Anomaly Detection Method and System of Spatiotemporal Graph Neural Network for Marine Ecology

Abstract

The present invention relates to the technical field of marine environmental monitoring, and in particular to a spatio-temporal graph neural network anomaly detection method and system for marine ecology. The method includes obtaining marine monitoring data; performing data preprocessing on the obtained marine monitoring data; constructing a spatio-temporal graph structure based on the preprocessed marine monitoring data; using an anomaly detection model to extract spatio-temporal features from the constructed spatio-temporal graph structure; performing feature fusion on the extracted spatio-temporal features; and performing local and global anomaly scoring for the marine ecosystem based on a multi-scale anomaly scoring mechanism. The present invention designs a unified data processing framework, which can simultaneously process multi-source heterogeneous data such as satellite remote sensing data, water quality monitoring data, and biological survey data, fully utilize the information contained in various types of data, and improve the comprehensiveness and accuracy of anomaly detection. Through the construction of a spatio-temporal graph structure, the correlation relationships between different data sources are effectively captured, providing a more comprehensive feature expression for anomaly detection.

Description

Technical Field

[0001] The present invention relates to the technical field of marine environment monitoring, and in particular to a spatio-temporal graph neural network anomaly detection method and system for marine ecology. Background Art

[0002] With the influence of biological activities and climate change, the marine ecosystem is facing many threats, such as frequent abnormal events like seawater acidification, eutrophication, red tide outbreaks, and marine litter pollution. Currently, the anomaly detection of the marine ecosystem mainly adopts traditional detection methods based on statistical thresholds, machine learning methods, and deep learning methods. Among them, the traditional detection method based on statistical thresholds determines whether the system is abnormal by setting the thresholds of key parameters, but this method is difficult to handle the complex correlation relationships between parameters, and the fixed thresholds are difficult to adapt to the dynamic change characteristics of the marine environment. Machine learning methods have advantages in processing multi-dimensional features, but existing methods often analyze temporal and spatial features separately and cannot make full use of the spatio-temporal correlation characteristics of data. Although deep learning methods perform well in processing temporal data or spatial data, it is difficult to effectively model temporal dependence and spatial correlation simultaneously. In addition, there are still many deficiencies in existing detection methods in aspects such as data fusion, model construction, detection mechanism, and system implementation, such as difficulties in effectively integrating multi-source heterogeneous data, lack of explicit modeling of the system network structure, and difficulties in realizing early warning of anomalies. Therefore, there is an urgent need to propose a new anomaly detection method for the marine ecosystem to solve the above problems. Graph neural networks have received extensive attention in recent years due to their powerful structured data processing capabilities. Combining graph neural networks with time series analysis can effectively handle the complex spatio-temporal dependence relationships in the marine ecosystem and provide a new solution for marine ecosystem anomaly detection.

[0003] However, for the problems existing in the traditional marine ecosystem anomaly detection methods, there are still the following technical problems:

[0004] (1) How to effectively fuse multi-source heterogeneous data in the marine ecosystem. When current detection systems process multi-source heterogeneous data such as satellite remote sensing data, water quality monitoring data, and biological survey data, they lack a unified data processing framework and are difficult to make full use of the information contained in various types of data, affecting the comprehensiveness and accuracy of anomaly detection.

[0005] (2) How to effectively fuse multi-source heterogeneous data in the marine ecosystem. When current detection systems process multi-source heterogeneous data such as satellite remote sensing data, water quality monitoring data, and biological survey data, they lack a unified data processing framework and are difficult to make full use of the information contained in various types of data, affecting the comprehensiveness and accuracy of anomaly detection.

[0006] (3) How to construct a reasonable anomaly detection mechanism to achieve early identification and warning of abnormal states of marine ecosystems. Existing detection methods mostly focus on discovering already-occurred anomalies, lacking the ability to predict potential anomalies and unable to provide timely warning information for the management and protection of marine ecosystems. Summary of the Invention

[0007] To solve the above-mentioned problems, the present invention provides a spatio-temporal graph neural network anomaly detection method and system for marine ecology.

