Methods, apparatus, devices, and media for determining a de-centralized sensor network

Abstract

Embodiments of the present application provide a method, device, equipment and medium for determining a decentralized sensor network, comprising: obtaining working state data and networking information data of sensors in an Internet of Things; screening working state sample data and networking information sample data from the working state data and the networking information data respectively according to preset screening rules; clustering the working state sample data and the networking information sample data by using a clustering algorithm to obtain a plurality of sensor clustering clusters; and selecting at least one sensor clustering cluster satisfying a preset correlation relationship from the plurality of sensor clustering clusters to obtain the decentralized sensor network. Embodiments of the present application solve the problem of complexity of Internet of Things management.

Description

Technical Field

[0001] This application belongs to the field of transmission and transmission technology, and in particular relates to a method, apparatus, device and medium for determining a decentralized sensor network. Background Technology

[0002] The Internet of Things (IoT) provides services based on cloud servers. Cloud servers receive and process data, then send the received data and processing results to terminal devices. As the demand for real-time data processing increases, centralized processing systems face the problem of struggling to effectively handle the growing data processing needs. Therefore, distributed processing is required, that is, decentralizing centralized processing systems to alleviate the pressure on them.

[0003] One approach to building a decentralized sensor network for the Internet of Things (IoT) involves dividing the IoT into multiple layers, with each layer's sensors processing different types of data, thus improving data processing efficiency. However, multi-layered decentralized sensor networks require management and functional maintenance for each layer, increasing the complexity of IoT management and hindering efficient network management and operation. Summary of the Invention

[0004] This application provides a method, apparatus, device, and medium for determining a decentralized sensor network, which solves the problem of the complexity of Internet of Things management.

[0005] In a first aspect, embodiments of this application provide a method for determining a decentralized sensor network, the method comprising:

[0006] Acquire operational status data and network information data from sensors in the Internet of Things (IoT);

[0007] Select sample data of working status data and sample data of network information data from the working status data and network information data respectively according to the preset filtering rules;

[0008] Clustering algorithms were used to cluster the working status sample data and network information sample data to obtain multiple sensor clusters;

[0009] A decentralized sensor network is obtained by selecting at least one sensor cluster that satisfies a preset association relationship from multiple sensor clusters.

[0010] In one optional implementation of the first aspect, the preset filtering rules include a first rule and a second rule. The first rule includes a time period, first filtering information, and a first sampling frequency. The second rule includes a classification rule and a second sampling frequency.

[0011] Select sample data for working status and network information from the working status data and network information data respectively according to preset filtering rules, including:

[0012] The working status data is filtered according to the time period and the first screening information to obtain the first working status data, and the first working status data is sampled according to the first sampling frequency to obtain the working status sample data.

[0013] The network information data is classified according to the classification rules to obtain classified network information data, and then sampled according to the second sampling frequency to obtain network information sample data.

[0014] In one optional implementation of the first aspect, sampling the first operating state data according to a first sampling frequency to obtain operating state sample data includes:

[0015] The first working state data is sampled according to the first sampling frequency to obtain the second working state data;

[0016] Delete the target field's working status data and the working status data that does not meet the target logical requirements from the second working status data to obtain working status sample data.

[0017] In one optional implementation of the first aspect, the first sampling frequency includes a random sampling frequency and a target fixed sampling frequency; sampling the first operating state data according to the first sampling frequency to obtain the second operating state data includes:

[0018] The first working state data is sampled according to the random sampling frequency to obtain random working state data;

[0019] The first working state data is sampled according to the target fixed sampling frequency to obtain fixed sampling working state data;

[0020] Duplicate working state data are removed from both the random working state data and the fixed sample working state data to obtain the second working state data.

[0021] In one optional implementation of the first aspect, the classification rules include multiple classification value ranges; the network information data is classified according to the classification rules to obtain classified network information data, including:

[0022] The network information data is classified according to multiple value ranges to obtain the network information data within each value range.

[0023] In an optional implementation of the first aspect, before classifying the network information data according to multiple classification value ranges to obtain the classified network information data within each classification value range interval, the method further includes:

[0024] A frequency distribution curve was plotted based on network information data.

[0025] The classification range is defined by the inflection point of the frequency distribution curve from low to high.

[0026] In one optional implementation of the first aspect, a clustering algorithm is used to cluster the working status data and network information sample data to obtain multiple sensor clusters, including:

[0027] The K-nearest centroid nearest neighbor classification algorithm was used to cluster the working status sample data and the network information sample data to obtain multiple sensor clusters.

[0028] In an optional implementation of the first aspect, before selecting at least one sensor cluster satisfying a preset association relationship from the sensor clusters to obtain a decentralized sensor network, the method further includes:

[0029] Acquire the operational status and network connectivity information characteristics of IoT sensors;

[0030] Filter the first set of features in the work status information features according to the aggregation rules, and filter the second aggregated features in the network information features respectively;

[0031] Based on the first aggregation feature and the second aggregation feature, a knowledge graph of working status information features and network information features is created;

[0032] Calculate the correlation value of each feature in the knowledge graph;

[0033] Identify target features whose correlation values ​​are greater than the correlation threshold;

[0034] Preset association relationships are generated based on the relationships between target features.

[0035] In an optional implementation of the first aspect, the step of selecting at least one sensor cluster satisfying a preset association relationship from multiple sensor clusters to obtain a decentralized sensor network includes:

[0036] For each of the multiple sensor clusters, obtain a list of functional categories for the sensor cluster, the list of functional categories including functional description information and sensor data description information;

[0037] The first keyword is extracted based on the functional description information, and the second keyword is extracted based on the sensor data description information;

[0038] Calculate the similarity between the first keyword and the second keyword, and establish association relationships by filtering target keywords whose similarity is greater than a similarity threshold;

[0039] Select at least one sensor cluster whose association relationship satisfies a preset association relationship from the plurality of sensor clusters.

