Method and device for identifying unsafe behavior of drilling operation personnel

Abstract

The embodiment of the application provides a kind of unsafe behavior identification method and device of drilling operation personnel, it is related to data analysis technical field, including: by eye movement data sensor, the eye movement data information of drilling operation personnel is acquired, eye movement data information is handled, obtains the eye movement characteristic parameter corresponding to eye movement data information, eye movement characteristic parameter is input into identification model, the output result of identification model is acquired;According to output result, the operation behavior of operation personnel is determined, and when operation behavior is unsafe operation behavior, warning is sent to operation personnel.The eye movement data information of drilling operation personnel can be analyzed in real time, whether the operation behavior of drilling operation personnel conforms to the regulation can be judged, and the real-time monitoring and identification of the unsafe operation behavior of drilling operation personnel are realized.

Description

Technical Field

[0001] This application relates to the field of data analysis, and in particular to a method and apparatus for identifying unsafe behaviors of drilling workers. Background Technology

[0002] Drilling operations are one of the high-risk areas in the oil and gas industry. Studies have found that more than 80% of drilling accidents are related to unsafe human behavior. Therefore, identifying unsafe behaviors of drilling personnel is key to reducing drilling accidents.

[0003] Current research on methods for identifying unsafe behaviors of drilling workers mainly falls into two categories: one is to use manual inspection or post-incident investigation and analysis methods, and the other is to use video surveillance to collect unsafe behaviors of workers and then use artificial intelligence algorithms for identification.

[0004] However, neither of the above two technical solutions can achieve real-time monitoring and identification of unsafe behaviors of drilling personnel throughout the entire process. Summary of the Invention

[0005] This application provides a method and apparatus for identifying unsafe behaviors of drilling workers. By using an eye-tracking sensor worn by the drilling worker to capture their eye movement data in real time, and analyzing this data, it is possible to determine whether the worker's work behavior is normal. This enables real-time monitoring and identification of unsafe behaviors by drilling workers.

[0006] Firstly, the application provides a method for identifying unsafe behaviors of drilling workers, including:

[0007] Eye movement data of the operator's gaze at each area within the drilling operation scenario is obtained from an eye movement data sensor, wherein the operation scenario includes multiple sensitive and non-sensitive areas;

[0008] The eye movement data is processed to obtain the corresponding eye movement feature parameters, which include: sensitive area fixation point parameters, sensitive area fixation duration parameters, time parameters for the first fixation to the sensitive area, time parameters for the second fixation to the sensitive area, and average fixation duration parameters.

[0009] Input the eye-tracking feature parameters into the recognition model and obtain the output results of the recognition model;

[0010] The operator's work behavior is determined based on the output results, and a warning is issued to the operator when the work behavior falls into a preset category.

[0011] Optional eye-tracking data information includes: total number of fixation points, time of a single eye-tracking data acquisition, average fixation duration, time of first fixation on the sensitive area, time of second fixation on the sensitive area, number of fixation points in the sensitive area, and fixation time in the sensitive area.

[0012] Optionally, the eye-tracking data is processed to obtain eye-tracking feature parameters corresponding to the eye-tracking data, including:

[0013] The sensitive area fixation point parameters are obtained by the ratio of the sum of the number of fixation points in each sensitive area within the work scene to the total number of fixation points in the work scene.

[0014] The fixation duration parameter of sensitive areas is obtained by the ratio of the sum of fixation duration of each sensitive area in the work scene to the duration of a single eye movement data acquisition.

[0015] The time parameter for the first fixation on a sensitive area is obtained by the ratio of the minimum first fixation time of each sensitive area in the work scenario to the duration of a single eye movement data acquisition.

[0016] The time parameter for the second fixation on the sensitive area is obtained by comparing the ratio of the second fixation time to the duration of a single eye movement data acquisition within the described work scenario.

[0017] The average fixation duration parameter is determined by the ratio of the sum of fixation durations at N fixation points within the task scenario to N times the duration of a single eye movement data acquisition.

[0018] Optionally, the eye-tracking feature parameters are input into the recognition model to obtain the output results of the recognition model, including: inputting the eye-tracking feature parameters into the recognition model to obtain multiple classification values, wherein the recognition model includes multiple sub-classification models, and each classification model outputs a classification value; determining the operator's work behavior based on the output results, including: determining the operator's work behavior based on the multiple classification values.

[0019] Optionally, the operator's work behavior can be determined based on multiple classification values, including:

[0020] Among multiple classification values, the largest classification value is obtained, where the classification value ranges from 0 to 1; the classification corresponding to the largest classification value is determined as the operator's work behavior.

