A method and system for constructing a pixel spectral library for natural resource surface coverage

Through pattern screening and spatial superposition analysis, combined with hyperspectral satellite remote sensing image preprocessing, the reliability and homogeneity of cell spectral acquisition are solved, and a high-quality cell spectral library is formed to support the fine surface coverage of natural resources. Identify.

CN118585589BActive Publication Date: 2024-12-06CHONGQING GEOMATICS & REMOTE SENSING CENT
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410623241.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-05-20
Publication Date
2024-12-06
Estimated Expiration
2044-05-20

AI Technical Summary

Technical Problem

The reliability, homogeneity and accuracy of cell spectral acquisition results are challenged in the prior art, especially in the radiation correction and atmospheric correction of hyperspectral remote sensing images, resulting in heterogeneous and heterogeneous phenomena. , affects the fine identification of natural resource surface coverage.

Method used

By obtaining the initial spatial vector identification data, performing spectral screening and spatial overlay analysis, calculating the centroid point as the initial cell spectrum acquisition point, and performing buffering and minimum envelope rectangle judgment, selecting homogeneous cell acquisition point, combining with the hyperspectral Preprocessing of satellite remote sensing images forms a cell spectrum library.

Benefits of technology

A cell acquisition point and a massive spectral library were obtained with good homogeneity, which reduced the impact of hyperspectral image radiation correction and atmospheric correction errors, and provided finely identified data support for natural resource surface coverage.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118585589B_ABST
    Figure CN118585589B_ABST
Patent Text Reader

Abstract

The present invention provides a method for constructing a pixel spectral library for natural resource surface coverage. The present invention is based on the third national land survey or geographical conditions or other natural resource surface coverage spatial vector data, and adopts methods such as spot screening and pixel collection point purification to obtain pixel collection points with good homogeneity, and form a pixel spectral library containing attribute information of different dimensions such as time, space, spectrum, and identification. The method is not affected by the superposition of errors in hyperspectral image radiation correction and atmospheric correction, and provides a massive pixel spectral library for fine identification of natural resource surface coverage.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of pixel spectral library construction, and in particular to a pixel spectral library construction method and system for natural resource surface coverage. Background Art

[0002] Hyperspectral remote sensing has the ability to identify subtle differences in ground objects due to its unique spectral characteristics. In recent years, it has been increasingly used in the fields of ground object classification and image interpretation. Spectral libraries are key markers for accurately interpreting hyperspectral remote sensing image information. They can provide effective sample data for fine identification of natural resource surface coverage, mainly including ground spectral libraries and pixel spectral libraries.

[0003] Ground spectra are mainly constructed by using ground spectrometers to collect surface and local spectra of objects in the field, focusing on the spectral characteristics of the parts of the objects, which can well reflect the fine spectral characteristics of the objects. However, the collection of ground object spectra is time-consuming and labor-intensive, and the amount of collected spectral data is limited, which cannot meet the needs of modern machine learning algorithms for large samples. In recent years, with the explosive growth of hyperspectral remote sensing images, pixel spectra have become a better way to build a massive spectral library.

[0004] Traditional pixel spectrum collection requires further screening of spectrum collection points through methods such as hyperspectral image calculation and setting empirical thresholds. This collection point selection method has three problems: First, based on hyperspectral remote sensing images, uniform ground objects are selected through image reflectance calculation and other methods, which will lead to the accumulation of errors in the hyperspectral radiation correction and atmospheric correction processes, directly reducing the reliability of the spectrum calculation results. Second, since the reflectance of hyperspectral remote sensing images is greatly affected by factors such as the environment and terrain, hyperspectral remote sensing images have a common phenomenon of different spectra for the same object and different objects for the same spectrum, making it difficult to extract homogeneous land types. Third, the ground object identification information of the pixel spectrum is missing, which seriously restricts the promotion and application of the spectral library in the fine identification of natural resource surface coverage.

[0005] As a result, the reliability, homogeneity and accuracy of current pixel spectral acquisition results are challenged. Summary of the invention

[0006] In view of the problems existing in the prior art, the present invention provides a method and system for constructing a pixel spectral library for natural resource surface coverage, so as to solve the technical problem that the reliability and accuracy of the results of current pixel spectral acquisition in the prior art are challenged.

