An alpine meadow biomass acquisition method and system, and a storage medium

Abstract

The application provides a high-cold grassland biomass acquisition method and system and a storage medium, and belongs to the technical field of data processing. The method comprises the following steps: acquiring the vegetation height, vegetation coverage and aboveground biomass of each sample region in each aerial image; obtaining the aboveground biomass prediction value of each sampling region of all aerial images according to the measured aboveground biomass and the spectral information of the pixel points in each sampling region; and obtaining the aboveground biomass of the region corresponding to the pixel points in the remote sensing image according to the aboveground biomass prediction value of all sampling regions in each remote sensing image and the spectral information of the pixel points corresponding to the sampling regions. The application aims to solve the problem that when the current aboveground biomass upscaling estimation model is established, the scale difference between the remote sensing satellite and the ground sampling plot is too large, and the saturation phenomenon occurs when a single vegetation index is used to fit the aboveground biomass, so that the difference between the prediction result and the actual result is too large.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of data processing, and particularly relates to a method and system for obtaining high-cold grassland biomass and a storage medium. BACKGROUND

[0002] As one of the most important vegetation types in the third pole of the world, high-cold grassland is the most unique grassland ecosystem in the world, and plays an important role in carbon cycle, climate regulation, water and soil conservation, and maintenance of biodiversity. However, due to the influence of climate change and human activities, about 19-60% of the high-cold grassland in the Qinghai-Tibet Plateau is in different degrees of degradation, and accurate monitoring of the health status of high-cold grassland is the basis for maintaining its ecological and production functions. Aboveground biomass reflects the productivity and carrying capacity of grassland, and is a key indicator for evaluating the health status of grassland. Estimating aboveground biomass is of great significance for grassland resource management and adaptive decision-making.

[0003] When obtaining the aboveground biomass of high-cold grassland by ground investigation combined with multi-source satellite remote sensing inversion, first, the satellite takes a picture of the measured area to obtain a satellite remote sensing image, and then the spectral information of the region corresponding to each pixel point in the satellite remote sensing image is obtained. A number of small sample regions are divided in the region corresponding to a number of pixel points in the satellite remote sensing image, and the aboveground biomass in each sample region is obtained manually. Using a random forest algorithm, an aboveground biomass upscaling estimation model from small sample regions to the region corresponding to the pixel points in the satellite remote sensing image is constructed according to the aboveground biomass of the small sample regions divided in the region corresponding to a number of pixel points in the satellite remote sensing image and the spectral information of the pixel points. Due to the large difference between the area of the ground region corresponding to each pixel point in the remote sensing image and the area of each sample region, and the poor resolution of the remote sensing image, the aboveground biomass obtained by ground investigation combined with multi-source satellite remote sensing inversion deviates too much from the actual value. In addition, the vegetation index calculated from the vegetation spectral information is prone to sensitivity reduction in areas with high vegetation coverage, which leads to early saturation of the estimated biomass result, and thus it is urgent to improve the process of ground investigation combined with multi-source satellite remote sensing inversion. SUMMARY

[0004] In order to solve the problem that the difference between the biomass predicted by the existing method and the actual biomass is too large due to the large difference between the scale of the ground sampling region and the scale of the satellite when obtaining the aboveground biomass of high-cold grassland, the application provides a method for obtaining the aboveground biomass of high-cold grassland.

[0005] In order to achieve the above-mentioned purpose, the application provides the following technical scheme:

[0006] Obtaining remote sensing images of alpine grassland regions, obtaining a plurality of sampling regions from the remote sensing images, obtaining a plurality of aerial photograph regions from each sampling region, and obtaining a plurality of sample regions from each aerial photograph region; obtaining the vegetation height, vegetation coverage and aboveground biomass of each sample region, the vegetation coverage of each aerial photograph region, the vegetation coverage of each sampling region, and the spectral information of each pixel point in the remote sensing image;

[0007] According to the vegetation height, vegetation coverage and aboveground biomass of all sample regions in each aerial photograph region, the prediction relationship of the aboveground biomass of each aerial photograph region and the vegetation height are obtained; according to the predicted aboveground biomass, vegetation height and vegetation coverage of all aerial photograph regions in each sampling region, the prediction relationship of the aboveground biomass of each sampling region and the vegetation height are obtained; according to the spectral information of each pixel point in each remote sensing image and the predicted aboveground biomass of the corresponding sampling region, the prediction model of the aboveground biomass of the corresponding region of the pixel point in each remote sensing image is obtained;

[0008] The vegetation coverage and vegetation height of each aerial photograph region are input into the prediction relationship of the aboveground biomass of the aerial photograph region, and the predicted aboveground biomass of each aerial photograph region is obtained; the vegetation coverage and vegetation height of each sampling region are input into the prediction relationship of the aboveground biomass of the sampling region, and the predicted aboveground biomass of the sampling region is obtained; the spectral information of each pixel point in each remote sensing image is input into the prediction model of the aboveground biomass of the corresponding region of the pixel point in the remote sensing image, and the predicted aboveground biomass of the corresponding region of each pixel point in the remote sensing image is obtained.

[0009] According to the vegetation height, vegetation coverage and aboveground biomass of all sample regions in each aerial photograph region, the prediction relationship of the aboveground biomass of each aerial photograph region and the vegetation height are obtained; combining the vegetation coverage and vegetation height of the aerial photograph region, the predicted aboveground biomass of each aerial photograph region is obtained.

[0010] According to the predicted aboveground biomass, vegetation height and vegetation coverage of all aerial photograph regions in each sampling region, the prediction relationship of the aboveground biomass of each sampling region and the vegetation height are obtained.

[0011] The vegetation coverage and vegetation height of each sampling region are input into the prediction relationship of the aboveground biomass of the sampling region, and the predicted aboveground biomass of the sampling region is obtained.

[0012] According to the spectral information of each pixel point in each remote sensing image and the predicted aboveground biomass of the corresponding sampling area, an aboveground biomass prediction model of the region corresponding to each pixel point in each remote sensing image is obtained;

[0013] The spectral information of each pixel point in each remote sensing image is input into the aboveground biomass prediction model of the region corresponding to each pixel point in the remote sensing image, and the predicted aboveground biomass of the region corresponding to each pixel point in the remote sensing image is obtained.

[0014] Further, the specific steps of obtaining the sampling areas from the remote sensing image are as follows:

[0015] The remote sensing image of the research area is obtained using a remote sensing satellite, and the remote sensing image is divided into a plurality of sampling areas in the first step. The regions corresponding to a plurality of pixel points in the remote sensing image are randomly selected.

