A method and apparatus for calibrating the location of a communication dumb resource

Abstract

The application discloses a communication dumb resource position calibration method and device, which is used to improve the position calibration effectiveness of the communication dumb resource. The scheme comprises the following steps: acquiring building distribution information in a target area and geographical position information of a communication dumb resource to be identified; inputting the building distribution information and the geographical position information of the communication dumb resource to be identified into a pre-trained identification model, and determining whether the communication dumb resource is a communication dumb resource to be calibrated according to an output result of the identification model, wherein the pre-trained identification model is used to identify whether the position of the communication dumb resource corresponding to the geographical position information is to be calibrated according to the building distribution information; acquiring network device relationship features and resource address data features of the communication dumb resource to be calibrated; and performing calibration on the geographical position information of the communication dumb resource to be calibrated in the target area according to the network device relationship features and the resource address data features, so as to obtain calibrated target geographical position information.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of communication, and in particular to a communication dumb resource position calibration method and device. BACKGROUND

[0002] With the development of broadband optical fiber technology and the increase of the scale of cable broadband users, users have higher dependence on the network, and there is a higher demand for the precision and refinement of network resource planning and construction of major operators. However, there are a large number of dumb resources in network transmission equipment, and the position information thereof is usually manually entered by network administrators of each city in the comprehensive resource management system of the operator. The accuracy of the basic device information and the data update frequency are difficult to guarantee. In the long run, there will be data error problems of part of the basic network devices, which greatly reduces the precision of network planning and construction and cannot timely find the resource uncovered area that needs to be constructed.

[0003] In addition, since the functional attributes of the communication dumb resource involve many aspects such as provided communication resources, covered ranges, and device functions, it is difficult to effectively calibrate the position of the communication dumb resource by only a single rule determination or manual adjustment.

[0004] How to improve the effectiveness of the position calibration of the communication dumb resource is a technical problem to be solved by the present application. SUMMARY

[0005] The purpose of the embodiments of the present application is to provide a communication dumb resource position calibration method and device to improve the effectiveness of the position calibration of the communication dumb resource.

[0006] In a first aspect, a communication dumb resource position calibration method is provided, comprising:

[0007] obtaining building distribution information in a target area and geographical position information of a to-be-identified communication dumb resource;

[0008] inputting the building distribution information and the geographical position information of the to-be-identified communication dumb resource into a pre-trained identification model, and determining whether the communication dumb resource is a to-be-calibrated communication dumb resource according to an output result of the identification model, wherein the pre-trained identification model is used to identify whether the position of the communication dumb resource corresponding to the geographical position information is to be calibrated according to the building distribution information;

[0009] obtaining network device relationship features and resource address data features of the to-be-calibrated communication dumb resource, wherein the network device relationship features represent the network connection relationship of the upper connection device corresponding to the to-be-calibrated communication dumb resource, and the resource address data features represent the address of the network resource corresponding to the to-be-calibrated communication dumb resource;

[0010] According to the network device relationship feature and the resource address data feature, the geographic location information of the to-be-calibrated communication dumb resource in the target area is calibrated to obtain calibrated target geographic location information.

[0011] In a second aspect, a communication dumb resource location calibration apparatus is provided, which comprises:

[0012] A first obtaining module is configured to obtain building distribution information in a target area and geographic location information of a to-be-identified communication dumb resource.

[0013] An identifying module is configured to input the building distribution information and the geographic location information of the to-be-identified communication dumb resource into a pre-trained identification model, and determine whether the communication dumb resource is a to-be-calibrated communication dumb resource according to an output result of the identification model, wherein the pre-trained identification model is configured to identify whether a location of the communication dumb resource corresponding to the geographic location information is to be calibrated according to the building distribution information.

[0014] A second obtaining module is configured to obtain network device relationship features and resource address data features of the to-be-calibrated communication dumb resource, wherein the network device relationship features represent network connection relationships of an upper-link device corresponding to the to-be-calibrated communication dumb resource, and the resource address data features represent addresses of network resources corresponding to the to-be-calibrated communication dumb resource.

[0015] A calibration module is configured to calibrate the geographic location information of the to-be-calibrated communication dumb resource in the target area according to the network device relationship features and the resource address data features, to obtain calibrated target geographic location information.

[0016] In a third aspect, an electronic device is provided, which comprises a processor, a memory, and a computer program stored in the memory and executable on the processor, and when the computer program is executed by the processor, the steps of the method according to the first aspect are implemented.

[0017] In a fourth aspect, a computer readable storage medium is provided, which stores a computer program, and when the computer program is executed by a processor, the steps of the method according to the first aspect are implemented.

[0018] In a fifth aspect, a computer program product is provided, which comprises a non-transitory computer readable storage medium storing a computer program, and the computer program is operable to cause a computer to perform some or all of the steps of the method according to the first aspect.

[0019] In the embodiment of the present application, the building distribution information in the target area and the geographic location information of the to-be-identified communication dumb resource are acquired; the building distribution information and the geographic location information of the to-be-identified communication dumb resource are input into a pre-trained identification model, whether the communication dumb resource is a to-be-calibrated communication dumb resource is determined according to the output result of the identification model; the network device relationship feature and the resource address data feature of the to-be-calibrated communication dumb resource are acquired; the geographic location information of the to-be-calibrated communication dumb resource in the target area is calibrated according to the network device relationship feature and the resource address data feature, and the calibrated target geographic location information is obtained, so as to identify the to-be-calibrated communication dumb resource through the pre-trained identification model, and then calibrate the geographic location information according to the network device relationship and the resource address data of the to-be-calibrated communication dumb resource, thereby improving the effectiveness of the calibrated target geographic location information. BRIEF DESCRIPTION OF DRAWINGS

[0020] The drawings described herein are used to provide further understanding of the present application, and form a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application, and do not constitute improper limitations on the present application. In the drawings:

[0021] Figure 1 is one of the flowcharts of a communication dumb resource location calibration method according to an embodiment of the present application;

[0022] Figure 2 is another flowchart of a communication dumb resource location calibration method according to an embodiment of the present application;

[0023] Figure 3a is a third flowchart of a communication dumb resource location calibration method according to an embodiment of the present application;

[0024] Figure 3b is a flowchart of judging the rationality of distribution of a to-be-identified optical fiber distribution box through an identification model according to an embodiment of the present application.

