Method for monitoring degradation degree of alpine grassland based on image data

Abstract

The invention discloses an alpine grassland degradation degree monitoring method based on image data, and relates to the technical field of alpine grassland degradation degree monitoring, and the method comprises the following steps: collecting image data of an alpine grassland with a known degradation degree as a sample image; converting the sample image into a processed image with high contrast; extracting surface features of the alpine grassland with different degradation degrees; analyzing color division intervals in the processed image according to the surface features, extracting greening features of the processed image, and analyzing a relationship between the greening features and the division intervals to obtain an interval calibration relationship equation; the degradation degree of the alpine grassland is calibrated according to the interval calibration relation equation; the method is used for solving the problem that an analysis result of the degradation degree of the alpine grassland has deviation due to the fact that an existing alpine grassland degradation degree monitoring technology cannot dynamically adjust judgment standards of different alpine grassland.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of alpine grassland degradation monitoring, in particular to an alpine grassland degradation monitoring method based on image data. BACKGROUND

[0002] Alpine grassland degradation monitoring technology refers to a series of technical methods for periodically, quantitatively observing, evaluating and grading the negative and continuous changes in the structure, function and service value of alpine grassland ecosystems by comprehensively utilizing remote sensing, geographic information systems, global positioning systems and modern ground surveying techniques.

[0003] The existing alpine grassland degradation monitoring technology usually calculates the degradation degree of alpine grassland by manual evaluation. In addition, a preset threshold is usually used to calculate and judge the degradation degree, which cannot dynamically adapt to different alpine grasslands. Moreover, when using image data to identify the degradation degree, the existing alpine grassland degradation monitoring technology usually uses remote sensing images for analysis. However, the resolution of remote sensing images is low, which cannot effectively identify the grassland degradation in a small range, and can only identify the approximate value of the overall degradation degree of alpine grassland, resulting in low data accuracy. For example, in the patent application with the publication number CN114639012A, a "grassland degradation evaluation method based on unmanned aerial vehicle hyperspectral remote sensing" is disclosed. This method analyzes the differences in the composition characteristics of plant communities with different degradation degrees to determine the threshold of evaluation indicators between different degradation levels. However, the specific analysis process of the threshold is not provided, which cannot ensure the accuracy of the threshold and whether it can dynamically adapt to different alpine grasslands. The existing alpine grassland degradation monitoring technology also has the problem of being unable to dynamically adjust the evaluation criteria for different alpine grasslands, resulting in deviation in the analysis results of the degradation degree of alpine grassland. SUMMARY

[0004] The present application aims to at least solve one of the technical problems in the prior art, by collecting image data of alpine meadow with known degradation degree as sample image, then pre-processing the sample image, converting the sample image into processed image with high contrast, analyzing the difference relationship between pixel points in the processed image of alpine meadow, extracting the surface features of alpine meadow with different degradation degrees, then analyzing the division interval of different processed images according to the surface features, extracting the greening features of the processed image and analyzing the relationship between the greening features and the division interval, obtaining the interval calibration relationship equation, dividing the alpine meadow to be monitored into different monitoring areas, finally shooting image data of the monitoring area by the unmanned aerial vehicle and analyzing the surface features of the image data, calibrating the degradation degree of the alpine meadow according to the interval calibration relationship equation, to solve the problem that the existing alpine meadow degradation degree monitoring technology cannot dynamically adjust the judgment standard of different alpine meadows, resulting in deviation in the analysis result of the degradation degree of alpine meadow.

[0005] To achieve the above-mentioned purpose, the present application provides an alpine meadow degradation degree monitoring method based on image data, comprising the following steps: Collecting image data of alpine meadow with known degradation degree as sample image; Pre-processing the sample image, converting the sample image into processed image with high contrast; Analyzing the difference relationship between pixel points in the processed image of alpine meadow, extracting the surface features of alpine meadow with different degradation degrees; According to the surface features, analyzing the division interval of color in the processed image, extracting the greening features of the processed image and analyzing the relationship between the greening features and the division interval, obtaining the interval calibration relationship equation; Dividing the alpine meadow to be monitored into different monitoring areas, shooting image data of the monitoring area by the unmanned aerial vehicle and analyzing the surface features of the image data, calibrating the degradation degree of the alpine meadow according to the interval calibration relationship equation.

[0006] Further, collecting image data of alpine meadow with known degradation degree as sample image specifically refers to providing image data of alpine meadow with known degradation degree as sample image by internet or experimenters, and the degradation degree is in percentage format.

[0007] Further, pre-processing the sample image, converting the sample image into processed image with high contrast specifically refers to image enhancement of the sample image by histogram equalization algorithm to obtain the processed image.