[0016] Further, the specific steps of obtaining the prediction relationship of the aboveground biomass of each aerial photography region and the vegetation height according to the vegetation height, the vegetation coverage and the aboveground biomass of all sample regions in each aerial photography region are as follows:

[0017] The product of the vegetation height and the vegetation coverage of the sample region in the first aerial photography region is taken as the horizontal axis, and the aboveground biomass of the sample region is taken as the vertical axis to construct a two-dimensional sample space. The product of the vegetation height and the vegetation coverage of the first sample region in the first aerial photography region is taken as the horizontal coordinate of the first data point in the two-dimensional sample space.

[0018] The aboveground biomass of the first sample region in the first aerial photography region is taken as the vertical coordinate of the first data point in the two-dimensional sample space, and the first data point in the two-dimensional sample space is obtained. All sample regions in the first aerial photography region are mapped into the two-dimensional sample space to obtain a plurality of data points in the two-dimensional sample space. The random forest algorithm is used to fit the data points in the two-dimensional sample space, and the obtained fitting relationship is recorded as the prediction relationship of the aboveground biomass of the first aerial photography region. The mean value of the vegetation height of all sample regions in the first aerial photography region is recorded as the vegetation height of the first aerial photography region.

[0019] The mean value of the vegetation height of all sample regions in the first aerial photography region is recorded as the vegetation height of the first aerial photography region. The mean value of the vegetation height of all sample regions in the first aerial photography region is recorded as the vegetation height of the first aerial photography region.

[0020] The random forest algorithm is used to fit the data points in the two-dimensional sample space, and the obtained fitting relationship is recorded as the prediction relationship of the aboveground biomass of the first aerial photography region.

[0021] The mean value of the vegetation height of all sample regions in the first aerial photography region is recorded as the vegetation height of the first aerial photography region. The mean value of the vegetation height of all sample regions in the first aerial photography region is recorded as the vegetation height of the first aerial photography region.​​​​​​

[0022] Further, the specific steps of obtaining the above-ground biomass prediction value of each aerial region are as follows:

[0023] The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region. The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region.

[0024] Further, the specific steps of obtaining the above-ground biomass prediction value of each aerial region are as follows:

[0025] The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region.

[0026] The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region. The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region. The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region. The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region. The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region.

[0027] The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region.

[0028] The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region. The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region.

[0029] The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region. The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region.

[0030] The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region.

[0031] The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region. The product of the vegetation height and the vegetation coverage of the first aerial region is taken as the input of the prediction relationship of the above-ground biomass of the first aerial region, and the output is recorded as the above-ground biomass prediction value of the first aerial region.

[0032] ​​​​​​​​​​ The product of the vegetation height and the vegetation coverage of the first sampling area is taken as the input of the prediction relationship of the aboveground biomass of the first sampling area, and the output is taken as the aboveground biomass prediction value of the first sampling area.

[0033] Further, the specific steps of obtaining the prediction model of the aboveground biomass of the pixel point corresponding area in each remote sensing image according to the spectral information of each pixel point in each remote sensing image and the aboveground biomass prediction value of the corresponding sampling area are as follows:

[0034] The aboveground biomass prediction value of the first sampling area in the alpine grassland region corresponding to the first remote sensing image and the spectral information of the pixel point in the first remote sensing image corresponding to the first sampling area in the alpine grassland region corresponding to the first remote sensing image are taken as the data values of two dimensions of the first target point in the first remote sensing image, respectively, to obtain all target points in the first remote sensing image.

[0035] A model is trained according to the data values of two dimensions of all target points in the first remote sensing image using a random forest algorithm, and the model is taken as the prediction model of the aboveground biomass of the pixel point corresponding area in the first remote sensing image.

[0036] The application further provides an alpine grassland biomass acquisition system, which comprises:

[0037] A data acquisition module is configured to acquire remote sensing images of an alpine grassland region, acquire a plurality of sampling areas from the remote sensing images, acquire a plurality of aerial photograph areas from each sampling area, acquire a plurality of sample areas from each aerial photograph area, and acquire the vegetation height, the vegetation coverage and the aboveground biomass of each sample area, the vegetation coverage of each aerial photograph area, the vegetation coverage of each sampling area and the spectral information of each pixel point in the remote sensing images.

[0038] A prediction model construction module is configured to obtain the prediction relationship of the aboveground biomass and the vegetation height of each aerial photograph area according to the vegetation height, the vegetation coverage and the aboveground biomass of all sample areas in each aerial photograph area.

[0039] The prediction relationship of the aboveground biomass and the vegetation height of each sampling area is obtained according to the aboveground biomass prediction value, the vegetation height and the vegetation coverage of all aerial photograph areas in each sampling area.

[0040] ​​​​​​​​​​​​According to the spectral information of each pixel point in each remote sensing image and the aboveground biomass prediction value of the corresponding sampling area, an aboveground biomass prediction model of the corresponding area of each pixel point in each remote sensing image is obtained.

[0041] The prediction model uses a module for inputting the vegetation coverage and vegetation height of each aerial photograph area into the prediction relationship of the aboveground biomass of the aerial photograph area to obtain the aboveground biomass prediction value of each aerial photograph area.

[0042] The vegetation coverage and vegetation height of each sampling area are input into the prediction relationship of the aboveground biomass of the sampling area to obtain the aboveground biomass prediction value of the sampling area.

[0043] The spectral information of each pixel point in each remote sensing image is input into the prediction model of the aboveground biomass of the corresponding area of the pixel point in the remote sensing image to obtain the aboveground biomass prediction value of the corresponding area of each pixel point in the remote sensing image.

[0044] The application also provides a computer readable storage medium, and the storage medium stores a computer program.

[0045] The application provides a method for obtaining the aboveground biomass of an alpine grassland.

[0046] When the aboveground biomass of the alpine grassland is obtained by ground investigation and multi-source satellite remote sensing inversion, the adaptation of the aerial photograph image and the sampling area scale and the adaptation of the aerial photograph image and the satellite scale are better than the adaptation of the satellite and the sampling scale. The aboveground biomass of the aerial photograph image is obtained by the aboveground biomass of a plurality of small sampling areas in the aerial photograph area. The aboveground biomass prediction value of the aerial photograph area is obtained according to the vegetation coverage of the aerial photograph area. However, the aboveground biomass of an area is related to the vegetation coverage and the vegetation height due to the complex vertical structure of the grassland. Therefore, the aboveground biomass prediction value of the aerial photograph area is obtained by the product of the vegetation height and the vegetation coverage of a plurality of sampling areas in the aerial photograph area and the aboveground biomass, the accuracy of the aboveground biomass of the aerial photograph area is enhanced, and the accuracy of the aboveground biomass prediction value of the corresponding area of the satellite is improved. BRIEF DESCRIPTION OF DRAWINGS

[0047] To more clearly illustrate the embodiments and design schemes of the present invention, the accompanying drawings required for this embodiment will be briefly described below. 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.