[0025] Figure 4 is a fourth flowchart of a communication dumb resource location calibration method according to an embodiment of the present application;

[0026] Figure 5a is a fifth flowchart of a communication dumb resource location calibration method according to an embodiment of the present application;

[0027] Figure 5b is a sixth flowchart of a communication dumb resource location calibration method according to an embodiment of the present application;

[0028] Figure 6 is a seventh flowchart of a communication dumb resource location calibration method according to an embodiment of the present application;

[0029] Figure 7 is a flowchart of a method for calibrating a location of a communication dumb resource according to an embodiment of the present application;

[0030] Figure 8 is a structural diagram of a device for calibrating a location of a communication dumb resource according to an embodiment of the present application.

[0031] Figure 9 is a logic architecture diagram of a device for calibrating a location of a communication dumb resource according to an embodiment of the present application. DETAILED DESCRIPTION

[0032] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of the present application. The figure numbers in the present application are only used to distinguish each step in the solutions, and are not used to limit the execution order of each step, which is subject to the description in the specification.

[0033] To solve the problems in the prior art, the embodiments of the present application provide a method for calibrating a location of a communication dumb resource, as shown in Figure 1 , which comprises the following steps.

[0034] S11: Obtain building distribution information in a target area and geographical position information of a communication dumb resource to be identified.

[0035] In this example, it is assumed that the communication dumb resource is an optical fiber distribution box for illustration. In actual applications, the communication dumb resource can also be other types of passive communication devices.

[0036] In actual applications, the above target area can be set according to actual needs. For example, a larger area is divided into multiple microgrids, and each microgrid is taken as a target area to perform location calibration of a communication dumb resource. The microgrid can also be referred to as a microgrid, which can be used for constructing a microgrid in a city according to a microgrid to divide the area into fragments, which can effectively improve the operation and management efficiency. Specifically, the backbone optical cable line can be constructed according to the integrated service access area as a unit, and the introduced optical cable line can be constructed according to the above microgrid as a unit. The optical fiber access service in the microgrid is converged to the distribution optical exchange through the introduced optical cable, and then connected to the backbone optical exchange through the distribution optical cable, and converged to the convergence point through the backbone optical cable.

[0037] The building distribution information can include a plurality of parameters such as the number, spacing, and density of the buildings in the target area, and is used to represent the distribution of the buildings in the target area. The building distribution information can be two-dimensional information based on latitude and longitude, or three-dimensional information based on latitude, longitude, and altitude.

[0038] The geographic location information of the to-be-identified communication dumb resource can include coordinate information of the location, such as the location coordinates of the to-be-identified optical distribution box in the target area. In addition, information of other associated communication dumb resources in the target area where the to-be-identified communication dumb resource is located can also be included, such as the distance, spacing standard deviation, and the like of the to-be-identified optical distribution box relative to other optical distribution boxes in the target area.

[0039] S12: input the building distribution information and the geographic location information of the to-be-identified communication dumb resource into a pre-trained identification model, and determine whether the communication dumb resource is a to-be-calibrated communication dumb resource according to the output result of the identification model, wherein the pre-trained identification model is used to identify whether the location of the communication dumb resource corresponding to the geographic location information needs to be calibrated according to the building distribution information.

[0040] In this step, the pre-trained identification model is used to determine whether the location of the to-be-identified communication dumb resource needs to be calibrated according to the building distribution information. In actual applications, the building distribution is closely related to the location of the communication dumb resource, and the location of the communication dumb resource determines the communication environment of the buildings in a certain area. This step efficiently identifies whether the location of the communication dumb resource needs to be calibrated through the pre-trained identification model, which can comprehensively consider the characteristics of communication and geographic location, and has the advantages of high efficiency and high accuracy.

[0041] S13: obtain network device relationship features and resource address data features of the to-be-calibrated communication dumb resource, wherein the network device relationship features represent the network connection relationship of the upper connection device corresponding to the to-be-calibrated communication dumb resource, and the resource address data features represent the address of the network resource corresponding to the to-be-calibrated communication dumb resource.

[0042] In this step, features are obtained from two aspects of network device relationship and resource address data for the to-be-calibrated communication dumb resource. The network device relationship represents the network connection relationship of the upper connection device of the to-be-calibrated communication dumb resource, and the resource address data can represent the actual address of the network resource. For example, the network device relationship features can include the node in the transmission network where the to-be-calibrated optical distribution box is located, which can include the upper secondary distribution point, the primary distribution point, and various types of transmission machine rooms. The resource address data features can include the installation address of the broadband user under the to-be-calibrated optical distribution box, the standard address of the resource point corresponding to the optical distribution box, and the like. In actual applications, the above features can be obtained based on the name identifier of the to-be-calibrated optical distribution box.

[0043] S14: Calibrate the geographic location information of the to-be-calibrated communication dumb resource in the target area according to the network device relationship feature and the resource address data feature, to obtain calibrated target geographic location information.

[0044] In this step, the geographic location information of the to-be-calibrated communication dumb resource is calibrated according to the network device relationship and the resource address data, which can comprehensively calibrate the geographic location of the communication dumb resource from the network relationship and the resource address.

