest spatial evolution is generated, and the corrected completion time of cultivated forest spatial evolution is determined, including: Based on the historical rate of change of the apparent canopy closure state and the historical rate of change of the root knot locking state of the field ridge, predict the apparent canopy closure state and the root knot locking state of the field ridge at the target time. When the apparent closure state of the forest canopy at the target time does not reach the normalized upper limit, the digital twin unit of the sloping farmland is divided into the apparent expansion stage of farmland to forest conversion. When the apparent closure state of the forest canopy at the target time reaches the normalized upper limit state and the root knot locking state of the field ridge at the target time does not reach the normalized upper limit state, the digital twin unit of the sloping farmland is divided into the stage of apparent forest land but field ridge not locked. When the root knot locking state of the sloping farmland reaches the normalized upper limit state at the target time, the digital twin unit of the sloping farmland is divided into the root knot locking stage. A stage map of the spatial evolution of cultivated forests is generated based on the stage division results of each slope cultivated land digital twin unit; The modified completion time of cultivated forest spatial evolution is obtained by superimposing the apparent canopy closure node with the evolutionary lag time.

[0014] Compared with the prior art, the above-mentioned technical solution of the present invention has the following beneficial technical effects: This invention constructs a digital twin unit for sloping farmland, unifying multi-period remote sensing images, digital elevation models, historical farmland patch boundaries with land use time-series attributes, and current land use survey patches into a single spatial object. Furthermore, it determines the proximity weight of field ridges based on the spatial line set formed by the original field ridges, contour farmland zones, and drainage ditches. This allows the simulation of farmland-forest spatial evolution to no longer rely solely on forest patch boundaries or vegetation cover enhancement results. It can simultaneously express the apparent changes in surface vegetation and the influence of the original sloping farmland's micro-topographic structure during the transition from farmland to forest in mountainous sloping farmland, ensuring that the digital twin simulation object corresponds to the actual sloping farmland's field ridges, slope, and drainage spatial framework, thereby improving the targeting and accuracy of identifying the stages of farmland-forest spatial evolution. This invention further divides the spatial evolution process of cultivated forest into two asynchronous state nodes: the apparent closure state of the canopy and the root-knot locking state of the field ridge. The apparent closure node of the canopy is used as the starting point for the cumulative calculation of the root-knot locking of the field ridge. The root-knot locking state of the field ridge is calculated by the adjacent weight of the field ridge, the apparent increase in vegetation, and the normalized value of the slope effect. Then, the evolutionary lag time is determined based on the root-knot locking node of the field ridge and the apparent closure node of the canopy. This invention can identify the intermediate stage in which the apparent closure of the canopy precedes the root-knot locking of the field ridge, avoiding the direct use of the apparent forestation time on remote sensing images as the completion time of the spatial evolution of cultivated forest. This improves the reliability of judging the difficulty of restoring sloping cultivated land in mountainous areas, dividing the stages of the spatial evolution of cultivated forest, and determining the corrected completion time of the evolution. Attached Figure Description

[0015] Figure 1 This is a flowchart of the digital twin-based method for simulating the spatial evolution of cultivated forests proposed in this invention; Figure 2 This is a block diagram of the digital twin-based simulation system for the spatial evolution of cultivated forests proposed in this invention. Detailed Implementation

[0016] Example 1, as Figure 1 As shown, this embodiment provides a digital twin-based simulation method for the spatial evolution of cultivated land and forest, which is used to simulate the process of the evolution of sloping cultivated land in mountainous areas from cultivated land to forest. This embodiment uses the canopy apparent closure first and the field ridge root knot locking lag as the core constraint of digital twin simulation. It breaks down the process of converting mountainous sloping farmland from farmland to forest into the canopy apparent closure process and the field ridge root knot locking process, avoiding the judgment that the spatial evolution of farmland and forest has been completed based solely on the canopy coverage enhancement or the conversion of land parcels to forest land in remote sensing images. In this embodiment, the target area is a mountainous or hilly area where there is historical sloping farmland that has been transformed into forest land or forested land. The data for the target area includes multi-period remote sensing images, digital elevation models, historical farmland patch boundaries with land use time-series attributes, and current land use survey patches. Multi-period remote sensing images are used to extract the apparent vegetation status, digital elevation models are used to extract slope, aspect, contour line direction, and local slope break lines, historical farmland patch boundaries with land use time-series attributes are used to determine the extent of historical sloping farmland, the period of historical farmland status, the number of years of farmland conversion, and the spatial location of the original field ridges and contour farming zones, and current land use survey patches are used to determine whether the target area is currently forest land or has undergone forested land transformation. It should be noted that, in this embodiment, canopy apparent closure refers to the state in which the remote sensing vegetation apparent value of the digital twin unit of sloping farmland reaches the normalized upper limit corresponding to the apparent reference value of stable forest vegetation; field ridge root knot locking refers to the state in which the continuous vegetation occupation, slope action and root consolidation accumulation near the original field ridge, contour tillage zone and drainage ditch reach the normalized upper limit corresponding to the stable converted farmland to forest sample; canopy apparent closure is used to characterize the advance changes in vegetation cover status on remote sensing images, while field ridge root knot locking is used to characterize the delayed formation process of micro-topography and recultivation obstacles inside the original sloping farmland.

[0017] This embodiment of a method for simulating the spatial evolution of cultivated forests based on digital twins includes the following steps; S1. Construct constrained digital twin units of sloping farmland with field ridges and determine the neighbor weights of field ridges; Specifically, the process involves acquiring multiple remote sensing images of the target area, a digital elevation model (DEM), historical cultivated land parcel boundaries with land use time-series attributes, and current land use survey parcels. The multiple remote sensing images utilize images from the same sensor or images from different sensors after radiometric normalization, selecting images with the same or similar phenological periods to reduce the impact of seasonal crop growth, leaf fall, and short-term bare soil on vegetation appearance. The DEM and the multiple remote sensing images are unified to the same projection coordinate system and the same pixel grid. The historical cultivated land parcel boundaries with land use time-series attributes and the current land use survey parcels are superimposed onto the pixel grid after coordinate transformation. Based on the boundaries of historical cultivated land parcels with land use time-series attributes and current land use survey parcels, target parcels that were historically sloping cultivated land and are currently forest land are selected. For target parcels that are not yet clearly recorded as forest land in the current land use survey but whose current vegetation apparent value is higher than that of historical cultivated land in multiple remote sensing images, they are also included in the selection as target parcels that are currently in a state of forestation. Specifically, in this embodiment, the increase in the current apparent vegetation value relative to the historical cultivated land state means that the current apparent vegetation value of the target patch is greater than the vegetation baseline value of the target patch under the historical cultivated land state. The current apparent vegetation value is obtained by taking the median of the normalized vegetation index of each pixel in the target patch in the remote sensing image at the current time. The vegetation baseline value under the historical cultivated land state is obtained by taking the median of the apparent vegetation values ​​of each period during the period when the target patch was under the historical cultivated land state. The increase in the current apparent vegetation value relative to the historical cultivated land state is the result of the difference in apparent vegetation between the current time and the historical cultivated land state period for the same target patch.

