A method and system for assessing the vegetation cover status of slopes using aerial imagery

Abstract

This invention relates to a method and system for assessing the vegetation cover status of slopes using aerial imagery, belonging to the field of image processing technology. The method includes: acquiring slope images using a drone; calculating a green index and classifying pixels; identifying connected vegetation regions; extracting suspected shadow areas through recursive segmentation and brightness comparison; calculating the structural complexity of vegetation; assessing vegetation health based on relative values ​​of RGB channels; analyzing vegetation distribution uniformity by dividing the data into grids; and finally fusing four indicators—coverage, structural complexity, health, and distribution uniformity—to generate quantitative parameters for slope vegetation cover. These parameters are then compared with threshold values ​​to determine whether the vegetation cover status is good. This invention makes slope vegetation cover assessment more accurate.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, specifically to a method and system for assessing the vegetation cover status of slopes in aerial photographs. Background Technology

[0002] Slope vegetation coverage is crucial, not only for aesthetic purposes but also as a core measure to ensure project safety and restore the ecosystem. Dense vegetation strengthens the soil through its root system and reduces soil moisture through transpiration, effectively preventing slope erosion and shallow landslides, and significantly improving slope stability. Simultaneously, it can restore the ecological environment, purify the air, reduce rainwater erosion, and significantly reduce long-term maintenance costs, making it a green and sustainable approach that integrates safety, ecological, and economic benefits.

[0003] Traditional methods for determining vegetation cover based on aerial images have the core advantage of being able to quickly, macroscopically, and cost-effectively assess the "area" of vegetation cover over large areas, efficiently grasping the overall scale and distribution of greening. However, these methods only consider a two-dimensional perspective, neglecting three-dimensional "quality" information such as the species type and health status of the vegetation, and therefore cannot accurately assess the true stability risk of slopes.

[0004] This method systematically and comprehensively analyzes multi-dimensional indicators such as vegetation hierarchy, spatial distribution, and growth status, thereby revealing the intrinsic relationship between vegetation and slope stability more profoundly and significantly improving the accuracy and authenticity of the assessment results. Summary of the Invention

[0005] This invention provides a method and system for assessing the vegetation cover status of slopes using aerial imagery, in order to solve the existing problem that existing methods are inaccurate in determining the vegetation cover status of slopes.

[0006] The aerial imagery-based method and system for assessing slope vegetation cover status of the present invention adopts the following technical solution:

[0007] In a first aspect, the present invention provides a method for assessing the vegetation cover status of slopes using aerial imagery, the method comprising the following steps:

[0008] Step S01: Use drone aerial photography to acquire multiple images of slope areas in different regions, and traverse the slope images according to the order in which they were captured;

[0009] Step S02: Calculate the green index based on the RGB three-channel values ​​of each pixel in the slope area image, filter all pixels with a green index greater than 0, divide the pixels with a green index greater than zero into two categories of pixels with the greatest difference in green index, record the pixel with the larger green index as the first category of pixels, and record the pixel with the smaller green index as the second category of pixels, and obtain the ratio of the number of first category pixels in each image to the total number of pixels in the image;

[0010] Step S03: Filter out all connected components in the first type of pixels, and denote the connected components in the first type of pixels with a number of pixels greater than N as green plant connected components. Compare the brightness of each green plant connected component with the brightness of the global green plant connected components, and divide the green plant connected components into multiple partitions based on the comparison results. Mark the suspected shadow areas of the green plant connected components. Calculate the green plant structure complexity of the corresponding shooting area for each image based on the number and brightness of the suspected shadow area pixels in each green plant connected component and the number and brightness of all pixels in each green plant connected component.

[0011] Step S04: Obtain the health status of the plants in each image based on the relative values ​​of the G channel, R channel, and B channel of the pixels in the connected component of the plants in each image.

[0012] Step S05: Divide each image into grids of the same size, calculate the ratio of the number of pixels in the connected component of green plants in each grid to the number of pixels in the grid, and obtain the uniformity of green plant distribution in each image based on the standard deviation and mean of the ratio of the number of pixels in the connected component of green plants in each grid to the number of pixels in the grid.

[0013] Step S06: Based on the ratio of the number of first-class pixels to the total number of pixels in each image, the structural complexity of the vegetation in the corresponding shooting area of ​​each image, the health of the vegetation in each image, and the uniformity of vegetation distribution in each image, obtain the slope vegetation coverage quantification parameters, and set a threshold K to compare with the slope vegetation coverage quantification parameters to determine whether the slope vegetation coverage is good.

[0014] Furthermore, multiple images of slope areas in different regions were acquired using drone aerial photography, and the slope images were traversed according to the order in which they were captured. Specific methods included:

[0015] During the period of highest solar altitude, images of the slope area were captured using a drone. The drone maintained a distance of Y meters from the slope surface, and each image was Z*Z pixels in size. A total of Q images were captured, and all images were iterated through according to the time sequence of the images captured by the drone.

[0016] Furthermore, the specific method for calculating the green index based on the RGB three-channel values ​​of each pixel in the slope area image, filtering all pixels with a green index greater than 0, dividing the pixels with a green index greater than zero into two categories with the largest difference in green index, designating the category with the larger green index as the first category and the category with the smaller green index as the second category, includes:

[0017]

[0018] In the formula, This represents the green index of the pixel in the i-th row and j-th column of the q-th image; This represents the G channel value of the pixel in the i-th row and j-th column of the q-th image; This represents the R channel value of the pixel in the i-th row and j-th column of the q-th image; This represents the B channel value of the pixel in the i-th row and j-th column of the q-th image;

[0019] Calculate the ratio of the number of pixels of type I to the total number of pixels in each image. The ratio of the number of pixels of type I to the total number of pixels in the q-th image is denoted as... ;

[0020] Record all pixels in each image with a green index greater than 0, and divide all pixels in each image into two categories. The classification criterion is that the absolute value of the difference between the standard deviations of the green indices of the two categories of pixels in each image after classification, and the ratio of the sum of the standard deviations of the green indices of the two categories of pixels in each image, reach a maximum value. The category of pixels with a larger mean green index after classification is recorded as the first category of pixels in each image, and the category of pixels with a smaller mean green index is recorded as the second category of pixels in each image.

