Multi-data fusion ecological environment monitoring method and system

Abstract

The invention discloses a multi-data fusion ecological environment monitoring method and system. The method comprises the following steps: acquiring ecological image data of a target area and corresponding non-image environment data; performing environment state inversion based on the ecological image data and the non-image environment data to obtain a first environment state parameter and a second environment state parameter; and performing state consistency verification on the first environment state parameter and the second environment state parameter to obtain a consistency verification result, and determining the weight of the ecological image data and the non-image environment data in data fusion, or determining the credibility level of an ecological environment monitoring result. According to the method, data fusion is improved from a result level to a state characterization level by inverting the environment state parameters corresponding to the physical meanings, so that the monitoring process better fits the inherent law of ecological evolution, wrong monitoring results are remarkably reduced through state consistency verification, and the accuracy and reliability of ecological environment monitoring are greatly improved.

Description

Technical Field

[0001] This invention relates to the field of ecological environment monitoring technology, specifically to a method and system for ecological environment monitoring that integrates multiple data sources. Background Technology

[0002] In the engineering practice of ecological and environmental monitoring, monitoring data mainly includes image-based data and non-image-based data. Image-based data includes remote sensing images, fixed monitoring images, and periodically acquired images, while non-image-based data includes meteorological data, historical statistical data, and regional ecological basic data. Traditional methods typically use feature stitching, weighted fusion, or joint modeling to process these multi-source data and then generate monitoring conclusions.

[0003] However, in practical applications, traditional methods have significant problems. Image recognition results may appear valid numerically or superficially, but lack the support of ecological processes. For example, a change in the color of a water body image might be interpreted as an anomaly in water quality, but the hydrological conditions, rainfall, and historical trends during the same period do not support such changes. Non-image data and image results may contradict each other in terms of time scale or magnitude of change. For instance, vegetation images may show a degradation trend, but the regional climate, soil moisture, and historical growth cycle indicate that the conditions for such a change are not present in the short term. Existing technologies assume that the output results of various types of data are consistent in a physical and ecological sense, but lack an effective verification mechanism to verify whether the fused results conform to ecological evolution mechanisms and environmental response patterns. This leads to misjudgments or false alarms, especially in engineering applications, particularly in scenarios with high requirements for result reliability, such as supervision, law enforcement, and long-term trend analysis.

[0004] Therefore, how to overcome the shortcomings of traditional methods with effective methods has become a technical problem that urgently needs to be solved. Summary of the Invention

[0005] In response, the present invention provides a method and system for ecological environment monitoring that integrates multiple data sources, so as to at least partially solve the above-mentioned technical problems.

[0006] This invention provides a multi-data fusion method for ecological environment monitoring, comprising the following steps: S1: Acquire ecological image data and corresponding non-image environmental data of the target area, wherein the non-image environmental data includes at least one of meteorological data, hydrological data or historical statistical data; S2: Based on the ecological image data, perform environmental state inversion to obtain the first environmental state parameter; the first environmental state parameter is a continuous or semi-continuous numerical parameter describing the continuous evolution trend of ecological elements; S3: Based on the non-image environment data, perform environmental state inversion to obtain a second environmental state parameter; the second environmental state parameter corresponds to the first environmental state parameter in physical meaning; S4: Perform a state consistency check on the first environmental state parameter and the second environmental state parameter to obtain a consistency check result; the state consistency check includes determining whether the deviation between the first environmental state parameter and the second environmental state parameter is within a preset tolerance range; S5: Based on the consistency verification results, determine the weights of the ecological image data and the non-image environment data in data fusion, or determine the credibility level of the ecological environment monitoring results.

[0007] In one aspect of this application, in step S2, the environmental state inversion based on the ecological image data includes: preprocessing the ecological image data, the preprocessing including geometric correction and radiometric normalization; extracting image features from the preprocessed ecological image data, and mapping the image features to the first environmental state parameter.

[0008] In one aspect of this application, extracting image features from preprocessed ecological image data includes: extracting intermediate layer feature vectors using a convolutional neural network; mapping the image features to a first environmental state parameter includes: performing a linear or nonlinear transformation on the intermediate layer feature vectors to obtain the first environmental state parameter.

[0009] In one aspect of this application, in step S3, the environmental state inversion based on the non-image environment data includes: performing time scale unification and outlier processing on the non-image environment data; processing the processed non-image environment data using a trend analysis model or a conditional constraint model to obtain the second environmental state parameters; wherein the trend analysis model is a time series sliding window statistical model or a trend fitting model; and the conditional constraint model is a logical judgment model constructed based on ecological experience rules.

[0010] In one aspect of this application, the state consistency verification result is a multi-level result, including a high consistency level, an acceptable inconsistency level, and a significant inconsistency level.

[0011] In one aspect of this application, the first environmental state parameter and the second environmental state parameter include at least one of the following: water eutrophication trend parameter, vegetation growth stage offset parameter, or surface humidity change direction parameter.

[0012] Another aspect of this application provides a multi-data fusion ecological environment monitoring system, including the following modules: The data acquisition module is used to acquire ecological image data and corresponding non-image environmental data of the target area. The non-image environmental data includes at least one of meteorological data, hydrological data, or historical statistical data. The first state inversion module is used to perform environmental state inversion based on the ecological image data to obtain a first environmental state parameter; the first environmental state parameter is a continuous or semi-continuous numerical parameter describing the continuous evolution trend of ecological elements. The second state inversion module is used to perform environmental state inversion based on the non-image environment data to obtain a second environmental state parameter; the second environmental state parameter corresponds to the first environmental state parameter in physical meaning. The verification module is used to perform a state consistency verification between the first environmental state parameter and the second environmental state parameter to obtain a consistency verification result; the state consistency verification includes determining whether the deviation between the first environmental state parameter and the second environmental state parameter is within a preset tolerance range. The monitoring module is used to determine the weights of the ecological image data and the non-image environmental data in data fusion, or to determine the reliability level of the ecological environment monitoring results, based on the consistency verification results.

