CN113252586B

CN113252586B - Hyperspectral image reconstruction method, terminal equipment and computer readable storage medium

Info

Publication number: CN113252586B
Application number: CN202110467817.XA
Authority: CN
Inventors: 梁正平; 王志强; 林万鹏
Original assignee: Shenzhen University
Current assignee: Shenzhen University
Priority date: 2021-04-28
Filing date: 2021-04-28
Publication date: 2023-04-28
Anticipated expiration: 2041-04-28
Also published as: CN113252586A

Abstract

The application is applicable to the field of hyperspectral image reconstruction, and provides a hyperspectral image reconstruction method, a hyperspectral image reconstruction device, terminal equipment and a computer readable storage medium, wherein the hyperspectral image reconstruction method comprises the following steps: establishing an objective function in a decision space according to an objective variable of the hyperspectral image reconstruction task; establishing an initial population in a decision space; estimating the curvature of the pareto front according to the initial population; generating a first reference vector for matching the pareto front surface according to the curvature; and performing environment selection processing on the initial population according to the curvature and the first reference vector to obtain the elite population. By the method, the matching degree of the reference vector and the real pareto front surface of the problem to be solved can be improved, and then the multi-objective optimization result is perfected, so that the comprehensive performance of hyperspectral image reconstruction is improved.

Description

Hyperspectral image reconstruction method, terminal equipment and computer readable storage medium

Technical Field

The application belongs to the technical field of hyperspectral image reconstruction, and particularly relates to a hyperspectral image reconstruction method, a hyperspectral image reconstruction device, terminal equipment and a computer readable storage medium.

Background

A spectral image with a spectral resolution in the order of 10 lambda is called a hyperspectral image (Hyperspectral Image). The spectral imaging technology is based on the imaging technology, and utilizes a light splitting element, such as a prism, a grating and the like to acquire image information of an object in each wave band. The spectral information of the object can be extracted from the image information, so that the spectral curve and the image of the target can be obtained. The spectral curve can be used for detecting the components of the object, the image information provides the characteristics of the appearance and the like of the object, and the combination of the two provides great convenience for analyzing the biochemical characteristics of the target object. With the development of spectral imaging technology, the spectral resolution of the system is higher and the system is smaller, and the system is developed from a chromatic dispersion type imaging spectrometer which is huge and needs fine scanning to a portable device which does not need scanning. According to different scanning wave bands, the imaging system can be divided into multispectral, hyperspectral and hyperspectral imaging systems, and the wave bands of the imaging system are respectively 10-20, 100-200 and 1000-2000nm.

A Liquid Crystal Tunable Filter (LCTF) imaging system adopts a structure of single-sided array CCD and liquid crystal tunable filter, LCTF is a novel optical device manufactured according to the electric control birefringence effect of liquid crystal and the interference principle of polarized light, and has the advantages of narrow bandwidth, low power consumption, wide tuning range, low driving voltage, simple structure, wide angle of view, large aperture, no moving parts and the like as a filter. LCTF is a device that changes the transmittance of a liquid crystal filter for each band of light by changing the voltage, so that one voltage corresponds to one transmittance curve, the horizontal axis is the band, and the vertical axis is the transmittance. Therefore, the spectrum information can be compressed, and then a more complete hyperspectral image with larger information content can be reconstructed in fewer measurement times (one voltage corresponds to one measurement) (sampling times and photographing times) by using a compressed sensing theory. Therefore, we need to find a combination of as few measurement times as possible to reconstruct the hyperspectral image as complete as possible, and the decomposition-based multi-objective evolutionary algorithm can provide a novel and effective solution.

Multi-objective optimization refers to the process of finding the optimal solution for multiple objectives in one feasible domain. A difficulty with multi-objective optimization, as compared to single-objective optimization, is that it focuses on not the optimal solution of a single objective, but the equilibrium relationship between the optimal solutions of multiple objectives, respectively. The result of the multi-objective optimization is a set, wherein a plurality of groups of optimal solutions are included, and the values of objective functions corresponding to the optimal solutions form the pareto front. The evolutionary algorithm is used as a heuristic algorithm derived from the biological evolutionary idea, has the characteristics of high robustness and wide adaptability, and provides a novel and effective solving idea for solving the multi-objective optimization problem. In the prior art, the evolution algorithm based on decomposition has a great competitive advantage, and the core idea is to divide a target space by using a group of reference vectors, so that the multi-target optimization problem is decomposed into a plurality of single-target optimization problems which can be solved independently. The performance of the algorithm depends on the matching degree of the adopted reference vector and the real pareto front of the problem to be solved to a great extent, and the reference vector generated by the existing method cannot be effectively matched with the real pareto front, so that the optimization result is poor.

Disclosure of Invention

The embodiment of the application provides a hyperspectral image reconstruction method, a hyperspectral image reconstruction device, terminal equipment and a computer readable storage medium, which can improve the matching degree of a reference vector and a real pareto front surface of a problem to be solved, further perfect a multi-objective optimization result and improve the comprehensive performance of hyperspectral image reconstruction.

In a first aspect, an embodiment of the present application provides a hyperspectral image reconstruction method, including:

establishing an objective function in a decision space according to an objective variable of the hyperspectral image reconstruction task;

establishing an initial population in the decision space, each individual in the initial population representing a set of initial solutions of decision variables of the hyperspectral image reconstruction task;

estimating the curvature of the pareto front according to the initial population;

generating a first reference vector for matching the pareto front according to the curvature;

and carrying out environment selection processing on the initial population according to the curvature and the first reference vector to obtain an elite population, wherein individuals in the elite population represent a group of pareto optimal solutions of the decision variables of the hyperspectral image reconstruction task.

In the embodiment of the application, the curvature of the pareto front surface is estimated according to the initial population, and then a first reference vector for matching the pareto front surface is generated according to the estimated curvature; because the estimated curvature can more accurately reflect the shape of the real pareto front surface, the first reference vector generated according to the estimated curvature can more accurately match the real pareto front surface; and finally, carrying out environment selection processing on the initial population according to the estimated curvature and the first reference vector to obtain the elite population. By the method, the matching degree of the reference vector and the real pareto front surface of the problem to be solved can be effectively improved, and then the multi-objective optimization result is perfected, so that the comprehensive performance of hyperspectral image reconstruction is improved.