[0048] Figure 1 A flowchart illustrating the steps involved in obtaining biomass from alpine grasslands;

[0049] Figure 2 Spatial distribution of grassland types and sampling areas in the Qilian Mountains;

[0050] Figure 3 A schematic diagram for scaling up

[0051] Figure 4 A flowchart illustrating the process of scaling up;

[0052] Figure 5 A collection of aerial images;

[0053] Figure 6 A collection of aerial photographs of the sample area;

[0054] Figure 7 A schematic diagram illustrating the process of constructing aboveground biomass estimation models at the quadrat and aerial photograph scales and resampling to the quadrat scale.

[0055] Figure 8 To obtain sampling information on aboveground biomass, vegetation cover, and vegetation height;

[0056] Figure 9 Regression fit plots of aboveground biomass versus vegetation cover and vegetation cover × vegetation height at quadrat and aerial photograph scales;

[0057] Figure 10 Accuracy maps for estimating aboveground biomass at the quadrat and aerial photograph scales.

[0058] Figure 11 This is a map showing the distribution of aboveground biomass in the Qilian Mountains region. Detailed Implementation

[0059] To enable those skilled in the art to better understand and implement the technical solutions of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and should not be construed as limiting the scope of protection of the present invention.

[0060] Example 1

[0061] This invention provides a method for obtaining biomass from alpine grasslands, specifically as follows: Figure 1 As shown, it includes the following steps:

[0062] Step S001: Obtain a remote sensing image of an alpine grassland region, obtain a plurality of sampling regions from the remote sensing image, obtain a plurality of aerial photograph regions from each sampling region, and obtain a plurality of sample regions from each aerial photograph region; obtain the vegetation height, vegetation coverage and aboveground biomass of each sample region, the vegetation coverage of each aerial photograph region, the vegetation coverage of each sampling region, and the spectral information of each pixel point in the remote sensing image.

[0063] It should be noted that, with the development of science and technology, small unmanned aerial vehicles can quickly obtain high-resolution vegetation coverage and other spectral information at each aerial photograph scale, and the coverage range of the unmanned aerial vehicle aerial image can effectively match the spatial scale of ground investigation and satellite pixels, so that the unmanned aerial vehicle aerial image becomes an intermediate bridge connecting the estimation of aboveground biomass from the sample region scale to the pixel point scale in the remote sensing image. Therefore, in the upscaling process of the present application, a plurality of regions corresponding to the pixel points in the remote sensing image are randomly selected as sampling regions, and then each sampling region is divided into a plurality of small aerial photograph regions, and a plurality of smaller sample regions are divided in each aerial photograph region. The upscaling process is as shown in Figure 3 . Figure 3 The aerial photograph image of the sampling region corresponding to one pixel point in the remote sensing image is shown in (a) of Figure 3 The flight route of the unmanned aerial vehicle in the sampling region is shown by the line formed by a plurality of points on (b) of Figure 3 The distribution of sample regions in each aerial photograph region is shown in (c) of Figure 7 The flowchart of constructing the aboveground biomass estimation model at the quadrat scale and the aerial photograph scale and resampling to the sample plot scale, Figure 7 The image at the quadrat scale is shown in (a) of Figure 7 The distribution of quadrats at the aerial photograph scale is shown in (b) of Figure 7 Each small block in (c) corresponds to an aerial photograph scale.

[0064] It's important to further clarify that current methods for acquiring aboveground biomass within a region fall into two categories: one involves manual collection, and the other divides a large region into smaller subregions. The relationship between aboveground biomass and vegetation cover in the larger region is then derived by analyzing the aboveground biomass and vegetation cover data from these smaller subregions. However, due to the complex vertical structure of grasslands, the aboveground biomass of each region is related not only to vegetation cover but also to vegetation height. Therefore, during upscaling, in addition to acquiring the vegetation cover, the vegetation height of each sample region is also obtained. Then, based on the acquired data, the predicted aboveground biomass value for each pixel in the remote sensing image is obtained using the upscaling approach. Figure 4 This is a schematic diagram of the scaling process of the present invention, wherein... Figure 4 Figure (a) shows the aboveground biomass, vegetation height, and vegetation cover obtained from multiple sample areas within each aerial photography area using existing methods. Figure 4 Figure (b) shows the relationship between aboveground biomass and the product of vegetation height and vegetation cover for all sample areas within each aerial survey area, obtained using the random forest algorithm based on the data collected from each sample area. Figure 4 Figure (c) shows the predicted aboveground biomass of the aerial photography area based on vegetation height and vegetation cover. Figure 4 Figure (d) shows the predicted aboveground biomass for each pixel in the remote sensing image, obtained by continuing the upscaling approach. Figure 8 Figure (a) shows the sampling information for aboveground biomass. Figure 8 Figure (b) shows the sampling information for vegetation height. Figure 8 Figure (c) shows the sampling information of vegetation height.

[0065] Specifically, remote sensing satellites were used to acquire remote sensing images of the study area, obtaining the spectral information of each pixel in each image. The study area selected here is the Qilian Mountains region on the northeastern edge of the Qinghai-Tibet Plateau. In the... Random selection within Zhang remote sensing image The region corresponding to each pixel is denoted as the sampling region. The area of ​​the ground region corresponding to one pixel in the remote sensing image is... In this embodiment, the number of sampling regions is preset within the region corresponding to each remote sensing image. Preset pixel scale This will be described using this as an example; other values ​​can be set in other implementations. Figure 2This is a spatial distribution map of grassland types and sampling areas in the Qilian Mountains. The acquisition of spectral information for each pixel in the remote sensing image is a well-known technique and will not be elaborated upon in this embodiment.

[0066] Furthermore, each sampling area is divided into... The size is The aerial photography area is defined by using a drone to photograph each area, obtaining aerial images of each area, and thus determining the vegetation cover of each area. The drone is used to photograph each sampling area to obtain the vegetation cover of that area. Determining the vegetation cover of an area based on its aerial images is a well-known technique and will not be elaborated upon in this embodiment. The number of aerial photography areas within each sampling area is preset in this embodiment. The length of the aerial photography area The width of the aerial photography area This example will be used to illustrate the concept; other values ​​can be used in other implementations. Figure 3 Figure (a) shows the top-down camera interface of the drone. Figure 3 Figure (b) shows the flight path of the drone within the sampling area, and the drone takes a picture at each digital node. Figure 5 This is a collection of aerial images.

[0067] Furthermore, each aerial photography area is divided into... Three sample regions were selected, and vegetation height and aboveground biomass were obtained for each region using existing methods. Aerial photography of the sample regions using drones was also conducted to obtain the vegetation cover for each region. The size of each sample region was [size missing]. The number of sample areas within each aerial photography area preset in this implementation is... The length of the preset sample region Meters, width The value of meter is used as an example in this description; other values ​​can be set in other implementations. Figure 4 Figure (c) shows the distribution of the sample area within the aerial photography area. Figure 6 This is a collection of aerial photographs of the sample area.