[0013] In another aspect, this application also provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method described above.

[0014] In another aspect, this application provides a computer-readable storage medium having stored thereon computer program instructions that can be executed by a processor to implement the method described above.

[0015] In another aspect of this application, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the method described above.

[0016] The solution provided in this application elevates data fusion from the result level to the state representation level by inverting environmental state parameters corresponding to physical meaning from ecological image data and non-image environmental data. This makes the monitoring process more in line with the inherent laws of ecological evolution, and significantly reduces erroneous monitoring results through state consistency verification, thereby greatly improving the accuracy and reliability of ecological environment monitoring. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a schematic diagram of an ecological environment monitoring method based on multi-data fusion provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of an ecological environment monitoring system with multi-data fusion provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of a device provided in an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] In a typical configuration of this application, the terminal and the service network devices each include one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0021] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0022] Computer-readable media include permanent and non-permanent, removable and non-removable media, which can store information by any method or technology. Information can be computer program instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, read-only optical disc (CD-ROM), digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0023] This application proposes a multi-data fusion method for ecological environment monitoring. The technical solution of this application will be described in detail below with reference to various embodiments.

[0024] like Figure 1 As shown in the figure, this invention discloses a multi-data fusion method for ecological environment monitoring, including the following steps: Step S1: Obtain ecological image data and corresponding non-image environmental data of the target area, wherein the non-image environmental data includes at least one of meteorological data, hydrological data or historical statistical data; In practice, the ecological image data can be acquired by deploying multispectral remote sensing monitoring equipment, drone aerial photography equipment, and ground-based fixed-point high-definition camera equipment. Periodic or continuous image acquisition is carried out on the target area. During the acquisition process, the resolution, color space, and shooting time reference of the images need to be unified so that the image data obtained by different acquisition equipment and different acquisition time periods have consistency and comparability. The non-image environmental data includes at least one of meteorological data, hydrological data, or historical statistical data. Meteorological data can be collected in real-time by meteorological monitoring stations strategically deployed within and around the target area. Collected indicators may include temperature, humidity, light intensity, wind speed, and precipitation, supplemented by long-term monitoring data stored in the existing meteorological observation network within the area. Hydrological data can be obtained from hydrological monitoring stations within the target area, including water flow velocity, water level, and water quality indicators. For target areas with rivers, lakes, or other water bodies, related data such as soil moisture content around the water bodies should also be collected simultaneously. Historical statistical data can be obtained by retrieving archived ecological statistics of the target area from agencies such as ecological environment management departments and natural resource management departments. This data may include historical records of the population size and distribution of biological species within the target area, land use change data, and ecosystem service function assessment data. After collecting both ecological image data and non-image environmental data, a data correspondence needs to be established, matching ecological image data collected at the same time and spatial location with non-image environmental data to form a one-to-one corresponding dataset.

[0025] Step S2: Perform environmental state inversion based on the ecological image data to obtain a first environmental state parameter; the first environmental state parameter is a continuous or semi-continuous numerical parameter describing the continuous evolution trend of ecological elements; specifically, the environmental state inversion based on the ecological image data includes: preprocessing the ecological image data, the preprocessing including geometric correction and radiometric normalization; extracting image features from the preprocessed ecological image data, and mapping the image features to the first environmental state parameter. Extracting image features from the preprocessed ecological image data includes: using a convolutional neural network to extract intermediate layer feature vectors; mapping the image features to the first environmental state parameter includes: performing linear or nonlinear transformations on the intermediate layer feature vectors to obtain the first environmental state parameter.

[0026] Here, to address the spatial geometric distortion of ecological images caused by changes in shooting angle, shooting position, or sensor attitude shift, a polynomial correction method can be used in some specific implementations to achieve geometric correction. First, select no fewer than 20 evenly distributed ground control points in the image. These ground control points should be geographically identifiable features within the target area, such as road intersections, bridge endpoints, and corner points of landmark buildings, and the coordinate accuracy error of the control points in the image and geographic coordinate system must not exceed 0.5 pixels. Based on the selected ground control points, a quadratic polynomial transformation model is constructed, the mathematical expression of which is: Where (x,y) are the original pixel coordinates of the image, , () represents the corrected geographic coordinates. ~ , These are polynomial coefficients. The control point coordinate data are fitted using the least squares method to obtain the polynomial coefficients. The entire image is then transformed using pixel coordinates to achieve geometric correction. After correction, the error between the actual distance and the pixel distance between any two points in the image must be controlled within 5%. This ensures that the extracted features accurately correspond to the ecological conditions of the actual geographical area.