In a possible implementation manner of the first aspect, the generating a first reference vector for matching the pareto front according to the curvature includes:

calculating the number of first sampling points according to the target dimension and the number of individuals of the initial population, wherein the target dimension represents the number of target functions;

if the number of the first sampling points is greater than or equal to the target dimension, generating the first reference vector according to a first preset method, wherein the first preset method is a method for generating a vector according to outer layer data, and the outer layer data is data on an edge line of the pareto front surface calculated according to the curvature;

If the number of the first sampling points is smaller than the target dimension, generating the first reference vector according to a second preset method, wherein the second preset method is a method for generating a vector according to outer layer data and inner layer data, and the inner layer data is data on a non-edge line of the pareto front surface calculated according to the curvature.

In a possible implementation manner of the first aspect, the generating the first reference vector according to a first preset method includes:

sampling data on the arc segments corresponding to the curvatures according to the number of the first sampling points to obtain first sampling data;

generating a second reference vector according to the first sampling data;

the second reference vector is determined as the first reference vector.

In a possible implementation manner of the first aspect, the generating the first reference vector according to the second preset method includes:

calculating the number of second sampling points according to the target dimension, the number of individuals of the initial population and the number of first sampling points;

sampling data on the arc segment corresponding to the curvature according to the number of the second sampling points to obtain second sampling data;

Generating a third reference vector according to the second sampling data;

the first reference vector is determined from the second reference vector and the third reference vector.

In a possible implementation manner of the first aspect, the performing an environmental selection process on the initial population according to the curvature and the first reference vector to obtain an elite population includes:

acquiring a cross variation population of the initial population, wherein the cross variation population comprises a parent population and a offspring population of the initial population;

selecting a candidate population from the cross variant population according to the curvature and the first reference vector;

if the number of individuals of the candidate population is greater than the number of individuals of the crossed variant population, selecting the elite population meeting a second preset condition from the candidate population;

and if the number of individuals of the candidate population is smaller than or equal to the number of individuals of the crossed variant population, determining the elite population according to the candidate population and the curvature.

In a possible implementation manner of the first aspect, the selecting a candidate population from the cross variant population according to the curvature and the first reference vector includes:

Screening a fourth reference vector from the first reference vector that is correlated with non-dominant solutions in the cross variant population;

for each fourth reference vector, if the curvature is within a preset numerical range, calculating an aggregate function value between each non-dominant solution in the cross variation population and the fourth reference vector based on a first aggregate function;

if the curvature is not in the preset numerical range, calculating an aggregation function value between each non-dominant solution in the cross variation population and the fourth reference vector based on a second aggregation function;

and adding non-dominant solutions in the cross variation population corresponding to the minimum value in the aggregation function value to the candidate population.

In a possible implementation manner of the first aspect, the determining the elite population according to the candidate population and the curvature if the number of individuals of the candidate population is less than or equal to the number of individuals of the cross variant population includes:

if the number of individuals of the candidate population is smaller than the number of individuals of the cross variation population, respectively calculating the minimum distance between each non-dominant solution in the cross variation population and the candidate population;

Obtaining a fifth reference vector, wherein the fifth reference vector is a non-dominant solution corresponding to the maximum value in the minimum distance in the cross variation population;

screening target individuals from non-dominant solutions of the cross variant population according to the curvature and the fifth reference vector;

adding the target individual to the candidate population and deleting the target individual in the cross variant population.

In a possible implementation manner of the first aspect, after performing an environmental selection process on the initial population according to the curvature and the first reference vector to obtain an elite population, the method includes:

and continuing to perform multi-objective optimization processing on the elite population until the preset evolution times are met.

In a second aspect, embodiments of the present application provide a hyperspectral image reconstruction apparatus, including:

the modeling unit is used for establishing an objective function in the decision space according to the objective variable of the hyperspectral image reconstruction task;

a population acquisition unit for establishing an initial population in the decision space, each individual in the initial population representing a set of initial solutions of decision variables of the hyperspectral image reconstruction task;

The curvature estimation unit is used for estimating the curvature of the pareto front surface according to the initial population;

a vector generation unit, configured to generate a first reference vector for matching the pareto front surface according to the curvature;

and the environment selection unit is used for carrying out environment selection processing on the initial population according to the curvature and the first reference vector to obtain an elite population, and individuals in the elite population represent a group of pareto optimal solutions of the decision variables of the hyperspectral image reconstruction task.

In a third aspect, an embodiment of the present application provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor implements the hyperspectral image reconstruction method as in any one of the first aspects when executing the computer program.

In a fourth aspect, embodiments of the present application provide a computer readable storage medium, where a computer program is stored, where the computer program is executed by a processor to implement a method for reconstructing a hyperspectral image according to any of the first aspects.

In a fifth aspect, embodiments of the present application provide a computer program product, which, when run on a terminal device, causes the terminal device to perform the hyperspectral image reconstruction method as described in any one of the first aspects above.

It will be appreciated that the advantages of the second to fifth aspects may be found in the relevant description of the first aspect, and are not described here again.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required for the embodiments or the description of the prior art will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a schematic flow chart of a hyperspectral image reconstruction method according to an embodiment of the present application;

FIG. 2 is a schematic diagram of a tiered harvest point provided in an embodiment of the present application;

FIG. 3 is a schematic diagram of a reference vector generation process provided by an embodiment of the present application;

FIG. 4 is a schematic illustration of an environment selection provided by an embodiment of the present application;

Fig. 5 is a block diagram of a hyperspectral image reconstruction apparatus provided in an embodiment of the present application;

fig. 6 is a schematic structural diagram of a terminal device provided in an embodiment of the present application.

Detailed Description

In the following description, for purposes of explanation and not limitation, specific details are set forth, such as particular system configurations, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

It should be understood that the terms "comprises" and/or "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.

As used in this specification and the appended claims, the term "if" may be construed as "when..once" or "in response to a determination" or "in response to detection" depending on the context.

In addition, in the description of the present application and the appended claims, the terms "first," "second," "third," and the like are used merely to distinguish between descriptions and are not to be construed as indicating or implying relative importance.