[0068] Thus, the spectral features of each pixel in the remote sensing image, the vegetation cover of each sampling area, the vegetation cover of each aerial photography area, and the vegetation height, vegetation cover, and aboveground biomass of each sample area are obtained.

[0069] Step S002: Based on the vegetation height, vegetation cover, and aboveground biomass of all sample areas within each aerial photography area, obtain the prediction formula for aboveground biomass and vegetation height for each aerial photography area; input the vegetation cover and vegetation height of each aerial photography area into the prediction formula for aboveground biomass and vegetation height for the aerial photography area to obtain the predicted value of aboveground biomass for each aerial photography area.

[0070] It should be noted that due to the complex vertical structure of grasslands, aboveground biomass can vary significantly even within sample areas with the same vegetation cover. Therefore, the aboveground biomass of each sample area is fitted with the product of vegetation height and vegetation cover to obtain a predictive formula for the aboveground biomass of each pixel within each aerial photography area. Figure 9 Figure (a) shows the regression fit diagram of aboveground biomass and vegetation cover at the quadrat scale and the aerial photograph scale. Figure 9 Figure (b) shows the regression fitting plots of aboveground biomass and vegetation cover × vegetation height at both quadrat and aerial photograph scales. It can be observed that compared to fitting using the product of vegetation height and vegetation cover, directly fitting using vegetation cover and aboveground biomass results in overfitting earlier. Figure 9 Figure (a) shows the accuracy of aboveground biomass estimation at the quadrat scale. Figure 9 Figure (b) shows the accuracy of aboveground biomass estimation at the scale of aerial photographs. Figure 9 In the graph, N represents the aboveground biomass involved in the modeling, the black line represents a 1:1 ratio, the red line represents the fitted line, the gray line represents the 95% confidence interval, and the bar chart represents the scatter density distribution.

[0071] Specifically, taking the first A two-dimensional sample space is constructed by using the product of vegetation height and vegetation cover in the sample area within the aerial photography area as the horizontal axis and the aboveground biomass of the sample area as the vertical axis.

[0072] Furthermore, with the first The first aerial photography area The product of vegetation height and vegetation cover in each sample region is used as the product of the vegetation height and vegetation cover in the two-dimensional sample space. The x-coordinate of the nth data point; with the nth The first aerial photography area The aboveground biomass of the sample area is used as the first sample area in the two-dimensional sample space. The ordinate of the nth data point is obtained; the nth data point in the two-dimensional sample space is obtained. Data points.

[0073] Furthermore, the first All sample areas within a single aerial photography area are mapped into a two-dimensional sample space to obtain several data points within the two-dimensional sample space.

[0074] Furthermore, the random forest algorithm is used to fit the data points in the two-dimensional sample space, and the resulting fitting relationship is denoted as the th... The prediction formula for the aboveground biomass of the aerial photography area is as follows. The random forest algorithm is used to fit the data points in the two-dimensional sample space, which is a well-known existing technique and will not be elaborated upon in this embodiment.

[0075] Furthermore, the first The mean vegetation height of all sample areas within the aerial photography area is denoted as the i-th. The vegetation height of the aerial photography area. (The first...) The product of vegetation height and vegetation cover in each aerial photography area is used as the first... The input to the predictive formula for the aboveground biomass of each aerial photography area will be denoted as the output of the i-th... Predicted aboveground biomass values ​​for the aerial photography area.

[0076] Thus, the predicted ground biomass value for each aerial photography area is obtained.

[0077] Step S003: Based on the predicted values ​​of aboveground biomass, vegetation height, and vegetation cover of all aerial photography areas within each sampling area, obtain the prediction formula for aboveground biomass and vegetation height of each sampling area; input the vegetation cover and vegetation height of each sampling area into the prediction formula for aboveground biomass of the sampling area to obtain the predicted value of aboveground biomass of the sampling area.

[0078] Specifically, taking the first A two-dimensional coordinate system is constructed by using the product of vegetation height and vegetation cover in the aerial photography area within each sampling area as the horizontal axis and the predicted aboveground biomass value of the aerial photography area as the vertical axis.

[0079] Furthermore, with the first Within the sampling area, the first The product of vegetation height and vegetation cover in each aerial photography area is used as the product of the vegetation height and vegetation cover in the two-dimensional coordinate system. The x-coordinate of the nth sample point; with the nth... Within the sampling area, the first The above-ground biomass of the aerial photograph area is used as the first... The ordinate of the nth sample point is obtained; the nth sample point in the two-dimensional coordinate system is obtained. One sample.

[0080] Furthermore, the first All aerial photography areas within a sampling area are mapped into a two-dimensional coordinate system, resulting in several data points within the two-dimensional coordinate system.

[0081] Furthermore, the random forest algorithm is used to fit the data points in the two-dimensional coordinate system, and the resulting fitting relationship is denoted as the first... The prediction formula for aboveground biomass in the sampling area is as follows. The random forest algorithm is used to fit the data points in the two-dimensional sample space, which is a well-known technique and will not be elaborated upon in this embodiment.

[0082] Furthermore, the first The average vegetation height of all aerial photography areas within a sampling area is denoted as the nth. The vegetation height of each sampling area. The first... The product of vegetation height and vegetation cover in each sampling area is used as the product of the vegetation height and vegetation cover in the first sampling area. The input of the prediction formula for aboveground biomass in each sampling area will be denoted as the output of the formula. Predicted aboveground biomass values ​​for each sampling area.

[0083] Thus, the predicted aboveground biomass value for each sampling area is obtained.

[0084] Step S004: Based on the spectral information of each pixel in each remote sensing image and the predicted value of the aboveground biomass of the corresponding sampling area, obtain the prediction model of the aboveground biomass of the area corresponding to each pixel in the remote sensing image; input the spectral information of each pixel in each remote sensing image into the prediction model of the aboveground biomass of the area corresponding to the pixel in the remote sensing image to obtain the predicted value of the aboveground biomass of the area corresponding to each pixel in the remote sensing image.

[0085] Specifically, the first The first remote sensing image corresponds to the alpine grassland region. The predicted aboveground biomass value for the first sampling area, and the predicted value for the second sampling area. The first remote sensing image corresponds to the alpine grassland region. The first sampling region corresponds to the first The spectral information of pixels in the remote sensing image is used as the first... Zhang remote sensing image The data values ​​of the two dimensions of the target point are obtained to get the first... All target points within the remote sensing image.

[0086] Furthermore, using the random forest algorithm, based on the... Using the two-dimensional data values ​​of all target points within a remote sensing image, a model is trained, denoted as the... A predictive model for aboveground biomass in the region corresponding to a pixel in a remote sensing image. The model uses a random forest algorithm trained on data, a well-known existing technique; therefore, it will not be described in detail in this embodiment.

