Risk source identification model construction method and communication base station environment risk source identification method
By constructing risk source classification rules and optimizing deep learning algorithms, the problems of accuracy and efficiency in risk source identification in communication base station environments have been solved, realizing automated and systematic risk source identification, reducing costs and supporting one-stop security briefings.
Patent Information
- Application Number
- CN202411040959.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-31
- Publication Date
- 2024-11-19
AI Technical Summary
Existing technologies rely on manual processing, resulting in poor accuracy, low efficiency, and high cost in identifying environmental risk sources at communication base stations, and lack systematic intelligent identification methods.
Risk source classification rules are constructed, and deep learning target detection algorithms such as YOLOv5 are used for training and optimization. Combined with the CBS module and the improved IMIOU loss function, a risk source identification model is established to automatically identify risk sources in the communication base station environment.
It improves the accuracy and efficiency of risk source identification, reduces human and material costs, achieves more objective and comprehensive risk source identification, and supports one-stop safety briefing.
Smart Images

Figure CN118982108A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of artificial intelligence and communication technology, and in particular to a risk source identification model construction method, a communication base station environment risk source identification method, system, device, computer-readable storage medium, computer program product and terminal equipment. Background Art
[0002] The security risk assessment of communication base stations, computer room interiors, and surrounding environment information is not only the most important step before the construction of communication base stations, but also the key assessment basis for ensuring the safe and stable operation and management of communication base stations after the construction of communication base stations. Base station construction is mainly divided into two categories: new construction at a new site and new construction at a shared site. Among them, new construction at a shared site accounts for a higher proportion. However, whether it is new construction at a shared site or new site, the investigation, identification, and assessment and analysis of base station environmental risk hazards are crucial steps in the base station construction process.
[0003] At present, the investigation, identification and evaluation of base station environmental risk hazards mainly rely on the personal experience of on-site survey and design personnel. These survey and design personnel first survey various on-site environments, equipment installations, computer rooms, etc., and then record the potential risks and hidden dangers in the corresponding base station environment based on their own experience. However, this approach requires a lot of manual operation and manual analysis and judgment. Relying on manual processing often leads to omissions, resulting in poor accuracy. In addition, due to differences in the skill level, execution level and judgment of potential risks and hidden dangers of different survey and design personnel, the assessment of base station environmental risk sources cannot reflect objectivity and comprehensiveness. Even through multiple on-site surveys and reviews, it is difficult to solve these problems. It will also waste a lot of manpower and material resources to repeatedly verify and confirm the safety risk sources, resulting in low efficiency and poor real-time performance. Summary of the invention
[0004] The purpose of the embodiments of the present invention is to provide a risk source identification model construction method, a communication base station environmental risk source identification method, system, device, computer-readable storage medium, computer program product and terminal equipment, which can solve the problems existing in the prior art that relies on manual processing, effectively improve the accuracy of communication base station environmental risk source identification, have more objective and comprehensive characteristics, and can significantly reduce manpower and material costs, and improve identification efficiency and real-time performance.
[0005] In order to achieve the above-mentioned object, a first aspect of the present invention provides a method for constructing a risk source identification model, comprising:
[0006] According to different types of risk sources in the communication base station environment, a risk source classification rule is constructed; wherein the risk source classification rule is used to represent the corresponding relationship between risk information of the risk source, and the risk information includes risk category, risk link, risk point name, risk type and risk disposal measures;
[0007] According to the risk source classification rules, original data with risk source characteristics under different classification rules are obtained, and a training data set is constructed according to the original data;
[0008] Training and optimizing a preset target detection algorithm according to the training data set to obtain an optimized target detection algorithm; wherein the optimization includes network structure optimization and loss function optimization of the target detection algorithm;
[0009] A risk source identification model is constructed according to the risk source classification rule and the optimized target detection algorithm to identify risk sources in a communication base station environment according to the risk source identification model.
[0010] Furthermore, constructing a training data set based on the original data specifically includes:
[0011] Preprocessing the original data to obtain standard data under different classification rules;
[0012] Performing format conversion processing on the standard data to obtain preset format data under different classification rules;
[0013] According to the risk source targets involved in the risk source classification rules, feature annotation is performed on the preset format data, and a training data set is constructed based on all the annotated data.
[0014] Furthermore, the target detection algorithm is YOLOV5, and the network structure optimization includes: designing a CBS module in the Neck part of the original network structure of YOLOV5, and constructing an auxiliary branch based on the original network structure of YOLOV5, and the auxiliary branch adopts a ResNet structure to perform communication base station environment scene recognition.
[0015] Furthermore, the loss function optimization includes: replacing the original CIOU loss function of YOLOV5 with the IMIOU loss function including overlap loss, center distance loss and width-height loss, and integrating the Focal Loss loss function into the IMIOU loss function.
[0016] Furthermore, constructing a risk source identification model according to the risk source classification rule and the optimized target detection algorithm specifically includes:
[0017] Detect the communication base station environment scenes under different classification rules according to the optimized target detection algorithm, and remove duplicates from all detected risk source targets to obtain a risk source target set;
[0018] Uniquely combining the risk source targets in the risk source target set according to the risk source classification rule, and constructing a mapping relationship between each group of risk source target combinations and the risk information involved in the risk source classification rule; wherein each group of risk source target combinations includes at least one risk source target;
[0019] All risk source target combinations after the mapping relationship is constructed are deduplicated according to the mapping relationship, and a risk source identification model is constructed according to the mapping relationship between the risk source target combinations after deduplication and the risk information involved in the risk source classification rules.
[0020] In order to achieve the above-mentioned object, a second aspect of the present invention provides a method for identifying a communication base station environment risk source, comprising:
[0021] Acquire data to be identified of a communication base station environment to be identified;
[0022] The data to be identified is identified according to a preset risk source identification model, a risk source identification result in the communication base station environment to be identified is obtained, and the risk source identification result is fed back to the client for display; wherein, the risk source identification model is obtained by adopting the risk source identification model construction method described in any one of the first aspects above.
