A method and system for monitoring power construction safety based on Beidou positioning assistance

Abstract

The present invention provides a method and system for monitoring the safety of electric construction based on Beidou positioning assistance, and relates to the technical field of construction safety monitoring. The method includes: obtaining the real-time collected positioning data and construction status data of electric construction personnel, constructing a construction feature set including multiple electric construction features of electric construction personnel; determining abnormal features in the construction feature set, determining a target feature database according to the type of electric construction currently performed by the electric construction personnel, performing feature abnormality detection on the abnormal features based on the target feature database, and generating abnormal detection results of the abnormal features; if the abnormal detection result contains the type of construction safety abnormality to which the abnormal feature belongs, a construction safety alarm is issued to the electric construction personnel based on the construction safety abnormality class to which the abnormal feature belongs. The present invention realizes the improvement of the detection accuracy and intelligence level of electric construction safety monitoring.

Description

Technical Field

[0001] The present invention relates to the technical field of construction safety monitoring, and in particular to a method and system for monitoring electric power construction safety based on Beidou positioning assistance. Background Art

[0002] The complexity of power construction scenes and environments makes the construction process prone to construction risks. Safety monitoring of power construction personnel is an important means to ensure construction safety. Traditional safety monitoring methods rely more on human management and manual inspection, which has many shortcomings, such as delayed response, limited coverage, and low monitoring accuracy. With the development of intelligent and automated technologies, monitoring systems based on sensors and the Internet of Things have gradually been introduced into the field of power construction to improve the automation level of construction safety monitoring.

[0003] In the existing construction safety monitoring technology, Beidou positioning system and sensor equipment are usually used to collect data such as location information, equipment status and physiological parameters of power construction personnel, so as to accurately monitor the construction safety of power construction personnel through the collected data. However, some technologies are simply used for real-time location tracking or separate monitoring of equipment status when processing these data, lacking the deep fusion of multi-source data and comprehensive safety monitoring capabilities in complex scenarios. For example, the detection of abnormal parameters relies more on preset rules or simple threshold comparisons. This method may lead to high false alarm rates or missed reports in the changeable and complex environment of power construction, and cannot accurately reflect the potential safety hazards in the construction process. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a method and system for power construction safety monitoring based on Beidou positioning assistance, which adopts multi-feature fusion analysis and context-based intelligent anomaly detection mechanism to improve the accuracy and coverage of anomaly detection while making the anomaly detection process more flexible and efficient, and can better adapt to the complex and changeable power construction environment, thereby improving the detection accuracy and intelligence level of power construction safety monitoring.

[0005] To achieve the above-mentioned object, the first aspect of the present invention provides a method for monitoring power construction safety based on Beidou positioning assistance, comprising:

[0006] Acquire the real-time positioning data and construction status data of the electric power construction personnel, extract the trajectory characteristics and motion characteristics of the electric power construction personnel from the positioning data, extract the equipment characteristics and physiological characteristics of the electric power construction personnel from the construction status data, and construct a construction feature set that includes multiple electric power construction characteristics of the electric power construction personnel;

[0007] Determine abnormal features in the construction feature set, determine a target feature database according to the type of power construction currently performed by the power construction personnel, perform feature anomaly detection on the abnormal features based on the target feature database, and generate anomaly detection results of the abnormal features;

[0008] Among them, performing feature anomaly detection on the abnormal feature based on the target feature database includes determining multiple construction safety anomaly types associated with the abnormal feature in the target feature database, performing context feature-based anomaly detection on the abnormal feature according to the construction feature fusion rule corresponding to the construction safety anomaly type, and generating anomaly detection results of the abnormal feature;

[0009] If the anomaly detection result contains the type of construction safety anomaly to which the anomaly feature belongs, a construction safety alarm is issued to the power construction personnel based on the construction safety anomaly class to which the anomaly feature belongs.

[0010] Preferably, the abnormal features are subjected to context-based abnormality detection according to the construction feature fusion rule corresponding to the construction safety abnormality type, including:

[0011] Determine multiple construction safety anomaly types associated with the abnormal feature, extract multiple groups of first abnormal sample data under each construction safety anomaly type from the target feature database, and extract second abnormal sample data under each construction safety anomaly type for each construction feature in the abnormal feature from the multiple groups of first abnormal sample data;

[0012] According to the construction safety anomaly type to which each group of second abnormal sample data belongs, a construction feature fusion rule for performing context feature-based anomaly detection on the abnormal features under each construction safety anomaly type is determined, and an abnormal risk value of each group of second abnormal sample data is determined according to the construction feature fusion rule;

