Spectral feature extraction methods, terminal equipment and storage media for remote sensing images

Abstract

This invention discloses a method, terminal device, and storage medium for extracting spectral features from remote sensing images. Drawing on the principle of projection pursuit algorithm, it employs a genetic acceleration algorithm to fully mine the features in spectral information, which can be used for quantitative inversion modeling. This invention proposes a computational encoding form, transforming the feature construction process into a spatial vector feature optimization process, thereby realizing genetic and mutation processing, greatly reducing the time required for feature mining and improving the efficiency of feature mining.

Description

Technical Field

[0001] This invention relates to the field of remote sensing technology, and in particular to a method for extracting spectral features from remote sensing images, a terminal device, and a storage medium. Background Technology

[0002] Currently, commonly used remote sensing image feature parameters mainly originate from existing exponential features or band reflectance, without deeply exploring significant feature parameters within the spectral feature space. The prerequisite for establishing an empirical model is obtaining feature parameters significantly correlated with the target parameters from remote sensing images. Because current technologies have not deeply explored significant feature parameters within the spectral feature space, the accuracy of qualitative remote sensing classification remains low. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a method, terminal device and storage medium for extracting spectral features from remote sensing images, which fully exploits the features in spectral information, in order to address the shortcomings of the existing technology.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a method for extracting spectral features from remote sensing images, comprising the following steps:

[0005] S1. Acquire remote sensing images and obtain information B for the band to be selected;

[0006] S2. Construct a repeatable array AL = [x1, x2, x3, ..., x...] m ],{x i |x i ∈[1,b],x i ∈Z}, AL represents the position of b in the corresponding array arrangement in the characteristic expression. i The selected band information, b represents the number of bands in the selected band information B, and Z is a set of integers; where the characteristic expression f is: m represents the number of characteristic expression terms. This indicates the rules of operation before each term, namely +, -, *, / . Simulate whether any of the items are within the scope defined by parentheses "()"; λ i It represents one of the terms in a polynomial formula;

[0007] Construct a repeatable array AS to represent the terms in the characteristic expression at the corresponding array positions. The operator it represents;

[0008] Construct an array AP to set the starting position of the brackets in the feature expression;

[0009] Construct an array AC to represent the λ values ​​in each term at the corresponding array position in the characteristic expression. i The selected coefficients;

[0010] S3. Set m, obtain all combinations of arrays AL, AS, and AP, and initialize array AC.

[0011] S4. Randomly combine arrays AL, AS, AP, and AC without repetition to obtain a random combination set X. Reverse encode the random combination set X to obtain a feature calculation expression set represented by the random combination set X. Obtain the corresponding feature parameter set F based on the feature calculation expression set.

[0012] S5. Preprocess the feature parameter set F to obtain preprocessed feature parameters; calculate the correlation coefficient between the preprocessed feature parameters and the ground point target parameters in the remote sensing image, and arrange the preprocessed feature parameters in ascending order according to the absolute value of the correlation coefficient.

[0013] S6. Extract the preprocessed feature parameters that are in the top M% after sorting and have a correlation coefficient > N. Perform cross-validation between the extracted feature parameters and the ground point target parameters. Calculate the mean relative error in the cross-validation. If the maximum relative error is less than the first set threshold and the mean relative error is less than the second set threshold, output the feature expression and end the process. Otherwise, use the random combination set X as the input of the genetic acceleration algorithm (genetic algorithm) to obtain the updated array AC and return to step S4.

[0014] This invention constructs and encodes feature expressions, and by drawing on the principles of projection pursuit algorithms, transforms the feature construction process into a spatial vector feature optimization process. This enables genetic and mutation processing, significantly reducing the time required for feature mining and fully exploiting features within spectral information. The extracted features can be used for inversion modeling in quantitative remote sensing, providing decision-making information for identifying objects with the same spectrum but different characteristics in land cover classification, thereby improving the accuracy of qualitative remote sensing classification.

[0015] In order to obtain ground point spectral information that can be used for feature mining and to better extract spectral features, the specific implementation process of obtaining the selected band information B in step S1 includes:

[0016] 1) Remove outliers from the remote sensing image data and read the ground point target parameters from the remote sensing image;

[0017] 2) Based on the longitude and latitude information in the ground point target parameters, locate the pixel points in the remote sensing image and obtain the spectral data of each ground point;

[0018] 3) Perform continuous de-statistic processing on the spectral data obtained in step 2);

[0019] 4) Calculate the first coefficient of variation of the ground point target parameters and the second coefficient of variation of the reflectance of the same band as the spectral data after continuous de-systematization. Compare the magnitudes of the first coefficient of variation and the second coefficient of variation. If the first coefficient of variation is greater than a times the second coefficient of variation, retain the spectral data of that band; otherwise, discard that band. a∈(0,1];

[0020] 5) Calculate the correlation coefficient between each band for the band information retained after processing in step 4). When the correlation coefficient between two bands is greater than the third set threshold, remove the band data with the longer wavelength between the two bands to obtain the band information B to be selected.