[0023] In order to achieve the above-mentioned purpose, the third aspect of the present invention also provides a communication base station environment risk source identification system, including a client and a server; wherein,
[0024] The client is used to collect the data to be identified of the communication base station environment to be identified, and send the data to be identified to the server;
[0025] The server is used to identify the data to be identified according to a preset risk source identification model, obtain the risk source identification result in the communication base station environment to be identified, and feed back the risk source identification result to the client; wherein the risk source identification model is obtained by adopting the risk source identification model construction method described in any one of the first aspects above;
[0026] The client is also used to display the risk source identification result.
[0027] In order to achieve the above-mentioned purpose, the fourth aspect of the present invention also provides a risk source identification model construction device, including:
[0028] A risk source classification rule construction module is used to construct risk source classification rules according to different types of risk sources in the communication base station environment; wherein the risk source classification rules are used to represent the corresponding relationship between risk information of risk sources, and the risk information includes risk category, risk link, risk point name, risk type and risk disposal measures;
[0029] A training data set construction module is used to obtain original data with risk source characteristics under different classification rules according to the risk source classification rules, and construct a training data set based on the original data;
[0030] A target detection algorithm training and optimization module, used to train and optimize a preset target detection algorithm according to the training data set to obtain an optimized target detection algorithm; wherein the optimization includes network structure optimization and loss function optimization of the target detection algorithm;
[0031] The risk source identification model construction module is used to construct a risk source identification model according to the risk source classification rules and the optimized target detection algorithm, so as to identify the risk sources in the communication base station environment according to the risk source identification model.
[0032] In order to achieve the above-mentioned purpose, the fifth aspect of the present invention further provides a communication base station environment risk source identification device, including:
[0033] A module for acquiring data to be identified, used to acquire data to be identified of a communication base station environment to be identified;
[0034] A risk source identification module is used to identify the data to be identified according to a preset risk source identification model, obtain the risk source identification result in the communication base station environment to be identified, and feed back the risk source identification result to the client for display; wherein, the risk source identification model is obtained by adopting the risk source identification model construction method described in any one of the first aspects above.
[0035] The sixth aspect embodiment of the present invention also provides a computer-readable storage medium, including a stored computer program, which, when running, controls the device where the computer-readable storage medium is located to execute the risk source identification model construction method described in any one of the first aspect above, or to execute the communication base station environment risk source identification method described in any one of the second aspect above.
[0036] The seventh aspect embodiment of the present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the risk source identification model construction method described in any one of the first aspect above, or implements the communication base station environment risk source identification method described in any one of the second aspect above.
[0037] The eighth aspect embodiment of the present invention also provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the processor implements the risk source identification model construction method described in any one of the first aspect above, or implements the communication base station environment risk source identification method described in any one of the second aspect above.
[0038] Compared with the prior art, the embodiment of the present invention provides a risk source identification model construction method, a communication base station environment risk source identification method, a system, a device, a computer-readable storage medium, a computer program product and a terminal device. First, according to different types of risk sources in the communication base station environment, a risk source classification rule is constructed, wherein the risk source classification rule is used to represent the corresponding relationship between the risk information of the risk source, and the risk information includes risk category, risk link, risk point name, risk type and risk disposal measures; then, according to the risk source classification rule, the original data with risk source characteristics under different classification rules is obtained, and a training data set is constructed according to the original data; then, according to the training data set, the preset target detection algorithm is trained and optimized to obtain an optimized target detection algorithm, wherein the optimization includes network structure optimization and loss function optimization of the target detection algorithm; finally, according to the risk source classification rule and the optimized target detection algorithm, a risk source identification model is constructed to identify the risk source in the communication base station environment according to the risk source identification model. The embodiment of the present invention can solve the problems existing in the prior art that rely on manual processing, effectively improve the accuracy of risk source identification in the communication base station environment, have more objective and more comprehensive characteristics, and can significantly reduce manpower and material costs, and improve identification efficiency and real-time performance. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] Figure 1 It is a flowchart of a preferred embodiment of a method for building a risk source identification model provided by the first aspect of the present invention;
[0040] Figure 2 It is a schematic diagram of the correspondence relationship between the classification rules of the communication base station environment risk sources provided by an embodiment of the present invention;
[0041] Figure 3 It is a schematic diagram of a mapping relationship between a target detection result and a risk source classification rule provided by an embodiment of the present invention;
[0042] Figure 4 It is a flowchart of a preferred embodiment of a communication base station environmental risk source identification method provided by the second aspect of the present invention;
[0043] Figure 5 It is a structural schematic diagram of a preferred embodiment of a communication base station environmental risk source identification system provided by the third aspect of the present invention;
[0044] Figure 6 It is a structural block diagram of a preferred embodiment of a terminal device provided in the eighth aspect of the present invention. DETAILED DESCRIPTION
[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this technical field without creative work are within the scope of protection of the present invention.
[0046] It should be noted that the existing communication base station environmental risk source identification methods mainly have the following problems:
[0047] Problem 1: The existing risk sources are highly concealed, and require a high level of professionalism from on-site survey and design personnel. There are also problems such as low operating efficiency, high error rate, incompleteness and poor real-time performance. Due to the lack of efficient risk identification tools, existing methods often require the dispatch of a large number of experienced professional and technical personnel to conduct on-site environmental surveys. However, due to differences in the professional skills of on-site survey and design personnel, and differences in the level of detail of on-site operations and records, it is impossible to accurately and completely give the safety risk points of the corresponding communication base stations in the corresponding environment. Therefore, the corresponding identification results cannot meet expectations and it is difficult to meet the requirements of one-stop safety briefings. In addition, if further review is required, it is mainly based on the large amount of image data collected by the surveyors at the time, which requires a lot of manual analysis and judgment. Due to the low efficiency and poor real-time performance of manual analysis, reliance on manual processing often leads to omissions, and there is a lack of corresponding means and methods to improve its efficiency and accuracy.