[0007] The present invention provides a method for constructing a pixel spectral library for natural resource surface coverage, comprising:

[0008] S1. Acquire initial spatial vector identification data, perform spot screening on the initial spatial vector identification data to obtain first screening spots and second screening spots respectively, perform spatial superposition analysis on the first screening spots and the second screening spots, and select the spatial intersection of the first screening spots and the second screening spots as the third screening spots;

[0009] S2, calculating the centroid of each minimum spot unit in the third screening spot, and using it as the initial pixel spectrum collection point, calculating the position coordinates of the initial pixel spectrum collection point, and assigning the identification information of the third screening spot to the initial pixel spectrum collection point;

[0010] S3, taking the initial pixel spectrum acquisition point as the center, buffering outward point by point to obtain a buffered spatial geometric figure, calculating the minimum envelope rectangle of the spatial geometric figure, and judging spot by spot whether the third screening spot completely contains the minimum envelope rectangle, if so, retaining the third screening spot, otherwise eliminating the third screening spot to form a fourth screening spot;

[0011] S4, performing spatial superposition analysis on the initial pixel spectrum collection point and the fourth screening pattern, and selecting the pixel collection point overlapping the fourth screening pattern as the final pixel collection point;

[0012] S5, obtaining a hyperspectral satellite remote sensing image of the same period as the final pixel acquisition point, and preprocessing the image to obtain surface reflectance data of the hyperspectral satellite remote sensing image;

[0013] S6. Extracting spectral information of the surface reflection data according to the final pixel acquisition point to form a pixel spectral library.

[0014] Optionally, the first screening pattern includes:

[0015] The first screening spots are selected according to the spot area larger than 3*3 pixel size, which is expressed as:

[0016] S min =3*3*P 2

[0017] S>S min

[0018] Among them, S is the image patch area, and P is the pixel size of any hyperspectral satellite remote sensing image that is intended to collect pixel spectra.

[0019] Optionally, the second screening pattern includes:

[0020] The second screening spot is screened out using the ratio of the spot perimeter to the area, expressed as:

[0021] T=L / A

[0022] Wherein, L represents the perimeter of the pattern, A represents the area of ​​the pattern, and T represents the ratio of the perimeter to the area of ​​the pattern. When T is less than 0.05, the pattern is selected as the second screening pattern.

[0023] Optionally, calculating the position coordinates of the initial pixel spectrum acquisition point includes:

[0024] The calculation formula of the position coordinates of the initial pixel spectrum acquisition point is expressed as:

[0025]

[0026]

[0027]

[0028] Wherein, i=1, 2, ... m, m represents the total number of spots in the third screening spot, j=1, 2, ... n, n represents the number of nodes of spot i in the third screening spot, X ij and Y ij They represent the geometric space coordinate value of the jth node of the third screening spot i, W ij represents the weight of the jth node of the third screening patch i, W i It is the sum of the weights of the i-th spot in the third screening spot.

[0029] Optionally, the calculating the minimum envelope rectangle of the spatial geometric figure includes:

[0030] (X max , Y max )(X max , Y min )(X min , Y max )(X min , Y min ) as the corner point, construct the minimum enveloping rectangle; where X and Y are the spatial coordinates of all the endpoints of the geometric figure, respectively. max and X min are the maximum X-axis coordinate and the minimum X-axis coordinate of the geometric figure under the spatial coordinate, Y max and Y min They are the maximum Y-axis coordinate and the minimum Y-axis coordinate of the geometric figure in the space coordinate.

[0031] Optionally, the preprocessing to obtain the surface reflectance data of the hyperspectral satellite remote sensing image includes:

[0032] The preprocessing at least includes performing radiation correction, atmospheric correction and geometric correction on the hyperspectral satellite remote sensing image.

[0033] Optionally, the final pixel acquisition point includes:

[0034] The attributes of the final pixel acquisition point include at least time information, identification information and location information;

[0035] The time information is the acquisition time of the initial space vector identification data, the identification information is the land use classification type of the final pixel collection point, and the position information is the space coordinate of the final pixel collection point in the plane coordinate system.