[0040] Secondly, embodiments of this application provide an apparatus for determining a decentralized sensor network, the apparatus comprising:

[0041] The acquisition module is used to acquire the working status data and network information data of sensors in the Internet of Things.

[0042] The filtering module is used to filter working status sample data and network information sample data from working status data and network information data according to preset filtering rules, respectively.

[0043] The clustering module is used to cluster working status sample data and network information sample data using clustering algorithms to obtain multiple sensor clusters;

[0044] The selection module is used to select at least one sensor cluster that satisfies a preset association relationship from multiple sensor clusters to obtain a decentralized sensor network.

[0045] In a third aspect, an electronic device is provided, comprising: a memory for storing computer program instructions; and a processor for reading and executing the computer program instructions stored in the memory to perform a method for determining a decentralized sensor network provided in any optional embodiment of the first aspect.

[0046] Fourthly, a computer storage medium is provided, on which computer program instructions are stored, which, when executed by a processor, implement the method for determining a decentralized sensor network provided by any optional embodiment of the first aspect.

[0047] Fifthly, a computer program product is provided, wherein instructions in the computer program product, when executed by a processor of an electronic device, cause the electronic device to perform a method for determining a decentralized sensor network provided by any optional embodiment of the first aspect.

[0048] In this embodiment, operational status data and network information data of sensors in the Internet of Things (IoT) can be acquired. Sample operational status data and sample network information data are then filtered from this data according to preset filtering rules. A clustering algorithm can then be used to cluster the sample operational status data and sample network information data, resulting in multiple sensor clusters. By selecting at least one sensor cluster that satisfies a preset association relationship from these clusters, a decentralized sensor network is obtained. This allows for the determination of the decentralized sensor cluster networking method, reducing the complexity of IoT management and enabling flexible and efficient implementation of IoT network functions. Attached Figure Description

[0049] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a flowchart illustrating a method for determining a decentralized sensor network according to an embodiment of this application;

[0051] Figure 2 This is a flowchart illustrating another method for determining a decentralized sensor network provided in an embodiment of this application;

[0052] Figure 3 This is a flowchart illustrating another method for determining a decentralized sensor network provided in an embodiment of this application;

[0053] Figure 4 This is a schematic diagram of the structure of a device for determining a decentralized sensor network provided in an embodiment of this application;

[0054] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0055] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0056] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0057] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0058] To address the problems in the prior art, this application provides a method, apparatus, device, and medium for determining a decentralized sensor network. This method acquires operational status data and network information data from sensors in the Internet of Things (IoT), and filters operational status sample data and network information sample data from these data according to preset filtering rules. Then, a clustering algorithm is used to cluster the operational status sample data and network information sample data to obtain multiple sensor clusters. Finally, at least one sensor cluster satisfying a preset association relationship is selected from these multiple sensor clusters to obtain the decentralized sensor network. This allows for the determination of the decentralized sensor cluster networking method, reducing the complexity of IoT management and enabling flexible and efficient implementation of IoT network functions.

[0059] It should be noted that the method for determining a decentralized sensor network provided in this application embodiment can be executed by a device for determining a decentralized sensor network, or by a control module within the device for determining a decentralized sensor network that performs the method for determining a decentralized sensor network. This application embodiment uses the example of a device for determining a decentralized sensor network performing the method for determining a decentralized sensor network to illustrate the method for determining a decentralized sensor network provided in this application embodiment.

[0060] The method for determining a decentralized sensor network provided in this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0061] Figure 1 This is a flowchart illustrating a method for determining a decentralized sensor network provided in an embodiment of this application.

[0062] like Figure 1 As shown, the execution entity of this method for determining a decentralized sensor network can be a device for determining a decentralized sensor network, and the method may specifically include the following steps:

[0063] S110 acquires the working status data and network information data of sensors in the Internet of Things.

[0064] The aforementioned sensors may include multiple sensors in the Internet of Things (IoT), without specific limitations. The aforementioned operational status data may include N1 operational status data points from each of the multiple sensors, where N1 is a positive integer, without further limitations. Correspondingly, the aforementioned network information data may include N2 network information data points from each of the aforementioned sensors, where N2 is a positive integer, without further limitations.

[0065] It should also be noted that each of the above-mentioned working status data includes data on temperature, pressure, gas, humidity, flow rate, light intensity, rainfall, vibration, and rotation speed obtained by the corresponding sensor. Correspondingly, each of the above-mentioned network information data may include data such as network speed, bandwidth, and network address, without specific limitations here.

[0066] S120, respectively filter working status sample data and network information sample data from working status data and network information data according to preset filtering rules.

[0067] Among them, the preset filtering rules can be filtering rules that are set in advance based on actual experience or circumstances. Specifically, the preset filtering rules can include filtering rules for filtering working status sample data from working status data, and filtering rules for filtering network information sample data from network information data.

[0068] In addition, the aforementioned operational status sample data may include M1 operational status data points for each sensor, where M1 ≤ N1 and M1 is a positive integer, without specific limitations. Similarly, the aforementioned network information sample data may include M2 ​​network information data points for each sensor, where M2 ≤ N2 and M2 is a positive integer, without specific limitations.

[0069] Specifically, the device for determining a decentralized sensor network can, for each sensor, select working status sample data from the working status data corresponding to that sensor according to a preset selection rule, and select network information sample data from the network information data corresponding to that sensor.

[0070] S130 uses a clustering algorithm to cluster the working status sample data and network information sample data to obtain multiple sensor clusters.