[0027] To reduce image radiometric distortion caused by factors such as variations in illumination intensity, differences in imaging time, and inconsistent sensor responses, histogram matching can be used for radiometric normalization in some specific implementations. First, an image of the target area taken under standard illumination conditions and with stable ecological conditions is selected as a reference image, and the gray-level histogram distribution function of the reference image is calculated. Then, the gray-level histogram distribution function of the ecological image to be processed is calculated. A gray-level mapping relationship T(z) is constructed such that the histogram distribution of the image to be processed is consistent with the histogram distribution of the reference image. The mathematical expression of the mapping relationship is: Where z is the original gray level of the image to be processed. These are candidate gray levels for the reference image during the mapping process. The mapping is a result of the mapping, and its value is satisfied. smallest For color ecological images, the above histogram matching process needs to be performed on the three RGB channels separately to ensure that the radiometric characteristics of each channel are consistent with the reference image. After radiometric normalization, the difference in grayscale values ​​of corresponding pixels in images of the same area taken at different times should be controlled within 10% to avoid misjudgment of state parameters due to non-ecological factors such as lighting.

[0028] To reduce image radiation distortion caused by factors such as changes in light intensity, differences in imaging time, and inconsistent sensor responses, the acquisition time of ecological images can be calibrated and aligned in some specific implementations. For periodically acquired ecological images, the acquisition time of each image is recorded and kept consistent with the time recording format of non-image environmental data. In practice, UTC time or local standard time can be used. If the non-image environmental data is aggregated data at the hourly or daily level, all ecological images acquired within that time period need to be labeled as image data at the corresponding time scale. If multiple ecological images exist within the same time scale, the average feature of the multiple images is calculated as the image input data for that time scale, ensuring that the time granularity of the image data matches that of the non-image data and avoiding deviations in the state inversion results due to time asynchrony.

[0029] The specific structure of the convolutional neural network is as follows: The convolutional neural network adopts a conventional structure with multiple convolutional and pooling layers. Its input is a pre-processed ecological image, with a uniform size of 256×256 pixels. The network has 8-12 layers, including 6-8 convolutional layers and 3-4 pooling layers. The input layer receives 256×256×3 color image data. The first convolutional layer can use 32 convolutional kernels of size 3×3, with a stride of 1, SamePadding padding, and ReLU activation function. The output feature map size is 256×256 pixels. The first pooling layer can use a 2×2 max-pooling kernel with a stride of 2 and ValidPadding padding, resulting in an output feature map size of 128×128×32. The second convolutional layer can use 64 3×3 convolutional kernels with a stride of 1, SamePadding padding, and ReLU activation, resulting in an output feature map size of 128×128×64. The second pooling layer can also use a 2×2 max-pooling kernel with a stride of 2 and ValidPadding padding. g, the output feature map size is 64×64×64; the number of convolutional kernels in the third to eighth convolutional layers increases sequentially to 128, 256, 256, and 512, respectively, with a kernel size of 3×3, a stride of 1, SamePadding padding, and ReLU activation function; a pooling layer is set after every two convolutional layers, with a pooling kernel size of 2×2 and a stride of 2, and the output feature map size is reduced sequentially to 32×32×128, 16×16×256, and 8×8×512; the 8×8×512 feature map output from the last pooling layer is processed... The data is converted into a one-dimensional feature vector through a flattening layer, with a vector dimension of 32768. A fully connected layer is set as an intermediate feature layer with an output dimension of 1024. The activation function is ReLU. The output of the intermediate feature layer is interpretable and mappable. Therefore, the output of the intermediate feature layer is the image feature extracted from the ecological image data. It can be understood that by gradually increasing the number of convolutional kernels and reducing the feature map size, the extraction of high-level semantic features, such as water body distribution and vegetation cover features, can be achieved from low-level visual features such as image edges and textures. Moreover, the high-level semantic features are related to changes in environmental state.

[0030] In practice, the training set of the convolutional neural network can use historical ecological image data of the target region. The specific training process is as follows: the ecological images in the training set are subjected to geometric correction, radiometric normalization, and temporal alignment according to the aforementioned image preprocessing steps, and the image size is unified to 256×256 pixels; the preprocessed image data is divided into training set, validation set, and test set according to a 7:2:1 ratio; the mean square error loss function is used, and the loss is calculated based on the mapping error between the feature vector output by the intermediate feature layer and the environmental state reference data. The expression of the mean square error loss function is: Where N is the number of training samples. For the i-th training sample image, The feature vector output by the intermediate feature layer. Let be the environmental state reference data vector corresponding to the i-th sample.

[0031] Here, the Adam optimizer can be used, with an initial learning rate of 0.001 and a learning rate decay strategy of decreasing to 0.9 every 10 epochs; the batch size is set to 32, and the number of training epochs is set to 50; the training set data is input into the convolutional neural network, and the network parameters are updated through the backpropagation algorithm. The model performance is evaluated using a validation set after each training epoch. When the validation set loss does not decrease for 5 consecutive epochs, training is stopped to avoid overfitting; the trained model is tested using a test set. If the mapping error between the intermediate feature vector and the environmental state reference data exceeds a preset threshold, which can be a mean squared error of 0.05, the network structure or training parameters are adjusted by increasing the number of convolutional layers or adjusting the convolutional kernel size, and training is retrained until the error requirement is met.

[0032] In practice, the state parameter mapping specifically involves determining the mapping function type based on different environmental monitoring objectives. This mapping function type can be linear or nonlinear, and can be determined based on the correlation between the feature vector and the environmental state parameters. For cases where the features and state parameters are linearly correlated, a linear mapping function is used; for cases where they are nonlinearly correlated, a neural network mapper consisting of two fully connected layers is used to achieve the nonlinear transformation.