[0018] Each selected target patch is constructed as a digital twin unit of sloping farmland, denoted as . ;in, This represents the j-th digital twin unit of sloping farmland, where j is the serial number of the digital twin unit. A digital twin unit of sloping farmland corresponds to a historical sloping farmland map patch with continuous spatial boundaries or a target area formed by merging adjacent historical sloping farmland map patches. (The term "digital twin unit of sloping farmland" is used to describe a specific type of digital twin unit.) It is the basic spatial object for simulating the subsequent canopy apparent closure state, field ridge root knot locking state, and the stage of cultivated forest spatial evolution; It should be noted that the digital twin unit of sloping farmland in this embodiment refers to a computational unit formed by overlaying multiple periods of remote sensing images, digital elevation models, historical farmland patch boundaries with land use time-series attributes, and current land use survey patches on a unified pixel grid, using a historical sloping farmland patch or a continuous spatial area formed by merging adjacent historical sloping farmland patches as the object. The digital twin unit of sloping farmland is not simply a patch boundary, but a digital spatial object used to carry the apparent closure state of the forest canopy, the root knot locking state of the field ridge, the evolutionary lag time, and the stage of the evolution of the cultivated forest space. Digital twin units of sloping farmland were calculated based on digital elevation models. The slope of each pixel x is calculated using the following formula: ; Where x represents a digital twin unit of sloping farmland The pixels inside, This represents the elevation value corresponding to the location of pixel x. This represents the first horizontal direction in the projected coordinate system. This represents the second horizontal direction in the projected coordinate system. Indicates elevation value The rate of change in the first horizontal direction, This represents the rate of change of the elevation value z(x) in the second horizontal direction. This represents the slope of pixel x; this slope is used in subsequent calculations of the normalized value of the slope effect. Further, based on the historical cultivated land parcel boundaries with land use time-series attributes, contour line directions extracted from the digital elevation model, local slope break lines, and linear feature responses from multiple remote sensing images, spatial line sets of original field ridges, contour cultivated zones, and drainage ditches are extracted, denoted as... ; Specifically, digital twin units of sloping farmland are first generated based on the digital elevation model. The contour lines within the area are calculated, and the angles between each boundary segment of the historical cultivated land parcel with land use time-series attributes and the directions of adjacent contour lines are calculated. The boundary segments of the historical cultivated land parcel with land use time-series attributes are sorted in ascending order of angle, and the boundary segments that can form a continuous field boundary framework with adjacent boundary segments are selected as candidate contour cultivated land zones. The slope variation between adjacent pixels is calculated using the digital elevation model, and the continuous slope break segments are sorted in descending order of slope variation, and the segments that intersect or connect with the candidate contour cultivated land zones are selected. Alternatively, parallel continuous line segments can be used as candidate line segments for local slope breaks; low vegetation response line segments that are connected to the boundaries of historical cultivated land parcels with land use time-series attributes, extend linearly, and maintain a consistent position in multiple images can be extracted from multi-period remote sensing images and used as candidate line segments for drainage ditches; finally, the candidate line segments for contour cultivated land, local slope breaks, and drainage ditches are spatially superimposed, duplicate line segments are deleted, broken line segments with endpoint spacing not greater than the ground resolution of the input image pixels are connected, and isolated line segments with length less than the ground resolution of the input image pixels are deleted to obtain a spatial line set. ; It should be noted that, in this embodiment, the spatial line set refers to a collection of linear objects used to characterize the spatial framework of the original field ridges, contour farmland, and drainage ditches; the spatial line set is not used to re-identify the land use type of the target area, but rather to determine the spatial proximity relationship between each pixel within the digital twin unit of sloping farmland and the original field ridges, contour farmland, and drainage ditches; the spatial line set The spatial framework used to represent the constraints of field ridges, contour tillage zones, and drainage ditches on vegetation occupation, litter retention, slope runoff redistribution, and root consolidation in the original sloping farmland; the subsequent field ridge proximity weights are determined by the shortest distance from the pixel to the spatial line set, so the spatial line set is used to limit the spatial influence range of the field ridge root-locking process on the micro-topographic structure of the original sloping farmland.

[0019] For digital twin units of sloping farmland For any pixel x within the set of spatial lines, calculate the distance from that pixel to the set of spatial lines. The shortest distance is calculated as follows: ; in, Represents the distance from pixel x to the spatial line set. The shortest distance, y represents the set of spatial lines. any point on, This represents the planar distance between pixel x and point y; The smaller the value, the closer the pixel x is to the original field ridge, contour farmland, or drainage ditch; Calculating digital twin units of sloping farmland All pixels within the space line set The median of the shortest distance, denoted as ;based on and The neighbor weights of the field ridges are constructed, and the calculation formula is as follows: ; in, This represents the neighbor weights of cell x. Represents the distance from pixel x to the spatial line set. The shortest distance, Digital twin units representing sloping farmland All pixels within the space line set The shortest distance median, where exp represents an exponential function with the natural constant as the base; ridge proximity weights. Follow Increases and decreases, used to indicate the degree to which pixel x is affected by root consolidation and slope reshaping near the original field ridge, contour tillage zone and drainage ditch; Digital twin units of sloping farmland The spatial scale of the field ridges themselves is determined; It should be noted that the proximity weight of the field ridge in this embodiment refers to the spatial weight of the degree to which a pixel is affected by root consolidation, litter accumulation, and slope runoff reshaping near the original field ridge, contour tillage zone, and drainage ditch. The proximity weight of the field ridge decreases as the shortest distance from the pixel to the spatial line set increases, indicating that the closer the pixel is to the original field ridge, contour tillage zone, or drainage ditch, the higher its contribution to the root-locking state of the field ridge. The scale factor of the proximity weight of the field ridge is determined by the median of the shortest distance of the digital twin unit of the sloping farmland. If the digital twin unit of sloping farmland All pixels within are associated with the spatial line set Overlap and resulting If the value is zero, then the pixel ground resolution of the aforementioned multi-period remote sensing images or digital elevation model is used as the baseline. The value of is used to avoid division by zero, and the ground resolution of the pixel is derived from the input data itself.

