A forest carbon sink monitoring system and method based on remote sensing images

Abstract

The present application relates to carbon sink monitoring technical field, disclose a kind of forest carbon sink monitoring system and method based on remote sensing image, a kind of forest carbon sink monitoring system based on remote sensing image, comprising: construct slope direction development coordinate;Extract carbon sink sensitive profile response field;Calculate structure scale;Determine curvature inversion grain center;Calculate inversion grain warping function;Calculate reference profile and reference warping function;Calculate false turning point response field;Output annual variation and turning intensity.The present application is aimed at the time sequence false turning point problem caused by interlayer radiation curvature inversion grain of slope direction lock, effectively separates the false signal generated by real carbon sink change and image structure interference, retains the real characteristics of mountain forest carbon sink;Annual carbon sink characterization, annual variation and turning intensity obtained by monitoring are more in line with actual situation, improve the reliability of mountain forest carbon sink monitoring.

Description

Technical Field

[0001] This invention relates to the field of carbon sink monitoring technology, and more specifically, to a forest carbon sink monitoring system and method based on remote sensing imagery. Background Technology

[0002] Against the backdrop of global climate change, forest carbon sink monitoring is a core component of ecological environment assessment and carbon balance regulation. Remote sensing technology, due to its advantages of wide coverage, high timeliness, and low cost, has become the mainstream method for forest carbon sink monitoring. Mountain forests, as important terrestrial carbon sinks, exhibit dramatic topographic relief and distinct vertical vegetation distribution; their carbon sink storage and dynamic changes have a significant impact on regional and even global carbon cycles. However, the complex topography of mountain forests alters the incident path of solar radiation and the canopy radiative transfer process, posing a serious challenge to the accuracy of remote sensing monitoring.

[0003] In mountain forests, multiple scattering from the upper canopy, lower vegetation, and forest background creates complex interlayer radiation. Combined with the coupling relationship between slope aspect and solar projection direction, this produces slope-locked interlayer radiation curvature reversal fringes. These reversal fringes cause the carbon sink sensitive response to exhibit a non-monotonic profile, and their positional drift is influenced by cloud shadow displacement and seasonal changes in solar angle at different observation phases. These factors work together to cause false turning points in carbon sink time-series data; what appear to be trends of carbon sink growth or decline are actually false features created by image structure interference, rather than genuine changes in the carbon sink process.

[0004] Existing remote sensing carbon sequestration monitoring methods are mostly based on planar image processing logic or only perform simple shape correction. These methods struggle to effectively separate real carbon sequestration changes from spurious signals, leading to biases in the calculation of key indicators such as annual changes and transition intensities of carbon sequestration in mountain forests, and failing to accurately reflect dynamic trends in carbon sequestration. This problem seriously affects the reliability of carbon sequestration monitoring results, hindering the provision of scientific data support for the management and assessment of carbon sequestration in mountain forests, and becoming a bottleneck restricting the development of remote sensing monitoring technology for carbon sequestration in mountain forests. Summary of the Invention

[0005] This invention provides a forest carbon sequestration monitoring system and method based on remote sensing imagery, which solves the technical problems mentioned in the background.

[0006] This invention provides a forest carbon sequestration monitoring system based on remote sensing imagery, comprising: The first module converts the two-dimensional image coordinates into slope-oriented unfolded coordinates based on the elevation and the direction of solar projection. The second module uses the red light band reflectance and near-infrared band reflectance to calculate and extract the carbon sink sensitive profile response field from the slope aspect unfolded coordinates. The third module uses the directional autocorrelation function to determine the first zero-crossing position of the response field of the carbon sink sensitive profile as the structural scale. The fourth module obtains a smooth response based on the response field of the carbon sink sensitive profile at the structural scale, and defines the zero point of the second derivative closest to the steepest darkening point in the smooth response as the center of the curvature inversion ripple. The fifth module constructs a chord reference based on the curvature inversion ripple center and structural scale on the smooth response, and calculates the deviation of the smooth response relative to the chord reference to obtain the inversion ripple warping function; The sixth module establishes a reference coordinate system with the center of the curvature inversion pattern as the zero point, and translates the smooth response and the inversion pattern warping function to the reference coordinate system to obtain the reference profile and the reference warping function. The seventh module generates support weights based on the reference coordinates and structural scale, subtracts the product of the support weights and the reference warping function from the reference profile, and maps the subtraction result back to the two-dimensional image coordinates to obtain the pseudo-turn response field. The eighth module performs annual integral processing on the pseudo-transition response field to generate annual integral values, and performs logarithmic compression and continuous difference calculation on the annual integral values ​​to output the annual change and transition intensity.

[0007] This invention provides a method for monitoring forest carbon sequestration based on remote sensing imagery, comprising the following steps: Step S1: Convert the two-dimensional image coordinates into slope unfolded coordinates based on the elevation and solar projection direction; Step S2: Calculate and extract the carbon sink sensitive profile response field using the red band reflectance and near-infrared band reflectance in the slope aspect unfolded coordinates; Step S3: Calculate the first zero-crossing position of the response field of the carbon sink sensitive profile based on the directional autocorrelation function as the structural scale; Step S4: Obtain a smooth response based on the response field of the carbon sink sensitive profile at the structural scale, and define the zero point of the second derivative closest to the steepest darkening point in the smooth response as the center of the curvature inversion ripple. Step S5: Construct a chord reference on the smooth response using the curvature inversion ripple center and structural scale, and calculate the deviation of the smooth response relative to the chord reference to obtain the inversion ripple warping function. Step S6: Establish a reference coordinate system with the center of the curvature inversion pattern as the zero point, and translate the smooth response and the inversion pattern warping function to the reference coordinate system to obtain the reference profile and the reference warping function. Step S7: Generate support weights based on the reference coordinates and structural scale, subtract the product of the support weights and the reference warping function from the reference profile, and map the subtraction result back to the two-dimensional image coordinates to obtain the pseudo-turn response field. Step S8: Perform annual integral processing on the pseudo-transition response field to generate annual integral quantity, and perform logarithmic compression and continuous difference calculation on the annual integral quantity to output the annual change and transition intensity.

[0008] The beneficial effects of this invention are as follows: This invention constructs aspect-deployed coordinates, extracts the response field of carbon sink sensitive profiles, adaptively acquires structural scale, locates the center of curvature inversion fringes, quantifies the warping characteristics of inversion fringes, establishes benchmark coordinates to eliminate positional drift, and removes false transitions through weighted decoupling. Finally, it calculates carbon sink-related indicators through annual integration and difference. This invention specifically addresses the temporal false transition problem caused by aspect-locked interlayer radiation curvature inversion fringes, effectively separating real carbon sink changes from false signals generated by image structure interference, and preserving the true characteristics of mountain forest carbon sinks. The annual carbon sink characterization, annual variation, and transition intensity obtained from monitoring are more consistent with actual conditions, improving the reliability and consistency of mountain forest carbon sink monitoring and providing reliable data support for mountain forest carbon sink assessment. Attached Figure Description

[0009] Figure 1 This is a flowchart of a forest carbon sequestration monitoring method based on remote sensing imagery according to the present invention. Detailed Implementation