[0021] In this invention, the third set threshold is 0.7.

[0022] Coefficient of variation (CV) x The calculation formula is: σ x μ represents the standard deviation of dataset X. x This represents the mean of dataset X; dataset X is a set of ground point target parameters or a set of reflectance values ​​for the same band of spectral data after continuous de-systematization.

[0023] To further extract more spectral information, step S5 involves preprocessing the feature parameter set F, specifically including the following steps:

[0024] Remove duplicate feature parameters from the feature parameter set F;

[0025] Calculate the coefficient of variation of the remaining characteristic parameters;

[0026] Feature parameters with a coefficient of variation greater than the fourth set threshold are removed to obtain the preprocessed feature parameter set.

[0027] In step S5, the fourth threshold is set to 15%.

[0028] In this invention, M = 1; N = 0.7; the first set threshold is 20%; and the second set threshold is 10%.

[0029] The present invention also provides a terminal device, including a memory, a processor, and a computer program stored in the memory; the processor executes the computer program to implement the steps of the method described above.

[0030] The present invention also provides a computer-readable storage medium having a computer program / instructions stored thereon; when the computer program / instructions are executed by a processor, they implement the steps of the method described above.

[0031] Compared with the prior art, the beneficial effects of this invention are as follows: This invention draws on the principle of projection pursuit algorithm and adopts genetic acceleration algorithm to fully mine the features in spectral information, which can be used for quantitative inversion modeling; This invention proposes a form of operation encoding, which transforms the feature construction process into a spatial vector feature optimization process, thereby realizing genetic and mutation processing, greatly reducing the time required for feature mining and improving the efficiency of feature mining. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of the method in Embodiment 1 of the present invention;

[0033] Figure 2(a) and Figure 2(b) show the water quality parameter distribution curves at various sampling points in Dongting Lake in Example 1 of the present invention. Figure 2(a) shows the distribution curves of potassium permanganate concentration, total nitrogen concentration, and transparency, while Figure 2(b) shows the distribution curves of total phosphorus concentration and chlorophyll a concentration.

[0034] Figures 3(a) and 3(b) are remote sensing images of the Dongting Lake Basin from Sentinel-2, where Figure 3(a) shows the East Dongting Lake area and Figure 3(b) shows the West Dongting Lake area.

[0035] Figure 4 This is a schematic diagram of water area pixels in the Dongting Lake basin.

[0036] Figures 5(a1) to 5(d2) The distribution of the coefficient of determination and relative error of the cross-validation results of the total phosphorus concentration inversion model is shown in Figure 1 (coefficient of determination) and Figure 2 (relative error). In Figure 5(a1) and Figure 5(a2), the training set:validation set ratio is 9:2; in Figure 5(b1) and Figure 5(b2), the training set:validation set ratio is 8:3; in Figure 5(c1) and Figure 5(c2), the training set:validation set ratio is 7:4; and in Figure 5(d1) and Figure 5(d2), the training set:validation set ratio is 6:5.

[0037] Figure 6(a) and Figure 6(b) are histograms of the cross-linear fitting accuracy distribution of the total phosphorus concentration inversion model, where Figure 6(a) is the histogram of the coefficient of determination distribution and Figure 6(b) is the histogram of the average relative error distribution.

[0038] Figure 7 This is a diagram showing the distribution of residuals from the total phosphorus concentration inversion model. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] In this document, the terms "first," "second," and other similar words are not intended to imply any order, quantity, or importance, but are merely used to distinguish different elements. The terms "one," "a," and other similar words are not intended to indicate the existence of only one of the stated things, but rather that the description pertains to only one of the two stated things, which may include one or more. The terms "comprising," "including," and other similar words are intended to indicate a logical relationship, not a spatial relationship. For example, "A includes B" means that logically B belongs to A, not that spatially B is located inside A. Furthermore, the meanings of the terms "comprising," "including," and other similar words should be considered open-ended, not closed. For example, "A includes B" means that B belongs to A, but B does not necessarily constitute all of A; A may also include other elements such as C, D, and E.