[0013] The calculation of the abnormal risk value includes: completing the contextual features of the abnormal features according to the construction feature set, generating the target abnormal features, matching the target abnormal features with the second abnormal sample data by feature type, determining multiple construction feature combinations, calculating the similarity between the construction features in each construction feature combination, determining the feature importance factor of each construction feature combination according to the construction feature fusion rule corresponding to the second abnormal sample data, performing weighted summation on the similarity between the construction features in each construction feature combination according to the multiple feature importance factors, and calculating the abnormal risk value of the second abnormal sample data;

[0014] Determine the target sample data with the highest abnormal risk value among multiple groups of second abnormal sample data, and judge whether the abnormal risk value of the target sample data is greater than the preset risk threshold. If so, mark the abnormal features for construction safety abnormalities according to the construction safety abnormality type corresponding to the target sample data, and generate an abnormal detection result for the abnormal features.

[0015] Preferably, the construction feature fusion rules corresponding to the construction safety anomaly types further include:

[0016] Acquire multiple groups of normal sample data without construction safety anomalies, perform difference analysis on multiple groups of first abnormal sample data corresponding to each construction safety anomaly type in the target feature database based on the multiple groups of normal sample data, and extract multiple difference factors between each group of first abnormal sample data corresponding to each construction safety anomaly type and the multiple groups of normal sample data, wherein each construction feature corresponds to one of the difference factors;

[0017] The feature importance factor of each construction feature is determined according to the difference factor of each construction feature in multiple groups of first abnormal sample data, and the construction feature fusion rule corresponding to each construction safety abnormality type is constructed according to the feature importance factor of each construction feature under each construction safety abnormality type.

[0018] Preferably, the extraction of the difference factor comprises:

[0019] Determine multiple groups of sub-sample data for each construction feature in multiple groups of normal sample data, extract the normal construction feature vector corresponding to each group of sub-sample data, perform feature fusion on multiple normal construction feature vectors of the same construction feature, and generate a first target reference vector for each construction feature;

[0020] Extract multiple groups of first abnormal sample data corresponding to each construction safety abnormality type, and multiple abnormal construction feature vectors corresponding to each construction feature. Calculate the difference factor of each group of first abnormal sample data with respect to each construction feature based on the first target reference vector of each construction feature, including extracting the abnormal difference vector of the first target reference vector under the construction feature compared with each abnormal construction feature vector, calculating the difference reference value of each abnormal difference vector, fusing the multiple difference reference values ​​of each construction feature, and generating the difference factor of each construction feature compared with multiple groups of normal sample data.

[0021] Preferably, extracting second abnormal sample data of each construction feature in the abnormal feature under each construction safety abnormality type from the multiple groups of first abnormal sample data includes:

[0022] Determine the first construction feature vector corresponding to each construction feature in the abnormal feature, extract the second construction feature vector corresponding to each construction feature of each group of first abnormal sample data, calculate the distance parameters between the first construction feature vector and multiple second construction feature vectors respectively, select multiple groups of first abnormal sample data whose distance parameters are greater than a preset distance threshold, and obtain the second abnormal sample data of each construction feature under each construction safety abnormality type.

[0023] Preferably, extracting an abnormal difference vector between the first target reference vector under the construction feature and each abnormal construction feature vector comprises:

[0024] For the first target reference vector and multiple abnormal construction feature vectors corresponding to each construction feature, each abnormal construction feature vector is processed, including calculating the difference between each element in the abnormal construction feature vector and the first target reference vector by subtracting each other item by item, and normalizing the multiple differences to generate an abnormal difference vector corresponding to each abnormal construction feature vector.

[0025] A second aspect of the present invention provides a power construction safety monitoring system based on Beidou positioning assistance, which is used to implement any of the above-mentioned power construction safety monitoring methods based on Beidou positioning assistance, including:

[0026] A data acquisition module is used to acquire real-time positioning data and construction status data of power construction personnel;

[0027] The feature processing module is used to extract the trajectory features and motion features of the electric power construction personnel from the positioning data, and to extract the equipment features and physiological features of the electric power construction personnel from the construction status data, so as to construct a construction feature set including multiple electric power construction features of the electric power construction personnel;

[0028] An anomaly detection module is used to determine abnormal features in the construction feature set, determine a target feature database according to the type of power construction currently performed by the power construction personnel, perform feature anomaly detection on the abnormal features based on the target feature database, and generate anomaly detection results of the abnormal features;

[0029] The construction safety early warning module is used to issue a construction safety alarm to power construction personnel based on the construction safety anomaly class to which the abnormal feature belongs when the abnormal detection result contains the construction safety anomaly type to which the abnormal feature belongs.