[0048] Problem 2: The existing risk sources are classified into many categories and have complex association rules, which bring challenges to the operation and management of communication base stations. Faced with the diversity and complex association rules of communication base station environmental risk sources, there is a lack of systematic intelligent identification means and methods for communication base station environmental risk sources, and the functions are single: the operation and management of communication base stations are faced with diversified and complex risk sources. These risk sources are not only numerous, but also have a large number of complex associations between them. The diversity of communication base station environmental risk sources requires operators and management teams to deal with multiple potential threats at the same time, and these risk sources include but are not limited to equipment failures and natural disasters. Each risk may have an adverse effect on the connectivity and availability of communication services, and therefore requires careful consideration and management; in addition, communication base station environmental risk sources not only have individual risks, but also a large number of associated risks, which increases the complexity and difficulty of risk source identification.
[0049] In order to solve the above problems, the first embodiment of the present invention provides a method for constructing a risk source identification model. Figure 1 FIG. 1 is a flowchart of a preferred embodiment of a method for constructing a risk source identification model provided by the first aspect of the present invention. The method comprises steps S11 to S14:
[0050] Step S11: construct risk source classification rules according to different types of risk sources in the communication base station environment; wherein the risk source classification rules are used to represent the corresponding relationship between risk information of risk sources, and the risk information includes risk category, risk link, risk point name, risk type and risk disposal measures;
[0051] Step S12: according to the risk source classification rule, original data with risk source characteristics under different classification rules are obtained, and a training data set is constructed according to the original data;
[0052] Step S13, training and optimizing a preset target detection algorithm according to the training data set to obtain an optimized target detection algorithm; wherein the optimization includes network structure optimization and loss function optimization of the target detection algorithm;
[0053] Step S14: construct a risk source identification model according to the risk source classification rule and the optimized target detection algorithm, so as to identify the risk sources in the communication base station environment according to the risk source identification model.
[0054] Specifically, first, according to the different types of risk sources existing in the communication base station environment, the various risk information corresponding to each risk source (including risk category, risk link, risk point name, risk type and corresponding risk disposal measures) are classified and summarized in detail to establish the risk source classification rule corresponding to the communication base station environment, and the risk source classification rule can represent the correspondence between the risk information of different types of risk sources; then, according to the established risk source classification rule, for each risk source classification rule (that is, each risk source), by collecting relevant base station survey data (for example, through professional base station survey information collection platform and historical survey database and other resources to collect basic data such as various base station environments, computer rooms, equipment, etc.), and collecting and sorting the original materials such as photos and videos containing the corresponding risk source characteristics according to each risk source classification rule, corresponding risk information under different classification rules is obtained. The original data of the risk source characteristics are obtained, and a training data set is constructed based on the obtained original data; then, based on the constructed training data set, a preset deep learning-based target detection algorithm (for example, an existing one-stage target detection algorithm, a two-stage target detection algorithm, etc.) is trained, and the target detection algorithm is optimized according to the training results, and an optimized target detection algorithm is obtained accordingly; finally, according to the established risk source classification rules and the obtained optimized target detection algorithm, a risk source identification model for the communication base station environment is constructed through the mapping relationship between the risk source target and the risk information involved in the risk source classification rules, so as to identify the risk sources in the communication base station environment according to the constructed risk source identification model, for example, in the survey and design stage, various potential risk sources such as the surrounding environment of the communication base station and the inside of the computer room are identified, and potential risk sources that may exist during the construction process are identified.
[0055] Among them, the optimization of the target detection algorithm mainly includes optimizing the network structure of the target detection algorithm and optimizing the loss function of the target detection algorithm. The accuracy of risk source target detection can be improved through network structure optimization, and the convergence speed of algorithm training can be improved through loss function optimization, thereby improving the algorithm training efficiency.
[0056] It should be noted that when actually establishing the risk source classification rules for the communication base station environment, the environmental risk points under different professions can be sorted out according to the differences in professions involved in the construction of communication base stations (for example, wireless, wired equipment, wired lines, etc.). For each profession, the environmental risk points are classified into corresponding risk categories and risk links according to the operating procedures and the corresponding links involved in the procedures, and the risk types corresponding to the environmental risk points and the corresponding risk treatment measures are further determined. Therefore, based on the correspondence between the risk categories, risk links, risk point names, risk types and risk treatment measures of environmental risk points under different professions, corresponding risk source classification rules are established.
[0057] For example, see Figure 2 As shown, it is a schematic diagram of the correspondence relationship of a communication base station environmental risk source classification rule provided by an embodiment of the present invention. Here, taking the wireless profession as an example, according to the establishment process of the above-mentioned risk source classification rule, the risk category is first divided into two standardized dimensions of on-site environmental safety risk and construction operation safety risk according to the operation process of the wireless profession. Then, according to the differences in professions, the corresponding risk links of the on-site environmental safety risk and the construction operation safety risk are listed respectively, and corresponded with the corresponding environmental risk points (i.e., the names of the risk points). After that, the corresponding specific risk types and risk disposal measures are determined by the environmental risk points, and finally a good risk source classification rule is established. Figure 2 shown.
[0058] by Figure 2 Taking the risk category of on-site environmental safety risk as an example, the corresponding risk links mainly include: site environment risk, roof environment risk, computer room environment risk, computer room equipment and cable environment risk, and fire environment risk. The corresponding risk point names mainly include: stations in low-lying areas, stations near rivers, rivers, and lakes, stations with a history of basement flooding, and stations with a history of ground flooding in the early stage. The corresponding risk types mainly include: floods. The corresponding risk management measures mainly include: 1. Improve the sealing performance of doors and windows in the computer room; 2. Regularly comb the drainage ditch outside the computer room, and prepare sandbags in the computer room to deal with rainy season floods; 3. For existing sites that have been flooded for a long time, relocation is required; 4. New base station sites should be selected in areas that are not easily affected by floods.