[0008] Further, analyzing the difference relationship between pixel points in the processed image of alpine meadow, extracting the surface features of alpine meadow with different degradation degrees comprises the following sub-steps: The pixel in the i-th row and j-th column of the processed image is labeled as R(i,j), where i and j are both non-zero natural numbers, i is the row number of the pixel, and j is the column number of the pixel. The color code of R(i,j) is labeled as C(i,j), and the color code is a hexadecimal format number encoding. Obtain the ranges of green, yellow, and brown color codes, and name them Green Code Range, Yellow Code Range, and Brown Code Range, respectively. At the same time, merge the Green Code Range and Yellow Code Range and name them Greening Range, and name the Brown Code Range Degradation Range. The green areas and degraded areas are sorted in ascending order. If there is no degraded area between two green areas, the two green areas are merged into one green area. If there is no green area between two degraded areas, the two degraded areas are merged into one degraded area. The merged green areas and degraded areas are sorted in ascending order. Two adjacent green areas and degraded areas are grouped into a calibration group. The criterion for determining adjacency is that there are no other green areas or degraded areas between the green areas and the degraded areas. Each group to be calibrated is analyzed independently. The R(i,j) corresponding to C(i,j) within the green range and degradation range of the group to be calibrated is labeled as GG(i,j) and DN(i,j) respectively. The R(i,j) corresponding to C(i,j) between the green range and degradation range of the group to be calibrated is labeled as OTH(i,j). The number of GG(i,j), DN(i,j) and OTH(i,j) is counted and labeled as NG, ND and NO respectively. NG, ND and NO are surface features.

[0009] Furthermore, based on the surface feature analysis, the color division intervals in the processed image are identified, the greening features of the processed image are extracted, and the relationship between the greening features and the division intervals is analyzed to obtain the interval calibration relationship equation, which includes the following sub-steps: The division intervals of different processed images are analyzed based on surface features; The greening features of the processed image are extracted and the relationship between the greening features and the divided intervals is analyzed to obtain the interval calibration relationship equation.

[0010] Furthermore, analyzing the segmentation of different processed images based on surface features includes the following sub-steps: The processed images are numbered using the symbol PR. n This indicates that n is a non-zero natural number and n is the index of PR; Each group to be calibrated was analyzed independently, and the PR was... n NO is divided into greening error number and degradation error number, respectively, by α n and βn This indicates the degree of degradation of the processed image, denoted as DOD. n ; There exists a system of equations Solving for α, we obtain n and β n , among which, NG n and ND n That is, PR n NG and ND; Mark the C(i,j) corresponding to OTH(i,j) as OCH(i,j), name the C(i,j) within the degradation range as the degradation color code, and obtain the minimum value of the difference between OCH(i,j) and the degradation color code, and name it as the degradation deviation color code; Calculate β n / NO n The calculation result is labeled as Q. n Sort the degradation deviation color codes in ascending order, and select the top Q... n The color code for a degenerate deviation is named a degenerate suspected code. The suspected degraded codes are included in the degraded range, and the range of the degraded range is recalculated to obtain the divided intervals. Each PR n Each group to be calibrated has a defined interval, and the PR n The partitioned intervals are uniformly marked as OP. n .

[0011] Furthermore, the greening features of the processed image are extracted and the relationship between the greening features and the divided intervals is analyzed to obtain the interval calibration relationship equation, which includes the following sub-steps: Integrate R(i,j) with the same color code within the greening area into a greening analysis group, count the number of R(i,j) in the greening analysis group and name it the greening analysis number, name the color code corresponding to the greening analysis group with the largest greening analysis number as the greening reference value, name the value with the smallest difference between the greening area and the degradation area as the edge reference value, and calculate the difference between the greening reference value and the edge reference value to obtain the greening characteristics; Get the minimum value of the divided interval, and mark it as TI. At the same time, get the minimum value of the original degradation range, and mark it as WI. Calculate TI-WI and mark the calculation result as FI. FIs with similar greening characteristics are grouped into a calibration analysis group. The corresponding NOs for each FI are also obtained. The FIs in the calibration analysis group are normalized and converted into normalized values, labeled as SIs. SIs are then numbered using the symbol KS. h This indicates that, where h is a non-zero natural number and h is the index of KS, KS h The corresponding NO label is KN h ; A two-dimensional coordinate system is established with the greening characteristics as the X-axis and SI as the Y-axis, named the Preliminary Calibration Analysis Chart. FI is entered into the Preliminary Calibration Analysis Chart according to the greening characteristics. With KN h For the X-axis, KS h Establish a two-dimensional coordinate system for the Y-axis, named the Depth Calibration Analysis Chart, and set KS... h According to KN h Enter the depth calibration analysis chart; Functional regression is performed on the preliminary calibration analysis plot and the deep calibration analysis plot to obtain the preliminary calibration relationship equation and the deep calibration relationship equation, respectively. The preliminary calibration relationship equation and the deep calibration relationship equation are the interval calibration relationship equations.

[0012] Furthermore, the alpine grasslands to be monitored are divided into different monitoring areas. Image data of the monitoring areas are captured by drones, and the surface features of the image data are analyzed. The degree of degradation of the alpine grasslands is calibrated according to the interval calibration relationship equation, including the following sub-steps: The alpine grasslands that need to be monitored are divided into different monitoring areas; By capturing image data of the monitored area using drones and analyzing the surface features of the image data, the degree of degradation of alpine grasslands is calibrated according to the interval calibration relationship equation.

[0013] Furthermore, the alpine grasslands that need to be monitored are divided into different monitoring areas, including the following sub-steps: Obtain the shooting area of ​​the drone, which is the actual area occupied by the image data captured by the drone; The alpine grasslands that need to be monitored are named the grasslands to be monitored. The grasslands to be monitored are divided into different monitoring areas using a grid division method. The area of ​​each monitoring area is the same as the area to be photographed, and each monitoring area is a square.