[0030] Preferably, the anomaly detection module performs feature anomaly detection on the abnormal features based on the target feature database, including:

[0031] Determine multiple construction safety anomaly types associated with the abnormal features in the target feature database, perform context feature-based anomaly detection on the abnormal features according to the construction feature fusion rules corresponding to the construction safety anomaly types, and generate anomaly detection results of the abnormal features.

[0032] The present invention has the following beneficial effects:

[0033] The present invention analyzes the real-time positioning data of construction personnel obtained based on the Beidou positioning system in combination with construction status data, constructs a construction feature set containing multi-source feature data, so as to fully reflect the actual situation of the construction site; screens out potential abnormal features through preliminary analysis, and performs more in-depth abnormal detection based on the target feature database to determine the construction safety abnormality type associated with the abnormal features. By completing and fusing the context features of the abnormal features, it is possible to carry out fine-grained mining and detection of complex and potential safety risks, which is helpful to timely discover and deal with hidden dangers in the construction process; by modeling the association between different construction features and constructing fusion rules, the abnormal features can be detected based on context features, and the overall information of multiple construction features can be combined to improve the accuracy and context perception of feature abnormality detection, and avoid the limitations of single feature detection; in the process of constructing the construction feature fusion rules, the importance of each construction feature can be identified through the calculation of the difference factor and the determination of the feature importance factor, and the weight in the abnormal detection can be dynamically adjusted according to the degree of deviation of different features, thereby enhancing the flexible response ability to abnormal situations and improving the accuracy of abnormal analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 A flowchart of a method for power construction safety monitoring based on Beidou positioning assistance is provided in an embodiment of the present invention.

[0035] Figure 2 A structural block diagram of a power construction safety monitoring system based on Beidou positioning assistance provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0036] In order to make those skilled in the art better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0037] Figure 1 A flowchart of a method for monitoring power construction safety based on Beidou positioning assistance is provided in an embodiment of the present invention, such as Figure 1 As shown, a method for monitoring power construction safety based on Beidou positioning assistance specifically includes the following steps:

[0038] Step S1, obtaining real-time positioning data and construction status data of electric power construction personnel, extracting multiple electric power construction features from the positioning data and construction status data, and constructing a construction feature set including multiple electric power construction features of electric power construction personnel.

[0039] Specifically, the real-time positioning data and construction status data of the power construction personnel are obtained from the power construction site. The positioning data can be the real-time position data of the power construction personnel recorded at a certain frequency from the Beidou positioning system, and the construction status data is collected through various sensors, including the operation status data of the equipment and the physiological monitoring data of the personnel. The power construction features such as trajectory features and motion features can be extracted from the positioning data. By analyzing the location information and motion status of the construction personnel, the relevant trajectory features such as path curve, speed, acceleration, etc., as well as motion features such as moving direction, path length, moving frequency, etc. can be extracted. The power construction features such as equipment features and physiological features can be extracted from the construction status data. By analyzing the usage status data of the construction equipment and the physiological data of the construction personnel, the relevant equipment features such as operating time, fault warning, etc., as well as physiological features such as heart rate, body temperature, etc. can be extracted. The multiple power construction features extracted from the positioning data and the construction status data are integrated to construct a construction feature set containing multi-source feature data.

[0040] Step S2, determine the abnormal features in the construction feature set, determine the target feature database according to the type of power construction currently performed by the power construction personnel, perform feature anomaly detection on the abnormal features based on the target feature database, and generate anomaly detection results of the abnormal features.

[0041] Specifically, the constructed construction feature set is analyzed, and the multiple power construction features contained therein are first analyzed to identify potential abnormal features. These abnormal features may indicate that some construction activities do not meet expectations or there are potential safety risks. Exemplarily, a threshold detection or rule detection method is used, such as setting a reasonable threshold, interval range, etc. to determine whether the feature is abnormal, such as heart rate exceeding a certain range, abnormal acceleration of movement speed, etc. It should be noted that in the specific application process, the preliminary abnormality analysis can be used to analyze some minor anomalies. For example, after the conventional detection of the movement speed reaching a certain value, it can be directly regarded as an abnormal movement and an alarm is issued. Then, in the preliminary analysis process, a lower threshold can be set to play the role of early warning. Since potential minor anomalies are not easy to identify, they may be included in multiple data, and comprehensive analysis can be used for mining. Therefore, in view of this situation, the following scheme is used to perform feature anomaly detection to mine potential anomalies in the data.

[0042] Then, according to the type of power construction corresponding to the power construction task currently performed by the power construction personnel, the corresponding target feature database is determined. The target feature database contains multiple types of construction safety anomalies that may appear in the construction tasks under this type and their corresponding sample data, as well as the construction feature fusion rules corresponding to each type of construction safety anomaly. The construction feature fusion rules are constructed through the data analysis process. The determined abnormal features are subjected to feature anomaly detection through the target feature database, specifically including determining multiple construction safety anomaly types associated with the abnormal features in the target feature database, and then performing context feature-based anomaly detection on the abnormal features according to the construction feature fusion rules corresponding to the construction safety anomaly types, generating anomaly detection results of the abnormal features, which are used to indicate whether the abnormal features have their corresponding construction safety anomaly types, so as to take corresponding safety measures.