[0059] It should be noted that according to the classification rules for communication base station environmental risk sources, the communication base station environmental risk sources under the wireless profession include two standard risk categories (i.e., on-site environmental safety risks and construction operation safety risks), mainly including 23 risk links, 68 specific risk point names, 33 risk types, and 49 risk disposal measures. Figure 2The following are only some of the risk source classification rules for the communication base station environment under the wireless profession, and other risk source classification rules are not fully shown. It can be understood that other professions can also establish corresponding communication base station environment risk source classification rules according to methods and rules similar to those of the wireless profession.
[0060] A method for constructing a risk source identification model provided in an embodiment of the first aspect of the present invention establishes a risk source classification rule for a communication base station environment, and constructs a risk source identification model in combination with a target detection algorithm based on deep learning, so as to identify risk sources in the communication base station environment scenario according to the constructed risk source identification model. This method can solve the problems existing in the prior art that relies on manual processing, effectively improve the accuracy of identifying risk sources in the communication base station environment, has more objective and comprehensive characteristics, and can significantly reduce manpower and material costs, and improve the efficiency and real-time performance of risk source identification.
[0061] In one of the optional embodiments, constructing a training data set according to the original data specifically includes:
[0062] Preprocessing the original data to obtain standard data under different classification rules;
[0063] Performing format conversion processing on the standard data to obtain preset format data under different classification rules;
[0064] According to the risk source targets involved in the risk source classification rules, feature annotation is performed on the preset format data, and a training data set is constructed based on all the annotated data.
[0065] Specifically, in combination with the above embodiments, when constructing a training data set based on the original data, the original data can be cleaned and preprocessed first to remove invalid data and erroneous data in the original data, fill in missing values, and process outliers, etc., so as to obtain standard data under different classification rules; then, the obtained standard data under different classification rules are format converted to obtain preset format data under different classification rules, so as to convert the standard data into a data format suitable for deep learning model learning. For example, standard image data can be converted into .JPG format data; thereafter, in combination with all risk source target objects involved in the risk source classification rules, all the obtained preset format data are annotated with key features using annotation software, and all the annotated data are added to the same data set, so as to form a communication base station environment risk source training data set suitable for deep learning models.
[0066] It can be understood that when annotating key features for each preset format data, if a single preset format data only contains a single risk source target object, it can be directly annotated and added to the training data set; if a single preset format data contains multiple risk source target objects, all the contained risk source target objects need to be annotated separately and then added to the training data set, so as to ensure the integrity, relevance and accuracy of the training data set.
[0067] In one of the optional embodiments, the target detection algorithm is YOLOV5, and the network structure optimization includes: designing a CBS module in the Neck part of the original network structure of YOLOV5, and constructing an auxiliary branch based on the original network structure of YOLOV5, and the auxiliary branch adopts a ResNet structure to perform communication base station environment scene recognition.
[0068] In one of the optional embodiments, the loss function optimization includes: replacing the original CIOU loss function of YOLOV5 with an IMIOU loss function including overlap loss, center distance loss and width-height loss, and integrating the Focal Loss loss function into the IMIOU loss function.
[0069] Specifically, in combination with the above embodiments, after obtaining the communication base station environment risk source training data set, it is necessary to use the target detection algorithm for training on the basis of the training data set. In combination with the characteristics of the communication base station environment risk source training data set, taking the existing one-stage target detection algorithm as an example, in the embodiment of the present invention, the target detection algorithm adopts YOLOV5 (YouOnly Look Once version 5), that is, YOLOV5 is trained according to the obtained training data set, and YOLOV5 is optimized accordingly according to the training results to improve the accuracy of the risk source target detection algorithm and the training convergence speed, wherein:
[0070] In order to improve the accuracy of risk source target detection, the embodiment of the present invention optimizes the network structure of YOLOV5. On the one hand, a convolutional blending domain attention module CBS (Convolutional Blending Domain Attention) of the attention mechanism can be designed in the Neck part of the original network structure of YOLOV5. The CBS module combines channel attention and spatial attention. On the other hand, an auxiliary branch can be constructed based on the original network structure of YOLOV5. The auxiliary branch adopts the ResNet (deep residual network) structure to recognize the communication base station environment scene.
[0071] In order to improve the training convergence speed of the risk source target detection algorithm, some problems were found when testing the original CIOU (Complete Intersection over Union) loss function of YOLOV5. For example, the description of the aspect ratio was relatively vague, and the balance problem of difficult and easy samples was not considered. Based on this, the embodiment of the present invention optimizes the loss function of YOLOV5, and proposes to use an improved IMIOU (Improved Mean Intersection over Union) loss function to replace the original CIOU loss function of YOLOV5. The IMIOU loss function includes overlap loss, center distance loss and width and height loss, which can effectively improve the convergence speed of algorithm training. Furthermore, in order to solve the balance problem of difficult and easy samples, the embodiment of the present invention also designs a Focal Loss loss function and integrates it into the final IMIOU loss function.
[0072] Through the improvement and optimization of the above-mentioned network structure and loss function, the accuracy and efficiency of the target detection algorithm in detecting risk sources in the communication base station environment are significantly enhanced.
[0073] It should be noted that the embodiment of the present invention only takes YOLOV5 as an example of a target detection algorithm (not limited to YOLOV5), takes YOLOV5 as an improvement object, and optimizes and improves the target detection algorithm to achieve accurate and efficient identification of risk sources. It can be understood that the target detection algorithm in the embodiment of the present invention can also adopt other existing target detection algorithms (for example, SSD (Single Shot MultiBox Detector) in the one-stage target detection algorithm, R-CNN (Region-based Convolutional Neural Networks) in the two-stage target detection algorithm, etc.), and the embodiment of the present invention does not make specific limitations.