[0033] The first environmental state parameter includes at least one of the following: eutrophication trend parameter, vegetation growth stage offset parameter, or surface humidity change direction parameter. Specifically, for parameters linearly correlated with the feature vector, such as the eutrophication trend parameter and the vegetation growth stage offset parameter, the mapping function expression is: Where p is the first environmental state parameter, W is the weight matrix, f is the feature vector output by the intermediate feature layer, and b is the bias vector. The weight matrix W and the bias vector b are obtained by the least squares method, based on the fitting of the feature vectors in the training set with the corresponding environmental state reference data.

[0034] For parameters such as the direction of surface humidity change, which are nonlinearly related to the feature vector, a two-layer fully connected neural network can be used as a mapper. The first fully connected layer has the same input dimension as the output dimension of the intermediate feature layer, 1024, and an output dimension of 256, with ReLU activation function. The second fully connected layer has an input dimension of 256 and an output dimension of 1, corresponding to a single state parameter, with Sigmoid activation function used to map the parameter value to the [0,1] interval. The training of the mapper is performed simultaneously with the training of the convolutional neural network, and the mean squared error loss function is also used, calculated based on the error between the mapped output state parameter and the reference data.

[0035] Depending on the monitoring target, one or more first environmental state parameters can be obtained through the above mapping method. The eutrophication trend parameter ranges from [0,1], with values ​​closer to 0 indicating a weaker eutrophication trend and closer to 1 indicating a stronger trend. It is obtained through a linear mapping function. During the mapping process, calibration is required using eutrophication level data (oligotrophic, mesotrophic, eutrophic, and hypereutrophic) from field monitoring to ensure the parameter values ​​align with the actual eutrophication trend. The vegetation growth stage offset parameter ranges from [-0.5,0.5], with negative values ​​indicating that the vegetation growth stage lags behind the historical average. The average level is positive, indicating that it is ahead of the historical average level for the same period. The larger the absolute value, the more significant the deviation. It is obtained through a nonlinear mapping function. Historical vegetation growth cycle data is introduced as a constraint during the mapping process to make the calculation of the deviation conform to the vegetation growth law. The value range of the surface humidity change direction parameter is [0,1]. The parameter value is less than 0.3, indicating that the surface humidity is decreasing. 0.3~0.7 indicates that the humidity is stable. The parameter value is greater than 0.7, indicating that the humidity is increasing. It is obtained through a linear mapping function. During the mapping process, regional hydrological data such as rainfall and groundwater recharge are used for verification to ensure that the parameter can accurately reflect the direction of humidity change.

[0036] Step S3: Perform environmental state inversion based on the non-image environment data to obtain a second environmental state parameter; the second environmental state parameter corresponds to the first environmental state parameter in physical meaning; specifically, the environmental state inversion based on the non-image environment data includes: performing time scale unification and outlier processing on the non-image environment data; and processing the processed non-image environment data using a trend analysis model or a conditional constraint model to obtain the second environmental state parameter; To eliminate the impact of missing data, inconsistent timescales, and outliers on state inversion, some specific implementations format the non-image environment data, such as storing it in CSV format. Each data record includes fields such as data type, collection time, data value, and data source. In practice, due to differences in the time granularity of different types of non-image data, all data needs to be unified to a timescale consistent with the ecological image data. For missing values, different processing methods are used depending on the missing rate. If the missing rate is <5%, the average of adjacent data can be used for imputation. For consecutive missing values ​​in the time series, the average of the three valid data points before and after the missing period is taken as the imputation value. If 5% ≤ missing rate < 20%, a missing value imputation model based on random forest can be used. Other relevant types of non-image data are used as input features to construct a random forest model to predict missing values. The model is trained using historical data without missing values, and the prediction error needs to be controlled within 10%. If the missing rate is ≥20%, and the missing data has a significant impact on state inversion, similar data from other data sources are collected. If the impact is minor, the missing data types are removed, and state inversion is performed based on the remaining data. Outliers can be identified and removed from the data using box plots. Subsequently, based on the characteristics of the non-image data and the monitoring targets, an environmental state inversion is performed using a trend analysis model or a conditional constraint model. The trend analysis model is used to extract the direction and rate of change of the environmental state from the non-image data in the form of a time series, thereby obtaining the second environmental state parameters. The trend analysis model is a time series sliding window statistical model or a trend fitting model; the conditional constraint model is a logical judgment model constructed based on ecological experience rules.

[0037] Specifically, the time-series sliding window statistical model uses a fixed-length sliding window to perform local statistical analysis on time-series data, capturing short-term trends in environmental conditions. The window length is determined based on the data timescale and the evolution cycle of environmental elements. For daily data, the window length is typically set to a 7-day weekly window or a 30-day monthly window, reflecting short-term trends while avoiding excessive data smoothing.

[0038] Taking the inversion of eutrophication trend parameters of water bodies as an example, a 30-day sliding window is used to statistically analyze the changes in water quality indicators such as ammonia nitrogen concentration and total phosphorus concentration, as well as data such as rainfall and water level within the window, including statistical quantities such as mean change rate, maximum change range, and standard deviation.

[0039] Using historical non-image data and corresponding field monitoring environmental status data as training samples, a mapping relationship between statistics and environmental status parameters is constructed. A multiple linear regression method is used to train the mapping model, and the expression for the mapping model is: Where q is the second environmental state parameter. ~ For statistics within the sliding window, ~ The regression coefficients are obtained by solving the least squares method. The inversion accuracy of the model is evaluated using validation set data. The correlation coefficient between the inversion results and the field monitoring data should be greater than 0.85 and the mean square error should be less than 0.05. If these conditions are not met, the sliding window length or the type of statistic should be adjusted, and the model should be retrained.