[0043] Step S3: If the abnormality detection result contains the construction safety abnormality type to which the abnormality feature belongs, a construction safety alarm is issued to the power construction personnel based on the construction safety abnormality type to which the abnormality feature belongs.

[0044] Specifically, check the abnormal detection results to determine whether there are any identified types of construction safety anomalies. If the detection results show that the abnormal characteristics do correspond to a known construction safety risk, then according to the confirmed type of construction safety anomaly, trigger the corresponding construction safety alarm, such as on-site sound and light alarms, push notifications on mobile devices, alarm log records in the background management system, etc. At the same time as the alarm is issued, corresponding emergency measures can also be recommended. For example, if an abnormal physiological parameter of a person is detected, it can be recommended to take a rest immediately or seek medical help; if an abnormal operation of the equipment is detected, it can be recommended to stop the equipment immediately for inspection to ensure the safety of the construction environment.

[0045] The above steps can be used to finely manage and dynamically monitor the safety of the power construction site, thereby improving the level of construction safety. The above content briefly summarizes the overall solution of this application, and some specific implementation details will be further elaborated below.

[0046] In one implementation process, for the process of performing feature anomaly detection on abnormal features based on the target feature database in step S2, performing context feature-based anomaly detection on abnormal features according to the construction feature fusion rule corresponding to the construction safety anomaly type, specifically includes:

[0047] Step S21, determining multiple construction safety anomaly types associated with the abnormal feature, extracting multiple groups of first abnormal sample data under each construction safety anomaly type from the target feature database, and extracting second abnormal sample data under each construction safety anomaly type for each construction feature in the abnormal feature from the multiple groups of first abnormal sample data;

[0048] Specifically, multiple construction safety anomaly types associated with abnormal features are determined, such as equipment failure, abnormal physical condition of construction personnel, violation of work process, etc. The abnormal features contain one or more construction features, and different construction safety anomaly types contain different construction features, so that multiple construction safety anomaly types containing one or more construction features in the abnormal features can be determined. Then, multiple groups of first abnormal sample data under each construction safety anomaly type are extracted from the target feature database. The multiple groups of first abnormal sample data contained in each construction safety anomaly type record the relevant feature information in the construction safety event corresponding to the construction safety anomaly type, such as the location data and construction status data of the power construction personnel collected in this process. Then, based on each construction feature, multiple groups of second abnormal sample data are further extracted from the multiple groups of first abnormal sample data. The second abnormal sample data is specifically data in the first abnormal sample data that has a strong correlation with a certain construction feature in the abnormal feature. For example, for a construction feature such as equipment feature, among multiple sets of first abnormal sample data, only some of the first abnormal sample data may be similar to the equipment feature involved in the abnormal feature in terms of equipment feature and have a strong correlation. In this way, multiple sets of second abnormal sample data with a strong correlation with each construction feature in the abnormal feature can be determined among multiple sets of first abnormal sample data.

[0049] In this embodiment, for the selection of multiple groups of second abnormal sample data, the first construction feature vector corresponding to each construction feature in the abnormal feature is first determined, and the second construction feature vector corresponding to each construction feature of each group of first abnormal sample data is extracted. Then, the distance parameters between the first construction feature vector and the multiple second construction feature vectors are calculated respectively, for example, the Euclidean distance between the first construction feature vector and the multiple second construction feature vectors is calculated, and multiple groups of first abnormal sample data with distance parameters greater than a preset distance threshold are selected. Each group of first abnormal sample data selected may come from data under different construction safety abnormality types, and it is recorded as the second abnormal sample data under the construction safety abnormality type, so as to obtain the second abnormal sample data of each construction feature under each construction safety abnormality type.

[0050] Step S22: According to the construction safety anomaly type to which each group of second abnormal sample data belongs, determine the construction feature fusion rule for performing context feature-based anomaly detection on the abnormal features under each construction safety anomaly type, and determine the abnormal risk value of each group of second abnormal sample data according to the construction feature fusion rule.

[0051] Specifically, the target feature database contains construction feature fusion rules under each construction safety anomaly type, which are used to indicate the correlation between different construction features under different construction safety anomaly types, so as to improve the detection accuracy and context perception ability. Then, based on the construction feature fusion rules, the determined abnormal features can be subjected to context feature-based anomaly detection, and the abnormal risk value of the abnormal features for each group of second abnormal sample data can be calculated.