[0021] Optionally, methods for identifying unsafe behaviors of drilling workers may also include:

[0022] Obtain training data of various types; if the training data corresponds to the first sub-classification model, the label value of the first sub-classification model is 1, and the label value of other second sub-classification models is 0; train multiple sub-classification models in the recognition model to obtain the trained recognition model.

[0023] Optionally, the work behavior can be one of the following: normal behavior, violation of regulations, wrong behavior, or safety hazard.

[0024] Secondly, this application provides a device for identifying unsafe behaviors of drilling workers, comprising:

[0025] The data acquisition module is used to collect eye-tracking data of drilling personnel within the operational scenario.

[0026] The data processing module is used to filter and process the eye movement data of the workers to obtain eye movement feature parameters;

[0027] The analysis module is used to analyze eye movement feature parameters and obtain analysis results;

[0028] The alert module is used to determine the operator's work behavior based on the analysis results, and to issue a reminder to the operator when the work behavior falls into a preset category.

[0029] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0030] The memory is used to store computer instructions; the processor is used to execute the computer instructions stored in the memory to implement the method of any one of the first aspects.

[0031] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement the method of any of the first aspects.

[0032] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the method of any one of the first aspects.

[0033] The method and apparatus for identifying unsafe behaviors of drilling workers provided in this application acquire eye movement data information of drilling workers through eye movement data sensors worn by the workers, process the eye movement data information to obtain eye movement feature parameters corresponding to the eye movement data information, input the eye movement feature parameters into the recognition model, obtain multiple classification values ​​output by the recognition model, determine the work behavior based on the maximum classification value, and issue a warning to the workers when the work behavior is one of the abnormal work behaviors. By analyzing the eye movement data information of drilling workers in real time, it is determined whether the work behavior of drilling workers complies with regulations, thereby realizing real-time monitoring and identification of unsafe work behaviors of drilling workers. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of a scenario provided for an embodiment of this application;

[0035] Figure 2 A flowchart illustrating the method for identifying unsafe behaviors of drilling workers provided in this application embodiment;

[0036] Figure 3 This is a schematic diagram of the sensitive area gridding provided in an embodiment of this application;

[0037] Figure 4 A schematic diagram illustrating the amorphous sensitive region division provided for embodiments of this application;

[0038] Figure 5 A flowchart illustrating the method for identifying unsafe behaviors of drilling workers provided in this application embodiment;

[0039] Figure 6 This is a schematic diagram of the recognition model provided in the embodiments of this application;

[0040] Figure 7 A schematic diagram of an unsafe behavior identification device for drilling workers provided in an embodiment of this application;

[0041] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0042] To facilitate a clear description of the technical solutions in the embodiments of this application, some terms and technologies involved in the embodiments of this application will be briefly introduced below:

[0043] 1) Eye-tracking technology is a technique that uses eye-tracking data sensors to record information such as the gaze position, saccade path, and gaze duration of a subject to understand their attention levels. Eye-tracking technology can not only monitor and identify unsafe behaviors of workers in real time, but also effectively identify erroneous actions that have not yet occurred and safety hazards caused by the worker's own actions.

[0044] 2) Sensitive areas refer to the areas where important equipment units are located during drilling operations, typically including the areas where drilling equipment instruments, valves, drill bits, equipment control panels, etc. are located.

[0045] 3) Other terms

[0046] In the embodiments of this application, the terms "first" and "second" are used to distinguish identical or similar items with essentially the same function and effect, without limiting their order. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that the terms "first" and "second" do not necessarily imply that they are different.

[0047] It should be noted that, in the embodiments of this application, the terms "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design scheme described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0048] The method provided in the embodiments of this application will be described in detail below with reference to the accompanying drawings. It should be noted that "at...time" in the embodiments of this application can refer to the instant at which a certain situation occurs, or to a period of time after the occurrence of a certain situation. The embodiments of this application do not specifically limit this.

[0049] Drilling operations are one of the high-risk areas in the oil and gas industry. Studies have found that more than 80% of drilling accidents are related to unsafe human behavior. Therefore, identifying unsafe behaviors of drilling personnel is key to reducing drilling accidents.

[0050] Current research on methods for identifying unsafe behaviors of drilling workers mainly falls into two categories: one is to use manual inspection or post-incident investigation and analysis methods, and the other is to use video surveillance to collect unsafe behaviors of workers and then use artificial intelligence algorithms for identification.

[0051] However, neither of the above two technical solutions can achieve real-time monitoring and identification of drilling personnel, especially when there are blind spots in monitoring or when inspections are not timely.

