Satellite hyperspectral channel selection method and device

Abstract

The invention provides a satellite hyperspectral channel selection method and device, and the method comprises the steps: calculating the drop rate of a value function corresponding to each satellite hyperspectral channel, and taking the channel with the maximum drop rate as a first selection channel; an alternative channel is screened through an included angle range of Jacobin vectors of an unselected channel and a selected channel after linear transformation, so that atmosphere information contained in the alternative channel is remarkably different from that contained in the selected channel; combining each alternative channel with the selected (n-1) channels, calculating the drop rate of the value function by using a recursion method, and taking the channel with the maximum drop rate as the nth selected channel; and calculating a weight inverse matrix after the combination of the selected n channels by using a recursive algorithm for selecting the next channel, and repeating the above process until n is equal to a preset value. According to the method, the maximum decline rate of the value function is used as a channel selection standard, the assimilation reliability and accuracy of the variational method are ensured, and due to the calculation complexity and the parallel function of O (n2), the method has high calculation efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a method and device for satellite hyperspectral channel selection. Background Art

[0002] With the rapid development of hyperspectral remote sensing technology, modern meteorological satellites can obtain observation data of thousands of spectral channels. For example, the Infrared Atmospheric Sounding Interferometer (IASI) has 8461 channels, and the GIIRS imager has 1690 channels. Although this data provides rich information for the accurate inversion of atmospheric parameters, its huge data volume and complex correlation pose great challenges to storage, transmission, and calculation. In this context, it is extremely necessary to eliminate redundant data and improve the efficiency of satellite data processing. Therefore, the channel selection method is one of the core issues in hyperspectral remote sensing applications.

[0003] The sensitivities of different spectral channels to atmospheric parameters vary significantly at different altitudes. If all channels are used without selection, it may lead to uneven height distribution or signal overlap in the spectral sensitive area. In addition, the Jacobian vectors of adjacent channels may be highly correlated due to spectral overlap. If directly used for inversion, it will lead to ill-conditioning of the matrix and be difficult to invert. Therefore, the channel selection method can not only improve the extraction accuracy of the effective information of the data, but also be an important link to ensure the numerical stability of the assimilation system.

[0004] The core of channel selection lies in defining a reasonable information quantization index. The atmospheric retrievability index and entropy subtraction comprehensively consider the background field error and observation error, and are relatively widely used methods in current operations. Both of them essentially use the reduction ratio of the determinant value of the error covariance matrix obtained by variational decomposition as the information quantization index. Compared with the Jacobian matrix channel selection method, although the efficiency of the above two algorithms is slightly lower, by dynamically updating the covariance matrix, redundant information can be eliminated more accurately, and the DFS (Degrees of Freedom for Signal) value is significantly improved. When computing resources permit, the atmospheric retrievability index and entropy subtraction are more suitable for scenarios with high-precision requirements.

[0005] However, the deficiencies of the atmospheric retrievable index and entropy subtraction are also relatively obvious. First, only the information of the background error covariance matrix of the atmosphere and the Jacobin matrix of the satellite channels based on the atmospheric background field are used. When the error of the background field is large, the selected channels are likely to be unreliable. Second, among all the channel combinations, the atmospheric retrievable index and entropy subtraction obtain the channel combination with the largest probability that can minimize the background error variance of the variational solution, but this does not mean that the probability value of this combination being the optimal combination is large. For example, when this probability value is less than 0.5, it indicates that most of the optimal combinations are scattered among other channel combinations, and the channel combinations obtained by these two methods are likely to cause the cost function not to decrease significantly, resulting in the inability to fully obtain satellite observation information. Summary of the Invention

[0006] The present invention provides a method and device for selecting satellite hyperspectral channels to solve the deficiencies in the prior art that the selected channels are unreliable and not the optimal combination. For the variational assimilation problem of satellite data, the present invention directly estimates the contribution of each channel to the cost function, selects the satellite channel combination that can make the cost function decrease the most, can consider real-time satellite observation information, analyze specific problems specifically, dynamically select channels according to observation data, and is applicable to satellite channel selection under various weather conditions.