[0021] Further, the process involves filtering out all connected components in the first category of pixels, and designating connected components in the first category with a number of pixels greater than N as "green plant connected components." The brightness of each green plant connected component is compared with the brightness of the global green plant connected components. Based on the comparison results, the green plant connected components are partitioned multiple times, and suspected shadow areas are marked. The complexity of the green plant structure in the corresponding shooting area of ​​each image is calculated based on the number and brightness of pixels in the suspected shadow areas within each green plant connected component and the number and brightness of all pixels within each green plant connected component. The specific methods include:

[0022] For each image, denote the connected components with more than N pixels among all connected components formed by the first type of pixels as each green plant connected component. Extract the V channel value of each pixel in the HSV color space for each green plant connected component. Then, for each green plant connected component in each image, perform the following recursive segmentation and shadow discrimination process sequentially. For the d-th green plant connected component in the q-th image, calculate the average value of the V channel value of each pixel in that connected component. Standard deviation of V channel values The recursive segmentation and shadow discrimination process is as follows: Locate the two points with the greatest Euclidean distance on the contour boundary, connect these two points to form a principal axis segment, then construct the perpendicular bisector of this principal axis segment. The principal axis segment and the perpendicular bisector together divide the original green plant connected domain into several secondary sub-regions. Calculate the standard deviation and mean of the V channel values ​​of the green plant connected domain in each secondary sub-region. If the standard deviation of the V channel of a pixel in a secondary sub-region is greater than... If the average V channel value of all pixels in the connected component of the greenery in a certain sub-region is lower than the threshold value, then the recursive segmentation and shadow discrimination process will continue to be performed on the next sub-region. Then, each connected component of green plants in the next lower-level sub-region is marked as a suspected shadow region; this continues until the standard deviation of the V channel of all pixels in the connected components of green plants in each sub-region is less than or equal to... Furthermore, for each sub-region, the suspected shadow area is marked, the recursive segmentation and shadow discrimination process is stopped, and all suspected shadow areas of the q-th image are traversed;

[0023] The method for calculating the structural complexity of vegetation in the area corresponding to the q-th image is as follows:

[0024]

[0025] In the formula, This represents the structural complexity of the vegetation in the area captured by the q-th image. Indicates that the q-th image has a total of A suspected shadow area, The q-th image represents the... The number of pixels in the suspected shadow area The q-th image represents the... The total number of pixels in the unsegmented connected components of greenery at the locations corresponding to the suspected shadow areas. The q-th image represents the... The average V channel value of each pixel in the suspected shadow area Represents the q-th image. The average value of the V channel of the pixels in the unsegmented connected domain of greenery at the location corresponding to the suspected shadow area.

[0026] Furthermore, the specific method for obtaining the health status of plants in each image based on the relative magnitudes of the G-channel, R-channel, and B-channel values ​​of pixels in the connected component of the greenery in each image includes:

[0027] ;

[0028] ;

[0029] In the formula, This indicates the health status of the plants in the q-th image; This represents the health status of the plants in the m-th connected component of the q-th image; Indicates that the q-th image has a total of A green plant connecting area; This represents the value of the G channel at the r-th pixel in the m-th connected component of the q-th plant in the image. This represents the value of the R channel of the r-th pixel in the m-th connected component of the q-th image; This represents the value of the B channel of the r-th pixel in the m-th connected component of the green plant in the q-th image; This indicates that the total number of connected components for the m-th green plant in the q-th image is... Each pixel.

[0030] Furthermore, the specific method for dividing each image into a grid of equal size, calculating the ratio of the number of pixels in the connected component of greenery in each grid to the total number of pixels in the grid, and obtaining the uniformity of greenery distribution in each image based on the standard deviation and mean of the ratio of the number of pixels in the connected component of greenery in each grid to the total number of pixels in the grid, includes:

[0031] Set a grid of size H*H, with the number of grid cells in each image being [number missing]. Z represents the side length of the captured image. The ratio of the number of pixels in the connected component of the greenery in each grid to the total number of pixels in the grid is calculated. The uniformity of the greenery distribution in each image is obtained based on the standard deviation and mean of this ratio. The specific method is as follows:

[0032] ;

[0033] ;

[0034] Formula parameters: This represents the ratio of the number of pixels in the connected component of the greenery in the s-th grid of the q-th image to the total number of pixels in the grid; the count() function extracts the number of pixels from its pixel set; This represents the set of pixels in the s-th grid of the q-th image that are covered by the green plant mask; This represents the set of all pixels in the s-th grid of the q-th image; This indicates the uniformity of the vegetation distribution in the q-th image; This represents all images in the q-th image. The average of the ratio of the number of pixels in the connected component of greenery in each grid to the total number of pixels in the grid; Represents all in the q-th image The standard deviation of the ratio of the number of pixels in the connected component of the greenery in a grid to the total number of pixels in the grid.

[0035] Furthermore, based on the ratio of the number of first-class pixels to the total number of pixels in each image, the structural complexity of the vegetation in the corresponding shooting area of ​​each image, the health status of the vegetation in each image, and the uniformity of vegetation distribution in each image, slope vegetation cover quantification parameters are obtained. A threshold K is then set and compared with the slope vegetation cover quantification parameters to determine whether the slope vegetation cover is good. The specific method is as follows:

[0036]

[0037] In the formula, This represents the quantification parameter for slope vegetation cover, where Q indicates the total number of Q images taken for the current slope. This represents the ratio of the number of pixels of type 1 in the t-th image to the total number of pixels in the image. This represents the structural complexity of the vegetation in the area corresponding to the t-th image. This represents the health status of the green plants in the t-th image. This indicates the uniformity of the vegetation distribution in the t-th image;

[0038] Set appropriate thresholds based on the regional characteristics of different areas. If the current slope vegetation cover quantification parameter I is greater than or equal to K, the current slope vegetation cover is considered to be good; if the current slope vegetation cover quantification parameter I is less than K, the vegetation cover is considered to be poor.

[0039] A second aspect of the present invention provides an aerial imagery slope vegetation cover status assessment system, comprising a UAV image acquisition module, a pixel classification module, a vegetation structure complexity calculation module, a vegetation health status calculation module, a vegetation distribution uniformity calculation module, and a slope vegetation cover status judgment module, wherein:

[0040] The UAV image acquisition module is used to acquire multiple images of slope areas in different regions using UAV aerial photography, and to traverse the slope images in the order they were captured.

[0041] The pixel classification module is used to calculate the green index based on the RGB three-channel values ​​of each pixel in the slope area image, filter all pixels with a green index greater than 0, divide the pixels with a green index greater than 0 into two categories of pixels with the greatest difference in green index, record the pixel with the larger green index as the first category of pixels, record the pixel with the smaller green index as the second category of pixels, and obtain the ratio of the number of first category pixels in each image to the total number of pixels in the image;

[0042] The green plant structure complexity calculation module is used to filter out all connected components in the first type of pixels, and to mark the connected components in the first type of pixels with a number of pixels greater than N as green plant connected components. The brightness of each green plant connected component is compared with the brightness of the global green plant connected components. Based on the comparison results, the green plant connected components are divided into multiple partitions, and the suspected shadow areas of the green plant connected components are marked. The green plant structure complexity of the corresponding shooting area of ​​each image is calculated based on the number and brightness of the suspected shadow area pixels in each green plant connected component and the number and brightness of all pixels in each green plant connected component.