[0052] The calculation of the abnormal risk value specifically includes: completing the context feature of the abnormal feature according to the construction feature set, generating the target abnormal feature, matching the target abnormal feature with the second abnormal sample data by feature type, and determining multiple construction feature combinations.

[0053] It is worth noting that the abnormal feature may not completely include all the power construction features. In the process of feature screening, reference data is screened based on some key features, namely the determined abnormal features, and then in the process of data analysis, the remaining power construction features are added thereto through feature completion operations to form a complete target abnormal feature, which includes multiple power construction features. The second abnormal sample data also includes multiple power construction features. After matching the target abnormal feature with the second abnormal sample data by feature type, multiple construction feature combinations can be determined. Each construction feature combination corresponds to one of the power construction features and includes the feature vectors corresponding to the power construction feature in the target abnormal feature and the second abnormal sample data respectively.

[0054] Then, the similarity between the construction features in each construction feature combination, that is, the similarity between two feature vectors, is calculated. Finally, according to the construction feature fusion rule corresponding to the second abnormal sample data, multiple calculated similarities are processed to achieve feature fusion. Specifically, the feature importance factor of each construction feature combination is determined according to the construction feature fusion rule, so that the multiple feature importance factors are used as weight parameters for weighting the similarity between the construction features of each construction feature combination. The similarity between the construction features in each construction feature combination is weighted and summed according to the multiple feature importance factors, and finally the abnormal risk value of the second abnormal sample data is calculated.

[0055] Step S23, determine the target sample data with the highest abnormal risk value among multiple groups of second abnormal sample data, and judge whether the abnormal risk value of the target sample data is greater than the preset risk threshold. If so, mark the abnormal feature for construction safety abnormality according to the construction safety abnormality type corresponding to the target sample data, and generate an abnormal detection result of the abnormal feature.

[0056] Specifically, after obtaining the abnormal risk values ​​of all second abnormal sample data, determine the target sample data with the highest abnormal risk value among multiple groups of second abnormal sample data. Then, determine whether the abnormal risk value of the target sample data is greater than the preset risk threshold. If the risk value of the target sample data exceeds the preset threshold, indicating that there is a significant safety risk, the abnormal feature is marked as a construction safety abnormality according to the construction safety abnormality type corresponding to the target sample data, and finally generate an abnormal detection result of the abnormal feature so that corresponding safety measures can be taken in time.

[0057] In one implementation process, the construction feature fusion rules corresponding to the construction safety anomaly types are constructed in the following way:

[0058] A plurality of groups of normal sample data without construction safety anomalies are obtained, and a difference analysis is performed on a plurality of groups of first abnormal sample data corresponding to each construction safety anomaly type in the target feature database based on the plurality of groups of normal sample data.

[0059] Specifically, first obtain multiple groups of normal sample data without construction safety anomalies. These normal sample data come from various characteristic manifestations in the normal state during the power construction process, including trajectory characteristics, movement characteristics, equipment characteristics, physiological characteristics, etc. of construction personnel. These normal sample data are used as a benchmark to evaluate the differences in features under different types of construction safety anomalies. Next, a difference analysis is performed on multiple groups of first abnormal sample data corresponding to each construction safety anomaly type in the target feature database. Specifically, it includes extracting multiple difference factors from each group of first abnormal sample data corresponding to each construction safety anomaly type compared to multiple groups of normal sample data, where each construction feature corresponds to one of the difference factors.

[0060] In this embodiment, the extraction of the difference factor specifically includes:

[0061] Determine multiple groups of sub-sample data for each construction feature in multiple groups of normal sample data, which contain power construction feature information collected under normal construction conditions, including trajectory characteristics, movement status, equipment usage, physiological parameters, etc. of personnel. Extract the normal construction feature vector corresponding to each group of sub-sample data, each construction feature contains a normal construction feature vector, perform feature fusion on multiple normal construction feature vectors of the same construction feature, and generate the first target reference vector of each construction feature. Specifically, the weighted average method can be used to achieve the fusion of multiple normal construction feature vectors to construct a benchmark feature representing the construction feature under normal construction conditions.

[0062] Extract multiple groups of first abnormal sample data corresponding to each construction feature, and multiple abnormal construction feature vectors corresponding to each construction feature. The abnormal construction feature vector reflects the abnormal performance of each construction feature under a specific construction safety abnormality type, which may include feature changes that significantly deviate from the normal state, such as abnormal vibration frequency of equipment, abnormal movement trajectory of personnel, etc. Calculate the difference factor of each group of first abnormal sample data on each construction feature based on the first target reference vector of each construction feature.