[0007] The present invention provides a method for selecting satellite hyperspectral channels, including:

[0008] Calculate the decrease rate of the cost function corresponding to each channel in the satellite hyperspectral channel set, and take the channel with the largest decrease rate as the first selected channel;

[0009] Use the background error covariance matrix to perform a linear transformation on the Jacobin vectors of the unselected channels and the selected channels in the satellite hyperspectral channel set to obtain vectors, and calculate the included angle between the vectors of the unselected channels and the selected channels. Screen out a subset of candidate channels from the unselected channels by setting the included angle range, so that the atmospheric vertical structure information observed by the channels in the subset of candidate channels is significantly different from that of the selected channels;

[0010] Combine each channel in the subset of candidate channels with n - 1 selected channels, where n≥2, and calculate the decrease rate of the cost function of each combination using a recursive algorithm based on the (n - 1)×(n - 1) weight inverse matrix of the selected channels. Take the channel corresponding to the combination with the largest decrease rate as the nth selected channel;

[0011] After obtaining the nth selected channel, use a recursive algorithm to calculate the n×n weight inverse matrix of the n selected channels for continuing to select the next channel.

[0012] A method for selecting satellite hyperspectral channels provided by the present invention determines the set of observation channels of the satellite through the following formula The rate of decline of the value functions corresponding to M channels in

[0013]

[0014]

[0015]

[0016]

[0017]

[0018]

[0019]

[0020] Among them, Are respectively the radiation observation value, radiation simulation value, Jacobin vector of radiation simulation, and error variance of observed radiation value of the j-th channel in the satellite observation channel set ; , where Represents the analytical solution of the atmospheric profile when using 1 channel, Is its background vector, and m is the number of vertical layers of the atmospheric profile. The error covariance matrix of is B; Represents the diagonal matrix of the error variances of the observed radiation values in n channels; Represents the rate of decline of the value function when there is only one channel; the superscripts T and Represent matrix transpose and diagonal matrix respectively.

[0021] A method for selecting satellite hyperspectral channels provided by the present invention constructs a set of candidate channels for screening the n-th channel from the unselected channels through the following formula :

[0022]

[0023]

[0024]

[0025] Among them, Represents the angle between the Jacobin vectors of the j-th unselected channel and the i-th selected channel after linear transformation; Are respectively the subsets The radiative observation value, radiative simulation value, Jacobian vector of radiative simulation, and observation error variance of the j-th channel in The Jacobian vector of the i-th channel representing the selected channels; is the adjustable angular increment parameter, represents the subset The number of channels in. When is an empty set, the alternative set is reconstructed by increasing the value.

[0026] According to a satellite hyperspectral channel selection method provided by the present invention, each channel in the subset is combined with the previous n - 1 selected channels respectively through the following recurrence formula, and the decline rate of the value function corresponding to each combination is calculated:

[0027]

[0028]

[0029]

[0030]

[0031]

[0032]

[0033]

[0034]

[0035]

[0036]

[0037]

[0038] Among them, are respectively the radiative observation value, radiative simulation value, Jacobian vector of radiative simulation, and observation error variance of the j-th channel in the subset ; represents the number of channels in the subset ; , where represents the analytical solution of the atmospheric profile when using n channels (n > 1), is its background vector, and the error covariance matrix is B; the subscript m is the vertical layer number of the atmospheric profile; H nm is the radiative simulation with respect to Jacobian matrix; denote the observation error variance on n channels, which is a diagonal matrix; denote the value function decrease rate when using n channels.

