A method for establishing a mixed ground object spectral library

Abstract

The present invention discloses a method for establishing a mixed ground object spectral library, comprising: obtaining a hyperspectral remote sensing image of a detected area, dividing the hyperspectral remote sensing image into a plurality of target areas, each of the target areas including a plurality of survey pixels, and different target areas having different vegetation coverage; setting a sample area for each of the target areas, obtaining a sample spectral image containing the sample area; classifying the ground objects in each of the sample spectral images, and obtaining the area proportion of each ground object in each of the sample spectral images; extracting the spectral curve of each of the sample areas on the hyperspectral remote sensing image, and constructing a spectral library of grassland mixed ground objects according to the area proportion of each ground object in the corresponding sample area. The present invention can establish a spectral library of mixed ground objects in areas with strong surface heterogeneity.

Description

Technical Field

[0001] The present application relates to the technical field of hyperspectral remote sensing applications, and in particular, to a method for establishing a mixed ground object spectral library. Background Art

[0002] Building a spectral library is an important basic work. The spectral library is the basis for remote sensing interpretation, object recognition, and accurate image classification. Therefore, it is of great significance to the development of remote sensing science. At present, the mainstream spectral libraries include the object spectral library built by the United States Geological Survey (USGS), the spectral library of 160 minerals built by the Jet Propulsion Laboratory (JPL), the ASTER spectral library established by the California Institute of Technology, and the object-oriented spectral database established by the Institute of Remote Sensing, Chinese Academy of Sciences.

[0003] However, most of the existing spectral libraries are constructed using ground hyperspectral technology for single objects, while the actual ground conditions corresponding to the pixels on remote sensing images are mostly the coexistence of several objects, and there are rarely only one object. Therefore, it is difficult to achieve sub-pixel classification using existing spectral libraries, and it is impossible to point out the specific proportion of each object in the pixel. Each pixel can only be simply classified into one object category, which is obviously not accurate enough. Summary of the invention

[0004] The embodiment of the present application aims to provide a method for establishing a mixed ground object spectral library to solve the technical problem of low accuracy of the spectral library established by the existing method.

[0005] To solve the above problems, some embodiments of the present application provide a method for establishing a mixed ground object spectral library, comprising the following steps:

[0006] Acquire a hyperspectral remote sensing image of the detected area;

[0007] Segmenting the hyperspectral remote sensing image into a plurality of target areas, each of the target areas including a plurality of survey pixels, and different target areas have different vegetation coverages;

[0008] A sample area is set for each target area, and a sample spectral image including the sample area is obtained;

[0009] Classifying the ground objects in each of the sample plot spectral images to obtain the area proportion of each ground object in each of the sample plot spectral images;

[0010] The spectral curve of each sample area on the hyperspectral remote sensing image is extracted, and a spectral library of grassland mixed objects is constructed according to the area proportion of each object in the corresponding sample area.

[0011] In the method for establishing a mixed ground object spectral library described in some embodiments of the present application, the step of obtaining a hyperspectral remote sensing image of the detected area includes:

[0012] Acquire an aerial hyperspectral remote sensing image of the detected area obtained by aerial photography, and preprocess the image to obtain a hyperspectral remote sensing image of the detected area; the preprocessing includes:

[0013] The aerial hyperspectral remote sensing image is radiometrically calibrated, and the pixel brightness values ​​of the original data of the aerial hyperspectral remote sensing image are converted into radiometric brightness values ​​to obtain a radiometrically calibrated image. The radiometric calibration formula is as follows:

[0014]

[0015] Where L is the total radiance of all objects in the aerial hyperspectral remote sensing image; calibration_ gain is the radiation correction factor; image_ ND The gray value of the ground object recorded by the pixel brightness value of the original data of the aerial hyperspectral remote sensing image; dark_ ND is the dark current brightness value of the corresponding pixel; Mean is the clustering function; FSS_ ND is the brightness value of scattered light in the aerial hyperspectral remote sensing image, and integration_time is the integration time;

[0016] Perform atmospheric correction on the radiometrically calibrated image, convert the radiometric brightness value into the true reflectivity of the surface to obtain the atmospherically corrected image;

[0017] Performing geometric correction on the atmospherically corrected image to obtain the true geographic coordinates of each point in the atmospherically corrected image;

[0018] The hyperspectral remote sensing image is obtained by splicing all images that have undergone radiation calibration, atmospheric correction and geometric correction.