[0014] Furthermore, by capturing image data of the monitored area using drones and analyzing the surface features of the image data, the degradation degree of the alpine grassland is calibrated according to the interval calibration relationship equation, including the following sub-steps: Image data of the monitored area captured by drones are named images to be analyzed. Surface features of the images to be analyzed are extracted and named features to be analyzed. NG, ND and NO in the features to be analyzed are marked as UG, UD and UO respectively. Obtain the preliminary calibration analysis diagram, preliminary calibration relation equation, and depth calibration relation equation. Substitute UO into the depth calibration relation equation and label the solution as Z. Extract the greening features from the image to be analyzed and name them as greening analysis features. Name the coordinate points in the preliminary calibration analysis image where X equals the greening analysis features as calibration reference points. Mark the minimum and maximum values ​​of Y in the calibration reference points as YI and YA, respectively. Calculate (YA-YI)×Z+YI+WI, name the calculation result as the calibration threshold, replace the minimum value of the degradation range with the calibration threshold, and obtain the calibration interval; All pixels in the image to be analyzed whose color code falls within the calibration range are named degradation points, and all other pixels are named green points. The number of degradation points and green points are counted and labeled as MD and MG, respectively. MG / (MG+MD) is calculated to obtain the degree of degradation of the image to be analyzed.

[0015] The beneficial effects of this invention are as follows: This invention collects image data of alpine grasslands with known degradation levels as sample images, then preprocesses the sample images to convert them into processed images with high contrast, and then analyzes the differences between pixels in the processed images of alpine grasslands to extract surface features of alpine grasslands with different degradation levels. Then, based on the surface features, it analyzes the division intervals of different processed images. The advantage is that different alpine grasslands have different surface features, and since the statistically obtained surface features do not match the known degradation level, it is necessary to calibrate the degradation range when analyzing the surface features to obtain the division intervals, which provides reliable basic data for subsequent analysis and improves the accuracy and effectiveness of monitoring the degradation level of alpine grasslands. This invention extracts and processes the greening features of images and analyzes the relationship between these features and the defined intervals to obtain an interval calibration equation. This allows for the division of alpine grasslands into different monitoring areas. Finally, image data of the monitoring areas is captured by a drone, and the surface features of the image data are analyzed. The degree of degradation of the alpine grassland is then calibrated according to the interval calibration equation. The advantage lies in the fact that there are small, indistinguishable areas between the green and degraded parts of alpine grasslands, which cannot be directly judged by color. Therefore, it is necessary to analyze the interval calibration equation to divide these indistinguishable areas and divide the alpine grassland into different monitoring areas. The degree of degradation within these small areas can be assessed, and subsequent visualization technology can be used to display the degree of degradation in different areas of an alpine grassland for precise maintenance, further improving the accuracy and effectiveness of alpine grassland degradation monitoring. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the steps of the method of the present invention; Figure 2 This is a schematic diagram of the preliminary calibration analysis diagram of the present invention; Figure 3This is a schematic diagram of the depth calibration analysis diagram of the present invention; Figure 4 This is a schematic diagram of the electronic device of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1, please refer to Figure 1 As shown, this application provides a method for monitoring the degradation degree of alpine grasslands based on image data, including the following steps: Step S1: Collect image data of alpine grassland with a known degree of degradation as sample images; specifically, collect image data of alpine grassland with a known degree of degradation provided by the Internet or by the experimenter as sample images, with the degree of degradation in percentage format; In practice, the sample images and their degradation levels are known data. This embodiment constructs an assessment method that can adapt to different degradation levels of different alpine grasslands based on the known degradation levels of different alpine grasslands. There are multiple ways to obtain known data, such as downloading from the Internet or obtaining data from the Forestry Bureau. This embodiment does not limit the method of obtaining known data.

[0019] Step S2: Perform image preprocessing on the sample image to convert it into a processed image with high contrast. Specifically, perform image preprocessing on the sample image to convert it into a processed image with high contrast by using a histogram equalization algorithm to enhance the sample image and obtain the processed image. In practice, image enhancement can make the color contrast in the sample image more obvious, and it is an existing technical means, so this embodiment will not describe it in detail.

[0020] Step S3 involves analyzing the differences between pixels in the processed image of alpine grassland and extracting surface features of alpine grassland with different degrees of degradation. Step S3 includes the following sub-steps: Step S301: Mark the pixel in the i-th row and j-th column of the processed image as R(i,j), where i and j are both non-zero natural numbers and i is the row number of the pixel and j is the column number of the pixel. Mark the color code of R(i,j) as C(i,j), where the color code is a hexadecimal number encoding. In practice, for example, R(12,165) is presented as green. In its RGB three primary colors, R is 163, G is 242, and B is 60, which means that the color values ​​in the red, green, and blue channels are 163, 242, and 60, respectively, which is visually light green. Different RGB three primary colors are stored in the computer through hexadecimal codes, i.e., color codes. When R is 163, G is 242, and B is 60, the color code in the computer is a3f23c, that is, C(12,165) is a3f23c, which is a hexadecimal number.