[0045] Specifically, the network device relationship feature and the resource address data feature can be integrated to determine the target geographic location information of the to-be-calibrated communication dumb resource in the map address of the target area. One or more optional geographic locations can be determined based on the text of the map address using a text similarity analysis method, and the target geographic location can be further selected from the optional geographic locations based on the network relationship and the network resource by calculating the distance, to obtain the calibrated target geographic location information.

[0046] Through the scheme provided in the examples of the present application, the to-be-calibrated communication dumb resource is identified by the pre-trained identification model, and then the geographic location information calibration is performed on the two aspects of the network device relationship and the resource address data of the to-be-calibrated communication dumb resource, thereby improving the effectiveness of the calibrated target geographic location information. The present scheme can be widely applied to the maintenance and management of city-level communication dumb resources, improve the calibration effectiveness, and is conducive to improving the overall communication environment quality of the region.

[0047] Based on the scheme provided in the above embodiments, optionally, before the step S12, that is, before the building distribution information and the geographic location information of the to-be-identified communication dumb resource are input into the pre-trained identification model to obtain the to-be-calibrated communication dumb resource output by the identification model, as shown in the following step S11, the method further comprises: Figure 2

[0048] S21: Obtain a training sample and a corresponding training label, the training sample comprising geographic location information of a sample communication dumb resource and building distribution information in a sample area where the sample communication dumb resource is located, and the training label representing whether the location of the corresponding sample communication dumb resource needs to be calibrated.

[0049] In actual application, a plurality of micro grids can be first divided in a certain area, a training sample is generated for each micro grid, and a label corresponding to the sample is generated according to the distribution rationality of the optical distribution fiber box in the micro grid. For the sample with reasonable distribution, the label does not need to be calibrated, and for the sample with unreasonable distribution, the label needs to be calibrated. The distribution rationality of the optical distribution fiber box in the micro grid can be determined according to various ways such as fixed-point detection, user reporting, and inspection feedback. ​

[0050] S22: training the identification model using the training samples and corresponding training labels.

[0051] In this step, the identification model is trained using the training samples and corresponding training labels. To improve the training quality of the identification model, sample features can be constructed based on the optical distribution box location features, building distribution features, and the association features of the optical distribution box location and building distribution. As an example, the modeling features can be as shown in the following table:

[0052]

[0053]

[0054] Through the scheme provided in this example, the identification model is pre-trained based on the training samples and corresponding labels, which can enable the trained identification model to determine whether the communication dumb resource to be calibrated according to the geographic location information of the communication dumb resource to be identified and the building distribution information of the target area where the communication dumb resource is located, and has the advantages of high efficiency and accuracy.

[0055] According to the scheme provided in the above embodiments, optionally, the identification model is a stacked classification model, and the stacked classification model includes at least two layers of learner structures, the first layer of learner structures includes a plurality of first classifiers, and the second layer of learner structures includes a second classifier.

[0056] In this example, the identification model is a stacked classification model. The stacked classification model is a Stacking integrated model, which realizes the classification function through the Stacking integrated algorithm. The Stacking model can integrate the optimal parameters of multiple classification models, thereby improving the generalization ability and prediction accuracy of the model as a whole.

[0057] As shown in the above step S22, the training of the identification model using the training samples and corresponding training labels includes: Figure 3a

[0058] S31: training the plurality of first classifiers in the first layer of learner structures using the training samples and corresponding training labels.

[0059] In this example, the stacked classification model includes at least two layers of learners, and the first layer of learner structures can also be referred to as a learning layer. A plurality of training samples and corresponding training labels are used to train a plurality of first classifiers.

[0060] ​Optionally, for the obtained samples and corresponding labels, the full amount of samples is divided into three parts as a training set, a validation set and a test set respectively, and feature construction and feature cleaning are performed on the full amount of samples to optimize sample quality and improve model training effect. In this step, the training set is used to train the plurality of first classifiers, and the validation set is used to verify the plurality of first classifiers.

[0061] S32: Training the second classifier in the second layer learner structure using the prediction results of the training samples by the plurality of first classifiers.

[0062] The second layer learner is trained based on the training result of the first layer learner. Based on the plurality of first classifiers trained in the above step, the validation set result of the first layer plurality of first classifiers is merged into a new feature set as training data of the second classifier, and then the test set is used for verification.

[0063] S33: Parameter tuning of the classifiers in the stacked classification model is performed through cross-validation to obtain a pre-trained recognition model.

[0064] In this step, parameter tuning of each classifier is performed through cross-validation to optimize the classification effect and generalization ability of the model.

[0065] Through the scheme provided in the example of the present application, the stacked classification model can fuse the optimal parameters of multiple classification models, thereby improving the overall generalization ability and classification prediction accuracy of the trained model.

[0066] Based on the scheme provided in the above embodiment, the plurality of first classifiers in the first layer learner structure can include at least one of the following: a random forest classifier, a gradient boosting decision tree, and an extreme gradient ascent algorithm learner.

[0067] The second classifier in the second layer learner structure includes a logistic regression algorithm classifier.

[0068] In this example, the plurality of first classifiers in the first layer learner structure can use a random forest classifier Random Forest, a gradient boosting decision tree GBDT (Gradient Boosting Decision Tree), and an extreme gradient ascent XGBoost (eXtreme Gradient Boosting) algorithm, and the second classifier in the second layer learner structure can select a logistic regression Logistic algorithm. In addition, in the step of cross-validation, grid search GridSearchCV can be used to tune the model, and the best effect parameters of each layer are determined through cross-validation.