[0041] Example 1

[0042] Embodiment 1 of this invention proposes a spectral feature extraction method for establishing an empirical model for quantitative inversion of optical remote sensing images. The specific process is as follows: Figure 1 As shown, it includes two main parts: data preprocessing and feature mining.

[0043] The data preprocessing work aims to obtain ground point spectral information that can be used in the feature mining module. Its main steps are as follows:

[0044] Step 1: Read the hyperspectral remote sensing image and complete preprocessing work such as radiometric calibration, geometric correction, georegistration, and atmospheric correction (preprocessing can be done using ENVI software). Read the target parameters of the ground point samples and remove outliers from the sampled data based on the actual situation and statistical analysis.

[0045] Step 2: Based on the ground point data filtered in Step 1, combined with the hyperspectral remote sensing image preprocessed in Step 1, and using the latitude and longitude information in the measured ground point data, locate the pixel points in the remote sensing image to obtain the spectral data of each ground point.

[0046] Step 3: Perform continuous de-statistic processing on the ground point spectral data obtained in Step 2 (continuous de-statistic processing can be performed using ENVI software);

[0047] Step 4: Calculate the target parameter T of the ground point samples after the screening in Step 1. x Coefficient of Variation (CV) target Calculate the reflectance S of each surface pixel in the same band after continuous de-scaling in step three. x coefficient of variation CVspect Compare the relative magnitudes of the coefficients of variation of the two, if CV spect >a*CV target If a∈(0,1], then retain the band; otherwise, discard it.

[0048] The formula for calculating the coefficient of variation is as follows:

[0049]

[0050] In the formula σ x μ represents the standard deviation of dataset X. x This represents the mean of dataset X;

[0051] Step 5: Calculate the correlation coefficient ρ between the remaining bands based on the band information retained after processing in Step 4. xy When the correlation coefficient between two band datasets is greater than 0.7, the band data with the longer wavelength is removed to obtain the band information B to be selected.

[0052] The correlation coefficient used is the Pearson correlation coefficient:

[0053]

[0054]

[0055] In the formula, Cov(x,y) represents the covariance of datasets x and y, and μ x Let σ represent the mean of dataset x, n represent the number of elements in dataset x, and σ represent the mean of dataset x. x This represents the standard deviation of dataset X.

[0056] The feature mining part of this embodiment uses computational encoding processing. The corresponding encoding is then processed using projection pursuit and genetic acceleration algorithms for data mining. The feature parameters are then decoded to obtain the feature parameters, and the correlation between the feature parameters and the target parameters is analyzed. The specific steps are as follows:

[0057] Step 1: Encode the algorithm according to the arrangement of the feature expressions from left to right, assuming that the feature expressions conform to the following form:

[0058]

[0059] In the formula, m represents the number of characteristic expression terms. Indicate the rules of operation for "+", "-", "*", and " / " before each term. Simulate whether each item exists within the parentheses "()". Based on this, the feature expression operation is encoded as follows:

[0060] Construct a repeatable array AL = [o1, o2, o3 ... o m ],{oi |o i ∈[1,b],and o i ∈Z}, used to represent the terms in the characteristic expression corresponding to the array arrangement positions, where b i The selected band information, where b represents the number of bands of the selected band information B, and Z is a set of integers;

[0061] Construct a repeatable array AS = [p1, p2, p3 ... p m ],p i ∈[1,2,3,4], used to represent the terms in the corresponding array arrangement positions in the characteristic expression. The operators represented are [1,2,3,4], which correspond to "+, -, *, / " in the formula.

[0062] Construct an array AP = [l1, l2, l3 ... l n ],{l i |l i ∈[1,m-1],and l i ∈Z}, where n is the largest even number less than m, Z is the set of integers, x i +x i+1 ≤m, where i is an odd number, used to set the starting position of the brackets in the characteristic expression;

[0063] Construct an array AC = [k1, k2, k3 ... k m ],k i ∈[0,1], used to represent the λ in each term of the corresponding array arrangement position in the characteristic expression. i The selected coefficients;

[0064] Step 2: Set the size of m in the characteristic expression in Step 1 according to the requirements, enumerate all combinations of integer arrays AL, AS, and AP, and initialize a certain number of coefficient arrays AC;

[0065] Step 3: Take all the forms of AL, AS, AP, and AC enumerated in Step 2 and combine them randomly without repetition to obtain a random combination set X. Decode each combination in the random combination set X in reverse according to the encoding principle in Step 1 to obtain the feature calculation expression set represented by the random combination set X. Calculate the corresponding feature parameter set F according to the expression.