[0087] Furthermore, the first In the remote sensing image, the first The spectral information of the nth pixel is used as the spectral information of the nth pixel. The input to the prediction model for aboveground biomass in the region corresponding to a pixel in the remote sensing image is denoted as the i-th. In the remote sensing image, the first The predicted aboveground biomass value corresponds to each pixel in the region. This yields the aboveground biomass for all regions of the Qilian Mountains. Figure 7 This is a map showing the distribution of aboveground biomass in the Qilian Mountains region.

[0088] This concludes the embodiment.

[0089] Example 2:

[0090] The method proposed in this invention will be further explained below, taking the Qilian Mountains region on the northeastern edge of the Qinghai-Tibet Plateau as an example.

[0091] Step 1: Study Area and Data Acquisition

[0092] The invention area is located in the Qilian Mountains region on the northeastern edge of the Qinghai-Tibet Plateau, which is an important ecological security barrier and water conservation area in Northwest China. The Qilian Mountains have an elevation of approximately 1749 meters. At an altitude of 5767m, the terrain gradually rises from southeast to northwest. This region has a plateau continental climate, characterized by high-altitude, high-latitude mountainous climates, with distinct zonation of temperature and precipitation. The average annual temperature is approximately -3.24℃, and the average annual precipitation is about 300mm. This unique geographical location and climatic conditions have resulted in a variety of typical grassland types. Figure 2 ), including alpine meadows, alpine steppes, temperate steppes, and alpine deserts, with common plant species being Kentucky bluegrass (Poa chinensis). Poa pratensis Stipa purpurea ), Purple Needlegrass ( Elymus nutans ), Lysimachia christinae ( Carex parvula ), Alpine Songgrass ( Carex moorcroftii Sibbaldianthe bifurca ), Qinghai-Tibet sedge ( Figure 2 ), Potentilla biloba ( Figure 3 )wait.

[0093] Based on the fixed observation plots of the UAV aerial photography-ground collaborative survey previously deployed by the team of this invention in the Qilian Mountains ( Figure 3 ), selecting 65 sample plots based on drone aerial photography system (FragMap) in 2023 Field surveys were conducted in the Qilian Mountains region from July to August 2024. At each fixed 200m × 200m observation plot, new or existing grid flight paths were created using FragMap. Figure 3Each flight line can obtain 16 aerial photos (Fig. 1(b)). After the flight line is added, the Mavic 2 Pro (DJI, China) drone is set to fly at a height of 20 m, a speed of 4 m / s, and camera parameters (white balance, sensitivity, etc.) according to the aerial photography requirements. Subsequently, the drone will start autonomous flight and shooting according to the preset parameters. Figure 5 Each aerial photo covers an area of 25 m x 36 m on the ground with a resolution of 1 cm. At the same time, ground surveys are conducted in sample plots matching the aerial photography scale. Six 0.5 m x 0.5 m quadrat frames are randomly placed in each sample plot (Fig. 1(c)), and vegetation height, coverage, and aboveground biomass are investigated. Figure 6 The specific steps of vegetation investigation are as follows: 1) measure the vegetation height (10 repeated measurements of the height of the reproductive branches and the top of the leaves) within the quadrat frame; 2) take a vertical photo of the vegetation within the quadrat frame; 3) cut the aboveground vegetation within the quadrat frame (cut green branches for shrubs) and put it into a paper bag.

[0094] Based on the GEE (Google Earth Engine) platform, the 2023 In July-August 2024, the MOD13Q1 (61) product with a resolution of 30 m x 30 m was used to calculate the vegetation index (Equations 1 and 2) after removing the cloud. The normalized difference vegetation index (NDVI_mean), normalized difference vegetation index (NDVI_max), normalized difference vegetation index (NDVI_min), and enhanced vegetation index (EVI) were included. Based on the geographic spatial data cloud platform, ASTER GDEM 30 m resolution elevation data (DEM) was downloaded, and the surface analysis tool was used to extract the slope (slope) and aspect (aspect).

[0095]

[0096]

[0097] where NIR is the near-infrared band, R is the red band, and B is the blue band.

[0098] Second step, aerial photo and vegetation sample analysis

[0099] The sample plot scale photos were converted from pixel type to double precision type and normalized to the range of 0-1 using matlab software. The threshold value was calculated according to the green index (Excess Green Index, EGI = 2G-R-B, R, G, and B represent red, green, and blue bands, respectively), and the accuracy of the threshold value was adjusted in combination with GIS10.8 software to obtain the vegetation coverage of the sample plot scale (the proportion of pixels less than the threshold value in the total number of pixels) Figure 5 With Figure 6 ). The vegetation coverage of the aerial photo scale was based on the grassland vegetation coverage analysis software (Pixel Classifier Manual) independently developed by the author's team. Green vegetation pixels or bare ground pixels were outlined according to the threshold value using the OpenCV package, and the original photo was manually corrected to calculate the proportion of vegetation patch area in the entire photo area, which was the vegetation coverage Figure 7 With Figure 7 ). The above-ground part of the vegetation needs to be placed in a 65℃ oven for 24h, and the air is dried to a constant weight (dry weight). The above-ground biomass is weighed.

[0100] Step 3, modeling and estimation of above-ground biomass at different scales

[0101] Random forest (Random Forest) is a machine learning algorithm widely used in simulation and prediction processes, which combines multiple decision trees to improve the prediction ability and accuracy of the model, and is widely used in the estimation of above-ground biomass. For model construction, the number of random forest regression trees (ntree) is set to 500 by default, and is adjusted by 10-fold cross-validation (cv). The ranger package of the regression model is used to train and predict the optimized parameters, and 70% of the data is randomly selected for model training, and 30% of the data is used for validation. Then the determination coefficient (R 2 ) and the root mean square error (RMSE) are calculated to evaluate the accuracy of the model (equations 3 and 4).

[0102]

[0103]

[0104] In the formula, n is the number of samples, and represent the measured above-ground biomass and the predicted above-ground biomass, respectively, is the average value of the measured above-ground biomass.

[0105] A random forest algorithm was used to construct an upscaling estimation model for aboveground biomass based on measured aboveground biomass, UAV imagery, and MODIS image data. The specific steps were: 1) A quadrat-scale (0.5m × 0.5m) random forest model was established using vegetation cover, vegetation height, and measured aboveground biomass. Figure 7 (a) Figure 1) Estimating aboveground biomass (25m × 36m) at the aerial photograph scale based on vegetation cover and vegetation height; 2) Establishing a random forest model at the aerial photograph scale based on the estimated aboveground biomass at the aerial photograph scale and MODIS image data. Figure 8 (b) Figure 1) The aboveground biomass at the pixel scale (30m × 30m) was estimated based on MODIS image data; 3) The estimated aboveground biomass at the pixel scale was resampled to the plot scale (250m × 250m) to obtain the spatial distribution of aboveground biomass in the entire Qilian Mountains grassland. Figure 8 (Figure (c) in the middle).

