A method for detecting changes in ecological restoration status based on multi-temporal slope image registration

Abstract

This invention discloses a method for detecting changes in ecological restoration status based on multi-temporal slope image registration, relating to the field of image processing technology. The method includes the following steps: S1, acquiring multi-temporal slope images and determining slope monitoring areas: using images of the same slope at different acquisition times as input, and determining the same slope monitoring area from each temporal image. The slope monitoring area is divided into slope units corresponding to the location, ensuring that slope units in different acquisition times have consistent spatial numbers. This step outputs a multi-temporal slope unit image sequence. S2, performing image layering on each temporal slope image: this invention extracts the most clearly visible base texture fragments of each slope unit from the multi-temporal slope images, generating a unified slope base reference map, and registering each temporal slope image to the slope base reference map before performing change detection.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, specifically to a method for detecting changes in ecological restoration status based on multi-temporal slope image registration. Background Technology

[0002] Multi-temporal image change detection is an important technique in image processing for identifying differences in the state of the same target area at different acquisition times. For slope ecological restoration areas, slope images at different times can be acquired through UAV images, fixed-position images, or inspection images. After image registration, pixel-level or unit-level comparisons are performed on vegetation cover areas, bare base areas, and areas of local abnormal changes. Image registration unifies slope images from different acquisition times to the same image coordinate reference, which is the foundation for subsequent identification of ecological cover changes, base exposure changes, and stable areas. Existing image registration methods typically determine the spatial correspondence between images from different times through image feature points, texture structure, edge information, or local image similarity, and then perform change detection based on the registered images.

[0003] In multi-temporal image detection scenarios for slope ecological restoration, the visible texture of the same image region changes significantly with the acquisition time: rock surface texture, soil surface texture, and substrate texture visible in early images may be obscured by vegetation cover in later images; in later images, due to seasonal changes, post-rain erosion, local sparse cover, and changes in illumination, the substrate texture may reappear in some areas. Existing multi-temporal image registration usually relies on stable texture features commonly visible between adjacent temporal phases. When there is insufficient commonly visible substrate texture in the core restoration area, the registration transformation is easily dominated by edge areas, uncovered areas, or a small amount of residual texture, resulting in local registration drift within the image. This local registration drift can be misidentified as changes in vegetation cover, substrate exposure, or abnormal restoration status in subsequent change detection, causing the change detection results to be mixed with false changes caused by inaccurate image alignment. Summary of the Invention

[0004] The purpose of this invention is to provide a method for detecting changes in ecological restoration status based on multi-temporal slope image registration, so as to solve the problems mentioned in the background art.

[0005] To address the aforementioned technical problems, this invention provides the following technical solution: a method for detecting changes in ecological restoration status based on multi-temporal slope image registration, comprising the following steps: S1, acquire multi-temporal slope images and determine slope monitoring areas: take the images of the same slope under different acquisition time phases as input, and determine the same slope monitoring area from each temporal image. Divide the slope monitoring area into slope units corresponding to the location, so that the slope units in different acquisition time phases have consistent spatial numbers. This step outputs a multi-temporal slope unit image sequence. S2, perform image layering for each phase of the slope image: identify the visible base area, ecological cover area and transient disturbance area in each slope unit, use the visible base area as a candidate registration area, and remove the ecological cover area and transient disturbance area from the registration calculation. This step outputs the visible base result of each slope unit at each acquisition phase. S3, Establish the base exposure record of the slope unit: Compare the base visibility results of the same slope unit in multiple acquisition phases to determine which acquisition phase of the slope unit has the clearest base texture. Only one base texture fragment is selected as the reference source for each slope unit to avoid duplicate references at the same location due to vegetation decay and shadow changes. This step outputs the base fragment source phase of each slope unit. S4, Generate a slope base reference map: According to the position of the slope unit, stitch the base texture fragments corresponding to each slope unit into the same reference map. The slope base reference map is not the original image of any single acquisition phase, but is composed of the base texture fragments most suitable as the registration reference in different acquisition phases. This step outputs a slope base reference map for subsequent unified registration. S5. Register the slope images of each time phase to the slope base reference map and perform change detection: For a slope unit with a visible base area in a certain acquisition time phase, perform local registration based on the matching relationship between the visible base area and the slope base reference map. For slope units that are covered by ecological cover and lack a visible base area, complete the location transfer based on the registration relationship of adjacent registered slope units. After registration, detect the changes in ecological cover and base exposure of each slope unit and output the results of ecological restoration status changes.

[0006] According to the above technical solution, step S1 includes the following steps: S1-1, Image Input: Input the same slope surface in the... Slope images at the time of acquisition ,in, For the phase numbering of data acquisition, The value is to , The total number of time phases collected; the slope image To display a two-dimensional image of the slope ecological restoration area; S1-2, Slope monitoring area determination: In each slope image Determining the slope monitoring area The slope monitoring area is defined based on the slope boundary, slope top edge, slope toe edge, and fixed structure edge. Initial alignment is performed; the initial alignment is used to establish a correspondence between slope monitoring areas under different acquisition phases, and is not used as the registration result for the final detection of changes in ecological restoration status; S1-3, Slope Unit Division: Dividing the initially aligned slope monitoring area Divided into Slope unit ,in, Number the slope unit. The value is to , The total number of slope units; the first The first acquisition phase The image fragment of each slope unit is denoted as The image fragment For slope images Extracted from the slope unit The corresponding local image; S1-4, Unit Scale Determination: Slope Unit The size is set to be larger than the image scale of a single blade of grass and local gravel, but smaller than the image scale of local degraded patches on the slope; by limiting the size, the slope unit... Capable of including identifiable base textures while preserving locational differences in local ecological restoration status, slope units The size is set to be greater than the average projected width of a single blade of grass in the image, and less than the image width of a preset minimum degraded patch; the average projected width of a single blade of grass and the image width of the preset minimum degraded patch are determined by the sample image annotation results.