[0071] The clustering algorithms mentioned above can be selected based on the actual situation, and no specific limitations are imposed here. Furthermore, each of the multiple sensor clusters mentioned above can include at least one sensor, and no further restrictions are imposed here.

[0072] Specifically, the device for determining decentralized sensors can use clustering algorithms to cluster the aforementioned working status sample data and network information sample data, thereby clustering the aforementioned sensors and obtaining multiple sensor clusters.

[0073] S140: Select at least one sensor cluster that satisfies a preset association relationship from multiple sensor clusters to obtain a decentralized sensor network.

[0074] Among them, the preset association relationship can be a relationship set based on actual situation or experience, which is used to filter sensor clusters that meet actual needs from multiple sensor clusters.

[0075] After obtaining multiple sensor clusters, the device for determining a decentralized sensor network can select at least one sensor cluster that satisfies a preset association relationship from the multiple sensor clusters to obtain a decentralized sensor network.

[0076] In this embodiment, operational status data and network information data of sensors in the Internet of Things (IoT) can be acquired. Sample operational status data and sample network information data are then filtered from this data according to preset filtering rules. A clustering algorithm can then be used to cluster the sample operational status data and sample network information data, resulting in multiple sensor clusters. By selecting at least one sensor cluster that satisfies a preset association relationship from these clusters, a decentralized sensor network is obtained. This allows for the determination of the decentralized sensor cluster networking method, reducing the complexity of IoT management and enabling flexible and efficient implementation of IoT network functions.

[0077] In one embodiment, the aforementioned preset filtering rules may include a first rule and a second rule. The first rule may include a time period, first filtering information, and a first sampling frequency, used to filter working status sample data from working status data. The second rule may include a classification rule and a second sampling frequency, used to filter network information sample data from network information sample data. Based on this, as... Figure 2 As shown, the S120 mentioned above may specifically include the following steps:

[0078] S210, the working status data is filtered according to the time period and the first filtering information to obtain the first working status data, and the first working status data is sampled according to the first sampling frequency to obtain the working status sample data.

[0079] The duration of this time period can be determined according to specific circumstances; for example, it can be set to one day or one week, without specific limitations. The aforementioned first filtering information can be used to filter out data in the working status data that is greater than or within the data threshold range. The aforementioned data threshold or data threshold range can be determined based on practical experience or circumstances, without further limitations. Additionally, the aforementioned first sampling frequency can be preset based on practical experience or circumstances, without further limitations.

[0080] It should also be noted that the above first working state data may include L1 working state data, M1≤L1≤N1, where L1 is a positive integer and is not specifically limited here.

[0081] Specifically, since the aforementioned preset filtering rules may include a first rule, which may include a time period, first filtering information, and a first sampling frequency, the device for determining the decentralized sensor network can filter the working status data based on the time period and the first filtering information to obtain the first working status data, and then sample the first working status data according to the first sampling frequency to obtain working status sample data.

[0082] It should be noted that, in order to ensure the accuracy of the filtered data, for the working status data measured by a single sensor, the device for determining the decentralized sensor network can first determine the normal range of values ​​for the working status data based on the historical working status data of that sensor, and then set the data threshold or data threshold range according to the normal range of values ​​for the working status data.

[0083] For multiple operating status data obtained by multiple sensors or sensor groups, the device for determining the decentralized sensor network can first determine an error compensation value based on the historical normal average value of the historical operating status data measured by each sensor. This error compensation value is used to compensate the currently acquired operating status data to the corresponding historical normal average value. In this way, the corresponding operating status data can be corrected based on the error compensation value to obtain multiple corrected operating status data.

[0084] S220: Classify the network information data according to the classification rules to obtain classified network information data, and sample the classified network information data according to the second sampling frequency to obtain network information sample data.

[0085] The aforementioned classification rules are used to categorize the network information data. Classified network information data based on these rules can include general network information data, qualified network information data, and unqualified network information data. Furthermore, the second sampling frequency can also be a pre-set sampling frequency based on practical experience or circumstances; no further limitations are imposed here.

[0086] Specifically, since the aforementioned preset filtering rules may include a second rule, which may include a classification rule and a second sampling frequency, the device for determining the decentralized sensor network can classify the network information data based on the classification rule to obtain classified network information data, and then sample the classified network information data obtained by the classification according to the second sampling frequency to obtain network information sample data.

[0087] In this embodiment, the preset filtering rules can include a first rule and a second rule. The first rule can include a time period, first filtering information, and a first sampling frequency, which are used to filter working status sample data from working status data. The second rule can include a classification rule and a second sampling frequency, which are used to filter network information sample data from network information data. Thus, sample data that meets the requirements can be effectively filtered from working status data and network information data.

[0088] In one embodiment, the step of sampling the first operating state data according to the first sampling frequency to obtain operating state sample data may specifically include the following steps:

[0089] The first working state data is sampled according to the first sampling frequency to obtain the second working state data;

[0090] Delete the target field's working status data and the working status data that does not meet the target logical requirements from the second working status data to obtain working status sample data.

[0091] The second working state data may include one working state data point K1, where M1≤K1≤L1≤N1, and K1 is a positive integer without further restrictions. Each working state data point in the second working state data may include multiple fields, and the aforementioned target field may be any number of these fields without further restrictions. Furthermore, the aforementioned target logical requirement information may be pre-set logical requirement information based on practical experience or circumstances without further restrictions.

[0092] Specifically, the device that determines the decentralized sensor network can sample the first operating state data at a first sampling frequency to obtain the second operating state data. Then, the operating state data can be cleaned by deleting the target field operating state data and the operating state data that does not meet the target logical requirements, that is, deleting unnecessary data fields in the second operating state data and cleaning the data in the second operating state data by using simple logical reasoning to find that there are logical errors, so as to obtain the operating state sample data.

