A method and device for quantitative inversion of surface parameters using hyperspectral remote sensing

Abstract

The present invention provides a method and device for quantitative inversion of surface parameters of a hyperspectral remote sensing, wherein the method comprises: obtaining hyperspectral remote sensing data of a to-be-measured object; performing feature enhancement on the hyperspectral remote sensing data to obtain a feature enhancement result; performing weighted processing on the hyperspectral remote sensing data based on the feature enhancement result to obtain weighted hyperspectral remote sensing data; inputting the weighted hyperspectral remote sensing data into a quantitative inversion model of surface parameters to obtain an inversion value of the surface parameters of the to-be-measured object output by the quantitative inversion model of surface parameters; the quantitative inversion model of surface parameters is used to invert the surface parameter value of the object. The present invention can fully utilize hyperspectral information according to the response of the surface parameters in different spectral bands by performing feature enhancement and weighted processing on the hyperspectral remote sensing data of the to-be-measured object, thereby improving the accuracy of the surface parameters of the to-be-measured object inverted by the quantitative inversion model of surface parameters.

Description

Technical Field

[0001] The present invention relates to the technical field of hyperspectral remote sensing, and in particular to a method, device, electronic equipment and storage medium for quantitative inversion of hyperspectral remote sensing surface parameters. Background Art

[0002] The emergence of remote sensing technology has provided new technologies and means for the detection and perception of objects. Multispectral remote sensing is a major remote sensing technology that has been widely used in the monitoring of vegetation, soil, water bodies and other objects, mainly focusing on the monitoring of the area, distribution, and changes of vegetation, soil, water bodies and other objects. The number of bands of multispectral remote sensing is generally less than 10, and the spectral resolution is about 100nm-200nm. Due to the small number of bands and low spectral resolution, it is difficult for multispectral remote sensing to detect the fine spectral characteristics of objects, which limits the application of multispectral remote sensing. The development of remote sensing technology has enhanced the detection capability of remote sensing technology and expanded the breadth and depth of remote sensing applications. Hyperspectral remote sensing technology is a combination of remote sensing technology and spectral technology. While performing remote sensing imaging of objects, it can obtain the spectral curve of objects. The spectral range of hyperspectral remote sensing covers visible light, near infrared and short-wave infrared, and can usually obtain the reflectance spectrum of objects in 400-2400nm, and the spectral resolution is generally better than 10nm. The emergence of hyperspectral remote sensing technology enables remote sensing to obtain fine reflectance spectra, more precisely characterize the spectral characteristics of objects, and comprehensively improve the level of remote sensing monitoring.

[0003] At present, hyperspectral remote sensing can quantitatively invert and monitor vegetation biochemical parameters, soil physical and chemical parameters, water quality parameters, etc. in the monitoring of vegetation, soil, water bodies and other land objects. Among the surface parameters such as vegetation, soil, and water bodies, except for parameters such as vegetation chlorophyll and soil organic matter, which have clear and detectable spectral characteristics, the spectral responses of the remaining parameters are mostly presented in a broad and mixed form, making it difficult to extract specific characteristic bands for quantitative inversion of surface parameters. Usually, a certain spectral interval is extracted for quantitative inversion of surface parameters, resulting in spectral information loss, which in turn leads to reduced accuracy of the inversion results. Summary of the invention

[0004] In view of the problem in the prior art that the accuracy of inversion results is reduced, the present invention provides a method and device for quantitative inversion of hyperspectral remote sensing surface parameters, and establishes a surface parameter inversion model according to the spectral response of the surface parameters in different bands to fully utilize the spectral information and improve the accuracy of quantitative inversion of surface parameters.

[0005] The present invention provides a method for quantitative inversion of surface parameters using hyperspectral remote sensing, comprising:

[0006] Obtaining hyperspectral remote sensing data of the object to be measured;

[0007] Performing feature enhancement on the hyperspectral remote sensing data to obtain a feature enhancement result;

[0008] Based on the feature enhancement result, weighted processing is performed on the hyperspectral remote sensing data to obtain weighted hyperspectral remote sensing data;

[0009] The weighted hyperspectral remote sensing data is input into a quantitative inversion model for surface parameters to obtain inversion values ​​of surface parameters of the object to be measured output by the quantitative inversion model for surface parameters; the quantitative inversion model for surface parameters is used to invert the surface parameter values ​​of the object.

[0010] According to a method for quantitative inversion of hyperspectral remote sensing surface parameters provided by the present invention, the hyperspectral remote sensing data includes a reflectance spectrum curve obtained by a ground object spectrometer;

[0011] Before performing feature enhancement on the hyperspectral remote sensing data to obtain a feature enhancement result, the method further includes:

[0012] The reflectivity spectrum curve is subjected to noise removal and band extraction processing to obtain a preprocessed reflectivity spectrum curve.

[0013] According to a method for quantitative inversion of hyperspectral remote sensing surface parameters provided by the present invention, the hyperspectral remote sensing data includes a hyperspectral remote sensing image;

[0014] Before performing feature enhancement on the hyperspectral remote sensing data to obtain a feature enhancement result, the method further includes:

[0015] Preprocessing the hyperspectral remote sensing image to obtain a preprocessed remote sensing image;

[0016] Extracting a reflectance spectrum from the preprocessed remote sensing image to obtain a preprocessed reflectance spectrum curve;

[0017] The preprocessing of the hyperspectral remote sensing image includes at least one of the following: performing radiation calibration on the hyperspectral remote sensing image, performing atmospheric correction on the hyperspectral remote sensing image, performing geometric precision correction on the hyperspectral remote sensing image, and performing noise removal on the hyperspectral remote sensing image.

[0018] According to a method for quantitative inversion of hyperspectral remote sensing surface parameters provided by the present invention, the feature enhancement of the hyperspectral remote sensing data is performed to obtain a feature enhancement result, including:

[0019] The preprocessed reflectance spectrum curve is subjected to feature enhancement based on a feature enhancement algorithm to obtain a feature enhancement result.