[0019] The method for establishing a mixed ground object spectral library described in some embodiments of the present application includes the steps of segmenting the hyperspectral remote sensing image into a plurality of target areas, each of the target areas includes a plurality of survey pixels, and the vegetation coverage of different target areas is different.

[0020] Obtaining a normalized difference vegetation index value of the hyperspectral remote sensing image;

[0021] The hyperspectral remote sensing image is segmented into a plurality of regions of different vegetation coverage types according to the normalized difference vegetation index value, and each region of the vegetation coverage type includes a plurality of survey pixels.

[0022] In some embodiments of the present application, the method for establishing a mixed ground object spectral library includes setting a sample area for each target area, and obtaining a sample spectral image containing the sample area, including:

[0023] The center coordinates of the investigated pixels in the target area are used as the coordinates of the sample area;

[0024] Taking the coordinates of the sample area as the center, a sample device with a height-adjustable hyperspectral camera is arranged in the detected area, and the lens of the hyperspectral camera is vertically downward to ensure that the field of view of the lens is completely filled by the sample area;

[0025] The image captured by the hyperspectral camera is obtained as the sample plot spectral image.

[0026] In some embodiments of the present application, the method for establishing a mixed ground object spectral library includes the steps of segmenting the hyperspectral remote sensing image into a plurality of regions of different vegetation coverage types according to the normalized difference vegetation index value:

[0027]

[0028] VFC is the vegetation coverage, NDVI is the normalized difference vegetation index value; NDVI max and NDVI min These are the maximum and minimum NDVI values ​​in the region, respectively.

[0029] In the method for establishing a mixed ground object spectral library described in some embodiments of the present application, a quadrat device with a height-adjustable hyperspectral camera is arranged in the detected area with the quadrat area coordinates as the center, and the lens of the hyperspectral camera is vertically downward to ensure that the field of view of the lens is completely filled with the quadrat area:

[0030] The quadrat device comprises a quadrangular pyramid frame structure, the bottom and top of the quadrangular pyramid frame structure are both square structures, and the top and the bottom are connected via a tripod;

[0031] A support plate is installed on the top, and the hyperspectral camera is arranged on the support plate; the area covered by the bottom of the sample plot device is used as a sample plot area.

[0032] In the method for establishing a mixed ground feature spectral library described in some embodiments of the present application, the tripod and the bottom in the quadrangular pyramid frame structure are both formed by telescopic rods; a horizontal bubble is provided on the support plate, and the horizontal bubble is used to ensure that the top surface is in a horizontal state.

[0033] The method for establishing a mixed ground feature spectral library in some embodiments of the present application, the bottom side length of the four-sided pyramid frame structure Where x is the image spatial resolution.

[0034] The method for establishing a mixed ground object spectral library described in some embodiments of the present application further includes, before the step of acquiring a hyperspectral remote sensing image of the detected area:

[0035] Control points are arranged in the detected area, and the coordinates of the control points are recorded. The coordinates of the control points are used for later coordinate point correction of the hyperspectral remote sensing image.

[0036] In the method for establishing a mixed ground object spectral library described in some embodiments of the present application, the spectral curve of each sample area on the hyperspectral remote sensing image is extracted, and the spectral library of grassland mixed ground objects is constructed according to the area proportion of each ground object in the corresponding sample area:

[0037] If there are n sample areas with the same ground object mixing conditions, the average spectral curve of the spectral curves of the n sample areas is taken as the final spectral curve of the n sample areas.

[0038] Compared with the prior art, the above technical solution provided by the present application has at least the following beneficial effects: the present invention obtains a hyperspectral remote sensing image of the detected area, divides the hyperspectral remote sensing image into multiple target areas, each of which includes multiple survey pixels, and the vegetation coverage of different target areas is different; a sample area is set for each target area, and a sample spectral image containing the sample area is obtained; the objects in each sample spectral image are classified to obtain the area proportion of each object in each sample spectral image; the spectral curve of each sample area on the hyperspectral remote sensing image is extracted, and a spectral library of mixed objects in grassland is constructed according to the area proportion of each object in the corresponding sample area. The present invention can establish a spectral library of mixed objects in areas with strong surface heterogeneity. Moreover, the present invention effectively reduces the workload because the detected area is selected in advance on the hyperspectral image. By obtaining the sample spectral image containing the sample area, the accurate proportion of each object in the sample area is obtained, thereby obtaining the area proportion of each object in the survey pixel, and the mixed object spectral library is established based on this, thereby realizing the sub-pixel classification of aerial images. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present application.