Reference in the specification to "one embodiment" or "some embodiments" or the like means that a particular feature, structure, or characteristic described in connection with the embodiment is included in one or more embodiments of the application. Thus, appearances of the phrases "in one embodiment," "in some embodiments," "in other embodiments," and the like in the specification are not necessarily all referring to the same embodiment, but mean "one or more but not all embodiments" unless expressly specified otherwise.

In recent decades, with social progress and technological development, hyperspectral remote sensing plays an increasingly important role in various fields. The hyperspectral imaging technology is a very important part, and compared with the common remote sensing technology, hyperspectral data has the characteristics of large information quantity and high dimension, and because of the characteristics, hyperspectral remote sensing can acquire ground object information more accurately. At the same time, however, these features present some problems in data processing, and how to process data more effectively is a leading edge of current hyperspectral technology research.

Hyperspectral remote sensing technology has been developed over twenty years ago to date, but it still has many problems to be solved. The most prominent aspect is that hyperspectral data, due to the high number of bands it contains, increases the difficulty in processing the data. Moreover, different environmental factors and different remote sensing platforms can cause non-ideal quality of hyperspectral data, and the degradation of data quality further influences the spectrum matching effect. In order to solve the problem that the difficulty in processing data is increased due to too many wave bands, the method is solved by introducing a multi-objective optimization algorithm, the algorithm has the characteristics of high robustness and wide adaptability, and the method can still ensure a better solution effect on the discontinuous and non-tiny complex multi-objective optimization problem, and is very suitable for processing the hyperspectral image reconstruction problem.

The following describes a multi-objective optimization algorithm:

in general, a multi-objective optimization problem (MOP-objective optimization problem) can be expressed in the following mathematical form:

wherein x= (x) ₁ ,x ₂ ,…,x _n ) Is an n-dimensional decision vector in a decision space, and is M objective functions to be solved. The objective of the multi-objective optimization is to find N groups of optimal solutions from feasible solutions according to a plurality of objective functions and preset constraint conditions, and the numerical values of the objective functions corresponding to the N groups of optimal solutions form the pareto front.

When M is greater than 3, this is also known as the super Multi-objective optimization problem (Many-objective optimization problems, maOPs). The evolution algorithm based on decomposition has a larger competitive advantage for solving the MaOPs, and has been widely applied to solving various types of MaOPs. The core of the decomposition evolution algorithm is to divide a target space by utilizing a reference vector, decompose a single super-multi-target optimization problem into a plurality of single-target optimization problems, and then solve the single-target optimization problem. A key factor affecting the performance of such algorithms is the degree of matching of the reference vector employed with the real Pareto Front (PF) of the problem to be solved.

Common reference vector generation methods in decomposition-based super multi-objective evolutionary algorithm (many-objective evolutionary algorithm based on decomposition, maOEA/D) include three types, NBI, K-layer, and Mixture Uniform Design (MUD). Among them, NBI is the most widely used reference vector generation method, and is very suitable for the treatment of linear PF optimization problem. The K-layer method has a good effect on the optimization problem of the concave PF. The MUD method is advantageous in that any number of reference vectors are supported, but the diversity of the generated reference vectors is not as good as that of NBI and K-layer. In general, the three types of reference vector generation methods are respectively suitable for processing some specific regular type PF optimization problems, and for the irregular type PF optimization problems such as disconnection, degradation, preference and the like, good effects are difficult to obtain due to uneven distribution of the generated reference vectors on the PF.

To support the handling of the different shape PF optimization problem, existing MaOEA/D-type algorithms typically introduce strategies for adaptively adjusting reference vectors during population evolution. According to the difference of heuristic information sources used in the adjustment process, the method can be divided into random reference vectors, fitting adaptive reference vectors, adaptive reference vectors guided by local population, adaptive reference vectors guided by local archiving, adaptive reference vectors based on adjacent reference vectors, adaptive reference vectors based on preference and the like. Although the above reference vector adjustment strategy can achieve a better effect in solving various irregular PF optimization problems, it is easy to cause performance degradation in handling regular PF optimization problems.

In order to solve the problem of PF optimization of different shapes at the same time and better balance the diversity and convergence of populations, the embodiment of the application provides a hyperspectral image reconstruction method.

In the hyperspectral image reconstruction method provided by the application, firstly, an objective function in a decision space is established according to an objective variable of a hyperspectral image reconstruction task, and the decision variable is determined.

For example, the voltage value may be determined as a decision variable. The corresponding multi-objective optimization problem is then: the most suitable at least 2 voltage values are selected from the plurality of voltage values. At least 2 of the number of voltages, sparsity, euclidean distance, spectral angle, and matrix coherence may be taken as target variables. Specific:

1. Number of voltages: one voltage represents a photo sample measurement, each spectrum can have different transmittance at each specific voltage, the smaller the number of voltages, the shorter the total measurement time, and more information can be compressed.

2. Sparseness: the smaller the sparsity of the sparse coefficient of the original complete spectrum signal, the shorter the image reconstruction time.

3. Euclidean distance: the smaller the Euclidean distance between the spectrum signal in the reconstructed image and the spectrum signal in the original image, the better the reconstruction effect is, and the smaller the error is.

4. Spectral angle: the smaller the included angle between the spectrum signal in the reconstructed image and the spectrum signal in the original image, the better the reconstruction effect is, and the smaller the error is.

5. Matrix coherence: the coherence between the measurement matrix and the sparse matrix is smaller, which indicates that the reconstruction effect is better and more information can be compressed.

The following describes an algorithm framework of the hyperspectral image reconstruction method provided in the embodiment of the present application, as shown below.

Algorithm 1: algorithm framework of hyperspectral image reconstruction method

1-1 initial population Q (number of individuals N)

1-2 while does not reach the preset number of evolutions do

1-3 estimating the curvature p of the PF based on the current population Q

1-4 generating a first reference vector W from the curvature p

1-5 Generation of progeny O using crossover variation

1-6 Environment selection of Q.u.O based on p and W

1-7:end while

1-8, output elite population P (number of individuals N)

Algorithm 1 described above is described below. Referring to fig. 1, a schematic flow chart of a hyperspectral image reconstruction method according to an embodiment of the present application is provided, and the method may include the following steps, by way of example and not limitation.

S101, obtaining an initial population.

Wherein each individual in the initial population represents a set of initial solutions for decision variables of the hyperspectral image reconstruction task. In the model shown in the above formula (1), each set of initial solutions includes n solutions, where n is the number of target variables of the problem to be solved.