[0039] According to a satellite hyperspectral channel selection method provided by the present invention, the inverse weight matrix of the selected channel combination is determined by the following formula. According to the and corresponding to the nth selected channel, calculate the inverse weight matrix according to the following recurrence formula , and the algorithm is as follows:

[0040]

[0041] The present invention also provides a satellite hyperspectral channel selection device, including:

[0042] The first channel selection module calculates the decrease rate of the value function corresponding to each channel in the satellite hyperspectral channel set, and takes the channel with the largest decrease rate as the first selected channel

[0043] The alternative channel screening module is used to linearly transform the Jacobin vectors of the unselected channels and the selected channels in the satellite hyperspectral channel set with the background error covariance matrix to obtain a vector, and calculate the angle between the vectors of the unselected channels and the selected channels. By setting the angle range, an alternative channel subset is screened out from the unselected channels, so that the atmospheric vertical structure information observed by the channels in the alternative channel subset is significantly different from that of the selected channels;

[0044] The nth channel selection module is used to combine each channel in the alternative channel subset with n - 1 selected channels, n ≥ 2, and calculate the decrease rate of the value function of each combination according to the (n - 1)×(n - 1) inverse weight matrix of the selected channels by using a recurrence algorithm, and take the channel corresponding to the combination with the largest decrease rate as the nth selected channel;

[0045] The inverse weight matrix calculation module is used to calculate the n×n inverse weight matrix of n selected channels by using a recurrence algorithm after obtaining the nth selected channel, for continuing to select the next channel.

[0046] The present invention also provides an electronic device, including a memory, a processor, and a parallel computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the satellite hyperspectral channel selection method as described in any one of the above.

[0047] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the satellite hyperspectral channel selection method described in any one of the above is implemented.

[0048] The present invention also provides a computer program product, including a computer program, and when the computer program is executed by a processor, the satellite hyperspectral channel selection method described in any one of the above is implemented.

[0049] A satellite hyperspectral channel selection method and device provided by the present invention directly estimate the contribution of each channel to the value function for the variational assimilation problem of satellite data, and select a satellite channel combination that can make the value function decrease the most; considering real-time satellite observation information, analyzing specifically according to the specific weather, and dynamically selecting channels based on the observation data, which is applicable to satellite channel selection in various weather conditions, thereby improving the reliability and accuracy of channel selection. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0051] FIG. 1 is a schematic flow chart of the satellite hyperspectral channel selection method provided by the present invention;

[0052] Figure 2 is a schematic diagram of the comparison of the channels selected by the satellite hyperspectral channel selection method provided by the present invention and the entropy subtraction method and the corresponding simulated brightness temperature error;

[0053] Figure 3 is a schematic diagram of the comparison of the relative values of the value functions of the channel combinations selected by the satellite hyperspectral channel selection method provided by the present invention and the entropy subtraction method;

[0054] Figure 4 is a schematic diagram of the comparison of the entropy reduction change rates of the satellite hyperspectral channel selection method provided by the present invention and the entropy subtraction method;

[0055] Figure 5 is a schematic structural diagram of the satellite hyperspectral channel selection device provided by the present invention;

[0056] Figure 6 is a schematic structural diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0058] The selection of satellite hyperspectral channels can be formulated as the following problem: The satellite has a total of M hyperspectral observation channels. Given their radiation observation values, radiation simulation values, Jacobin vectors of radiation simulation, and observation error variances, which are defined in sequence as follows:

[0059]

[0060] Satellite channel selection is about how to select a subset from the above information to form a channel combination for data assimilation, so as to extract as much effective information of atmospheric variables as possible from the observed radiation data, and thus improve the initial field of the numerical model.