[0020] S2. Determine the apparent closure status and nodes of the canopy; Specifically, the remote sensing images from multiple periods are registered in the same seasonal phase. The same seasonal phase registration includes spatial registration and temporal registration. Spatial registration is used to make the same pixel position in the remote sensing images from different periods correspond to the same ground position. Temporal registration is used to make the images being compared as close as possible to the same crop phenological period or the same month, so as to reduce the interference of seasonal vegetation changes on the apparent closure state of the forest canopy. For each remote sensing image, the normalized vegetation index (NDI) of pixel x at time t is calculated using the following formula: ; in, This represents the normalized vegetation index of pixel x at time t. This represents the near-infrared reflectance of pixel x at time t. The reflectance of pixel x in the red band at time t represents the time corresponding to the remote sensing image; the normalized vegetation index is used to characterize the vegetation greenness and apparent vegetation cover of pixel x at time t. For digital twin units of sloping farmland Calculate the apparent vegetation value of this unit at time t using the following formula: ; in, Digital twin units representing sloping farmland The apparent value of vegetation at time t, This indicates taking the median. This indicates that cell x belongs to the digital twin unit of sloping farmland. Using the median can reduce the impact of local shading, bare soil residue, boundary mixed pixels, and anomalous reflectance values ​​on the overall vegetation apparent value of a unit. It should be noted that the apparent vegetation value in this embodiment refers to the unit-level vegetation status characterization value obtained by summing the normalized vegetation index of each pixel within the same digital twin unit of sloping farmland. The apparent vegetation value is used to describe the vegetation cover and greenness level of the digital twin unit of sloping farmland on the remote sensing image, and is not directly equivalent to the measured canopy closure or ground vegetation coverage. This embodiment uses the median to determine the apparent vegetation value in order to reduce the influence of local shadows, bare soil residue, cloud edges, boundary mixed pixels and anomalous reflectance values ​​on the judgment of the overall vegetation status of the unit.

[0021] Digital twin units of sloping farmland are determined based on the boundaries of historical cultivated land patches with land use time-series attributes. The collection of periods that were still historically cultivated land is denoted as ;based on The formula for calculating the vegetation baseline value under historical cultivated land conditions is as follows: ; in, Digital twin units representing sloping farmland Vegetation baseline values ​​under historical cultivated land conditions Digital twin units representing sloping farmland The collection of lands that were still historically cultivated land. Indicates that time t belongs to the set of historical cultivated land status periods; vegetation baseline value. Used to characterize the apparent vegetation level of the digital twin unit of sloping farmland in the same phenological phase before forestation; Select a reference sample set of stable forest land, denoted as Stable forest land reference sample set Digital twin units of sloping farmland Stable forest map patches are formed by sorting and matching them according to slope difference, aspect difference, and image temporal phase difference within the same target area. Specifically, map patches that continuously maintain forest status in the land use time series within the target area are selected as candidate stable forest map patches. For each candidate stable forest map patch, its average slope and digital twin unit of sloping farmland are calculated. The difference between the average slopes is denoted as the slope difference; the average aspect is calculated and compared with the digital twin unit of the sloping farmland. The difference between the average slope aspects is denoted as the slope aspect difference; the image acquisition time phase and the digital twin unit of the sloping farmland are calculated. The difference between the acquisition time phases of the images used is denoted as the image phase difference. The slope difference, aspect difference, and image temporal difference of each candidate stable forest map patch are sorted in ascending order to obtain the slope matching order, aspect matching order, and image temporal matching order. The slope matching order, aspect matching order, and image temporal matching order of the same candidate stable forest map patch are added together to obtain the comprehensive matching order of that candidate stable forest map patch. Based on the comprehensive matching order in ascending order, the candidate stable forest map patch with the smallest comprehensive matching order is selected to form a stable forest reference sample set. If multiple candidate stable forest map patches have the same overall matching order and are all the smallest, then these multiple candidate stable forest map patches will be included in the stable forest reference sample set. ; Stable forest land reference sample set Used to provide a vegetation appearance reference for stable forest land within a target area; due to the stable forest land reference sample set The determination is based on the sorting results of slope difference, aspect difference and image phase difference, which can adapt to the terrain conditions and remote sensing image phase conditions of different target areas; It should be noted that the stable forest land reference sample set in this embodiment refers to the set of reference patches obtained by maintaining the forest land status continuously in the land use time series within the target area and sorting and matching them with the digital twin units of sloping farmland according to the slope difference, aspect difference, and image time phase difference; the stable forest land reference sample set is used to determine the apparent reference value of stable forest land vegetation, so that the normalized calculation of the apparent closure state of the canopy has a comparable benchmark within the region; the stable forest land reference sample set is not used to train the classification model, nor is it used to replace the current land use survey patches.

[0022] The apparent reference value of vegetation in stable forest land is calculated using the following formula: ; in, Digital twin units representing sloping farmland The corresponding apparent reference value for stable forest vegetation, Let r represent the stable forest land reference sample set. Stable forest map patches in Represents the apparent vegetation value of a stable forest patch r at time t; apparent reference value of stable forest vegetation. Used to represent the apparent vegetation level of stable forest land within the same target area under similar topographic and image temporal conditions; Based on vegetation apparent value Vegetation baseline values ​​under historical cultivated land conditions and apparent reference values ​​of stable forest vegetation Calculate digital twin units of sloping farmland At time t, in the apparent closed state of the canopy, the calculation formula is: ; in, Digital twin units representing sloping farmland At time t, the canopy appears to be closed. Digital twin units representing sloping farmland The apparent value of vegetation at time t, This represents the baseline value of vegetation under historical cultivated land conditions. This represents the apparent reference value for stable forest vegetation. This indicates taking the maximum value. This indicates taking the minimum value; through and Processing, making Limited to between zero and one; when The closer to a given time, the more accurate the representation of the digital twin unit of sloping farmland. The closer the remotely sensed vegetation appearance is to a stable forest state; It should be noted that, in this embodiment, the apparent closure state of the canopy refers to the degree of normalization of the vegetation apparent value of the digital twin unit of sloping farmland as it approaches the apparent reference state of stable forest vegetation from the historical farmland state. The apparent closure state of the canopy is used to characterize the apparent closure process of the canopy or shrub cover on the remote sensing image. When its value reaches the upper limit of normalization, it indicates that the digital twin unit of sloping farmland has reached or exceeded the apparent reference value of stable forest vegetation in the remote sensing vegetation appearance.

[0023] like and The fact that they are equal indicates that the apparent reference value of stable forest vegetation cannot be distinguished from the vegetation baseline value under historical cultivated land conditions. In this case, the digital twin unit of the sloping cultivated land... This method will not be used for determining the apparent closure nodes of the forest canopy, and a new reference sample set for stable forest land will be selected. If a distinguishable result still cannot be formed after reselection; and If the digital twin unit of the sloping farmland is marked as a unit with insufficient reference, it will not participate in the output of the completion time of the current round of farmland spatial evolution; this process is used to ensure that the apparent closure state of the forest canopy is computable and interpretable. Canopy apparent closure state The moment when the normalized upper limit of the apparent reference value of stable forest vegetation is first reached is defined as the canopy apparent closure node, denoted as […]. The calculation formula is: ; in, Digital twin units representing sloping farmland Canopy apparent closure nodes, This represents the apparent canopy closure state at time t; due to It has been limited to between zero and one, therefore Digital twin units representing sloping farmland The apparent state of vegetation has reached or exceeded the normalized upper limit state corresponding to the apparent reference value of stable forest vegetation. It should be noted that the canopy apparent closure node in this embodiment refers to the moment when the canopy apparent closure state first reaches the normalized upper limit state. This node is used to start the subsequent cumulative calculation of field ridge root knot locking. For those that have already reached the observation time series Digital twin units of sloping farmland directly determine and with This serves as the starting point for the cumulative calculation of the root knot locking in step S3; for those not yet reached in the observation time series However, for digital twin units of sloping farmland that require future evolution simulation, the cumulative calculation of root knot locking on the field ridge is not initiated within the observation time series. In step S5, the apparent closure state of the canopy is predicted at the target time based on the historical rate of change of the apparent closure state of the canopy, and the predicted canopy apparent closure node is determined in the prediction sequence. When the canopy apparent closure node is reached for the first time in the prediction sequence, the predicted canopy apparent closure node is used as the starting point to continue predicting the root knot locking state of the field ridge. Thus, the calculation or prediction of the root knot locking state of the field ridge always occurs after the canopy apparent closure node, which conforms to the evolutionary order of canopy apparent closure first and root knot locking of the field ridge lagging behind.