[0063] In this process, for the first target reference vector and multiple abnormal construction feature vectors under any construction feature, the abnormal difference vector of the first target reference vector under the construction feature compared with each abnormal construction feature vector is extracted. The abnormal difference vector contains the deviation of the abnormal construction feature vector compared with the reference feature. Specifically, the abnormal difference vector of each abnormal construction feature vector is generated by normalizing the difference of each element by subtracting each other. Then, each abnormal difference vector is subjected to feature conversion, specifically, the difference reference value of each abnormal difference vector is calculated to realize the conversion of the vector into a difference factor representing the comprehensive level. In this embodiment, the difference reference value of each abnormal difference vector is obtained by square summing each element of the abnormal difference vector, thereby reflecting the degree of deviation of the construction feature under abnormal conditions. Finally, multiple difference reference values ​​of each construction feature are fused, for example, the fusion process adopts a weighted average method to generate a difference factor of each construction feature compared with multiple groups of normal sample data. The difference factor reflects the overall deviation degree of the construction feature under abnormal conditions relative to the normal state, and can be used as an important indicator for judging the abnormality of the construction feature.

[0064] After determining the difference factors of each construction feature in multiple groups of first abnormal sample data in the above manner, the difference factors of each construction feature are used as the feature importance factors of each construction feature, wherein the larger the feature importance factor, the greater the deviation of the corresponding construction feature from the benchmark feature, and the greater the importance of the construction feature in the subsequent abnormal analysis. After determining the feature importance factor of each construction feature, the construction feature fusion rule corresponding to each construction safety abnormality type is constructed according to the feature importance factor of each construction feature under each construction safety abnormality type. The construction feature fusion rule can be used to indicate the fusion of information between different construction features, thereby realizing a fine-grained evaluation of abnormal features and improving the accuracy of power construction safety monitoring.

[0065] The power construction safety monitoring method based on Beidou positioning assistance provided by the embodiment of the present invention analyzes the real-time positioning data of construction personnel obtained based on the Beidou positioning system in combination with construction status data, and constructs a construction feature set containing multi-source feature data to fully reflect the actual situation of the construction site; through preliminary analysis, potential abnormal features are screened out, and more in-depth abnormal detection is performed based on the target feature database to determine the construction safety abnormality type associated with the abnormal feature. By completing and fusing the context features of the abnormal features, fine-grained mining and detection of complex and potential safety risks can be performed, which helps to timely discover and deal with hidden dangers in the construction process; by modeling the association between different construction features and constructing fusion rules, the abnormal features can be detected based on context features, and the overall information of multiple construction features can be combined to improve the accuracy and context perception of feature abnormality detection, avoiding the limitations of single feature detection; in the process of constructing the construction feature fusion rule, the importance of each construction feature can be identified by calculating the difference factor and determining the feature importance factor, and the weight in abnormal detection can be dynamically adjusted according to the degree of deviation of different features, thereby enhancing the flexible response to abnormal situations and improving the accuracy of abnormal analysis. The Beidou positioning system provides high-precision real-time support for the positioning data of construction personnel. Combined with the construction status data obtained by sensors, the entire safety monitoring process has a high degree of dynamics. This real-time nature can promptly reflect changes on the construction site and provide a guarantee for rapid safety intervention.

[0066] Figure 2 A structural block diagram of a power construction safety monitoring system based on Beidou positioning assistance provided by an embodiment of the present invention, such as Figure 2 As shown, a power construction safety monitoring system based on Beidou positioning assistance includes:

[0067] A data acquisition module is used to acquire real-time positioning data and construction status data of power construction personnel;

[0068] The feature processing module is used to extract the trajectory features and motion features of the electric power construction personnel from the positioning data, and to extract the equipment features and physiological features of the electric power construction personnel from the construction status data, so as to construct a construction feature set including multiple electric power construction features of the electric power construction personnel;

[0069] An anomaly detection module is used to determine abnormal features in the construction feature set, determine a target feature database according to the type of power construction currently performed by the power construction personnel, perform feature anomaly detection on the abnormal features based on the target feature database, and generate anomaly detection results of the abnormal features;

[0070] Among them, feature anomaly detection of abnormal features based on the target feature database includes:

[0071] Determine multiple construction safety anomaly types associated with the abnormal features in the target feature database, perform context feature-based anomaly detection on the abnormal features according to the construction feature fusion rules corresponding to the construction safety anomaly types, and generate anomaly detection results of the abnormal features.