[0007] According to the above technical solution, step S2 includes the following steps: S2-1, Pixel Category Recognition: Image Fragment Recognition The pixels in the image are categorized to obtain base pixels, ecological cover pixels, and transient interference pixels. The base pixels include rock surface texture pixels, soil surface texture pixels, and hydroseeding substrate texture pixels. The ecological cover pixels include herbaceous vegetation pixels, shrub vegetation pixels, and withered grass pixels. The transient interference pixels include shadow pixels, water accumulation pixels, wet streaks pixels, and highly reflective pixels. The effective pixels are image fragments. Pixels located within the slope monitoring area that are not affected by image boundary cropping, image occlusion, or obvious imaging defects; S2-2, Calculate the base visibility value: In order to select the most suitable slope unit image segment as the registration reference from multiple acquisition phases, the image segment is first processed... The exposed area of ​​the substrate, the clarity of the substrate texture, and the degree of transient interference are uniformly quantified, and the substrate visibility value is calculated based on the quantization results. The calculation method is as follows: ,in, For the first The first acquisition phase The visible value of the base of each slope element. For image fragments The proportion of base pixels to the total number of effective pixels. For image fragments The texture sharpness normalization value of the mid-base pixel. From image fragments The grayscale gradient intensity and texture edge continuity of the base pixel are normalized; when the grayscale gradient intensity of the base pixel is high and the texture edge continuity is high... As the value of increases, For image fragments The proportion of transient interference pixels to the total number of effective pixels; the substrate visibility value The larger the value, the larger the image segment. The more suitable it is as a registration reference; S2-3, Determine candidate basis segments: When the basis visibility value Not less than the preset substrate visible threshold At that time, the image fragment Determined as a slope unit In the Candidate substrate segments under each acquisition phase; the preset substrate visibility threshold Used to exclude image segments with too small a base area, too weak texture, and too strong transient interference; S2-4, Remove non-registration regions: Mark the regions corresponding to ecological cover pixels and transient interference pixels as non-registration regions, so that grass leaf morphology, vegetation withering and flourishing, shadow boundaries, water accumulation boundaries and strong reflective boundaries do not participate in subsequent registration transformation calculations; this processing is used to avoid dynamic surface textures being mistakenly used as stable slope features.

[0008] According to the above technical solution, step S3 specifically includes: S3-1, Establish a base exposure record: For the same slope unit According to the collection phase number Record its base visibility value and candidate basal fragments The base exposure record is used to represent the same slope unit. The exposure of substrate texture under different acquisition phases; S3-2, Determine the phase of origin of the basal segment: This is done when obtaining the same slope unit... Baseline visible values ​​at each acquisition phase Subsequently, the acquisition phase with the highest base visibility value was determined as the base segment origin phase of that slope unit. The calculation method is as follows: ,in, For slope unit The formula represents the phase of the basal segment origin; Baseline visible values ​​were selected from each acquisition phase. The largest acquisition phase, as a slope unit Reference source, when two or more acquisition phases are used to obtain the base visibility value. When the proportions are the same, prioritize the proportion of transient interference pixels. If the smaller acquisition time phase is still the same, select the acquisition time phase with the smaller acquisition time phase number r. S3-3, Determine the base reference segment: For slope elements with at least one candidate base segment Image fragments Determined as a slope unit Base reference fragment ,in, This represents a local base image used to generate a slope base reference map; the base visibility values ​​for all acquisition phases. All are less than the preset substrate visibility threshold slope unit No base reference fragment is generated. The slope unit Marked as a unit to be transferred; S3-4, Constraint: Unique element source for each slope element. Only one base reference segment is retained. This restriction is used to avoid multiple conflicting registration references for the same slope location due to different seasonal vegetation conditions, humidity conditions, and shade conditions.

[0009] According to the above technical solution, step S4 specifically includes: S4-1, Establish reference map coordinates: based on the slope monitoring area at the first acquisition time phase. The unit division results are used to establish reference map coordinates, and the common overlapping area after initial alignment of each acquisition phase is used as the slope base reference map. The coordinate range; each position in the reference map coordinates corresponds to a slope unit. Correspondingly, it is used to determine the reference segments of each base. Reference diagram of slope base The placement position in the middle; S4-2, Reference diagram for assembling slope base: according to slope unit The base reference segment is positioned in the reference graph coordinates. Place it in the corresponding position to generate a slope base reference map. The reference diagram of the slope base. Basis reference fragments selected from different acquisition phases This composite image serves as a registration benchmark for slope images acquired at various time points. Through this stitching method, a slope base reference map is created. Instead of relying on the complete exposed substrate in any single acquisition phase, it uses the substrate textures that are scattered and exposed in multiple acquisition phases to form a unified registration reference. S4-3, Processing the unit to be transferred: For segments without a base reference... The element to be transferred is based on the base reference fragments already obtained in its adjacent slope elements. Determine the reference diagram of the slope base where the unit to be transferred is located. The boundary positions and cell ranges in the data are determined; this step does not generate false base textures, but only determines the positional constraints of the cells to be transferred in subsequent registration. S4-4, Retain reference fragment source: in the slope base reference diagram In the middle, there is a reference segment for each base. Record its origin phase and corresponding slope unit number This record allows for tracing the origin of each local reference area during subsequent registration error checks.