[0106] Step 4: Statistical Analysis

[0107] Linear / nonlinear regression was used to fit the relationships between aboveground biomass and vegetation cover, as well as between aboveground biomass and vegetation height × vegetation cover, and the explanatory coefficients were calculated. MODIS image data was preprocessed and thematic maps were drawn using GIS 10.8 software. The significance of the aboveground biomass upscaling estimation model was tested using a permutation test. All other statistical analyses and modeling were performed using R 4.3.1 software, and all plotting was done using Origin software and the ggplot2 package.

[0108] The calculation results will be analyzed below.

[0109] 1. Vegetation parameter characteristics and their fitting relationship with aboveground biomass

[0110] Aboveground biomass is concentrated in the range of 0 275g·m -2 It exhibits an approximately normal distribution, with an average value of approximately 110 g·m³. -2 And it appears most frequently ( Figure 8 (Figure (a)). There are significant differences in vegetation cover; both low and high cover areas have a large number of samples, while the number of samples in areas with intermediate cover (approximately 60%-80%) is relatively small. Figure 9 (Figure (b)). The vegetation height shows a clear right skewness, with the main distribution range being 0. Vegetation heights exceeding 7cm and 15cm are relatively few. Figure 9 (Figure (c)). At the quadrat and aerial photograph scales, vegetation cover exhibited a power function relationship with aboveground biomass, explaining 39% and 61% of the variation in aboveground biomass, respectively.Figure 10 (See Figure (a)). Notably, aboveground biomass reaches saturation with increasing vegetation cover. In contrast, vegetation height × vegetation cover shows a linear relationship with aboveground biomass, explaining 47% and 74% of the variation in aboveground biomass, respectively. Figure 10 (Figure (b) in the middle).

[0111] 2. Accuracy assessment of aboveground biomass estimation

[0112] The estimation model constructed at the scale of quadrat and aerial photographs can estimate aboveground biomass relatively well. P <0.001, Figure 10 At the sample plot scale, the training set (R0) of the random forest model... 2 =0.85, RMSE=27.97 g·m -2 ) and test set (R 2 =0.72, RMSE=38.77g·m -2 The accuracy is relatively high. Figure 11 (Figure (a) in the image). Compared to the quadrat scale, the training set (R) of the random forest model at the aerial photograph scale. 2 =0.93, RMSE=13.5 g·m -2 ) and test set (R 2 =0.78, RMSE=25.41g·m -2 It has higher accuracy and smaller error. Figure 8 (See Figure (b)). Secondly, the number of samples used in model training (N=283) and testing (N=122) at the aerial photo scale was also higher than the number of samples used in model training (N=646) and testing (N=278) at the quadrat scale.

[0113] 3. Spatial distribution of aboveground biomass

[0114] The prediction results show that the estimated range of aboveground biomass in the Qilian Mountains grasslands is 0. 250g·m -2 The average value is approximately 110 g·m -2 The estimated results are close to the measured range. Spatially, areas with higher aboveground biomass in the Qilian Mountains grasslands are mainly distributed in the southeast, while areas with lower aboveground biomass are mainly concentrated in the northwest, showing an overall decreasing trend from southeast to northwest. Figure 10 ).

[0115] The importance of vegetation height in estimating aboveground biomass:

[0116] The results of the present application show that the aboveground biomass at the sample plot and aerial photo scale increases with the increase of vegetation coverage and shows an early saturation phenomenon. Combined with the vegetation height parameter (vegetation coverage x vegetation height), the aboveground biomass can be objectively estimated. The vegetation height reflects the spatial heterogeneity of different grassland types in the vertical direction. Compared with the model constructed by using vegetation coverage as a single indicator, the addition of vegetation height can effectively eliminate the early saturation phenomenon of estimating aboveground biomass and improve the estimation accuracy. Among the extraction methods of vegetation height (including three-dimensional reconstruction technology, canopy height model, etc.), existing studies have shown that unmanned aerial vehicle aerial photography has high accuracy in identifying forest aboveground height or high vegetation area. However, for alpine meadow with small plants, the prediction effect is poor (R 2 = 0.33). The results of the present application show that the vegetation height of most alpine meadow is lower than 7cm (FIG. c in the Figure 1 ), and even if the aerial photo has a centimeter-level resolution, it will also produce a large error in obtaining the vegetation height. Therefore, the present application measures the vegetation height to ensure the objectivity of the vegetation height at the sample plot scale. However, compared with the measured biomass, the results of the present application show that the estimated aboveground biomass at the aerial photo scale has a maximum underestimation phenomenon (FIG. b in the Figure 1 ). This is mainly because on the one hand, the number of samples with aboveground biomass exceeding 250g·m -2 -2 is small; on the other hand, the average method is used to unify the vegetation height of the aerial photo scale in the same flight route, which leads to the saturation phenomenon in the area with the same vegetation coverage but high vegetation height, resulting in the underestimation of the high aboveground biomass.

[0117] The estimation accuracy of the aboveground biomass shows that:

[0118] Satellite remote sensing inversion can usually obtain periodic remote sensing images at fixed time intervals, but it has a lag for real-time monitoring of ground biomass. In contrast, unmanned aerial vehicles have higher flexibility, are less affected by seasons and climate, and can quickly and efficiently obtain large-scale, high-precision vegetation information, which has more advantages in estimating alpine grassland aboveground biomass. The results of the present application show that the model constructed at the plot and aerial photo scales can accurately estimate the aboveground biomass, indicating that unmanned aerial vehicles can fill the spatial scale mismatch between ground samples and satellite pixels and further improve the estimation accuracy of aboveground biomass. Notably, the estimation model of aboveground biomass constructed at the aerial photo scale has higher precision and sample size than the estimation model of aboveground biomass constructed at the plot scale, and the increase in scale and the increase in the number of samples involved in modeling significantly improves the estimation accuracy. Although the regression model currently established based on unmanned aerial vehicle photography at a small scale can better estimate aboveground biomass, due to the limitations of scale and spatial heterogeneity, the applicability and scalability of these models are limited in different grassland types and lack of verification. The aboveground biomass estimation model constructed in the present application couples the aboveground biomass of multiple grassland types, effectively avoiding the influence of spatial heterogeneity. At the same time, the model considers the horizontal and vertical information of vegetation, and has high estimation accuracy at the plot and aerial photo scales, and the estimated value and spatial distribution of aboveground biomass of large-scale grasslands are similar to the measured results.