[0040] The trend fitting model uses curve fitting to fit the time series data, obtaining a trend curve. By analyzing the slope, curvature, and other characteristics of the trend curve, environmental state parameters are retrieved. Depending on the data trend, linear fitting, exponential fitting, or polynomial fitting can be selected. In practice, taking the retrieval of vegetation growth stage offset parameters as an example, non-image data such as daily average temperature, rainfall, and sunshine duration from the past three years are selected as input, and field-monitored vegetation growth stage data is used as output to construct a polynomial trend fitting model. The fitting function is: Where y(t) is the predicted vegetation growth stage at time t, and t is the time variable in days. ~ The fitting coefficients are given by k, which is the order of the polynomial. It can often be taken as 2 or 3 to avoid overfitting. The trend fitting model is trained using the least squares method to solve for the fitting coefficients. By adjusting the order of the polynomial, the fitting curve can fit the historical data to the maximum extent while ensuring the smoothness of the curve. The fitting model is validated using an independent test dataset. The mean absolute error (MAE) between the predicted value and the actual monitored value is calculated. The MAE is required to be less than 0.03. If it is not satisfied, the fitting function type is adjusted and the fitting is performed again.

[0041] The conditional constraint model is constructed based on ecological common sense and engineering experience rules. It determines whether non-image data meets preset environmental state constraints and inversely obtains the second environmental state parameters. Logical constraint rules are constructed according to the ecological characteristics and environmental evolution patterns of the target area, with each rule corresponding to a specific environmental state scenario. In practice, the water body state constraint rule could be: if daily rainfall is greater than 50 mm and the water level rises by more than 10 cm compared to the previous day, the eutrophication trend parameter should be less than 0.3, indicating a weakening eutrophication trend; if there is no rainfall for 7 consecutive days and the ammonia nitrogen concentration is greater than 1.0 mg / L, the eutrophication trend parameter should be greater than 0.7, indicating an increasing eutrophication trend. The vegetation growth constraint rules are as follows: If, during the vegetation growing season (April-September), the average daily temperature is between 20-30℃ and the daily rainfall is between 10-50mm, then the vegetation growth stage offset parameter should be between -0.1 and 0.1, indicating that the growth stage basically conforms to the historical level for the same period. If the average daily temperature remains below 15℃ for more than 10 days, then the vegetation growth stage offset parameter should be less than -0.3, indicating that the growth stage is lagging. The surface humidity constraint rules are as follows: If the daily rainfall is greater than 30mm, then the surface humidity change direction parameter should be greater than 0.7, indicating that the humidity is increasing. If the sunshine duration exceeds 8 hours for 5 consecutive days and the relative humidity is less than 40%, then the surface humidity change direction parameter should be less than 0.3, indicating that the humidity is decreasing.

[0042] Since different constraint rules have varying degrees of impact on the environmental state, in some specific implementations, the analytic hierarchy process (AHP) can be used to assign weights to each rule. Five to ten experts in the field of ecological and environmental monitoring are invited to conduct pairwise comparisons of the importance of each rule, constructing a judgment matrix. By calculating the eigenvectors of the judgment matrix, the weight values ​​of each rule are obtained, with the sum of the weight values ​​being 1. The constraint rules and their corresponding weights are embedded into a conditional constraint model. The input is preprocessed non-image data, and the output is a second environmental state parameter. The operating logic of the conditional constraint model is as follows: for the input non-image data, it is determined whether each constraint rule is satisfied; for satisfied rules, the corresponding environmental state reference value is multiplied by the rule weight to obtain a weighted reference value; the sum of all weighted reference values ​​is calculated as the initial value of the second environmental state parameter; a correction coefficient is introduced to calibrate the initial value. The correction coefficient is obtained based on the statistical analysis of inversion errors from historical data to ensure the accuracy of the parameter value. Field monitoring data is collected regularly to update and optimize the constraint rules and weights. If the inversion error of a certain rule exceeds the preset threshold multiple times in a row, the rationality of the rule will be re-evaluated and the rule conditions or weight values ​​will be adjusted. If a new environmental state scenario appears, corresponding constraint rules will be added so that the model can adapt to different environmental changes.

[0043] The preprocessed non-image data is processed using the aforementioned trend analysis model or conditional constraint model to generate a second environmental state parameter. The type of the second environmental state parameter corresponds to the first environmental state parameter, and includes at least one of the following: water eutrophication trend parameter, vegetation growth stage offset parameter, or surface humidity change direction parameter.

[0044] Step S4: Perform a state consistency check on the first environmental state parameter and the second environmental state parameter to obtain a consistency check result. The state consistency check result is a multi-level result, including a high consistency level, an acceptable inconsistency level, and a significant inconsistency level. The state consistency check includes determining whether the deviation between the first environmental state parameter and the second environmental state parameter is within a preset tolerance range.

[0045] Here, the state consistency verification maps the first environmental state parameter and the second environmental state parameter to the same state space through normalization processing, and then determines whether the deviation between the two is within the preset tolerance range to obtain multi-level consistency verification results. Furthermore, it performs anomaly source analysis on inconsistent states, thereby effectively identifying reasonable but inconsistent contradictions and solving the problem of the lack of ecological rationality verification in existing technologies.