[0010] According to the above technical solution, step S5 specifically includes: S5-1, Local Matching: For an already generated base reference segment And the slope unit with base pixels in the r-th acquisition phase Image fragments Reference diagram of slope base Basis reference fragment at the corresponding position in the middle Local matching is performed; the pixels participating in local matching are limited to the base pixels determined in step S2, and ecological coverage pixels and transient interference pixels do not participate in local matching; through the local matching, image segments for evaluation are determined. Reference diagram of slope base Local registration transformation for alignment ; S5-2, Calculate local registration error: To determine the local registration transformation... Is it possible to make the base pixels match the slope base reference image? Alignment with the corresponding base texture in the reference pixel position set Non-empty and When the value is greater than 0, the local registration error needs to be calculated. When the reference pixel position set When empty, local registration error is not calculated. and the corresponding slope unit The r-th acquisition phase is marked as a registration unit to be transferred, and the pixel difference in the base pixel set is used to calculate the... The first acquisition phase Local registration error of individual slope units The calculation method is as follows: ,in, For the first The first acquisition phase Local registration error of individual slope units Reference diagram for slope base The set of reference pixel positions that participate in the local error calculation, and the set of reference pixel positions Each pixel position in the image fragment It has corresponding base pixels, For reference pixel position set The number of pixels in Reference diagram for slope base pixel position in To provide a reference diagram of the slope base The reference pixel position x in the image is mapped to the slope image. Local sampling transformation at the corresponding pixel position in the image. Representing a slope image After local registration transformation Later in position The corresponding pixel value, Reference diagram of slope base In position The formula is used to evaluate local registration results by the differences between base pixels, avoiding the use of vegetation surface texture to evaluate registration results, and addressing local registration errors. Not greater than the preset registration error threshold At that time, local registration transformation is preserved. Local registration error Greater than the preset registration error threshold At that time, the corresponding slope unit The unit to be registered and transferred is marked in the r-th acquisition phase; S5-3, complete the position transfer of occlusion units: for sets of pixels occluded by ecological coverage and lacking reference pixel locations. slope unit And due to local registration errors Greater than the preset registration error threshold The marked registration units to be transferred do not directly use their own image texture to calculate the local registration transformation; instead, they are selected based on the slope unit. Adjacent and with local registration transformation preserved slope unit As a source of transmission, among them To transmit the source unit number; based on the slope unit With the source unit The relative positional relationship in the reference diagram coordinates will convey the source unit. Local registration transformation Transfer to slope unit Determine the slope unit Reference diagram of slope base The corresponding position in the slope element There are more than two source units in the vicinity. When selecting slope units, priority should be given to those that match the slope. The source unit with the longest shared boundary length When the shared boundary lengths are the same, select the source unit number. Smaller slope units When the slope unit When there are no source cells in the surrounding area that have preserved the local registration transformation, the slope cell is... The untransmitted units are marked as non-transmitted units and retained in the results of ecological restoration status changes so that they do not participate in the determination of effective change areas. The above processing is used to avoid the slope units that are covered by ecological cover from producing incorrect matching due to the lack of stable base texture. At the same time, it is clear that the registration relationship is transmitted from the adjacent slope units that have completed local matching to the slope units to be transmitted. S5-4, Detecting changes in ecological restoration status: Slope images at each acquisition time... Reference diagram of slope base After registration, calculate the slope element. In the Ecological coverage ratio at each collection time point and the proportion of exposed substrate ;in, For image fragments The proportion of ecological coverage pixels to the total number of effective pixels. For image fragments The proportion of base pixels to the total number of effective pixels; S5-5, Output Change Results: This will change the same slope unit in adjacent acquisition phases. ecological coverage ratio and the proportion of exposed substrate By comparing the data, regions of increased coverage, decreased coverage, increased basement exposure, and stable conditions are obtained. The final output includes registered multi-temporal slope images, slope unit change categories, and ecological restoration status change maps. This output characterizes changes in ecological restoration status and reduces the impact of spurious changes caused by variations in image acquisition perspective. When the increase in ecological coverage ratio from the previous acquisition phase to the subsequent acquisition phase reaches a preset coverage change threshold, the corresponding slope unit is identified as an area of ​​increased coverage. When the decrease in ecological coverage ratio from the previous acquisition phase to the next acquisition phase reaches a preset coverage change threshold, the corresponding slope unit is identified as an area of ​​decreased coverage. When the increase in basement exposure ratio from the previous acquisition phase to the subsequent acquisition phase reaches a preset exposure change threshold, the corresponding slope unit is identified as an area of ​​increased basement exposure. When neither the change in ecological coverage ratio nor the change in basement exposure ratio reaches the corresponding threshold, the corresponding slope unit is identified as a stable condition region.

[0011] An ecological restoration state change detection system based on multi-temporal slope image registration includes an image layering module for unifying slope regions, identifying visible base areas, identifying ecological cover areas, and identifying transient disturbance areas in multi-temporal slope images to obtain the visible base areas that can participate in registration in each temporal slope image; a slope base reference map generation module for selecting the image segment with the clearest base texture for each slope unit in multiple acquisition temporal phases and stitching the image segments into a slope base reference map; and a registration change detection module for registering each temporal slope image to the slope base reference map and outputting the ecological restoration state change result after registration is completed.

[0012] According to the above technical solution, the image layering module includes a slope area determination submodule, a substrate visibility recognition submodule, and an interference removal submodule. The slope area determination submodule is used to determine the same slope monitoring area from each temporal slope image and divide the slope monitoring area into slope units corresponding to the location. The substrate visibility recognition submodule is used to identify the rock surface texture, soil surface texture, and sprayed substrate texture in each slope unit to obtain the substrate visibility area. The interference removal submodule is used to identify vegetation cover, shadows, water accumulation, wet marks, and strong reflective areas and remove them from the image areas participating in the registration. The slope base reference map generation module includes a base exposure recording submodule, a base segment selection submodule, and a reference map stitching submodule. The base exposure recording submodule is used to record the base visibility status of each slope unit in each acquisition phase. The base segment selection submodule is used to select the image segment with the highest base visibility from the image segments of each acquisition phase of each slope unit. The reference map stitching submodule is used to stitch the selected image segments according to the slope unit position to obtain the slope base reference map. The registration change detection module includes a local registration submodule, a cover change recognition submodule, and a change result output submodule. The local registration submodule is used to locally match the visible base area in each temporal slope image with the slope base reference map and transfer the matching relationship to the slope units that are covered by ecological cover. The cover change recognition submodule is used to identify the ecological cover change and base exposure change of each slope unit after registration. The change result output submodule is used to output the registered multi-temporal slope images, slope unit change categories, and ecological restoration status change maps.

[0013] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: This invention identifies base pixels, ecological cover pixels, and transient interference pixels through image layering, and selects the most suitable base texture fragment as the registration benchmark for each slope unit from multi-temporal images, thereby generating a unified slope base reference map. Each temporal slope image uses this slope base reference map as the registration benchmark, which reduces errors caused by unstable image content such as vegetation cover, local occlusion, shadows, and wet streaks participating in registration, making the registration calculation more reliant on stable base textures. After registration, comparing the ecological cover ratio and the base exposure ratio reduces the risk of local registration drift being misjudged as state changes, improving the spatial consistency and verifiability of ecological restoration state change detection results in multi-temporal slope images. Attached Figure Description