[0119] The present application is based on ground investigation, unmanned aerial vehicle photography and remote sensing image to construct the aboveground biomass upscaling estimation model from plot to aerial photo to pixel, and the results show that the estimation accuracy of the model is high. Vegetation height is an important parameter for estimating aboveground biomass, and considering vegetation height can solve the problem of over-saturation of aboveground biomass when fitting single vegetation coverage. The amplitude of aboveground biomass estimated based on the model is 0 250g· m -2 , the amplitude of measured aboveground biomass is 0 275g·m -2 , and the estimated value is similar to the measured result. Higher grassland aboveground biomass is mainly distributed in the southeast region of Qilian Mountain, and gradually decreases towards the northwest. The results of the present application show that the aboveground biomass upscaling estimation model constructed based on unmanned aerial vehicle photography can be used for accurate inversion of large-scale alpine grassland aboveground biomass, and provides a theoretical basis and practical reference for grassland resource management and adaptive decision-making.

[0120] Among them, another embodiment of the present application provides a system for obtaining alpine grassland biomass, which comprises:

[0121] The data acquisition module is configured to acquire remote sensing images of the alpine grassland region, acquire a plurality of sampling areas from the remote sensing images, acquire a plurality of aerial photograph areas from each sampling area, and acquire a plurality of sample areas from each aerial photograph area; and acquire the vegetation height, the vegetation coverage, and the aboveground biomass of each sample area, the vegetation coverage of each aerial photograph area, the vegetation coverage of each sampling area, and the spectral information of each pixel point in the remote sensing images.

[0122] The prediction model construction module is configured to obtain a prediction relationship between the aboveground biomass and the vegetation height of each aerial photograph area according to the vegetation height, the vegetation coverage, and the aboveground biomass of all sample areas in each aerial photograph area.

[0123] The prediction model construction module is configured to obtain a prediction relationship between the aboveground biomass and the vegetation height of each sampling area according to the aboveground biomass prediction value, the vegetation height, and the vegetation coverage of all aerial photograph areas in each sampling area.

[0124] The prediction model construction module is configured to obtain a prediction model of the aboveground biomass of the corresponding area of each pixel point in each remote sensing image according to the spectral information of each pixel point in each remote sensing image and the aboveground biomass prediction value of the corresponding sampling area.

[0125] The prediction model using module is configured to input the vegetation coverage and the vegetation height of each aerial photograph area into the prediction relationship between the aboveground biomass and the vegetation height of the aerial photograph area, and obtain the aboveground biomass prediction value of each aerial photograph area.

[0126] The prediction model using module is configured to input the vegetation coverage and the vegetation height of each sampling area into the prediction relationship between the aboveground biomass and the vegetation height of the sampling area, and obtain the aboveground biomass prediction value of the sampling area.

[0127] The prediction model using module is configured to input the spectral information of each pixel point in each remote sensing image into the prediction model of the aboveground biomass of the corresponding area of the pixel point in the remote sensing image, and obtain the aboveground biomass prediction value of the corresponding area of each pixel point in the remote sensing image.

[0128] Those skilled in the art will understand that embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0129] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 one or more flowcharts and / or blocks

[0130] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart block or blocks. Figure 1 one or more flowcharts and / or blocks Figure 1 one or more flowcharts and / or blocks

[0131] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. ​ one or more flowcharts and / or blocks ​ one or more flowcharts and / or blocks

[0132] It should be noted that the above detailed description is merely illustrative of the present application and is not intended to limit the present application in any way. Thus, although the present application has been described in considerable detail with reference to certain specific embodiments thereof, other versions will become apparent to those skilled in the art to which the present application pertains upon examination of the drawings and detailed description. It is therefore desired that the scope of the present application be defined by the appended claims rather than by the above detailed description.

Claims

1. A method for obtaining biomass from alpine grasslands, characterized in that, include: Remote sensing images of alpine grassland regions are acquired. Several sampling areas are obtained from the remote sensing images. Several aerial photography areas are obtained from each sampling area. Several sample areas are obtained from each aerial photography area. Vegetation height, vegetation cover and aboveground biomass of each sample area, vegetation cover of each aerial photography area, vegetation cover of each sampling area, and spectral information of each pixel in the remote sensing image are obtained. Based on the vegetation height, vegetation cover, and aboveground biomass of all sample areas within each aerial photography area, a prediction formula for aboveground biomass and vegetation height for each aerial photography area is obtained; based on the predicted aboveground biomass, vegetation height, and vegetation cover of all aerial photography areas within each sampling area, a prediction formula for aboveground biomass and vegetation height for each sampling area is obtained; based on the spectral information of each pixel in each remote sensing image and the predicted aboveground biomass value of the corresponding sampling area, a prediction model for the aboveground biomass of the area corresponding to each pixel in each remote sensing image is obtained. The vegetation cover and vegetation height of each aerial photography area are input into the prediction formula of the aboveground biomass of the aerial photography area to obtain the predicted value of the aboveground biomass of each aerial photography area; the vegetation cover and vegetation height of each sampling area are input into the prediction formula of the aboveground biomass of the sampling area to obtain the predicted value of the aboveground biomass of the sampling area. The spectral information of each pixel in each remote sensing image is input into the prediction model of the aboveground biomass of the corresponding area in the remote sensing image to obtain the predicted value of the aboveground biomass of the corresponding area in the remote sensing image. The specific steps for obtaining the predictive formula for aboveground biomass and vegetation height for each aerial photography area based on vegetation height, vegetation cover, and aboveground biomass of all sample areas within each aerial photography area are as follows: With the first A two-dimensional sample space is constructed by using the product of vegetation height and vegetation cover in the sample area within the aerial photography area as the horizontal axis and the aboveground biomass of the sample area as the vertical axis. With the first The first aerial photography area The product of vegetation height and vegetation cover in each sample region is used as the product of the vegetation height and vegetation cover in the two-dimensional sample space. The x-coordinate of the nth data point; with the nth The first aerial photography area The aboveground biomass of the sample area is used as the first sample area in the two-dimensional sample space. The ordinate of the nth data point is obtained; the nth data point in the two-dimensional sample space is obtained. One data point; The first All sample areas within a single aerial photography area are mapped into a two-dimensional sample space to obtain several data points within the two-dimensional sample space. The random forest algorithm is used to fit the data points in the two-dimensional sample space. The resulting fitting relationship is denoted as the i-th Predictive formula for aboveground biomass in an aerial photography area; The first The mean vegetation height of all sample areas within the aerial photography area is denoted as the i-th. The vegetation height of the aerial photography area; The specific steps for obtaining the predictive relationship between aboveground biomass and vegetation height for each sampling area are as follows: With the first A two-dimensional coordinate system is constructed with the product of vegetation height and vegetation cover in the aerial photography area within each sampling area as the horizontal axis and the predicted aboveground biomass of the aerial photography area as the vertical axis. With the first Within the sampling area, the first The product of vegetation height and vegetation cover in each aerial photography area is used as the product of the vegetation height and vegetation cover in the two-dimensional coordinate system. The x-coordinate of the nth sample point; with the nth... Within the sampling area, the first The above-ground biomass of the aerial photograph area is used as the first... in a two-dimensional coordinate system. The ordinate of the nth sample point is obtained; the nth sample point in the two-dimensional coordinate system is obtained. One sample; The first All aerial photography areas within a sampling area are mapped into a two-dimensional coordinate system, resulting in several data points within the two-dimensional coordinate system. The random forest algorithm is used to fit the data points in the two-dimensional coordinate system. The resulting fitting relationship is denoted as the i-th... Predictive formula for aboveground biomass in each sampling area; The first The average vegetation height of all aerial photographed areas within a sampling area is denoted as the nth. Vegetation height in each sampling area.