[0043] The plant health calculation module is used to obtain the plant health level in each image based on the relative magnitudes of the G channel values, R channel values, and B channel values ​​of the pixels in the connected domain of the plant.

[0044] The green plant distribution uniformity calculation module is used to divide each image into grids of the same size, calculate the ratio of the number of pixels in the connected component of green plants in each grid to the number of pixels in the grid, and obtain the green plant distribution uniformity of each image based on the standard deviation and mean of the ratio of the number of pixels in the connected component of green plants in each grid to the number of pixels in the grid.

[0045] The slope vegetation cover determination module is used to obtain slope vegetation cover quantification parameters based on the ratio of the number of first-class pixels to the total number of pixels in each image, the complexity of the vegetation structure in the corresponding shooting area of ​​each image, the health of the vegetation in each image, and the uniformity of vegetation distribution in each image. It then sets a threshold K and compares it with the slope vegetation cover quantification parameters to determine whether the slope vegetation cover is good.

[0046] In a third aspect, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described method for assessing the vegetation cover status of a slope based on aerial imagery.

[0047] In a fourth aspect, a computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements the steps of the above-described method for assessing the vegetation cover status of a slope based on aerial imagery.

[0048] The beneficial effects of the technical solution of the present invention are:

[0049] This paper provides a comprehensive method for assessing the vegetation cover status of slopes using aerial imagery. It comprehensively judges the vegetation status through multiple indicators: coverage rate, structural complexity, health level, and distribution uniformity, which significantly improves the accuracy of assessment and engineering applicability.

[0050] By selecting the period of maximum sun altitude for aerial photography, the interference of terrain shadows is effectively reduced, ensuring the consistency of image illumination and providing a high-quality data foundation for subsequent vegetation identification and index calculation.

[0051] An adaptive classification method is used to perform secondary segmentation on pixels with a green index greater than 0, which effectively distinguishes between dense vegetation and sparse / misjudged areas, improving the accuracy and robustness of vegetation recognition.

[0052] By recursively segmenting and comparing brightness to identify shadow areas, the three-dimensional structure and hierarchical complexity of vegetation can be indirectly reflected, enhancing the assessment dimensions of vegetation ecological functions, such as soil and water conservation capacity.

[0053] A health index is constructed based on the phenomenon of red and blue light inhibition, which can sensitively capture changes in the physiological state of vegetation, avoid healthy vegetation from covering degraded areas, and improve the comprehensiveness and accuracy of health assessment.

[0054] By analyzing the uniformity of vegetation distribution through gridding, areas with coverage defects can be identified, which makes up for the inadequacy of relying solely on the overall coverage rate and enhances the spatial dimension assessment of slope ecological stability.

[0055] A comprehensive quantitative parameter I is constructed, which integrates coverage, structural complexity, health status and uniformity. The "barrel effect" is reflected through a product model, which scientifically reflects the overall effectiveness of vegetation cover and supports the judgment of regional adaptability threshold. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1This is a flowchart illustrating the steps of an aerial image slope vegetation cover status assessment method according to the present invention.

[0058] Figure 2 This is a structural block diagram of an aerial image slope vegetation cover status assessment system according to the present invention. Detailed Implementation

[0059] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an aerial image slope vegetation cover status assessment method and system proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0060] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0061] The following description, in conjunction with the accompanying drawings, details the specific scheme of the aerial image slope vegetation cover status assessment method and system provided by the present invention.

[0062] Please see Figure 1 It illustrates the first objective of the present invention, a flowchart of a method for assessing the vegetation cover status of slopes using aerial imagery, the method comprising the following steps:

[0063] Step S01: Use drones to take aerial photos to acquire multiple images of slope areas in different regions, and then traverse the slope images in the order they were taken.

[0064] To assess the vegetation cover of slopes, images are essential for judgment. Drones can quickly and safely acquire high-resolution images of slopes, providing an extremely accurate and comprehensive data foundation for vegetation cover assessment. Compared to traditional manual surveys, drones not only significantly improve efficiency and reduce safety risks associated with steep terrain, but also transcend the limitations of human vision through multispectral analysis and other methods, quantitatively monitoring vegetation health and cover density.

[0065] Specifically, multiple images of slope areas in different regions are obtained using drone aerial photography, and the slope images are traversed according to the order in which they were captured. The specific method is as follows:

[0066] During the period of highest solar altitude, images of the slope area were captured using a drone. The drone maintained a distance of Y meters from the slope surface, and each image was Z*Z pixels in size. A total of Q images were captured, and all images were iterated through according to the time sequence of the images captured by the drone.

[0067] It should be noted that the time of day with the highest solar altitude angle varies in different regions. This embodiment does not limit the image acquisition time, image size, number of images, or the distance Y between the drone and the slope surface. In this embodiment, the image acquisition starts at noon, the image size is 1000*1000 pixels, 100 images are captured, and the distance between the drone and the slope surface is 30 meters. The aerial photography is conducted during the period with the highest solar altitude angle because the sunlight shines almost perpendicularly to the ground during this period, which can minimize the shadows cast by the slope surface due to terrain undulations. This ensures that the vegetation color characteristics in the image are minimally affected by the angle of illumination, improves the accuracy of vegetation identification and green index calculation, and provides a data foundation with consistent illumination and low distortion for subsequent quantitative assessment of vegetation cover status.

[0068] Step S02: Calculate the green index based on the RGB three-channel values ​​of each pixel in the slope area image, filter all pixels with a green index greater than 0, divide the pixels with a green index greater than zero into two categories of pixels with the largest difference in green index, record the pixel with the larger green index as the first category of pixels, and record the pixel with the smaller green index as the second category of pixels.

[0069] It should be noted that when assessing slope vegetation cover, green vegetation typically constitutes the main body of the cover. To accurately determine the spatial distribution and coverage of vegetation, it is necessary to first effectively identify and delineate green areas using image analysis methods, thereby providing a reliable basis for further quantitative assessment of the cover density, composition type, and ecological benefits of these areas.