The hyperspectral image reconstruction provided by the embodiment of the application is based on an evolutionary algorithm, namely, an initial population is evolved to obtain an elite population after evolution. In the first evolution process, steps include S102-S106. The details are as follows.

S102, estimating the curvature of the pareto front surface according to the initial population.

This corresponds to steps 1-3 in algorithm 1 above. In one embodiment, the curvature estimation method may include:

I. normalization

Since scaling of the PF affects the estimation of the curvature p, the generated first reference vector does not match well the shape of the real PF. For this purpose, the initial population is normalized before estimating the curvature of the PF. In the embodiment of the application, the minimum value of the non-dominant layer of the initial population Q on each dimension component is taken as the corresponding component of the ideal point z. The maximum value of the corner solution on each dimension component is taken as the corresponding component of the range point znad. The corner solution is an individual nearest to each direction axis (coordinate axis in the coordinate system where the pareto front is located), and the calculation formula of the ith dimension component value is as follows:

Where x is a target vector corresponding to the corner solution, e is a unit direction vector of each axis, and dist is a Euclidean distance from x to the unit direction vector e of the axis.

The above dimension refers to the number of objective functions in the hyperspectral image reconstruction problem. For example: when there are 2 objective functions, there are 2 dimensions, and the pareto front is a curve. When there are 3 objective functions, there are 3 dimensions, and the pareto front surface is a curved surface.

After the ideal point z and the range point znad are calculated, the formula for normalizing each dimension of x is as follows:

II. Estimating curvature

The estimation of curvature is easily misleading due to the poor quality of the dominant individuals in the population. For this purpose, the existing 2REA method is used herein to select only non-dominant individuals in the initial population and based on their L _p The distance is used to adaptively estimate the curvature p. Wherein L is _p The calculation formula of the distance is as follows:

the specific process of curvature estimation is as follows: the p value is limited in a proper value area, sampling is carried out based on a certain interval, and then the p value is calculated in the value areaL of all non-dominant individuals at each value _p Standard deviation corresponding to the distance. The closer the curved surface corresponding to the selected p value is to the approximate PF formed by the non-dominant individual, the smaller the corresponding standard deviation is. Therefore, the p-value curved surface with the smallest standard deviation is closest to the approximate PF formed by non-dominant individuals, so that the curved surface corresponding to the p-value can be used for fitting the initial population, and the p-value is used as the predicted value of the PF curvature corresponding to the initial population.

And S103, generating a first reference vector for matching the pareto front surface according to the curvature.

Step S103 corresponds to steps 1-4 of algorithm 1 above. The specific method for generating the first reference vector can be referred to as algorithm 2 described in the following embodiments, and will not be described herein.

S104, performing environment selection processing on the initial population according to the curvature and the first reference vector to obtain elite population.

Each individual in the elite population represents a set of pareto optimal solutions for decision variables for the task of hyperspectral image reconstruction.

Step S104 corresponds to steps 1-5 and 1-6 in algorithm 1 above. The specific environment selection processing method can be referred to as algorithm 3 described in the following embodiments, and will not be described herein.

S105, if the current evolution times reach the preset evolution times, determining the elite population obtained at present as the final elite population, and stopping evolution.

And S106, if the current evolution times do not reach the preset evolution times, continuing to perform multi-objective optimization processing on the elite population obtained currently.

Here, continuing to perform the multi-objective optimization process on the elite population currently obtained means that the steps S102 to S106 described above are performed on the elite population, specifically:

re-estimating the curvature of the pareto front according to elite population; regenerating a first reference vector for matching the pareto front according to the re-estimated curvature; performing environment selection processing on the elite population according to the re-estimated curvature and the regenerated first reference vector to obtain a new elite population; if the current evolution times reach the preset evolution times, determining the new elite population obtained currently as a final elite population, and stopping evolution; if the current evolution times do not reach the preset evolution times, continuing to perform multi-objective optimization processing on the new elite population obtained currently.

In the embodiment of the application, the curvature of the pareto front is estimated according to the current population, then a reference vector for matching the pareto front is generated according to the estimated curvature, and the curvature is continuously estimated again in the evolution process and the reference vector is adjusted, so that the generated reference vector gradually matches the shape of the real pareto front. By the method, the matching degree of the reference vector and the real pareto front surface of the problem to be solved can be effectively improved, and then the multi-objective optimization result is improved; meanwhile, the self-adaptability of the method can be improved, and the generated reference vector can be self-adaptively matched with pareto front faces in various shapes.

In one embodiment, an algorithmic framework for generating a first reference vector from curvature in S103 is shown below.

And 2. Generating a reference vector.

2-1 calculating H, H 'using equations (5) and (6), respectively'

2-2, uniformly dividing an arc line segment with the curvature p into H parts by a formula (7), and calculating a corresponding { t } ₀ ,t ₁ ,…,t _H }

2-3 according to { t ] ₀ ,t ₁ ,…,t _H Building a second reference vector W by sum of equations (8), (9) ₁

2-4:if H≥M then

W＝W ₁

2-5:else

2-6, uniformly dividing an arc line segment with curvature p into H parts by a formula (7), and calculating a corresponding { t } ₀ ,t ₁ ,…,t _H }

2-7 according to { t ] ₀ ,t ₁ ,…,t _H Building a third reference vector W by sum of equations (8), (9) ₂

2-8W is calculated using equation (10) ₂ Scaling 1/2 to the inside

2-9:W＝W ₁ ∪W ₂

2-10:end if

2-11 outputting the first reference vector W

The core idea of algorithm 2 is to calculate the number of divisions H of each dimension using equation (5) based on the target dimension M and the number of individuals N of the initial population. If H is more than or equal to M, a first reference vector is generated by using a first preset method (such as a single-layer sampling point method); if H < M, a second preset method (such as a double-layer sampling method) is used for generating a first reference vector. Referring to fig. 2, a schematic diagram of a hierarchical sampling point is provided in an embodiment of the present application. Fig. 2 (a) shows a single-layer acquisition point, wherein the connection line between the solid point and the origin of coordinates represents a second reference vector generated according to the first preset method. Fig. 2 (b) shows a double-layer sampling point, wherein a connection line between the solid point of the outer layer and the origin of coordinates represents a second reference vector, and a connection line between the solid point of the inner layer and the origin of coordinates represents a third reference vector generated according to a second preset method. The reason for adopting the layering strategy is that for the high-dimensional problem, the sparse number of the mining points in the inner area is easily caused by using only a single-layer mining point, and the diversity is difficult to ensure.