[0010] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0011] It should be noted that, unless otherwise defined, the technical or scientific terms used in one or more embodiments of the present invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in one or more embodiments of the present invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" indicate that the element or object preceding the term encompasses the elements or objects listed following the term and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0012] like Figure 1 As shown, a forest carbon sequestration monitoring system based on remote sensing imagery includes: The first module converts the two-dimensional image coordinates into slope-oriented unfolded coordinates based on the elevation and the direction of solar projection. The second module uses the red light band reflectance and near-infrared band reflectance to calculate and extract the carbon sink sensitive profile response field from the slope aspect unfolded coordinates. The third module uses the directional autocorrelation function to determine the first zero-crossing position of the response field of the carbon sink sensitive profile as the structural scale. The fourth module obtains a smooth response based on the response field of the carbon sink sensitive profile at the structural scale, and defines the zero point of the second derivative closest to the steepest darkening point in the smooth response as the center of the curvature inversion ripple. The fifth module constructs a chord reference based on the curvature inversion ripple center and structural scale on the smooth response, and calculates the deviation of the smooth response relative to the chord reference to obtain the inversion ripple warping function; The sixth module establishes a reference coordinate system with the center of the curvature inversion pattern as the zero point, and translates the smooth response and the inversion pattern warping function to the reference coordinate system to obtain the reference profile and the reference warping function. The seventh module generates support weights based on the reference coordinates and structural scale, subtracts the product of the support weights and the reference warping function from the reference profile, and maps the subtraction result back to the two-dimensional image coordinates to obtain the pseudo-turn response field. The eighth module performs annual integral processing on the pseudo-transition response field to generate annual integral values, and performs logarithmic compression and continuous difference calculation on the annual integral values ​​to output the annual change and transition intensity.

[0013] In one embodiment of the present invention, converting two-dimensional image coordinates into slope-oriented unfolded coordinates based on elevation and solar projection direction includes: Calculate the slope gradient components in two-dimensional image coordinates. Obtaining elevation Calculate elevation Two-dimensional image coordinates direction and The partial derivatives in the direction yield the slope gradient components. and : ; ; in For position Elevation value, For elevation along Rate of change of direction For elevation along Rate of change of direction; Calculate the unit vector of the sun's projection direction at the th Each observation phase is based on the solar zenith angle. With solar azimuth Calculate the unit vector of the sun's projection direction. : ; in For the first The solar zenith angle at each observation phase. For the first The solar azimuth angle at each observation phase. and Unit vectors of the sun's projection direction In two-dimensional image coordinates direction and Component of direction; Calculate the slope aspect unfolded coordinates based on the unit vector of the solar projection direction. Two-dimensional image coordinates Mapped to slope aspect expansion coordinates : ; ; in For two-dimensional image coordinates, These are the principal axis components of the slope aspect expansion coordinate system. These are the orthogonal components of the slope aspect expansion coordinates.

[0014] It should be noted that 2D image coordinates are the planar positioning identifiers of each pixel in a remote sensing image, reflecting the pixel's specific location within the image. These coordinates can be obtained through georegistration information of the remote sensing image. Commonly used coordinate systems include the Universal Mercator coordinate system and the World Geodetic System 1984 coordinate system. Elevation values ​​are the elevation data corresponding to the Earth's surface location in the 2D image coordinates, reflecting the surface's undulation. These can be obtained through digital elevation model data. The rate of change of elevation values ​​with respect to the horizontal axis of the 2D image coordinates is the instantaneous degree of change of elevation values ​​along the horizontal axis, reflecting the slope steepness trend of the Earth's surface in the horizontal direction. The rate of change of elevation values ​​with respect to the vertical axis of the 2D image coordinates is the instantaneous degree of change of elevation values ​​along the vertical axis, reflecting the slope steepness trend of the Earth's surface in the vertical direction. The slope gradient component is vector data composed of the rates of change of elevation values ​​with respect to the horizontal and vertical axes, reflecting comprehensive information about the slope's tilt direction and steepness. The observation phase is the time marker for the acquisition of a remote sensing image, reflecting the specific moment the image was acquired. It can be obtained from the image's metadata, which typically records the date and time of acquisition. The solar zenith angle is the angle between the sun's rays and the direction of the Earth's surface normal, reflecting the sun's altitude in the sky. It can be calculated using astronomical algorithms, inputting parameters such as the observation phase and the latitude and longitude of the observation area; alternatively, it can be extracted directly from the image's metadata. The solar azimuth angle is the angle between the projection of the sun's rays onto the Earth's surface and true north, reflecting the sun's horizontal projection direction on the Earth's surface. It is obtained in the same way as the solar zenith angle. The unit direction vector of the sun's projection onto the 2D image coordinate plane is the standardized vector of the sun's illumination direction on the 2D image plane, reflecting the direction information of the sun's projection direction, and its length is 1. The solar projection direction unit vector is a shorthand for the unit direction vector of the sun's projection onto the 2D image coordinate plane, reflecting the standardized direction information of the sun's projection. The principal axis components of the slope aspect unfolded coordinates are the coordinate components in the slope aspect unfolded coordinates that are consistent with the sun's projection direction, reflecting the relative position of the pixel along the sun's projection direction. The orthogonal component of the slope aspect unfolded coordinates is the coordinate component perpendicular to the direction of the sun projection in the slope aspect unfolded coordinates, reflecting the relative position of the pixel perpendicular to the direction of the sun projection.

[0015] It should be noted that the elevation values ​​are preferentially selected from the Space Shuttle Radar Topographic Mapping Mission Data with a resolution of 30 meters or the Advanced Spaceborne Thermal Emission and Reflection Radiometer Global Digital Elevation Model. This resolution is sufficient for the analysis of mountainous forest slope structures, with an accuracy error within 10 meters. The solar zenith angle and solar azimuth angle are preferentially calculated using astronomical algorithms. The steps are as follows: obtain the latitude and longitude of the observation area and the Coordinated Universal Time (UTC) of the observation phase; calculate astronomical parameters such as solar declination and hour angle; and then calculate using spherical trigonometric formulas. Alternatively, they can be directly extracted from the metadata of remote sensing images. For example, if the observation area has a longitude of 110.5 degrees and a latitude of 28.5 degrees, and the observation phase is 10:00 on July 15, 2023, which is converted to UTC 2:00, the calculated solar zenith angle is approximately 35 degrees and the azimuth angle is approximately 120 degrees. The quantification criteria for the principal axis components and orthogonal components of the slope aspect unfolded coordinates are that the coordinate unit is consistent with the original image coordinates, and the value range is determined by the size of the original image and the direction of the sun projection. For example, if the original image range is 1000m × 1000m and the sun projection direction is northeast, the principal axis component value range is approximately -707m to 1414m, and the orthogonal component value range is approximately -707m to 707m. Pixels on the same slope are continuously distributed in the direction of the principal axis component, and the difference in the principal axis component of adjacent pixels does not exceed 1.5 times the pixel spacing of the image.

[0016] It should be noted that terrain elevation determines the slope's inclination, while the solar projection direction reflects the horizontal direction of sunlight. Combining these two factors allows the converted coordinates to highlight the slope's structural features along the solar projection direction. In mountainous forests, aspect-locked interlayer radiation curvature inversion fringes are stably distributed along the solar projection direction. Converting the two-dimensional image coordinates to aspect-unfolded coordinates based on elevation and solar projection direction concentrates the originally dispersed inversion fringes along the principal axis component, facilitating subsequent analysis. Rotation mapping aligns the original image coordinates with the solar projection direction using a coordinate rotation matrix. The original coordinate axes are fixed, and the rotation matrix is ​​constructed based on the components of the unit vector along the solar projection direction. After rotation, the principal axis component aligns with the solar projection direction, while the orthogonal component is perpendicular to this direction. This rotation does not change the relative pixel positions, only adjusting the coordinate reference direction, making the slope structure along the solar projection direction easier to identify. For example, if the original coordinate axis is east and the vertical axis is north, and the solar projection direction is southeast, after rotation, the principal axis component extends along the southeast direction, allowing the slope's structural features along this direction to be continuously presented.