[0069] Through the scheme provided by the present example, the model output result is the probability score of the position of the to-be-identified optical distribution frame in the microgrid being correct. Further, based on a preset score threshold, it is determined whether the position of the to-be-identified optical distribution frame corresponding to the probability score needs to be calibrated. For example, an optical distribution frame with a score greater than the preset value is determined to not need to be calibrated, and an optical distribution frame with a score less than or equal to the preset value is determined to need to be calibrated.

[0070] Optionally, in a scenario in which a microgrid contains multiple to-be-identified optical distribution frames, the model can be used to determine whether the distribution of the multiple to-be-identified optical distribution frames in the microgrid is reasonable. The output result of the model represents a probability score of the position distribution of the multiple to-be-identified optical distribution frames in the microgrid being reasonable. If the score is greater than a preset value, it indicates that the positions of the multiple optical distribution frames in the microgrid do not need to be calibrated.

[0071] Figure 3b A flowchart of the process of determining whether the distribution of a to-be-identified optical distribution frame is reasonable by using a recognition model is shown. The training input is based on a stacking integrated algorithm stack classification model, and the training model is trained to determine whether the position of the optical distribution frame needs to be calibrated, using the training sample and the corresponding training label as the training data.

[0072] The microgrid features are constructed based on the training data, which can include basic features such as the identification and geographic position of the microgrid, the distribution features of the optical distribution frames in the microgrid, the distribution features of the buildings in the microgrid, and the derived association features associated with the optical distribution frames and the buildings in the microgrid.

[0073] The first layer of the stack classification model is trained by calculating the above-mentioned multiple microgrid features and performing data normalization processing, and the output result of the first layer is used to train the second layer of the stack classification model. The first layer of the stack classification model can include, for example, Random Forest, GBDT, and XGBoost, and the second layer of the stack classification model can include, for example, Logistic.

[0074] Based on the training results of the layers of the classification model, the optimal parameters are selected through cross-validation based on grid search parameter tuning, so as to optimize the generalization ability and classification prediction accuracy of the model as a whole. The trained model can output a microgrid distribution probability score. Further, the output can be in the form of an optical distribution frame list, which can include the name of the optical distribution frame, the microgrid to which the optical distribution frame belongs, and the original position of the optical distribution frame.

[0075] Based on the scheme provided by the above-mentioned embodiments, the network device relationship features and the resource address data features are used to calibrate the geographic position information of the to-be-calibrated communication dumb resource in the target area, as shown in Figure 4 In step S14, the geographic position information of the to-be-calibrated communication dumb resource in the target area is calibrated based on the network device relationship features and the resource address data features, to obtain calibrated target geographic position information.

[0076] S41: Obtain a plurality of configurable region information of the communication dumb resource in the target region.

[0077] In this example, the geographical position calibration is performed based on the characteristics of the network device relationship and the resource address data of the communication dumb resource to be calibrated. In this step, the configurable region information of the communication dumb resource can be collected in various ways. For example, based on the target region information, a search is performed to collect the standard map address data of the target region.

[0078] Optionally, the obtained map address is data cleaned, and the hierarchical region address information such as province, county, and city in the map address is extracted to construct a dictionary based on the address information.

[0079] S42: Perform word segmentation on the plurality of configurable region information and the resource address data feature corresponding to the to-be-calibrated position information respectively by using a preset word segmentation algorithm.

[0080] In this example, the preset word segmentation algorithm can be pre-set according to actual needs. For example, the Jieba word segmentation algorithm can be used to perform word segmentation. In this step, the configurable region information is segmented, and the to-be-calibrated position information is segmented.

[0081] The resource address data feature can be the name of the optical distribution box, the corresponding resource point standard address, the resource coverage address, and the installation address of the broadband user installed under the optical distribution box, and the like information obtained through the operator big data. In addition, the Internet crawler technology can be used to crawl the actual position point information and the provincial map address point data corresponding to the resource address library list to optimize the quality of the resource address data feature.

[0082] S43: Perform feature extraction on the word segmentation results corresponding to the plurality of configurable region information and the word segmentation result of the to-be-calibrated position information, to obtain a first word frequency vector corresponding to the plurality of configurable region information and a second word frequency vector of the to-be-calibrated position information.

[0083] In this step, feature extraction is performed based on the above word segmentation result. Optionally, the TF-IDF (term frequency-inverse document frequency) algorithm is used to perform text feature extraction, and the TfidfVectorizer parameter can be used to realize word frequency vectorization statistics of the word segmentation, so as to express each word segmentation feature by using a vector.

[0084] S44: Determine a target first word frequency vector with the highest similarity to the second word frequency vector from the plurality of first word frequency vectors.

[0085] The similarity of the above word frequency vectors is compared in this step. Optionally, the similarity between the address texts corresponding to the two vectors is expressed by the cosine similarity between the vectors. In this step, the second word frequency vector corresponding to the to-be-calibrated position information is determined from the plurality of first word frequency vectors corresponding to the plurality of configurable area information.

[0086] S45: generating calibrated target geographic position information according to the target configurable area information corresponding to the target first word frequency vector.

[0087] The target first word frequency vector corresponds to a new position most similar to the to-be-calibrated position information. In this step, the calibrated target geographic position is generated based on the target configurable area information corresponding to the target first word frequency vector, the microcell to which the target geographic position belongs is determined according to the target geographic position, and the target geographic position information containing the geographic position and the microcell is generated.

[0088] Based on the scheme provided in the above embodiment, optionally, as shown in Figure 5a In step S45, the calibrated target geographic position information is generated according to the target configurable area information corresponding to the target first word frequency vector, which includes:

[0089] S51: obtaining a plurality of resource addresses of communication dumb resources in a target configurable area corresponding to the target configurable area information and a plurality of map addresses in the target configurable area.