[0020] According to a method for quantitative inversion of hyperspectral remote sensing surface parameters provided by the present invention, the feature enhancement algorithm is used to perform feature enhancement on the preprocessed reflectance spectrum curve to obtain a feature enhancement result, including:

[0021] In the case where the preprocessed reflectance spectrum curve has an identifiable spectrum feature of a ground parameter, extracting spectrum data corresponding to the identifiable spectrum feature from the preprocessed reflectance spectrum curve;

[0022] The spectral data is feature enhanced using a feature enhancement algorithm to obtain a feature enhancement result.

[0023] According to a method for quantitative inversion of hyperspectral remote sensing surface parameters provided by the present invention, the feature enhancement algorithm is used to perform feature enhancement on the preprocessed reflectance spectrum curve to obtain a feature enhancement result, including:

[0024] Based on a continuum removal algorithm or a spectrum derivation algorithm, performing continuum removal or spectrum derivation on the preprocessed reflectance spectrum curve to obtain a continuum removal result or a spectrum derivation result;

[0025] The continuum removal result or the spectrum derivative result is used as a feature enhancement result.

[0026] According to a method for quantitative inversion of hyperspectral remote sensing surface parameters provided by the present invention, the quantitative inversion model of surface parameters is obtained by training based on the following steps:

[0027] Acquire sample data of ground objects; the sample data includes surface parameter samples and hyperspectral remote sensing data samples;

[0028] Performing feature enhancement on the hyperspectral remote sensing data sample to obtain a feature enhancement result sample;

[0029] Based on the feature enhancement result sample, weighted processing is performed on the hyperspectral remote sensing data sample to obtain a weighted hyperspectral remote sensing data sample;

[0030] A surface parameter quantitative inversion model is constructed based on the surface parameter samples and the weighted hyperspectral remote sensing data samples to obtain the surface parameter quantitative inversion model.

[0031] The present invention also provides a hyperspectral remote sensing surface parameter quantitative inversion device, comprising:

[0032] An acquisition unit, used for acquiring hyperspectral remote sensing data of the object to be measured;

[0033] A first obtaining unit is used to perform feature enhancement on the hyperspectral remote sensing data to obtain a feature enhancement result;

[0034] A second obtaining unit is used to perform weighted processing on the hyperspectral remote sensing data based on the feature enhancement result to obtain weighted hyperspectral remote sensing data;

[0035] The third obtaining unit is used to input the weighted hyperspectral remote sensing data into the quantitative inversion model of surface parameters to obtain the inversion value of the surface parameters of the object to be measured output by the quantitative inversion model of surface parameters; the quantitative inversion model of surface parameters is used to invert the surface parameter values ​​of the object.

[0036] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the above-mentioned methods for quantitative inversion of hyperspectral remote sensing surface parameters is implemented.

[0037] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for quantitative inversion of hyperspectral remote sensing surface parameters as described in any one of the above is implemented.

[0038] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for quantitative inversion of hyperspectral remote sensing surface parameters as described above is implemented.

[0039] The present invention provides a method and device for quantitative inversion of hyperspectral remote sensing surface parameters. By performing feature enhancement on the hyperspectral remote sensing data of an object to be measured and performing weighted processing on the hyperspectral remote sensing data based on the feature enhancement result, the hyperspectral information can be fully utilized according to the response of the surface parameters in different spectral bands. The weighted hyperspectral remote sensing data is input into a quantitative inversion model for surface parameters, thereby improving the accuracy of the surface parameters of the object to be measured inverted by the quantitative inversion model for surface parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 This is one of the flow charts of a method for quantitative inversion of surface parameters of hyperspectral remote sensing provided by the present invention;

[0042] Figure 2 This is the second flow chart of a method for quantitative inversion of hyperspectral remote sensing surface parameters provided by the present invention;

[0043] Figure 3 This is the third flow chart of a method for quantitative inversion of hyperspectral remote sensing surface parameters provided by the present invention;

[0044] Figure 4 is a schematic diagram of a reflectivity spectrum curve after pretreatment provided by the present invention;

[0045] Figure 5 This is a fourth flow chart of a method for quantitative inversion of hyperspectral remote sensing surface parameters provided by the present invention;

[0046] Figure 6 This is a schematic diagram of a hyperspectral remote sensing soil reflectance spectrum continuum removal result provided by the present invention;

[0047] Figure 7 It is a schematic diagram of the soil reflectance spectrum continuum removal transformation result provided by the present invention;

[0048] Figure 8 This is the fifth flow chart of a method for quantitative inversion of hyperspectral remote sensing surface parameters provided by the present invention;

[0049] Fig. 9 This is a schematic diagram of a quantitative inversion result of a hyperspectral remote sensing surface parameter provided by the present invention;

[0050] Fig.10 It is a structural schematic diagram of a hyperspectral remote sensing surface parameter quantitative inversion device provided by the present invention;

[0051] Fig.11 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] Combine the following Figure 1-Figure 9 A method for quantitatively inverting surface parameters using hyperspectral remote sensing is described in the present invention.

[0054] Figure 1 This is one of the flow charts of a method for quantitative inversion of hyperspectral remote sensing surface parameters provided by the present invention, such as Figure 1 As shown in FIG. 1 , the method for quantitative inversion of hyperspectral remote sensing surface parameters includes the following steps:

[0055] Step 101: Acquire hyperspectral remote sensing data of the object to be measured.

[0056] The objects to be measured may be vegetation, soil, water and other objects to be measured; in the monitoring of vegetation, soil, water and other objects, hyperspectral remote sensing can monitor the biochemical parameters of vegetation, the physical and chemical parameters of soil, and the water quality parameters.

[0057] Among them, the hyperspectral remote sensing data can be a reflectance spectral curve measured by a ground object spectrometer, an aerial hyperspectral remote sensing image obtained by an aerial hyperspectral instrument, or a satellite hyperspectral remote sensing image obtained by a satellite hyperspectral instrument.

[0058] Among them, the ground feature spectrometer can be an ASD FieldSpec series and PSR series ground feature spectrometer; the aviation hyperspectrometer can be a HyMap aviation hyperspectrometer; the satellite hyperspectrometer can be an EO-1Hyperion, GF-5 AHSI and other space-borne hyperspectrometers.

[0059] Step 102: Perform feature enhancement on the hyperspectral remote sensing data to obtain a feature enhancement result.