[0093] In this embodiment, after sampling the first working state data at a first sampling frequency to obtain the second working state data, the second working state data can be cleaned by deleting unnecessary data fields or data with logical errors. Thus, by reviewing and verifying the sampled working state data, errors can be corrected, ensuring data consistency and providing accurate and effective working state data samples.

[0094] In one embodiment, the first sampling frequency mentioned above may include a random sampling frequency and a target fixed sampling frequency. The random sampling frequency may be based on a preset duration as sampling points, randomly determining the corresponding working state data from the first working state data. This preset duration may be a duration pre-set based on actual experience or circumstances; for example, the preset duration may be one hour, without specific limitation. Furthermore, the target fixed sampling frequency may be determined by determining the corresponding working state data in the first working state data according to an arithmetic or geometric sequence.

[0095] Based on this, the steps mentioned above, which involve sampling the first operating state data according to the first sampling frequency to obtain the second operating state data, may specifically include the following steps:

[0096] The first working state data is sampled according to the random sampling frequency to obtain random working state data;

[0097] The first working state data is sampled according to the target fixed sampling frequency to obtain fixed sampling working state data;

[0098] Duplicate working state data are removed from both the random working state data and the fixed sample working state data to obtain the second working state data.

[0099] Specifically, since the first sampling frequency mentioned above can include a random sampling frequency and a target fixed sampling frequency, the device for determining the decentralized sensor network can first sample the first operating state data according to the random sampling frequency to obtain random operating state data, and then sample the first operating state data according to the target fixed sampling frequency to obtain fixed sampling operating state data. Based on this, by deleting duplicate operating state data from the random operating state data and the fixed sampling operating state data, the second operating state data is obtained.

[0100] In this embodiment, a combination of random and fixed sampling methods can be used to sample the first working state data to obtain the second working state data. This facilitates obtaining accurate and effective sample data subsequently.

[0101] In one embodiment, the classification rules mentioned above may include multiple classification value ranges. Based on this, the steps of classifying network information data according to the classification rules to obtain classified network information data may specifically include the following steps:

[0102] The network information data is classified according to multiple value ranges to obtain the network information data within each value range.

[0103] Specifically, since the classification rules mentioned above can include multiple classification value ranges, the device for determining a decentralized sensor network can classify network information data according to multiple classification value ranges to obtain classified network information data belonging to each of the above classification value ranges.

[0104] In this embodiment, since the classification rules can include multiple classification value ranges, network information data can be classified based on these multiple classification value ranges to obtain classified network information data belonging to each of the aforementioned classification value ranges. Thus, network information data can be accurately and effectively classified, thereby accurately obtaining classified network information data.

[0105] In one embodiment, before classifying the network information data according to multiple classification value ranges to obtain the classified network information data belonging to each classification value range, the method for determining the decentralized sensor network mentioned above may further include the following steps:

[0106] A frequency distribution curve was plotted based on network information data.

[0107] The classification range is defined by the inflection point of the frequency distribution curve from low to high.

[0108] Specifically, the device for determining a decentralized sensor network can plot a frequency distribution curve based on network information data, and then use the inflection point of the frequency distribution curve from low to high as the boundary to divide the range of classification values.

[0109] It should also be noted that the method for determining the decentralized sensor provided in this application embodiment can clean the above-mentioned network information data to obtain network information samples. The method of cleaning the above-mentioned network information data is the same as the method of cleaning the working status data, and will not be elaborated further here.

[0110] In this embodiment, the changes in network information data can be visually displayed by plotting a frequency distribution curve of the network information data, thereby accurately determining the range of classification values ​​so that the network information data can be accurately distinguished in the future.

[0111] In one embodiment, the above-mentioned S140 may specifically include the following steps:

[0112] The K-nearest centroid nearest neighbor classification algorithm was used to cluster the working status sample data and the network information sample data to obtain multiple sensor clusters.

[0113] Specifically, the device for determining a decentralized sensor network can use the K-nearest centroid nearest neighbor classification algorithm to cluster operational status sample data and network information sample data to obtain multiple sensor clusters.

[0114] In one embodiment, the step of using the K-nearest centroid nearest neighbor classification algorithm to cluster the working state sample data and network information sample data to obtain multiple sensor clusters may specifically include the following steps:

[0115] Step 1: For the working status sample data or network information sample data, obtain the training sample set and test samples; calculate the Euclidean distance between the test sample and all training samples in the training sample set.

[0116] Step 2: Calculate the K nearest centroid neighbors of each training sample with respect to the test sample using the nearest centroid neighbor criterion; calculate the K local means of the K nearest centroid neighbors.

[0117] Step 3: Calculate the cooperative representation coefficients corresponding to the K nearest centroid local mean points in each class, set the cooperative representation coefficients as the weights of the nearest centroid local mean points, and perform weighted representation on the nearest centroid local mean points.

[0118] Step 4: Calculate the residual between the test sample and the sample represented by the K nearest centroid local mean points of each class, and assign the class label corresponding to the smallest residual to the test sample.

[0119] Step 5: Based on the category labels, classify the working status sample dataset and the network information sample dataset respectively to obtain several sensor clusters.

[0120] Thus, the noise problem existing in the standard k-nearest neighbor algorithm can be taken into account. The embodiments of this application can more intuitively reflect the geometric distribution of different neighbors around the test sample by calculating the nearest centroid neighbors of the test sample.

[0121] Furthermore, if only the nearest centroid neighbor method and local mean are used to calculate nearest neighbors for classification, setting a fixed K value for all test samples during the classification process can lead to K-value sensitivity issues. This is because the number of local mean nearest neighbors selected for each class and their corresponding contribution values ​​should be different for different test samples.