[0036] Optionally, forming a pixel spectrum library comprises:

[0037] The attributes of the pixel spectrum library include at least time information, identification information, location information and spectrum information.

[0038] The present invention also provides a system for constructing a pixel spectral library for natural resource surface coverage, comprising:

[0039] A data acquisition and preprocessing module is used to obtain initial spatial vector identification data, perform spot screening on the initial spatial vector identification data to obtain a first screening spot and a second screening spot respectively, perform spatial superposition analysis on the first screening spot and the second screening spot, select the spatial intersection of the first screening spot and the second screening spot as a third screening spot; calculate the centroid point of each minimum spot unit in the third screening spot, and use it as an initial pixel spectrum collection point, calculate the position coordinates of the initial pixel spectrum collection point, and assign the identification information of the third screening spot to the initial pixel spectrum collection point;

[0040] A spatial analysis module, which is used to buffer outward point by point with the initial pixel spectrum acquisition point as the center to obtain a buffered spatial geometric figure, calculate the minimum envelope rectangle of the spatial geometric figure, and determine spot by spot whether the third screening spot completely contains the minimum envelope rectangle, if it does, retain the third screening spot, otherwise remove the third screening spot to form a fourth screening spot; and is also used to perform spatial superposition analysis on the initial pixel spectrum acquisition point and the fourth screening spot, and select the pixel acquisition point overlapping the fourth screening spot as the final pixel acquisition point;

[0041] A remote sensing image reflectance calculation module is used to obtain the hyperspectral satellite remote sensing image corresponding to the time node in the final pixel acquisition point, and preprocess it to obtain the surface reflectance data of the hyperspectral satellite remote sensing image;

[0042] The pixel spectrum library building module is used to extract the spectrum information of the surface reflection data according to the attributes of the final pixel acquisition point to form a pixel spectrum library.

[0043] Compared with the existing spectral library construction technology, the present invention has the following beneficial effects:

[0044] Based on the third national land survey or geographical conditions or other natural resource surface coverage spatial vector data, the method of spot screening and pixel collection point purification is used to obtain pixel collection points with good homogeneity and form a pixel spectral library containing attribute information of different dimensions. This method is not affected by the superposition of errors in hyperspectral image radiation correction and atmospheric correction, and provides a massive pixel spectral library for fine identification of natural resource surface coverage. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0047] Figure 1 It is a schematic diagram of the method flow of the present invention;

[0048] Figure 2 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other implementation cases obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application. The functional units with the same labels in the examples of the present invention have the same and similar structures and functions.

[0050] See also Figure 1 The present invention provides a method for constructing a pixel spectral library for natural resource surface coverage, comprising:

[0051] S1. Acquire initial spatial vector identification data, perform spot screening on the initial spatial vector identification data to obtain first screening spots and second screening spots respectively, perform spatial superposition analysis on the first screening spots and the second screening spots, and select the spatial intersection of the first screening spots and the second screening spots as the third screening spots;

[0052] S2, calculating the centroid of each minimum spot unit in the third screening spot, and using it as the initial pixel spectrum collection point, calculating the position coordinates of the initial pixel spectrum collection point, and assigning the identification information of the third screening spot to the initial pixel spectrum collection point;

[0053] S3, taking the initial pixel spectrum acquisition point as the center, buffering outward point by point to obtain a buffered spatial geometric figure, calculating the minimum envelope rectangle of the spatial geometric figure, and judging spot by spot whether the third screening spot completely contains the minimum envelope rectangle, if so, retaining the third screening spot, otherwise eliminating the third screening spot to form a fourth screening spot;

[0054] S4, performing spatial superposition analysis on the initial pixel spectrum collection point and the fourth screening pattern, and selecting the pixel collection point overlapping the fourth screening pattern as the final pixel collection point;

[0055] S5, obtaining a hyperspectral satellite remote sensing image of the same period as the final pixel acquisition point, and preprocessing the image to obtain surface reflectance data of the hyperspectral satellite remote sensing image;

[0056] S6. Extracting spectral information of the surface reflection data according to the final pixel acquisition to form a pixel spectral library.

[0057] In this embodiment, S1, obtain initial spatial vector identification data, perform spot screening on the initial spatial vector identification data to obtain first screening spots and second screening spots respectively, perform spatial superposition analysis on the first screening spots and the second screening spots, and select the spatial intersection of the first screening spots and the second screening spots as the third screening spots.