[0060] For example, the hyperspectral remote sensing data is feature enhanced according to the spectral response characteristics (spectral characteristics) of the surface parameters; taking the hyperspectral remote sensing data of the object to be measured as a reflectance spectrum curve as an example, when the spectral characteristics of the surface parameters can be identified on the reflectance spectrum curve, the spectral curve with the spectral characteristics is extracted from the reflectance spectrum curve, and the extracted spectral curve is feature enhanced using a feature enhancement algorithm to obtain a feature enhancement result; when the spectral characteristics of the surface parameters cannot (or are not easy to) be identified on the reflectance spectrum curve, the reflectance spectrum curve is feature enhanced using a feature enhancement algorithm to obtain a feature enhancement result.

[0061] Step 103: performing weighted processing on the hyperspectral remote sensing data based on the feature enhancement result to obtain weighted hyperspectral remote sensing data.

[0062] For example, the hyperspectral remote sensing data is multiplied by the corresponding feature enhancement result to implement weighted processing of the hyperspectral remote sensing data and obtain weighted hyperspectral remote sensing data.

[0063] Step 104: input the weighted hyperspectral remote sensing data into a quantitative inversion model for surface parameters to obtain inversion values ​​of surface parameters of the object to be measured output by the quantitative inversion model for surface parameters.

[0064] The quantitative inversion model of surface parameters is obtained by training based on hyperspectral remote sensing data samples and surface parameter samples of ground objects, and the quantitative inversion model of surface parameters is used to invert the surface parameter values ​​of ground objects. The modeling process of the quantitative inversion model of surface parameters is as follows: firstly, the hyperspectral remote sensing data samples are feature enhanced, and based on the feature enhancement results, the hyperspectral remote sensing data samples are weighted to obtain weighted hyperspectral remote sensing data samples. The surface parameter samples and the weighted hyperspectral remote sensing data samples can be combined to establish a quantitative inversion model of surface parameters based on hyperspectral remote sensing by using partial least squares regression (PLSR).

[0065] It should be noted that based on the surface parameter samples and weighted hyperspectral remote sensing data samples, algorithms such as machine learning, deep learning, curve fitting or other regression analysis can also be used to establish a quantitative inversion model for surface parameters.

[0066] For example, the weighted hyperspectral remote sensing data is input into a surface parameter quantitative inversion model, and the output of the surface parameter quantitative inversion model is the surface parameter inversion value of the object to be measured, so that the surface parameter value of the object to be measured can be obtained through the surface parameter quantitative inversion model.

[0067] It should be noted that the land features may be vegetation, soil, water bodies and the like, and the surface parameters may be the physical and chemical parameters of the land features, such as the chlorophyll content of vegetation, the heavy metal content of soil, the chemical oxygen demand of water bodies, etc.; the surface parameter values ​​may be the contents of the physical and chemical parameters of the land features, such as the chlorophyll content of vegetation, the heavy metal content of soil, the chemical oxygen demand content of water bodies, etc.

[0068] The method for quantitative inversion of surface parameters of hyperspectral remote sensing provided by the embodiment of the present invention can fully utilize hyperspectral information according to the response of surface parameters in different spectral bands by performing feature enhancement on the hyperspectral remote sensing data of the object to be measured and weighting the hyperspectral remote sensing data based on the feature enhancement result; and input the weighted hyperspectral remote sensing data into the quantitative inversion model of surface parameters, thereby improving the accuracy of the surface parameters of the object to be measured inverted by the quantitative inversion model of surface parameters.

[0069] Optionally, Figure 2 This is the second flow chart of a method for quantitative inversion of hyperspectral remote sensing surface parameters provided by the present invention, such as Figure 2 As shown, before executing step 102, the method further includes the following steps:

[0070] Step 105: preprocess the hyperspectral remote sensing data to obtain preprocessed hyperspectral remote sensing data.

[0071] Among them, when the hyperspectral remote sensing data of the measured object are different, the preprocessing process is also different. Since the hyperspectral remote sensing data include the reflectance spectral curve measured by the object spectrometer and the hyperspectral remote sensing image obtained by the imaging spectrometer, the preprocessing process of the reflectance spectral curve and the hyperspectral remote sensing image are different.

[0072] Optionally, when the hyperspectral remote sensing data includes a reflectance spectrum curve acquired by a ground object spectrometer, step 105 is specifically implemented in the following manner:

[0073] The reflectivity spectrum curve is subjected to noise removal and band extraction processing to obtain a preprocessed reflectivity spectrum curve.

[0074] For example, the reflectance spectrum curve obtained by the ground object spectrometer is preprocessed as needed to obtain a preprocessed reflectance spectrum curve, wherein the preprocessing includes but is not limited to noise removal, band extraction and other processing.

[0075] Optionally, Figure 3 This is the third flow chart of a method for quantitative inversion of hyperspectral remote sensing surface parameters provided by the present invention. Figure 4 is a schematic diagram of the reflectivity spectrum curve after pretreatment provided by the present invention; Figure 3 As shown, in the case where the hyperspectral remote sensing data includes a hyperspectral remote sensing image, step 105 is implemented by the following steps:

[0076] Step 1051: preprocess the hyperspectral remote sensing image to obtain a preprocessed remote sensing image.

[0077] The preprocessing of the hyperspectral remote sensing image includes at least one of the following: performing radiation calibration on the hyperspectral remote sensing image, performing atmospheric correction on the hyperspectral remote sensing image, performing geometric precision correction on the hyperspectral remote sensing image, and performing noise removal on the hyperspectral remote sensing image.

[0078] For example, for hyperspectral remote sensing images, corresponding preprocessing is performed according to the processing level of the acquired hyperspectral remote sensing images and in combination with data characteristics. The preprocessing process includes but is not limited to radiation calibration, atmospheric correction, geometric precision correction, noise removal, etc.

[0079] Specifically, the embodiment of the present invention selects the Gaofen-5 (GF-5) satellite hyperspectral sensor (Advanced Hyperspectral Imaging, AHSI) to obtain a hyperspectral remote sensing image of the object to be measured. The obtained AHSI hyperspectral remote sensing image is a level 1 remote sensing product. In order to obtain a reflectance spectrum curve, the hyperspectral remote sensing image is preprocessed as follows:

[0080] Firstly, the radiometric calibration coefficient is used to perform radiometric calibration on the hyperspectral remote sensing image to obtain the radiometric brightness remote sensing image.