[0074] In one of the optional embodiments, constructing a risk source identification model according to the risk source classification rule and the optimized target detection algorithm specifically includes:
[0075] Detect the communication base station environment scenes under different classification rules according to the optimized target detection algorithm, and remove duplicates from all detected risk source targets to obtain a risk source target set;
[0076] Uniquely combining the risk source targets in the risk source target set according to the risk source classification rule, and constructing a mapping relationship between each group of risk source target combinations and the risk information involved in the risk source classification rule; wherein each group of risk source target combinations includes at least one risk source target;
[0077] All risk source target combinations after the mapping relationship is constructed are deduplicated according to the mapping relationship, and a risk source identification model is constructed according to the mapping relationship between the risk source target combinations after deduplication and the risk information involved in the risk source classification rules.
[0078] Specifically, in combination with the above embodiments, when constructing a risk source identification model according to the established risk source classification rules and the optimized target detection algorithm, first, for specific communication base station environment scenarios under different risk source classification rules of different professions (wireless, wired equipment, wired lines, etc.), the optimized target detection algorithm can be used to perform target detection on the specific scenario to detect all risk source targets in the scenario, and all detected risk source targets are deduplicated to eliminate the same risk source targets detected in the specific scenario, and a risk source target set is obtained accordingly, and the risk source target set contains multiple different risk source targets; secondly, in combination with the communication base station environment risk source classification rules, the risk source target set is enumerated. Various combinations of risk source targets contained in the combination, and ensuring the uniqueness of the risk source target combination, correspondingly obtain at least one group of risk source target combinations, each group of risk source target combinations contains at least one risk source target, and then establish a mapping relationship between each group of risk source target combinations after the corresponding combination and the risk information involved in the communication base station environment risk source classification rules (that is, risk category, risk link, risk point name, risk type and risk disposal measures); finally, according to the established mapping relationship, duplicate removal processing is performed again. The duplicate removal processing here mainly refers to the partial overlap of risk sources after the mapping relationship is established, and the overlapping risk sources need to be removed to achieve the output of all unique risk sources according to the communication base station environment risk source classification rules.
[0079] For example, see Figure 3 As shown, it is a schematic diagram of a mapping relationship between a target detection result and a risk source classification rule provided by an embodiment of the present invention. Here, a base station tower commonly used in the wireless profession is taken as an example. In the environmental scenario where the base station tower is located, detection and identification are performed according to the optimized target detection algorithm to obtain the corresponding target detection result, that is, a single-dimensional base station tower and a two-dimensional combination of the base station tower and the surrounding environment, buildings, etc. (for example, base station tower + mountain, base station tower + river, base station tower + high-voltage transmission line, base station tower + gas station, base station tower + building, base station tower + road, base station tower + flat land, etc.) can be identified. A multi-dimensional combination method higher than two dimensions is essentially composed of multiple two-dimensional combinations. Only the two-dimensional combination is explained here. According to the above-mentioned mapping relationship establishment process, for the single-dimensional and two-dimensional combinations, the mapping relationship established with the communication base station environment risk source classification rule is as follows: Figure 3 As shown, Figure 3Shown are only the mapping relationships between the two combinations of base station tower (basic common part) and base station tower + building (combined individual part) and the communication base station environmental risk source classification rules. The mapping relationships between other combinations and the communication base station environmental risk source classification rules are not fully shown.
[0080] by Figure 3 Taking the combination of base station towers in as an example, the mapping relationship established with the classification rules of environmental risk sources of communication base stations is as follows: the risk categories corresponding to base station towers include on-site environmental safety risks and construction safety risks. Among them, the risk links corresponding to the risk category of on-site environmental safety risks mainly include: roof environmental risks, and the corresponding risk point names mainly include: tower tilt, loose antenna / RRU (Remote Radio Unit) / outer cover, and power cables are randomly bundled, tied, and flying wires, etc. The corresponding risk types mainly include: collapse / falling objects, falling objects, and electric shock / fire. The corresponding risk disposal measures mainly include: 1. The status of the tower needs to be evaluated and identified, and those with problems must be rectified or demolished. 2. Before rectification, the construction unit shall not construct on this tower. 3. Eliminate safety hazards in a timely manner. If they cannot be handled temporarily, set up warning signs or fence on the ground or pull safety cordons. 4. Strengthen daily inspection work and feedback to relevant departments in a timely manner for problems found. 5. The site of the new base station should be selected in an area that is not easily affected by flooding.
[0081] A second aspect of the present invention provides a method for identifying a communication base station environment risk source, see Figure 4 FIG. 1 is a flowchart of a preferred embodiment of a communication base station environmental risk source identification method provided by the second aspect of the present invention, wherein the identification method comprises steps S21 to S22:
[0082] Step S21, obtaining the data to be identified of the communication base station environment to be identified;
[0083] Step S22: identify the data to be identified according to a preset risk source identification model, obtain the risk source identification result in the communication base station environment to be identified, and feed back the risk source identification result to the client for display; wherein, the risk source identification model is obtained by adopting the risk source identification model construction method described in any embodiment of the first aspect above.
[0084] Specifically, after obtaining the risk source identification model of the communication base station environment by adopting the risk source identification model construction method described in any embodiment of the first aspect above, the data to be identified of the communication base station environment to be identified can be obtained first, and then the constructed risk source identification model can be used to detect and identify the data to be identified, and the risk source identification results in the communication base station environment to be identified can be obtained accordingly, and the obtained risk source identification results can be fed back to the client for display.