[0072] According to the construction feature fusion rules corresponding to the construction safety anomaly type, the anomaly features are detected based on context features, including:

[0073] Determine multiple construction safety anomaly types associated with the abnormal feature, extract multiple groups of first abnormal sample data under each construction safety anomaly type from the target feature database, and extract second abnormal sample data under each construction safety anomaly type for each construction feature in the abnormal feature from the multiple groups of first abnormal sample data;

[0074] According to the construction safety anomaly type to which each group of second abnormal sample data belongs, a construction feature fusion rule for performing context feature-based anomaly detection on the abnormal features under each construction safety anomaly type is determined, and an abnormal risk value of each group of second abnormal sample data is determined according to the construction feature fusion rule;

[0075] The calculation of the abnormal risk value includes: completing the contextual features of the abnormal features according to the construction feature set, generating the target abnormal features, matching the target abnormal features with the second abnormal sample data by feature type, determining multiple construction feature combinations, calculating the similarity between the construction features in each construction feature combination, determining the feature importance factor of each construction feature combination according to the construction feature fusion rule corresponding to the second abnormal sample data, performing weighted summation on the similarity between the construction features in each construction feature combination according to the multiple feature importance factors, and calculating the abnormal risk value of the second abnormal sample data;

[0076] Determine the target sample data with the highest abnormal risk value among multiple groups of second abnormal sample data, and judge whether the abnormal risk value of the target sample data is greater than the preset risk threshold. If so, mark the abnormal features for construction safety abnormalities according to the construction safety abnormality type corresponding to the target sample data, and generate an abnormal detection result for the abnormal features.

[0077] The construction safety early warning module is used to issue a construction safety alarm to power construction personnel based on the construction safety anomaly class to which the abnormal feature belongs when the abnormal detection result contains the construction safety anomaly type to which the abnormal feature belongs.

[0078] Preferably, the anomaly detection module performs feature anomaly detection on the abnormal features based on the target feature database, including:

[0079] Determine multiple construction safety anomaly types associated with the abnormal features in the target feature database, perform context feature-based anomaly detection on the abnormal features according to the construction feature fusion rules corresponding to the construction safety anomaly types, and generate anomaly detection results of the abnormal features.

[0080] The above-mentioned power construction safety monitoring system based on Beidou positioning assistance is specifically used to implement the power construction safety monitoring method based on Beidou positioning assistance as described in any of the above-mentioned implementation modes.

[0081] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A method for monitoring power construction safety based on Beidou positioning assistance, characterized in that: include: Acquire the real-time positioning data and construction status data of the electric power construction personnel, extract the trajectory characteristics and motion characteristics of the electric power construction personnel from the positioning data, extract the equipment characteristics and physiological characteristics of the electric power construction personnel from the construction status data, and construct a construction feature set that includes multiple electric power construction characteristics of the electric power construction personnel; Determine abnormal features in the construction feature set, determine a target feature database according to the type of power construction currently performed by the power construction personnel, perform feature anomaly detection on the abnormal features based on the target feature database, and generate anomaly detection results of the abnormal features; Among them, performing feature anomaly detection on the abnormal feature based on the target feature database includes determining multiple construction safety anomaly types associated with the abnormal feature in the target feature database, performing context feature-based anomaly detection on the abnormal feature according to the construction feature fusion rule corresponding to the construction safety anomaly type, and generating anomaly detection results of the abnormal feature; If the anomaly detection result contains the type of construction safety anomaly to which the anomaly feature belongs, a construction safety alarm is issued to the power construction personnel based on the construction safety anomaly class to which the anomaly feature belongs.

2. The method for monitoring power construction safety based on Beidou positioning assistance according to claim 1 is characterized in that: The performing of anomaly detection based on context features on the abnormal features according to the construction feature fusion rules corresponding to the construction safety abnormality types includes: Determine multiple construction safety anomaly types associated with the abnormal feature, extract multiple groups of first abnormal sample data under each construction safety anomaly type from the target feature database, and extract second abnormal sample data under each construction safety anomaly type for each construction feature in the abnormal feature from the multiple groups of first abnormal sample data; According to the construction safety anomaly type to which each group of second abnormal sample data belongs, a construction feature fusion rule for performing context feature-based anomaly detection on the abnormal features under each construction safety anomaly type is determined, and an abnormal risk value of each group of second abnormal sample data is determined according to the construction feature fusion rule; The calculation of the abnormal risk value includes: completing the contextual features of the abnormal features according to the construction feature set, generating the target abnormal features, matching the target abnormal features with the second abnormal sample data by feature type, determining multiple construction feature combinations, calculating the similarity between the construction features in each construction feature combination, determining the feature importance factor of each construction feature combination according to the construction feature fusion rule corresponding to the second abnormal sample data, performing weighted summation on the similarity between the construction features in each construction feature combination according to the multiple feature importance factors, and calculating the abnormal risk value of the second abnormal sample data; Determine the target sample data with the highest abnormal risk value among multiple groups of second abnormal sample data, and judge whether the abnormal risk value of the target sample data is greater than the preset risk threshold. If so, mark the abnormal features for construction safety abnormalities according to the construction safety abnormality type corresponding to the target sample data, and generate an abnormal detection result for the abnormal features.