[0081] Secondly, atmospheric correction is performed on the radiometric brightness remote sensing image to obtain a reflectivity remote sensing image.

[0082] Third, geometrically precisely correct the reflectivity remote sensing image to obtain a corrected reflectivity remote sensing image.

[0083] Among them, geometric correction can be divided into two types: rough geometric correction and fine geometric correction; rough geometric correction is a correction performed on the cause of distortion; fine geometric correction is a geometric correction performed using control points, which uses a mathematical model to approximate the geometric distortion process of the remote sensing image, and uses some corresponding points (i.e., control point data pairs) between the distorted remote sensing image and the reference image to obtain the geometric distortion model, and then uses this model to correct the geometric distortion. This correction does not consider the cause of distortion; therefore, the embodiment of the present invention uses fine geometric correction to process the reflectivity remote sensing image.

[0084] For example, a high-precision remote sensing image is used as a reference image, ground control points (GCPs) are selected, and a geometric correction transformation model is constructed to perform geometric precision correction on the reflectivity remote sensing image to obtain a corrected reflectivity remote sensing image.

[0085] Step 1052: extracting a reflectance spectrum from the preprocessed remote sensing image to obtain a preprocessed reflectance spectrum curve.

[0086] For example, according to the sampling point coordinates recorded during sampling, the soil reflectance spectrum is extracted from the corrected reflectance remote sensing image to obtain the preprocessed reflectance spectrum curve, see Figure 4 shown.

[0087] It should be noted that if Figure 4 As shown, since the noise in the reflectance spectrum in the embodiment of the present invention is relatively small, a noise removal algorithm is not used to remove the noise; however, in another embodiment, if the noise in the reflectance spectrum is relatively large, the reflectance spectrum needs to be subjected to noise removal processing.

[0088] Optionally, step 102 is specifically implemented in the following manner:

[0089] The preprocessed reflectance spectrum curve is subjected to feature enhancement based on a feature enhancement algorithm to obtain a feature enhancement result.

[0090] Furthermore, the implementation of step 102 is different depending on whether the pre-processed reflectance spectrum curve has identifiable spectral features of surface parameters. Step 102 can be implemented in the following two ways:

[0091] The first method is to extract spectral data corresponding to the identifiable spectral features of surface parameters from the preprocessed reflectance spectral curve when the preprocessed reflectance spectral curve has the identifiable spectral features; and perform feature enhancement on the spectral data using a feature enhancement algorithm to obtain a feature enhancement result.

[0092] For example, in the case of surface parameters having identifiable spectral features on the spectral curve, spectral data of the spectral features corresponding to the surface parameters are extracted from the reflectance spectral curve preprocessed in step 105, and the extracted spectral curve (formed by the extracted spectral data) is feature enhanced using a feature enhancement algorithm to obtain a feature enhancement result.

[0093] The second method is to use a feature enhancement algorithm to perform feature enhancement on the reflectance spectrum curve preprocessed in step 105 when the spectrum feature of the surface parameter does not exist (cannot be or is difficult to identify) on the spectrum curve to obtain a feature enhancement result.

[0094] It should be noted that the specific implementation method of step 102 is selected according to the specific surface parameters. For example, the heavy metal Cu has no clear spectral characteristics on the soil reflectance spectrum curve, so the characteristic spectrum cannot be extracted. The second method mentioned above is used to directly enhance the characteristics of the reflectance spectrum curve.

[0095] The hyperspectral remote sensing surface parameter quantitative inversion method provided by the embodiment of the present invention extracts spectral data from the preprocessed reflectance spectral curve when there are identifiable spectral features of the surface parameters on the preprocessed reflectance spectral curve, which can reduce the amount of computational processing of the surface parameter quantitative inversion model, thereby improving the inversion efficiency of the surface parameters of the object to be measured.

[0096] Furthermore, when the spectral features of the surface parameters cannot be identified on the spectral curve, before the feature enhancement algorithm is used to perform feature enhancement on the preprocessed reflectance spectral curve, the method further includes:

[0097] Select the target spectrum from the preprocessed reflectance spectrum curve.

[0098] For example, according to the characteristic spectral position of the material related to the surface parameters of the object to be measured, the spectral segment corresponding to the characteristic spectral position is selected from the preprocessed reflectance spectrum curve as the target spectrum. For example, the object to be measured is soil, and the surface parameter to be measured (inverted) is the heavy metal Cu content in the soil; considering the influence of water vapor in the atmosphere on the satellite-borne hyperspectral remote sensing image, and the characteristic spectra of soil spectrally active substances such as soil organic matter and iron oxides related to the inversion of heavy metal Cu content are mainly located in the visible and near-infrared (Visible and Near-Infrared, VNIR) range, the VNIR spectrum of 390-1029nm is selected from the soil reflectance spectrum (preprocessed reflectance spectrum curve, 390-2400) as the target spectrum for the inversion of soil heavy metal Cu.

[0099] The hyperspectral remote sensing surface parameter quantitative inversion method provided by an embodiment of the present invention extracts a section of the spectrum from the preprocessed reflectance spectrum curve as the target spectrum when there is no identifiable spectral feature of the surface parameters on the preprocessed reflectance spectrum curve, thereby further reducing the calculation amount of the surface parameter quantitative inversion model and further improving the inversion efficiency of the surface parameters of the object to be measured.

[0100] Further, Figure 5 This is the fourth flow chart of a method for quantitative inversion of hyperspectral remote sensing surface parameters provided by the present invention. Figure 6 This is a schematic diagram of the removal result of a hyperspectral remote sensing soil reflectance spectrum continuum provided by the present invention. Figure 7 Schematic diagram of soil reflectance spectrum continuum removal transformation result provided by the present invention; Figure 5 As shown, step 102 specifically includes the following steps:

[0101] Step 1021: perform continuum removal or spectral derivation on the preprocessed reflectance spectrum curve based on a continuum removal algorithm or a spectral derivation algorithm to obtain a continuum removal result or a spectral derivation result.