[0085] It should be noted that the data to be identified of the communication base station environment to be identified can be collected and acquired by the client, and the obtained data to be identified can be sent to the server; the server is deployed with a risk source identification model obtained by adopting the risk source identification model construction method described in any embodiment of the first aspect above. After receiving the data to be identified sent by the client, the server can use the risk source identification model to identify the received data to be identified, and obtain the risk source identification result accordingly, and feed back the obtained risk source identification result to the client; after receiving the risk source identification result fed back by the server, the client can display it.
[0086] It should be noted that when the server uses the risk source identification model to identify the received data to be identified, if the corresponding security risk source is not detected from all the data to be identified, the server will promptly feedback the corresponding identification result to the client, prompting the user that it cannot be identified, or the identification result is abnormal, and prompting the user to enter the corresponding object to be identified with potential risk source characteristics; if the corresponding security risk source is detected from part of the data to be identified, it is necessary to further re-analyze the remaining unidentifiable data in the data to be identified, and combine the training data set used in the existing algorithm training to determine whether the training data is missing. If there is a missing, further supplement the missing type of data, and continuously improve the training data set to enhance the recognition ability of the model; after obtaining the risk source identification result, the server can use the risk source identification result according to Figure 3 The format shown is aggregated, classified, and deduplicated to form a risk source identification result list in a standard format, which is stored in the relevant data variables of the back-end server.
[0087] It should be noted that before displaying the risk source identification results received from the server, the client can first perform necessary processing on the variable information of the received risk source identification results, such as eliminating outliers, performing format conversion and other operations to ensure the accuracy and consistency of the data. Afterwards, the corresponding identification results are presented in an intuitive and friendly manner based on the specific content of the processed variable information. The presentation method mainly includes the identification result feedback of the uploaded data to be identified, abnormal result feedback and overall result presentation, and can export the corresponding identification results in the standard format required for communication engineering design to form a one-stop safety briefing design plan.
[0088] A communication base station environment risk source identification method provided in an embodiment of the second aspect of the present invention can identify risk sources in communication base station environment scenarios with the help of a built risk source identification model, and can effectively and intelligently identify risk sources in various communication base station environments under existing scenarios, and output accurate, comprehensive and objective various hidden danger risk sources in an intuitive presentation manner, thereby solving the problems existing in the existing technology that relies on manual processing, effectively improving the accuracy of communication base station environment risk source identification, having more objective and comprehensive characteristics, and being able to significantly reduce manpower and material costs, and improve the efficiency and real-time performance of risk source identification.
[0089] The third aspect of the present invention also provides a communication base station environment risk source identification system, which is used to implement the communication base station environment risk source identification method described in any embodiment of the second aspect, see Figure 5 , which is a schematic diagram of a preferred embodiment of a communication base station environment risk source identification system provided by the third aspect of the present invention, wherein the system includes a client and a server; wherein,
[0090] The client is used to collect the data to be identified of the communication base station environment to be identified, and send the data to be identified to the server;
[0091] The server is used to identify the data to be identified according to a preset risk source identification model, obtain the risk source identification result in the communication base station environment to be identified, and feed back the risk source identification result to the client; wherein the risk source identification model is obtained by adopting the risk source identification model construction method described in any embodiment of the first aspect above;
[0092] The client is also used to display the risk source identification result.
[0093] Exemplary, combined Figure 5As shown, an APP application is installed in the client, and the APP application includes a data acquisition module. The data acquisition module can meet the communication base station environment risk source data acquisition function in a variety of selected forms (including uploading data from local files, capturing data in real time through a camera, or selecting mobile terminal album data, etc.), and can process the collected data in real time, and upload the corresponding collected data to the server. The server is deployed with a risk source identification model and a training data set, and stores the client user registration information; when risk source detection and identification is required, the user selects the image data to be analyzed through the data selection interface of the front-end APP application and sends it to the server through the network, and waits for a response from the server. After receiving the image data sent by the client, the server uses the risk source identification model to intelligently identify the received image data, and combines the risk source classification rules of the communication base station environment to promptly feedback the identification results to the client, and presents the identification results on the front-end APP application interface. The user can query and export the final identification results based on the feedback results and the presented results.
[0094] It should be noted that the communication base station environmental risk source identification system provided in the embodiment of the third aspect of the present invention can implement all the processes of the communication base station environmental risk source identification method described in any embodiment of the second aspect above. The roles of the client and server in the system and the technical effects achieved are respectively the same as the roles of the client and server involved in the communication base station environmental risk source identification method described in the embodiment of the second aspect above and the technical effects achieved, which will not be repeated here.
[0095] The fourth aspect of the present invention further provides a risk source identification model construction device, which is used to implement the risk source identification model construction method described in any embodiment of the first aspect, and the construction device includes:
[0096] A risk source classification rule construction module is used to construct risk source classification rules according to different types of risk sources in the communication base station environment; wherein the risk source classification rules are used to represent the corresponding relationship between risk information of risk sources, and the risk information includes risk category, risk link, risk point name, risk type and risk disposal measures;
[0097] A training data set construction module is used to obtain original data with risk source characteristics under different classification rules according to the risk source classification rules, and construct a training data set based on the original data;
[0098] A target detection algorithm training and optimization module, used to train and optimize a preset target detection algorithm according to the training data set to obtain an optimized target detection algorithm; wherein the optimization includes network structure optimization and loss function optimization of the target detection algorithm;
[0099] The risk source identification model construction module is used to construct a risk source identification model according to the risk source classification rules and the optimized target detection algorithm, so as to identify the risk sources in the communication base station environment according to the risk source identification model.
[0100] Preferably, the training data set construction module specifically includes:
[0101] A raw data preprocessing unit, used to preprocess the raw data to obtain standard data under different classification rules;
[0102] A data format conversion unit, used to perform format conversion processing on the standard data to obtain preset format data under different classification rules;
[0103] The training data set construction unit is used to perform feature annotation on the preset format data according to the risk source target involved in the risk source classification rule, and construct a training data set according to all the annotated data.