[0090] In this step, after the correct microcell of the optical distribution box is calibrated based on the foregoing steps, the optical distribution box resource short address and the map short address under the microcell are obtained. The optical distribution box resource short address can be the remaining address information after removing the administrative region information such as city, county, and street from various resource addresses of the optical distribution box. The map short address can be the remaining address information after removing the administrative region information such as city, county, and street from the map address.

[0091] S52: determining a target resource address from the plurality of resource addresses and a target map address from the plurality of map addresses by using an edit distance algorithm, wherein the similarity between the target resource address and the target map address is greater than or equal to the similarity between any resource address in the plurality of resource addresses and any map address in the plurality of map addresses.

[0092] In this step, the edit distance algorithm (Levenshtein) is used to calculate the text similarity of the optical distribution box resource short address and the map short address under the microcell. The position information corresponding to the map address with the highest similarity to the optical distribution box resource address under the microcell is selected as the new position information of the optical distribution box.

[0093] S53: generating calibrated target geographic location information according to the target configurable area information, the target resource address and the target map address.

[0094] In this step, the target geographic location information is generated based on the target configurable area information, the target resource address and the target map address. For example, Figure 5b The flowchart of the present scheme is shown.

[0095] Firstly, the resource address text of the optical distribution box is obtained according to the resource address data of the optical distribution box, and the configurable area information is obtained, which specifically includes the provincial map address text to which the optical distribution box belongs.

[0096] The above text is preprocessed to clean up dirty characters to improve the quality of the text.

[0097] Subsequently, the above text is segmented by using the Jieba segmentation based on the text feature self-defined dictionary, and the above dictionary includes the address cleaned and segmented based on the above step.

[0098] Further, the word frequency statistics vectorization is performed on the segmentation result, and the text feature extraction is performed based on the TfidfVectorizer parameter to express the segmentation features in the form of a vector.

[0099] Subsequently, the similarity between the segmentation feature vector of text 1 corresponding to the resource address text of the optical distribution box and the segmentation feature vector of text 2 corresponding to the provincial map address text is calculated. Wherein, the cosine similarity parameter is used to calculate the vector cosine similarity.

[0100] Next, the map address with the highest similarity to the resource address of the optical distribution box is selected.

[0101] Then, the original position of the optical distribution box is attributed to the microgrid 1, the most similar map address position point recognized in the above step is attributed to the microgrid 2, and the resource address crawler position of the optical distribution box is attributed to the microgrid 3. Wherein, the microgrid 2 to which the original position of the full-amount optical distribution box belongs can be determined by using the position function relationship in the spatial database PostGIS. In addition, the uplink distance of the optical distribution box is obtained, which specifically includes the distance between the optical distribution box and the uplink device.

[0102] Based on the above identification of the micro-grid and the distance to the upper link, it is determined whether the micro-grid 1, the micro-grid 2 and the micro-grid 3 are the same, and the distance to the upper link is less than 2 kilometers. If yes, it is determined that the original position of the optical distribution fiber box belongs to the micro-grid correctly. If no, it is indicated that the optical distribution fiber box does not belong to the micro-grid 1, and it is continued to determine whether the address of the crawler is greater than or equal to 80% and the accuracy is 1. If yes, it is indicated that the address of the optical distribution fiber box resource crawler belongs to the micro-grid with high credibility, and it is determined that the optical distribution fiber box belongs to the micro-grid 3 of the crawler position. If no, it is indicated that the optical distribution fiber box does not belong to the micro-grid 1 or the micro-grid 3, and it is continued to determine whether the similarity between the map address and the resource address is greater than 70%. If yes, it is indicated that the most similar map address identified above is the address corresponding to the micro-grid to which the optical distribution fiber box belongs, that is, the optical distribution fiber box belongs to the micro-grid 2.

[0103] It should be noted that the discrimination criteria in the present example are only examples, and various discrimination criteria such as the distance to the upper link being less than 2 kilometers, the credibility being greater than or equal to 80%, the accuracy being 1, and the similarity being greater than 70% can be flexibly set according to actual needs.

[0104] Based on the scheme provided in the above embodiment, optionally, as shown in Figure 6 In the step S13, the network device relationship feature and the resource address data feature of the to-be-calibrated communication dumb resource are acquired, including:

[0105] S61: acquiring passive optical network link data of an upper link device of the to-be-calibrated communication dumb resource and at least one of resource address information: resource point standard address, resource coverage address, and downstream broadband installation address of the to-be-calibrated communication dumb resource.

[0106] In the present example, the upper link device feature of the optical distribution fiber box can be acquired from the transmission network data of the resource network system of the operator big data. The optical distribution fiber box can be connected to a secondary distribution point, a primary distribution point and various transmission machine rooms in the actual optical fiber transmission link. The above resource address information can be acquired by a crawler technology, including resource address crawler position data corresponding to the optical distribution fiber box.

[0107] S62: determining the network device relationship feature according to the passive optical network link data of the upper link device, and determining the resource address data feature according to the resource address information.

[0108] The scheme abstracts various device points (optical distribution fiber boxes, secondary distribution points, primary distribution points, and various machine rooms) into points according to the passive optical network (PON) link data of the transmission equipment, abstracts the local optical fibers in the transmission link into edges, and the edges represent the relationship between the points (the local optical fibers are connected at both ends of various transmission equipment). Finally, the optical distribution fiber box all uplink equipment information is obtained by traversing and querying the graph database technology, and the distance between the optical distribution fiber box and the uplink equipment is calculated as the main feature of the optical distribution fiber box uplink equipment.

[0109] For the resource address crawler position data corresponding to the distribution box, the scheme can perform cleaning and normalization on the data, and express it in a matching manner with the network equipment relationship features described above, so as to perform the microgrid determination to which the optical distribution fiber box belongs.