[0024] S3. Calculate the root knot locking status of the field embankment; Canopy apparent closure nodes Starting with the digital twin unit of sloping farmland, calculate the digital twin unit. The intensity of sustained vegetation occupation and cumulative locking degree near the inner field ridges, contour cultivated zones and drainage ditches after the apparent closure of the forest canopy; Based on the obtained slope The normalized value of slope action is constructed, and the calculation formula is as follows: ; in, Digital twin units representing sloping farmland Normalized value of slope effect for inner pixel x. This represents the slope of pixel x. This represents the sine value of the slope of pixel x. Digital twin units representing sloping farmland Median of the sine values ​​of slope for each pixel; normalized value of slope action It is used to characterize the effects of slope on slope runoff, deposition, and root soil stabilization; because it uses digital twin units of sloping farmland. The median of the internal slope sine value is normalized, thus enabling it to adapt to the terrain differences of different digital twin units of sloping farmland; It should be noted that the normalized slope action value in this embodiment refers to the normalized slope influence quantity constructed based on the slope differences within the digital twin unit of sloping farmland; the normalized slope action value is used to characterize the differences in slope runoff, sediment retention, and root-soil consolidation caused by different slopes in different pixels; the normalized slope action value is normalized using the median of the sine value of the slope within the digital twin unit of sloping farmland to make the calculation results between different digital twin units of sloping farmland comparable; like If the value is zero, it means that the unit does not meet the slope attribute requirements of the digital twin unit of sloping farmland, or that there is insufficient effective slope data in the unit. In this case, the unit is marked as a unit with insufficient slope data and the field ridge root lock status calculation is not performed.

[0025] At the canopy apparent closure node Then, for each remote sensing time phase τ, the apparent increase in vegetation of pixel x relative to the historical cultivated land state is calculated using the following formula: ; in, Digital twin units representing sloping farmland The apparent increase in vegetation of internal pixel x relative to historical cultivated land state in remote sensing time phase τ. This represents the normalized vegetation index of pixel x in the remote sensing phase τ. Digital twin units representing sloping farmland Vegetation baseline values ​​under historical cultivated land conditions; through... The processing method records the portion of vegetation below the historical cultivated land status benchmark value as zero, so that the apparent increase in vegetation only reflects the enhanced vegetation occupation relative to the historical cultivated land status. It should be noted that the apparent increase in vegetation in this embodiment refers to the non-negative increase of the normalized vegetation index of a pixel in the remote sensing time phase relative to the vegetation baseline value under the historical cultivated land state of the digital twin unit of the sloping cultivated land. The apparent increase in vegetation is used to characterize the degree of vegetation occupation of the pixel relative to the historical cultivated land state after the canopy apparent closure. The part lower than the vegetation baseline value under the historical cultivated land state is recorded as zero so that the root knot locking state of the field ridge only accumulates the vegetation enhancement contribution related to the cultivated land to forest process, and does not take the short-term bare soil, shadow or low vegetation response as negative locking contribution.

[0026] Weighting the proximity of the field ridge Apparent increase in vegetation Normalized value of slope action Multiply and perform digital twin unit analysis on sloping farmland The median of the multiplication results of each pixel is taken to obtain the intensity of persistent vegetation occupation near the field ridge under the remote sensing time phase τ. The calculation formula is as follows: ; in, Digital twin units representing sloping farmland Intensity of persistent vegetation cover near field ridges under remote sensing time phase τ This represents the neighbor weights of cell x. This represents the apparent increase in vegetation of pixel x at the remote sensing time phase τ relative to its historical cultivated land state. This represents the normalized value of the slope effect for pixel x; The larger the value, the stronger the basis for sustained vegetation occupation and slope action accumulation near the original field ridges, contour farmland and drainage ditches under the remote sensing time phase. It should be noted that the persistent vegetation occupation intensity near the field ridge in this embodiment refers to the temporal proxy quantity formed by the increase in vegetation appearance and slope effect in the area near the original field ridge, contour tillage zone and drainage ditch within the digital twin unit of sloping farmland under a remote sensing time phase. The persistent vegetation occupation intensity near the field ridge is determined by the field ridge proximity weight, the increase in vegetation appearance and the normalized value of slope effect. Only when the pixel simultaneously satisfies the conditions of being close to the spatial line set, having an increase in vegetation appearance relative to the historical farmland state and having a contribution from slope effect will it make a high contribution to the subsequent cumulative amount of field ridge root knot locking. It should be further clarified that the root knot locking status of the field ridge is not directly measured by the underground root system. Instead, it uses the amount of persistent vegetation occupation obtainable through remote sensing, the spatial location of the field ridge, and the intensity of slope action as proxy quantities to characterize the cumulative formation process of obstacles to reclamation within the original sloping farmland; among which, Used to characterize the increase in vegetation occupation of pixel x after the apparent closure of the canopy relative to the historical cultivated land state, the increased vegetation occupation corresponding to the root development basis provided by the continuous growth of trees or shrubs; Used to characterize the spatial proximity relationship between pixel x and the original field ridge, contour tillage zone and drainage ditch, which corresponds to the spatial constraint that roots, litter and slope sediments preferentially accumulate along the original field ridge, contour tillage zone and drainage ditch. Used to characterize the effects of slope on slope runoff, sediment retention, and root-soil consolidation; , and Multiplication indicates that a pixel contributes significantly to the accumulation of root-knot locking on the field ridge only when it simultaneously meets the following conditions: proximity to the original field ridge spatial framework, enhanced vegetation occupation relative to historical cultivated land conditions, and contribution from slope action. Therefore, It can serve as a temporal proxy for the combined effects of continuous vegetation occupation and slope reshaping near original field ridges, contour cultivated zones, and drainage ditches.