[0061] To distinguish from the members in , the member symbols in remove the identifier "^" and are used to represent the relevant information of the selected channels, and are renumbered and defined as follows:

[0062]

[0063] Facing thousands of observation channels of the satellite, it is necessary to develop an efficient algorithm to solve the computational amount problem of the cost function and its analytical solution. The cost function is a direct criterion for measuring the quality of the satellite channel combination, because a good channel combination must be able to significantly reduce the cost function. For a combination containing n satellite channels, , its one-dimensional cost function and cost function decline rate are as follows:

[0064]

[0065] The vector is the atmospheric vertical structure information to be solved, such as the vertical profiles of temperature and humidity in the atmosphere, divided into m layers; is the background vector of , usually provided by the numerical model; B is the error covariance matrix of , obtained by statistical analysis of historical data; the vector is the vector composed of the radiation observation values y of the n channels of . Correspondingly, represents the radiation simulation of this vector. denotes the observation error variance of the vector, which is a diagonal matrix. Under the first-order approximation condition formula (4), we can obtain the variational decomposition formula (5) of, as follows:

[0066]

[0067]

[0068] where H nm is the Jacobin matrix of, with the dimension of n×m. E is the entropy reduction index of the error covariance matrix of , which is consistent with the "atmospheric retrievability index" and the channel selection index of the "entropy subtraction method". Here, det represents the determinant of the matrix.

[0069] How to efficiently estimate the value function and its analytical solution is one of the main innovation points of the present invention and also one of the important difficult problems to be solved by the present invention. Currently, the "entropy subtraction method" is widely used in the business to select satellite channels, but the entropy subtraction method obtains a channel combination that can minimize the background error variance of the variational solution and has the largest probability theoretically compared to other combinations.

[0070] In this embodiment, for the variational assimilation problem of satellite data, the contribution of each channel to the value function is directly estimated, and the satellite channel combination that can make the value function decrease the most is selected; considering the real-time satellite observation information, specific weather is analyzed specifically, and channels are dynamically selected according to the observation data, which is applicable to the satellite channel selection under various weather conditions, thereby improving the reliability and accuracy of channel selection.

[0071] Next, in combination with Figure 1 describe a satellite hyperspectral channel selection method of the present invention, including:

[0072] Step 101: Determine the descent rate of the value function corresponding to each channel in the satellite hyperspectral observation channel set There are a total of M, and the channel with the largest descent rate is used as the first selected channel;

[0073] In this embodiment, the descent rate of the value function corresponding to each satellite channel is determined by the following formula:

[0074]

[0075] where are respectively the radiance observation value, radiance simulation value, radiance simulation Jacobin vector, and observation radiance value error variance of the j-th channel in the satellite observation channel set ; , where Denotes the analytical solution of the atmospheric profile when using 1 channel, is its background vector, and m is the number of vertical layers of the atmospheric profile. The error covariance matrix of is B; Denotes the diagonal matrix of the error variances of the observed radiation values in n channels; Denotes the rate of decrease of the cost function when there is only one channel; the superscripts T and denote matrix transpose and diagonal matrix respectively.

[0076] A high-quality channel combination must make significantly decrease. Therefore, use the rate of decrease of as the criterion for measuring the channel combination. First, in find the channel with the largest using the enumeration method as the first selected channel. When selecting the first channel, is a one-dimensional vector, is a number, so , and are all easy to calculate, obtaining with only one member and the corresponding , and for the selection of the next channel.

[0077] Step 102: Calculate the angle between the Jacobin vectors of the unselected channels and the selected channels after linear transformation, and construct a subset of candidate channels for screening the nth selected channel from the unselected channels , so that the subset contains information on the multi-level vertical structure of the atmosphere, where n > 1;

[0078] In this embodiment, the subset of candidate channels is constructed by the following formula :

[0079]

[0080] where, denotes the angle between the Jacobin vectors of the jth unselected channel and the ith selected channel after linear transformation; are respectively the radiation observation value, radiation simulation value, Jacobin vector of radiation simulation, and observation error variance of the jth channel in the subset ; denotes the Jacobin vector of the ith channel of the selected channels; is an adjustable angle increment parameter, denotes the number of channels in the subset When When it is an empty set, by increasing the value to reconstruct the alternative set.