[0122] This scheme first finds the K nearest centroid neighbors of each training sample with respect to the test sample, and then finds the K local multimean neighbors through the K nearest centroid neighbors.

[0123] Meanwhile, in order to overcome the problems of neighbors with different contribution values ​​having the same weight and the simple maximum voting principle in the k-nearest neighbor algorithm, this scheme further calculates the cooperative representation coefficients corresponding to the K local multi-means neighbors. The cooperative representation coefficients are equivalent to assigning different weights to the neighbors.

[0124] Finally, based on the category residual as the classification decision function of the algorithm, the category residuals of the test sample and the predicted sample after the local multi-mean nearest neighbor cooperative representation of each class are calculated respectively, and the category label corresponding to the minimum residual is assigned to the test sample.

[0125] This embodiment can overcome the problems caused by noise points and K-value sensitivity. At the same time, it uses cooperative representation coefficients to assign different weights to different nearest neighbors, so that different neighbors have different contribution proportions in the classification process, which can improve the effectiveness and robustness of the classification algorithm.

[0126] In this embodiment, the noise problem present in the conventional K-nearest neighbor algorithm is taken into account. The K-nearest centroid nearest neighbor classification algorithm is used to cluster the working state sample data and network information sample data, resulting in multiple sensor clusters. This avoids the influence of noise and allows for the accurate acquisition of multiple sensor clusters.

[0127] In one embodiment, prior to S140 above, as Figure 3 As shown, the method for determining a decentralized sensor network mentioned above may specifically include the following steps:

[0128] Acquire the operational status and network connectivity information characteristics of IoT sensors;

[0129] Filter the first set of features in the work status information features according to the aggregation rules, and filter the second aggregated features in the network information features respectively;

[0130] Based on the first aggregation feature and the second aggregation feature, a knowledge graph of working status information features and network information features is created;

[0131] Calculate the correlation value of each feature in the knowledge graph;

[0132] Identify target features whose correlation values ​​are greater than the correlation threshold;

[0133] Relationships are generated based on the relationships between target features.

[0134] The clustering rules can be based on the functions to be achieved by the Internet of Things (IoT), integrating sensor features. For example, temperature, humidity, and light intensity data can be integrated to obtain a first aggregated feature, which can comprehensively reflect data features of the atmospheric, greenhouse, or indoor environment. Network information features are similar; for example, network speed and bandwidth features can be integrated to reflect network performance data features, generating a second aggregated feature, without specific limitations here. Furthermore, the association thresholds mentioned above can be thresholds set based on practical experience or circumstances, and are not subject to further restrictions here.

[0135] Specifically, the device of the decentralized sensor network is able to acquire the working status information features and network information features of IoT sensors. Then, it can filter the first set of features in the working status information features and the second cluster features in the network information features according to the aggregation rules. Then, it can establish the relationship between the working status information features and the network information features based on the first cluster features and the second aggregation features, that is, create a knowledge graph of the working status information features and the network information features. Then, it can calculate the correlation value of each feature in the knowledge graph, determine the target features with a correlation value greater than the correlation threshold, and generate a preset correlation relationship based on the relationship between the target features.

[0136] Furthermore, it should be noted that after obtaining the first clustering feature and the second aggregation feature, based on preset association rules, a first association feature from the working status information features and a second association feature from the network information features can be selected. The union of the first aggregation feature and the second aggregation feature is taken as the first feature; the union of the first association feature and the second association feature is taken as the second feature. The sensor clusters corresponding to the sensor working status information features, network information features, the first feature, and the second feature are then assigned to a blank cluster, generating several sensor cluster clusters. For example, according to the requirements of the monitored environmental data features, the humidity feature, which is of the same class as the temperature feature and reflects the environmental data status, can be selected as the association feature of the temperature feature.

[0137] In this embodiment, the technology of knowledge graphs can be used to generate preset association relationships, which effectively improves the accuracy and comprehensiveness of subsequent association feature selection.

[0138] In one embodiment, such as Figure 3 As shown, the step of selecting at least one sensor cluster that satisfies a preset association relationship from multiple sensor clusters to obtain a decentralized sensor network may specifically include the following steps:

[0139] S310: For each sensor cluster in multiple sensor clusters, obtain a list of functional categories for the sensor clusters.

[0140] S320 extracts the first keyword based on the functional description information and the second keyword based on the sensor data description information.

[0141] S330, calculate the similarity between the first keyword and the second keyword, and select target keywords with similarity greater than the similarity threshold to establish association relationships.

[0142] S340, Select at least one sensor cluster from multiple sensor clusters whose association relationship satisfies a preset association relationship.

[0143] The aforementioned list of functional categories may include functional description information and sensor data description information. Functional description information can describe the functions the sensor can perform, while sensor data description information can describe the sensor's own data; no specific limitations are imposed here. The similarity threshold can be a pre-set threshold based on actual conditions or experience; no further limitations are imposed here.

[0144] Specifically, the device defining the decentralized sensor network can obtain a list of functional categories for each of the multiple sensor clusters mentioned above. It extracts a first keyword based on functional description information and a second keyword based on sensor data description information. Then, it can calculate the similarity between the first and second keywords. Since each sensor cluster can include multiple sensors, it can calculate a one-to-one similarity with multiple sensors. Target keywords with similarity greater than a similarity threshold can then be selected, and association relationships established. Finally, at least one sensor cluster whose association relationship satisfies a preset association relationship can be selected from the multiple sensor clusters. It should be noted that the association model refers to the branch functions to be implemented in the Internet of Things, such as environmental data acquisition; extracting keywords describing environmental data such as temperature and humidity; associating these keywords with sensor types, i.e., with temperature sensors and humidity sensors, ultimately generating an association model between the environmental data acquisition function and temperature and humidity sensors; the association model is established according to different branch functions; the similarity threshold refers to the similarity relationship between keywords such as temperature, body temperature, and room temperature, set to satisfy a certain number of identical keywords.