[0052] In view of this, this application proposes a method and apparatus for identifying unsafe behaviors of drilling workers based on eye-tracking data. The aim is to achieve the monitoring and identification of unsafe behaviors of drilling workers by combining the ability of eye-tracking sensors to capture human eye-tracking data in real time, the wearability of eye-tracking sensors, and the data classification advantages of integrated learning algorithms.

[0053] The technical solutions of the present invention and how they solve the above-mentioned technical problems will be described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present invention will now be described with reference to the accompanying drawings.

[0054] Figure 1 This is a schematic diagram of a scenario according to an embodiment of this application, such as... Figure 1 As shown, it includes an operating console, drilling equipment, drilling personnel, a server, and the work glasses and wristbands worn by the drilling personnel during operation.

[0055] Among them, the work glasses and wristbands can interact with the server.

[0056] Drilling workers are required to wear work glasses and wristbands during operations. The work glasses are equipped with eye-tracking sensors to collect eye movement data from the workers. The wristbands can receive alert messages from the server to remind workers of unsafe behaviors.

[0057] When drilling operations are underway, operators need to focus on the control panel and drilling equipment. Therefore, their line of sight needs to switch back and forth between the control panel and the drilling equipment. By analyzing the eye movement data of drilling operators, it can be determined whether they are paying attention to the control panel and drilling equipment, and thus determine whether the operators are engaging in any unsafe behaviors.

[0058] It is understood that the work glasses and wristband can be other wearable devices, such as helmets, hats, goggles, etc., which can realize the functions of collecting eye-tracking data and receiving message reminders in the embodiments of this application. The embodiments of this application do not limit this.

[0059] The application scenarios of the embodiments of this application have been described above. The following is a detailed description of how the embodiments of this application identify unsafe behaviors of drilling personnel.

[0060] Figure 2 This is a flowchart illustrating the method for identifying unsafe behaviors of drilling workers provided in an embodiment of this application, applied to... Figure 1 Taking the server in the example, the following steps are included:

[0061] S201. Obtain eye movement data information of the operator's gaze at each area in the drilling operation scene from the eye movement data sensor.

[0062] Eye-tracking data information includes information such as the gaze position, saccade path, and gaze duration of drilling personnel, recorded by eye-tracking data sensors.

[0063] Drilling operations encompass multiple sensitive and non-sensitive areas, such as... Figure 1 The work scenario shown includes three sensitive areas, with the remaining areas being non-sensitive areas.

[0064] A common method for delineating sensitive areas in a work environment is to divide the work environment into a grid, with all units containing key equipment designated as sensitive areas. Figure 3As shown. However, due to the large number of important equipment units in drilling field operations, their irregular shapes, and significant differences in placement, the grid adaptability is poor (too much blank space in the grid, incomplete grid coverage, or too many grids), which can lead to inaccurate data recording or slow calculation speed. Therefore, this application's embodiment adopts an amorphous sensitive area division method.

[0065] like Figure 4 As shown, the amorphous sensitive region division method used in this application embodiment can select different sensitive region division methods according to the function of each important unit and for different important types.

[0066] For example, important unit A belongs to the division method where important functional areas exist around it; important unit B belongs to the division method where all functional areas are located within the unit and not near the edge; important unit C belongs to the division method where important functional areas exist only on one or some faces.

[0067] The server can interact with eye-tracking data sensors to obtain eye-tracking data information of the worker's gaze at each area within the work scene.

[0068] The server can interact with the eye-tracking data sensor once per minute, or this can be configured according to actual needs.

[0069] S202. Process the eye movement data to obtain the eye movement feature parameters corresponding to the eye movement data.

[0070] The eye movement data recorded by the eye movement data sensor can be called raw eye movement data. Due to the large variety of types, in order to avoid the impact of raw eye movement data on the accuracy of recognition results, it is necessary to filter out eye movement data that is not affected or has a small impact on the number of sensitive areas based on the differences between multiple scenes and multiple sensitive areas, and calculate the corresponding eye movement feature parameters.

[0071] S203. Input the eye-tracking feature parameters into the recognition model and obtain the output results of the recognition model.

[0072] The recognition model is a pre-trained model used to identify unsafe behaviors of drilling workers based on the acquired eye-tracking data.

[0073] The server inputs the calculated eye-tracking feature parameters into the recognition model and obtains the corresponding results output by the recognition model.

[0074] S204. Determine the operator's work behavior based on the output results, and issue a warning to the operator when the work behavior is a preset category.

[0075] The output of the identification model can be 0 or 1. The preset category can be abnormal work behavior, which can also be called unsafe work behavior.

[0076] For example, when the output of the recognition model is 0, the behavior of the drilling operator is normal and not a preset category, and the server can proceed with the next data collection and processing.