[0014] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart illustrating the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Please see Figure 1 The present invention provides a technical solution: a method for detecting changes in ecological restoration status based on multi-temporal slope image registration, comprising the following steps: S1, acquire multi-temporal slope images and determine slope monitoring areas: take the images of the same slope under different acquisition time phases as input, and determine the same slope monitoring area from each temporal image. Divide the slope monitoring area into slope units corresponding to the location, so that the slope units in different acquisition time phases have consistent spatial numbers. This step outputs a multi-temporal slope unit image sequence. S2, perform image layering for each phase of the slope image: identify the visible base area, ecological cover area and transient disturbance area in each slope unit, use the visible base area as a candidate registration area, and remove the ecological cover area and transient disturbance area from the registration calculation. This step outputs the visible base result of each slope unit at each acquisition phase. S3, Establish the base exposure record of the slope unit: Compare the base visibility results of the same slope unit in multiple acquisition phases to determine which acquisition phase of the slope unit has the clearest base texture. Only one base texture fragment is selected as the reference source for each slope unit to avoid duplicate references at the same location due to vegetation decay and shadow changes. This step outputs the base fragment source phase of each slope unit. S4, Generate slope base reference map: According to the position of slope unit, stitch the base texture fragments corresponding to each slope unit into the same reference map. The slope base reference map is not the original image of any single acquisition phase, but is composed of the base texture fragments most suitable as the registration reference in different acquisition phases. This step outputs the slope base reference map for subsequent unified registration. S5. Register the slope images of each time phase to the slope base reference map and perform change detection: For a slope unit with a visible base area in a certain acquisition time phase, perform local registration based on the matching relationship between the visible base area and the slope base reference map. For slope units that are covered by ecological cover and lack a visible base area, complete the location transfer based on the registration relationship of adjacent registered slope units. After registration, detect the changes in ecological cover and base exposure of each slope unit and output the results of ecological restoration status changes. Step S1 includes the following steps: S1-1, Image Input: Input the same slope surface in the... Slope images at the time of acquisition ,in, For the phase numbering of data acquisition, The value is to , Total number of phases collected; slope image To display a two-dimensional image of the slope ecological restoration area; S1-2, Slope monitoring area determination: In each slope image Determining the slope monitoring area The slope monitoring area is determined based on the slope boundary, slope top edge, slope toe edge, and fixed structure edge. Initial alignment is performed to establish a correspondence between slope monitoring areas at different acquisition times, and is not used as the registration result for the final detection of changes in ecological restoration status. S1-3, Slope Unit Division: Dividing the initially aligned slope monitoring area Divided into Slope unit ,in, Number the slope unit. The value is to , The total number of slope units; the first The first acquisition phase The image fragment of each slope unit is denoted as Image fragments For slope images Extracted from the slope unit The corresponding local image; S1-4, Unit Scale Determination: Slope Unit The size is set to be larger than the image scale of a single blade of grass and local gravel, but smaller than the image scale of local degraded patches on the slope; by limiting the scale, the slope unit... Capable of including identifiable base textures while preserving locational differences in local ecological restoration status, slope units The size is set to be greater than the average projected width of a single blade of grass in the image, and less than the image width of the preset minimum degradation patch; the average projected width of a single blade of grass and the image width of the preset minimum degradation patch are determined by the annotation results of the sample images; The purpose of this step is not to directly determine the repair status, but rather to place images from different time periods into the same comparable slope unit framework. Since slope images may come from different acquisition dates, shooting angles, and lighting conditions, directly comparing pixel changes could easily misinterpret differences in shooting perspective as changes in slope condition. Therefore, this step first identifies the same slope monitoring area and divides it into corresponding slope units, ensuring that each subsequent local image processing action has a clear target. After this processing, subsequent steps do not compare differences at arbitrary locations in two original images, but rather the differences in the image representation of the same slope unit at different time periods, improving the spatial correspondence of change detection from the source.

[0017] Step S2 includes the following steps: S2-1, Pixel Category Recognition: Image Fragment Recognition The pixels in the image are categorized to obtain base pixels, ecological cover pixels, and transient interference pixels. Base pixels include rock surface texture pixels, soil surface texture pixels, and hydroseeding substrate texture pixels. Ecological cover pixels include herbaceous vegetation pixels, shrub vegetation pixels, and withered grass pixels. Transient interference pixels include shadow pixels, water accumulation pixels, wet marks pixels, and highly reflective pixels. Valid pixels are image fragments. Pixels located within the slope monitoring area that are not affected by image boundary cropping, image occlusion, or obvious imaging defects; S2-2, Calculate the base visibility value: In order to select the most suitable slope unit image segment as the registration reference from multiple acquisition phases, the image segment is first processed... The exposed area of ​​the substrate, the clarity of the substrate texture, and the degree of transient interference are uniformly quantified, and the substrate visibility value is calculated based on the quantization results. The calculation method is as follows: ,in, For the first The first acquisition phase The visible value of the base of each slope element. For image fragments The proportion of base pixels to the total number of effective pixels. For image fragments The texture sharpness normalization value of the mid-base pixel. From image fragments The grayscale gradient intensity and texture edge continuity of the base pixel are normalized; when the grayscale gradient intensity of the base pixel is high and the texture edge continuity is high... As the value of increases, For image fragments The proportion of transient interference pixels to the total number of effective pixels; substrate visibility value The larger the value, the larger the image segment. The more suitable it is as a registration reference; S2-3, Determine candidate basis segments: When the basis visibility value Not less than the preset substrate visible threshold At that time, the image fragment Determined as a slope unit In the Candidate substrate segments at each acquisition time phase; preset substrate visibility threshold. Used to exclude image segments with too small a base area, too weak texture, and too strong transient interference; S2-4, Remove non-registration regions: Mark the regions corresponding to ecological cover pixels and transient interference pixels as non-registration regions, so that grass leaf morphology, vegetation withering and flourishing, shadow boundaries, water accumulation boundaries and strong reflective boundaries do not participate in subsequent registration transformation calculations; this process is used to avoid dynamic surface textures being mistakenly used as stable slope features. This step is a crucial preprocessing step for ensuring the reliability of subsequent registration. Conventional image registration often treats grass edges, shadow boundaries, wet streaks, and highly reflective areas as matchable textures, but these are not stable slope references and change with seasons, weather, and lighting. This step divides the image content into visible base areas, ecological cover areas, and transient disturbance areas, allowing only the visible base areas to participate in the subsequent registration reference selection process. Its working principle is to first distinguish between content that represents the fixed texture of the slope and surface content that only appears temporarily at a certain time. Compared to the conventional method of directly extracting feature points from the entire image, this step does not pursue more feature points but actively reduces unstable feature points, making subsequent registration rely more on relatively stable content such as rock surfaces, soil surfaces, and substrate textures.