Algorithm 2 is described below. S103 specifically comprises the following steps:

S301, calculating the number H of first sampling points according to the target dimension M and the number N of individuals of the initial population.

Wherein the target dimension represents the number of target functions.

Calculating the number H of the first sampling points through a formula (5):

s302, if the number H of the first sampling points is greater than or equal to the target dimension M, a first reference vector is generated according to a first preset method.

In the first preset method (i.e., single-layer dot picking method), outer layer data (i.e., data on the edge line of the pareto front calculated from curvature, as shown by solid dots in fig. 2 (a)) is used. Optionally, generating the first reference vector according to a first preset method includes:

sampling data on the arc segments corresponding to the curvature according to the number of the first sampling points to obtain first sampling data; generating a second reference vector according to the first sampling data; the second reference vector is determined as the first reference vector.

Specifically, an arc line segment with curvature p is uniformly divided into H portions, and first sampling data { t } ₀ ,t ₁ ,…,t _H As component values in the dimensions of the second reference vector. Wherein t is ₀ ＝0，t _H =1, the rest t _k (k ε {1,2, …, H-1 }) satisfies the following formula:

a second reference vector is then generated from the first sampled data using equation (8):

Wherein u is _i Representing the i-th second reference vector,

represents u _i Component values in the j-th dimension and satisfy the following formula:

the generated second reference vector is determined as the first reference vector.

S303, if the number H of the first sampling points is smaller than the target dimension M, generating a first reference vector according to a second preset method.

In the second preset method (i.e., the two-layer sampling method), outer layer data (i.e., data on the edge line of the pareto front calculated from the curvature, as shown by solid points of the outer layer in fig. 2 (b)) and inner layer data (i.e., data on the non-edge line of the pareto front calculated from the curvature, as shown by solid points of the inner layer in fig. 2 (b)) are used. Optionally, generating the first reference vector according to a second preset method includes:

calculating the number of second sampling points according to the target dimension, the number of individuals of the initial population and the number of first sampling points; sampling data on the arc segments corresponding to the curvature according to the number of the second sampling points to obtain second sampling data; generating a third reference vector according to the second sampling data; the first reference vector is determined from the second reference vector and the third reference vector.

For the high-dimensional problem of the double-layer sampling points, the arc sections of the inner layer are further uniformly divided into H' parts according to a formula (7), and then the { t }, which is obtained ₀ ,t ₁ ,…,t _H’ Building a reference vector W by sum of equations (8), (9) ₂ . Because the outer layer sampling points of the high-dimensional problem have dense edges and sparse interiors, W can be used for improving the overall diversity of the reference vector ₂ Scaling 1/2 inwardly to supplement the number of inner layer reference vectors. The scaling formula is as follows:

finally, the second reference vector W ₁ And a scaled third reference vector W ₂ And merging to obtain the first reference vector of the high-dimensional problem.

Exemplary, referring to fig. 3, a schematic diagram of a reference vector generation process provided in an embodiment of the present application is shown. Fig. 3 intuitively illustrates a three-dimensional problem processing procedure of the reference vector generation method provided in the embodiment of the present application, where the initial population size N is set to 28, and the estimated curvature p is set to 2. From equation (5), H is found to be 6. Due to H>And M, generating all reference vectors by using single-layer sampling points. FIG. 3 (a) is a diagram of a uniform segmentation arc segment, and { t } is derived according to equation (7) ₀ ,t ₁ ,t ₂ ,t ₃ ,t ₄ ,t ₅ ,t ₆ Respectively {0,0.2588,0 } correspond to.5000,0.7071,08660,0.9659,1}. FIG. 3 (b) shows a process of constructing reference vector coordinates (0, 1), (0,0.2588,0.9659), …, (0.9659,0.2588,0), (1, 0) according to formula (9), where u ¹ 、u ² 、u ³ The components of the reference vector in the first, second and third dimensions, respectively. Fig. 3 (c) shows the reference vectors that are finally generated and uniformly distributed on the curved surface corresponding to the estimated curvature.

In the embodiment of the application, based on the estimated curvature p generated in an adaptive manner, f with different curvatures is used for the true PF with different problems ₁ ^p +f ₂ ^p +...+f _M ^p The method comprises the steps of fitting an=1 curved surface, and generating uniformly distributed reference vectors with good diversity based on the curved surface, so that the treatment of various PF optimization problems can be well supported.

In one embodiment, an algorithm framework for performing an environmental selection process on the initial population according to the curvature and the first reference vector in S104 is as follows.

Algorithm 3 Environment selection

3-1 find corner solutions from across variant populations R to put into S

3-2 normalized Cross variant population R

3-3, eliminating the reference vector without non-dominant decorrelation in the first reference vector W to obtain W ₃

3-4 according to the curvatures p and W ₃ Selection of candidate populations S 'from the crossover variant population R'

3-5:S＝S∪S’，R＝R\S

3-6:if|S|>N then

3-7, eliminating the solution with the maximum local density of |S| -N in S

3-8:P＝S

3-9:end if

3-10:if|S|<N then

3-11 determining elite population P based on curvature P, current candidate population S and current crossover variation population R

3-12:end if

3-13, output elite population P

The core of the algorithm 3 is that an elite population is selected by adaptively selecting an aggregation function based on the estimated curvature and the generated reference vector, and meanwhile, various irregular PF optimization problems are better supported by dynamically adjusting the reference vector.

Algorithm 3 is described below. S104 specifically comprises the following steps:

s401, acquiring a cross variation population of the initial population.

The crossover variation population comprises a parent population and a offspring population of the initial population.

The method of determining the parent and offspring of the initial population may employ existing genetic algorithms. Alternatively, a binary tournament method may be used to select the parent population of the initial population, and a simulated binary crossover method and a polynomial variation method may be used to generate the offspring population of the initial population. The union of the parent and offspring populations is then used as the crossover variation population for the initial population.

And S402, selecting a candidate population S from the crossed variant populations R according to the curvature p and the first reference vector W.