[0070] Specifically, the green index is calculated based on the RGB three-channel values ​​of each pixel in the slope area image. All pixels with a green index greater than 0 are filtered out. These pixels are then divided into two categories with the largest differences in green index. The category with the larger green index is designated as the first category, and the category with the smaller green index is designated as the second category. The methods include the following:

[0071]

[0072] In the formula, This represents the green index of the pixel in the i-th row and j-th column of the q-th image; This represents the G channel value of the pixel in the i-th row and j-th column of the q-th image; This represents the R channel value of the pixel in the i-th row and j-th column of the q-th image; This represents the B channel value of the pixel in the i-th row and j-th column of the q-th image;

[0073] Calculate the ratio of the number of pixels of type I to the total number of pixels in each image. The ratio of the number of pixels of type I to the total number of pixels in the q-th image is denoted as... .

[0074] Record all pixels in each image with a green index greater than 0, and divide all pixels in each image into two categories. The classification criterion is that the absolute value of the difference between the standard deviations of the green indices of the two categories of pixels in each image after classification, and the ratio of the sum of the standard deviations of the green indices of the two categories of pixels in each image, reach a maximum value. The category of pixels with a larger mean green index after classification is recorded as the first category of pixels in each image, and the category of pixels with a smaller mean green index is recorded as the second category of pixels in each image.

[0075] It's important to note that selecting pixels with a green index greater than 0 for subsequent adaptive classification is based on the core logic of constructing a clean dataset centered on "suspected vegetation" pixels. This lays a reliable foundation for accurately distinguishing vegetation states later. This step is crucial because it's based on the fundamental physical characteristics of vegetation spectral reflectance: healthy green vegetation reflects more green light and absorbs more red and blue light, ensuring its green index value is greater than 0. This initial screening efficiently eliminates a large number of obviously non-vegetation background interference pixels such as soil, rocks, shadows, and man-made structures, significantly improving the signal-to-noise ratio. Subsequent adaptive classification based on pixels with a green index greater than 0 further extracts vegetation areas. This is because pixels with a green index greater than zero but a small value may contain overly sparse vegetation or soil misclassified as vegetation, while the green index of vegetation is significantly higher than that of overly sparse vegetation or misclassified soil.

[0076] Step S03: Filter out all connected components in the first type of pixels, and denote the connected components in the first type of pixels with a number of pixels greater than N as green plant connected components. Compare the brightness of each green plant connected component with the brightness of the global green plant connected components, and divide the green plant connected components into multiple partitions based on the comparison results. Mark the suspected shadow areas in the green plant connected components. Calculate the green plant structure complexity of the corresponding shooting area for each image based on the number and brightness of the suspected shadow area pixels in each green plant connected component and the number and brightness of all pixels in each green plant connected component.

[0077] It should be noted that when assessing slope vegetation cover, relying solely on the area of ​​green vegetation is insufficient to fully reflect its ecological function and structural stability. To achieve a scientifically accurate evaluation, multi-dimensional indicators such as vegetation hierarchy, health status, and spatial distribution characteristics need to be introduced. This step aims to integrate these factors to construct a holistic vegetation cover assessment model, thereby more systematically characterizing the actual coverage and ecological benefits of green vegetation.

[0078] It is important to further explain that when assessing slope vegetation cover, especially when considering not only cover but also soil and water conservation capacity, the layered structure of vegetation is crucial. A single herbaceous plant cover is far less effective at conserving soil and water than a layered vegetation area combining trees, shrubs, and herbs. Furthermore, the more complex the layered structure, the stronger its resistance to interference. Therefore, this step calculates the proportion of complex structures to provide a more scientific basis for assessing cover. The complexity of vegetation structure is usually manifested in obvious vertical stratification, such as a multi-layered canopy formed by tall trees, lower shrubs, and surface herbs. This layered structure produces a significant shading effect, causing the lower vegetation and ground surface to be shaded by the upper canopy, creating shadows. During the image acquisition period selected for this study, the solar altitude angle was relatively high, resulting in more concentrated shadow distribution, which often completely covered the corresponding vegetation area below. Therefore, under these imaging conditions, the proportion of shadowed areas in the image can reflect the spatial distribution ratio of complex vegetation structures to a considerable extent.

[0079] Specifically, all connected components in the first category of pixels are selected, and those connected components with more than N pixels are denoted as "green plant connected components." The brightness of each green plant connected component is compared with the brightness of the global green plant connected components. Based on the comparison results, the green plant connected components are divided into multiple partitions, and suspected shadow areas are marked. The structural complexity of the green plants in the corresponding shooting area of ​​each image is calculated based on the number and brightness of suspected shadow area pixels in each green plant connected component and the number and brightness of all pixels in each green plant connected component. The specific methods include:

[0080] For each image, denote the connected components with more than N pixels among all connected components formed by the first type of pixels as each green plant connected component. Extract the V channel value of each pixel in the HSV color space for each green plant connected component. Then, for each green plant connected component in each image, perform the following recursive segmentation and shadow discrimination process sequentially. For the d-th green plant connected component in the q-th image, calculate the average value of the V channel value of each pixel in that connected component. Standard deviation of V channel values The recursive segmentation and shadow discrimination process is as follows: Locate the two points with the greatest Euclidean distance on the contour boundary, connect these two points to form a principal axis segment, then construct the perpendicular bisector of this principal axis segment. The principal axis segment and the perpendicular bisector together divide the original green plant connected domain into several secondary sub-regions. Calculate the standard deviation and mean of the V channel values ​​of the green plant connected domain in each secondary sub-region. If the standard deviation of the V channel of a pixel in a secondary sub-region is greater than... If the average V channel value of all pixels in the connected component of the greenery in a certain sub-region is lower than the threshold value, then the recursive segmentation and shadow discrimination process will continue to be performed on the next sub-region. Then, each connected component of green plants in the next lower-level sub-region is marked as a suspected shadow region; this continues until the standard deviation of the V channel of all pixels in the connected components of green plants in each sub-region is less than or equal to... Furthermore, for each sub-region, the suspected shadow area is marked, the recursive segmentation and shadow discrimination process is stopped, and all suspected shadow areas of the q-th image are traversed;

[0081] The method for calculating the structural complexity of vegetation in the area corresponding to the q-th image is as follows:

[0082]

[0083] In the formula, This represents the structural complexity of the vegetation in the area captured by the q-th image. Indicates that the q-th image has a total of A suspected shadow area, The q-th image represents the... The number of pixels in the suspected shadow area The q-th image represents the... The total number of pixels in the unsegmented connected components of greenery at the locations corresponding to the suspected shadow areas. The q-th image represents the... The average V channel value of each pixel in the suspected shadow area Represents the q-th image. The average value of the V channel of the pixels in the unsegmented connected domain of greenery at the location corresponding to the suspected shadow area.