[0081] How to make the channel combination contain as much high-precision information on atmospheric temperature and humidity at different heights as possible is another important and difficult problem to be faced. Equation (5) shows that is the background vector plus a linear combination of the n column vectors of . Therefore, whether multi-level atmospheric vertical structure information can be obtained depends on whether there is a certain included angle between so that the peaks of each are scattered in different height layers. For this purpose, the nth channel should be selected from

[0082] When , it means that the two vectors are orthogonal, and the difference in their vertical structure information is the largest. Although perfect orthogonality is the most ideal choice, the actual situation cannot meet it. Therefore, by adjusting the parameter so that remains within a certain angular range. When is an empty set, by increasing the value to reconstruct the subset , and the specific method can be seen in the numerical experiment part.

[0083] Step 103: Combine each channel in the subset with the first n - 1 selected channels respectively, calculate the decline rate of the value function corresponding to each combination, and use the channel corresponding to the combination with the largest decline rate as the nth selected channel. In this embodiment, the decline rate of the value function is calculated by the following formula:

[0084]

[0085] Among them, are respectively the radiation observation value, radiation simulation value, radiation simulation Jacobin vector and observation error variance of the jth channel in the subset , represents the number of channels in the subset , H nm is the matrix composed of n Jacobin vectors, and H （n-1）m is the Jacobin matrix of the selected channels; , where represents the analytical solution of the atmospheric profile when using n channels, is its background vector, and m is the number of vertical layers of the atmospheric profile. The error covariance matrix is B; represents the diagonal matrix of the error variance of the observed radiation value; represents the rate of decline of the value function when using the combination of n channels.

[0086] Step 104: Obtain the corresponding , , and , and then use the corresponding to calculate according to the recurrence formula, and then let n = n + 1 for the next channel selection process. The channel selection loops between Step 102 and Step 104 and ends when n = N. The recurrence algorithm of

[0087]

[0088] How to design an efficient solution method according to the matrix characteristics of the matrix to be inverted in the variational solution is one of the important challenges faced by the present invention. The core idea of the present invention for selecting satellite channels is to select a channel combination that can significantly reduce the value function. Therefore, it is necessary to face how to solve the variational analytical solution of the value function, and the solution of the inverse matrix is the key.

[0089] Taking the GIIRS imager of Fengyun 4B satellite (FY4B) as an example, there are 1690 channels. If n - 1 satellite channels have been selected, it is necessary to select an optimal channel from the remaining channels as the nth channel by the enumeration method, then it is necessary to solve the inverse matrix 1691 - n times. The computational complexity of the common inverse algorithms such as LU decomposition or Cholesky decomposition is , when the value of n is large, the solution of the inverse matrix requires a very large amount of computation, which is difficult to implement in actual operations. Therefore, solving the algorithm of the inverse matrix is the key problem.

[0090] Comprehensively analyzing the above recurrence formula, it can be seen that the computational complexity of formula (8) is about , mainly from . Using LU decomposition to calculate , and then calculating the value function, the computational complexity is about , and the efficiency is very low.

[0091] The present invention only needs to calculate once when selecting the nth channel, and its computational complexity is . Therefore, the amount of computation mainly comes from formula (8), and it is necessary to calculate times. The entropy subtraction and the atmospheric retrievable index also select channels by enumeration, and the computational complexity is , although this algorithm is not as efficient as these two algorithms, its computational complexity is quadratic. Moreover, when calculating the rate of decrease of the value function, the calculations between different channels are independent of each other and are suitable for parallel computing. Therefore, the computational amount is completely acceptable.