[0040] Figure 1 A flowchart of a method for establishing a mixed ground object spectral library according to an embodiment of the present application;

[0041] Figure 2 This is a schematic diagram of a quadrangular pyramid frame structure according to an embodiment of the present application. DETAILED DESCRIPTION

[0042] In the description of the present application, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the simplified description of the present application, and do not indicate or imply that the device or component referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0043] In the description of this application, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0044] The scheme in the following embodiments of the present application takes Xilin Gol Grassland as the detected area object. Xilin Gol Grassland is located in Inner Mongolia Autonomous Region, China, with geographical coordinates of 115°50'39"-116°15'19"E, 43°56'7"-44°4'20"N, and an average altitude of 970-1202m. The main species of the grassland include Stipa krylovii, Cleistogenes scabra, Leymus chinensis, Icegrass, Salsola, etc., with rich species, and various species grow in a mixed and staggered manner. Therefore, in response to the current situation of mixed growth of grassland species, this embodiment provides a method for establishing a mixed ground feature spectral library in the following embodiments, establishes a grassland mixed ground feature spectral library based on aerial hyperspectral images, and uses aerial hyperspectral images to establish a mixed ground feature spectral library to achieve refined classification of grassland species.

[0045] This embodiment provides a method for establishing a mixed ground object spectral library. Figure 1 As shown, the following steps are included:

[0046] S101: Obtain the hyperspectral remote sensing image of the detected area; aerial hyperspectral remote sensing images of the detected area can be obtained by aerial photography, and then the hyperspectral remote sensing images can be obtained after preprocessing. According to the situation of the detected area, the route is planned to be in the east-west direction, and the drone is equipped with the Headwall hyperspectral sensor for aerial photography. The band range of the Headwall hyperspectral sensor is 402.82-960.20nm, a total of 188 bands, 2.96nm spectral resolution, and the spatial resolution is set to 1.24m. The flight time is from 10 am to 3 pm, with a total of 8 flight bands. The images obtained from each flight band can be spliced ​​to obtain the image of the detected area.

[0047] S102: Segment the hyperspectral remote sensing image into a plurality of target areas, each of which includes a plurality of survey pixels, and different target areas have different vegetation coverage. Specifically, the target area can be obtained by the following steps: obtaining a normalized difference vegetation index value of the hyperspectral remote sensing image; segmenting the hyperspectral remote sensing image into a plurality of areas of different vegetation coverage types according to the normalized difference vegetation index value, each area of ​​the vegetation coverage type includes a plurality of survey pixels.

[0048] S103: Setting a sample plot area for each of the target areas, and obtaining a sample plot spectral image including the sample plot area; wherein, preferably, the center coordinates of the investigated pixels in the target area are used as the sample plot area coordinates; with the sample plot area coordinates as the center, a sample plot device equipped with a height-adjustable hyperspectral camera is arranged in the detected area, and the lens of the hyperspectral camera is pointed vertically downward to ensure that the field of view of the lens is completely filled by the sample plot area; and the image captured by the hyperspectral camera is obtained as the sample plot spectral image.

[0049] S104: Classifying the ground objects in each of the sample plot spectral images to obtain the area proportion of each ground object in each of the sample plot spectral images.

[0050] S105: extracting the spectral curve of each sample area on the hyperspectral remote sensing image, and constructing a spectral library of grassland mixed objects according to the area proportion of each object in the corresponding sample area.

[0051] The above scheme effectively reduces the workload because the detected area is selected in advance on the hyperspectral image. By obtaining the sample plot spectral image that includes the sample plot area, the accurate proportion of each type of land object in the sample plot area is obtained, thereby obtaining the area proportion of each type of land object in the survey pixel. Based on this, a mixed land object spectral library is established, thereby realizing sub-pixel classification of aerial images.