[0102] For example, taking soil as an example, the continuum removal algorithm is used to remove the preprocessed reflectance spectrum curve continuum, and the soil reflectance spectrum continuum removal result is obtained, such as Figure 6 shown.

[0103] Step 1022: Use the continuum removal result or the spectrum derivative result as a feature enhancement result.

[0104] Furthermore, step 103 is specifically implemented in the following manner:

[0105] The feature enhancement result obtained in step 102 is used to perform weighted processing on the preprocessed reflectance spectrum curve obtained in step 105 .

[0106] For example, first, according to application requirements, the continuum removal result can be subjected to transformation processing such as translation and flipping. Specifically, the embodiment of the present invention transforms the continuum removal result based on formula (1) to obtain the soil reflectance spectrum continuum removal transformation result, such as Figure 7 shown.

[0107] D T =-1×D+1 (1)

[0108] Wherein, D represents the continuum removal result, D T It is the continuum removal transformation result, that is, the feature enhancement result.

[0109] Secondly, each reflectivity spectrum of the reflectivity spectrum curve is multiplied by the corresponding spectrum curve continuum removal transformation result to achieve weighted processing of the reflectivity spectrum curve. Each reflectivity spectrum is specifically based on formula (2) to obtain weighted hyperspectral remote sensing data. After all reflectivity spectra are weighted, weighted hyperspectral remote sensing data is obtained.

[0110] R T =R×D T (2)

[0111] Where R is the reflectance spectrum, D T is the result of the continuum removal transformation, R T It is the result of reflectance spectrum weighting processing.

[0112] It should be noted that in the prior art, when hyperspectral remote sensing is used to monitor vegetation, soil, water bodies and other land objects, the entire spectrum is usually used directly for quantitative inversion of surface parameters. Since the spectral range of hyperspectral remote sensing covers visible light, near infrared and short-wave infrared, the characteristics or responses of surface parameters in these bands are different. If all spectral bands are directly used for quantitative inversion of surface parameters, the response differences of spectral bands cannot be reflected, which will also affect the accuracy of the inversion results; the present invention weights each band of the reflectance spectrum curve by the response of the surface parameters in each band of the reflectance spectrum, so as to make full use of the spectral information and thus improve the inversion accuracy of the surface parameters.

[0113] It should be noted that the spectral derivative of the preprocessed reflectance spectral curve is performed based on the spectral derivative algorithm to obtain the spectral derivative result. When the preprocessed reflectance spectral curve is weighted, it is necessary to construct a corresponding weighting formula according to the weighting requirements and apply it to the reflectance spectral curve to obtain weighted hyperspectral remote sensing data.

[0114] The method for quantitative inversion of hyperspectral remote sensing surface parameters provided by an embodiment of the present invention adopts a continuum removal algorithm to perform continuum removal on the preprocessed reflectance spectral curve, and transforms the continuum removal result to obtain a feature enhancement result, so as to facilitate subsequent weighted processing of the hyperspectral remote sensing data based on the feature enhancement result; and by weighting each band of the hyperspectral remote sensing data to reflect the spectral response of the surface parameters in different spectral bands, the hyperspectral information is fully utilized to improve the quantitative inversion accuracy of the surface parameters.

[0115] Optionally, Figure 8 This is a fifth flow chart of a method for quantitative inversion of hyperspectral remote sensing surface parameters provided by the present invention, such as Figure 8 As shown in Figure 1, the quantitative inversion model of surface parameters is trained based on the following steps:

[0116] Step 801, obtaining sample data; the sample data includes surface parameter samples and hyperspectral remote sensing data samples.

[0117] The ground feature may be vegetation, soil, water, etc., and the surface parameter may be the physical and chemical parameters of the ground feature, such as the chlorophyll content of the vegetation, the heavy metal content of the soil, the chemical oxygen demand content of the water, etc. In the embodiment of the present invention, the ground feature is soil as an example, and the physical and chemical parameter is the heavy metal copper (Cu) content of the soil.

[0118] For example, a soil sample field collection plan is formulated according to the soil sampling rules, and soil is collected according to the collection plan to obtain soil samples.

[0119] According to the chemical analysis method of soil samples, inductively coupled plasma optical emission spectrometer (ICP-OES) was used to determine the heavy metal copper (Cu) content of soil samples, thereby obtaining surface parameter samples of soil samples.

[0120] According to the soil sample collection time, the hyperspectral remote sensing images of the sampling area acquired by the Gaofen-5 (GF-5) satellite hyperspectral sensor (Advanced Hyperspectral Imaging, AHSI) with the same or similar collection time as the surface parameter samples are selected to obtain the hyperspectral remote sensing data samples of the soil samples.

[0121] It should be noted that soil sampling rules refer to the method of collecting soil samples. If the metal content in the soil is to be analyzed, the use of metal utensils for sampling should be avoided. When sampling in the field, due to the heterogeneity of the spatial distribution of the soil itself, it is necessary to take samples at multiple points in units of plots and then mix them into a mixed sample, so as to better represent the soil properties of the sampling area. There are generally the following soil sampling rules: diagonal sampling method: suitable for sewage irrigation plots, sampling at the central points of each diagonal division; plum blossom sampling method: suitable for plots with small area, flat terrain and uniform soil; chessboard sampling method: suitable for medium-sized plots with flat terrain, basically intact terrain and uneven soil; snake sampling method: suitable for plots with small area, uneven terrain, uneven soil, and more sampling points.

[0122] Step 802: Perform feature enhancement on the hyperspectral remote sensing data sample to obtain a feature enhancement result sample.

[0123] Among them, the heavy metal Cu has no clear spectral features in the soil reflectance spectrum of the hyperspectral remote sensing image, so feature extraction is not performed. The soil reflectance spectrum samples obtained based on the hyperspectral remote sensing image samples are feature enhanced to obtain feature enhanced result samples.

[0124] Step 803: Perform weighted processing on the hyperspectral remote sensing data samples based on the feature enhancement result samples to obtain weighted hyperspectral remote sensing data samples.

[0125] Step 804: construct a quantitative inversion model for surface parameters based on the surface parameter samples and the weighted hyperspectral remote sensing data samples to obtain the quantitative inversion model for surface parameters.