[0092] More importantly, the channel combination found by the present invention is the combination with the highest rate of decrease of the value function, which is consistent with the goal of variational decomposition and can obtain more atmospheric information. The present invention has developed a recursive algorithm for solving On this basis, a recursive algorithm for variational decomposition is derived, and then the rate of decrease of the value function is obtained, reducing the computational complexity to , which is one of the main core technologies of the present invention

[0093] To test whether the channel combination can reflect the complete atmospheric vertical structure information, this experiment considers the channel selection under clear sky conditions. Taking the GIIRS imager of the Fengyun 4B (FY4B) satellite as an example to test the method of the present invention. GIIRS has a total of M = 1690 hyperspectral channels, among which channels 1 to 723 are long-wave channels and channels 724 - 1690 are medium-wave channels. To make the experiment more representative, 11 site samples are used for testing and compared with the results of entropy subtraction. The background field contains two variables, temperature T and absolute humidity q, and has 41 layers in the vertical direction, that is , m = 82; the radiation model uses RTTOV - 4.0, and can be obtained. The simulated background error covariance matrix B is the statistical result under clear sky conditions; the radiation observation value and the observation error variance are provided by the GIIRS imager. The parameter θ of the alternative channel space is set to , where the initial value of k is 0, and when is empty, k = k + 1.

[0094] Figure 2 are the channels selected by the method of the present invention (right figure) and entropy subtraction (left figure) in the 11-sample experiment and the corresponding simulated brightness temperature errors. Figure 2 It shows that there are very significant differences in the channels selected by the present invention and entropy subtraction. Entropy subtraction does not select any channels in three channel intervals, especially in the medium-wave channels of 1350 - 1690. The common point is that the channels selected in the medium-wave channels (to the right of the green line) are all channels with relatively small simulated errors.

[0095] Since the present invention uses the rate of decrease of the value function as the standard for channel selection, Figure 3The relative value of the display value function fluctuates upward between 1 - 100 and 150 - 300 channels. However, after reaching the peak, the relative value of the value function steadily decreases and tends to converge, with the maximum decrease of the value function reaching 0.75. In contrast, the value function obtained by the entropy subtraction method does not decrease significantly and is concentrated around 0.1. This is because the entropy subtraction method only obtains the combination with the smallest variance among numerous channel combinations that has the highest probability. When this "highest probability value" is relatively small, for most samples, the selected channels do not meet the conditions for the value function to decrease significantly.

[0096] Figure 4 It is shown that both channel selection algorithms can make the entropy reduction (Entropy Reduction, i.e., the E value in formula (5)) monotonically decrease as the number of channels increases. Although the entropy subtraction method is better than the result of the present invention, the advantage is less than 0.1.

[0097] The significance of the present invention lies in:

[0098] 1) Variational assimilation obtains the analysis solution by solving the minimum value of the value function. The method proposed by the present invention forms a satellite channel combination by selecting those channels that can cause the largest decrease in the value function. Therefore, the present invention is of great significance for the variational assimilation of satellite data.

[0099] 2) The present invention takes into account real - time satellite observation data and analyzes it according to specific weather conditions. Therefore, it can dynamically select channels based on the quality of observation data, etc., and is applicable to the satellite channel selection before data assimilation in various weather conditions. This is superior to the satellite channel selection method based on statistical concepts. For example, the entropy subtraction method only considers the information of the background field, etc.

[0100] 3) The method of the present invention has effectively solved the problem of inverse matrix calculation in covariance matrix update, and its computational complexity is , so it can be used in combination with other satellite channel selection algorithms. For example, the entropy subtraction method is used before (after) the 50th channel, and the method of the present invention is used after (before) that, so that the value function can decrease rapidly.

[0101] The satellite hyperspectral channel selection device provided by the present invention will be described below. The satellite hyperspectral channel selection device described below can be mutually corresponding and referenced with the satellite hyperspectral channel selection method described above.

[0102] As Figure 5 shown, the device includes a first channel selection module 501, an alternative channel screening module 502, an nth channel selection module 503, and a weight inverse matrix calculation module 504, where:

[0103] The first channel selection module 501 is used to determine the satellite hyperspectral channel set For each of the M channels, the decline rate of the value function corresponding to each channel is calculated, and the channel with the largest decline rate is selected as the first selected channel.

[0104] The alternative channel screening module 502 is used to construct a subset of alternative channels for screening the nth channel , which contains atmospheric vertical structure information that is significantly different from the previously selected channels, where n > 1 and is an integer.