[0058] First, obtain the spatial vector data of all elements and land types of the third national land survey or geographical conditions or other natural resource surface coverage interpretation. The third national land survey or geographical conditions or other natural resource surface coverage data is obtained based on high spatial resolution remote sensing images with high spatial accuracy. This type of data is spatially divided and typed according to land use classification. A single spot has good homogeneity and is used as the initial spatial vector identification data for the selection of pixel spectral collection points.

[0059] Secondly, the initial spatial vector identification data is screened from the two aspects of patch area and graphic morphology to obtain patches that meet the requirements. According to the patch area screening, the area S of each patch screened is required to be greater than 3*3 pixels. The pixel size is based on the acquired hyperspectral satellite image, and the patches that meet the patch area screening requirements are used as the first screening patches.

[0060] S min =3*3*P 2

[0061] S>S min

[0062] P is the pixel size of any hyperspectral satellite remote sensing image that is intended to collect pixel spectra.

[0063] Taking the Zhuhai-1 hyperspectral satellite with a spatial resolution of 10 meters as an example, P = 10 meters, S min =900 square meters, and the spots with a spot area S greater than 900 square meters are used as the first screening spots. The screening method can use ArcMap10.6 software to calculate the plane area of ​​the spots, and use the selection tool to screen out the vector spots with an area greater than 900 square meters.

[0064] By using graphic morphology screening and the ratio of the perimeter to the area of ​​the patch, the narrow and long patches or irregular patches in the original data are eliminated. The ratio T of the perimeter to the area of ​​the patch can effectively reflect the geometric shape of the graphic and effectively eliminate the narrow and long patches or irregular patches that cannot extract effective pixel collection points in the initial spatial vector identification data. The initial spatial vector identification data calculates the perimeter L and the area A of each patch, and further calculates the perimeter to area ratio T of each patch.

[0065] T=L / A

[0066] Generally, when T≥0.05, it is a narrow or irregular pattern, and the pattern with a perimeter-to-area ratio T less than 0.05 is selected as the second screening pattern.

[0067] The first screening spots and the second screening spots are spatially superimposed and analyzed, and the spatial intersection part is selected. This part of the spots not only meets the requirements of the spot area, but also meets the requirements of the spot shape, and is used as the third screening spot.

[0068] S2. Calculate the centroid point of each minimum spot unit in the third screening spot and use it as the initial pixel spectrum collection point, calculate the position coordinates of the initial pixel spectrum collection point, and assign the identification information of the third screening spot to the initial pixel spectrum collection point.

[0069] Obtain the initial pixel acquisition point. Based on the third screening spot, calculate the centroid of each minimum spot unit in the third screening spot as the initial pixel spectrum acquisition point, calculate and obtain the position coordinates of the initial pixel spectrum acquisition point, and assign the identification information of the third screening spot to the initial pixel acquisition point.

[0070] The third screening spot data in the i-th spot centroid point G i (X i , Y i ) coordinate calculation method:

[0071]

[0072]

[0073]

[0074] Wherein, i=1, 2, ... m, m represents the total number of spots in the third screening spot, j=1, 2, ... n, n represents the number of nodes of spot i in the third screening spot, X ij and Y ij They represent the geometric space coordinate value of the jth node of the third screening spot i, W ij represents the weight of the jth node of the third screening patch i, W i It is the sum of the weights of the i-th spot in the third screening spot.

[0075] S3. Taking the initial pixel spectrum acquisition point as the center, buffer outward point by point to obtain a buffered spatial geometric figure, calculate the minimum envelope rectangle of the spatial geometric figure, and determine spot by spot whether the third screening spot completely contains the minimum envelope rectangle. If it does, retain the third screening spot; otherwise, remove the third screening spot to form a fourth screening spot.

[0076] Hyperspectral remote sensing images are raster data, and the third screening spots are vector data. In order to overcome the acquisition errors caused by the different data types of the two, the initial pixel acquisition points need to be purified.