[0017] It should be noted that this invention combines elevation and solar projection direction to transform the coordinates of two-dimensional images, converting the original coordinates into slope aspect unfolded coordinates. This allows the slope aspect-locked interlayer radiation curvature inversion ripple structures distributed along the solar projection direction to be concentrated and presented. This solves the problem of scattered and difficult-to-identify inversion ripple structures in the original two-dimensional image coordinates, laying the foundation for subsequent extraction of the response field of carbon sequestration sensitive profiles. The coordinate transformation process does not change the relative positional relationship of pixels, preserving the integrity of the slope structure. The transformed principal axis components and orthogonal components clearly distinguish between the solar projection direction and the vertical direction, facilitating subsequent targeted analysis of profile features along the solar projection direction and improving the targeting and effectiveness of false inversion removal in mountain forest carbon sequestration monitoring.

[0018] In one embodiment of the present invention, the carbon sink sensitive profile response field is calculated and extracted from the slope aspect unfolded coordinates using red light band reflectance and near-infrared band reflectance, including: Calculate the carbon sink sensitive response quantity, at the first Each observation phase, for any two-dimensional image coordinate Based on the reflectivity of the red light band Near-infrared band reflectivity Calculate the carbon sink sensitive response quantity : ; in For the first Each observation phase in two-dimensional image coordinates Reflectivity in the red band at that location For the first Each observation phase in two-dimensional image coordinates Near-infrared reflectance at that location For the first Each observation phase in two-dimensional image coordinates Carbon sequestration sensitive response quantity at the location; Extract the response field of the carbon sink sensitive profile and select a fixed profile location marker in the slope aspect expansion coordinate system. Extracting continuous position variables within the profile Response field of carbon sink sensitive profile : ; in To meet and Two-dimensional image coordinates under constraints, For the first Two-dimensional image coordinates of each observation phase Principal axis components in the aspect ratio expansion coordinate system For the first Two-dimensional image coordinates of each observation phase Orthogonal components in the aspect ratio expansion coordinate system For fixed section location markings, For continuous positional variables within the profile, For the first Each observation phase is marked at the profile location. The response field of the carbon sink sensitive profile on the surface.

[0019] It should be noted that red band reflectance is the surface reflectance coefficient corresponding to the red band in remote sensing images, reflecting the surface's ability to reflect red light. It can be extracted from multispectral remote sensing images, commonly using satellites such as Sentinel-2 and Landsat-8. The images need to undergo atmospheric correction and radiometric calibration. Near-infrared band reflectance is the surface reflectance coefficient corresponding to the near-infrared band in remote sensing images, reflecting the surface's ability to reflect near-infrared light. It is obtained in the same way as red band reflectance, extracted from multispectral remote sensing images and then corrected. The preset weighting coefficients are fixed numerical coefficients set to highlight carbon sink sensitive information, reflecting the contribution weight of different band reflectance in the calculation of carbon sink sensitive response. Preferred values ​​are 2.5, 2.4, and 1. This set of values ​​can effectively suppress soil background interference and high canopy saturation problems, aligning with the calculation logic of Enhanced Vegetation Index II and adapting to the extraction of carbon sink sensitive features in mountain forests. The carbon sequestration sensitive response is a quantitative indicator calculated using red band reflectance, near-infrared band reflectance, and preset weighting coefficients, reflecting the sensitivity of surface vegetation to carbon sequestration activity. The profile location markers are fixed values ​​selected on the orthogonal components of the aspect-expanded coordinate system, reflecting the specific location of the carbon sequestration sensitive profile. Preferred values ​​are evenly spaced values ​​on the orthogonal components of the aspect-expanded coordinate system, with an interval distance of 5 to 10 times the pixel spacing of the remote sensing image, ensuring uniform coverage of the entire image area and comprehensive capture of slope structure features at different locations. Continuous position variables within the profile are continuous values ​​on the principal axis components of the aspect-expanded coordinate system, reflecting the pixel's position along the carbon sequestration sensitive profile. The carbon sequestration sensitive profile response field is a one-dimensional data field formed by the distribution of carbon sequestration sensitive response quantities along the profile in the aspect-expanded coordinate space, reflecting the continuous changes in carbon sequestration sensitive characteristics along the profile direction.

[0020] It should be noted that the carbon sequestration activity of vegetation is closely related to the chlorophyll content of leaves and the canopy structure. Chlorophyll strongly absorbs red light and strongly reflects near-infrared light, while non-vegetated surfaces (such as soil and rocks) show little difference in reflectance between the two. By highlighting vegetation information through the difference between near-infrared and red light reflectance, and then adjusting their proportions in the calculation by preset weighting coefficients, the interference of non-vegetated background and the reflection saturation phenomenon of high-coverage canopies are suppressed. The resulting carbon sequestration sensitive response can accurately correlate with vegetation carbon sequestration activity. The orthogonal component of the aspect unfolded coordinate is perpendicular to the solar projection direction. Fixing the profile position marker in this direction can lock a slope profile extending along the solar projection direction. The continuous position variables within the profile correspond to the principal axis component of the aspect unfolded coordinate. Extracting the carbon sequestration sensitive response along this component can form a continuous profile data field. The interlayer radiation curvature reversal fringes locked by the aspect in mountain forests are distributed along the solar projection direction. This extraction method can completely present the continuous structure of the reversal fringes in the profile response field, which is convenient for subsequent identification and analysis. The red light reflectance is preferentially selected in the 620 nm to 670 nm band, which is a sensitive band for chlorophyll absorption in vegetation, effectively distinguishing between vegetated and non-vegetated areas and meeting the needs of carbon sink sensitive feature extraction. The near-infrared reflectance is preferentially selected in the 780 nm to 2500 nm band, which is a sensitive band for vegetation reflection, accurately reflecting differences in vegetation canopy structure and carbon sink activity. The preset weighting coefficients are 2.5, 2.4, and 1. 2.5 is used to amplify the reflectance difference between near-infrared and red light, 2.4 is used to suppress background interference in the red light band, and 1 is a stabilizing factor to avoid calculation errors due to an excessively small denominator. The selected interval for profile location markers is 5 to 10 times the pixel spacing of the remote sensing image. For example, when the image pixel spacing is 10 meters, the interval is 50 to 100 meters. The selected number must ensure that the coverage of adjacent profiles does not overlap or omit any, covering the entire range of orthogonal components of the image. In addition, the range of values ​​for continuous positional variables within the profile is consistent with the range of values ​​for the principal axis components of the slope aspect unfolded coordinate system, extending from the minimum to the maximum value of the principal axis components. The step size is consistent with the pixel spacing of the remote sensing image, ensuring the continuity and integrity of the profile response field.

[0021] It should be noted that this invention utilizes the difference in reflection of red and near-infrared light by vegetation, combined with preset weighting coefficients, to calculate the carbon sink sensitive response, highlighting vegetation carbon sink activity information. Then, based on slope aspect unfolded coordinates, the profile response field is extracted through fixed profile location markers and continuous location variables, centrally presenting the carbon sink sensitive characteristics along the solar projection direction. This suppresses interference from soil background and canopy saturation, making the carbon sink sensitive information more prominent. The profile response field extracts one-dimensional structural features from two-dimensional images, facilitating subsequent structural scale calculations and curvature inversion fringe localization. Adapting to the distribution characteristics of slope aspect-locked interlayer radiation curvature inversion fringes provides an analytical basis for subsequent false inversion removal, improving the targeting and reliability of mountain forest carbon sink monitoring.