[0126] For example, the surface parameter samples and weighted hyperspectral remote sensing data samples can be combined to establish a quantitative inversion model of surface parameters based on hyperspectral remote sensing using partial least squares regression (PLSR). The surface parameter quantitative inversion model can also be established using machine learning, deep learning, curve fitting or other regression analysis methods.

[0127] It should be noted that steps 801 to 804 may be performed before step 101, or after step 101, or after step 102, and the present invention is not limited to this.

[0128] It should be noted that the acquisition and processing process of the hyperspectral remote sensing data samples in steps 801 to 804 must be the same and similar (almost the same) as the acquisition and processing process of steps 101 to 103 and step 105.

[0129] The method for quantitative inversion of surface parameters using hyperspectral remote sensing provided by an embodiment of the present invention obtains hyperspectral remote sensing data and surface parameters, pre-processes the hyperspectral remote sensing data, enhances the features of the hyperspectral data according to the spectral response characteristics of the surface parameters, weights the hyperspectral remote sensing data using the feature enhancement results, and combines the weighted hyperspectral remote sensing data and surface parameters to establish a quantitative inversion model for surface parameters based on hyperspectral remote sensing. The acquisition and processing process of the hyperspectral remote sensing data of the object to be measured is the same or similar to the acquisition and processing process in constructing the quantitative inversion model for surface parameters, thereby ensuring the accuracy of the quantitative inversion results of the surface parameters of the object to be measured.

[0130] Optionally, after executing step 804, the method further includes the following steps:

[0131] The quantitative inversion model of surface parameters was evaluated.

[0132] For example, Fig. 9 FIG. 1 is a schematic diagram of quantitative inversion results of surface parameters provided by hyperspectral remote sensing of the present invention, such as Fig. 9 As shown in the figure, the measured object takes soil as an example, and the measured surface parameter takes the soil heavy metal Cu content as an example. First, a verification sample is obtained, and the verification sample includes the soil hyperspectral remote sensing data verification sample and the soil heavy metal Cu content verification sample. Then, the accuracy of the surface parameter quantitative inversion model is evaluated based on the hyperspectral remote sensing data verification sample and the soil heavy metal Cu content verification sample; RMSE (Root mean square error) is the root mean square error, which represents the root mean square error of the surface parameter quantitative inversion model on the verification sample; R 2 (Coefficient of determination) is the determination coefficient, which indicates the determination coefficient of the quantitative inversion model of surface parameters on the verification sample. Fig. 9 The measured values ​​(actually obtained sample data) and predicted values ​​(output values ​​of the quantitative inversion model of surface parameters) show that the root mean square error RMSE of the quantitative inversion model of surface parameters of the present invention is 18.34 mg / kg, and the determination coefficient is 0.71, indicating that the quantitative inversion model of surface parameters of the present invention has high inversion accuracy and can be used for quantitative inversion of surface parameters of measured objects.

[0133] Optionally, the specific steps of a method for quantitative inversion of surface parameters of hyperspectral remote sensing of the present invention are as follows:

[0134] S1. Obtain the AHSI hyperspectral remote sensing image of the object to be measured;

[0135] S2, classifying the AHSI hyperspectral remote sensing image obtained in S1 by using a supervised classification method to obtain a remote sensing image classification result;

[0136] S3, extracting soil area hyperspectral remote sensing images from AHSI hyperspectral remote sensing images according to remote sensing image classification results;

[0137] S4, preprocessing, feature enhancement and weighting the extracted soil area hyperspectral remote sensing image to obtain a weighted soil area hyperspectral remote sensing image;

[0138] S5. Apply the established quantitative inversion model of surface parameters to the extracted weighted hyperspectral remote sensing image of the soil area to obtain a soil area heavy metal Cu content map output by the quantitative inversion model of surface parameters.

[0139] Furthermore, after executing step S5, the method further comprises the following steps:

[0140] S6. According to the needs, spatial interpolation and other geo-statistical methods can be used to perform spatial interpolation based on the inversion results of the heavy metal Cu content in the soil area to obtain the soil heavy metal Cu content distribution map of the entire area.

[0141] The hyperspectral remote sensing surface parameter quantitative inversion method provided in the embodiment of the present invention uses the established surface parameter quantitative inversion model for surface parameter quantitative inversion and mapping, which is conducive to promoting the application of hyperspectral remote sensing in the monitoring of surface parameters such as vegetation, soil, and water bodies, and can improve the quantitative monitoring capability of surface parameters.

[0142] A hyperspectral remote sensing surface parameter quantitative inversion device provided by the present invention is described below. The hyperspectral remote sensing surface parameter quantitative inversion device described below and the hyperspectral remote sensing surface parameter quantitative inversion method described above can be referred to each other.

[0143] Fig.10 Schematic diagram of the structure of a device for quantitative inversion of surface parameters of hyperspectral remote sensing provided by the present invention. Fig.10 As shown, the quantitative inversion of surface parameters includes an acquisition unit 1001, a first obtaining unit 1002, a second obtaining unit 1003 and a third obtaining unit 1004; wherein:

[0144] An acquisition unit 1001 is used to acquire hyperspectral remote sensing data of an object to be measured;

[0145] A first obtaining unit 1002 is used to perform feature enhancement on the hyperspectral remote sensing data to obtain a feature enhancement result;

[0146] A second obtaining unit 1003 is used to perform weighted processing on the hyperspectral remote sensing data based on the feature enhancement result to obtain weighted hyperspectral remote sensing data;

[0147] The third obtaining unit 1004 is used to input the weighted hyperspectral remote sensing data into the quantitative inversion model of surface parameters to obtain the inversion value of the surface parameters of the object to be measured output by the quantitative inversion model of surface parameters; the quantitative inversion model of surface parameters is used to invert the surface parameter values ​​of the object.

[0148] The hyperspectral remote sensing surface parameter quantitative inversion device provided by the present invention can fully utilize the hyperspectral information according to the response of the surface parameters in different spectral bands by performing feature enhancement on the hyperspectral remote sensing data of the object to be measured and performing weighted processing on the hyperspectral remote sensing data based on the feature enhancement result; and input the weighted hyperspectral remote sensing data into the surface parameter quantitative inversion model, thereby improving the accuracy of the surface parameter inversion value of the object to be measured output by the surface parameter quantitative inversion model.