[0066] Step 4: Remove duplicate feature parameters from the feature parameter set F; calculate each feature parameter F in the remaining feature parameter set. x The coefficient of variation (CV) is calculated. If the coefficient of variation is greater than 15%, the feature parameter is considered abnormal and is discarded. The remaining feature parameters are then compared with the target parameters T of the ground points obtained in step one of the data preprocessing module. xPearson correlation coefficient ρ xy Sort them in descending order of the absolute value of the correlation coefficient;

[0067] Step 5: Extract the top 1% of sorted values ​​and |ρ xy The feature parameters with a value greater than 0.7 are compared with the target parameter T. x Perform cross-validation and calculate the mean relative error in the cross-validation. If the maximum relative error is less than 20% and the mean relative error is less than 10%, return the feature parameter expression and end the feature optimization. Otherwise, use the "Genetics and Mutation" process in the Genetic Acceleration Algorithm (Reynolds CW, "Flocks, Herds, and Schools: A Distributed Behavioral Model", Computer Graphics, Vol. 21, No. 4, 1987, pp. 25-34) to obtain the updated AC, i.e., AC′, and repeat steps three to five.

[0068] Taking the water quality inversion of East Dongting Lake as an example, the application of the present invention will be illustrated by way of example.

[0069] Data introduction:

[0070] 1. Field measurement data:

[0071] The average water quality monitoring data for November 2020 was obtained from 11 monitoring points set up in the Dongting Lake area, including 6 in East Dongting Lake and 5 in South (West) Dongting Lake. The water quality parameters were permanganate index, total phosphorus, total nitrogen, chlorophyll a, and transparency, as shown in Figures 2(a) and 2(b).

[0072] 2. Remote sensing data:

[0073] Sentinel-2 MSI data has ground resolutions of 10m, 20m, and 60m, with a revisit period of 10 days, which can be shortened to 5 days with two complementary satellites. A single image can cover either the South Dongting Lake or the East Dongting Lake area. Three images can be acquired per month for the South Dongting Lake area and two images per month for the East Dongting Lake area. The image illustrations are shown in Figures 3(a) and 3(b).

[0074] First, the data preprocessing method of Embodiment 1 of this invention is used to perform atmospheric correction, geometric fine correction, and georegistration on the remote sensing image; then, water index is used to extract pixels of the Dongting Lake area, and the results are as follows. Figure 4 As shown.

[0075] Based on the geographical information of the monitoring points, reflectance information of each band of the Sentinel-2 image data is obtained. Then, through data filtering, namely steps three and four in the data preprocessing process, the self-similarity dimensionality reduction of the spectral data is completed, and finally the bands to be selected for feature construction are obtained.

[0076] For the selected wavelength reflectance, the data from the screened monitoring points are combined with module two of the workflow: feature mining, to obtain the optimal feature combination. Taking the total phosphorus concentration in water quality parameters as an example, the cross-validation results of each combination are as follows: Figures 5(a1) to 5(d2) As shown.

[0077] Finally, the relevant variable X for total phosphorus concentration in the water body was obtained. TP = (B1*B1) / B4, where B1 is the B1 band in the Sentinel-2 data and B4 corresponds to the B4 band in the Sentinel-2 data. The inversion models established using different combinations of training and test sets were analyzed for inversion accuracy. The results are shown in Figures 6(a) and 6(b). The weighted sum of the coefficients of determination R² in all inversion models is 0.8320, and the weighted sum of the average relative errors is 20.77%. This indicates that the quantitative inversion model features constructed in Embodiment 1 of this invention (i.e., the spectral features extracted in this embodiment) have strong generalization ability and robust performance for any combination of training and test sets.

[0078] Select one of the better-performing inversion models and perform residual analysis on the test set inversion, such as... Figure 7 As shown, from Figure 7 It can be seen that the residual of the inversion model is small (only the inversion residual value of the 11th test data is large), and it can accurately invert the total phosphorus concentration value of the water in the test set.

[0079] Example 2

[0080] Embodiment 2 of the present invention provides a terminal device corresponding to Embodiment 1 above. The terminal device can be a processing device for a client, such as a mobile phone, a laptop, a tablet computer, a desktop computer, etc., to execute the method of the above embodiments.

[0081] The terminal device in this embodiment includes a memory, a processor, and a computer program stored in the memory; the processor executes the computer program in the memory to implement the steps of the method in Embodiment 1 described above.

[0082] In some implementations, the memory may be high-speed random access memory (RAM), and may also include non-volatile memory, such as at least one disk storage device.

[0083] In other implementations, the processor can be any type of general-purpose processor, such as a central processing unit (CPU) or a digital signal processor (DSP), and there is no limitation here.