[0145] IoT platforms primarily employ centralized rule engines to determine and execute rules, thereby enabling functions such as data forwarding and device linkage. However, in large-scale IoT scenarios, there are issues with high computational pressure and resource consumption during rule determination and execution.

[0146] This application adopts distributed rules. A rule base is set up in the IoT cloud, a rule engine is deployed on the edge gateway, the rule engine executes the rules issued by the cloud, and then the gateway connects to the sensor cluster network.

[0147] In this embodiment, by obtaining a list of functional categories for each sensor cluster in multiple sensor clusters, and based on the list of functional categories, the association relationship of the sensor cluster can be determined. Then, at least one sensor cluster whose association relationship satisfies a preset association relationship can be selected from multiple sensor clusters. In this way, a decentralized sensor network can be flexibly and efficiently constructed according to actual needs.

[0148] Based on this, in one embodiment, after obtaining the decentralized sensor network, sensor networking can be performed based on the new functions of the decentralized sensor network, which may specifically include the following steps:

[0149] Step 1: Obtain new functional description information of the decentralized sensor network and extract new information keywords to form a keyword library; divide the keyword library into a sub-library of working status data keywords and a sub-library of network information data keywords.

[0150] Step 2: Based on the preset keyword search rules, search for 3 first sensor clusters with high relevance in the working status data keyword sub-library; search for 3 second sensor clusters with high relevance in the network information data keyword sub-library; combine the first sensor clusters and the second sensor clusters to generate 9 sensor cluster combination clusters.

[0151] Step 3: Based on the preset test standards, test the network performance of the combined cluster of 9 sensor clusters, and network the sensors in the combined cluster of sensor clusters with the best network performance test results.

[0152] In one embodiment, after obtaining the decentralized sensor network, testing the generated decentralized sensor network may specifically include the following steps:

[0153] Step 1: Obtain the working status data and network information data of the network under test. The network under test is the decentralized sensor network mentioned above. The working status data and network information data mentioned above are similar to the working status data and network information data mentioned above, and no further restrictions are imposed here.

[0154] Step 2: Obtain historical operating status data and historical network information data of the sensors in the network under test, and extract the corresponding first normal value range and second normal value range respectively;

[0155] Step 3: Compare the working status data to be tested with the historical working status data. If the working status data to be tested exceeds the first normal value range, mark the network to be tested corresponding to the working status data to be tested with a first mark. Compare the network information data to be tested with the historical network information data. If the network information data to be tested exceeds the second normal value range, mark the network to be tested corresponding to the network information data to be tested with a second mark.

[0156] Step 4: Perform network repair or reconfiguration on the network under test that has been marked with the first and second tags.

[0157] In one embodiment, the method for determining a decentralized sensor network provided by this application can also locate and eliminate faults in the decentralized sensor network, specifically including the following steps:

[0158] Step 1: Obtain the normal operating status data of sensor nodes in the decentralized sensor network, and aggregate the sensor nodes and their corresponding normal operating status data to generate a node-status database.

[0159] Step 2: Obtain the temporary working status data of sensor nodes in the decentralized sensor network, and match the temporary working status data with the data in the node-status database; if the match fails, the temporary working status data is determined as pending working status data.

[0160] Step 3: Perform a joint analysis on the data to be determined for operation. If the failure is not due to a hardware device, add the data to the node-status database. If the failure is due to a hardware device, locate and troubleshoot the sensor node corresponding to the data to be determined for operation.

[0161] In one embodiment, the method for determining a decentralized sensor network provided in this application determines the deployment location of newly joined sensor nodes, which may specifically include the following steps:

[0162] Step 1: Obtain the sensor network to be joined by the newly joined sensor node, and obtain the third working status data and third network data of the nodes in the same sensor network as the newly joined sensor node;

[0163] Step 2: Replace the same node with the newly joined sensor node, and obtain the working status data and network data of the newly joined sensor node at several candidate deployment locations;

[0164] Step 3: Compare the first difference between the working status data and the third working status data of the newly joined sensor node at several candidate deployment locations, and the second difference between the network data and the third network data. If the first difference of the newly joined sensor node at the first deployment location is the smallest and the second difference is small, then the first deployment location is taken as the target deployment location.

[0165] Based on the same inventive concept, embodiments of this application also provide an apparatus for determining a decentralized sensor network. (Specifically combined with...) Figure 4 The apparatus for determining a decentralized sensor network provided in the embodiments of this application will be described in detail.

[0166] Figure 4 This is a schematic diagram of the structure of a device for determining a decentralized sensor network provided in an embodiment of this application.

[0167] like Figure 4 As shown, the apparatus 400 for determining a decentralized sensor network may include: an acquisition module 410, a filtering module 420, a clustering module 430, and a selection module 440.

[0168] The acquisition module 410 is used to acquire the working status data and network information data of sensors in the Internet of Things;

[0169] The filtering module 420 is used to filter working status sample data and network information sample data from working status data and network information data according to preset filtering rules respectively.

[0170] Clustering module 430 is used to cluster working status sample data and network information sample data using clustering algorithms to obtain multiple sensor clusters;

[0171] Selection module 440 is used to select at least one sensor cluster that satisfies a preset association relationship from multiple sensor clusters to obtain a decentralized sensor network.