[0110] The scheme provided by the embodiment of the present application can be used to calibrate the optical distribution fiber box and other network dumb resources. Taking the optical distribution fiber box as an example, the scheme is further described in combination with the flowchart shown in Figure 7

[0111] First, the basic information of the optical distribution fiber box is obtained, including but not limited to the name of the optical distribution fiber box, the original longitude and latitude of the optical distribution fiber box, the belonging microgrid, the belonging resource point, the standard address of the resource point, and the resource coverage address. In addition, the building boundary information of the whole province map to which the optical distribution fiber box belongs, the whole province map address information, and the optical distribution fiber box resource address crawler information are obtained, including but not limited to the building boundary, the building belonging microgrid, the map address name, the map address longitude and latitude, the map address belonging microgrid, and the optical distribution fiber box resource address crawler longitude and latitude.

[0112] Then, the original position distribution characteristics of the optical distribution fiber box in the microgrid and the building distribution characteristics are used to determine whether the original position of the optical distribution fiber box in the microgrid is distributed reasonably by using a pre-trained classification and recognition model.

[0113] Next, for the to-be-calibrated optical distribution fiber box with unreasonable distribution, the optical distribution fiber box is calibrated to the correct microgrid and the correct position by using the map address data characteristics and the optical distribution fiber box resource address data characteristics, in combination with a text similarity algorithm and business rules.

[0114] Finally, the optical distribution fiber box calibration list is output, and the list content includes, for example, the city, county, longitude and latitude, and the belonging microgrid of the optical distribution fiber box before calibration, the new longitude and latitude, the new microgrid, whether the new and old microgrids are consistent, and the distance before and after calibration of the optical distribution fiber box.

[0115] ​The scheme fully utilizes the advantages of big data, fuses multiple means such as machine learning algorithm, graph database technology, NLP (Natural Language Processing) natural language processing technology, image recognition, and Internet crawler technology, and realizes optical distribution fiber box position calibration. First, the basic information of the optical distribution fiber box equipment and the attribution microgrid information of the operator management system are obtained, the position distribution of the optical distribution fiber box in the microgrid is comprehensively judged by using a machine learning classification model in combination with the building data of the whole province, and the optical distribution fiber box corresponding to the resource address data and the optical distribution fiber box uplink equipment information is further obtained, the optical distribution fiber box calibration position list is output by using the NLP text similarity algorithm and superimposing business rules in combination with the provincial map address data, and the optical distribution fiber box calibration list is actively pushed to the city technical personnel, so that the city technical personnel can timely maintain and manage the dumb resources and ensure the high accuracy of subsequent network planning.

[0116] The scheme provided by the present application example has data feature configurability. In the actual application process, different cities can configure network resource equipment position features and building features according to actual network equipment calibration requirements, and obtain a network resource equipment list with correct original positions.

[0117] The scheme provided by the present application example has business development difference. In the actual application process, different optical distribution fiber box calibration lists can be output according to the actual network resource business development status of each city.

[0118] The scheme can effectively improve the optical distribution fiber box position calibration efficiency, so that the city front-line personnel only need to locate the optical distribution fiber box position information according to the calibration list, reduce a large amount of repeated field investigation work of the front-line personnel, and improve the optical distribution fiber box position calibration work efficiency.

[0119] In addition, the scheme can also improve the optical distribution fiber box position calibration accuracy. By combining multi-dimensional data fusion such as actual network transmission characteristics of the optical distribution fiber box, map address characteristics, and building characteristics, and using machine learning models, text processing technology, and business rules for multi-technology superposition optimization, the optical distribution fiber box position calibration list is automatically output, manual intervention is reduced, and the optical distribution fiber box position calibration accuracy is improved.

[0120] The present application proposal has strong universality and generalizability. The models and algorithm models used and constructed in the proposal are developed based on operator big data, have high convenience for promotion and deployment, and can be widely applied to optical distribution fiber box calibration planning scenes and related work in various provinces and cities to improve planning efficiency and realize cost reduction and efficiency improvement.

[0121] In order to solve the problems in the prior art, an embodiment of the present application provides a communication dumb resource position calibration device 80, as shown in Figure 8 , which comprises:

[0122] The first obtaining module 81 obtains building distribution information in a target area and geographical position information of a communication dumb resource to be identified.

[0123] The identification module 82 inputs the building distribution information and the geographical position information of the communication dumb resource to be identified into a pre-trained identification model, and determines whether the communication dumb resource is a communication dumb resource to be calibrated according to an output result of the identification model, wherein the pre-trained identification model is used to identify whether a position of a communication dumb resource corresponding to the geographical position information is to be calibrated according to the building distribution information.

[0124] The second obtaining module 83 obtains network device relationship features and resource address data features of the communication dumb resource to be calibrated, wherein the network device relationship features represent a network connection relationship of an upper-link device corresponding to the communication dumb resource to be calibrated, and the resource address data features represent an address of a network resource corresponding to the communication dumb resource to be calibrated.

[0125] The calibration module 84 performs calibration on the geographical position information of the communication dumb resource to be calibrated in the target area according to the network device relationship features and the resource address data features, and obtains calibrated target geographical position information.

[0126] The device provided by the embodiment of the present application obtains building distribution information in a target area and geographical position information of a communication dumb resource to be identified, inputs the building distribution information and the geographical position information of the communication dumb resource to be identified into a pre-trained identification model, determines whether the communication dumb resource is a communication dumb resource to be calibrated according to an output result of the identification model, obtains network device relationship features and resource address data features of the communication dumb resource to be calibrated, and performs calibration on the geographical position information of the communication dumb resource to be calibrated in the target area according to the network device relationship features and the resource address data features, and obtains calibrated target geographical position information, so as to identify the communication dumb resource to be calibrated through the pre-trained identification model, and then perform geographical position information calibration on the network device relationship and the resource address data of the communication dumb resource to be calibrated, thereby improving the effectiveness of the calibrated target geographical position information.