[0149] Based on any of the above embodiments, the hyperspectral remote sensing data includes a reflectance spectrum curve obtained by a ground object spectrometer; and the device further includes:

[0150] The fourth obtaining unit is used to perform noise removal and band extraction processing on the reflectance spectrum curve to obtain a pre-processed reflectance spectrum curve.

[0151] Based on any of the above embodiments, the hyperspectral remote sensing data includes a hyperspectral remote sensing image; and the device further includes:

[0152] A fifth obtaining unit is used to preprocess the hyperspectral remote sensing image to obtain a preprocessed remote sensing image;

[0153] A sixth obtaining unit, used for extracting a reflectance spectrum from the preprocessed remote sensing image to obtain a preprocessed reflectance spectrum curve;

[0154] The preprocessing of the hyperspectral remote sensing image includes at least one of the following: performing radiation calibration on the hyperspectral remote sensing image, performing atmospheric correction on the hyperspectral remote sensing image, performing geometric precision correction on the hyperspectral remote sensing image, and performing noise removal on the hyperspectral remote sensing image.

[0155] Based on any of the above embodiments, the first obtaining unit 1002 is specifically configured to:

[0156] The preprocessed reflectance spectrum curve is subjected to feature enhancement based on a feature enhancement algorithm to obtain a feature enhancement result.

[0157] Based on any of the foregoing embodiments, the first obtaining unit 1002 is further specifically configured to:

[0158] In the case where the preprocessed reflectance spectrum curve has an identifiable spectrum feature of a ground parameter, extracting spectrum data corresponding to the identifiable spectrum feature from the preprocessed reflectance spectrum curve;

[0159] The spectral data is feature enhanced using a feature enhancement algorithm to obtain a feature enhancement result.

[0160] Based on any of the foregoing embodiments, the first obtaining unit 1002 is further specifically configured to:

[0161] Based on a continuum removal algorithm or a spectrum derivation algorithm, performing continuum removal or spectrum derivation on the preprocessed reflectance spectrum curve to obtain a continuum removal result or a spectrum derivation result;

[0162] The continuum removal result or the spectrum derivative result is used as a feature enhancement result.

[0163] Based on any of the above embodiments, the surface parameter quantitative inversion model is obtained by training based on the following steps:

[0164] Acquire sample data of ground objects; the sample data includes surface parameter samples and hyperspectral remote sensing data samples;

[0165] Performing feature enhancement on the hyperspectral remote sensing data sample to obtain a feature enhancement result sample;

[0166] Based on the feature enhancement result sample, weighted processing is performed on the hyperspectral remote sensing data sample to obtain a weighted hyperspectral remote sensing data sample;

[0167] A surface parameter quantitative inversion model is constructed based on the surface parameter samples and the weighted hyperspectral remote sensing data samples to obtain the surface parameter quantitative inversion model.

[0168] Fig.11 An example of a physical structure diagram of an electronic device is shown in FIG. Fig.11 As shown, the electronic device may include: a processor 1110, a communication interface 1120, a memory 1130 and a communication bus 1140, wherein the processor 1110, the communication interface 1120 and the memory 1130 communicate with each other through the communication bus 1140. The processor 1110 may call the logic instructions in the memory 1130 to execute the hyperspectral remote sensing surface parameter quantitative inversion method, which includes: obtaining hyperspectral remote sensing data of the object to be measured;

[0169] Performing feature enhancement on the hyperspectral remote sensing data to obtain a feature enhancement result;

[0170] Based on the feature enhancement result, weighted processing is performed on the hyperspectral remote sensing data to obtain weighted hyperspectral remote sensing data;

[0171] The weighted hyperspectral remote sensing data is input into a quantitative inversion model for surface parameters to obtain inversion values ​​of surface parameters of the object to be measured output by the quantitative inversion model for surface parameters; the quantitative inversion model for surface parameters is used to invert the surface parameter values ​​of the object.

[0172] In addition, the logic instructions in the above-mentioned memory 1130 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0173] On the other hand, the present invention further provides a computer program product, the computer program product includes a computer program, the computer program can be stored in a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute the hyperspectral remote sensing surface parameter quantitative inversion method provided by the above methods, the method includes: obtaining hyperspectral remote sensing data of a surface object to be measured;

[0174] Performing feature enhancement on the hyperspectral remote sensing data to obtain a feature enhancement result;

[0175] Based on the feature enhancement result, weighted processing is performed on the hyperspectral remote sensing data to obtain weighted hyperspectral remote sensing data;

[0176] The weighted hyperspectral remote sensing data is input into a quantitative inversion model for surface parameters to obtain inversion values ​​of surface parameters of the object to be measured output by the quantitative inversion model for surface parameters; the quantitative inversion model for surface parameters is used to invert the surface parameter values ​​of the object.

[0177] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to execute the hyperspectral remote sensing surface parameter quantitative inversion method provided by the above methods, the method comprising: obtaining hyperspectral remote sensing data of a surface object to be measured;

[0178] Performing feature enhancement on the hyperspectral remote sensing data to obtain a feature enhancement result;

[0179] Based on the feature enhancement result, weighted processing is performed on the hyperspectral remote sensing data to obtain weighted hyperspectral remote sensing data;

[0180] The weighted hyperspectral remote sensing data is input into a quantitative inversion model for surface parameters to obtain inversion values ​​of surface parameters of the object to be measured output by the quantitative inversion model for surface parameters; the quantitative inversion model for surface parameters is used to invert the surface parameter values ​​of the object.

[0181] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0182] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0183] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention 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 invention.