This step corresponds to steps 3-1 through 3-5 in algorithm 3. Specific:

finding corner solutions from the variant cross population R and putting the corner solutions into S; normalizing the variant crossover population R; then, the reference vector W without non-dominant decorrelation is provided to obtain a filtered fourth reference vector W ₃ The method comprises the steps of carrying out a first treatment on the surface of the Based on the curvatures p and W ₃ Selecting a candidate population S' from R; finally, S is updated according to S '(s=s ∈s'), and R is updated according to S after updating (r=r\s).

Because the corner solutions cover a large space and are far apart from each other, in order to ensure the overall diversity of the population, the corner solutions in the population are first taken as default elite solutions. Wherein, according to the curvatures p and W ₃ The algorithm framework for selecting the candidate population S' from R is shown below.

Algorithm 4 curvature p and W ₃ Selection of candidate populations S 'from R'

1:S’＝[]

2:for w∈Wdo

3:if p≥1then

4 calculating the PBI aggregation function value of each non-dominant solution and the reference vector w in R

5:else

6, calculating TCH aggregation function values of each non-dominant solution and the reference vector w in R

7:end if

8, finding out the optimal solution s of the aggregation function value

9:S’＝S’∪s

10:end for

11 outputting candidate population S'

As shown in the algorithm framework above, for each of the fourth reference vectors w: if the curvature p is in a preset numerical range (for example, p is more than or equal to 1), calculating an aggregation function value between each non-dominant solution in the cross variation population R and a fourth reference vector w based on a first aggregation function (for example, a boundary penalty aggregation function PBI); if the curvature p is not within the preset value range (p < 1), calculating an aggregation function value between each non-dominant solution in the cross variation population R and the fourth reference vector w based on a second aggregation function (such as chebyshev aggregation function TCH); and adding the non-dominant solution in the cross variation population R corresponding to the minimum value in the aggregation function value to the candidate population S'.

In algorithm 4, for each valid reference vector (i.e., the fourth reference vector w), a candidate solution is selected from the non-dominant solutions as an environmental choice by optimizing the aggregate function value. For the optimization problem that the PF shape is not convex, the PBI can provide better diversity while ensuring the population convergence, and for the optimization problem that the PF shape is convex, the TCH can better balance the population convergence and diversity. Therefore, the adaptive selection of the aggregation function through the pre-estimated curvature is beneficial to ensuring that elite solutions with better quality can be selected according to effective reference vectors when various PF shapes are fitted to the curved surface, and avoiding the preference existing in the use of only a single aggregation function, thereby effectively balancing the convergence and diversity of the population.

S403, if the number of individuals of the candidate population S is greater than the number of individuals N of the initial population, selecting elite population P meeting a second preset condition from the candidate population S.

This step corresponds to steps 3-7 and 3-8 in algorithm 3.

Alternatively, the first N individuals with the greatest local density may be selected from the candidate population S, and these N individuals may be grouped into elite population P. N is the number of individuals in the initial population.

S404, if the number of individuals of the candidate population S is smaller than or equal to the number of individuals N of the initial population, determining an elite population P according to the candidate population S and the curvature P.

This step corresponds to steps 3-11 in algorithm 3.

Optionally, an algorithmic framework for elite population P is determined from candidate population S and curvature P in S404, as shown below.

Algorithm 5 determining the final elite population based on curvature and the current candidate population

5-1:while|S|<N do

5-2 calculating the minimum distance between each non-dominant solution in R and all solutions in S

5-3 find out the solution l corresponding to the maximum value in the minimum distance in R

5-4 selecting a candidate solution u from the non-dominant solution set of R based on the curvature p and the reference vector l

5-5:S＝S∪u,R＝R\u

5-6 find out the solution c with worst convergence index in R

5-7:R＝R\c

5-8:end while

5-9:P＝S

5-10, output elite population P

As shown in the algorithm framework above, S404 may include the steps of:

If the number of individuals of the candidate population S is smaller than the number of individuals N of the initial population, respectively calculating the minimum distance between each non-dominant solution in the cross variation population R and the candidate population S; obtaining a fifth reference vector l, wherein the fifth reference vector is a non-dominant solution corresponding to the maximum value in the minimum distance in the cross variation population R; screening target individuals u from non-dominant solutions of the cross variant population R according to the curvature p and the fifth reference vector l; target individuals u are added to the candidate population S, target individuals u in the crossover variation population R are deleted, and the solution c with the worst convergence in R is deleted. And sequentially cycling until the number of individuals in the candidate population S is equal to N, and determining the candidate population S at the moment as elite population P.

Wherein, step 5-4 may determine the target individual u using the method in algorithm 4.

To ensure more accurate definition of the minimum distance between solutions for different PF shapes, adaptive calculations are also performed for the distance between solutions based on the estimated curvature. If the PF is fitted into a plane in the curvature estimation process, namely, p=1, the Euclidean distance between the mapping points of the two solutions on the hyperplane is used as the distance between the solutions; if PF is fitted into a concave curved surface, namely p >1, the included angle formed between the two solutions and the origin is used as the distance between the solutions; if PF is fitted to a convex surface, i.e. p <1, the angle between the two solutions and the point of extreme difference znad is taken as the distance between the solutions. For any one of the non-dominant solutions in R, a distance between the non-dominant solution and each of the solutions in S is calculated based on the distance calculation method, and a minimum value in the distances is determined as a minimum distance between the non-dominant solution and S.

The solution l corresponding to the maximum value of the minimum distance is the solution with the best diversity among the non-dominant solutions of the current population R, and is therefore most suitable as a new reference vector to choose elite solutions. In addition, in order to further improve the convergence of the population, at the end of each round of selection, the solution with the worst convergence in the current residual population is deleted. Because the components and the method have the advantages of simplicity, high efficiency, strong searching performance and the like, the components and the method can be adopted as convergence indexes of each solution in the population.