[0172] In one embodiment, the preset filtering rules include a first rule and a second rule. The first rule includes a time period, first filtering information, and a first sampling frequency. The second rule includes a classification rule and a second sampling frequency.

[0173] The aforementioned filtering module is specifically used to: filter the working status data according to the time period and the first filtering information to obtain the first working status data, and sample the first working status data according to the first sampling frequency to obtain the working status sample data;

[0174] The network information data is classified according to the classification rules to obtain classified network information data; and the classified network information data is sampled according to the second sampling frequency to obtain network information sample data.

[0175] In one embodiment, the filtering module mentioned above is specifically used to: sample the first working state data according to the first sampling frequency to obtain the second working state data;

[0176] Delete the target field's working status data and the working status data that does not meet the target logical requirements from the second working status data to obtain working status sample data.

[0177] In one embodiment, the first sampling frequency includes a random sampling frequency and a target fixed sampling frequency; based on this, the aforementioned filtering module is specifically used for:

[0178] The first working state data is sampled according to the random sampling frequency to obtain random working state data;

[0179] The first working state data is sampled according to the target fixed sampling frequency to obtain fixed sampling working state data;

[0180] Duplicate working state data are removed from both the random working state data and the fixed sample working state data to obtain the second working state data.

[0181] In one embodiment, the classification rule includes multiple classification value ranges;

[0182] Based on this, the aforementioned filtering module is specifically used for:

[0183] The network information data is classified according to multiple value ranges to obtain the network information data within each value range.

[0184] In one embodiment, the apparatus for determining a decentralized sensor network may further include a drawing module and a partitioning module.

[0185] Based on this, the drawing module is used to classify the network information data according to multiple classification value ranges before obtaining the classified network information data within each classification value range interval.

[0186] A frequency distribution curve was plotted based on network information data.

[0187] The segmentation module is used to divide the range of values ​​for classification based on the inflection point of the frequency distribution curve from low to high frequency.

[0188] In one embodiment, the clustering module mentioned above is specifically used for

[0189] The K-nearest centroid nearest neighbor classification algorithm was used to cluster the working status sample data and the network information sample data to obtain multiple sensor clusters.

[0190] In one embodiment, the apparatus for determining a decentralized sensor network may further include a creation module, a calculation module, a determination module, and a generation module.

[0191] The acquisition module is also used to acquire the working status information features and network information features of IoT sensors before selecting at least one sensor cluster that satisfies a preset association relationship from the sensor cluster clusters to obtain the decentralized sensor network.

[0192] The filtering module is also used to filter the first set of features in the working status information features and the second aggregated features in the network information features according to the aggregation rules.

[0193] A creation module is used to create a knowledge graph of working status information features and network information features based on the first aggregation feature and the second aggregation feature;

[0194] The calculation module is used to calculate the correlation value of each feature in the knowledge graph;

[0195] The determination module is used to identify target features whose correlation values ​​are greater than the correlation threshold.

[0196] The generation module is used to generate preset association relationships based on the relationships between target features.

[0197] In one embodiment, at least one sensor cluster satisfying a preset association relationship is selected from multiple sensor clusters to obtain a decentralized sensor network, including:

[0198] For each sensor cluster in multiple sensor clusters, obtain a list of functional categories for the sensor cluster. The list of functional categories includes functional description information and sensor data description information.

[0199] The first keyword is extracted based on the functional description information, and the second keyword is extracted based on the sensor data description information;

[0200] Calculate the similarity between the first keyword and the second keyword, and then select target keywords with similarity greater than the similarity threshold to establish association relationships;

[0201] Select at least one sensor cluster from multiple sensor clusters whose association relationship satisfies a preset association relationship.

[0202] In this embodiment, operational status data and network information data of sensors in the Internet of Things (IoT) can be acquired. Sample operational status data and sample network information data are then filtered from this data according to preset filtering rules. A clustering algorithm can then be used to cluster the sample operational status data and sample network information data, resulting in multiple sensor clusters. By selecting at least one sensor cluster that satisfies a preset association relationship from these clusters, a decentralized sensor network is obtained. This allows for the determination of the decentralized sensor cluster networking method, reducing the complexity of IoT management and enabling flexible and efficient implementation of IoT network functions.

[0203] The various modules in the apparatus for determining a decentralized sensor network provided in this application embodiment can achieve... Figures 1 to 3 The method steps of any of the embodiments shown herein, and the corresponding technical effects thereof, will not be described in detail here for the sake of brevity.

[0204] Figure 5 A schematic diagram of the hardware structure of the electronic device provided in an embodiment of this application is shown.

[0205] An electronic device may include a processor 501 and a memory 502 storing computer program instructions.

[0206] Specifically, the processor 501 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0207] Memory 502 may include mass storage for data or instructions. For example, and not limitingly, memory 502 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 502 may include removable or non-removable (or fixed) media. Where appropriate, memory 502 may be internal or external to the integrated gateway disaster recovery device. In a particular embodiment, memory 502 is non-volatile solid-state memory.

[0208] Memory may include read-only memory (ROM), random access memory (RAM), disk storage media devices, optical storage media devices, flash memory devices, and electrical, optical, or other physical / tangible memory storage devices. Therefore, typically, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the methods according to one aspect of this disclosure.

[0209] The processor 501 implements any of the methods for determining a decentralized sensor network in the above embodiments by reading and executing computer program instructions stored in the memory 502.

[0210] In one example, the electronic device may also include a communication interface 503 and a bus 510. Wherein, as... Figure 5 As shown, the processor 501, memory 502, and communication interface 503 are connected through bus 510 and complete communication with each other.

[0211] The communication interface 503 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0212] Bus 510 includes hardware, software, or both, that couples components of an online data traffic metering device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 510 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, any suitable bus or interconnect is contemplated herein.