[0022] In one embodiment of the present invention, the first zero-crossing position of the response field of the carbon sink sensitive profile is obtained as the structural scale based on the directional autocorrelation function, including: Calculate the mean value of the profile to satisfy the condition for the first... Carbon sink sensitive profile response field at each observation time phase In the domain of the section definition Calculate the mean value of the above profile : ; in For the first Each observation phase is marked at the profile location. Carbon sink sensitive profile response field on Mark the location of the section. The corresponding cross-sectional domain, Define the domain of the profile Length, The average value is the cross-sectional value. Construct a directional autocorrelation function that satisfies the position displacement along the profile direction. Calculate the directional autocorrelation function for the variable. : ; in This represents the positional displacement along the cross-sectional direction. Here is the directional autocorrelation function; Extract the first zero-crossing position as the structural scale, satisfying the condition that... Under the condition of [condition], determine the directional autocorrelation function. The minimum displacement position when the initial value is zero is used as the structural dimension. : ; in For structural scale.

[0023] It should be noted that the profile domain is the effective range of values ​​for the response field of the carbon sink sensitive profile, reflecting the actual coverage of the carbon sink sensitive features in the profile. The profile mean reflects the overall baseline level of the profile response quantity. The positional displacement is the distance offset value set along the direction of the carbon sink sensitive profile, reflecting the interval length between the original position and the comparison position of the profile. The directional autocorrelation function is a function describing the structural similarity of the response field of the carbon sink sensitive profile under different positional displacements, reflecting the repetitive pattern and spatial correlation of the profile structure. The structural scale is the minimum positional displacement when the directional autocorrelation function first takes a value of zero, reflecting the natural width of the aspect-locked interlayer radiation curvature reversal fringes in the carbon sink sensitive profile. The reversal patterns in the carbon sink sensitive profile are structures with locally repeating characteristics. When the displacement is small, the profile response fields at the original position and the displacement position are within the same reversal pattern range, exhibiting high structural similarity and an autocorrelation function value close to 1. As the displacement increases, when it exceeds the natural width of the reversal pattern, the profile response field at the displacement position leaves the original reversal pattern range and enters an adjacent structural region, significantly reducing similarity and gradually decreasing the autocorrelation function value. When the displacement reaches the natural width of the reversal pattern, the similarity completely disappears, and the autocorrelation function value drops to zero. Furthermore, the natural structural scale of the reversal pattern, i.e., its maximum width, maintains continuity and similarity within this width range. Beyond this width, the structural characteristics undergo a fundamental change, and the similarity disappears. The minimum displacement at which the directional autocorrelation function first reaches zero corresponds precisely to the critical position where the reversal pattern transitions from structural continuity to structural breakage. This displacement is consistent with the natural width of the reversal pattern and is therefore determined as the structural scale.

[0024] It should be noted that the core of this study is based on the effective data of the carbon sink sensitive profile response field, removing regions with anomalous jumps in response values ​​at both ends. The criterion for anomalous jumps is that the difference in carbon sink sensitive response values ​​between adjacent pixels exceeds 0.5. The starting position of the domain is the first pixel position on the left end of the profile that does not exhibit an anomalous jump, and the ending position is the first pixel position on the right end of the profile that does not exhibit an anomalous jump. If there are no anomalous jump regions in the profile, the domain is the entire pixel coverage area of ​​the profile. The step size of the position displacement is consistent with the pixel spacing of the remote sensing image. For example, if the pixel spacing of the remote sensing image is 10 meters, the step size of the position displacement is 10 meters. The value range starts from 10 meters, and the maximum value does not exceed half the length of the profile domain to avoid excessive displacement causing the integration area to exceed the profile domain, resulting in invalid calculations. The threshold for determining the first zero-crossing position is that the directional autocorrelation function value is considered zero if it falls within the range of -0.01 to 0.01, with an allowable error range of 0.01. As the position displacement gradually increases, if the function value corresponding to a certain displacement falls into this range for the first time, and the function values ​​corresponding to subsequent adjacent displacements remain in this range or far from zero, then this displacement is the first zero-crossing position. The validity verification of the structural scale adopts the second derivative zero-point verification method. The number of second derivative zeros of the response field of the carbon sink sensitive profile is calculated within the interval corresponding to the structural scale (i.e., the range from zero displacement to the structural scale). If at least one zero exists, the structural scale is considered valid. If there are no second derivative zeros in the interval, the maximum value of the position displacement needs to be expanded to two-thirds of the length of the profile domain, the directional autocorrelation function is recalculated, and the first zero-crossing position is extracted until a valid structural scale is obtained.

[0025] It should be noted that this invention eliminates the overall offset of the response field of carbon sink sensitive profiles by calculating the profile mean, constructs a directional autocorrelation function to capture the variation law of profile structural similarity with positional displacement, and extracts the first zero-crossing position to adaptively obtain the natural structural scale of the inversion ripples. This avoids the subjectivity of manually preset structural scales, ensuring that the structural scale matches the actual shape of the inversion ripples. The calculation logic of the directional autocorrelation function conforms to the spatial distribution characteristics of the profile structure, ensuring accurate characterization of similarity changes. The determination threshold of the first zero-crossing position and the validity verification standard of the structural scale guarantee the reliability and stability of the structural scale, providing a quantitative basis for subsequent smoothing processing and warp function construction, thus improving the operability and reliability of the entire carbon sink monitoring system.

[0026] In one embodiment of the present invention, a smooth response is obtained based on the response field of the carbon sink sensitive profile at the structural scale, and the zero point of the second derivative closest to the steepest darkening point in the smooth response is defined as the center of curvature inversion ripples, including: Calculate the smooth response to satisfy the condition for the th Observation time phase and profile location markers Corresponding carbon sink sensitive profile response field , in terms of structural scale Calculate the smooth response for the smooth window half-width : ; in For the response field of carbon sink sensitive profile, For structural scale. For integration variables, For a smooth response; Determine the steepest darkening point to satisfy the calculation of smooth response. first derivative And determine the steepest darkening point. : ; ; in The first derivative of the smooth response, This is the steepest point of darkening; Determine the center of curvature inversion fringes to satisfy the calculation of smooth response. The second derivative And determine the center of the curvature reversal pattern. : ; ; in The second derivative of the smooth response, The center of the curvature reversal pattern.

[0027] It should be noted that the half-width of the smoothing window is half the width of the smoothing calculation window set based on the structural scale. The smoothed response is the smoothed data obtained after local average integration of the carbon sink sensitive profile response field, reflecting the overall trend of the profile response field and eliminating noise interference. The first derivative is the rate of change of the smoothed response with respect to the positional variables within the profile, reflecting the slope change of the smoothed response. The steepest darkening point is the location with the smallest value in the first derivative distribution, reflecting the region with the fastest darkening rate in the profile response field. The second derivative is the rate of change of the first derivative with respect to the positional variables within the profile, reflecting the curvature change of the smoothed response. The set of locations where the second derivative is zero is the set of all locations where the second derivative of the smoothed response is zero, reflecting all potential locations of curvature inversion in the smoothed response. The center of the curvature inversion ripple is the location with the smallest distance from the steepest darkening point in the set of locations where the second derivative is zero, reflecting the core location of the aspect-locked interlayer radiation curvature inversion ripple.