Exemplary, referring to fig. 4, a schematic diagram of an environment selection provided by an embodiment of the present application. Fig. 4 is an example of environment selection in two-dimensional space using algorithm 3 of an embodiment of the present application. The population size N is set to 7, the curve represents the fitted PF curve with curvature p=0.5, v ₁ -v ₇ For the reference vector generated based on algorithm 2, x ₁₁ -x ₁₄ Is a candidate solution for a population, where x ₁ -x ₇ For the first layer non-dominant solution, x ₈ -x ₁₁ Is a second layer non-dominant solution. As can be seen from fig. 4, x ₁ And x ₇ For corner solution, add essence directlyIn the english population. v ₂ And v ₆ No non-dominant solution is associated with it, they are culled from the reference vector. Due to p<1, selecting TCH as an aggregation function to obtain residual reference vectors v respectively ₁ 、v ₃ 、v ₄ 、v ₅ And v ₇ X is corresponding one by one and the aggregation function value is optimal ₁ 、x ₃ 、x ₄ 、x ₆ And x ₇ As newly obtained elite solutions. At this time, the elite population is 5 in size, and the elite solution needs to be selected continuously. In the remaining population, x ₂ And x ₅ Is a non-dominant solution. As can be seen from the figure, compared with x ₂ ，x ₅ The minimum distance from each solution in the current elite population is greater, so x will be ₅ As a new reference vector. It can further be found that the ratio of x ₂ Candidate solution x ₅ TCH value is better than that of new reference vector, and x is calculated as follows ₅ Elite populations were added. Thereafter, the newly selected elite solution x is removed from the population ₅ And the worst convergence candidate solution x ₂ . Continuing to process the remaining population, x can be found ₉ Is elite solution. At this time, the number of elite solutions is equal to the number of individuals of the initial population, and the environment selection is ended.

Compared with the existing method, the hyperspectral image reconstruction method provided by the embodiment of the application has good diversity of the generated reference vector and can be more accurately matched with the pareto front surfaces with various shapes; meanwhile, in the environment selection process, the diversity and the convergence of the PF can be well balanced by adaptively selecting the aggregation function. In addition, the time complexity of the method is low (max { O (MN) ² ) O (DN). Therefore, compared with other algorithms, the hyperspectral image reconstruction method provided by the application has better performance.

It should be understood that the sequence number of each step in the foregoing embodiment does not mean that the execution sequence of each process should be determined by the function and the internal logic of each process, and should not limit the implementation process of the embodiment of the present application in any way.

Corresponding to the hyperspectral image reconstruction method described in the above embodiments, fig. 5 is a block diagram of the hyperspectral image reconstruction apparatus provided in the embodiment of the present application, and for convenience of explanation, only the portions related to the embodiment of the present application are shown.

Referring to fig. 5, the apparatus includes:

a modeling unit 51 for establishing an objective function in the decision space from the objective variables of the hyperspectral image reconstruction task.

A population acquisition unit 52 for acquiring an initial population, each individual of the initial population representing a set of initial solutions of decision variables of the hyperspectral image reconstruction task.

A curvature estimating unit 53 for estimating the curvature of the pareto front surface based on the initial population.

A vector generation unit 54 for generating a first reference vector for matching the pareto front surface according to the curvature.

An environment selection unit 55, configured to perform an environment selection process on the initial population according to the curvature and the first reference vector, to obtain an elite population, where individuals in the elite population represent a set of pareto optimal solutions of the decision variables of the hyperspectral image reconstruction task.

Optionally, the vector generation unit 54 is further configured to:

calculating the number of first sampling points according to the target dimension and the number of individuals of the initial population, wherein the target dimension represents the number of target functions; if the number of the first sampling points is greater than or equal to the target dimension, generating the first reference vector according to a first preset method, wherein the first preset method is a method for generating a vector according to outer layer data, and the outer layer data is data on an edge line of the pareto front surface calculated according to the curvature; if the number of the first sampling points is smaller than the target dimension, generating the first reference vector according to a second preset method, wherein the second preset method is a method for generating a vector according to outer layer data and inner layer data, and the inner layer data is data on a non-edge line of the pareto front surface calculated according to the curvature.

Optionally, the vector generation unit 54 is further configured to:

sampling data on the arc segments corresponding to the curvatures according to the number of the first sampling points to obtain first sampling data; generating a second reference vector according to the first sampling data; the second reference vector is determined as the first reference vector.

Optionally, the vector generation unit 54 is further configured to:

calculating the number of second sampling points according to the target dimension, the number of individuals of the initial population and the number of first sampling points; sampling data on the arc segment corresponding to the curvature according to the number of the second sampling points to obtain second sampling data; generating a third reference vector according to the second sampling data; the first reference vector is determined from the second reference vector and the third reference vector.

Optionally, the environment selection unit 55 is further configured to:

acquiring a cross variation population of the initial population, wherein the cross variation population comprises a parent population and a offspring population of the initial population; selecting a candidate population from the cross variant population according to the curvature and the first reference vector; if the number of individuals of the candidate population is greater than the number of individuals of the crossed variant population, selecting the elite population meeting a second preset condition from the candidate population; and if the number of individuals of the candidate population is smaller than or equal to the number of individuals of the crossed variant population, determining the elite population according to the candidate population and the curvature.

Optionally, the environment selection unit 55 is further configured to:

screening a fourth reference vector from the first reference vector that is correlated with non-dominant solutions in the cross variant population; for each fourth reference vector, if the curvature is within a preset numerical range, calculating an aggregate function value between each non-dominant solution in the cross variation population and the fourth reference vector based on a first aggregate function; if the curvature is not in the preset numerical range, calculating an aggregation function value between each non-dominant solution in the cross variation population and the fourth reference vector based on a second aggregation function; and adding non-dominant solutions in the cross variation population corresponding to the minimum value in the aggregation function value to the candidate population.

Optionally, the environment selection unit 55 is further configured to:

if the number of individuals of the candidate population is smaller than the number of individuals of the cross variation population, respectively calculating the minimum distance between each non-dominant solution in the cross variation population and the candidate population; obtaining a fifth reference vector, wherein the fifth reference vector is a non-dominant solution corresponding to the maximum value in the minimum distance in the cross variation population; screening target individuals from non-dominant solutions of the cross variant population according to the curvature and the fifth reference vector; adding the target individual to the candidate population and deleting the target individual in the cross variant population.

Optionally, the apparatus 5 further comprises:

the number judgment unit 56 is configured to, after performing an environmental selection process on the initial population according to the curvature and the first reference vector, continue performing a multi-objective optimization process on the elite population until a preset number of evolutions is satisfied.