[0027] The cumulative root-locking amount of the field ridge is obtained by accumulating the intensity of continuous vegetation occupation near the field ridge according to the time interval between adjacent remote sensing phases. The calculation formula is as follows: ; in, Digital twin units representing sloping farmland From the apparent closure node of the forest canopy The cumulative amount locked at the root knot of the field at time t. This represents the intensity of persistent vegetation cover near the field ridge under the remote sensing time phase τ. Σ represents the time interval between adjacent remote sensing phases. Accumulate the remote sensing phases between time t; cumulative amount locked at the root knot of the field ridge. It is used to characterize the cumulative locking degree formed near the original field ridge, contour tillage zone and drainage ditch after the canopy has closed due to continuous vegetation occupation, root consolidation, litter accumulation and slope runoff reshaping. Select a stable sample set of land converted from farmland to forest, denoted as Stable sample set of farmland converted to forest Digital twin units of sloping farmland The stable converted farmland to forest plots are composed of plots that are sorted and matched in terms of slope, aspect and conversion period; the conversion period is determined by the boundary of historical farmland plots with land use time series attributes and the current land use survey plots, that is, the time length between the time when the plot changes from farmland to forest in the land use time series and the current land use survey time is used as the conversion period. Specifically, land parcels that transitioned from cultivated land to forest land in the land use time series and maintained their forest land status in subsequent periods are designated as candidate stable conversion plots. For each candidate stable conversion plot, its average slope and digital twin unit of sloping cultivated land are calculated. The difference between the average slopes is denoted as the slope difference; the average aspect is calculated and compared with the digital twin unit of the sloping farmland. The difference between the average slope aspects is denoted as the slope aspect difference; the number of years of land conversion from cultivation to farmland and the digital twin unit of sloping farmland are calculated. The difference between the years of land conversion is denoted as the difference in years of land conversion. For each candidate stable converted farmland to forest plot, the slope difference, aspect difference, and conversion year difference are sorted in ascending order to obtain the slope matching order, aspect matching order, and conversion year matching order, respectively. The slope matching order, aspect matching order, and conversion year matching order of the same candidate stable converted farmland to forest plot are added together to obtain the comprehensive matching order of that candidate stable converted farmland to forest plot. Based on the comprehensive matching order in ascending order, the candidate stable converted farmland to forest plot with the smallest comprehensive matching order is selected to form a stable converted farmland to forest sample set. If multiple candidate stable converted farmland to forest plots exist with the same comprehensive matching order and are all the smallest, then these multiple candidate stable converted farmland to forest plots will be included in the stable converted farmland to forest sample set. ; Stable sample set of farmland converted to forest Used to provide a reference for locking the root knots of field ridges in the target area that have reached a stable state of conversion from farmland to forest; due to the stable conversion of farmland to forest sample set The determination is based on the ranking results of slope difference, aspect difference, and difference in years of land conversion from cultivation, without relying on manually set fixed thresholds, thus enabling stable locking of the reference value. Digital twin units of sloping farmland The terrain conditions and the process of converting farmland back to forest should correspond; A stable sample set of converted farmland to forest Each stable converted farmland to forest plot m in the data is configured according to the digital twin unit of sloping farmland. In the same way, calculate the cumulative amount of root knot locking on the field ridge. Where m represents the stable sample set of farmland converted to forest. One of the stable areas where farmland has been converted into forest. This represents the stable forest state time of stable converted farmland to forest patch m in the available remote sensing time series; then, the stable locking reference quantity is calculated, using the following formula: ; in, Digital twin units representing sloping farmland The corresponding stable locking reference value, The image plot m represents the stable forest land status at any given time. The accumulated amount of root knots in the field ridges This represents a stable sample set of land converted from farmland to forest. This indicates that patch m, representing stable converted farmland to forest, belongs to the stable converted farmland to forest sample set. Stable locking of reference quantity Used to represent digital twin units of sloping farmland within the target area. The locked-in cumulative level achieved by stable afforestation plots that match topography and years of land conversion; It should be noted that the stable conversion of farmland to forest sample set in this embodiment refers to the set of sample patches that have changed from farmland to forest in the land use time series and have maintained the forest status in subsequent periods. These patches are obtained by sorting and matching them with the digital twin units of sloping farmland according to the differences in slope, aspect, and conversion years. The stable conversion of farmland to forest sample set is used to provide a regional stability reference for the field ridge root knot locking status. The stable locking reference quantity in this embodiment refers to the median of the cumulative amount of field ridge root knot locking of each sample patch in the stable conversion of farmland to forest sample set under the stable forest status. This is used to normalize the field ridge root knot locking status of different sloping farmland digital twin units to a uniform scale.

[0028] Based on the accumulated amount locked in the field ridge root knot and stable locking reference quantity The root knot locking state of the field embankment is calculated using the following formula: ; in, Digital twin units representing sloping farmland At time t, the root knot of the field is locked. Digital twin units representing sloping farmland From the apparent closure node of the forest canopy The cumulative amount locked at the root knot of the field at time t. This indicates that the reference value is locked stably, through... Processing, making No more than one; when The closer to a given time, the more accurate the representation of the digital twin unit of sloping farmland. The more closely the root consolidation, litter accumulation, and slope remodeling accumulation near the inner field ridges, contour tillage zones, and drainage ditches are to the locked state of stable converted farmland into forest samples; It should be noted that the root knot locking state of the field ridge in this embodiment refers to the state quantity that characterizes the degree of accumulation of obstacles to reclamation near the original field ridge, contour tillage zone and drainage ditch by using the normalized result of the cumulative amount of root knot locking relative to the stable locking reference quantity; the root knot locking state of the field ridge does not directly measure the underground root system, but uses the amount of continuous vegetation occupation, the spatial location of the field ridge and the intensity of slope action as proxy quantities to characterize the process of root consolidation, litter accumulation and slope runoff remodeling over time after the apparent closure of the canopy; like A value of zero indicates a stable sample set of land converted from farmland to forest. If a valid and stable reference value cannot be provided, a new stable sample set of converted farmland to forest should be selected. If you select again If the value is still zero, the digital twin unit of the sloping farmland is marked as a unit with insufficient stable locking reference, and the root knot locking node of the field ridge is not output.

[0029] S4. Determine the root knot locking node and evolutionary lag time of the field ridge; Specifically, based on the root knot locking status of the field ridge The moment when the root knot locking state first reaches the normalized upper limit state is determined as the root knot locking node, and the calculation formula is: ; in, Digital twin units representing sloping farmland The root knot of the field ridge is locked. Digital twin units representing sloping farmland At time t, the root knot of the field is locked; due to It has been limited to between zero and one, therefore Digital twin units representing sloping farmland The cumulative state of the root knot locking in the field has reached the normalized upper limit state corresponding to the stable locking reference quantity; It should be noted that the field ridge root knot locking node in this embodiment refers to the moment when the field ridge root knot locking state first reaches the normalized upper limit state. This node is used to indicate that the digital twin unit of sloping farmland has reached the state of completion of the farmland spatial evolution corresponding to the stable locking reference value. Based on the apparent closure nodes of the forest canopy Hotan Kangen Knot Locking Node The evolutionary lag time is calculated using the following formula: ; in, Digital twin units representing sloping farmland Evolutionary lag time, This indicates the root knot locking node of the field ridge. Indicates the apparent closure node of the forest canopy; if If the value is greater than zero, it indicates that the digital twin unit of the sloping farmland exhibits an evolutionary process in which the apparent closure of the forest canopy precedes the locking of the root knots on the field ridges; if... If the value is zero, it means that the digital twin unit of the sloping farmland does not exhibit an identifiable time lag difference under the current data time series. It should be noted that the evolutionary lag time in this embodiment refers to the time delay between the root-knot locking node of the field ridge and the apparent closure node of the canopy. The evolutionary lag time is used to characterize the length of time that the original field ridge, contour tillage zone and drainage ditch area still need to undergo root consolidation, litter accumulation and slope runoff remodeling accumulation process after the digital twin unit of sloping farmland has reached the apparent closure state of the canopy in the remote sensing image. The larger the evolutionary lag time, the higher the possibility of temporal misjudgment when judging the completion time of the farmland spatial evolution based solely on the apparent closure node of the canopy.