[0105] The nth channel selection module 503 is used to combine each channel in the subset with the first n - 1 previously selected channels respectively, calculate the decline rate of the value function corresponding to each combination using the recursive algorithm of the invention, and select the channel corresponding to the combination with the largest decline rate as the nth selected channel.

[0106] The weight inverse matrix calculation module 504 calculates the n×n - dimensional weight inverse matrix according to the Jacobin vector corresponding to the nth selected channel and the known (n - 1)×(n - 1) - dimensional weight inverse matrix using a recursive formula, reducing the computational complexity of solving the inverse matrix.

[0107] In this embodiment, for the variational assimilation problem of satellite data, the contribution of each channel to the value function is directly estimated, and a combination of satellite channels that can maximize the decline of the value function is selected; considering real - time satellite observation information, specific weather conditions are analyzed specifically, and channels are dynamically selected based on observation data, which is applicable to satellite channel selection in various weather conditions, thereby improving the reliability and accuracy of channel selection.

[0108] Figure 6 An example of the physical structure diagram of an electronic device is shown as Figure 6 shown. The electronic device may include: a processor 610, a communication interface 620, a memory 630, and a communication bus 640. Among them, the processor 610, the communication interface 620, and the memory 630 complete mutual communication through the communication bus 640. The processor 610 can call the logical instructions in the memory 630 to execute the satellite hyperspectral channel selection method. The functions of this device include: 1) Reading hyperspectral satellite observation data, radiation mode simulation data, background covariance matrix information, etc.; 2) Invoking the first channel selection module 501, the alternative channel screening module 502, the nth channel selection module 503, and the weight inverse matrix calculation module 504 of the computer program to implement parallel operations of the hyperspectral channel selection algorithm; 3) Storing the hyperspectral channel selection results on a medium.

[0109] In addition, when the logic instructions in the above-mentioned memory 630 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0110] On the other hand, the present invention also provides a computer program product based on the Fortran language. The computer program product includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, it can execute the above-mentioned satellite hyperspectral channel selection method. This method sequentially calls the first channel selection module 501, the alternative channel screening module 502, the nth channel selection module 503, and the weight inverse matrix calculation module 504, then adjusts n = n + 1, and then returns to call the alternative channel screening module 502 for cycling until n is equal to the preset value and the cycling stops.

[0111] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0112] Through the description of the above embodiments, those skilled in the art can clearly understand each embodiment. It can be realized by means of software plus a necessary general hardware platform, and of course, it can also be realized by hardware. Based on such an understanding, the technical solution, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disks, optical discs, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments.

[0113] 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 them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for selecting satellite hyperspectral channels, comprising: Calculating the decline rate of the value function corresponding to each channel in the satellite hyperspectral channel set, and taking the channel with the largest decline rate as the first selected channel; Performing a linear transformation on the Jacobin vectors of the unselected channels and the selected channels in the satellite hyperspectral channel set using the background error covariance matrix to obtain vectors, calculating the angles between the vectors of the unselected channels and the selected channels, and screening out a subset of alternative channels from the unselected channels by setting an angle range, so that the atmospheric vertical structure information observed by the channels in the subset of alternative channels is significantly different from that of the selected channels; Combining each channel in the subset of alternative channels with n - 1 selected channels, where n ≥ 2, and calculating the decline rate of the value function of each combination using a recursive algorithm according to the (n - 1)×(n - 1) weight inverse matrix of the selected channels, and taking the channel corresponding to the combination with the largest decline rate as the nth selected channel; After obtaining the nth selected channel, using a recursive algorithm to calculate the n×n weight inverse matrix of the n selected channels for continuing to select the next channel.