[0046] In some specific implementations, the min-max normalization method can be used for normalization. Specifically, based on the historical environmental state data of the target area, the actual value ranges of the first and second environmental state parameters are statistically analyzed and denoted as [min1, max1] and [min2, max2], respectively. For example, for the eutrophication trend parameter of water body, the historical value range of the first parameter is [0.05, 0.95], and the historical value range of the second parameter is [0.1, 0.9]. Normalization is then performed on each first environmental state parameter p1 and the corresponding second environmental state parameter p2, resulting in normalized parameter values. and The value range of is [0,1], and the calculation formula is: , Then, the normalized parameters were validated by selecting a set of sample data with known environmental states and calculating their normalized values. and If the mean difference between the two is greater than 0.1, the parameter range is readjusted and normalization is performed again until the mean difference is less than 0.1, so that the normalization process does not introduce additional bias.

[0047] The tolerance range is derived based on the ecological evolution patterns of the target area, historical data statistics, and engineering experience. Specifically, based on historical data of the first and second environmental state parameters of the target area over the past 3-5 years, the deviation value Δp = | -|. Statistical analysis is performed on the deviation value Δp, calculating its mean μ, standard deviation σ, median M, and other statistical quantities, and analyzing the distribution type of the deviation. Based on the deviation distribution characteristics and combined with the reliability requirements of ecological environment monitoring, multiple levels of tolerance intervals are set. Taking the normal distribution as an example, the tolerance intervals can be set as follows: high consistency tolerance interval [0, μ+1σ], acceptable inconsistency tolerance interval (μ+1σ, μ+2σ], and significant inconsistency tolerance interval (μ+2σ, +∞). In practice, if the mean μ of the historical deviation data is 0.05 and the standard deviation σ is 0.03, then the high consistency tolerance interval is [0, 0.08], the acceptable inconsistency tolerance interval is (0.08, 0.11], and the significant inconsistency tolerance interval is (0.11, +∞).

[0048] Here, the tolerance range needs to be dynamically adjusted according to environmental changes and data quality. New historical data can be collected every six months, and the statistical measures of the deviation can be recalculated. If the mean or standard deviation of the deviation changes by more than 20%, the tolerance range should be reset. For ecologically sensitive areas such as nature reserves, the tolerance range can be narrowed to increase the strictness of consistency verification. For areas with large environmental fluctuations, the tolerance range can be widened to avoid excessive false alarms.

[0049] The determination of the state consistency verification result is based on the normalized first environment state parameters. Second environmental state parameters The deviation value Δp is compared with the tolerance range. Specifically, when Δp∈[0,μ+1σ], it is judged as a high consistency level, indicating that the environmental state inverted from image data and non-image data is highly consistent, and the ecological rationality and reliability of the fusion result are high, which can be directly used to generate ecological environment monitoring conclusions. When Δp∈(μ+1σ,μ+2σ], it is judged as an acceptable inconsistency level, indicating that the deviation of the two types of parameters exceeds the normal range, but is still within the tolerable range. It may be caused by minor image quality problems, small-amplitude anomalies in non-image data, or short-term environmental fluctuations. The fusion result needs to be corrected in conjunction with anomaly source analysis. When Δp∈(μ+2σ,+∞), it is judged as a significant inconsistency level, indicating that the deviation of the two types of parameters is significant, there is a serious data contradiction, and the reliability of the fusion result is low.

[0050] In practice, if the monitoring target involves multiple environmental state parameters, each parameter needs to be assessed for consistency separately. The final overall consistency level is determined by combining the assessment results of each parameter. The comprehensive assessment rule uses a weighted voting method, assigning a weight to each parameter, with the sum of the weights equal to 1. The weighted score for each level is calculated, and the level with the highest score is the overall consistency level. For example, if the weight of the eutrophication trend parameter is 0.6, it is judged as high consistency; and the weight of the vegetation growth stage offset parameter is 0.4, it is judged as acceptable inconsistency. Then, the weighted score for high consistency is 0.6 × 1 = 0.6, and the weighted score for acceptable inconsistency is 0.4 × 1 = 0.4, resulting in a final overall consistency level of high consistency.

[0051] In practice, when the state consistency verification result is an acceptable inconsistency level or a significant inconsistency level, the method further includes: performing anomaly source analysis based on the deviation between the first environmental state parameter and the second environmental state parameter; the anomaly source analysis includes analyzing the possibility of image data degradation, sudden environmental events, or data loss.

[0052] S5: Based on the consistency verification results, determine the weights of the ecological image data and the non-image environment data in data fusion, or determine the credibility level of the ecological environment monitoring results.

[0053] Here, data sources with higher consistency and stronger data reliability are assigned greater weights; data sources with anomalies and lower reliability are assigned smaller weights. Specifically, when the consistency check result is at a high consistency level, an equal-weighted fusion scheme is adopted, with the weight w1 for ecological image data and the weight w2 for non-image environment data both set to 0.5. When the consistency check result is at an acceptable inconsistency level, the weights are adjusted based on the anomaly source analysis results. Specifically, when the anomaly source is image data degradation, the weight w1 for image data is reduced to 0.3~0.4, and the weight w2 for non-image data is increased to 0.6~0.7. The specific weight values ​​are determined based on the severity of image quality degradation; the more severe the degradation, the smaller the value of w1.

[0054] When the source of the anomaly is a sudden environmental event, the weight w1 of image data is increased to 0.6~0.7, and the weight w2 of non-image data is decreased to 0.3~0.4. For example, if a sudden water pollution event causes a significant increase in the eutrophication trend inverted from the image data, but the non-image data does not reflect this in time, w1 is set to 0.7 and w2 to 0.3.