[0077] Taking the initial pixel acquisition point as the center, buffer outward by 1.5*P point by point, and after obtaining the buffered spatial geometry, further calculate the minimum envelope rectangle of the geometry. P is the pixel size of any hyperspectral satellite remote sensing image to be collected.

[0078] The minimum envelope matrix is ​​calculated as follows:

[0079] 1. Calculate the spatial coordinates (X, Y) of all endpoints of the geometric figure

[0080] 2. Get the maximum X-axis coordinate X of the geometric figure under the spatial coordinate max , the minimum X-axis coordinate X min ;

[0081] 3. Get the maximum Y-axis coordinate Y of the geometric figure under the spatial coordinate max , minimum Y-axis coordinate Y min ;

[0082] 4. (X max , Y max )(X max , Y min )(X min , Ymax )(X min , Y min ) as the corner points, construct a rectangle to form the minimum enveloping rectangle.

[0083] Taking the 10-meter spatial resolution Zhuhai-1 hyperspectral satellite data as an example, P = 10 meters, the buffer range is 1.5*P, that is, 15 meters, and the minimum envelope rectangle is further calculated. It is determined by each spot whether the third screening spot completely contains the minimum envelope rectangle. If it does, the third screening spot is retained, otherwise the third screening spot is removed to form the fourth screening spot.

[0084] S4, performing spatial superposition analysis on the initial pixel spectrum collection point and the fourth screening pattern, and selecting the pixel collection point overlapping the fourth screening pattern as the final pixel collection point.

[0085] Calculate the initial pixel collection point and the fourth screening spot for spatial overlay analysis, and select the pixel collection point that intersects with the fourth screening spot as the final pixel collection point. The attribute information of the final pixel collection point is shown in the following table.

[0086]

[0087]

[0088] Among them, the X and Y coordinates are the location information of the final pixel collection point in the CSGS2000 plane coordinate system, the identification information is the land use classification type of the pixel collection point, the time information is the acquisition time of the initial spatial vector identification data, and the source data records the type of surface coverage spatial vector data referenced. Recording multiple attributes allows the pixel collection point to be distinguished in time, space and attribute dimensions.

[0089] S5. Acquire a hyperspectral satellite remote sensing image of the same period as the final pixel acquisition point, and preprocess the image to obtain surface reflectance data of the hyperspectral satellite remote sensing image.

[0090] The hyperspectral satellite remote sensing image of the corresponding time node of the final pixel acquisition point is obtained, and the hyperspectral satellite remote sensing image is preprocessed by radiation correction, atmospheric correction, geometric correction, etc. to obtain the surface reflectance data of the remote sensing image. The data adopts the conventional CSGS2000 plane coordinate system to maintain spatial consistency with the pixel acquisition point.

[0091] S6. Extracting spectral information of the surface reflection data according to the final pixel acquisition point to form a pixel spectral library.

[0092] According to the time information, location information and identification information of the final pixel collection point, the spectral information of the surface reflectance data of the hyperspectral satellite image, that is, the reflectance data of each band of the hyperspectral satellite remote sensing image, is extracted and recorded to form a pixel spectral library for the surface coverage of natural resources.

[0093] Taking the Zhuhai No. 1 OHS-3A CMOS2 hyperspectral satellite remote sensing data as an example, the band range is 400-1000nm, the number of bands is 32, and the central bands of the hyperspectral satellite remote sensing images are:

[0094] 443nm,466nm,490nm,500nm,510nm,531nm,550nm,560nm,580nm,596nm,620nm,640nm,665nm,670nm,686nm,700 nm,709nm,730nm,746nm,760nm,776nm,780nm,806nm,820nm,833nm,850nm,880nm,896nm,910nm,926nm,940nm.