[0104] Preferably, the target detection algorithm is YOLOV5, and the network structure optimization includes: designing a CBS module in the Neck part of the original network structure of YOLOV5, and constructing an auxiliary branch based on the original network structure of YOLOV5, and the auxiliary branch adopts a ResNet structure to perform communication base station environment scene recognition.
[0105] Preferably, the loss function optimization includes: replacing the original CIOU loss function of YOLOV5 with an IMIOU loss function including overlap loss, center distance loss and width-height loss, and integrating the Focal Loss loss function into the IMIOU loss function.
[0106] Preferably, the risk source identification model building module specifically includes:
[0107] A risk source target detection unit is used to detect the communication base station environment scenes under different classification rules according to the optimized target detection algorithm, and remove duplicates from all detected risk source targets to obtain a risk source target set;
[0108] A risk source target mapping unit, used to uniquely combine the risk source targets in the risk source target set according to the risk source classification rule, and to construct a mapping relationship between each group of risk source target combinations and the risk information involved in the risk source classification rule; wherein each group of risk source target combinations includes at least one risk source target;
[0109] The risk source identification model construction unit is used to remove duplicates from all risk source target combinations after the mapping relationship is constructed according to the mapping relationship, and to construct a risk source identification model according to the mapping relationship between the risk source target combination after the duplication removal and the risk information involved in the risk source classification rule.
[0110] It should be noted that a risk source identification model construction device provided in the fourth aspect of the present invention can implement all the processes of the risk source identification model construction method described in any embodiment of the first aspect above. The functions of each module and unit in the training device and the technical effects achieved are respectively the same as the functions and technical effects achieved by the risk source identification model construction method described in the first aspect above, and will not be repeated here.
[0111] The fifth aspect of the present invention further provides a communication base station environment risk source identification device, which is used to implement the communication base station environment risk source identification method described in any embodiment of the second aspect, and the identification device includes:
[0112] A module for acquiring data to be identified, used to acquire data to be identified of a communication base station environment to be identified;
[0113] A risk source identification module is used to identify the data to be identified according to a preset risk source identification model, obtain the risk source identification result in the communication base station environment to be identified, and feed back the risk source identification result to the client for display; wherein, the risk source identification model is obtained by adopting the risk source identification model construction method described in any embodiment of the first aspect above.
[0114] It should be noted that the communication base station environmental risk source identification device provided in the fifth aspect embodiment of the present invention can implement all the processes of the communication base station environmental risk source identification method described in any embodiment of the second aspect above, and the functions of each module in the identification device and the technical effects achieved are respectively the same as the functions and technical effects achieved by the communication base station environmental risk source identification method described in the second aspect embodiment above, and will not be repeated here.
[0115] The sixth aspect embodiment of the present invention also provides a computer-readable storage medium, including a stored computer program, which, when running, controls the device where the computer-readable storage medium is located to execute the risk source identification model construction method described in any embodiment of the first aspect above, or to execute the communication base station environment risk source identification method described in any embodiment of the second aspect above.
[0116] The seventh aspect embodiment of the present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the risk source identification model construction method described in any embodiment of the first aspect above, or implements the communication base station environment risk source identification method described in any embodiment of the second aspect above.
[0117] The eighth aspect of the present invention further provides a terminal device, see Figure 6 As shown, it is a structural block diagram of a preferred embodiment of a terminal device provided in the eighth aspect of the present invention, wherein the terminal device includes a processor 10, a memory 20, and a computer program stored in the memory 20 and configured to be executed by the processor 10, and when the processor 10 executes the computer program, it implements the risk source identification model construction method described in any embodiment of the first aspect above, or implements the communication base station environment risk source identification method described in any embodiment of the second aspect above.
[0118] Preferably, the computer program can be divided into one or more modules / units (such as computer program 1, computer program 2, ...), and the one or more modules / units are stored in the memory 20 and executed by the processor 10 to complete the present invention. The one or more modules / units can be a series of computer program instruction segments that can complete specific functions, and the instruction segments are used to describe the execution process of the computer program in the terminal device.
[0119] The processor 10 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor, or the processor 10 may be any conventional processor. The processor 10 is the control center of the terminal device, and various parts of the terminal device are connected using various interfaces and lines.
[0120] The memory 20 mainly includes a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function, etc., and the data storage area can store related data, etc. In addition, the memory 20 can be a high-speed random access memory, or a non-volatile memory, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, and a flash card (Flash Card), etc., or the memory 20 can also be other volatile solid-state storage devices.
[0121] It should be noted that the above terminal device may include, but is not limited to, a processor and a memory. Those skilled in the art will understand that Figure 6 The structural block diagram shown is only a structural example of the above-mentioned terminal device and does not constitute a structural limitation of the above-mentioned terminal device. The above-mentioned terminal device may include more or fewer components than shown in the figure, or combine certain components, or different components.
[0122] In summary, the risk source identification model construction method, communication base station environment risk source identification method, system, device, computer-readable storage medium, computer program product and terminal device provided by the embodiments of the present invention have the following beneficial effects:
[0123] (1) The embodiments of the present invention solve the problems that the existing communication base station environmental risk source identification mainly relies on manual means to conduct on-site surveys, analysis and recording, resulting in inefficiency and inaccuracy. It can effectively improve the accuracy of communication base station environmental risk source identification, has more objective and comprehensive characteristics, and can significantly reduce manpower and material costs, and improve identification efficiency and real-time performance;
[0124] (2) The embodiment of the present invention systematically establishes the classification rules of communication base station environment risk sources and establishes the corresponding risk source identification model, which can quickly and accurately identify various hidden risk sources in the existing communication base station environment. Compared with traditional manual identification or other existing intelligent analysis methods, it has the characteristics of being more comprehensive, systematic and efficient;
[0125] (3) The embodiment of the present invention realizes one-stop automatic output of personalized safety design briefing, achieving the goal of one-stop one safety briefing.