[0127] The device provided by the embodiment of the present application can include a plurality of functional modules, for example, a micro-lab indoor optical distribution frame box position feature module, a micro-lab indoor building feature module, an optical distribution frame box resource address feature module, an optical distribution frame box position discrimination module, and an optical distribution frame box position calibration module. Figure 9 A logical architecture diagram based on the device provided by the embodiment of the present application is shown.

[0128] The micro-grid optical distribution frame position feature module can be used to obtain the optical distribution frame position distribution features in the micro-grid, obtain the building distribution features in the micro-grid by using the micro-grid building position feature module, obtain the optical distribution frame corresponding address features by using the optical distribution frame resource address feature module, determine whether the original position of the optical distribution frame is correct by using the optical distribution frame position determination module, give the position information of the optical distribution frame after calibration when the original position is incorrect by using the optical distribution frame position calibration module, and finally give the final position list of the optical distribution frame according to the optical distribution frame position determination result and the position calibration result.

[0129] Specifically, the micro-grid optical distribution frame position feature module can be used to obtain the optical distribution frame original position data manually maintained by the resource management system and the micro-grid basic information according to the operator big data, determine the micro-grid to which the optical distribution frame belongs by using the position function relationship in the spatial database PostGIS, and the main features include the number of optical distribution frames in the micro-grid, the distance between the optical distribution frames in the micro-grid, the number of optical distribution frames in different threshold distances, etc.

[0130] The micro-grid building position feature module can be used to obtain the building boundary data and the provincial map address data of Zhejiang Province according to the image recognition, internet crawler and other technologies of the operator big data, and determine the micro-grid to which the building belongs by using the spatial database technology, and the main features include the number of buildings in the micro-grid, the distance between the buildings in the micro-grid, the number of buildings in different threshold values, the distribution of optical distribution frames around the buildings, etc.

[0131] The optical distribution frame position determination module can be used to combine the micro-grid optical distribution frame position features and the micro-grid building position features, construct a binary classification machine learning model, and determine whether the optical distribution frame position distribution in the micro-grid of the whole province is reasonable.

[0132] The modules in the device provided in the embodiments of the present application can also implement the method steps provided in the method embodiments. Alternatively, the device provided in the embodiments of the present application can also include other modules in addition to the above modules to implement the method steps provided in the method embodiments. The device provided in the embodiments of the present application can achieve the technical effects achieved by the method embodiments.

[0133] Preferably, the embodiments of the present application also provide an electronic device, which includes a processor, a memory, a computer program stored in the memory and executable on the processor, and the computer program is executed by the processor to implement the processes of each of the above communication dumb resource position calibration method embodiments and achieve the same technical effects. To avoid repetition, details are not described here.

[0134] The embodiment of the present application further provides a computer readable storage medium, and the computer readable storage medium stores a computer program. The computer program is executed by a processor to implement each process of the communication dumb resource position calibration method embodiment and achieve the same technical effects. To avoid repetition, details are not described herein. The computer readable storage medium includes a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0135] The embodiment of the present application further provides a computer program product, which includes a non-transitory computer readable storage medium storing a computer program. The computer program is operable to cause a computer to perform some or all of the steps of the communication dumb resource position calibration method embodiment and achieve the same technical effects. To avoid repetition, details are not described herein.

[0136] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system or a computer program product. Therefore, the present application can be in the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. In addition, the present application can be in the form of a computer program product implemented on one or more computer usable storage media (including but not limited to a magnetic disk storage, a CD-ROM, an optical storage, etc.) containing computer usable program code.

[0137] The present application is described with reference to the flowcharts and / or block diagrams of the method, device (system) and computer program product according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks can be implemented by computer program instructions. These computer program instructions can be provided to a general purpose computer, a special purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the computer or other programmable data processing device produce the functions described in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks Figure 1 The apparatus that performs the functions specified in one or more flows and / or blocks.

[0138] These computer program instructions can also be stored in a computer readable storage medium that can direct the computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable storage medium produce the manufactured product including the instruction apparatus that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks Figure 1 The apparatus that performs the functions specified in one or more flows and / or blocks.

[0139] These computer program instructions can also be loaded into computer or other programmable data processing devices to cause a series of operational steps to be performed on the computer or other programmable devices to produce a computer implemented process such that the instructions which execute on the computer or other programmable devices provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 Figure 1

[0140] In one typical configuration, the computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0141] The memory can include non-persistent memory and / or volatile memory, such as random access memory (RAM) about which the computer stores information about an operating system, application software, and / or the like. Memory is an example of computer readable media.

[0142] Computer readable media includes permanent and non-permanent, removable and non-removable media implemented in any method or technology for storage of information such as computer readable 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 technology, compact disc read only memory (CD-ROM), digital versatile disc (DVD), or other optical storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to computing devices. According to the definition herein, computer readable media does not include transitory media, such as modulated data signals and carrier waves.

[0143] It should also be noted that the terms "comprising", "comprises", "including", "includes" or any other variation thereof are intended to cover a non-exclusive inclusion, such that a process, method, article or apparatus that comprises a list of elements does not include only those elements recited, but can also include other elements not expressly listed or inherent to such process, method, article or apparatus. Without limitation, an element preceded by "comprises a" does not, without more constraints, foreclose the existence of additional identical elements in the process, method, article or apparatus that comprises the recited element.

[0144] ​​Those skilled in the art will appreciate that embodiments of the present application can be devised for a method, a system, or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer-readable program code thereon for use by or in connection with an instruction execution system. For the purposes of this description, a computer-usable or computer readable storage medium can be any apparatus that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.