[0028] It should be noted that the first derivative of the smoothing response directly quantifies the darkening rate. Due to interlayer radiative coupling, the darkening process in the inversion fringe region exhibits significant extrema. The location corresponding to the minimum value of the first derivative is the point with the fastest darkening rate. This point is physically closely related to the center of the inversion fringe and is a key feature correlation point for the inversion fringe, providing a reliable benchmark for subsequent center localization. The zero point of the second derivative is the critical position where the profile curvature reverses. The core curvature reversal point of the inversion fringe is necessarily physically related to the steepest darkening point; that is, the curvature reversal point near the point of fastest darkening is precisely the structural core of the inversion fringe. The local average integration operation adopts the sliding window integration method. The window is centered on the current pixel and covers all pixels from the current pixel minus half the width of the smoothing window to plus half the width of the smoothing window. During integration, the arithmetic mean of all carbon sink sensitive response values ​​within the window is calculated, and this average value is used as the smoothing response value of the current pixel. If the pixels at the window edge are insufficient to cover the window range, the average value of the effective pixels is used to ensure the reasonableness of the smoothing results in the edge region. The first derivative is calculated using the central difference method, with the positional interval between adjacent pixels being the step size of the positional variable within the profile. The calculation result is retained to four decimal places, with an error not exceeding 0.001. If an abnormal jump occurs, i.e., the difference between adjacent derivatives exceeds 0.1, the smoothed response needs to be re-integrated, and the pixel weights within the window adjusted before smoothing again. The second derivative is calculated based on the first derivative result using the central difference method. After calculation, the second derivative needs to be smoothed using a 3-pixel moving average method to avoid false zeros caused by noise and ensure the authenticity of the curvature inversion position.

[0029] It should be noted that this invention uses a smoothing window set at the structural scale to smooth the response field of the carbon sink sensitive profile to suppress noise interference. The steepest darkening point with the fastest darkening speed is located using the first derivative, and then the distance relationship between the zero point of the second derivative and this point is combined to accurately pinpoint the center of the curvature reversal ripple. The smoothing window matches the natural scale of the reversal ripple, preserving its core features while filtering out invalid noise. The steepest darkening point located by the first derivative provides a reliable benchmark for center positioning, reducing false position interference. The selection criteria for the zero point of the second derivative and the distance determination rules ensure the accuracy and stability of the reversal ripple center positioning, providing precise core position parameters for subsequent warp function construction and false inflection point removal, ensuring the consistency and reliability of the entire carbon sink monitoring process.

[0030] In one embodiment of the present invention, a chord reference is constructed on the smooth response using the curvature inversion ripple center and structural scale, and the deviation of the smooth response relative to the chord reference is calculated to obtain the inversion ripple warping function, including: Determine the reference interval for the chord, satisfying the condition for the first... Smooth response of each observation phase With the curvature reversal pattern center Centered on structural scale Define the interval for half-width constraint: ; in For the first The center of curvature reversal fringes corresponding to each observation time, For structural scale; Construct a chord reference that satisfies the following formula for calculating the chord reference within the defined interval. : ; in To smooth the response at the left endpoint, To smooth the value of the response at the right endpoint, Use the chord as a reference; Calculate the reverse fringe warpage function, satisfying the following formula within the defined interval. : ; in To smooth the response, This is the warping function for reverse ripples.

[0031] It should be noted that the chord reference definition interval is a computational range defined by the center of the curvature reversal fringes as the midpoint and the structural dimension as half-width, reflecting the effective calculation range of the reversal fringes warping function. The chord reference is a straight-line reference line obtained by linearly connecting the smooth response values ​​of the left and right endpoints of the chord reference definition interval, reflecting the ideal monotonic change trend of the section response when there is no reversal fringes interference. The reversal fringes warping function is the set of deviation values ​​between the smooth response and the chord reference within the chord reference definition interval, reflecting the degree of section curvature bending caused by the aspect-locked interlayer radial curvature reversal fringes. The curvature reversal line center is the core location of the reversal line, and the structural scale is the natural width of the reversal line. The interval defined by the combination of the two can completely cover the bending range of the reversal line. In an ideal state, the profile response changes along a monotonic trend when there is no reversal line. This ideal trend without interference can be simulated by constructing a chord reference by linearly connecting the values ​​at the two ends of the interval, providing a reference for quantifying the bending of the reversal line. For example, if the curvature reversal line center is located at 500 meters in the profile, the structural scale is 40 meters, and the defined interval is 460 meters to 540 meters, the chord obtained by connecting the smooth response values ​​at the two ends of this interval is the ideal trend line when there is no reversal line in this area.

[0032] It should be noted that the specific method for defining the range of the chord reference is to take the cross-sectional position corresponding to the center of the curvature reversal ripple as the midpoint, and extend it to the left and right by the number of pixels corresponding to the structural scale to determine the range boundary. The pixel-level determination rule is to divide the structural scale by the pixel spacing of the remote sensing image to obtain the number of extended pixels; for example, if the structural scale is 40 meters and the pixel spacing is 10 meters, extending 4 pixels to the left and right, the range boundary is the range of pixels from the center pixel position of the curvature reversal ripple minus 4 to plus 4 pixels. This invention delineates the effective range based on the core position and natural width of the curvature reversal ripple, constructs a chord reference that simulates an interference-free ideal trend, and quantifies the bending characteristics of the reversal ripple by calculating the deviation between the smooth response and the reference. The chord reference can simulate the ideal changing trend, providing a reliable reference for the calculation of the warp function; the warp function can clearly separate and quantify the bending caused by the reversal ripple, avoiding confusion between normal trends and pseudo-turns; the clarity of the defined range and calculation rules ensures the consistency and effectiveness of the warp function.

[0033] In one embodiment of the present invention, a reference coordinate system is established with the center of the curvature reversal pattern as the zero point. The smooth response and the warping function of the reversal pattern are translated to the reference coordinate system to obtain a reference profile and a reference warping function, including: Establish a reference coordinate system to satisfy the requirements of the first... Position change within the profile of each observation phase Center of curvature reversal pattern Calculate the reference coordinates for the zero point : ; in For positional variables within the profile, Center of curvature reversal pattern Used as the reference coordinates; Calculate the reference profile to satisfy the smooth response. Calculate the reference section using the following formula. : ; in To smooth the response, As the reference section; Calculate the baseline warp function to satisfy the reverse warp function. Calculate the baseline warp function using the following formula. : ; in For the reverse warp function, The baseline warping function is used.

[0034] It should be noted that the reference coordinates are relative coordinates established with the center of the curvature reversal pattern as the origin, reflecting the spatial offset relationship of each position within the profile relative to the center of the reversal pattern. The spatial position offset is the difference between any position variable within the profile and the position of the center of the curvature reversal pattern, reflecting the deviation distance and direction of that position relative to the center of the reversal pattern. The reference profile is the profile data obtained by translating the smooth response to the reference coordinate system, reflecting the relative distribution characteristics of the smooth response after translation, while preserving the original shape. The reference warp function is the function obtained by translating the reversal pattern warp function to the reference coordinate system, reflecting the relative distribution of the bending characteristics of the reversal pattern after translation, and precisely matching the reference profile. Under different observation phases, the absolute position of the same reversal pattern will drift due to cloud shadow displacement and changes in solar angle, but the distribution law of the core structure relative to its own center remains unchanged. Establishing the reference coordinates with the center of the reversal pattern as the origin can unify the reversal patterns of different phases into the same relative coordinate system, eliminating the influence of absolute position drift and making the structural characteristics of the reversal patterns of different phases comparable. The reference coordinates range from negative to positive structural scales, consistent with the definition range of the chord reference. The specific steps for translating the smooth response to the reference coordinates are as follows: First, extract the original position and corresponding value of each pixel in the smooth response; second, calculate the spatial offset between the original position of each pixel and the center position of the curvature inversion fringe, using this as the reference coordinate for that pixel; third, map the original values ​​to the reference coordinates to form a reference profile; during the translation process, the values ​​remain unchanged, only the coordinate labels are updated. Furthermore, the specific steps for translating the inversion fringe warp function to the reference coordinates are consistent with the smooth response translation logic and will not be elaborated here.