[0021] Step S302: Obtain the range of green and brown color codes, and name them as green code range and brown code range respectively. At the same time, name the green code range as greening range and the brown code range as degradation range. Step S303: Sort the greening area and the degradation area in ascending order. If there is no degradation area between two greening areas, merge the two greening areas into one greening area. If there is no greening area between two degradation areas, merge the two degradation areas into one degradation area. Step S304: Sort the merged greening range and degradation range in ascending order. Two adjacent greening ranges and degradation ranges form a group to be calibrated. The basis for judging adjacency is that there are no other greening ranges or degradation ranges between the greening range and the degradation range. Step S305: Analyze each group to be calibrated independently. Mark the R(i,j) corresponding to C(i,j) within the greening range and degradation range of the group to be calibrated as GG(i,j) and DN(i,j) respectively. Mark the R(i,j) corresponding to C(i,j) between the greening range and degradation range of the group to be calibrated as OTH(i,j). Step S306: Count the number of GG(i,j), DN(i,j) and OTH(i,j) and label them as NG, ND and NO respectively. NG, ND and NO are surface features. In practice, the green, yellow, and brown code ranges all contain multiple discontinuous ranges. For example, the green code range includes [6E8B3D, 698B22] and [00FF00, 00FF7F], which are not continuous. The green code range is named the "greening range" because in alpine grasslands, human activity is minimal, there are no other constructions, and only plants exist. Most grassland plants are typically greenish. If a pixel is green, it can be determined that grassland plants definitely exist there and have not degraded. Brown represents the soil color. If the grassland has degraded, the pixel should appear brownish. Other colors cannot accurately determine whether the grassland has degraded, so calibration is necessary. For ease of understanding in this embodiment, an assumption is made here. Let's assume the green areas include [a,b] and [g,k], while the degraded areas include [c,d] and [p,q], where a < b < c < d < g < k < p < q. The lowercase letters follow the alphabetical order, so [a,b] < [c,d] < [g,k] < [p,q]. Between [c,d] and [g,k] lies [e,f]. If the color code falls within [e,f], it's impossible to accurately determine if the grassland is degraded; calibration is needed. [e,f] should be divided and assigned to either the green or degraded areas. Furthermore, if there's no degraded area between two green areas, they are merged. The same logic applies to degraded areas. For example, consider [a,b] and [g,k]. [a,b] and [g,k] are both green areas. If there are no degraded areas within the range (b,g) between [a,b] and [g,k], then [a,b] and [g,k] are merged into [a,k]. However, if a degraded area is stored within this range, it is not merged. Taking [c,d] and [g,k] as another example, one is a degraded area and the other is a green area. Since there are no green or degraded areas within the range (d,g) between them, [c,d] and [g,k] are considered as a group for analysis. In this embodiment, this is called the calibration group. In this example, there are a total of 3 calibration groups: [a,b] and [c,d], and [c,d] and [g,k]. And [g,k] and [p,q], each analysis only analyzes one group to be calibrated. Taking the group to be calibrated [c,d] and [g,k] as an example, [c,d] is the degradation range and [g,k] is the greening range. The number of pixels in the image within [g,k] is used to obtain NG, the number of pixels in the image within [c,d] is used to obtain ND, and the number of pixels within (d,g) is used to obtain NO. Finally, the NG is 1,483,650, the ND is 438,824, and the NO is 151,126. The NG, ND and NO listed here are all data of the group to be calibrated [c,d] and [g,k]. The surface characteristics of each group to be calibrated are independent of each other.

[0022] Step S4 involves analyzing the color division intervals in the processed image based on surface features, extracting the greening features of the processed image, and analyzing the relationship between the greening features and the division intervals to obtain the interval calibration relationship equation. Step S4 includes the following sub-steps: Step S401: Analyze the division intervals of different processed images based on surface features; Step S401 includes the following sub-steps: Step S4011: Number the processed images using the symbol PR. n This indicates that n is a non-zero natural number and n is the index of PR; Step S4012: Analyze each group to be calibrated independently, and transfer the PR data. n NO is divided into greening error number and degradation error number, respectively, by α n and β n This indicates the degree of degradation of the processed image, denoted as DOD. n ; Step S4013, a system of equations exists. Solving for α, we obtain n and β n , among which, NG n and ND n That is, PR n NG and ND; Step S4014: Mark the C(i,j) corresponding to OTH(i,j) as OCH(i,j), name the C(i,j) within the degradation range as the degradation color code, obtain the minimum value of the difference between OCH(i,j) and the degradation color code, and name it as the degradation deviation color code; Step S4015, calculate β n / NO n The calculation result is labeled as Q. n Sort the degradation deviation color codes in ascending order, and select the top Q... n The color code for a degenerate deviation is named a degenerate suspected code. Step S4016: Include the suspected degraded codes in the degraded range, recalculate the range of the degraded range, obtain the divided intervals, and assign each PR... n Each group to be calibrated has a defined interval, and the PR n The partitioned intervals are uniformly marked as OP. n ; In a specific implementation, for example, in this embodiment there are 1268 processed images, then the PR is marked. n1≤n≤1268. Taking the data listed in this embodiment as an example, the processed image is PR1, NG1 is 1483650, ND1 is 438824, and NO1 is 151126. This is to analyze the calibration groups [c,d] and [g,k]. The analysis process of different calibration groups is completely independent. NO is divided into α1 and β1. The sum of α1 and β1 is 151126, and the degradation degree of PR1 is 35%, which means that 33% of the grassland area in the processed image has been degraded, that is, DOD1 is 30%. The solutions are α1 and β1, which are 71550 and 79576, respectively. The degradation deviation color code and Q are calculated. n Essentially, the goal is to find the OCH(i,j) closest to the degenerate range as a potential degenerate code. Specifically, within the range (d,g) between [c,d] and [g,k], the top 79576 OCH(i,j) closest to g are considered potential degenerate codes. The final statistical analysis yields a partitioned interval of [c,e], where each PR... n Each group to be calibrated has a division interval. For example, the division interval between the calibrated groups [c,d] and [g,k] of PR1 is [c,e]. The division interval represents the degradation range after calibration, while the green range after calibration is (e,k]. In other words, all color codes in the group to be calibrated are divided into green range or degradation range. The boundary between the degradation range and the green range after calibration is the accurate threshold for determining whether a pixel belongs to green or degradation.