[0030] S5. Generate simulation results of the spatial evolution of cultivated forests with hysteresis constraints; Based on the apparent closure state of the forest canopy Historical rate of change and field ridge root knot locking status The historical rate of change, predicting the apparent canopy closure state and the root knot locking state of the field ridge at the target time; For two adjacent remote sensing times t-1 and t, when When the value is less than one, calculate the historical rate of change of the apparent canopy closure state using the following formula: ; in, Digital twin units representing sloping farmland The rate of change in the apparent closure state of the forest canopy at time t. This represents the apparent canopy closure state at time t. This indicates the apparent canopy closure state at the previous remote sensing time; if If the value is one, it means that the canopy has reached the apparent closure state at the previous remote sensing time, and the canopy apparent closure state in the subsequent prediction will remain one. when When the value is less than one, calculate the historical rate of change of the root knot locking state of the field ridge. The calculation formula is: ; in, Digital twin units representing sloping farmland The rate of change of the historical locking state of the field ridge root knot at time t. This indicates the locked state of the root knot at time t. This indicates the root knot locking state of the field ridge at the previous remote sensing moment; if If the value is one, it means that the root knot of the field ridge has reached the locked state in the previous remote sensing moment, and the root knot locked state of the field ridge in the subsequent prediction will remain at one. The canopy apparent closure state at the next target time is predicted using a first-order saturation update method, calculated as follows: ; in, Digital twin units representing sloping farmland The predicted canopy apparent closure state at the next target time t+1. This represents the apparent canopy closure state at time t. This indicates the historical rate of change in the apparent closure state of the forest canopy. This indicates the remaining amount of time from the normalized upper limit state at the current moment; this update method causes the growth rate of the canopy apparent closure state to gradually decrease as it approaches the normalized upper limit state, which is consistent with the evolutionary characteristic of vegetation apparent value slowing down after the forest state stabilizes. The first-order saturated update method is used to predict the root knot locking state of the field at the next target time. The calculation formula is as follows: ; in, Digital twin units representing sloping farmland Predicted field ridge root knot locking state at the next target time t+1 This indicates the root knot locking state of the field at time t. This represents the historical rate of change of the root knot locking state in the field. This represents the remaining distance from the normalized upper limit state at the current moment; this update method is used to simulate the process by which the root knot locking state of the field ridge gradually approaches a stable locking state as continuous vegetation occupation and slope action accumulate; When multiple target times need to be predicted, the predicted canopy apparent closure state at the previous target time is used as the input for the next update, and the predicted root knot locking state at the previous target time is used as the input for the next update, iterating sequentially until the target time; if during the prediction process... If it is greater than one, then its value is restricted to one; if If the value is greater than one, then its value is limited to one; if it is due to image fluctuations... or If the value is negative, the corresponding historical rate of change will be treated as zero in this embodiment to avoid abnormal regression in the digital twin simulation results that is opposite to the direction of the continuous evolution of farmland conversion to forest. Digital twin units of sloping farmland When there are three or more consecutive remote sensing phases, the historical rate of change of the apparent canopy closure state can be determined by the median of the historical rate of change of the most recent consecutive non-negative canopy apparent closure state, and the historical rate of change of the root knot locking state of the field ridge can be determined by the median of the historical rate of change of the most recent consecutive consecutive non-negative root knot locking state. Using the median to determine the historical rate of change can reduce the interference of cloud shadows, shadows, boundary mixed pixels, or short-term vegetation disturbance on the target time prediction results when a single-period remote sensing image is affected by cloud shadows, shadows, boundary mixed pixels, or short-term vegetation disturbance. If there are only two consecutive remote sensing phases, the non-negative historical rate of change calculated from the adjacent phases is used for target time prediction. It should be noted that the cultivated forest spatial evolution stage map in this embodiment refers to the layer data formed after writing back the evolution stage results of each slope cultivated land digital twin unit in the target area to the spatial grid; the cultivated forest spatial evolution stage map includes at least the cultivated-to-forest apparent expansion stage, the apparent forestation but field ridge unlocked stage, and the field ridge root knot locking stage; the corrected cultivated forest spatial evolution completion time refers to the completion time obtained by superimposing the evolution lag time on the canopy apparent closure node. This completion time is consistent with the field ridge root knot locking node, which is used to avoid misjudging the canopy apparent closure time as the actual cultivated forest spatial evolution completion time.

[0031] Based on the apparent canopy closure state and the root knot locking state of the field ridges at the target time, the digital twin units of sloping farmland were analyzed. The spatial evolution stages of cultivated forests are divided; When the apparent canopy closure at the target time does not reach the normalized upper limit, the digital twin unit of sloping farmland will be... The land is divided into two stages: the apparent expansion stage of the farmland-to-forest conversion; this stage represents the digital twin unit of sloping farmland. The canopy or shrub cover is still in the process of remote sensing appearance enhancement and has not yet reached the upper limit state corresponding to the stable forest vegetation appearance reference value; When the apparent canopy closure at the target time reaches the normalized upper limit, and the root knot locking at the target time does not reach the normalized upper limit, the digital twin unit of the sloping farmland will be... The area is divided into a stage where the apparent forestation is not yet fully defined, but the field ridges are not locked; this stage represents the digital twin unit of sloping farmland. The vegetation appearance in remote sensing images is close to that of stable forest land, but the root consolidation, litter accumulation and slope remodeling accumulation near the original field ridges, contour tillage zones and drainage ditches have not yet reached the locking level of stable converted farmland into forest samples. When the root knot locking state of the sloping farmland reaches the normalized upper limit at the target time, the digital twin unit of the sloping farmland will be... The process is divided into a field ridge root knot locking stage; this stage represents the digital twin unit of sloping farmland. The root consolidation, litter accumulation, and slope reshaping processes near the central plains ridges, contour tillage zones, and drainage ditches have reached the normalized upper limit state corresponding to the stable locked reference value, and the completion time of the spatial evolution of cultivated forests is no longer determined by the canopy apparent closure node alone. After performing the aforementioned stage division on all digital twin units of sloping farmland within the target area, the stage division results of each digital twin unit are written back to the spatial grid of the target area to generate a farmland-forest spatial evolution stage map. This farmland-forest spatial evolution stage map includes at least three spatial states: the apparent expansion stage of farmland to forest conversion, the apparent forestation stage but unlocked field ridge stage, and the root-knot-locked field ridge stage. Each spatial state corresponds to a digital twin unit of sloping farmland. The apparent closure state of the forest canopy and the root knot locking state of the field ridge correspond to each other, which can reflect the evolutionary differences of different sloping farmland patches at the same target time. Based on the apparent closure nodes of the forest canopy and evolutionary lag time The revised completion time of the spatial evolution of cultivated forests is determined by the following formula: ; in, Digital twin units representing sloping farmland The revised completion time of the spatial evolution of cultivated forests Indicates the apparent closure node of the forest canopy. Indicates evolutionary lag time; due to ,therefore Locking nodes with field ridge roots To maintain consistency, the digital twin simulation results no longer use the canopy apparent closure node as the completion time of cultivated forest spatial evolution, but instead use the field ridge root knot locking node as the actual completion time of evolution, thus reflecting the evolutionary characteristics of canopy apparent closure preceding field ridge root knot locking.