[0084] It's important to note that the method of hierarchically segmenting and comparing sub-regions with global brightness is based on the fact that under natural light, the denser and more varied the vegetation canopy, the more numerous and complex the shadows it produces; conversely, sparse or low-lying vegetation casts fewer and simpler shadows. Recursive segmentation allows for precise separation of these structure-induced shadow regions, rather than simply using a fixed threshold, thus avoiding misinterpreting dark soil or the dark color of the vegetation itself as shadows. The advantage of this approach is that it transcends the two-dimensional image level, indirectly reflecting the three-dimensional structure and growth density of the vegetation, transforming three-dimensional morphological information that is difficult to measure directly into calculable parameters.

[0085] Step S04: Obtain the health status of the plants in each image based on the relative values ​​of the G channel, R channel, and B channel of the pixels in the connected component of the plants in each image.

[0086] It should be noted that when judging vegetation cover, not only the hierarchical structure of the vegetation should be considered, but the health of the vegetation also greatly affects the accuracy of the vegetation cover rate. Furthermore, unhealthy or malnourished plants also mean weaker soil and water conservation capabilities. Based on this, this step will calculate the health of the vegetation to provide a more scientific basis for judging the cover. The health of vegetation is often directly reflected in the spectral capacity of the leaves: healthy plant leaves can usually absorb red and blue light well, while plants with health problems often show abnormal phenotypes of reduced absorption of red and blue light. Based on this phenomenon, the average value of the red and blue inhibition index of pixels within the green vegetation range is normalized and calculated as the health of the plants.

[0087] Specifically, the health level of the plants in each image is obtained based on the relative values ​​of the G, R, and B channels of the pixels in the connected component of the plants. The specific methods include:

[0088] ;

[0089] ;

[0090] In the formula, This indicates the health status of the plants in the q-th image; This represents the health status of the plants in the m-th connected component of the q-th image; Indicates that the q-th image has a total of A green plant connecting area; This represents the value of the G channel at the r-th pixel in the m-th connected component of the q-th plant in the image. This represents the value of the R channel of the r-th pixel in the m-th connected component of the q-th image; This represents the value of the B channel of the r-th pixel in the m-th connected component of the green plant in the q-th image; This indicates that the total number of connected components for the m-th green plant in the q-th image is... Each pixel.

[0091] It should be noted that by calculating the health status of each vegetation connectivity domain and averaging it across the entire image, the physiological vitality of vegetation is quantified from local to global, providing a key health dimension for assessing the ecological stability of slopes. The calculation can effectively amplify the differences in spectral response between healthy green vegetation and unhealthy withered vegetation, thus capturing changes in the physiological state of vegetation more sensitively than simply using the green index. By first calculating the health of each connected domain and then taking the global average, it avoids the problem of large-area lush vegetation covering small-area vegetation degradation, ensuring the comprehensiveness of the assessment results.

[0092] Step S05: Divide each image into grids of equal size, calculate the ratio of the number of pixels in the connected component of green plants in each grid to the total number of pixels in the grid, and obtain the uniformity of green plant distribution in each image based on the standard deviation and mean of the ratio of the number of pixels in the connected component of green plants in each grid to the total number of pixels in the grid.

[0093] It should be noted that the above two steps basically achieve the calculation of vegetation layer and health status. However, when judging the vegetation cover of the slope, the distribution of vegetation must also be considered. The more uniform the vegetation distribution, the more accurate the area that can be covered, and it provides a prerequisite for extending to the uncovered area. Based on this, this step will calculate the vegetation distribution pattern to provide judgment conditions for a more scientific assessment of vegetation cover. The more uniform the vegetation distribution, the higher the probability of vegetation appearing at any location on the slope, proving that the vegetation cover is more comprehensive. Based on this, the vegetation distribution is calculated.

[0094] Specifically, the method of dividing each image into a grid of equal size, calculating the ratio of the number of pixels in the connected component of greenery in each grid to the total number of pixels in the grid, and obtaining the uniformity of greenery distribution in each image based on the standard deviation and mean of the ratio of the number of pixels in the connected component of greenery in each grid to the total number of pixels in the grid, includes the following specific methods:

[0095] Set a grid of size H*H, with the number of grid cells in each image being [number missing]. Z represents the side length of the captured image. The ratio of the number of pixels in the connected component of the greenery in each grid to the total number of pixels in the grid is calculated. The uniformity of the greenery distribution in each image is obtained based on the standard deviation and mean of this ratio. The specific method is as follows:

[0096] ;

[0097] ;

[0098] Formula parameters: This represents the ratio of the number of pixels in the connected component of the greenery in the s-th grid of the q-th image to the total number of pixels in the grid; the count() function extracts the number of pixels from its pixel set; This represents the set of pixels in the s-th grid of the q-th image that are covered by the green plant mask; This represents the set of all pixels in the s-th grid of the q-th image; This indicates the uniformity of the vegetation distribution in the q-th image; This represents all images in the q-th image. The average of the ratio of the number of pixels in the connected component of greenery in each grid to the total number of pixels in the grid; Represents all in the q-th image The standard deviation of the ratio of the number of pixels in the connected component of the greenery in a grid to the total number of pixels in the grid.

[0099] It should be noted that this step discretizes the spatial distribution of vegetation to quantitatively assess the stability and continuity of its cover, rather than focusing solely on the total amount. The use of local cover statistics within the grid, rather than global calculations, is because the effectiveness of slope protection is highly dependent on the uniformity of vegetation distribution—even with a high overall cover, the presence of large bare patches still poses a risk of soil erosion and landslides. Through calculations... Essentially, this calculation measures the ratio of the mean to the sum of the standard deviation and the mean. Statistically, this value reflects the relative dispersion of the data (its reciprocal is similar to the coefficient of variation). The more uniform the distribution, the closer this value is to 1; the more uneven the distribution, the smaller the value. The advantage of this approach is that it can accurately identify areas with deficient vegetation cover, supplementing indicators such as coverage and health from a spatial perspective, and providing crucial evidence for assessing the overall ecological protection performance of slopes.

[0100] Step S06: Based on the ratio of the number of first-class pixels to the total number of pixels in each image, the structural complexity of the vegetation in the corresponding shooting area of ​​each image, the health of the vegetation in each image, and the uniformity of vegetation distribution in each image, obtain the slope vegetation coverage quantification parameters, and set a threshold K to compare with the slope vegetation coverage quantification parameters to determine whether the slope vegetation coverage is good.