[0213] Furthermore, in conjunction with the method for determining a decentralized sensor network in the above embodiments, this application embodiment can provide a computer storage medium for implementation. This computer storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement the method for determining a decentralized sensor network provided in this application embodiment.

[0214] This application also provides a computer program product in which instructions, when executed by a processor of an electronic device, cause the electronic device to perform the method for determining a decentralized sensor network as provided in this application.

[0215] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0216] The functional blocks shown in the above block diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0217] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0218] The aspects of this disclosure have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable means for determining decentralized sensor networks to produce a machine such that these instructions, executable via the processor of the computer or other programmable means for determining decentralized sensor networks, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, an application-specific processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowchart illustrations, and combinations of blocks in the block diagrams and / or flowchart illustrations, can also be implemented by special-purpose hardware performing the specified functions or actions, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0219] The above are merely specific embodiments of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A method for determining a decentralized sensor network, characterized in that, The method includes: Acquire operational status data and network information data from sensors in the Internet of Things (IoT); Select sample data of working status and sample data of network information from the working status data and the network information data respectively according to preset filtering rules; Clustering algorithms are used to cluster the working status sample data and the network information sample data to obtain multiple sensor clusters; At least one sensor cluster that satisfies a preset association relationship is selected from the plurality of sensor clusters to obtain a decentralized sensor network.

2. The method according to claim 1, characterized in that, The preset filtering rules include a first rule and a second rule. The first rule includes a time period, first filtering information, and a first sampling frequency. The second rule includes a classification rule and a second sampling frequency. The step of filtering working status sample data and network information sample data from the working status data and the network information data according to preset filtering rules includes: The working status data is filtered according to the time period and the first filtering information to obtain the first working status data, and the first working status data is sampled according to the first sampling frequency to obtain the working status sample data. The network information data is classified according to the classification rules to obtain classified network information data, and the classified network information data is sampled according to the second sampling frequency to obtain network information sample data.

3. The method according to claim 2, characterized in that, The step of sampling the first working state data according to the first sampling frequency to obtain the working state sample data includes: The first working state data is sampled according to the first sampling frequency to obtain the second working state data; Delete the target field's working status data and the working status data that does not meet the target logical requirements from the second working status data to obtain the working status sample data.

4. The method according to claim 3, characterized in that, The first sampling frequency includes a random sampling frequency and a target fixed sampling frequency; the step of sampling the first working state data according to the first sampling frequency to obtain the second working state data includes: The first working state data is sampled according to the random sampling frequency to obtain random working state data; The first working state data is sampled according to the target fixed sampling frequency to obtain fixed sampling working state data; The second working state data is obtained by deleting duplicate working state data from the random working state data and the fixed sampled working state data.

5. The method according to claim 2, characterized in that, The classification rules include multiple classification value ranges; classifying the network information data according to the classification rules to obtain classified network information data includes: The network information data is classified according to the multiple classification value ranges to obtain the classified network information data belonging to each of the classification value ranges.

6. The method according to claim 5, characterized in that, Before classifying the network information data according to the multiple classification value ranges to obtain the classified network information data belonging to each of the classification value ranges, the method further includes: A frequency distribution curve was plotted based on the network information data. The range of classification values ​​is defined by using the inflection point of the frequency distribution curve from low to high as the boundary.

7. The method according to claim 1, characterized in that, The clustering algorithm is used to cluster the working status sample data and the network information sample data to obtain multiple sensor clusters, including: The working status sample data and the network information sample data are clustered using the K-nearest centroid nearest neighbor classification algorithm to obtain multiple sensor clusters.

8. The method according to claim 1, characterized in that, Before selecting at least one sensor cluster that satisfies a preset association relationship from the plurality of sensor clusters to obtain a decentralized sensor network, the method further includes: Acquire the operational status and network connectivity information characteristics of IoT sensors; The first aggregated feature in the working status information features and the second aggregated feature in the network information features are filtered according to the aggregation rules respectively. Based on the first aggregation feature and the second aggregation feature, a knowledge graph of the working status information feature and the network information feature is created; Calculate the correlation value of each feature in the knowledge graph; Identify target features whose correlation value is greater than the correlation threshold; The preset association relationship is generated based on the relationship between the target features.

9. The method according to claim 8, characterized in that, The step of selecting at least one sensor cluster that satisfies a preset association relationship from multiple sensor clusters to obtain a decentralized sensor network includes: For each of the multiple sensor clusters, obtain a list of functional categories for the sensor cluster, the list of functional categories including functional description information and sensor data description information; The first keyword is extracted based on the functional description information, and the second keyword is extracted based on the sensor data description information; Calculate the similarity between the first keyword and the second keyword, and establish association relationships by filtering target keywords whose similarity is greater than a similarity threshold; Select at least one sensor cluster whose association relationship satisfies a preset association relationship from the plurality of sensor clusters.

10. An apparatus for determining a decentralized sensor network, characterized in that, The device includes: The acquisition module is used to acquire the working status data and network information data of sensors in the Internet of Things. The filtering module is used to filter working status sample data and network information sample data from the working status data and the network information data respectively according to preset filtering rules; The clustering module is used to cluster the working status data and the network information sample data using a clustering algorithm to obtain multiple sensor clusters; The selection module is used to select at least one sensor cluster that satisfies a preset association relationship from the sensor clusters to obtain a decentralized sensor network.

11. An electronic device, characterized in that, The device includes: a processor and a memory storing computer program instructions; the processor reads and executes the computer program instructions to implement the method for determining a decentralized sensor network as described in any one of claims 1-9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer program instructions that, when executed by a processor, implement the method for determining a decentralized sensor network as described in any one of claims 1-9.

Images