[0008] In a first aspect, a spatio-temporal graph neural network anomaly detection method for marine ecology provided by the present invention adopts the following technical solution:

[0009] A spatio-temporal graph neural network anomaly detection method for marine ecology includes:

[0010] Obtain marine monitoring data;

[0011] Perform data preprocessing on the obtained marine monitoring data;

[0012] Construct a spatio-temporal graph structure based on the preprocessed marine monitoring data;

[0013] Use an anomaly detection model to extract spatio-temporal features from the constructed spatio-temporal graph structure;

[0014] Perform feature fusion on the extracted spatio-temporal features;

[0015] Perform local anomaly and global anomaly scoring of the marine ecosystem based on a multi-scale anomaly scoring mechanism;

[0016] Output the anomaly detection result.

[0017] Further, the performing data preprocessing on the obtained marine monitoring data includes performing outlier processing and missing value filling on the marine monitoring data. Among them, the outlier processing includes an improved 3 criterion based on a time series sliding window. By dynamically calculating local statistical features to adapt to the time-varying characteristics of the data, within the time window w, the outlier recognition criterion is expressed as:

[0018]

[0019] Wherein, is the processed data set, x is the observed value, is the mean of the data sequence, is the standard deviation of the data sequence, n is the data volume, and are calculated using a sliding window:

[0020]

[0021]

[0022] wherein, is the local mean at time t, is the local standard deviation at time t, and x i is the observation value at the i-th moment within the window, and t is the current time step.

[0023] Furthermore, the construction of the spatio-temporal graph structure based on the preprocessed ocean monitoring data includes constructing node features by fusing environmental, biological, and temporal information based on triple feature vectors; constructing the edge relationship of the graph based on the node features, wherein a dual edge relationship construction mechanism is used to consider both the spatial position relationship and the state similarity. The spatial position relationship includes the spatial adjacency relationship and the dynamic association relationship. For the spatial adjacency relationship, a Gaussian kernel function based on distance is used to define the association strength:

[0024]

[0025] wherein, is the spatial matrix, and d ij is the distance between node i and node j, is the distance scale parameter that controls the attenuation rate of the spatial association, is the distance threshold, and n is the total number of nodes.

[0026] Furthermore, the spatio-temporal feature extraction of the constructed spatio-temporal graph structure using the anomaly detection model includes using the graph attention network of the anomaly detection model for spatial feature learning, wherein the influence of important nodes is highlighted by adaptively learning the attention coefficients between nodes, expressed as:

[0027]

[0028] wherein, represents the output feature vector of node i, with a dimension of d, (·) is the activation function, is the attention coefficient, representing the influence weight of node j on node i, is the learnable feature transformation matrix, f is the dimension of the input feature, is the input feature vector of node j, and the attention coefficient is calculated as:

[0029]

[0030] wherein, , It is the representation after the transformation of node features. Softmax is a normalization function that ensures the sum of all attention coefficients is 1.

[0031] Furthermore, the use of the anomaly detection model to extract spatio-temporal features from the constructed spatio-temporal graph structure further includes extracting temporal features by combining causal convolution and long short-term memory network. Among them, the causal convolution network extracts local temporal patterns:

[0032]

[0033] Among them, is the temporal feature output at time t, is the parameter of the k-th convolutional kernel, is the input feature at time t-k, and then the long-term dependence relationship is processed using the LSTM network steps:

[0034]

[0035] Among them, is the output feature of the forward LSTM, is the output feature of the backward LSTM, is the forward LSTM function, is the backward LSTM function, is the concatenated temporal feature vector.

[0036] Furthermore, the feature fusion of the extracted spatio-temporal features includes using an adaptive feature fusion mechanism to achieve the optimal combination of features by learning the importance weights of different features, which is expressed as:

[0037]

[0038] Among them, z is the final feature vector after fusion, , are the weight matrices of the spatial feature and the temporal feature respectively, , are the spatial and temporal feature vectors, and b is the bias vector.

[0039] Furthermore, the local and global anomaly scoring of the marine ecosystem based on the multi-scale anomaly scoring mechanism includes first calculating the local anomaly score of each monitoring point, and reflecting the deviation degree of the current state of the monitoring point from its historical normal state through the local anomaly score; a global anomaly scoring mechanism is proposed based on local anomaly detection. Among them, the anomalies at the system level are identified by comparing the changes in feature distributions, which is expressed as:

[0040]

[0041] Among them, The probability distribution of the current feature The probability distribution of the reference normal state is the KL divergence loss. The distribution estimation uses the kernel density estimation method:

[0042]

[0043] where n is the number of samples, h is the smoothness degree of the control sum function, K(·) is the Gaussian kernel function, and z i is the i-th sample point.