[0084] Example 3

[0085] Embodiment 3 of the present invention provides a computer-readable storage medium corresponding to Embodiment 1 above, on which a computer program / instructions are stored. When the computer program / instructions are executed by a processor, they implement the steps of the method of Embodiment 1 above.

[0086] A computer-readable storage medium can be a tangible device that holds and stores instructions for use by an instruction execution device. A computer-readable storage medium can be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any combination thereof.

[0087] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0088] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0089] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0090] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0091] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for extracting spectral features from remote sensing images, characterized in that, Includes the following steps: S1. Acquire remote sensing images and obtain information B for the band to be selected; S2, Construct a repeatable array , This indicates the items in the corresponding array arrangement positions in the characteristic expression. The selected band information, b represents the number of bands in the selected band information B, and Z is a set of integers; where the characteristic expression f is: ; m represents the number of characteristic expression terms, This indicates the rules of operation before each term, namely +, -, *, / . Simulate whether any of the items are within the scope defined by parentheses "()"; It represents one of the terms in a polynomial formula; Construct repeatable arrays , used to represent the terms in the corresponding array arrangement position in the characteristic expression. The operator symbol represented by "; Build array , used to set the starting position of the brackets in the feature expression; Build array , used to represent the terms in the corresponding array arrangement position in the characteristic expression The selected coefficients; S3. Set m, obtain all combinations of arrays AL, AS, and AP, and initialize array AC. S4. Randomly combine arrays AL, AS, AP, and AC without repetition to obtain a random combination set X. Reverse encode the random combination set X to obtain a feature calculation expression set represented by the random combination set X. Obtain the corresponding feature parameter set F based on the feature calculation expression set. S5. Preprocess the feature parameter set F to obtain preprocessed feature parameters; calculate the correlation coefficient between the preprocessed feature parameters and the ground point target parameters in the remote sensing image, and arrange the preprocessed feature parameters in ascending order according to the absolute value of the correlation coefficient. S6. Extract the preprocessed feature parameters that are in the top M% after sorting and have a correlation coefficient > N. Perform cross-validation between the extracted feature parameters and the ground point target parameters. Calculate the mean relative error in the cross-validation. If the maximum relative error is less than the first set threshold and the mean relative error is less than the second set threshold, output the feature expression and end the process. Otherwise, use the random combination set X as the input of the genetic algorithm to obtain the updated array AC and return to step S4. In step S1, the specific implementation process of obtaining the information B of the band to be selected includes: 1) Remove outliers from the remote sensing image data and read the ground point target parameters from the remote sensing image; 2) Based on the longitude and latitude information in the ground point target parameters, locate the pixel points in the remote sensing image and obtain the spectral data of each ground point; 3) Perform continuous de-statistical processing on the spectral data obtained in step 2); 4) Calculate the first coefficient of variation of the ground point target parameters and the second coefficient of variation of the reflectance of the same band as the spectral data after continuous de-systematization. Compare the magnitudes of the first coefficient of variation and the second coefficient of variation. If the first coefficient of variation is greater than a times the second coefficient of variation, retain the spectral data of that band; otherwise, discard that band. ; 5) Calculate the correlation coefficient between each band for the band information retained after processing in step 4). When the correlation coefficient between two bands is greater than the third set threshold, remove the band data with the longer wavelength between the two bands to obtain the band information B to be selected.

2. The method for extracting spectral features from remote sensing images according to claim 1, characterized in that, The third set threshold is 0.

7.

3. The method for extracting spectral features from remote sensing images according to claim 1, characterized in that, coefficient of variation The calculation formula is: ; This represents the standard deviation of dataset X; This represents the mean of dataset X; Data set X is a set of ground point target parameters or a set of reflectance data in the same band of spectral data after continuous de-systematization.

4. The method for extracting spectral features from remote sensing images according to claim 1, characterized in that, Step S5 details the specific implementation process of preprocessing the feature parameter set F. include: Remove duplicate feature parameters from the feature parameter set F; Calculate the coefficient of variation of the remaining characteristic parameters; Feature parameters with a coefficient of variation greater than the fourth set threshold are removed to obtain the preprocessed feature parameter set.

5. The method for extracting spectral features from remote sensing images according to claim 4, characterized in that, In step S5, the fourth threshold is set to 15%.

6. The method for extracting spectral features from remote sensing images according to claim 1, characterized in that, M=1; N=0.7; the first threshold is set to 20%, and the second threshold is set to 10%.

7. A terminal device, comprising a memory, a processor, and a computer program stored in the memory; characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program / instructions stored thereon; characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method according to any one of claims 1 to 6.

Images