[0055] When the anomaly originates from missing data or time discrepancies, the weight w2 for non-image data is reduced to 0.3-0.4, and the weight w1 for image data is increased to 0.6-0.7. The specific weight values ​​are determined based on the severity of the missing data or the magnitude of the time difference; the more severe the missing data or the larger the time difference, the better.

[0056] When the consistency check result is of the obvious inconsistency level, if the source of the anomaly is image data degradation, the weight w1 of the image data is reduced to 0.1~0.2, and the weight w2 of the non-image data is increased to 0.8~0.9; if the image data degradation is extremely severe, the weight w1 of the image data is set to 0, and the monitoring result is generated only based on the second environmental state parameters inverted from the non-image data.

[0057] If the source of the anomaly is a sudden environmental event, the weight w1 of the image data is increased to 0.8~0.9, and the weight w2 of the non-image data is decreased to 0.1~0.2. At the same time, the impact of the sudden environmental event is marked in the monitoring results, prompting the user that the results are obtained based on the fusion dominated by image data.

[0058] If the source of the anomaly is missing data or time synchronization, when there is a serious lack of non-image data or a serious time synchronization, the weight w2 of the non-image data is reduced to 0.1~0.2, and the weight w1 of the image data is increased to 0.8~0.9; if the non-image data is completely missing, the weight w2 of the non-image data is set to 0, and the monitoring results are generated only based on the first environmental state parameters inverted from the image data, and the missing data situation is marked.

[0059] If the source of the anomaly cannot be determined after analysis, a conservative fusion scheme is adopted, setting the weight of both types of data to 0.5, and marking in the monitoring results that the consistency check is obviously inconsistent and the source is unknown, prompting the user that the result needs to be further confirmed by on-site verification.

[0060] The trust level is divided into five levels, from highest to lowest: extremely high trust, high trust, medium trust, low trust, and extremely low trust. Specifically, when the consistency check result is high consistency and there are no indications of any anomaly source, the trust level is extremely high trust. When the consistency check result is high consistency, but there are minor image quality issues or small data gaps, or the consistency check result is at an acceptable inconsistency level, the anomaly source is clear, and the deviation has been effectively corrected through weight adjustment, the trust level is high trust. When the consistency check result is at an acceptable inconsistency level, the anomaly source is clear, but the deviation cannot be completely corrected through weight adjustment; or the consistency check result is at a significant inconsistency level, the anomaly source is clear, and the fusion result is primarily based on a reliable data source, the trust level is medium trust. When the consistency check result is at a significant inconsistency level, the anomaly source is clear, but both types of data sources have some degree of reliability issues; or the anomaly source cannot be completely determined, but the risk has been minimized as much as possible through weight adjustment, the trust level is low trust. When the consistency check result is of the obviously inconsistent level, the source of the anomaly cannot be determined, and both types of data sources have serious reliability problems; or when the fusion result seriously conflicts with the known ecological evolution laws, the credibility level is extremely low.

[0061] Figure 2 An ecological environment monitoring system 200 with multi-data fusion is shown. This device embodiment is similar to... Figure 1 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.

[0062] like Figure 2 As shown, the ecological environment monitoring system 200 provided in this application embodiment includes the following modules: Data acquisition module 201 is used to acquire ecological image data and corresponding non-image environmental data of the target area, wherein the non-image environmental data includes at least one of meteorological data, hydrological data or historical statistical data; The first state inversion module 202 is used to perform environmental state inversion based on the ecological image data to obtain a first environmental state parameter; the first environmental state parameter is a continuous or semi-continuous numerical parameter describing the continuous evolution trend of ecological elements. The second state inversion module 203 is used to perform environmental state inversion based on the non-image environment data to obtain a second environmental state parameter; the second environmental state parameter corresponds to the first environmental state parameter in physical meaning. The verification module 204 is used to perform a state consistency verification between the first environmental state parameter and the second environmental state parameter to obtain a consistency verification result; the state consistency verification includes determining whether the deviation between the first environmental state parameter and the second environmental state parameter is within a preset tolerance range. The monitoring module 205 is used to determine the weight of the ecological image data and the non-image environment data in data fusion based on the consistency verification result, or to determine the credibility level of the ecological environment monitoring results.

[0063] Based on the same inventive concept, this application also provides an electronic device. The method corresponding to the electronic device can be the method in the foregoing embodiments, and its problem-solving principle is similar to that method. The electronic device provided in this application includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the methods and / or technical solutions of the foregoing embodiments of this application.

[0064] The electronic device can be a user device, or a device formed by integrating user devices and network devices through a network, or it can be an application running on the aforementioned devices. The user device includes, but is not limited to, various terminal devices such as computers, mobile phones, tablets, smartwatches, and wristbands. The network device includes, but is not limited to, network hosts, single network servers, multiple network server sets, or cloud computing-based computer sets, and can be used to implement some processing functions when setting an alarm clock. Here, the cloud consists of a large number of hosts or network servers based on cloud computing. Cloud computing is a type of distributed computing, consisting of a virtual computer composed of a group of loosely coupled computer sets.

[0065] Figure 3 The diagram illustrates the structure of an apparatus suitable for implementing the methods and / or technical solutions in the embodiments of this application. The apparatus 300 includes a Central Processing Unit (CPU) 301, which can perform various appropriate actions and processes based on a program stored in a Read Only Memory (ROM) 302 or a program loaded from a storage portion 308 into a Random Access Memory (RAM) 303. The RAM 303 also stores various programs and data required for system operation. The CPU 301, ROM 302, and RAM 303 are interconnected via a bus 304. An Input / Output (I / O) interface 305 is also connected to the bus 304.