[0032] This embodiment outputs a spatial evolution stage map of cultivated forest in the target area and the apparent canopy closure nodes of each digital twin unit of sloping cultivated land. , Field ridge root knot locking node Evolutionary lag time and the revised completion time of the spatial evolution of cultivated forests .

[0033] Example 2: Based on Example 1, this example illustrates an alternative implementation of the present invention; In an alternative implementation, if the multi-period remote sensing images of the target area contain other remote sensing vegetation indices that can characterize vegetation cover status, the enhanced vegetation index or the normalized difference water index can be used to replace the normalized vegetation index in the calculation of vegetation apparent value; however, the replaced vegetation index should still be used to calculate vegetation apparent value, vegetation baseline value under historical cultivated land status, vegetation apparent reference value of stable forest land, and forest canopy apparent closure status, and should not be separated from the lag relationship between the forest canopy apparent closure node and the field ridge root knot locking node. In another alternative implementation, spatial line sets It can be determined by one or more of the following: historical cultivated land patch boundaries with land type time-series attributes, contour line direction, local slope break lines, field ridge identification results, or drainage ditch identification results; however, spatial line sets The role should still be to characterize the spatial constraints of the original field ridge, contour tillage zone, and drainage ditch on the root knot locking process of the field ridge, and the adjacent weight of the field ridge. It should still be based on pixel to spatial line set Determining the shortest distance; In another alternative implementation, a stable forest land reference sample set and stable afforestation sample set Stable forest land can be determined through manual verification of samples, historical stable land cover samples, or stable samples interpreted by remote sensing over multiple years; however, a stable forest land reference sample set... Still used to determine the apparent reference value of stable forest vegetation, and the sample set of stable converted farmland to forest. They are still used to determine the stable locking reference value, but neither can replace the calculation process of the canopy apparent closure state and the field ridge root knot locking state; In another alternative implementation, if some digital twin units of sloping farmland in the target area have not yet reached the canopy apparent closure node or the field ridge root knot locking node in the observation time series, the first-order saturation update method in step S5 can be used to predict the future target time, and the canopy apparent closure node, field ridge root knot locking node and the corrected farmland spatial evolution completion time can be determined in the prediction sequence; this alternative implementation still takes canopy apparent closure first and field ridge root knot locking as the core constraint. In summary, this invention, by constructing digital twin units of sloping farmland, determining the adjacent weights of field ridges, calculating the apparent closure state of the forest canopy, calculating the root knot locking state of field ridges, determining the evolutionary lag time, and correcting the completion time of the spatial evolution of farmland and forest, enables the spatial evolution simulation of sloping farmland in mountainous areas after conversion from farmland to forest to no longer rely solely on remote sensing apparent forestation nodes. Instead, it can reflect the process of root consolidation and the delayed formation of slope reshaping in the original field ridges, contour tillage zones, and drainage ditch areas, thereby improving the realism and interpretability of the simulation of the spatial evolution of farmland and forest.

[0034] Example 3, as Figure 2As shown, the digital twin-based arable land spatial evolution simulation system proposed in this invention is used to execute the digital twin-based arable land spatial evolution simulation method described in Embodiment 1, including: The twin unit construction module is used to acquire multi-period remote sensing images, digital elevation models, historical cultivated land parcel boundaries with land type time-series attributes, and current land type survey parcels of the target area. It filters target areas that were historically sloping cultivated land and are currently forest land, or whose vegetation apparent value determined by the multi-period remote sensing images is higher than that of historical cultivated land. It constructs digital twin units of sloping cultivated land and determines the adjacent weight of field ridges. The canopy apparent closure module is used to calculate the canopy apparent closure status of the digital twin unit of the sloping farmland based on the multi-period remote sensing images, and to determine the canopy apparent closure node; The field ridge root knot locking module is used to calculate the field ridge root knot locking state starting from the canopy apparent closure node, using the field ridge proximity weight, the vegetation apparent increase relative to the historical cultivated land state, and the slope action normalization value determined by the digital elevation model. The evolutionary lag determination module is used to determine the root knot locking node of the field ridge based on the root knot locking state of the field ridge, and to determine the evolutionary lag time based on the canopy apparent closure node and the root knot locking node of the field ridge; The evolution stage output module is used to generate a farmland spatial evolution stage map based on the canopy apparent closure node, the field ridge root knot locking node, and the evolution lag time, and to determine the corrected farmland spatial evolution completion time.

[0035] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A method for simulating the spatial evolution of cultivated forests based on digital twins, characterized in that, Includes the following steps: S1. Acquire multi-period remote sensing images, digital elevation models, historical cultivated land parcel boundaries with land type time-series attributes, and current land type survey parcels of the target area. Filter out target areas that were historically sloping cultivated land and are currently forest land, or whose vegetation apparent value determined by the multi-period remote sensing images is higher than the historical cultivated land status. Construct digital twin units of sloping cultivated land and determine the adjacent weight of field ridges. S2. Calculate the apparent canopy closure status of the digital twin unit of the sloping farmland based on the multi-period remote sensing images, and determine the apparent canopy closure node; S3. Starting from the apparent closure node of the canopy, the root knot locking state of the field ridge is calculated using the adjacent weight of the field ridge, the apparent increase of vegetation relative to the historical cultivated land state, and the normalized value of slope action determined by the digital elevation model. S4. Determine the root knot locking node of the field ridge according to the root knot locking state, and determine the evolutionary lag time according to the canopy apparent closure node and the root knot locking node of the field ridge; S5. Based on the canopy apparent closure node, the field ridge root knot locking node, and the evolutionary lag time, generate a stage map of cultivated forest spatial evolution and determine the corrected completion time of cultivated forest spatial evolution.

2. The method for simulating the spatial evolution of cultivated forests based on digital twins according to claim 1, characterized in that, Construct digital twin units of sloping farmland and determine the proximity weights of field ridges, including: Based on the boundaries of the historical cultivated land parcels with land type time-series attributes and the current land type survey parcels, target parcels that were historically sloping cultivated land and are currently forest land, or whose vegetation apparent value determined by the multi-period remote sensing images is higher than the historical cultivated land status, are selected, and each target parcel is treated as a digital twin unit of sloping cultivated land. The slope of each pixel within the digital twin unit of the sloping farmland is calculated based on the digital elevation model. Based on the historical cultivated land patch boundaries, contour line directions, and local slope break lines with land type time-series attributes, the spatial line sets of original field ridges, contour cultivated zones, and drainage ditches are extracted. Calculate the shortest distance from each pixel within the digital twin unit of the sloping farmland to the spatial line set; The proximity weights of field ridges are determined based on the shortest distance and the median of the shortest distances from each pixel in the digital twin unit of the sloping farmland to the spatial line set.