[0018] Step S3 specifically includes: S3-1, Establish a base exposure record: For the same slope unit According to the collection phase number Record its base visibility value and candidate basal fragments The base exposure record is used to represent the same slope unit. The exposure of substrate texture under different acquisition phases; S3-2, Determine the phase of origin of the basal segment: This is done when obtaining the same slope unit... Baseline visible values ​​at each acquisition phase Subsequently, the acquisition phase with the highest base visibility value was determined as the base segment origin phase of that slope unit. The calculation method is as follows: ,in, For slope unit The formula represents the phase of the basal segment origin; Baseline visible values ​​were selected from each acquisition phase. The largest acquisition phase, as a slope unit Reference source, when two or more acquisition phases are used to obtain the base visibility value. When the proportions are the same, prioritize the proportion of transient interference pixels. If the smaller acquisition time phase is still the same, select the acquisition time phase with the smaller acquisition time phase number r. S3-3, Determine the base reference segment: For slope elements with at least one candidate base segment Image fragments Determined as a slope unit Base reference fragment ,in, This represents a local base image used to generate a slope base reference map; the base visibility values ​​for all acquisition phases. All are less than the preset substrate visibility threshold slope unit No base reference fragment is generated. The slope unit Marked as a unit to be transferred; S3-4, Constraint: Unique element source for each slope element. Only one base reference segment is retained. This restriction is used to avoid multiple conflicting registration references for the same slope location due to different seasonal vegetation conditions, humidity conditions, and shade conditions. This step best demonstrates one of the originalities of this solution. For slope ecological restoration images, the base texture of the same slope unit is not necessarily visible simultaneously in two adjacent time phases. Some areas are bare in the early stages and later covered by vegetation; other areas may see the base texture reappear later due to sparse vegetation, changes in vegetation growth and decay, or local exposure. Conventional registration methods usually require that there be a common and stable texture between two time phases to be registered. However, this step does not require that the same base texture be visible simultaneously between adjacent time phases. Instead, it records when each slope unit is most clearly visible in multiple time phases. This transforms the temporal differences that originally caused registration difficulties into a basis for finding the best base fragment, allowing each slope unit to obtain the most suitable base texture source as a reference from its own historical images.

[0019] Step S4 specifically includes: S4-1, Establish reference map coordinates: based on the slope monitoring area at the first acquisition time phase. The unit division results are used to establish reference map coordinates, and the common overlapping area after initial alignment of each acquisition phase is used as the slope base reference map. The coordinate range; refer to the coordinates of each position in the diagram and a slope unit. Correspondingly, it is used to determine the reference segments of each base. Reference diagram of slope base The placement position in the middle; S4-2, Reference diagram for assembling slope base: according to slope unit The base reference segment is positioned in the reference graph coordinates. Place it in the corresponding position to generate a slope base reference map. Slope base reference diagram Basis reference fragments selected from different acquisition phases This composite image serves as a registration benchmark for slope images acquired at various time points. Through this stitching method, a slope base reference map is created. Instead of relying on the complete exposed substrate in any single acquisition phase, it uses the substrate textures that are scattered and exposed in multiple acquisition phases to form a unified registration reference. S4-3, Processing the unit to be transferred: For segments without a base reference... The element to be transferred is based on the base reference fragments already obtained in its adjacent slope elements. Determine the reference diagram of the slope base where the unit to be transferred is located. The boundary positions and cell ranges in the data are determined; this step does not generate false base textures, but only determines the positional constraints of the cells to be transferred in subsequent registration. S4-4, Retain reference fragment source: in the slope base reference diagram In the middle, there is a reference segment for each base. Record its origin phase and corresponding slope unit number This record allows for tracing the origin of each local reference area during subsequent registration error checks. This step is the core component that distinguishes this approach from conventional multi-temporal image registration. Conventional methods typically select an image from a specific temporal phase as a reference image and then register other temporal images to that reference image. However, in slope ecological restoration scenarios, any single temporal phase may have issues such as localized vegetation shading, shadows, wet marks, or insufficient substrate exposure. Directly using a single original image as a reference would solidify these local defects throughout the entire registration process. The slope substrate reference image generated in this step is not the original image from a specific temporal phase, but rather a composite of the clearest substrate texture fragments from different acquisition temporal phases. Its purpose is to construct a more stable image registration reference, ensuring that slope images from different acquisition temporal phases are aligned around the same substrate reference, thereby avoiding the reference image itself being affected by a single acquisition temporal phase.