It should be noted that, because the content of information interaction and execution process between the above devices/units is based on the same concept as the method embodiment of the present application, specific functions and technical effects thereof may be referred to in the method embodiment section, and will not be described herein again.

In addition, the hyperspectral image reconstruction device shown in fig. 5 may be a software unit, a hardware unit, or a unit combining soft and hard, which are built in an existing terminal device, may be integrated into the terminal device as an independent pendant, or may exist as an independent terminal device.

It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of the functional units and modules is illustrated, and in practical application, the above-described functional distribution may be performed by different functional units and modules according to needs, i.e. the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-described functions. The functional units and modules in the embodiment may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit, where the integrated units may be implemented in a form of hardware or a form of a software functional unit. In addition, specific names of the functional units and modules are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working process of the units and modules in the above system may refer to the corresponding process in the foregoing method embodiment, which is not described herein again.

Fig. 6 is a schematic structural diagram of a terminal device provided in an embodiment of the present application. As shown in fig. 6, the terminal device 6 of this embodiment includes: at least one processor 60 (only one shown in fig. 6), a memory 61 and a computer program 62 stored in the memory 61 and executable on the at least one processor 60, the processor 60 implementing the steps in any of the various method embodiments described above when executing the computer program 62.

The terminal equipment can be computing equipment such as a desktop computer, a notebook computer, a palm computer, a cloud server and the like. The terminal device may include, but is not limited to, a processor, a memory. It will be appreciated by those skilled in the art that fig. 6 is merely an example of the terminal device 6 and is not meant to be limiting as to the terminal device 6, and may include more or fewer components than shown, or may combine certain components, or different components, such as may also include input-output devices, network access devices, etc.

The processor 60 may be a central processing unit (Central Processing Unit, CPU), the processor 60 may also be other general purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory 61 may in some embodiments be an internal storage unit of the terminal device 6, such as a hard disk or a memory of the terminal device 6. The memory 61 may in other embodiments also be an external storage device of the terminal device 6, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card) or the like, which are provided on the terminal device 6. Further, the memory 61 may also include both an internal storage unit and an external storage device of the terminal device 6. The memory 61 is used for storing an operating system, application programs, boot loader (BootLoader), data, other programs, etc., such as program codes of the computer program. The memory 61 may also be used for temporarily storing data that has been output or is to be output.

Embodiments of the present application also provide a computer readable storage medium storing a computer program which, when executed by a processor, implements steps that may implement the various method embodiments described above.

The present embodiments provide a computer program product which, when run on a terminal device, causes the terminal device to perform steps that enable the respective method embodiments described above to be implemented.

The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the present application implements all or part of the flow of the method of the above embodiments, and may be implemented by a computer program to instruct related hardware, where the computer program may be stored in a computer readable storage medium, where the computer program, when executed by a processor, may implement the steps of each of the method embodiments described above. Wherein the computer program comprises computer program code which may be in source code form, object code form, executable file or some intermediate form etc. The computer readable medium may include at least: any entity or device capable of carrying computer program code to an apparatus/terminal device, recording medium, computer Memory, read-Only Memory (ROM), random access Memory (RAM, random Access Memory), electrical carrier signals, telecommunications signals, and software distribution media. Such as a U-disk, removable hard disk, magnetic or optical disk, etc. In some jurisdictions, computer readable media may not be electrical carrier signals and telecommunications signals in accordance with legislation and patent practice.

In the foregoing embodiments, the descriptions of the embodiments are emphasized, and in part, not described or illustrated in any particular embodiment, reference is made to the related descriptions of other embodiments.

Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

In the embodiments provided in the present application, it should be understood that the disclosed apparatus/terminal device and method may be implemented in other manners. For example, the apparatus/terminal device embodiments described above are merely illustrative, e.g., the division of the modules or units is merely a logical function division, and there may be additional divisions in actual implementation, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection via interfaces, devices or units, which may be in electrical, mechanical or other forms.

The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

The above embodiments are only for illustrating the technical solution of the present application, and are not limiting; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application, and are intended to be included in the scope of the present application.

Claims

1. A hyperspectral image reconstruction method, characterized in that the hyperspectral image reconstruction method comprises:

performing environment selection processing on the initial population according to the curvature and the first reference vector to obtain an elite population, wherein individuals in the elite population represent a group of pareto optimal solutions of the decision variables of the hyperspectral image reconstruction task;

the generating a first reference vector for matching the pareto front according to the curvature comprises:

If the number of the first sampling points is smaller than the target dimension, generating the first reference vector according to a second preset method, wherein the second preset method is a method for generating a vector according to outer layer data and inner layer data, and the inner layer data is data on a non-edge line of the pareto front surface calculated according to the curvature;

and performing environment selection processing on the initial population according to the curvature and the first reference vector to obtain elite population, wherein the method comprises the following steps:

if the number of individuals of the candidate population is greater than the number of individuals of the crossed variant population, selecting the elite population meeting a second preset condition from the candidate population; wherein a front with the greatest local density is selected from the candidate populationNIndividuals, will do thisNThe individuals make up the elite population,Nthe number of individuals that are the initial population;

if the number of individuals of the candidate population is less than or equal to the number of individuals of the cross variation population, determining the elite population according to the candidate population and the curvature;

The selecting a candidate population from the cross variant population based on the curvature and the first reference vector comprises:

2. The hyperspectral image reconstruction method as claimed in claim 1, wherein the generating the first reference vector according to the first preset method includes:

generating a second reference vector according to the first sampling data;

The second reference vector is determined as the first reference vector.

3. The hyperspectral image reconstruction method as claimed in claim 2, wherein the generating the first reference vector according to the second preset method includes:

generating a third reference vector according to the second sampling data;

4. The method of hyperspectral image reconstruction as claimed in claim 1 wherein said determining said elite population based on said candidate population and said curvature if the number of individuals of said candidate population is less than or equal to the number of individuals of said crossover variation population comprises:

5. The hyperspectral image reconstruction method as claimed in claim 1, wherein after the initial population is subjected to an environmental selection process based on the curvature and the first reference vector, the method comprises:

6. A terminal device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the method according to any of claims 1 to 5 when executing the computer program.

7. A computer readable storage medium storing a computer program, characterized in that the computer program when executed by a processor implements the method according to any one of claims 1 to 5.