[0101] It should be noted that by multiplying and integrating four key indicators—coverage, structural complexity, health, and distribution uniformity—a comprehensive quantitative parameter is formed to evaluate the cover status of slope vegetation from multiple dimensions. This overcomes the limitations of single-indicator evaluation and yields more comprehensive and reliable conclusions on ecological protection effectiveness.

[0102] Specifically, based on the ratio of the number of first-class pixels to the total number of pixels in each image, the structural complexity of vegetation in the corresponding shooting area for each image, the health status of vegetation in each image, and the uniformity of vegetation distribution in each image, quantitative parameters of slope vegetation cover are obtained. A threshold K is then set and compared with the quantitative parameters of slope vegetation cover to determine whether the slope vegetation cover is good. The specific method is as follows:

[0103]

[0104] In the formula, This represents the quantification parameter for slope vegetation cover, where Q indicates the total number of Q images taken for the current slope. This represents the ratio of the number of pixels of type 1 in the t-th image to the total number of pixels in the image. This represents the structural complexity of the vegetation in the area corresponding to the t-th image. This represents the health status of the green plants in the t-th image. This indicates the uniformity of the vegetation distribution in the t-th image;

[0105] Set appropriate thresholds based on the regional characteristics of different areas. If the current slope vegetation cover quantification parameter I is greater than or equal to K, the current slope vegetation cover is considered to be good; if the current slope vegetation cover quantification parameter I is less than K, the vegetation cover is considered to be poor.

[0106] It's important to note that the product model, rather than a weighted average or other fusion method, is used because these four factors exhibit a mutually restrictive and synergistic relationship: extremely high vegetation cover, if composed of unhealthy vegetation or with highly uneven distribution, will significantly reduce its actual slope protection effect; conversely, healthy and evenly distributed vegetation, even with slightly lower cover, may possess good stability. The product model can keenly capture how any weakness in any dimension can have a significant negative impact on the final result. This aligns closely with the "barrel effect" in engineering practice. A dynamic threshold K is set for comparison based on different regional characteristics (such as climate, soil, and slope type), rather than using a fixed standard. This approach achieves a balance between standardization and regional adaptability in the evaluation results, making the evaluation method both scientifically universal and flexibly applicable to different regional engineering scenarios. It provides an objective, accurate, and operable decision-making basis for the acceptance of slope ecological restoration effects and long-term monitoring and maintenance.

[0107] It should be further noted that this invention does not specifically limit the value of K. In this embodiment, K is 0.55. This embodiment uses aerial images of "standard slopes" with known good maintenance conditions, such as lush vegetation, no erosion marks, and stable ecology, for retrospective calculation. The vegetation cover quantification parameter I values ​​are all above 0.65. At the same time, calculations are performed on "problem slopes" with known vegetation degradation and obvious bare patches, and their I values ​​are generally below 0.4. In other embodiments, the value of K can be determined based on a certain number of slope sample images in the region whose vegetation cover status has been authoritatively confirmed. The quantification parameter I values ​​are calculated for each sample, and the distribution range of I values ​​of these sample data is used as a statistical basis to select a critical value that can most effectively distinguish between "good" and "poor" samples. For example, the optimal discrimination point can be determined using an ROC curve as the applicable threshold K for the region. This method ensures the scientific nature and regionality of the threshold setting, so that the evaluation standard can adapt to the characteristics of different climates, soils, and vegetation types, and match the actual needs and expected goals of engineering management.

[0108] Please see Figure 2 It illustrates the second objective of the present invention, a structural block diagram of an aerial image slope vegetation cover status assessment system, which includes the following modules:

[0109] The UAV image acquisition module is used to acquire multiple images of slope areas in different regions using UAV aerial photography, and to traverse the slope images in the order they were captured.

[0110] The pixel classification module is used to calculate the green index based on the RGB three-channel values ​​of each pixel in the slope area image, filter all pixels with a green index greater than 0, divide the pixels with a green index greater than 0 into two categories of pixels with the greatest difference in green index, record the pixel with the larger green index as the first category of pixels, record the pixel with the smaller green index as the second category of pixels, and obtain the ratio of the number of first category pixels in each image to the total number of pixels in the image;

[0111] The green plant structure complexity calculation module is used to filter out all connected components in the first type of pixels, and to mark the connected components in the first type of pixels with a number of pixels greater than N as green plant connected components. The brightness of each green plant connected component is compared with the brightness of the global green plant connected components. Based on the comparison results, the green plant connected components are divided into multiple partitions, and the suspected shadow areas of the green plant connected components are marked. The green plant structure complexity of the corresponding shooting area of ​​each image is calculated based on the number and brightness of the suspected shadow area pixels in each green plant connected component and the number and brightness of all pixels in each green plant connected component.

[0112] The plant health calculation module is used to obtain the plant health level in each image based on the relative magnitudes of the G channel values, R channel values, and B channel values ​​of the pixels in the connected domain of the plant.

[0113] The green plant distribution uniformity calculation module is used to divide each image into grids of the same size, calculate the ratio of the number of pixels in the connected component of green plants in each grid to the number of pixels in the grid, and obtain the green plant distribution uniformity of each image based on the standard deviation and mean of the ratio of the number of pixels in the connected component of green plants in each grid to the number of pixels in the grid.

[0114] The slope vegetation cover determination module is used to obtain slope vegetation cover quantification parameters based on the ratio of the number of first-class pixels to the total number of pixels in each image, the complexity of the vegetation structure in the corresponding shooting area of ​​each image, the health of the vegetation in each image, and the uniformity of vegetation distribution in each image. It then sets a threshold K and compares it with the slope vegetation cover quantification parameters to determine whether the slope vegetation cover is good.

[0115] A third objective of this invention is to provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method for assessing the vegetation cover status of a slope using aerial imagery.

[0116] The fourth objective of this invention is to provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described method for assessing the vegetation cover status of a slope based on aerial imagery.

[0117] Beneficial effects of this invention:

[0118] This paper provides a comprehensive method for assessing the vegetation cover status of slopes using aerial imagery. It comprehensively judges the vegetation status through multiple indicators: coverage rate, structural complexity, health level, and distribution uniformity, which significantly improves the accuracy of assessment and engineering applicability.

[0119] By selecting the period of maximum sun altitude for aerial photography, the interference of terrain shadows is effectively reduced, ensuring the consistency of image illumination and providing a high-quality data foundation for subsequent vegetation identification and index calculation.

[0120] An adaptive classification method is used to perform secondary segmentation on pixels with a green index greater than 0, which effectively distinguishes between dense vegetation and sparse / misjudged areas, improving the accuracy and robustness of vegetation recognition.