[0052] In the above scheme, step S101 includes:

[0053] Acquire an aerial hyperspectral remote sensing image of the detected area obtained by aerial photography, and preprocess the image to obtain a hyperspectral remote sensing image of the detected area; the preprocessing includes:

[0054] S1.1: Perform radiometric calibration on the aerial hyperspectral remote sensing image, convert the pixel brightness values ​​of the original data of the aerial hyperspectral remote sensing image into radiometric brightness values ​​to obtain a radiometric calibrated image. The radiometric calibration formula is as follows:

[0055]

[0056] Where L is the total radiance of all ground objects in the aerial hyperspectral remote sensing image, that is, the total radiance of the ground target measured by the Headwall hyperspectral sensor, in units of uw / (cm 2 ·str·nm);calibration_ gain is the radiation correction factor; image_ ND The gray value of the ground object recorded by the pixel brightness value of the original data of the aerial hyperspectral remote sensing image; dark_ ND is the dark current brightness value of the corresponding pixel; Mean is the clustering function; FSS_ ND is the brightness value of scattered light in aerial hyperspectral remote sensing images, and integration_time is the integration time. When users need to calculate the spectral reflectance or spectral radiation brightness of objects, or need to compare images acquired at different times and by different sensors, they must convert the brightness grayscale value of the image into absolute radiation brightness. This process is called radiation calibration.

[0057] S1.2: Perform atmospheric correction on the radiometrically calibrated image, convert the radiometric brightness value into the true reflectivity of the surface to obtain the atmospherically corrected image; in this step, the radiometrically calibrated image is loaded into the image processing software (ENVI), and the image processing software (ENVI) runs the atmospheric correction function.

[0058] S1.3: geometrically correct the image after atmospheric correction to obtain the real geographic coordinates of each point in the image after atmospheric correction; in this step, the image after atmospheric correction is loaded into the image processing software (ENVI), and the image processing software (ENVI) completes the geometric correction of the image according to the coordinates of the control points arranged during the aerial photography process to obtain the real geographic coordinates of the image. Preferably, before step S101, the following steps are included: control points are arranged in the detected area, and the coordinates of the control points are recorded, and the coordinates of the control points are used for later coordinate point correction of the hyperspectral remote sensing image.

[0059] S1.4: The hyperspectral remote sensing image is obtained by stitching all images that have undergone radiometric calibration, atmospheric correction, and geometric correction. The remote sensing images of the eight flight strips that have undergone radiometric calibration, atmospheric correction, and geometric correction are stitched together to obtain a complete aerial hyperspectral remote sensing image of the study area.

[0060] In some embodiments, the specific method of calculating the normalized difference vegetation index NDVI of the image and estimating the vegetation coverage VFC using the NDVI is as follows:

[0061] S2.1: Obtain the reflectance of the near-infrared band and the red light band, and use the band calculation function of ENVI software to calculate NDVI. The calculation formula is as follows:

[0062]

[0063] Among them, NIR is the reflectance of the near-infrared band of the image, and R is the reflectance of the red band of the image. Check the calculation results, which should be between -1 and 1. Pixels less than -1 are assigned a value of -1, and pixels greater than 1 are assigned a value of 1.

[0064] S2.2: Use the band calculation function of ENVI software to estimate vegetation coverage VFC. The calculation formula is as follows:

[0065]

[0066] Among them, NDVI max and NDVI min To obtain the maximum and minimum values ​​within the 95% confidence range, NDVImax is 0.987 and NDVImin is -0.870. The confidence value is mainly determined according to the actual situation of the image. Specifically, for Xilin Gol grassland, according to the calculation results of the above formula, there are 10 vegetation coverage types (VFC 1 、VFC 2 、VFC 3 、…VFC 10 ).

[0067] S2.3: For each vegetation cover type (VFC 1 、VFC 2 、VFC 3 、…VFC 10 ), 8 survey pixels were evenly selected on the remote sensing image as the sample area for laying out the sample plots, and the coordinates of the survey pixel centers were recorded as the coordinates of the sample area, resulting in a total of 80 sample areas.

[0068] In some embodiments, the sample device 100 is as follows Figure 2As shown, it includes a quadrangular pyramid frame structure, the quadrangular pyramid frame structure includes a top 101 and a bottom 102, both of which are square structures, and the top 101 and the bottom 102 are connected via a tripod 103; the top 101 is installed with a support plate 104, and the hyperspectral camera is arranged on the support plate 104; the area covered by the bottom 102 of the sample plot device is used as a sample plot area. With the coordinates of the sample plot area as the center, a square sample plot with a height-adjustable hyperspectral camera is arranged on the ground according to the spatial resolution of the image, and the lens of the hyperspectral camera is vertically downward to ensure that the field of view in the lens is completely filled with the sample plot. The hyperspectral camera in this embodiment is preferably manufactured by SPECIM of Finland, with a spectral range of 400-1000nm and 204 bands.