[0023] Step S402: Extract the greening features of the processed image and analyze the relationship between the greening features and the divided intervals to obtain the interval calibration relationship equation; Step S402 includes the following sub-steps: Step S4021: Integrate R(i,j) with the same color code within the greening area into a greening analysis group, count the number of R(i,j) in the greening analysis group and name it the greening analysis number, name the color code corresponding to the greening analysis group with the largest greening analysis number as the greening reference value, name the value with the smallest difference between the greening area and the degradation area as the edge reference value, and calculate the difference between the greening reference value and the edge reference value to obtain the greening characteristics; Step S4022: Obtain the minimum value of the divided interval, marked as TI, and at the same time obtain the minimum value of the original degradation range, marked as WI. Calculate TI-WI and mark the calculation result as FI. In specific implementation, taking the calibration groups [c,d] and [g,k] as examples, the division range of the calibration groups [c,d] and [g,k] is [c,e]. The color code with the most occurrences in [c,d] is 5D8110, meaning the greening reference value is 5D8110, while the edge reference value is 5D8029. The calculated greening feature is E7, which is 231 in decimal. In step S4022, both TI and WI are the minimum values ​​of the degradation range. This is because the default degradation range is larger than the greening range, but in the calibration groups [c,d] and [g,k], [c,d] is the degradation range. Let [g,k] be the greening range. We need to obtain the maximum value of the degradation range, i.e., TI is the maximum value e of the interval [c,e], and WI is the maximum value d of the original degradation range [c,d]. The final calculated FI is ed. Assuming e is 5D8165 and d is 5D8135, the calculated FI is 30. This 30 is in hexadecimal format; converting it to decimal gives FI as 48. The values ​​of TI and WI depend on the size difference between the greening and degradation ranges, as the ultimate goal is to distinguish whether the color code between the greening and degradation ranges belongs to greening or degradation.

[0024] Step S4023: Integrate FIs with the same greening characteristics into a calibration analysis group, and simultaneously obtain the corresponding NO for each FI. Perform normalization calculations on the FIs in the calibration analysis group, convert the FIs into normalized values, label them as SIs, and number the SIs using the symbol KS. h This indicates that, where h is a non-zero natural number and h is the index of KS, KS h The corresponding NO label is KN h ; Please see Figure 2 As shown, in step S4024, a two-dimensional coordinate system is established with the greening features as the X-axis and SI as the Y-axis, named the preliminary calibration analysis diagram, and FI is entered into the preliminary calibration analysis diagram according to the greening features. Please see Figure 3 As shown, in step S4025, KN h For the X-axis, KS h Establish a two-dimensional coordinate system for the Y-axis, named the Depth Calibration Analysis Chart, and set KS... h According to KN h Enter the depth calibration analysis chart; Step S4026: Perform function regression on the preliminary calibration analysis chart and the deep calibration analysis chart to obtain the preliminary calibration relationship equation and the deep calibration relationship equation, which are the interval calibration relationship equations. In this specific implementation, the range (d,g) between [c,d] and [g,k] is temporarily referred to as the error portion. Since the greening feature is the difference between the most frequently occurring color code within the greening range and the edge reference value, and the most frequently occurring color code within the greening range represents that many pixels in the processed image have the same color as the greening feature, the color codes are concentrated in their vicinity. When the greening feature is smaller, more colors in the error portion will represent actual greening rather than degradation, and vice versa. Therefore, the greening feature can reflect the division of color codes within the error portion to a certain extent, and FI is the dividing boundary of color codes within the error portion. Thus, there is a certain relationship between the greening feature and FI. For ease of observation, FI is converted to a normalized SI for analysis. Based on this, a preliminary calibration analysis diagram is constructed as follows: Figure 2 As shown, when constructing the preliminary calibration analysis map and the deep calibration analysis map, all PRs need to be... n All groups to be calibrated are included in the reference range; FIs with the same greening characteristics are integrated into a single calibration analysis group. Figure 2 The coordinate points at different heights within the same X-magnitude range constitute a calibration analysis group. The corresponding NO for FI is 151126. For example, in all PRs... n In all the groups to be calibrated, the minimum value of FI is 20 and the maximum value is 100. If a certain FI is 40, then the corresponding SI is (40-20) / (100-20) = 0.25. Similarly, calculate all the SIs and number them to obtain KS1 to KS. 16584 At the same time, KN1 to KN were obtained. 16584 Through observation Figure 2 It is easy to see that for the same greening features, FI is not unique; that is, FI fluctuates, with some FI values ​​being larger and others smaller. NO represents the number of pixels within the error region. A larger NO indicates a smaller interval between the color codes of pixels within the error region, and a smaller NO indicates a larger interval between the color codes of pixels within the error region. In other words, a larger NO results in a smaller FI under the same greening features. Based on this, a depth calibration analysis map is constructed as follows: Figure 3 As shown, the depth calibration relationship equation obtained through function regression is YF=-10. -11 ×XF 2 -8×10 -6 ×XF+1.0852, where YF is SI and XF is KN. h Since the preliminary calibration relation equation is not required in the subsequent analysis process, but the preliminary calibration analysis chart is used directly, the preliminary calibration relation equation will not be explained here.