[0121] By recursively segmenting and comparing brightness to identify shadow areas, the three-dimensional structure and hierarchical complexity of vegetation can be indirectly reflected, enhancing the assessment dimensions of vegetation ecological functions, such as soil and water conservation capacity.

[0122] A health index is constructed based on the phenomenon of red and blue light inhibition, which can sensitively capture changes in the physiological state of vegetation, avoid healthy vegetation from covering degraded areas, and improve the comprehensiveness and accuracy of health assessment.

[0123] By analyzing the uniformity of vegetation distribution through gridding, areas with coverage defects can be identified, which makes up for the inadequacy of relying solely on the overall coverage rate and enhances the spatial dimension assessment of slope ecological stability.

[0124] A comprehensive quantitative parameter I is constructed, which integrates coverage, structural complexity, health status and uniformity. The "barrel effect" is reflected through a product model, which scientifically reflects the overall effectiveness of vegetation cover and supports the judgment of regional adaptability threshold.

[0125] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0126] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0127] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0128] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0129] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for assessing the vegetation cover status of slopes using aerial imagery, characterized in that, The method includes the following steps: Step S01: Use drone aerial photography to acquire multiple images of slope areas in different regions, and traverse the slope images according to the order in which they were captured; Step S02: Calculate the green index based on the RGB three-channel values ​​of each pixel in the slope area image, filter all pixels with a green index greater than 0, divide the pixels with a green index greater than zero into two categories of pixels with the greatest difference in green index, record the pixel with the larger green index as the first category of pixels, and record the pixel with the smaller green index as the second category of pixels, and obtain the ratio of the number of first category pixels in each image to the total number of pixels in the image; Step S03: Filter out all connected components in the first type of pixels, and denote the connected components in the first type of pixels with a number of pixels greater than N as green plant connected components. Compare the brightness of each green plant connected component with the brightness of the global green plant connected components, and divide the green plant connected components into multiple partitions based on the comparison results. Mark the suspected shadow areas of the green plant connected components. Calculate the green plant structure complexity of the corresponding shooting area for each image based on the number and brightness of the suspected shadow area pixels in each green plant connected component and the number and brightness of all pixels in each green plant connected component. Step S04: Obtain the health status of the plants in each image based on the relative values ​​of the G channel, R channel, and B channel of the pixels in the connected component of the plants in each image. Step S05: Divide each image into grids of the same size, calculate the ratio of the number of pixels in the connected component of green plants in each grid to the number of pixels in the grid, and obtain the uniformity of green plant distribution in each image based on the standard deviation and mean of the ratio of the number of pixels in the connected component of green plants in each grid to the number of pixels in the grid. Step S06: Based on the ratio of the number of first-class pixels to the total number of pixels in each image, the structural complexity of the vegetation in the corresponding shooting area of ​​each image, the health of the vegetation in each image, and the uniformity of vegetation distribution in each image, obtain the slope vegetation coverage quantification parameters, and set a threshold K to compare with the slope vegetation coverage quantification parameters to determine whether the slope vegetation coverage is good.

2. The method for assessing the vegetation cover status of slopes using aerial imagery according to claim 1, characterized in that, The method of acquiring multiple images of slope areas from different regions using drone aerial photography and then traversing the slope images in the order they were captured includes the following specific steps: During the period of highest solar altitude, images of the slope area were captured using a drone. The drone maintained a distance of Y meters from the slope surface, and each image was Z*Z pixels in size. A total of Q images were captured, and all images were iterated through according to the time sequence of the images captured by the drone.

3. The method for assessing the vegetation cover status of slopes using aerial imagery according to claim 1, characterized in that, The method for calculating the green index based on the RGB three-channel values ​​of each pixel in the slope area image, filtering all pixels with a green index greater than 0, and dividing these pixels into two categories with the largest differences in green index, designating the category with the larger green index as the first category and the category with the smaller green index as the second category, includes the following specific steps: In the formula, This represents the green index of the pixel in the i-th row and j-th column of the q-th image; This represents the G channel value of the pixel in the i-th row and j-th column of the q-th image; This represents the R channel value of the pixel in the i-th row and j-th column of the q-th image; This represents the B channel value of the pixel in the i-th row and j-th column of the q-th image; Calculate the ratio of the number of pixels of type I to the total number of pixels in each image. The ratio of the number of pixels of type I to the total number of pixels in the q-th image is denoted as... ; Record all pixels in each image with a green index greater than 0, and divide all pixels in each image into two categories. The classification criterion is that the absolute value of the difference between the standard deviations of the green indices of the two categories of pixels in each image after classification, and the ratio of the sum of the standard deviations of the green indices of the two categories of pixels in each image, reach a maximum value. The category of pixels with a larger mean green index after classification is recorded as the first category of pixels in each image, and the category of pixels with a smaller mean green index is recorded as the second category of pixels in each image.

4. The method for assessing the vegetation cover status of slopes using aerial imagery according to claim 1, characterized in that, The process involves filtering out all connected components in the first category of pixels, and then defining connected components with a number of pixels greater than N as "green plant connected components." The brightness of each green plant connected component is compared with the brightness of the global green plant connected components. Based on the comparison results, the green plant connected components are partitioned multiple times, and suspected shadow areas are marked. The complexity of the green plant structure in the corresponding shooting area of ​​each image is calculated based on the number and brightness of pixels in the suspected shadow areas within each green plant connected component and the number and brightness of all pixels within each green plant connected component. The specific methods include: For each image, denote the connected components with more than N pixels in all connected components formed by the first type of pixels as each green plant connected component. Extract the V channel value of each pixel in each green plant connected component in the HSV color space. Then, for each green plant connected component in each image, perform the following recursive segmentation and shadow discrimination process sequentially. For the d-th green plant connected component in the q-th image, calculate the average value of the V channel value of each pixel in that connected component. Standard deviation of V channel values The recursive segmentation and shadow discrimination process is as follows: Locate the two points with the greatest Euclidean distance on the contour boundary, connect these two points to form a principal axis segment, then construct the perpendicular bisector of this principal axis segment. The principal axis segment and the perpendicular bisector together divide the original green plant connected domain into several secondary sub-regions. Calculate the standard deviation and mean of the V channel values ​​of the green plant connected domain in each secondary sub-region. If the standard deviation of the V channel of a pixel in a secondary sub-region is greater than... If the average V channel value of all pixels in the connected component of the greenery in a certain sub-region is lower than the threshold value, then the recursive segmentation and shadow discrimination process will continue to be performed on the next sub-region. Then, each connected component of green plants in the next lower-level sub-region is marked as a suspected shadow region; this continues until the standard deviation of the V channel of all pixels in the connected components of green plants in each sub-region is less than or equal to... Furthermore, for each sub-region, the suspected shadow area is marked, the recursive segmentation and shadow discrimination process is stopped, and all suspected shadow areas of the q-th image are traversed; The method for calculating the structural complexity of vegetation in the area corresponding to the q-th image is as follows: In the formula, This represents the structural complexity of the vegetation in the area captured by the q-th image. Indicates that the q-th image has a total of A suspected shadow area, The q-th image represents the... The number of pixels in the suspected shadow area The q-th image represents the... The total number of pixels in the unsegmented connected components of greenery at the locations corresponding to the suspected shadow areas. The q-th image represents the... The average V channel value of each pixel in the suspected shadow area Represents the q-th image. The average value of the V channel of the pixels in the unsegmented connected domain of greenery at the location corresponding to the suspected shadow area.