3. The method for monitoring power construction safety based on Beidou positioning assistance according to claim 1 is characterized in that: The construction feature fusion rules corresponding to the construction safety anomaly types are obtained by the following steps: Acquire multiple groups of normal sample data without construction safety anomalies, perform difference analysis on multiple groups of first abnormal sample data corresponding to each construction safety anomaly type in the target feature database based on the multiple groups of normal sample data, and extract multiple difference factors between each group of first abnormal sample data corresponding to each construction safety anomaly type and the multiple groups of normal sample data, wherein each construction feature corresponds to one of the difference factors; The feature importance factor of each construction feature is determined according to the difference factor of each construction feature in multiple groups of first abnormal sample data, and the construction feature fusion rule corresponding to each construction safety abnormality type is constructed according to the feature importance factor of each construction feature under each construction safety abnormality type.

4. The method for monitoring power construction safety based on Beidou positioning assistance according to claim 3 is characterized in that: The step of extracting multiple difference factors between each group of first abnormal sample data corresponding to each construction safety abnormality type and multiple groups of normal sample data includes: Determine multiple groups of sub-sample data for each construction feature in multiple groups of normal sample data, extract the normal construction feature vector corresponding to each group of sub-sample data, perform feature fusion on multiple normal construction feature vectors of the same construction feature, and generate a first target reference vector for each construction feature; Extract multiple groups of first abnormal sample data corresponding to each construction safety abnormality type, and multiple abnormal construction feature vectors corresponding to each construction feature. Calculate the difference factor of each group of first abnormal sample data with respect to each construction feature based on the first target reference vector of each construction feature, including extracting the abnormal difference vector of the first target reference vector under the construction feature compared with each abnormal construction feature vector, calculating the difference reference value of each abnormal difference vector, fusing the multiple difference reference values ​​of each construction feature, and generating the difference factor of each construction feature compared with multiple groups of normal sample data.

5. The method for monitoring power construction safety based on Beidou positioning assistance according to claim 2 is characterized in that: The step of extracting second abnormal sample data of each construction feature in the abnormal features under each construction safety abnormality type from the multiple groups of first abnormal sample data includes: Determine the first construction feature vector corresponding to each construction feature in the abnormal feature, extract the second construction feature vector corresponding to each construction feature of each group of first abnormal sample data, calculate the distance parameters between the first construction feature vector and multiple second construction feature vectors respectively, select multiple groups of first abnormal sample data whose distance parameters are greater than a preset distance threshold, and obtain the second abnormal sample data of each construction feature under each construction safety abnormality type.

6. The method for monitoring power construction safety based on Beidou positioning assistance according to claim 4 is characterized in that: The abnormal difference vector between the first target reference vector under the extracted construction feature and each abnormal construction feature vector includes: For the first target reference vector and multiple abnormal construction feature vectors corresponding to each construction feature, each abnormal construction feature vector is processed, including calculating the difference between each element in the abnormal construction feature vector and the first target reference vector by subtracting each other item by item, and normalizing the multiple differences to generate an abnormal difference vector corresponding to each abnormal construction feature vector.

7. A power construction safety monitoring system based on Beidou positioning assistance, characterized in that: The system is used to implement a method for monitoring power construction safety based on Beidou positioning assistance as described in any one of claims 1 to 6, comprising: A data acquisition module is used to acquire real-time positioning data and construction status data of power construction personnel; The feature processing module is used to extract the trajectory features and motion features of the electric power construction personnel from the positioning data, and to extract the equipment features and physiological features of the electric power construction personnel from the construction status data, so as to construct a construction feature set including multiple electric power construction features of the electric power construction personnel; An anomaly detection module is used to determine abnormal features in the construction feature set, determine a target feature database according to the type of power construction currently performed by the power construction personnel, perform feature anomaly detection on the abnormal features based on the target feature database, and generate anomaly detection results of the abnormal features; The construction safety early warning module is used to issue a construction safety alarm to power construction personnel based on the construction safety anomaly class to which the abnormal feature belongs when the abnormal detection result contains the construction safety anomaly type to which the abnormal feature belongs.

8. The power construction safety monitoring system based on Beidou positioning assistance according to claim 7 is characterized in that: The anomaly detection module performs feature anomaly detection on the abnormal features based on the target feature database, including: Determine multiple construction safety anomaly types associated with the abnormal features in the target feature database, perform context feature-based anomaly detection on the abnormal features according to the construction feature fusion rules corresponding to the construction safety anomaly types, and generate anomaly detection results of the abnormal features.

Images