[0044] In a second aspect, a spatio-temporal graph neural network anomaly detection system for marine ecology includes:

[0045] A data acquisition module, configured to acquire marine monitoring data;

[0046] A preprocessing module, configured to perform data preprocessing on the acquired marine monitoring data;

[0047] A graph construction module, configured to construct a spatio-temporal graph structure based on the preprocessed marine monitoring data;

[0048] A feature extraction module, configured to extract spatio-temporal features from the constructed spatio-temporal graph structure by using an anomaly detection model;

[0049] A feature fusion module, configured to perform feature fusion on the extracted spatio-temporal features;

[0050] A scoring module, configured to perform local anomaly and global anomaly scoring on the marine ecosystem based on a multi-scale anomaly scoring mechanism, and output an anomaly detection result.

[0051] In a third aspect, the present invention provides a computer-readable storage medium, in which multiple instructions are stored, and the instructions are suitable for being loaded and executed by a processor of a terminal device to perform the above-mentioned spatio-temporal graph neural network anomaly detection method for marine ecology.

[0052] In a fourth aspect, the present invention provides a terminal device, including a processor and a computer-readable storage medium. The processor is used to implement each instruction; the computer-readable storage medium is used to store multiple instructions, and the instructions are suitable for being loaded and executed by the processor to perform the above-mentioned spatio-temporal graph neural network anomaly detection method for marine ecology.

[0053] In summary, the present invention has the following beneficial technical effects:

[0054] (1) The effective fusion of multi-source heterogeneous data is achieved. The present invention designs a unified data processing framework, which can simultaneously process multi-source heterogeneous data such as satellite remote sensing data, water quality monitoring data, and biological survey data, fully utilize the information contained in various types of data, and improve the comprehensiveness and accuracy of anomaly detection. By constructing a spatio-temporal graph structure, the correlation relationships between different data sources are effectively captured, providing a more comprehensive feature representation for anomaly detection.

[0055] (2) The ability to extract spatio-temporal features is improved. The present invention realizes the effective extraction of spatio-temporal features of the marine ecosystem by designing a temporal feature extraction module that combines causal convolution and bidirectional LSTM networks, and a spatial feature learning module based on graph attention mechanism. An adaptive feature fusion mechanism is adopted to automatically adjust the importance weights of spatial features and temporal features according to different scenarios, enhancing the adaptability of the model.

[0056] (3) The early identification and warning of anomalies in the marine ecosystem are realized. The present invention designs a multi-scale anomaly scoring mechanism and an adaptive warning mechanism, which can not only timely detect the occurred anomalies, but also predict potential anomaly risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a schematic diagram of an anomaly detection method based on a spatio-temporal graph neural network for marine ecology in Embodiment 1 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0058] The present invention will be further described in detail below with reference to the accompanying drawings.

[0059] Embodiment 1

[0060] Referring to Figure 1 , an anomaly detection method based on a spatio-temporal graph neural network for marine ecology in this embodiment includes:

[0061] Obtain marine monitoring data;

[0062] Perform data preprocessing on the obtained marine monitoring data;

[0063] Based on the preprocessed marine monitoring data, construct a spatio-temporal graph structure;

[0064] Use the anomaly detection model to extract spatio-temporal features from the constructed spatio-temporal graph structure;

[0065] Perform feature fusion on the extracted spatio-temporal features;

[0066] Based on the multi-scale anomaly scoring mechanism, score local anomalies and global anomalies of the marine ecosystem;

[0067] Output the anomaly detection result.

[0068] Specifically:

[0069] S1 Data acquisition and preprocessing;

[0070] First, in view of the characteristics of multi-source heterogeneous data in the marine ecosystem, the present invention designs a data preprocessing process for the system. Since marine monitoring data is susceptible to environmental, equipment, and human factors, the original data often contains outliers, which will interfere with the subsequent outlier detection effect. Therefore, it is necessary to identify and process outliers.