[0025] Step S5 involves dividing the alpine grassland to be monitored into different monitoring areas, capturing image data of the monitoring areas using a drone, analyzing the surface features of the image data, and calibrating the degree of degradation of the alpine grassland according to the interval calibration relationship equation. Step S5 includes the following sub-steps: Step S501: Divide the alpine grasslands that need to be monitored into different monitoring areas; Step S501 includes the following sub-steps: Step S5011: Obtain the shooting area of ​​the drone. The shooting area is the actual area occupied by the image data captured by the drone. Step S5012: The alpine grassland to be monitored is named the grassland to be monitored. The grassland to be monitored is divided into different monitoring areas by grid division method. The area of ​​each monitoring area is the shooting area, and the monitoring area is square. In practice, the alpine grassland can be divided using the existing grid division method. Controlling the shooting area is to ensure that all locations of the alpine grassland can be analyzed. It should be noted that when the drone is shooting the monitoring area, it should be positioned at the geometric center of the monitoring area to take a top-down shot.

[0026] Step S502: Take image data of the monitoring area by drone and analyze the surface features of the image data; calibrate the degree of degradation of alpine grassland according to the interval calibration relationship equation. Step S502 includes the following sub-steps: Step S5021: Take image data of the monitoring area by drone, name it as the image to be analyzed, extract the surface features of the image to be analyzed, name them as the features to be analyzed, and mark NG, ND and NO in the features to be analyzed as UG, UD and UO respectively; Step S5022: Obtain the preliminary calibration analysis diagram, the preliminary calibration relational equation, and the depth calibration relational equation. Substitute UO into the depth calibration relational equation and mark the solution as Z. Step S5023: Extract the greening features of the image to be analyzed and name them as greening analysis features. Name the coordinate points in the preliminary calibration analysis image where X equals the greening analysis features as calibration reference points. Mark the minimum and maximum values ​​of Y in the calibration reference points as YI and YA, respectively. Step S5024: Calculate (YA-YI)×Z+YI+WI, name the calculation result as the calibration threshold, replace the minimum value of the degradation range with the calibration threshold, and obtain the calibration interval; Step S5025: Name all pixels in the image to be analyzed whose color code is within the calibration range as degradation points, and name all the remaining pixels as green points. Count the number of degradation points and green points, and label them as MD and MG respectively. Calculate MG / (MG+MD) to finally obtain the degree of degradation of the image to be analyzed. In practice, assuming that UG, UD, and UO are obtained as 125864, 33622, and 24683 respectively, substitute XB=24683 into YB=-10 -11 ×XB 2 -8×10 -6 ×XB+1.0852, solving for Z yields 0.88. The result is rounded to two decimal places, and the extracted greening analysis feature is 200. Figure 2 The coordinate point with X=200 is named the calibration reference point. YI and YA are obtained as 0.59 and 0.77 respectively. When calculating the calibration threshold, (YA-YI)×Z is calculated because YA and YI at this point are... Figure 2 In the dimension of X=200, it is actually equivalent to Figure 3 The maximum value of SI in the equation is 1 and the minimum value is 0, because... Figure 2 The dataset with X=200 dimension only contains calibration reference points. Therefore, calculating (YA-YI)×Z gives the correct position of SI between YA and YI, considering NO. Adding this to YI yields the accurate SI of the image to be analyzed. This calculated result, SI, needs to be converted back to FI. A normalization algorithm is used to reconstruct SI, obtaining the accurate FI of the image. Adding FI to WI gives the boundary between the defined interval and the green area, i.e., the calibration threshold. It's important to note that in the formula (YA-YI)×Z... In the +YI+WI method, before adding WI, the calculation result of (YA-YI)×Z+YI+ must be restored to FI and then added to WI. Finally, each degradation range of the image to be analyzed is calibrated to obtain different calibration intervals. All pixels in the image to be analyzed whose color code is within the calibration interval are named degradation points, and all other pixels are named greening points. Finally, the degradation degree of the image to be analyzed is calculated. This method can adaptively calibrate the applicable degradation range in different image data to accurately calculate the degradation degree of different alpine grasslands.

[0027] Example 2, please refer to Figure 4 As shown, Figure 4A schematic diagram of an electronic device is provided, which may include a processor, a communication interface, a memory, and a communication bus. The processor, communication interface, and memory communicate with each other via the communication bus. The memory stores computer-readable instructions, which the processor can call. When the processor executes a computer-readable instruction, it performs steps such as those in the image data-based method for monitoring the degradation of alpine grasslands, to achieve the following functions: acquiring image data of alpine grasslands with known degradation levels as sample images; converting the sample images into processed images with high contrast; extracting surface features of alpine grasslands with different degradation levels; analyzing the color division intervals in the processed images based on the surface features, extracting the greening features of the processed images, and analyzing the relationship between the greening features and the division intervals to obtain an interval calibration relationship equation; calibrating the degradation level of the alpine grasslands based on the interval calibration relationship equation.