[0035] It should be noted that this invention addresses the issue of positional drift of inverted ripples at different times by establishing a reference coordinate system with the center of the inverted ripple as the origin. The smoothing response and the warping function of the inverted ripple are then translated to this coordinate system, achieving relative uniformity of structural characteristics. This effectively eliminates the interference of absolute positional drift, making the inverted ripple structures at different times comparable. After translation, the reference profile and the reference warping function are precisely matched, providing a unified coordinate basis for subsequent uncoiling and removal of false inflections. The operation only changes the coordinate reference, without destroying the original data form, ensuring the authenticity and accuracy of the analysis and improving the coherence and reliability of the entire carbon sequestration monitoring process.

[0036] In one embodiment of the present invention, support weights are generated based on reference coordinates and structural scales. The product of the support weights and the reference warping function is subtracted from the reference profile. The subtraction result is then mapped back to two-dimensional image coordinates to obtain a pseudo-corrected transition response field, including: Generate supporting weights to satisfy the requirements of the first... Observation time phase and profile location markers Corresponding reference coordinates , in terms of structural scale Calculate the support weight for the half-width of the support interval. : ; in As the reference coordinates, For structural scale To support the weight; Subtracting the product of the support weight and the datum warping function satisfies the datum profile. Warping function with reference Perform the operation to obtain the subtraction result : ; in As the reference section, As the baseline warping function, The result is the difference between the two. The mapping yields a pseudo-transition response field that satisfies the coordinates of the two-dimensional image. Corresponding slope aspect expansion coordinate principal axis components and the center of the curvature reversal pattern By performing mapping, we obtain the pseudo-transition response field. : ; in For two-dimensional image coordinates, The principal axis components of the slope aspect expansion coordinate system are... Center of curvature reversal pattern To eliminate false transition response fields.

[0037] It should be noted that the support weight is a linear attenuation value (1 minus the ratio of the absolute value of the reference coordinate to the structural scale) generated based on the reference coordinates and structural scale. This reflects the correction intensity at different locations of the inversion ridges, with higher weights in the central region and lower weights in the peripheral region. The product of the support weight and the reference warping function is a set of values ​​obtained by multiplying the support weight and the reference warping function position by position, reflecting the quantized value of the inversion ridge bending after correction intensity adjustment. The subtraction result reflects the ideal profile response trend after removing pseudo-turns of the inversion ridges. The reference coordinate input value is the difference between the principal component of the slope aspect unfolded coordinates corresponding to the two-dimensional image coordinates and the center of the curvature inversion ridge, reflecting the relative position of the image coordinates with respect to the center of the inversion ridge. The pseudo-turn response field is the response field formed after mapping the subtraction result back to the two-dimensional image coordinates, reflecting the true carbon sink sensitive response distribution after removing pseudo-turns of the slope aspect-locked interlayer radiation curvature inversion ridges.

[0038] It should be noted that the pseudo-turning effect of the inversion ridge is strongest in the central region and gradually weakens towards the edge. The linear decay weight can match this effect gradient. The weight of the reference coordinate origin (center of the inversion ridge) is 1, corresponding to the strongest correction. When the absolute value of the reference coordinate is equal to the structural scale, the weight is 0, corresponding to no correction, to avoid overcorrection of the normal slope trend. The product term is the quantized value of bending after adjustment of the correction intensity. The reference profile includes the bending of the inversion ridge and the normal trend. After subtracting the product term, the local bending caused by the inversion ridge is removed, and only the normal slope trend is retained. In addition, the reference coordinate input value establishes a one-to-one correspondence between the two-dimensional image coordinates and the reference coordinates. The subtraction result is distributed to each two-dimensional image coordinate according to this correspondence, which can restore the pseudo-removal response distribution of the entire image. The principal axis component of the slope aspect unfolded coordinate is associated with the image coordinates and the solar projection direction. The reference coordinate input value obtained after subtracting the center of the inversion ridge ensures that the subtraction result can accurately match the spatial position of the original image and avoid mapping misalignment.

[0039] It should be noted that this invention generates linearly decaying support weights that match the gradient of the inversion fringe influence. By weighted subtraction of the inversion fringe bending quantization value, an ideal profile with false turning points is obtained. This profile is then mapped back to two-dimensional image coordinates according to the relative position to obtain the true carbon sink sensitive response field. This ensures that the support weights accurately match the distribution of the inversion fringe influence, avoiding overcorrection or undercorrection. The subtraction operation effectively separates normal trends from false turning points, preserving the true carbon sink response. The mapping rule ensures accurate spatial matching without misalignment or distortion. The false turning point response field provides a reliable data foundation for subsequent carbon sink quantification calculations, improving the authenticity and accuracy of mountain forest carbon sink monitoring.

[0040] In one embodiment of the present invention, the annual integral of the pseudo-transition response field is processed by annual integration to generate an annual integral quantity, and the annual integral quantity is logarithmically compressed and continuously differencing is performed to output the annual change and transition intensity, including: Calculate the annual integral to satisfy the year labeling. Internal pseudo-transition response field Based on the observation date Using time as the coordinate, calculate the annual integral. : ; in For the first Each observation phase in two-dimensional image coordinates The false transition response field at the location, For year marking, Mark the year The covered set of observation phases, For the first The observation date corresponding to each observation time. The time interval between adjacent observation phases. Annual points; Calculate the annual carbon sequestration characterization quantity to meet the requirements of the annual integral quantity. Logarithmic compression is used to obtain the annual carbon sequestration characterization. : ; in For annual points, This is a characterization of annual carbon sequestration. Let be the natural logarithm function, a constant. For the offset term of logarithmic compression; Calculate the annual variation and the intensity of the transition to meet the requirements for characterizing the annual carbon sink. Perform continuous difference calculations to output the annual change. With the intensity of the turning point : ; ; in and This represents the annual carbon sequestration measure for adjacent years. This represents the annual change. This represents the annual change in the adjacent preceding year. The intensity of the transition.

[0041] It should be noted that the observation date is the specific calendar date of the remote sensing image acquisition, reflecting the temporal location information of the observation phase. This information can be obtained through the metadata of the remote sensing image, which clearly records the year, month, and day of the image acquisition. Commonly used satellites include Sentinel-2 and Landsat-8. The annual integral is a quantified value obtained by accumulating the pseudo-transition response fields of all observation phases within a single year, reflecting the cumulative level of carbon sequestration sensitive response throughout the year. The constant term is a fixed value set to avoid logarithmic aberrations, preferably set to 1. This value effectively avoids the problem of the natural logarithm being meaningless when the annual integral is 0, while not changing the relative magnitude of the annual integral, thus meeting the standardization requirements of carbon sequestration characterization. The annual carbon sequestration characterization is an indicator obtained by logarithmically compressing the sum of the annual integral and the constant term, reflecting the standardized quantification result of the annual carbon sequestration level. The annual change is the difference between the annual carbon sequestration characterization of the current year and the adjacent previous year, reflecting the annual increase or decrease trend of the carbon sequestration level. The transition strength is the difference between the annual change of the current year and the adjacent previous year, reflecting the degree of fluctuation in the annual change trend of carbon sequestration.