3. The method for simulating the spatial evolution of cultivated forests based on digital twins according to claim 1, characterized in that, Based on the multi-period remote sensing images, the apparent canopy closure status of the digital twin unit of the sloping farmland is calculated, and the apparent canopy closure nodes are determined, including: The multi-period remote sensing images were registered in the same seasonal phase. The normalized vegetation index of each pixel is calculated based on the near-infrared and red reflectance of each remote sensing image. The median of the normalized vegetation index of each pixel in the digital twin unit of the sloping farmland is taken to obtain the apparent vegetation value of the digital twin unit of the sloping farmland at each time. Based on the period when the digital twin unit of the sloping farmland was in the historical farmland state, the vegetation baseline value under the historical farmland state is determined. Based on the reference sample set of stable forest land, the apparent reference value of vegetation in stable forest land was determined; Based on the vegetation apparent value, the vegetation baseline value, and the stable forest vegetation apparent reference value, the canopy apparent closure state is calculated, and the moment when the canopy apparent closure state first reaches the normalized upper limit state corresponding to the stable forest vegetation apparent reference value is determined as the canopy apparent closure node.

4. The method for simulating the spatial evolution of cultivated forests based on digital twins according to claim 3, characterized in that, The stable forest land reference sample set consists of stable forest land patches located in the same target area as the digital twin unit of the sloping farmland, and obtained by matching according to slope difference, aspect difference and image temporal phase difference; the apparent vegetation reference value of the stable forest land is the median of the apparent vegetation value of each stable forest land patch in the stable forest land reference sample set.

5. The method for simulating the spatial evolution of cultivated forests based on digital twins according to claim 3, characterized in that, Starting from the apparent canopy closure node, the root-knot locking state of the field ridge is calculated using the proximity weights of the field ridge, the apparent increase in vegetation relative to the historical cultivated land state, and the normalized value of slope action determined by the digital elevation model, including: Based on the slope of each pixel within the digital twin unit of the sloping farmland, the normalized value of the slope effect is determined; Using the aforementioned canopy apparent closure node as the starting point, the increase in vegetation apparent value relative to the vegetation baseline value is calculated for each pixel in each remote sensing time phase. Multiply the adjacent weight of the field ridge, the increased apparent vegetation and the normalized value of the slope effect, and take the median of the multiplication results of each pixel in the digital twin unit of the sloping farmland to obtain the intensity of continuous vegetation occupation near the field ridge in each remote sensing time phase. The intensity of continuous vegetation occupation near the field ridge is accumulated according to the time interval between adjacent remote sensing phases to obtain the cumulative amount of root knot locking of the field ridge; The locking state of the field ridge root knot is calculated based on the cumulative locking amount and the stable locking reference amount.

6. The method for simulating the spatial evolution of cultivated forests based on digital twins according to claim 5, characterized in that, The normalized value of slope action is determined based on the sine value of the pixel slope and the median of the sine values ​​of the slope of each pixel in the digital twin unit of the sloping farmland, and is used to characterize the influence of slope on slope runoff, deposition and root soil stabilization.

7. The method for simulating the spatial evolution of cultivated forests based on digital twins according to claim 5, characterized in that, The stable locking reference value is determined based on the stable converted farmland to forest sample set. The stable converted farmland to forest sample set consists of stable converted farmland to forest patches that match the digital twin unit of the sloping farmland in terms of slope, aspect, and conversion period. The conversion period is determined by the boundary of the historical farmland patch with land type time-series attributes and the current land type survey patch. The field ridge root knot locking status is the normalized result of the cumulative field ridge root knot locking amount relative to the stable locking reference value.

8. The method for simulating the spatial evolution of cultivated forests based on digital twins according to claim 1, characterized in that, The root-knot locking node is determined based on the root-knot locking state of the field ridge, and the evolutionary lag time is determined based on the apparent canopy closure node and the root-knot locking node of the field ridge, including: The moment when the locking state of the field ridge root knot first reaches the normalized upper limit state is determined as the locking node of the field ridge root knot. The time difference between the root knot locking node of the field ridge and the apparent closure node of the forest canopy is determined as the evolutionary lag time.

9. The method for simulating the spatial evolution of cultivated forests based on digital twins according to claim 1, characterized in that, Based on the aforementioned canopy closure nodes, the aforementioned root knot locking nodes, and the aforementioned evolutionary lag time, a stage map of cultivated forest spatial evolution is generated, and the revised completion time of cultivated forest spatial evolution is determined, including: Based on the historical rate of change of the apparent canopy closure state and the historical rate of change of the root knot locking state of the field ridge, predict the apparent canopy closure state and the root knot locking state of the field ridge at the target time. When the apparent closure state of the forest canopy at the target time does not reach the normalized upper limit, the digital twin unit of the sloping farmland is divided into the apparent expansion stage of farmland to forest conversion. When the apparent closure state of the forest canopy at the target time reaches the normalized upper limit state and the root knot locking state of the field ridge at the target time does not reach the normalized upper limit state, the digital twin unit of the sloping farmland is divided into the stage of apparent forest land but field ridge not locked. When the root knot locking state of the sloping farmland reaches the normalized upper limit state at the target time, the digital twin unit of the sloping farmland is divided into the root knot locking stage. A stage map of the spatial evolution of cultivated forests is generated based on the stage division results of each slope cultivated land digital twin unit; The modified completion time of cultivated forest spatial evolution is obtained by superimposing the apparent canopy closure node with the evolutionary lag time.

10. A digital twin-based simulation system for the spatial evolution of cultivated forests, used to execute the digital twin-based simulation method for the spatial evolution of cultivated forests as described in any one of claims 1-9, characterized in that, Specifically, it includes: The twin unit construction module is used to acquire multi-period remote sensing images, digital elevation models, historical cultivated land parcel boundaries with land type time-series attributes, and current land type survey parcels of the target area. It filters target areas that were historically sloping cultivated land and are currently forest land, or whose vegetation apparent value determined by the multi-period remote sensing images is higher than that of historical cultivated land. It constructs digital twin units of sloping cultivated land and determines the adjacent weight of field ridges. The canopy apparent closure module is used to calculate the canopy apparent closure status of the digital twin unit of the sloping farmland based on the multi-period remote sensing images, and to determine the canopy apparent closure node; The field ridge root knot locking module is used to calculate the field ridge root knot locking state starting from the canopy apparent closure node, using the field ridge proximity weight, the vegetation apparent increase relative to the historical cultivated land state, and the slope action normalization value determined by the digital elevation model. The evolutionary lag determination module is used to determine the root knot locking node of the field ridge based on the root knot locking state of the field ridge, and to determine the evolutionary lag time based on the canopy apparent closure node and the root knot locking node of the field ridge; The evolution stage output module is used to generate a farmland spatial evolution stage map based on the canopy apparent closure node, the field ridge root knot locking node, and the evolution lag time, and to determine the corrected farmland spatial evolution completion time.

Images