[0028] Furthermore, when the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0029] Example 3: This application also provides a computer program product, which includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the image data-based alpine grassland degradation monitoring method provided by the above methods. The method includes: acquiring image data of alpine grassland with known degradation levels as sample images; converting the sample images into processed images with high contrast; extracting surface features of alpine grassland with different degradation levels; analyzing the color division intervals in the processed images based on the surface features, extracting the greening features of the processed images and analyzing the relationship between the greening features and the division intervals to obtain an interval calibration relationship equation; and calibrating the degradation level of the alpine grassland based on the interval calibration relationship equation.

[0030] Example 4: This application also provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it performs the steps of the above-described method for monitoring the degradation degree of alpine grassland based on image data, to achieve the following functions: acquiring image data of alpine grassland with known degradation degree as sample images; converting the sample images into processed images with high contrast; extracting surface features of alpine grassland with different degradation degrees; analyzing the color division intervals in the processed images based on the surface features, extracting the greening features of the processed images and analyzing the relationship between the greening features and the division intervals to obtain the interval calibration relationship equation; calibrating the degradation degree of alpine grassland based on the interval calibration relationship equation.

[0031] Based on the above description of the embodiments, the embodiments of the present invention can be provided as methods, systems, or computer program products. Based on this understanding, the above technical solutions, in essence or in terms of their contribution to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or certain parts of the embodiments.

[0032] In the embodiments provided in this application, it should be understood that the disclosed system or method can be implemented in other ways. The embodiments described above are merely illustrative. For example, the division of modules or units is only a logical functional division, and there may be other division methods in actual implementation. Furthermore, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the coupling or direct coupling or communication connection shown or discussed may be through some communication interfaces. The indirect coupling or communication connection between systems, modules, and units may be electrical, mechanical, or other forms.

[0033] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for monitoring the degree of degradation of alpine meadow based on image data, characterized in that, The method comprises the following steps: Collecting image data of alpine meadow with known degradation degree as sample images; Image preprocessing is performed on the sample images to convert the sample images into processed images with high contrast; Analyzing the difference relationship between pixel points in the processed images of the alpine meadow, and extracting surface features of the alpine meadow with different degradation degrees; According to the surface feature analysis, the division interval of the color in the processed image is obtained, the greening feature of the processed image is extracted, and the relationship between the greening feature and the division interval is analyzed to obtain an interval calibration relationship equation; The alpine meadow to be monitored is divided into different monitoring areas, image data of the monitoring areas are photographed by a UAV, and the surface features of the image data are analyzed, and the degradation degree of the alpine meadow is calibrated according to the interval calibration relationship equation.

2. The method for monitoring the degradation degree of alpine meadow based on image data according to claim 1, characterized in that, The image data of the alpine meadow with known degradation degree as sample images is specifically provided by the Internet or an experimenter, and the degradation degree is in percentage format.

3. The method for monitoring the degradation degree of alpine meadow based on image data according to claim 2, characterized in that, The image preprocessing of the sample images to convert the sample images into processed images with high contrast is specifically image enhancement of the sample images by a histogram equalization algorithm to obtain the processed images.

4. The method for monitoring the degradation degree of alpine meadow based on image data according to claim 3, characterized in that, Analyzing the difference relationship between pixel points in the processed images of the alpine meadow, and extracting surface features of the alpine meadow with different degradation degrees comprises the following sub-steps: Mark the pixel point in the i-th row and the j-th column of the processed image as R(i,j), wherein i and j are both non-zero natural numbers, i is the row number of the pixel point, and j is the column number of the pixel point, and the color code of R(i,j) is marked as C(i,j), which is a hexadecimal format of digital coding; Obtain the range of green, yellow and brown color codes, respectively named as green code range, yellow code range and brown code range, and combine the green code range and the yellow code range and name them as greening range, and name the brown code range as degradation range; Sort the greening range and the degradation range in ascending order, if there is no degradation range between two greening ranges, merge the two greening ranges into one greening range, if there is no greening range between two degradation ranges, merge the two degradation ranges into one degradation range; Sort the combined greening range and degradation range in ascending order, and form a group to be calibrated group consisting of two adjacent greening ranges and degradation ranges, and the adjacent basis is that there is no other greening range or degradation range between the greening range and the degradation range; Analyze each group to be calibrated group independently, mark R(i,j) corresponding to C(i,j) in the greening range and the degradation range in the group to be calibrated group as GG(i,j) and DN(i,j) respectively, and mark R(i,j) corresponding to C(i,j) between the greening range and the degradation range in the group to be calibrated group as OTH(i,j); Statistically count the number of GG(i,j), DN(i,j) and OTH(i,j), and mark them as NG, ND and NO respectively, which are surface features.

5. The method for monitoring the degradation degree of alpine meadow based on image data according to claim 4, characterized in that, Based on surface feature analysis, the color division intervals in the processed image are identified. The greening features of the processed image are extracted, and the relationship between these greening features and the division intervals is analyzed to obtain the interval calibration relationship equation. This process includes the following sub-steps: The division intervals of different processed images are analyzed based on surface features; The greening features of the processed image are extracted and the relationship between the greening features and the divided intervals is analyzed to obtain the interval calibration relationship equation.