[0042] It should be noted that carbon sequestration is a continuous process of vegetation throughout the year. The spurious-free transition response fields at different observation phases reflect the carbon sequestration sensitivity of the corresponding periods. The arithmetic mean of the response fields between adjacent phases can represent the average carbon sequestration sensitivity level of that period. Multiplying this by the time interval (number of days) yields the cumulative response for that period. Summing up the cumulative responses for all periods throughout the year provides a complete characterization of the cumulative effect of carbon sequestration sensitivity response throughout the year. For example, in a year with 23 observation phases and an interval of 16 days between adjacent phases, if the average response for a certain period is 0.5 and the cumulative response for that period is 8, summing up the cumulative responses for all 23 periods throughout the year yields the annual integral, which truly reflects the cumulative characteristics of carbon sequestration throughout the year. The annual integral may have a large dynamic range, with integral values ​​in different regions differing by several times or even tens of times. Direct analysis is easily affected by extreme values. The natural logarithm function has the characteristic of compressing the dynamic range of data, transforming large-scale fluctuating values ​​into relatively concentrated standardized indicators while preserving the relative magnitudes of the original values. In addition, the annual change directly quantifies the annual increase or decrease in carbon sink level by the difference between the values ​​of adjacent years. A positive value indicates that carbon sink is strengthening, and a negative value indicates that carbon sink is weakening. The turning point intensity further quantifies the fluctuation of the carbon sink change trend by the difference between the changes of adjacent years. A positive value indicates that carbon sink growth is accelerating or decreasing is slowing down, and a negative value indicates that carbon sink growth is slowing down or decreasing is accelerating.

[0043] It should be noted that the specific steps for summing and accumulating the annual integral are as follows: First, filter all observation phases within the year by year label and sort them by observation date; second, calculate the difference in observation dates between two adjacent phases to obtain the time interval (in days); third, calculate the arithmetic mean of the pseudo-transition response fields of adjacent phases; fourth, multiply the average response value by the corresponding time interval to obtain the cumulative response for that period; fifth, sum the cumulative response for all periods throughout the year to obtain the annual integral, with the calculation process retaining four decimal places. Logarithmic compression is performed using the natural logarithm function, with the result retained to four decimal places and an error not exceeding 0.0001. If the logarithmic result is greater than 5, it is necessary to check for anomalies in the annual integral and confirm the accuracy of the pseudo-transition response field calculation. The reasonable range for the annual carbon sink characterization is 0 to 5, the reasonable range for the annual change is -1 to 1, and the reasonable range for the transition intensity is -0.5 to 0.5. If these ranges are exceeded, it is necessary to investigate whether there are anomalies in the calculation of the pseudo-transition response field or the annual integral. The allowable error for the time interval between adjacent observation phases is ±1 day. If the interval deviation exceeds 1 day due to limitations in image acquisition, interpolation processing for that period needs to be supplemented. Linear interpolation is used to obtain the false transition response of the missing period to ensure the accuracy of the cumulative calculation.

[0044] It should be noted that this invention performs annual integration on the pseudo-transition response field to accumulate sensitive carbon sink information for the entire year. By logarithmically compressing and standardizing the data range, and then performing continuous differencing, the annual increase / decrease trend and fluctuation degree of carbon sink are obtained. The annual integration operation aligns with the time-accumulation characteristics of carbon sinks, completely preserving annual carbon sink information; logarithmic compression effectively optimizes the dynamic range of the data, facilitating cross-regional and cross-year comparisons; continuous differencing is used to quantify dynamic changes in carbon sinks, forming hierarchical monitoring indicators. This provides an intuitive and reliable quantitative basis for the annual assessment and trend analysis of mountain forest carbon sinks, enhancing the practicality of the monitoring results.

[0045] In one embodiment of the present invention, a method for monitoring forest carbon sequestration based on remote sensing imagery includes the following steps: Step S1: Convert the two-dimensional image coordinates into slope unfolded coordinates based on the elevation and solar projection direction; Step S2: Calculate and extract the carbon sink sensitive profile response field using the red band reflectance and near-infrared band reflectance in the slope aspect unfolded coordinates; Step S3: Calculate the first zero-crossing position of the response field of the carbon sink sensitive profile based on the directional autocorrelation function as the structural scale; Step S4: Obtain a smooth response based on the response field of the carbon sink sensitive profile at the structural scale, and define the zero point of the second derivative closest to the steepest darkening point in the smooth response as the center of the curvature inversion ripple. Step S5: Construct a chord reference on the smooth response using the curvature inversion ripple center and structural scale, and calculate the deviation of the smooth response relative to the chord reference to obtain the inversion ripple warping function. Step S6: Establish a reference coordinate system with the center of the curvature inversion pattern as the zero point, and translate the smooth response and the inversion pattern warping function to the reference coordinate system to obtain the reference profile and the reference warping function. Step S7: Generate support weights based on the reference coordinates and structural scale, subtract the product of the support weights and the reference warping function from the reference profile, and map the subtraction result back to the two-dimensional image coordinates to obtain the pseudo-turn response field. Step S8: Perform annual integral processing on the pseudo-transition response field to generate annual integral quantity, and perform logarithmic compression and continuous difference calculation on the annual integral quantity to output the annual change and transition intensity.

[0046] It should be noted that the final output of this invention consists of three types of rasterized spatial distribution images of the annual carbon sink characterization, annual change, and transition intensity within the monitoring area. The image resolution is consistent with the input original remote sensing image, clearly presenting the spatial differentiation characteristics of carbon sink-related indicators. Simultaneously, it outputs statistical data tables for each year and sub-region, including the mean, extreme values, variance, and area proportion of different numerical intervals, comprehensively reflecting the overall level and local differences of carbon sinks in the monitoring area. All output results are in a standardized geospatial data format, directly compatible with mainstream geographic information processing software. Furthermore, the practical application value of the above output results is reflected in three aspects: First, it completes the baseline survey of regional forest carbon sink resources. Forestry and ecological environment management departments can grasp the basic carbon sink levels of different vegetation types and terrain areas based on the spatial distribution of annual carbon sink characterization, forming a systematic carbon sink resource ledger. Second, it enables routine monitoring of the dynamic trends of forest carbon sinks. By identifying areas of annual increase or decrease in carbon sinks through annual change, and combining this with the transition intensity analysis of carbon sink change trends, areas with abnormal changes in carbon sink levels can be identified and their causes investigated in a timely manner. Third, it provides data support for ecological protection and forestry management decisions, and formulates measures such as forest tending and vegetation restoration for areas with low carbon sequestration levels or unfavorable trends. At the same time, the quantified carbon sequestration data can also verify the effectiveness of ecological protection measures, which will not be elaborated here.

[0047] It should be noted that the interval and threshold sizes are set for ease of comparison. The size of the threshold depends on the amount of sample data and the base number set by those skilled in the art for each set of sample data, as long as it does not affect the proportional relationship between the parameter and the quantized value. Furthermore, the above formulas are all dimensionless calculations, and the formulas are derived from software simulations using a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0048] The embodiments of this example have been described above. However, this example is not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms based on the guidance of this example, and all of them are within the protection scope of this example.

Claims

1. A forest carbon sequestration monitoring system based on remote sensing imagery, characterized in that, include: The first module converts the two-dimensional image coordinates into slope-oriented unfolded coordinates based on the elevation and the direction of solar projection. The second module uses the red light band reflectance and near-infrared band reflectance to calculate and extract the carbon sink sensitive profile response field from the slope aspect unfolded coordinates. The third module uses the directional autocorrelation function to determine the first zero-crossing position of the response field of the carbon sink sensitive profile as the structural scale. The fourth module obtains a smooth response based on the response field of the carbon sink sensitive profile at the structural scale, and defines the zero point of the second derivative closest to the steepest darkening point in the smooth response as the center of the curvature inversion ripple. The fifth module constructs a chord reference based on the curvature inversion ripple center and structural scale on the smooth response, and calculates the deviation of the smooth response relative to the chord reference to obtain the inversion ripple warping function; The sixth module establishes a reference coordinate system with the center of the curvature inversion pattern as the zero point, and translates the smooth response and the inversion pattern warping function to the reference coordinate system to obtain the reference profile and the reference warping function. The seventh module generates support weights based on the reference coordinates and structural scale, subtracts the product of the support weights and the reference warping function from the reference profile, and maps the subtraction result back to the two-dimensional image coordinates to obtain the pseudo-turn response field. The eighth module performs annual integral processing on the pseudo-transition response field to generate annual integral values, and performs logarithmic compression and continuous difference calculation on the annual integral values ​​to output the annual change and transition intensity.