[0071] (1) Outlier processing,

[0072] Considering the dynamic characteristics of marine ecological data, the present invention proposes an improved 3 criterion based on the time series sliding window on the basis of traditional statistical methods. This method can better adapt to the time-varying characteristics of data by dynamically calculating local statistical features. Specifically, within the time window w, the identification criterion for outliers is:

[0073]

[0074] Among them, is the processed data set, x is the observed value, is the mean of the data sequence, is the standard deviation of the data sequence, n is the data volume, and are calculated using a sliding window:

[0075]

[0076]

[0077] Among them, is the local mean at time t, is the local standard deviation at time t, x i is the observed value at the i-th moment within the window, and t is the current time step.

[0078] (2) Missing value filling,

[0079] Missing phenomena often occur in marine monitoring data, and simple interpolation methods are difficult to ensure the accuracy of the filling results. Therefore, the present invention considers a filling strategy with spatio-temporal dual constraints, makes full use of the temporal continuity and spatial correlation of data, and obtains more reliable filling values by combining the results of temporal prediction and spatial interpolation:

[0080]

[0081] Among them, x t and x s are the results of temporal interpolation and spatial interpolation respectively, and are the weight coefficients of the two items, and .

[0082] S2 Spatiotemporal graph structure construction;

[0083] After completing the data preprocessing, the next step is to construct a graph structure that can effectively express the complex associations of the marine ecosystem. Considering the obvious spatiotemporal coupling characteristics of the marine ecosystem, the present invention designs a dynamic spatiotemporal graph structure.

[0084] (1) Node feature construction. To comprehensively characterize the state features of the monitoring points, the present invention designs a triple feature vector, which integrates environmental, biological, and temporal information:

[0085]

[0086] Among them, is the complete feature vector, d is the total feature dimension, is the environmental parameter feature, is the biological index feature, is the temporal feature, and d = d1 + d2 + d3.

[0087] Based on the above feature representation, the edge relationship of the graph is further constructed. Using a dual-edge relationship construction mechanism, both the spatial position relationship and the state similarity are considered. First is the spatial adjacency relationship, and the association strength is defined using a distance-based Gaussian kernel function:

[0088]

[0089] Among them, is the spatial matrix, d ij is the distance between node i and node j, is the distance scale parameter, which controls the attenuation rate of the spatial association, is the distance threshold, and n is the total number of nodes.

[0090] Secondly is the dynamic association relationship, which is defined based on the similarity of node features:

[0091]

[0092] Among them, is the dynamic association matrix, x i and x j are the feature vectors of nodes i and j, cos(·) is the cosine similarity function, sim(i, j) is the comprehensive similarity between nodes i and j, is the similarity threshold.

[0093] (2) Temporal Dependence Modeling To capture the temporal evolution characteristics of the system state, the present invention designs a temporal correlation matrix. This matrix not only considers state similarity but also introduces a time decay factor to reflect the influence of time distance on the correlation strength:

[0094]

[0095] Wherein, is the state vector of node i at time t, is the state vector of node j at time t - k, is the time scale parameter, which controls the influence degree of state difference, is the time interval, is the time decay coefficient, which controls the duration of time dependence, and k is the time delay step.

[0096] S3 Anomaly Detection Model;

[0097] Based on the constructed spatio-temporal graph structure, the present invention designs a hierarchical anomaly detection model. This model extracts spatio-temporal features through a graph neural network and realizes anomaly detection by combining a multi-scale anomaly scoring mechanism.

[0098] (1) Spatial Feature Learning,

[0099] After constructing the spatio-temporal graph structure, an effective feature extraction mechanism needs to be designed. Considering the differences in the importance of different monitoring points in the marine ecosystem, the present invention uses a graph attention network for spatial feature learning. This network highlights the influence of important nodes by adaptively learning the attention coefficients between nodes:

[0100]

[0101] Wherein, represents the output feature vector of node i, with dimension d, (·) is the activation function, is the attention coefficient, indicating the influence weight of node j on node i, is the learnable feature transformation matrix, f is the dimension of the input feature, is the input feature vector of node j. The attention coefficient is calculated as:

[0102]

[0103] Wherein, , is the representation of the node feature after transformation, and softmax is the normalization function to ensure that the sum of all attention coefficients is 1.

[0104] Through the above steps, the model can automatically identify and focus on the key nodes and associated relationships in anomaly detection.