[0126] In addition, the embodiments of the present invention can be applied to government authorities, operators, etc. During on-site safety inspections, with the help of the embodiments of the present invention, the on-site base station environmental safety risk sources can be quickly identified, and various risk sources and risk management measures can be output in a relatively systematic, comprehensive and accurate manner. At the same time, the embodiments of the present invention can also be applied to design units to enhance the design units' ability to accurately reflect various safety risk sources in the design and achieve one-stop safety briefings.
[0127] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A method for constructing a risk source identification model, characterized in that: include: According to different types of risk sources in the communication base station environment, a risk source classification rule is constructed; wherein the risk source classification rule is used to represent the corresponding relationship between risk information of the risk source, and the risk information includes risk category, risk link, risk point name, risk type and risk disposal measures; According to the risk source classification rules, original data with risk source characteristics under different classification rules are obtained, and a training data set is constructed according to the original data; Training and optimizing a preset target detection algorithm according to the training data set to obtain an optimized target detection algorithm; wherein the optimization includes network structure optimization and loss function optimization of the target detection algorithm; A risk source identification model is constructed according to the risk source classification rule and the optimized target detection algorithm to identify risk sources in a communication base station environment according to the risk source identification model.
2. The risk source identification model construction method according to claim 1, characterized in that: The constructing of a training data set according to the original data specifically includes: Preprocessing the original data to obtain standard data under different classification rules; Performing format conversion processing on the standard data to obtain preset format data under different classification rules; According to the risk source targets involved in the risk source classification rules, feature annotation is performed on the preset format data, and a training data set is constructed based on all the annotated data.
3. The risk source identification model construction method according to claim 1, characterized in that: The target detection algorithm is YOLOV5, and the network structure optimization includes: designing a CBS module in the Neck part of the original network structure of YOLOV5, and constructing an auxiliary branch based on the original network structure of YOLOV5, and the auxiliary branch adopts a ResNet structure to perform communication base station environment scene recognition.
4. The method for constructing a risk source identification model according to claim 3, characterized in that: The loss function optimization includes: using the IMIOU loss function including overlap loss, center distance loss and width-height loss to replace the original CIOU loss function of YOLOV5, and integrating the FocalLoss loss function into the IMIOU loss function.
5. The method for constructing a risk source identification model according to any one of claims 1 to 4, characterized in that: The step of constructing a risk source identification model according to the risk source classification rule and the optimized target detection algorithm specifically includes: Detect the communication base station environment scenes under different classification rules according to the optimized target detection algorithm, and remove duplicates from all detected risk source targets to obtain a risk source target set; Uniquely combining the risk source targets in the risk source target set according to the risk source classification rule, and constructing a mapping relationship between each group of risk source target combinations and the risk information involved in the risk source classification rule; wherein each group of risk source target combinations includes at least one risk source target; All risk source target combinations after the mapping relationship is constructed are deduplicated according to the mapping relationship, and a risk source identification model is constructed according to the mapping relationship between the risk source target combinations after deduplication and the risk information involved in the risk source classification rules.
6. A method for identifying environmental risk sources of a communication base station, characterized in that: include: Acquire data to be identified of a communication base station environment to be identified; The data to be identified is identified according to a preset risk source identification model, a risk source identification result in the communication base station environment to be identified is obtained, and the risk source identification result is fed back to the client for display; wherein, the risk source identification model is obtained by adopting the risk source identification model construction method described in any one of claims 1 to 5.
7. A communication base station environmental risk source identification system, characterized in that: Including client and server; among them, The client is used to collect the data to be identified of the communication base station environment to be identified, and send the data to be identified to the server; The server is used to identify the data to be identified according to a preset risk source identification model, obtain the risk source identification result in the communication base station environment to be identified, and feed back the risk source identification result to the client; wherein the risk source identification model is obtained by adopting the risk source identification model construction method according to any one of claims 1 to 5; The client is also used to display the risk source identification result.
8. A risk source identification model construction device, characterized in that: include: A risk source classification rule construction module is used to construct risk source classification rules according to different types of risk sources in the communication base station environment; wherein the risk source classification rules are used to represent the corresponding relationship between risk information of risk sources, and the risk information includes risk category, risk link, risk point name, risk type and risk disposal measures; A training data set construction module is used to obtain original data with risk source characteristics under different classification rules according to the risk source classification rules, and construct a training data set based on the original data; A target detection algorithm training and optimization module, used to train and optimize a preset target detection algorithm according to the training data set to obtain an optimized target detection algorithm; wherein the optimization includes network structure optimization and loss function optimization of the target detection algorithm; The risk source identification model construction module is used to construct a risk source identification model according to the risk source classification rules and the optimized target detection algorithm, so as to identify the risk sources in the communication base station environment according to the risk source identification model.
9. A communication base station environmental risk source identification device, characterized in that: include: A module for acquiring data to be identified, used to acquire data to be identified of a communication base station environment to be identified; A risk source identification module is used to identify the data to be identified according to a preset risk source identification model, obtain the risk source identification result in the communication base station environment to be identified, and feed back the risk source identification result to the client for display; wherein, the risk source identification model is obtained by adopting the risk source identification model construction method described in any one of claims 1 to 5.
10. A computer-readable storage medium, characterized in that: It includes a stored computer program, which, when running, controls the device where the computer-readable storage medium is located to execute the risk source identification model construction method as described in any one of claims 1 to 5, or executes the communication base station environment risk source identification method as described in claim 6.
11. A computer program product, characterized in that It includes a computer program, which, when executed by a processor, implements the risk source identification model construction method as described in any one of claims 1 to 5, or implements the communication base station environment risk source identification method as described in claim 6.
12. A terminal device, characterized in that: It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the risk source identification model construction method as described in any one of claims 1 to 5, or implements the communication base station environment risk source identification method as described in claim 6.