[0066] The following components are connected to I / O interface 305: an input section 306 including a keyboard, mouse, touchscreen, microphone, infrared sensor, etc.; an output section 307 including a cathode ray tube (CRT), liquid crystal display (LCD), LED display, OLED display, etc., and speakers, etc.; a storage section 308 including one or more computer-readable media such as hard disk, optical disk, magnetic disk, semiconductor memory, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 309 performs communication processing via a network such as the Internet.

[0067] In particular, the methods and / or embodiments in this application can be implemented as computer software programs. For example, the embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowchart. When the computer program is executed by the central processing unit (CPU) 301, it performs the functions defined in the methods of this application.

[0068] Another embodiment of this application provides a computer-readable storage medium having computer program instructions stored thereon, which can be executed by a processor to implement the methods and / or technical solutions of any one or more embodiments of this application described above.

[0069] Specifically, this embodiment may employ any combination of one or more computer-readable media. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device.

[0070] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including—but not limited to—electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0071] The program code contained on a computer-readable medium may be transmitted using any suitable medium, including—but not limited to—wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0072] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as "C" or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0073] The flowcharts or block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of devices, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-specific system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0074] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0075] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or page components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between devices or units through some interfaces, and may be electrical, mechanical, or other forms.

[0076] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0077] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in a combination of hardware and software functional units.

[0078] The integrated units implemented as software functional units described above can be stored in a computer-readable storage medium. These software functional units, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0079] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

[0080] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in a device claim may also be implemented by a single unit or device through software or hardware. The terms "first," "second," etc., are used to indicate names and do not indicate any specific order.

Claims

1. A multi-data fusion method for ecological environment monitoring, characterized in that, Includes the following steps: S1: Acquire ecological image data and corresponding non-image environmental data of the target area, wherein the non-image environmental data includes at least one of meteorological data, hydrological data or historical statistical data; S2: Based on the ecological image data, perform environmental state inversion to obtain the first environmental state parameter; the first environmental state parameter is a continuous or semi-continuous numerical parameter describing the continuous evolution trend of ecological elements; S3: Based on the non-image environment data, perform environmental state inversion to obtain a second environmental state parameter; the second environmental state parameter corresponds to the first environmental state parameter in physical meaning; S4: Perform a state consistency check on the first environmental state parameter and the second environmental state parameter to obtain a consistency check result; the state consistency check includes determining whether the deviation between the first environmental state parameter and the second environmental state parameter is within a preset tolerance range; S5: Based on the consistency verification results, determine the weights of the ecological image data and the non-image environment data in data fusion, or determine the credibility level of the ecological environment monitoring results.

2. The ecological environment monitoring method according to claim 1, characterized in that, In step S2, the environmental state inversion based on the ecological image data includes: preprocessing the ecological image data, the preprocessing including geometric correction and radiometric normalization; extracting image features from the preprocessed ecological image data, and mapping the image features to the first environmental state parameter.

3. The ecological environment monitoring method according to claim 2, characterized in that, The step of extracting image features from the preprocessed ecological image data includes: using a convolutional neural network to extract intermediate layer feature vectors; the step of mapping the image features to a first environmental state parameter includes: performing a linear or nonlinear transformation on the intermediate layer feature vectors to obtain the first environmental state parameter.

4. The ecological environment monitoring method according to claim 1, characterized in that, In step S3, the environmental state inversion based on the non-image environment data includes: unifying the time scale and processing outliers in the non-image environment data; processing the processed non-image environment data using a trend analysis model or a conditional constraint model to obtain the second environmental state parameters; wherein, the trend analysis model is a time series sliding window statistical model or a trend fitting model; and the conditional constraint model is a logical judgment model constructed based on ecological experience rules.

5. The ecological environment monitoring method according to claim 1, characterized in that, The state consistency verification result is a multi-level result, including high consistency level, acceptable inconsistency level, and obvious inconsistency level.

6. The ecological environment monitoring method according to claim 1, characterized in that, The first environmental state parameter and the second environmental state parameter include at least one of the following: water eutrophication trend parameter, vegetation growth stage offset parameter, or surface humidity change direction parameter.

7. A multi-data fusion ecological environment monitoring system, characterized in that, Includes the following modules: The data acquisition module is used to acquire ecological image data and corresponding non-image environmental data of the target area. The non-image environmental data includes at least one of meteorological data, hydrological data, or historical statistical data. The first state inversion module is used to perform environmental state inversion based on the ecological image data to obtain a first environmental state parameter; the first environmental state parameter is a continuous or semi-continuous numerical parameter describing the continuous evolution trend of ecological elements. The second state inversion module is used to perform environmental state inversion based on the non-image environment data to obtain a second environmental state parameter; the second environmental state parameter corresponds to the first environmental state parameter in physical meaning. The verification module is used to perform a state consistency verification between the first environmental state parameter and the second environmental state parameter, and obtain a consistency verification result. The state consistency verification includes determining whether the deviation between the first environmental state parameter and the second environmental state parameter is within a preset tolerance range. The monitoring module is used to determine the weights of the ecological image data and the non-image environmental data in data fusion, or to determine the reliability level of the ecological environment monitoring results, based on the consistency verification results.

8. An electronic device, characterized in that, The electronic device includes: at least one processor; a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.

9. A computer-readable medium having computer program instructions stored thereon, characterized in that, The computer program instructions can be executed by a processor to implement the method as described in any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-6.

Images