[0095] The pixel spectral library attribute fields are as follows:

[0096]

[0097]

[0098] See also Figure 2 The present invention also provides a system for constructing a pixel spectrum library for natural resource surface coverage, comprising:

[0099] A data acquisition and preprocessing module is used to obtain initial spatial vector identification data, perform spot screening on the initial spatial vector identification data to obtain a first screening spot and a second screening spot respectively, perform spatial superposition analysis on the first screening spot and the second screening spot, select the spatial intersection of the first screening spot and the second screening spot as a third screening spot; calculate the centroid point of each minimum spot unit in the third screening spot, and use it as an initial pixel spectrum collection point, calculate the position coordinates of the initial pixel spectrum collection point, and assign the identification information of the third screening spot to the initial pixel spectrum collection point;

[0100] A spatial analysis module, which is used to buffer outward point by point with the initial pixel spectrum acquisition point as the center to obtain a buffered spatial geometric figure, calculate the minimum envelope rectangle of the spatial geometric figure, and determine spot by spot whether the third screening spot completely contains the minimum envelope rectangle, if it does, retain the third screening spot, otherwise remove the third screening spot to form a fourth screening spot; and is also used to perform spatial superposition analysis on the initial pixel spectrum acquisition point and the fourth screening spot, and select the pixel acquisition point overlapping the fourth screening spot as the final pixel acquisition point;

[0101] A remote sensing image reflectance calculation module is used to obtain the hyperspectral satellite remote sensing image corresponding to the time node in the final pixel acquisition point, and preprocess it to obtain the surface reflectance data of the hyperspectral satellite remote sensing image;

[0102] The pixel spectrum library building module is used to extract the spectrum information of the surface reflection data according to the attributes of the final pixel acquisition point to form a pixel spectrum library.

[0103] The present invention is based on the third national land survey or geographical conditions or other natural resource surface coverage spatial vector data, and adopts methods such as spot screening and pixel collection point purification to obtain pixel collection points with good homogeneity and form a pixel spectrum library containing attribute information of different dimensions. This method is not affected by the superposition of errors in hyperspectral image radiation correction and atmospheric correction, and provides a massive pixel spectrum library for fine identification of natural resource surface coverage.

[0104] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0105] The foregoing is merely a specific embodiment of the present invention, which enables those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for constructing a pixel spectral library for natural resource surface coverage. It is characterized in that include: S1. Using a single spot as the initial spatial vector identification data selected from the pixel spectrum collection point, the initial spatial vector identification data is screened from the two aspects of spot area and graphic morphology to obtain spots that meet the requirements, specifically: screening according to spot area, requiring each screened spot with an area S greater than a preset threshold to be a first screening spot, using graphic morphology screening, using the ratio method of spot perimeter and area, eliminating narrow and long spots or irregular spots in the original data to obtain a second screening spot, and performing spatial superposition analysis on the first screening spot and the second screening spot, and selecting the spatial intersection of the first screening spot and the second screening spot as the third screening spot; S2, calculating the centroid of each minimum spot unit in the third screening spot, and using it as the initial pixel spectrum collection point, calculating the position coordinates of the initial pixel spectrum collection point, and assigning the identification information of the third screening spot to the initial pixel spectrum collection point; S3, taking the initial pixel spectrum acquisition point as the center, buffering outward point by point to obtain a buffered spatial geometric figure, calculating the minimum envelope rectangle of the spatial geometric figure, and judging spot by spot whether the third screening spot completely contains the minimum envelope rectangle, if so, retaining the third screening spot, otherwise eliminating the third screening spot to form a fourth screening spot; S4, performing spatial superposition analysis on the initial pixel spectrum collection point and the fourth screening pattern, selecting the pixel collection point overlapping the fourth screening pattern as the final pixel collection point, wherein the attributes of the final pixel collection point include time information, identification information and location information; The time information is the acquisition time of the initial space vector identification data, the identification information is the land use classification type of the final pixel collection point, and the position information is the spatial coordinates of the final pixel collection point in the plane coordinate system; S5, obtaining a hyperspectral satellite remote sensing image of the same period as the final pixel acquisition point, and preprocessing the image to obtain surface reflectance data of the hyperspectral satellite remote sensing image; S6. Extracting spectral information of the surface reflection data according to the final pixel acquisition point to form a pixel spectral library.

2. The method for constructing a pixel spectral library for natural resource surface coverage according to claim 1, It is characterized in that The first screening pattern comprises: The first screening spots are selected according to the spot area larger than 3*3 pixel size, which is expressed as: , , Among them, S is the image patch area, and P is the pixel size of any hyperspectral satellite remote sensing image that is intended to collect pixel spectra.