6. The method for monitoring the degradation degree of alpine meadow based on image data according to claim 5, characterized in that, Analyzing the division of different processed images based on surface features includes the following sub-steps: The processed image is numbered by the symbol PR n where n is a non-zero natural number and n is the order number of PR. For each group to be calibrated group is analyzed independently, PR n The NO is divided into greening error and degradation error, respectively through α n And β n Indicate, get the degree of processing image degradation, marked as DOD n ; Equations exist , and solving for a n and β n , where NG n and ND n are the NG and ND of PR n ; Mark the C(i,j) corresponding to OTH(i,j) as OCH(i,j), name the C(i,j) within the degradation range as the degradation color code, and obtain the minimum value of the difference between OCH(i,j) and the degradation color code, and name it as the degradation deviation color code; Compute β n / NO n , mark the result as Q n , sort the degenerate bias color codes in ascending order, and name the top Q n degenerate bias color codes as degenerate suspect codes; The degenerate suspected codes are included in the degenerate range, the range of the degenerate range is re-counted, and each PR n partition interval is obtained Each group of to-be-calibrated groups of PR n has a partition interval, and the partition intervals of PR n are uniformly marked as OP .

7. The method for monitoring the degradation degree of alpine meadow based on image data according to claim 6, characterized in that, Extracting greening features from the processed image and analyzing the relationship between these features and the defined intervals to obtain the interval calibration relationship equation includes the following sub-steps: Integrate R(i,j) with the same color code within the greening area into a greening analysis group, count the number of R(i,j) in the greening analysis group and name it the greening analysis number, name the color code corresponding to the greening analysis group with the largest greening analysis number as the greening reference value, name the value with the smallest difference between the greening area and the degradation area as the edge reference value, and calculate the difference between the greening reference value and the edge reference value to obtain the greening characteristics; Get the minimum value of the divided interval, and mark it as TI. At the same time, get the minimum value of the original degradation range, and mark it as WI. Calculate TI-WI and mark the calculation result as FI. The FIs with the same greening features are integrated into a calibration analysis group, and the corresponding NOs of the FIs are obtained. The FIs in the calibration analysis group are normalized to convert the FIs into normalized values, which are marked as SIs. The SIs are numbered, and the number is represented by KS h , wherein h is a non-zero natural number, and h is the serial number of KS. The KS h corresponding to the NO is marked as KN h . A two-dimensional coordinate system is established with the greening characteristics as the X-axis and SI as the Y-axis, named the Preliminary Calibration Analysis Chart. FI is entered into the Preliminary Calibration Analysis Chart according to the greening characteristics. Take KN h as the X axis, KS h as the Y axis to establish a two-dimensional coordinate system, named depth calibration analysis chart, and record KS h according to KN h in the depth calibration analysis chart; Functional regression is performed on the preliminary calibration analysis plot and the deep calibration analysis plot to obtain the preliminary calibration relationship equation and the deep calibration relationship equation, respectively. The preliminary calibration relationship equation and the deep calibration relationship equation are the interval calibration relationship equations.

8. The method for monitoring the degradation degree of alpine meadow based on image data according to claim 7, characterized in that, The alpine grasslands to be monitored are divided into different monitoring areas. Image data of the monitoring areas are captured by drones and the surface features of the image data are analyzed. The degree of degradation of the alpine grasslands is calibrated according to the interval calibration relationship equation, including the following sub-steps: The alpine grasslands that need to be monitored are divided into different monitoring areas; By capturing image data of the monitored area using drones and analyzing the surface features of the image data, the degree of degradation of alpine grasslands is calibrated according to the interval calibration relationship equation.

9. The method for monitoring the degradation degree of alpine meadow based on image data according to claim 8, characterized in that, The process of dividing the alpine grasslands that need to be monitored into different monitoring areas includes the following sub-steps: Obtain the shooting area of ​​the drone, which is the actual area occupied by the image data captured by the drone; The alpine grasslands that need to be monitored are named the grasslands to be monitored. The grasslands to be monitored are divided into different monitoring areas using a grid division method. The area of ​​each monitoring area is the same as the area to be photographed, and each monitoring area is a square.

10. The method for monitoring the degradation degree of alpine meadow based on image data according to claim 9, characterized in that, By capturing image data of the monitored area using drones and analyzing the surface features of the image data, the degradation degree of alpine grassland is calibrated according to the interval calibration relationship equation, including the following sub-steps: Image data of the monitoring area is taken by the unmanned aerial vehicle, and is named as an image to be analyzed. Surface features of the image to be analyzed are extracted, and are named as features to be analyzed. NG, ND and NO in the features to be analyzed are marked as UG, UD and UO respectively; A preliminary calibration analysis graph, a preliminary calibration relationship equation and a depth calibration relationship equation are obtained. UO is substituted into the depth calibration relationship equation, and a result of solving is marked as Z; Greening features of the image to be analyzed are extracted, and are named as greening analysis features. A coordinate point with X equal to the greening analysis features in the preliminary calibration analysis graph is named as a calibration reference point. The minimum and maximum values of Y in the calibration reference point are marked as YI and YA respectively; (YA-YI)×Z+YI+WI is calculated, a calculation result is named as a calibration threshold, the minimum value of the degradation range is replaced by the calibration threshold, and a calibration interval is obtained; Pixel points with color codes in the calibration interval in the image to be analyzed are all named as degradation points. The remaining pixel points are all named as greening points. The number of the degradation points and the greening points are counted, and are marked as MD and MG respectively. MG / (MG+MD) is calculated, and finally the degradation degree of the image to be analyzed is obtained.

Images