2. The method for obtaining biomass in alpine grasslands according to claim 1, characterized in that, The specific steps for obtaining several sampling regions from the remote sensing image are as follows: Remote sensing images of the study area were acquired using remote sensing satellites. A region corresponding to several pixels was randomly selected within a remote sensing image.

3. The method for obtaining biomass in alpine grasslands according to claim 1, characterized in that, The specific steps for obtaining the predicted ground biomass value for each aerial photography area are as follows: The first The product of vegetation height and vegetation cover in each aerial photography area is used as the first... The input to the predictive formula for the aboveground biomass of each aerial photography area will be denoted as the output of the i-th... Predicted aboveground biomass values ​​for the aerial photography area.

4. The method for obtaining biomass in alpine grasslands according to claim 1, characterized in that, The specific steps for obtaining the predicted aboveground biomass value for each sampling area are as follows: The first The average vegetation height of all aerial photographed areas within a sampling area is denoted as the nth. Vegetation height in each sampling area; The first The product of vegetation height and vegetation cover in each sampling area is used as the product of the first sampling area and the second sampling area. The input to the predictive formula for aboveground biomass in each sampling area will be denoted as the output of the formula. Predicted aboveground biomass values ​​for each sampling area.

5. The method for obtaining biomass in alpine grasslands according to claim 1, characterized in that, The specific steps for obtaining the prediction model of the aboveground biomass corresponding to each pixel in each remote sensing image based on the spectral information of each pixel in each remote sensing image and the predicted value of aboveground biomass in the corresponding sampling area are as follows: The first The first remote sensing image corresponds to the alpine grassland region. The predicted aboveground biomass value for the first sampling area, and the predicted value for the second sampling area. The first remote sensing image corresponds to the alpine grassland region. The first sampling region corresponds to the first The spectral information of pixels in the remote sensing image is used as the first... Zhang remote sensing image The data values ​​of the two dimensions of the target point are used to obtain the first... All target points within the remote sensing image; Using the random forest algorithm, based on the... Using the two-dimensional data values ​​of all target points within a remote sensing image, a model is trained, denoted as the... A predictive model for aboveground biomass in the region corresponding to a pixel in a remote sensing image.

6. A system for obtaining biomass from alpine grasslands, characterized in that, include: The data acquisition module is used to acquire remote sensing images of alpine grassland areas, obtain several sampling areas from the remote sensing images, obtain several aerial photography areas from each sampling area, and obtain several sample areas from each aerial photography area; acquire vegetation height, vegetation cover and aboveground biomass of each sample area, vegetation cover of each aerial photography area, vegetation cover of each sampling area, and spectral information of each pixel in the remote sensing image. The prediction model building module is used to obtain the prediction formula for the aboveground biomass and vegetation height of each aerial photography area based on the vegetation height, vegetation cover and aboveground biomass of all sample areas in each aerial photography area; Based on the predicted values ​​of aboveground biomass, vegetation height, and vegetation cover of all aerial photography areas within each sampling area, the predicted relationship of aboveground biomass and vegetation height for each sampling area is obtained. Based on the spectral information of each pixel in each remote sensing image and the predicted value of aboveground biomass in the corresponding sampling area, a prediction model of aboveground biomass for the area corresponding to each pixel in the remote sensing image is obtained. The prediction model uses a module to input the vegetation cover and vegetation height of each aerial photography area into the prediction formula of the aboveground biomass of the aerial photography area, and obtain the predicted value of the aboveground biomass of each aerial photography area. The vegetation cover and vegetation height of each sampling area are input into the prediction formula of the aboveground biomass of the sampling area to obtain the predicted value of the aboveground biomass of the sampling area. The spectral information of each pixel in each remote sensing image is input into the prediction model of the aboveground biomass of the corresponding area in the remote sensing image to obtain the predicted value of the aboveground biomass of the corresponding area in the remote sensing image. The specific steps for obtaining the predictive formula for aboveground biomass and vegetation height for each aerial photography area based on vegetation height, vegetation cover, and aboveground biomass of all sample areas within each aerial photography area are as follows: With the first A two-dimensional sample space is constructed by using the product of vegetation height and vegetation cover in the sample area within the aerial photography area as the horizontal axis and the aboveground biomass of the sample area as the vertical axis. With the first The first aerial photography area The product of vegetation height and vegetation cover in each sample region is used as the product of the vegetation height and vegetation cover in the two-dimensional sample space. The x-coordinate of the nth data point; with the nth The first aerial photography area The aboveground biomass of the sample area is used as the first sample area in the two-dimensional sample space. The ordinate of the nth data point is obtained; the nth data point in the two-dimensional sample space is obtained. One data point; The first All sample areas within a single aerial photography area are mapped into a two-dimensional sample space to obtain several data points within the two-dimensional sample space. The random forest algorithm is used to fit the data points in the two-dimensional sample space. The resulting fitting relationship is denoted as the i-th Predictive formula for aboveground biomass in an aerial photography area; The first The mean vegetation height of all sample areas within the aerial photography area is denoted as the i-th. The vegetation height of the aerial photography area; The specific steps for obtaining the predictive relationship between aboveground biomass and vegetation height for each sampling area are as follows: With the first A two-dimensional coordinate system is constructed with the product of vegetation height and vegetation cover in the aerial photography area within each sampling area as the horizontal axis and the predicted aboveground biomass of the aerial photography area as the vertical axis. With the first Within the sampling area, the first The product of vegetation height and vegetation cover in each aerial photography area is used as the product of the vegetation height and vegetation cover in the two-dimensional coordinate system. The x-coordinate of the nth sample point; with the nth... Within the sampling area, the first The above-ground biomass of the aerial photograph area is used as the first... in a two-dimensional coordinate system. The ordinate of the nth sample point is obtained; the nth sample point in the two-dimensional coordinate system is obtained. One sample; The first All aerial photography areas within a sampling area are mapped into a two-dimensional coordinate system, resulting in several data points within the two-dimensional coordinate system. The random forest algorithm is used to fit the data points in the two-dimensional coordinate system. The resulting fitting relationship is denoted as the i-th... Predictive formula for aboveground biomass in each sampling area; The first The average vegetation height of all aerial photographed areas within a sampling area is denoted as the nth. Vegetation height in each sampling area.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is loaded by the processor, it is able to execute the steps of the method for obtaining biomass in alpine grasslands as described in any one of claims 1 to 5.

Images