[0020] Step S5 specifically includes: S5-1, Local Matching: For an already generated base reference segment And the slope unit with base pixels in the r-th acquisition phase Image fragments Reference diagram of slope base Basis reference fragment at the corresponding position in the middle Local matching is performed; the pixels participating in local matching are limited to the base pixels determined in step S2, and ecological cover pixels and transient interference pixels do not participate in local matching; through local matching, the image segments used for evaluation are determined. Reference diagram of slope base Local registration transformation for alignment ; S5-2, Calculate local registration error: To determine the local registration transformation... Is it possible to make the base pixels match the slope base reference image? Alignment with the corresponding base texture in the reference pixel position set Non-empty and When the value is greater than 0, the local registration error needs to be calculated. When the reference pixel position set When empty, local registration error is not calculated. and the corresponding slope unit The r-th acquisition phase is marked as a registration unit to be transferred, and the pixel difference in the base pixel set is used to calculate the... The first acquisition phase Local registration error of individual slope units The calculation method is as follows: ,in, For the first The first acquisition phase Local registration error of individual slope units Reference diagram for slope base The set of reference pixel locations involved in the local error calculation, and the set of reference pixel locations Each pixel position in the image fragment It has corresponding base pixels, For reference pixel position set The number of pixels in Reference diagram for slope base pixel position in To provide a reference diagram of the slope base The reference pixel position x in the image is mapped to the slope image. Local sampling transformation at the corresponding pixel position in the image. Representing a slope image After local registration transformation Later in position The corresponding pixel value, Reference diagram of slope base In position The formula is used to evaluate local registration results by the differences between base pixels, avoiding the use of vegetation surface texture to evaluate registration results, and addressing local registration errors. Not greater than the preset registration error threshold At that time, local registration transformation is preserved. Local registration error Greater than the preset registration error threshold At that time, the corresponding slope unit The unit to be registered and transferred is marked in the r-th acquisition phase; S5-3, complete the position transfer of occlusion units: for sets of pixels occluded by ecological coverage and lacking reference pixel locations. slope unit And due to local registration errors Greater than the preset registration error threshold The marked registration units to be transferred do not directly use their own image texture to calculate the local registration transformation; instead, they are selected based on the slope unit. Adjacent and with local registration transformation preserved slope unit As a source of transmission, among them To transmit the source unit number; based on the slope unit With the source unit The relative positional relationship in the reference diagram coordinates will convey the source unit. Local registration transformation Transfer to slope unit Determine the slope unit Reference diagram of slope base The corresponding position in the slope element There are more than two source units in the vicinity. When selecting slope units, priority should be given to those that match the slope. The source unit with the longest shared boundary length When the shared boundary lengths are the same, select the source unit number. Smaller slope units When the slope unit When there are no source cells in the surrounding area that have preserved the local registration transformation, the slope cell is... The untransmitted units are marked as non-transmitted units and retained in the results of ecological restoration status changes so that they do not participate in the determination of effective change areas. The above processing is used to avoid the slope units that are covered by ecological cover from producing incorrect matching due to the lack of stable base texture. At the same time, it is clear that the registration relationship is transmitted from the adjacent slope units that have completed local matching to the slope units to be transmitted. S5-4, Detecting changes in ecological restoration status: Slope images at each acquisition time... Reference diagram of slope base After registration, calculate the slope element. In the Ecological coverage ratio at each collection time point and the proportion of exposed substrate ;in, For image fragments The proportion of ecological coverage pixels to the total number of effective pixels. For image fragments The proportion of base pixels to effective pixels: S5-5, Output Change Results: This will change the same slope unit in adjacent acquisition phases. ecological coverage ratio and the proportion of exposed substrate By comparing the data, regions of increased coverage, decreased coverage, increased basement exposure, and stable conditions are obtained. The final output includes registered multi-temporal slope images, slope unit change categories, and ecological restoration status change maps. This output characterizes changes in ecological restoration status and reduces the impact of spurious changes caused by variations in image acquisition perspective. When the increase in ecological coverage ratio from the previous acquisition phase to the subsequent acquisition phase reaches a preset coverage change threshold, the corresponding slope unit is identified as an area of ​​increased coverage. When the decrease in ecological coverage ratio from the previous acquisition phase to the next acquisition phase reaches a preset coverage change threshold, the corresponding slope unit is identified as an area of ​​decreased coverage. When the increase in basement exposure ratio from the previous acquisition phase to the subsequent acquisition phase reaches a preset exposure change threshold, the corresponding slope unit is identified as an area of ​​increased basement exposure. When neither the change in ecological coverage ratio nor the change in basement exposure ratio reaches the corresponding threshold, the corresponding slope unit is identified as a stable condition region. The purpose of this step is to first eliminate spatial discrepancies in the images before assessing changes in ecological restoration status. For slope units where the substrate texture is still visible, this step directly utilizes the correspondence between that region and the slope substrate reference image to complete local registration. For slope units obscured by ecological cover or with unreliable local matching, the system does not forcibly use their own image texture for matching; instead, it transfers the registration relationship based on the positional relationship of adjacent reliably registered slope units. The unique aspect of this process is that it acknowledges the lack of stable texture in obscured areas, and instead of using grass blades, shadows, or wet streaks to create registration points, it transfers the registration relationship from reliable to unreliable areas. Comparing changes in ecological cover and substrate exposure after registration reduces the likelihood of misjudging misalignment caused by image misalignment as genuine restoration changes.

[0021] In summary, the key to this implementation method lies not in simply identifying the presence of vegetation or exposed substrate on the slope, but in establishing a stable registration benchmark for multi-temporal slope images. By eliminating unstable surface content through image layering and selecting the most suitable substrate segment for each slope unit as a reference through substrate exposure records, and then using the slope substrate reference map to uniformly connect the registration relationships of images from different temporal phases, this scheme can maintain a relatively stable registration benchmark even when ecological cover is constantly changing. Compared with directly selecting an original image from a certain temporal phase as a reference map or directly extracting common feature points from adjacent temporal phase images, this scheme is more suitable for handling slope image sequences where substrate texture is gradually obscured and intermittently exposed, thus reducing the interference of registration errors on change detection results.

[0022] An ecological restoration state change detection system based on multi-temporal slope image registration includes an image layering module for unifying slope regions, identifying visible base areas, identifying ecological cover areas, and identifying transient disturbance areas in multi-temporal slope images, thus obtaining the visible base areas that can participate in registration in each temporal slope image; a slope base reference map generation module for selecting the image segment with the clearest base texture for each slope unit in multiple acquisition temporal phases and stitching the image segments into a slope base reference map; and a registration change detection module for registering each temporal slope image to the slope base reference map and outputting the ecological restoration state change results after registration is completed. The image layering module includes a slope region determination submodule, a substrate visibility recognition submodule, and an interference removal submodule. The slope region determination submodule is used to determine the same slope monitoring area from each temporal slope image and divide the slope monitoring area into slope units corresponding to the location. The substrate visibility recognition submodule is used to identify the rock surface texture, soil surface texture, and sprayed substrate texture in each slope unit to obtain the substrate visible area. The interference removal submodule is used to identify vegetation cover, shadows, water accumulation, wet marks, and strong reflective areas and remove them from the image areas participating in the registration. The slope base reference map generation module includes a base exposure recording submodule, a base segment selection submodule, and a reference map stitching submodule. The base exposure recording submodule records the base visibility status of each slope unit in each acquisition phase. The base segment selection submodule selects the image segment with the highest base visibility from the image segments of each acquisition phase of each slope unit. The reference map stitching submodule stitches the selected image segments according to the slope unit position to obtain the slope base reference map. The registration change detection module includes a local registration submodule, a cover change recognition submodule, and a change result output submodule. The local registration submodule is used to locally match the visible base area in each temporal slope image with the slope base reference map and transfer the matching relationship to the slope units that are covered by ecological cover. The cover change recognition submodule is used to identify the ecological cover change and base exposure change of each slope unit after registration. The change result output submodule is used to output the registered multi-temporal slope images, slope unit change categories, and ecological restoration status change maps.