[0069] Preferably, the tripod 103 and the bottom 102 in the quadrangular pyramid frame structure are both formed by telescopic rods; the telescopic rods are provided with a telescopic portion 106, which can adjust the length. A horizontal bubble 105 is provided on the support plate 104, and the horizontal bubble 105 is used to ensure that the top surface 101 is in a horizontal state. With the coordinates of the sample plot area as the center, a square sample plot with a height-adjustable hyperspectral camera is arranged on the ground according to the spatial resolution of the image, and the lens of the hyperspectral camera is vertically downward to ensure that the field of view in the lens is completely filled with the sample plot. The hyperspectral camera in this embodiment is preferably manufactured by SPECIM of Finland, with a spectral range of 400-1000nm and 204 bands. The horizontal bubble 105 is always centered to ensure that the hyperspectral camera remains in a horizontal state, and the bottom side length of the sample plot device is adjusted so that the area covered by the bottom of each sample plot device 100 represents a sample plot. In specific implementation, the bottom side length of the quadrangular pyramid frame structure Wherein, x is the image spatial resolution. In this embodiment, the bottom side length of the sample plot device is adjusted to 1.80m according to the image spatial resolution of 1.24m, that is, the side length of the sample plot is 1.80m, and the lens of the hyperspectral camera is vertically downward to check whether the field of view in the lens is completely filled with the sample plot. If not, continue to adjust the height of the tripod 103 until the field of view in the lens is completely filled with the sample plot. Of course, the sample plot device can also be selected from other sizes as needed. It is preferred to ensure that the bottom side length of the sample plot device is 1.80m. The hyperspectral photos of each sample plot were obtained by a hyperspectral camera, and the hyperspectral photos of each sample plot were classified using ENVI software. According to the classification results, the area ratio of each feature in each sample plot was obtained. The spectral curve of each sample plot corresponding to the survey pixel on the hyperspectral remote sensing image was extracted in the image processing software ENVI, and the spectral library of grassland mixed features was constructed according to the area ratio of each feature in the sample plot.

[0070] In the above scheme provided by the present application, an aerial hyperspectral remote sensing image is obtained and preprocessed; the image normalized difference vegetation index (NDVI) value is calculated, and the vegetation coverage is estimated using NDVI; for each type of vegetation coverage, m survey pixels are uniformly selected on the remote sensing image, and the center coordinates of the pixel are recorded as the sample area for laying out the sample; with the above coordinates as the center, a square sample with a hyperspectral camera is laid out on site; a hyperspectral photo of the sample is obtained, and the objects in the photo are classified to obtain the area ratio of each object in each sample; the spectral curve of the surveyed pixel on the remote sensing image is extracted, and a spectral library of mixed objects is constructed according to the area ratio of each object in the sample. The present invention achieves sub-pixel classification of aerial hyperspectral images by establishing a spectral library of mixed objects, thereby obtaining the area ratio of each object in the pixel.

[0071] In some embodiments, in the above step S105, if there are n sample areas with the same ground object mixing conditions, the average spectral curve of the spectral curves of the n sample areas is taken as the final spectral curve of the n sample areas to make the constructed spectral library more accurate.