2. The satellite hyperspectral channel selection method according to claim 1, wherein Calculate the set of satellite hyperspectral observation channels using the following formula The decline rate of the value function corresponding to a total of M channels ； ； ； ； ； ； ； wherein, are respectively the radiation observation value, radiation simulation value, Jacobin vector of radiation simulation and observation error variance of the j-th channel in , where represents the analytical solution of the atmospheric profile when using 1 channel; is its background vector, the error covariance matrix is B, which is obtained by statistical analysis of historical data; the subscript m is the vertical layer number of the atmospheric profile; represents the descent rate of the cost function when there is only one channel; the superscripts T and respectively represent matrix transpose and diagonal matrix.

3. The satellite hyperspectral channel selection method according to claim 2, wherein A vector is obtained by linearly transforming the Jacobin vectors of the unselected channels and the selected channels in the set of satellite hyperspectral channels with the background error covariance matrix through the following formula, and the included angle between the vectors of the unselected channels and the selected channels is calculated. An alternative channel subset is screened out from the unselected channels by setting the included angle range such that the atmospheric vertical structure information observed by the channels in ； ； ； in, Represents the angle between the linearly transformed Jacobin vectors of the j-th unselected channel and the i-th selected channel; The candidate channel subsets The radiation observation value, radiation simulation value, radiation simulation Jacobin vector and observation error variance of the j-th channel in; Represents the Jacobin vector of the i-th channel of the selected channel; is the adjustable angle increment parameter, Indicates a subset of candidate channels The number of channels in When it is an empty set, by adding Value to reconstruct the subcollection .

4. The satellite hyperspectral channel selection method according to claim 3, wherein Combining each channel in the alternative channel subset with n - 1 selected channels through the following formula, and calculating the descent rate of the value function of each combination using a recursive algorithm according to the (n - 1)×(n - 1) weight inverse matrix of the selected channels: ； ； ； ； ； ； ； ； ； ； ； Among them, are the radiance observation value, radiance simulation value, Jacobian vector of radiance simulation, and observation error variance of the j-th channel in the alternative channel subset respectively; denotes the subset the number of channels in; , where denotes the analytical solution of the atmospheric profile when using n channels, n > 1, is its background vector, and the error covariance matrix is B; the subscript m is the vertical number of layers of the atmospheric profile; H nm is the Jacobian matrix of the radiance simulation with respect to ; denotes the observation error variance of n channels and is a diagonal matrix; denotes the rate of decrease of the cost function when using n channels.

5. The satellite hyperspectral channel selection method according to claim 4, wherein Inverse weight matrix of n selected channels According to the one corresponding to the nth selected channel and Calculate through the following recurrence formula: 。 6. A device for selecting satellite hyperspectral channels, characterized by including: A first channel selection module for calculating the decline rate of the value function corresponding to each channel in the satellite hyperspectral channel set, and taking the channel with the largest decline rate as the first selected channel; An alternative channel screening module for performing a linear transformation on the Jacobin vectors of the unselected channels and the selected channels in the satellite hyperspectral channel set using the background error covariance matrix to obtain vectors, calculating the angles between the vectors of the unselected channels and the selected channels, and screening out a subset of alternative channels from the unselected channels by setting an angle range, so that the atmospheric vertical structure information observed by the channels in the subset of alternative channels is significantly different from that of the selected channels; An nth channel selection module for combining each channel in the subset of alternative channels with n - 1 selected channels, where n ≥ 2, and calculating the decline rate of the value function of each combination using a recursive algorithm according to the (n - 1)×(n - 1) weight inverse matrix of the selected channels, and taking the channel corresponding to the combination with the largest decline rate as the nth selected channel; A weight inverse matrix calculation module for, after obtaining the nth selected channel, using a recursive algorithm to calculate the n×n weight inverse matrix of the n selected channels for continuing to select the next channel.

7. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein, When the processor executes the program, it implements the satellite hyperspectral channel selection method according to any one of claims 1 to 5.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the satellite hyperspectral channel selection method according to any one of claims 1 to 5.

9. A computer program product, comprising a computer program, characterized in that: When the computer program is executed by the processor, it implements the satellite hyperspectral channel selection method according to any one of claims 1 to 5.

Images