Claims

1. A method for quantitative inversion of surface parameters using hyperspectral remote sensing, characterized in that: include: Obtaining hyperspectral remote sensing data of the object to be measured; Preprocessing the hyperspectral remote sensing data to obtain preprocessed hyperspectral remote sensing data; Performing feature enhancement on the hyperspectral remote sensing data to obtain a feature enhancement result; Based on the feature enhancement result, weighted processing is performed on the hyperspectral remote sensing data to obtain weighted hyperspectral remote sensing data; Inputting the weighted hyperspectral remote sensing data into a quantitative inversion model for surface parameters to obtain inversion values ​​of surface parameters of the object to be measured output by the quantitative inversion model for surface parameters; the quantitative inversion model for surface parameters is used to invert the surface parameter values ​​of the object; The hyperspectral remote sensing data includes a reflectance spectrum curve obtained by a ground object spectrometer; the preprocessing of the hyperspectral remote sensing data to obtain the preprocessed hyperspectral remote sensing data includes: Performing noise removal and band extraction processing on the reflectivity spectrum curve to obtain a preprocessed reflectivity spectrum curve; The hyperspectral remote sensing data includes a hyperspectral remote sensing image; the preprocessing of the hyperspectral remote sensing data to obtain the preprocessed hyperspectral remote sensing data includes: Preprocessing the hyperspectral remote sensing image to obtain a preprocessed remote sensing image; extracting a reflectance spectrum from the preprocessed remote sensing image to obtain a preprocessed reflectance spectrum curve; the preprocessing of the hyperspectral remote sensing image includes at least one of the following: performing radiation calibration on the hyperspectral remote sensing image, performing atmospheric correction on the hyperspectral remote sensing image, performing geometric precision correction on the hyperspectral remote sensing image, and performing noise removal on the hyperspectral remote sensing image; The step of performing feature enhancement on the hyperspectral remote sensing data to obtain a feature enhancement result includes: Based on a continuum removal algorithm, continuum removal is performed on the preprocessed reflectance spectrum curve to obtain a continuum removal result; and the continuum removal result is used as a feature enhancement result; The step of performing weighted processing on the hyperspectral remote sensing data based on the feature enhancement result to obtain weighted hyperspectral remote sensing data includes: Based on formula (1), the continuum removal result is transformed to obtain the soil reflectance spectrum continuum removal transformation result; （1） In the formula, D represents the continuum removal result, D T Remove the transformation result for the continuum; Each reflectance spectrum of the reflectance spectrum curve is weighted based on formula (2) to obtain the reflectance spectrum weighted processing result, and all reflectance spectra are weighted to obtain weighted hyperspectral remote sensing data; （2） Where R is the reflectance spectrum, D T is the result of the continuum removal transformation, R T It is the result of reflectance spectrum weighting processing.

2. The method for quantitative inversion of surface parameters using hyperspectral remote sensing according to claim 1, characterized in that: Performing feature enhancement on the preprocessed reflectance spectrum curve to obtain a feature enhancement result, including: In the case where the preprocessed reflectance spectrum curve has an identifiable spectrum feature of a ground parameter, extracting spectrum data corresponding to the identifiable spectrum feature from the preprocessed reflectance spectrum curve; The spectral data is feature enhanced using a feature enhancement algorithm to obtain a feature enhancement result.

3. The method for quantitative inversion of surface parameters using hyperspectral remote sensing according to claim 1, characterized in that: The surface parameter quantitative inversion model is obtained by training based on the following steps: Acquire sample data of ground objects; the sample data includes surface parameter samples and hyperspectral remote sensing data samples; Performing feature enhancement on the hyperspectral remote sensing data sample to obtain a feature enhancement result sample; Based on the feature enhancement result sample, weighted processing is performed on the hyperspectral remote sensing data sample to obtain a weighted hyperspectral remote sensing data sample; A surface parameter quantitative inversion model is constructed based on the surface parameter samples and the weighted hyperspectral remote sensing data samples to obtain the surface parameter quantitative inversion model.

4. A device for quantitative inversion of surface parameters using hyperspectral remote sensing, characterized in that: include: An acquisition unit, used for acquiring hyperspectral remote sensing data of the object to be measured; A fifth obtaining unit is used to preprocess the hyperspectral remote sensing data to obtain preprocessed hyperspectral remote sensing data; A first obtaining unit is used to perform feature enhancement on the hyperspectral remote sensing data to obtain a feature enhancement result; A second obtaining unit is used to perform weighted processing on the hyperspectral remote sensing data based on the feature enhancement result to obtain weighted hyperspectral remote sensing data; The third obtaining unit is used to input the weighted hyperspectral remote sensing data into the quantitative inversion model of surface parameters to obtain the inversion value of the surface parameters of the object to be measured output by the quantitative inversion model of surface parameters; the quantitative inversion model of surface parameters is used to invert the surface parameter value of the object; The hyperspectral remote sensing data includes a reflectance spectrum curve obtained by a ground object spectrometer; the fifth obtaining unit is used to perform noise removal and band extraction processing on the reflectance spectrum curve to obtain a pre-processed reflectance spectrum curve; The hyperspectral remote sensing data includes a hyperspectral remote sensing image; the fifth obtaining unit is used to preprocess the hyperspectral remote sensing image to obtain a preprocessed remote sensing image; extract a reflectance spectrum from the preprocessed remote sensing image to obtain a preprocessed reflectance spectrum curve; the preprocessing of the hyperspectral remote sensing image includes at least one of the following: performing radiation calibration on the hyperspectral remote sensing image, performing atmospheric correction on the hyperspectral remote sensing image, performing geometric precision correction on the hyperspectral remote sensing image, and performing noise removal on the hyperspectral remote sensing image; The first obtaining unit is used to perform continuum removal on the preprocessed reflectance spectrum curve based on a continuum removal algorithm to obtain a continuum removal result; and use the continuum removal result as a feature enhancement result; The second obtaining unit is used to transform the continuum removal result based on formula (1) to obtain the soil reflectance spectrum continuum removal transformation result; each reflectance spectrum of the reflectance spectrum curve is used to obtain the reflectance spectrum weighted processing result based on formula (2), and all reflectance spectra are weighted to obtain weighted hyperspectral remote sensing data; （1） （2） In the formula, D represents the continuum removal result, D T is the result of continuum removal transformation; R is the reflectance spectrum, D T is the result of the continuum removal transformation, R T It is the result of reflectance spectrum weighting processing.

5. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method for quantitative inversion of hyperspectral remote sensing surface parameters as described in any one of claims 1 to 3 is implemented.

6. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for quantitative inversion of hyperspectral remote sensing surface parameters as claimed in any one of claims 1 to 3 is implemented.

Images