[0072] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for establishing a mixed ground feature spectral library. It is characterized in that The steps include: Acquire a hyperspectral remote sensing image of the detected area; including: planning an east-west route according to the situation of the detected area, using a drone equipped with a hyperspectral sensor for aerial photography, the band range of the hyperspectral sensor is 402.82-960.20nm, a total of 188 bands, 2.96nm spectral resolution, and setting the spatial resolution to 1.24m; a total of 8 flight bands, each flight band After the image is spliced, the image of the detected area can be obtained; after preprocessing, the hyperspectral remote sensing image can be obtained; the hyperspectral remote sensing image is divided into multiple target areas, each of the target areas includes multiple survey pixels, and the vegetation coverage of different target areas is different; A sample area is set for each target area, and a sample spectral image containing the sample area is obtained; including: taking the central coordinates of the investigated pixels in the target area as the sample area coordinates; taking the sample area coordinates as the center, a sample device with a height-adjustable hyperspectral camera is arranged in the detected area, and the lens of the hyperspectral camera is vertically downward to ensure that the field of view of the lens is completely filled by the sample area; the image captured by the hyperspectral camera is obtained as the sample spectral image; the sample device includes a quadrangular pyramid frame structure, the bottom and top of the quadrangular pyramid frame structure are both square structures, and the top and the bottom are connected via a tripod; a support plate is installed on the top, and the hyperspectral camera It is arranged on the support plate; the area covered by the bottom of the sample plot device is used as a sample plot area; the telescopic rod has a telescopic part, if the field of view in the lens of the hyperspectral camera is not completely filled by the sample plot area, the height of the tripod is adjusted until the field of view of the lens is completely filled by the sample plot area; the tripod and the bottom of the quadrangular pyramid frame structure are both formed by telescopic rods; a horizontal bubble is arranged on the support plate, and the horizontal bubble is used to ensure that the top surface is in a horizontal state; the bottom side length L in the quadrangular pyramid frame structure is ≥√2·x, where x is the image spatial resolution; the spectral range of the spectral camera is 400-1000nm, and the number of bands is 204; Classifying the ground objects in each of the sample plot spectral images to obtain the area proportion of each ground object in each of the sample plot spectral images; The spectral curve of each sample area on the hyperspectral remote sensing image is extracted, and a spectral library of mixed grassland objects is constructed according to the area proportion of each object in the corresponding sample area; if there are n sample areas with the same object mixing situation, the average spectral curve of the spectral curves of the n sample areas is taken as the final spectral curve of the n sample areas.

2. The method for establishing a mixed ground object spectral library according to claim 1, It is characterized in that The steps of obtaining a hyperspectral remote sensing image of the detected area include: Acquire an aerial hyperspectral remote sensing image of the detected area obtained by aerial photography, and preprocess the image to obtain a hyperspectral remote sensing image of the detected area; the preprocessing includes: The aerial hyperspectral remote sensing image is radiometrically calibrated, and the pixel brightness values ​​of the original data of the aerial hyperspectral remote sensing image are converted into radiometric brightness values ​​to obtain a radiometrically calibrated image. The radiometric calibration formula is as follows: Where L is the total radiance of all objects in the aerial hyperspectral remote sensing image; calibration _gain is the radiation correction factor; image -ND The gray value of the ground object recorded by the pixel brightness value of the original data of the aerial hyperspectral remote sensing image; dark_ ND is the dark current brightness value of the corresponding pixel; Mean is the clustering function; FSS_ ND is the brightness value of scattered light in the aerial hyperspectral remote sensing image, and integration_time is the integration time; Perform atmospheric correction on the radiometrically calibrated image, convert the radiometric brightness value into the true reflectivity of the surface to obtain the atmospherically corrected image; Performing geometric correction on the atmospherically corrected image to obtain the true geographic coordinates of each point in the atmospherically corrected image; The hyperspectral remote sensing image is obtained by splicing all images that have undergone radiation calibration, atmospheric correction and geometric correction.

3. The method for establishing a mixed ground object spectral library according to claim 1, It is characterized in that The step of segmenting the hyperspectral remote sensing image into a plurality of target areas, each of the target areas including a plurality of survey pixels, and different target areas having different vegetation coverages comprises: Obtaining a normalized difference vegetation index value of the hyperspectral remote sensing image; The hyperspectral remote sensing image is segmented into a plurality of regions of different vegetation coverage types according to the normalized difference vegetation index value, and each region of the vegetation coverage type includes a plurality of survey pixels.

4. The method for establishing a mixed ground object spectral library according to claim 3, It is characterized in that The step of segmenting the hyperspectral remote sensing image into a plurality of regions of different vegetation coverage types according to the normalized difference vegetation index value comprises: VFC is the vegetation coverage, NDVI is the normalized difference vegetation index value; NDVI max and NDVI min These are the maximum and minimum NDVI values ​​in the region, respectively.

5. The method for establishing a mixed ground object spectral library according to any one of claims 1 to 4, It is characterized in that Before the step of acquiring the hyperspectral remote sensing image of the detected area, the following steps are also included: Control points are arranged in the detected area, and the coordinates of the control points are recorded. The coordinates of the control points are used for later coordinate point correction of the hyperspectral remote sensing image.

Images