2. The forest carbon sequestration monitoring system based on remote sensing imagery according to claim 1, characterized in that, Elevation values ​​are obtained at the coordinate positions of the two-dimensional image. The slope gradient components are obtained by calculating the rate of change of the elevation values ​​with respect to the horizontal axis of the two-dimensional image coordinates and the rate of change of the elevation values ​​with respect to the vertical axis of the two-dimensional image coordinates. During the observation phase, the solar zenith angle and solar azimuth angle are obtained, and the unit direction vector of the solar projection onto the two-dimensional image coordinate plane is constructed as the unit vector of the solar projection direction. By using the unit vector of the solar projection direction to rotate and map the coordinates of the two-dimensional image, the principal axis components of the slope unfolded coordinates that are consistent with the solar projection direction and the orthogonal components of the slope unfolded coordinates that are perpendicular to the solar projection direction are generated.

3. A forest carbon sequestration monitoring system based on remote sensing imagery according to claim 1, characterized in that, The reflectance in the red band and near-infrared band of the two-dimensional image coordinates is obtained during the observation phase. The carbon sink sensitive response corresponding to the two-dimensional image coordinates is obtained by calculating the difference between the near-infrared band reflectance and the red band reflectance, the near-infrared band reflectance, the red band reflectance, and the preset weighting coefficient. A fixed profile location marker is selected within the slope aspect unfolded coordinate system. The two-dimensional image coordinates corresponding to the profile location marker and the continuous position variables within the profile are determined. The carbon sink sensitive response quantity is mapped to the slope aspect unfolded coordinate space to form a carbon sink sensitive profile response field that is distributed with the continuous position variables within the profile.

4. A forest carbon sequestration monitoring system based on remote sensing imagery according to claim 1, characterized in that, The mean value of the carbon sink sensitive profile is obtained by integrating the response field of the profile within the profile domain and dividing it by the length of the profile domain. Using the displacement along the profile direction as a variable, calculate the integral of the product of the original position and the displacement position of the carbon sink sensitive profile response field minus the profile mean, and divide it by the integral of the square of the carbon sink sensitive profile response field minus the profile mean to obtain the directional autocorrelation function. Within the range where the positional displacement is greater than zero, find the minimum positional displacement when the directional autocorrelation function is zero, and determine the minimum positional displacement as the structural dimension.

5. A forest carbon sequestration monitoring system based on remote sensing imagery according to claim 1, characterized in that, Using the structural scale as the half-width of the smoothing window, a local average integral operation is performed on the response field of the carbon sink sensitive section in the cross-sectional dimension to obtain the smooth response. Calculate the first derivative of the smooth response with respect to the positional variables within the profile, locate the position with the minimum value in the first derivative distribution, and determine it as the steepest darkening point; Calculate the second derivative of the smooth response with respect to the position variable within the profile, find the position with the smallest distance from the steepest darkening point in the set of positions where the second derivative is zero, and determine the selected position as the center of the curvature inversion pattern.

6. A forest carbon sequestration monitoring system based on remote sensing imagery according to claim 1, characterized in that, Using the center of curvature inversion texture as the interval center point and the structural scale as the interval half-width, the range of intervals in which smooth response participates in the calculation is defined. Obtain the values ​​of the smooth response at the left and right endpoints of the defined interval, and generate a chord reference by linearly connecting the values ​​of the left and right endpoints; Within the defined interval, calculate the deviation value obtained by subtracting the chord reference from the smooth response, and determine the reverse ripple warping function.

7. A forest carbon sequestration monitoring system based on remote sensing imagery according to claim 1, characterized in that, Using the center of the curvature inversion pattern as the origin of the coordinate system, the spatial offset of the positional variables within the profile relative to the center of the curvature inversion pattern is calculated, and a reference coordinate system is established. By replacing the in-section position variable in the smooth response with the summation of the curvature inversion pattern center and the reference coordinates, the smooth response is translated to the reference coordinate system to obtain the reference profile. By replacing the in-section position variable in the warping function of the inverted curve with the summation term of the center of the curvature inversion pattern and the reference coordinate, the warping function of the inverted curve is translated into the reference coordinate system to obtain the reference warping function.

8. A forest carbon sequestration monitoring system based on remote sensing imagery according to claim 1, characterized in that, Using reference coordinates and structural scales for calculation, within the range where the absolute value of the reference coordinates is less than the structural scale, a weight value that decreases linearly as the absolute value of the reference coordinates increases is constructed to determine the support weights; Calculate the product term of the support weight and the benchmark warping function, and subtract the product term from the benchmark profile to obtain the subtraction result; Read the principal axis components of the slope aspect unfolded coordinates corresponding to the two-dimensional image coordinates, subtract the curvature inversion fringe center from the principal axis components of the slope aspect unfolded coordinates to obtain the reference coordinate input value, map the subtraction result to the two-dimensional image coordinate position according to the reference coordinate input value, and output the pseudo-turn response field.

9. A forest carbon sequestration monitoring system based on remote sensing imagery according to claim 1, characterized in that, Within the range defined by the year marker, obtain the set of pseudo-transition response fields and the corresponding observation dates for each observation time. Then, use the arithmetic mean of the pseudo-transition response fields of adjacent observation time phases and the time interval between adjacent observation time phases to perform summation and accumulation operations to generate the annual integral. The annual carbon sequestration characterization is obtained by performing logarithmic compression on the sum of the annual integral and the constant term using the natural logarithm function. The annual change is obtained by subtracting the annual carbon sink of the previous year from the current year's annual carbon sink value; the turning point intensity is obtained by subtracting the annual change of the previous year from the current year's annual change.

10. A method for monitoring forest carbon sequestration based on remote sensing imagery, characterized in that, Implementing a forest carbon sequestration monitoring system based on remote sensing imagery as described in any one of claims 1 to 9 includes the following steps: Step S1: Convert the two-dimensional image coordinates into slope unfolded coordinates based on the elevation and solar projection direction; Step S2: Calculate and extract the carbon sink sensitive profile response field using the red band reflectance and near-infrared band reflectance in the slope aspect unfolded coordinates; Step S3: Calculate the first zero-crossing position of the response field of the carbon sink sensitive profile based on the directional autocorrelation function as the structural scale; Step S4: Obtain a smooth response based on the response field of the carbon sink sensitive profile at the structural scale, and define the zero point of the second derivative closest to the steepest darkening point in the smooth response as the center of the curvature inversion ripple. Step S5: Construct a chord reference on the smooth response using the curvature inversion ripple center and structural scale, and calculate the deviation of the smooth response relative to the chord reference to obtain the inversion ripple warping function. Step S6: Establish a reference coordinate system with the center of the curvature inversion pattern as the zero point, and translate the smooth response and the inversion pattern warping function to the reference coordinate system to obtain the reference profile and the reference warping function. Step S7: Generate support weights based on the reference coordinates and structural scale, subtract the product of the support weights and the reference warping function from the reference profile, and map the subtraction result back to the two-dimensional image coordinates to obtain the pseudo-turn response field. Step S8: Perform annual integral processing on the pseudo-transition response field to generate annual integral quantity, and perform logarithmic compression and continuous difference calculation on the annual integral quantity to output the annual change and transition intensity.

Images