3. The method for constructing a pixel spectral library for natural resource surface coverage according to claim 1, It is characterized in that The second screening pattern comprises: The second screening spot is screened out using the ratio of the spot perimeter to the area, expressed as: , Wherein, L represents the perimeter of the pattern, A represents the area of ​​the pattern, and T represents the ratio of the perimeter to the area of ​​the pattern. When T is less than 0.05, the pattern is selected as the second screening pattern.

4. The method for constructing a pixel spectral library for natural resource surface coverage according to claim 1, It is characterized in that The step of calculating the position coordinates of the initial pixel spectrum acquisition point comprises: The calculation formula of the position coordinates of the initial pixel spectrum acquisition point is expressed as: , , , in, , , m represents the total number of spots in the third screening spot, , n , n Indicates the third screening spot i The number of nodes, and Respectively represent the spots in the third screening spot i No. j The geometric space coordinates of the nodes, Indicates the third screening spot i No. j The weight of the node, For the third screening pattern i The sum of the weights of all nodes in a patch.

5. The method for constructing a pixel spectral library for natural resource surface coverage according to claim 1, It is characterized in that The step of calculating the minimum envelope rectangle of the spatial geometric figure comprises: by( , )( , ) ( , )( , ) as the corner point, construct the minimum enveloping rectangle; where X and Y are the spatial coordinates of all the endpoints of the geometric figure, and are the maximum X-axis coordinate and the minimum X-axis coordinate of the geometric figure in the space coordinate, and They are the maximum Y-axis coordinate and the minimum Y-axis coordinate of the geometric figure in the space coordinate.

6. The method for constructing a pixel spectral library for natural resource surface coverage according to claim 1, It is characterized in that The method of preprocessing the hyperspectral satellite remote sensing image to obtain the surface reflectance data includes: The preprocessing includes performing radiation correction, atmospheric correction and geometric correction on the hyperspectral satellite remote sensing image.

7. The method for constructing a pixel spectral library for natural resource surface coverage according to claim 1, It is characterized in that The forming of the pixel spectrum library comprises: The attributes of the pixel spectrum library include time information, identification information, location information and spectrum information.

8. A pixel spectral library construction system for natural resource surface coverage, It is characterized in that include: The data acquisition and preprocessing module is used to select a single spot as the initial spatial vector identification data of the pixel spectrum acquisition point, and screen the initial spatial vector identification data from the two aspects of spot area and graphic morphology to obtain spots that meet the requirements, specifically: screening according to the spot area, requiring each screened spot with an area S greater than a preset threshold to be used as the first screening spot, using graphic morphology screening, using the ratio method of the spot perimeter and area, eliminating narrow and long spots or irregular spots in the original data to obtain the second screening spot, and performing spatial superposition analysis on the first screening spot and the second screening spot, and selecting the spatial intersection of the first screening spot and the second screening spot as the third screening spot; Calculate the centroid of each minimum spot unit in the third screening spot and use it as the initial pixel spectrum collection point, calculate the position coordinates of the initial pixel spectrum collection point, and assign the identification information of the third screening spot to the initial pixel spectrum collection point; A spatial analysis module, which is used to buffer outward point by point with the initial pixel spectrum acquisition point as the center to obtain a buffered spatial geometric figure, calculate the minimum envelope rectangle of the spatial geometric figure, and determine spot by spot whether the third screening spot completely contains the minimum envelope rectangle, if it does, retain the third screening spot, otherwise remove the third screening spot to form a fourth screening spot; and is also used to perform spatial superposition analysis on the initial pixel spectrum acquisition point and the fourth screening spot, select the pixel acquisition point overlapping the fourth screening spot as the final pixel acquisition point, and the attributes of the final pixel acquisition point include time information, identification information and location information; The time information is the acquisition time of the initial space vector identification data, the identification information is the land use classification type of the final pixel collection point, and the position information is the spatial coordinates of the final pixel collection point in the plane coordinate system; A remote sensing image reflectance calculation module is used to obtain the hyperspectral satellite remote sensing image corresponding to the time node in the final pixel acquisition point, and preprocess it to obtain the surface reflectance data of the hyperspectral satellite remote sensing image; The pixel spectrum library building module is used to extract the spectrum information of the surface reflection data according to the attributes of the final pixel acquisition point to form a pixel spectrum library.