[0105] (2) Temporal feature learning,

[0106] Anomalies in the marine ecosystem often manifest as changes in temporal patterns. To capture such temporal anomaly features, the present invention designs a temporal feature extraction module that combines causal convolution and long short-term memory networks.

[0107] First, extract local temporal patterns through a causal convolution network:

[0108]

[0109] Among them, is the temporal feature output at time t, is the parameter of the k-th convolution kernel, is the input feature at time t-k, and then use the LSTM network to capture long-term dependencies:

[0110]

[0111] Among them, is the output feature of the forward LSTM, is the output feature of the backward LSTM, is the forward LSTM function, is the backward LSTM function, is the concatenated temporal feature vector.

[0112] (3) Feature fusion and anomaly detection,

[0113] To comprehensively utilize spatial and temporal features for anomaly detection, the present invention designs an adaptive feature fusion mechanism. This mechanism realizes the optimal combination of features by learning the importance weights of different features:

[0114]

[0115] Among them, z is the final fused feature vector, , are the weight matrices of spatial and temporal features respectively, , are spatial and temporal feature vectors, and b is the bias vector.

[0116] S4 Multi-scale anomaly scoring mechanism;

[0117] Considering that marine ecosystem anomalies may occur simultaneously at local and global scales, the present invention proposes a multi-scale anomaly scoring mechanism.

[0118] (1) Local anomaly scoring,

[0119] First, calculate the local anomaly score for each monitoring point. This score reflects the degree of deviation of the current state of the monitoring point from its historical normal state:

[0120]

[0121] Among them, is the importance weight of the i-th dimension feature, is the i-th dimension feature value of the current observation, and is the historical normal mean of the i-th dimension feature. To improve the stability of detection, the present invention adopts a local statistical feature update mechanism based on a sliding window:

[0122]

[0123] Among them, is the update rate parameter, used to control the retention degree of historical information, t is the current time step, w is the size of the sliding window, is the historical observation value of the i-th dimension feature within the window.

[0124] (2) Global anomaly scoring,

[0125] Based on the local anomaly detection, the present invention further designs a global anomaly scoring mechanism. This mechanism identifies system-level anomalies by comparing changes in feature distributions:

[0126]

[0127] Among them, is the probability distribution of the current feature, is the probability distribution of the reference normal state, is the KL divergence loss. The distribution estimation adopts the method of kernel density estimation:

[0128]

[0129] Among them, n is the number of samples, h is the smoothing degree of the control and function, K(·) is the Gaussian kernel function, z i is the i-th sample point.

[0130] The anomaly detection of the marine ecosystem is completed through the above steps S1-S4.

[0131] Embodiment 2

[0132] This embodiment provides a spatio-temporal graph neural network anomaly detection system for the marine ecosystem, including:

[0133] A data acquisition module, configured to:

[0134] A computer-readable storage medium stores multiple instructions, and the instructions are adapted to be loaded and executed by a processor of a terminal device for the above-mentioned spatio-temporal graph neural network anomaly detection method for ocean ecology.

[0135] A terminal device includes a processor and a computer-readable storage medium. The processor is used to implement each instruction; the computer-readable storage medium is used to store multiple instructions, and the instructions are adapted to be loaded and executed by the processor for the above-mentioned spatio-temporal graph neural network anomaly detection method for ocean ecology.

[0136] The above are all preferred embodiments of the present invention, and the protection scope of the present invention is not limited thereby. Therefore, all equivalent changes made according to the structure, shape, and principle of the present invention should be covered within the protection scope of the present invention.