[0023] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0024] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for detecting changes in ecological restoration status based on multi-temporal slope image registration, characterized in that: Includes the following steps: S1, acquire multi-temporal slope images and determine slope monitoring areas: take the images of the same slope under different acquisition time phases as input, and determine the same slope monitoring area from each temporal image. Divide the slope monitoring area into slope units corresponding to the location, so that the slope units in different acquisition time phases have consistent spatial numbers. This step outputs a multi-temporal slope unit image sequence. S2, perform image layering for each phase of the slope image: identify the visible base area, ecological cover area and transient disturbance area in each slope unit, use the visible base area as a candidate registration area, and remove the ecological cover area and transient disturbance area from the registration calculation. This step outputs the visible base result of each slope unit at each acquisition phase. S3, Establish the base exposure record of the slope unit: Compare the base visibility results of the same slope unit in multiple acquisition phases to determine which acquisition phase of the slope unit has the clearest base texture. Only one base texture fragment is selected as the reference source for each slope unit to avoid duplicate references at the same location due to vegetation decay and shadow changes. This step outputs the base fragment source phase of each slope unit. S4, Generate a slope base reference map: According to the position of the slope unit, stitch the base texture fragments corresponding to each slope unit into the same reference map. The slope base reference map is not the original image of any single acquisition phase, but is composed of the base texture fragments most suitable as the registration reference in different acquisition phases. This step outputs a slope base reference map for subsequent unified registration. S5. Register the slope images of each time phase to the slope base reference map and perform change detection: For a slope unit with a visible base area in a certain acquisition time phase, perform local registration based on the matching relationship between the visible base area and the slope base reference map. For slope units that are covered by ecological cover and lack a visible base area, complete the location transfer based on the registration relationship of adjacent registered slope units. After registration, detect the changes in ecological cover and base exposure of each slope unit and output the results of ecological restoration status changes.

2. The method for detecting changes in ecological restoration status based on multi-temporal slope image registration according to claim 1, characterized in that: Step S1 includes the following steps: S1-1, Image Input: Input the same slope surface in the... Slope images at the time of acquisition ,in, For the phase numbering of data collection, The value is to , The total number of time phases collected; the slope image To display a two-dimensional image of the slope ecological restoration area; S1-2, Slope monitoring area determination: In each slope image Determining the slope monitoring area The slope monitoring area is defined based on the slope boundary, slope top edge, slope toe edge, and fixed structure edge. Perform initial alignment; S1-3, Slope Unit Division: Dividing the initially aligned slope monitoring area Divided into Slope unit ,in, Number the slope unit. The value is to , The total number of slope units; the first The first acquisition phase The image fragment of each slope unit is denoted as The image fragment For slope images Extracted from the slope unit The corresponding local image; S1-4, Unit Scale Determination: Slope Unit The size is set to be larger than the image scale of a single blade of grass and local gravel, but smaller than the image scale of local degraded patches on the slope; by limiting the size, the slope unit... Capable of including identifiable base textures while preserving locational differences in local ecological restoration status, slope units The size is set to be greater than the average projected width of a single blade of grass in the image, but less than the image width of the preset minimum degradation patch.

3. The method for detecting changes in ecological restoration status based on multi-temporal slope image registration according to claim 2, characterized in that: Step S2 includes the following steps: S2-1, Pixel Category Recognition: Image Fragment Recognition The pixels in the image are classified to obtain base pixels, ecological cover pixels, and transient interference pixels. The base pixels include rock surface texture pixels, soil surface texture pixels, and sprayed substrate texture pixels. The ecological cover pixels include herbaceous vegetation pixels, shrub vegetation pixels, and withered grass pixels. The transient interference pixels include shadow pixels, water accumulation pixels, wet marks pixels, and strong reflective pixels. S2-2, Calculate the basis visibility value: For image segments The exposed area of ​​the substrate, the clarity of the substrate texture, and the degree of transient interference are uniformly quantified, and the substrate visibility value is calculated based on the quantization results. The calculation method is as follows: ,in, For the first The first acquisition phase The visible value of the base of each slope element. For image fragments The proportion of base pixels to the total number of effective pixels. For image fragments The texture sharpness normalization value of the mid-base pixel. From image fragments The gray-level gradient intensity and texture edge continuity of the base pixel are normalized; when the gray-level gradient intensity of the base pixel is high and the texture edge continuity is high... As the value of increases, For image fragments The proportion of transient interference pixels to the total number of effective pixels; the substrate visibility value The larger the value, the larger the image segment. The more suitable it is as a registration reference; S2-3, Determine candidate basis segments: When the basis visibility value Not less than the preset substrate visible threshold At that time, the image fragment Determined as a slope unit In the Candidate base segments at each acquisition phase; For image fragments The proportion of transient interference pixels to the total number of effective pixels; the substrate visibility value The larger the value, the larger the image segment. The more suitable it is as a registration reference; S2-3, Determine candidate basis segments: When the basis visibility value Not less than the preset substrate visible threshold At that time, the image fragment Determined as a slope unit In the Candidate base segments at each acquisition phase; S2-4, Remove non-registration areas: Mark the areas corresponding to ecological cover pixels and transient interference pixels as non-registration areas, so that grass leaf morphology, vegetation withering and flourishing, shadow boundaries, water accumulation boundaries and strong reflective boundaries do not participate in subsequent registration transformation calculations.

4. The method for detecting changes in ecological restoration status based on multi-temporal slope image registration according to claim 3, characterized in that: Step S3 specifically includes: S3-1, Establish a base exposure record: For the same slope unit According to the collection phase number Record its base visibility value and candidate base fragments The base exposure record is used to represent the same slope unit. The exposure of substrate texture under different acquisition phases; S3-2, Determine the phase of origin of the basal segment: This is done when obtaining the same slope unit... Baseline visible values ​​at each acquisition phase Subsequently, the acquisition phase with the highest base visibility value was determined as the base segment origin phase of that slope unit. The calculation method is as follows: ,in, For slope unit The formula represents the phase of the basal segment origin; Baseline visible values ​​were selected from each acquisition phase. The largest acquisition phase, as a slope unit Reference source, when two or more acquisition phases are used to obtain the base visibility value. When the proportions are the same, prioritize the proportion of transient interference pixels. If the smaller acquisition time phases are still the same, select the acquisition time phase with the smaller acquisition time phase number r. S3-3, Determine the base reference segment: For slope elements with at least one candidate base segment Image fragments Determined as a slope unit Base reference fragment ,in, This represents a local base image used to generate a slope base reference map; the base visibility values ​​for all acquisition phases. All are less than the preset substrate visibility threshold slope unit No base reference fragment is generated. The slope unit Marked as a unit to be transferred; S3-4, Constraint: Unique element source for each slope element. Only one base reference segment is retained. .