[0145] The foregoing is merely illustrative of the principles of the application and various modifications can be made by persons skilled in the art. The present application is not intended to be limited to the embodiments shown, but is to be accorded the full scope that resides in the art thereof. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for calibrating the location of communication dummy resources, characterized in that, include: Obtain information on the distribution of buildings within the target area and the geographic location information of the communication dummy resources to be identified; The building distribution information and the geographic location information of the communication dummy resource to be identified are input into a pre-trained identification model. The output of the identification model is used to determine whether the communication dummy resource is a communication dummy resource to be calibrated. The pre-trained identification model is used to identify whether the location of the communication dummy resource corresponding to the geographic location information needs to be calibrated based on the building distribution information. Obtain the network device relationship characteristics and resource address data characteristics of the communication dummy resource to be calibrated, wherein the network device relationship characteristics represent the network connection relationship of the uplink device of the corresponding communication dummy resource to be calibrated, and the resource address data characteristics represent the address of the network resource of the corresponding communication dummy resource to be calibrated; Based on the network device relationship characteristics and resource address data characteristics, the geographical location information of the communication dummy resource to be calibrated in the target area is calibrated to obtain the calibrated target geographical location information.

2. The method as described in claim 1, characterized in that, Before inputting the building distribution information and the geographic location information of the communication dummy resource to be identified into the pre-trained recognition model, the method further includes: Acquire training samples and corresponding training labels. The training samples include the geographical location information of the sample communication dummy resources and the building distribution information within the sample area where the sample communication dummy resources are located. The training labels indicate whether the location of the corresponding sample communication dummy resources needs to be calibrated. The recognition model is trained using the training samples and corresponding training labels.

3. The method as described in claim 2, characterized in that, The recognition model is a stacked classification model, which includes at least two learner structures. The first learner structure includes multiple first classifiers, and the second learner structure includes a second classifier. The process of training a recognition model using the training samples and corresponding training labels includes: The training samples and corresponding training labels are used to train multiple first classifiers in the first layer learner structure; The second classifier in the second-layer learner structure is trained using the prediction results of the multiple first classifiers on the training samples. By performing parameter tuning on the classifier in the stacked classification model through cross-validation, a pre-trained recognition model is obtained.

4. The method as described in claim 3, characterized in that, The first classifiers in the first layer learner structure include at least one of the following: random forest classifier, gradient boosting decision tree, extreme gradient ascent algorithm learner; The second classifier in the second-layer learner structure includes a logistic regression algorithm classifier.

5. The method as described in claim 1, characterized in that, Based on the network device relationship characteristics and resource address data characteristics, the geographical location information of the communication dummy resource to be calibrated within the target area is calibrated to obtain calibrated target geographical location information, including: Obtain information on multiple configurable regions of communication dummy resources within the target area; The multiple configurable region information and the location information to be calibrated corresponding to the resource address data features are segmented using a preset word segmentation algorithm. Feature extraction is performed on the word segmentation results corresponding to the multiple configurable region information and the word segmentation results of the position information to be calibrated, respectively, to obtain the first word frequency vector corresponding to the multiple configurable region information and the second word frequency vector of the position information to be calibrated; From multiple first word frequency vectors, determine the target first word frequency vector that has the highest similarity to the second word frequency vector; Based on the target configurable region information corresponding to the first word frequency vector of the target, calibrated target geographical location information is generated.

6. The method as described in claim 5, characterized in that, Based on the target configurable region information corresponding to the first word frequency vector of the target, calibrated target geographic location information is generated, including: Obtain multiple resource addresses of communication dummy resources within the target configurable area corresponding to the target configurable area information, and multiple map addresses within the target configurable area; A target resource address is determined from the plurality of resource addresses using an edit distance algorithm, and a target map address is determined from the plurality of map addresses, wherein the similarity between the target resource address and the target map address is greater than or equal to the similarity between any one of the plurality of resource addresses and any one of the plurality of map addresses; Based on the target configurable area information, the target resource address, and the target map address, calibrated target geographic location information is generated.

7. The method according to any one of claims 1 to 6, characterized in that, Obtaining the network device relationship characteristics and resource address data characteristics of the communication dummy resource to be calibrated includes: Obtain the passive optical network link data of the uplink device of the communication dummy resource to be calibrated and at least one of the following resource address information: resource point standard address, resource coverage address, and downstream broadband installation address of the communication dummy resource to be calibrated; The network device relationship characteristics are determined based on the passive optical network link data of the uplink device, and the resource address data characteristics are determined based on the resource address information.

8. A communication dummy resource location calibration device, characterized in that, include: The first acquisition module acquires information on the distribution of buildings within the target area and the geographical location information of the communication dummy resources to be identified. The identification module inputs the building distribution information and the geographical location information of the communication dummy resource to be identified into a pre-trained identification model, and determines whether the communication dummy resource is a communication dummy resource to be calibrated based on the output of the identification model. The pre-trained identification model is used to identify whether the location of the communication dummy resource corresponding to the geographical location information needs to be calibrated based on the building distribution information. The second acquisition module acquires the network device relationship characteristics and resource address data characteristics of the communication dummy resource to be calibrated, wherein the network device relationship characteristics represent the network connection relationship of the uplink device of the corresponding communication dummy resource to be calibrated, and the resource address data characteristics represent the address of the network resource of the corresponding communication dummy resource to be calibrated. The calibration module performs calibration on the geographical location information of the communication dummy resource to be calibrated within the target area based on the network device relationship characteristics and resource address data characteristics, thereby obtaining the calibrated target geographical location information.

9. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method as described in any one of claims 1 to 7.

11. A computer program product, characterized in that, The computer program product includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps of the method as described in any one of claims 1 to 7.

Images