Claims

1. A spatio-temporal graph neural network anomaly detection method for marine ecology, characterized in that Including: Obtain ocean monitoring data; Perform data preprocessing on the obtained ocean monitoring data; Construct a spatio-temporal graph structure based on the preprocessed ocean monitoring data; Use an anomaly detection model to extract spatio-temporal features from the constructed spatio-temporal graph structure; Perform feature fusion on the extracted spatio-temporal features; Perform local anomaly and global anomaly scoring for the marine ecosystem based on a multi-scale anomaly scoring mechanism; Output the anomaly detection result; The data preprocessing of the obtained marine monitoring data includes outlier processing and missing value filling for the marine monitoring data. Among them, the outlier processing includes an improved 3 criterion. By dynamically calculating local statistical features to adapt to the time-varying characteristics of the data, within the time window w, the identification criterion for outliers is expressed as: Among them, is the processed data set, x is the observed value, is the mean of the data sequence, is the standard deviation of the data sequence, n is the data volume, and are calculated using a sliding window: Among them, is the local mean at time t, is the local standard deviation at time t, and x i is the observation value at the i-th moment within the window, and t is the current time step; The constructing of the spatio-temporal graph structure based on the preprocessed ocean monitoring data includes constructing node features by fusing environmental, biological, and temporal information based on triple feature vectors; constructing the edge relationship of the graph based on the node features, wherein a dual edge relationship construction mechanism is used to consider both the spatial position relationship and the state similarity. For the spatial position relationship, a Gaussian kernel function based on distance is used to define the association strength: Among them, is a spatial matrix, and d ij is the distance between node i and node j, is a distance scale parameter that controls the decay rate of spatial correlation, is a distance threshold, and n is the total number of nodes; The using of the anomaly detection model to extract spatio-temporal features from the constructed spatio-temporal graph structure includes using the graph attention network of the anomaly detection model for spatial feature learning, wherein the influence of important nodes is highlighted by adaptively learning the attention coefficients between nodes, expressed as: Among them, represents the output feature vector of node i, with a dimension of d, (·) is the activation function, is the attention coefficient, indicating the influence weight of node j on node i, is a learnable feature transformation matrix, f is the dimension of the input feature, is the input feature vector of node j, and the attention coefficient is calculated as follows: Among them, , is the representation of the node features after transformation, and softmax is a normalization function that ensures the sum of all attention coefficients is 1; The using of the anomaly detection model to extract spatio-temporal features from the constructed spatio-temporal graph structure further includes combining causal convolution and long short-term memory network for temporal feature extraction, wherein the causal convolution network extracts local temporal patterns: Among them, is the temporal feature output at time t, is the parameter of the k-th convolutional kernel, is the input feature at time t - k, and then the long-term dependency relationship is processed using the LSTM network steps: Among them, is the output feature of the forward LSTM, is the output feature of the backward LSTM, is the forward LSTM function, is the backward LSTM function, is the concatenated time series feature vector; The performing of feature fusion on the extracted spatio-temporal features includes using an adaptive feature fusion mechanism to achieve the optimal combination of features by learning the importance weights of different features, expressed as: where z is the final fused feature vector, , are the weight matrices of the spatial feature and the temporal feature respectively, , are the spatial and temporal feature vectors, and b is the bias vector; The performing of local anomaly and global anomaly scoring for the marine ecosystem based on the multi-scale anomaly scoring mechanism includes first calculating the local anomaly score of each monitoring point, and reflecting the deviation degree of the current state of the monitoring point from its historical normal state through the local anomaly score; a global anomaly scoring mechanism is proposed based on local anomaly detection, wherein the anomaly at the system level is identified by comparing the changes in feature distributions, expressed as: Among them, is the probability distribution of the current feature, is the probability distribution of the reference normal state, is the KL divergence loss, and the distribution estimation adopts the method of kernel density estimation: where n is the number of samples, h is the smoothness of the control sum function, K(·) is the Gaussian kernel function, and z i is the i-th sample point.

2. A spatio-temporal graph neural network anomaly detection system for marine ecology, which executes the method described in claim 1, and is characterized in that, Including: A data acquisition module configured to obtain ocean monitoring data; A preprocessing module configured to perform data preprocessing on the obtained ocean monitoring data; A graph construction module configured to construct a spatio-temporal graph structure based on the preprocessed ocean monitoring data; A feature extraction module configured to use an anomaly detection model to extract spatio-temporal features from the constructed spatio-temporal graph structure; A feature fusion module configured to perform feature fusion on the extracted spatio-temporal features; A scoring module configured to perform local anomaly and global anomaly scoring for the marine ecosystem based on a multi-scale anomaly scoring mechanism and output the anomaly detection result.

3. A computer-readable storage medium storing a plurality of instructions, characterized in that, The instructions are adapted to be loaded and executed by a processor of a terminal device to perform the method according to claim 1.

4. A terminal device, comprising a processor and a computer-readable storage medium, the processor being configured to implement each instruction; the computer-readable storage medium being configured to store multiple instructions, characterized in that, The instructions are adapted to be loaded and executed by a processor to perform the method according to claim 1.

Images