5. The method for assessing the vegetation cover status of slopes using aerial imagery according to claim 1, characterized in that, The method for obtaining the health status of plants in each image based on the relative magnitudes of the G-channel, R-channel, and B-channel values ​​of pixels in the connected component of the greenery in each image includes the following specific methods: ； ； In the formula, This indicates the health status of the plants in the q-th image; This represents the health status of the plants in the m-th connected component of the q-th image; Indicates that the q-th image has a total of A green plant connecting area; This represents the value of the G channel at the r-th pixel in the m-th connected component of the q-th plant in the image. This represents the value of the R channel of the r-th pixel in the m-th connected component of the q-th image; This represents the value of the B channel of the r-th pixel in the m-th connected component of the green plant in the q-th image; This indicates that the total number of connected components for the m-th green plant in the q-th image is... Each pixel.

6. The method for assessing the vegetation cover status of slopes using aerial imagery according to claim 1, characterized in that, The method of dividing each image into a grid of equal size, calculating the ratio of the number of pixels in the connected component of greenery in each grid to the total number of pixels in the grid, and obtaining the uniformity of greenery distribution in each image based on the standard deviation and mean of the ratio of the number of pixels in the connected component of greenery in each grid to the total number of pixels in the grid, includes the following specific methods: Set a grid of size H*H, with the number of grid cells in each image being [number missing]. Z represents the side length of the captured image. The ratio of the number of pixels in the connected component of the greenery in each grid to the total number of pixels in the grid is calculated. The uniformity of the greenery distribution in each image is obtained based on the standard deviation and mean of this ratio. The specific method is as follows: ； ； Formula parameters: This represents the ratio of the number of pixels in the connected component of the greenery in the s-th grid of the q-th image to the total number of pixels in the grid; the count() function extracts the number of pixels from its pixel set; This represents the set of pixels in the s-th grid of the q-th image that are covered by the green plant mask; This represents the set of all pixels in the s-th grid of the q-th image; This indicates the uniformity of the vegetation distribution in the q-th image; This represents all images in the q-th image. The average of the ratio of the number of pixels in the connected component of greenery in each grid to the total number of pixels in the grid; Represents all in the q-th image The standard deviation of the ratio of the number of pixels in the connected component of the greenery in a grid to the total number of pixels in the grid.

7. The method for assessing the vegetation cover status of slopes using aerial imagery according to claim 1, characterized in that, The method involves obtaining slope vegetation cover quantification parameters based on the ratio of the number of first-class pixels to the total number of pixels in each image, the complexity of the vegetation structure in the corresponding shooting area of ​​each image, the health of the vegetation in each image, and the uniformity of vegetation distribution in each image. A threshold K is then set and compared with the slope vegetation cover quantification parameters to determine whether the slope vegetation cover is good. The specific method is as follows: In the formula, This represents the quantification parameter for slope vegetation cover, where Q indicates the total number of Q images taken for the current slope. This represents the ratio of the number of pixels of type 1 in the t-th image to the total number of pixels in the image. This represents the structural complexity of the vegetation in the area corresponding to the t-th image. This represents the health status of the green plants in the t-th image. This indicates the uniformity of the vegetation distribution in the t-th image; Set appropriate thresholds based on the regional characteristics of different areas. If the current slope vegetation cover quantification parameter I is greater than or equal to K, the current slope vegetation cover is considered to be good; if the current slope vegetation cover quantification parameter I is less than K, the vegetation cover is considered to be poor.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the aerial image slope vegetation cover status assessment method as described in any one of claims 1 to 7.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the aerial image slope vegetation cover status assessment method as described in any one of claims 1 to 7.

10. A system for assessing the vegetation cover status of slopes based on aerial imagery, characterized in that, The system includes the following modules: The UAV image acquisition module is used to acquire multiple images of slope areas in different regions using UAV aerial photography, and to traverse the slope images in the order they were captured. The pixel classification module is used to calculate the green index based on the RGB three-channel values ​​of each pixel in the slope area image, filter all pixels with a green index greater than 0, divide the pixels with a green index greater than 0 into two categories of pixels with the greatest difference in green index, record the pixel with the larger green index as the first category of pixels, record the pixel with the smaller green index as the second category of pixels, and obtain the ratio of the number of first category pixels in each image to the total number of pixels in the image; The green plant structure complexity calculation module is used to filter out all connected components in the first type of pixels, and to mark the connected components in the first type of pixels with a number of pixels greater than N as green plant connected components. The brightness of each green plant connected component is compared with the brightness of the global green plant connected components. Based on the comparison results, the green plant connected components are divided into multiple partitions, and the suspected shadow areas of the green plant connected components are marked. The green plant structure complexity of the corresponding shooting area of ​​each image is calculated based on the number and brightness of the suspected shadow area pixels in each green plant connected component and the number and brightness of all pixels in each green plant connected component. The plant health calculation module is used to obtain the plant health level in each image based on the relative magnitudes of the G channel values, R channel values, and B channel values ​​of the pixels in the connected domain of the plant. The green plant distribution uniformity calculation module is used to divide each image into grids of the same size, calculate the ratio of the number of pixels in the green plant connected region of each grid to the number of pixels in the grid, and obtain the green plant distribution uniformity of each image based on the standard deviation and mean of the ratio of the number of pixels in the green plant connected region of each grid to the number of pixels in the grid. The slope vegetation cover determination module is used to obtain slope vegetation cover quantification parameters based on the ratio of the number of first-class pixels to the total number of pixels in each image, the complexity of the vegetation structure in the corresponding shooting area of ​​each image, the health of the vegetation in each image, and the uniformity of vegetation distribution in each image. It then sets a threshold K and compares it with the slope vegetation cover quantification parameters to determine whether the slope vegetation cover is good.

Images