5. The method for detecting changes in ecological restoration status based on multi-temporal slope image registration according to claim 4, characterized in that: Step S4 specifically includes: S4-1, Establish reference map coordinates: based on the slope monitoring area at the first acquisition time phase. The unit division results are used to establish reference map coordinates, and the common overlapping area after initial alignment of each acquisition phase is used as the slope base reference map. The coordinate range; S4-2, Reference diagram for assembling slope base: according to slope unit The base reference segment is positioned in the reference graph coordinates. Place it in the corresponding position to generate a slope base reference map. ; S4-3, Processing the unit to be transferred: For segments without a base reference... The element to be transferred is based on the base reference fragments already obtained in its adjacent slope elements. Determine the reference diagram of the slope base where the unit to be transferred is located. Boundary locations and cell ranges within the range; S4-4, Retain reference fragment source: in the slope base reference diagram In the middle, there is a reference segment for each base. Record its origin phase and corresponding slope unit number This record allows for tracing the origin of each local reference area during subsequent registration error checks.

6. The method for detecting changes in ecological restoration status based on multi-temporal slope image registration according to claim 5, characterized in that: Step S5 specifically includes: S5-1, Local Matching: For an already generated base reference segment And the slope unit with base pixels in the r-th acquisition phase Image fragments Reference diagram of slope base Basis reference fragment at the corresponding position in the middle Perform local matching; the pixels participating in local matching are limited to the base pixels determined in step S2, and ecological coverage pixels and transient interference pixels do not participate in local matching; S5-2, Calculate local registration error: when the reference pixel position set Non-empty and When the value is greater than 0, the local registration error needs to be calculated. When the reference pixel position set When empty, local registration error is not calculated. and the corresponding slope unit The r-th acquisition phase is marked as a registration unit to be transferred, and the pixel difference in the base pixel set is used to calculate the... The first acquisition phase Local registration error of individual slope units The calculation method is as follows: ,in, For the first The first acquisition phase Local registration error of individual slope units Reference diagram for slope base The set of reference pixel positions that participate in the local error calculation, and the set of reference pixel positions Each pixel position in the image fragment It has corresponding base pixels, For reference pixel position set The number of pixels in Reference diagram for slope base pixel position in To provide a reference diagram of the slope base The reference pixel position x in the image is mapped to the slope image. Local sampling transformation at the corresponding pixel position in the image. Representing a slope image After local registration transformation Later in position The corresponding pixel value, Reference diagram of slope base In position The pixel values, where the local registration error Not greater than the preset registration error threshold At that time, local registration transformation is preserved. Local registration error Greater than the preset registration error threshold At that time, the corresponding slope unit The unit to be registered and transferred is marked in the r-th acquisition phase; S5-3, complete the position transfer of occlusion units: for sets of pixels occluded by ecological coverage and lacking reference pixel locations. slope unit And due to local registration errors Greater than the preset registration error threshold The marked registration units to be transferred do not directly use their own image texture to calculate the local registration transformation; instead, they are selected based on the slope unit. Adjacent and with local registration transformation preserved slope unit As a source of transmission, among them To transmit the source unit number; based on the slope unit With the source unit The relative positional relationship in the reference diagram coordinates will convey the source unit. Local registration transformation Transfer to slope unit Determine the slope unit Reference diagram of slope base The corresponding position in the slope element There are more than two source units in the vicinity. When selecting slope units, priority should be given to those that match the slope. The source unit with the longest shared boundary length When the shared boundary lengths are the same, select the source unit number. Small slope unit When the slope unit When there are no source cells in the surrounding area that have preserved the local registration transformation, the slope cell is... The units are marked as non-transmitted units, and these non-transmitted markers are retained in the results of ecological restoration status changes so that they do not participate in the determination of effective change areas.

7. The method for detecting changes in ecological restoration status based on multi-temporal slope image registration according to claim 6, characterized in that: The S5 also includes: S5-4, Detecting changes in ecological restoration status: Slope images at each acquisition time... Reference diagram of slope base After registration, calculate the slope element. In the Ecological coverage ratio at each collection time point and the proportion of exposed substrate ;in, For image fragments The proportion of ecological coverage pixels to the total number of effective pixels. For image fragments The proportion of base pixels to effective pixels: S5-5, Output Change Results: This will change the same slope unit in adjacent acquisition phases. ecological coverage ratio and the proportion of exposed substrate By comparing these data, regions of increased coverage, decreased coverage, increased base exposure, and stable conditions are identified. The final output includes registered multi-temporal slope images, slope unit change categories, and ecological restoration status change maps. These outputs characterize changes in ecological restoration status and reduce the impact of spurious changes caused by variations in image acquisition perspective. When the increase in ecological coverage ratio from the previous acquisition phase to the subsequent acquisition phase reaches a preset coverage change threshold, the corresponding slope unit is identified as an area of ​​increased coverage. Similarly, when the decrease in ecological coverage ratio from the previous acquisition phase to the next acquisition phase reaches a preset coverage change threshold, the corresponding slope unit is identified as an area of ​​decreased coverage. When the increase in base exposure ratio from the previous acquisition phase to the subsequent acquisition phase reaches a preset exposure change threshold, the corresponding slope unit is identified as an area of ​​increased base exposure. When neither the change in ecological coverage ratio nor the change in base exposure ratio reaches the corresponding threshold, the corresponding slope unit is identified as a stable condition region.

8. An ecological restoration state change detection system based on multi-temporal slope image registration, the system being based on the method of any one of claims 1-7, characterized in that: It includes an image layering module, which is used to unify the slope area, identify the visible base area, identify the ecological cover area and the transient disturbance area in multi-temporal slope images, so as to obtain the visible base area in each temporal slope image that can participate in registration; The slope base reference map generation module is used to select the image fragment with the clearest base texture for each slope unit in multiple acquisition phases and stitch the image fragments together to form a slope base reference map; The registration change detection module is used to register each temporal slope image to the slope base reference map and output the ecological restoration status change results after registration is completed.

Images