[0077] When the recognition model outputs a value of 1, the corresponding drilling worker's behavior is considered unsafe, matching the preset category. In this case, the server can use the wristband worn by the drilling worker to issue a warning, such as a vibration or sound alert, to correct the unsafe behavior.

[0078] In the above method, eye-tracking data of drilling workers is collected using eye-tracking sensors. This data is then processed to obtain corresponding eye-tracking feature parameters. A trained recognition model is used to analyze these parameters, generating analysis results. Based on these results, an alert is issued when the drilling worker's behavior is deemed unsafe. By using real-time eye-tracking data from drilling workers, it is possible to determine whether unsafe work behaviors exist, achieving real-time monitoring and identification of unsafe behaviors by drilling personnel.

[0079] Figure 5 A flowchart illustrating Embodiment 2 of the method for identifying unsafe behaviors of drilling workers provided in this application includes the following steps:

[0080] S501. Obtain eye movement data information of the operator's gaze at each area in the drilling operation scene from the eye movement data sensor.

[0081] In this embodiment, step S501 and Figure 2 Step S201 in the illustrated embodiment is similar and will not be repeated here.

[0082] In order to identify unsafe behaviors of drilling personnel in different drilling operation scenarios, it is necessary to maintain a uniform proportion of sensitive areas in each scenario to the total area of ​​each scenario. In this embodiment, the proportion of sensitive areas in each scenario to the total area of ​​each scenario is set to 10%.

[0083] The calculation method for the proportion of sensitive areas in each scene to the total area of ​​each scene satisfies the following formula:

[0084]

[0085] Among them, S os - The area of ​​the entire work scene image; - The area of ​​each sensitive area within the work scenario; - The proportion of sensitive areas in the entire work scene image.

[0086] S502. Filter the acquired eye-tracking data.

[0087] To avoid excessive computation caused by too much eye-tracking data and to adapt to the task of identifying unsafe behaviors of drilling workers in different work scenarios, it is necessary to filter the eye-tracking data of drilling workers collected by eye-tracking data sensors to obtain eye-tracking data that can accurately analyze the work behavior of drilling workers.

[0088] The filtered eye-tracking data includes the following types:

[0089] Total number of fixation points, time of single eye movement data acquisition, average fixation duration, time to first fixation on the sensitive area, time to second fixation on the sensitive area, number of fixation points in the sensitive area, and fixation time in the sensitive area.

[0090] The total number of gaze points refers to the total number of gaze positions of the drilling operator's eyes within a single data acquisition time. One gaze position is one gaze point. If the drilling operator's gaze lingers at a certain gaze position for more than a preset time, that gaze position is considered a gaze point. The preset time can be 250 milliseconds or can be set according to actual needs.

[0091] The single eye movement data acquisition time is the time during which the eye movement sensor collects eye movement data of the drilling operator within a preset time. For example, if 20 seconds is set as the single eye movement data acquisition time, the eye movement sensor can collect the eye movement data of the drilling operator in a 20-second cycle and send it to the server.

[0092] The average fixation duration is the average duration of fixation at each fixation point by the drilling operator; the time of first fixation on a sensitive area is the duration of the first fixation on a sensitive area by the drilling operator; the time of second fixation on a sensitive area is the duration of the second fixation on a sensitive area by the drilling operator; the number of fixation points in a sensitive area is the total number of fixation points in a sensitive area by the drilling operator during one data acquisition period; and the fixation time in a sensitive area is the total duration of fixation on a sensitive area by the drilling operator.

[0093] The server filters the eye movement data of drilling personnel recorded by the eye movement data sensor to obtain the above 7 types of eye movement data information, and can then execute the steps shown in S503.

[0094] S503. Process the filtered eye movement data to obtain the eye movement feature parameters corresponding to the filtered eye movement data.

[0095] To address the impact of differences in the number and location of sensitive areas across various drilling scenarios on the identification of unsafe behaviors by drilling personnel, it is necessary to calculate the selected eye-tracking data to obtain the corresponding eye-tracking feature parameters.

[0096] Among them, eye movement characteristic parameters include the following: sensitive area fixation point parameters, sensitive area fixation duration parameters, time parameters for the first fixation on the sensitive area, time parameters for the second fixation on the sensitive area, and average fixation duration parameters.

[0097] Specifically, the calculation method for each eye-tracking feature parameter satisfies the following formula:

[0098] Sensitive area fixation point parameters:

[0099]

[0100] Where sfp1, fp2, ..., sfp n - Number of fixation points in each sensitive area of ​​the scene; fp - Total number of fixation points in the scene; P sfp - Sensitive area fixation point parameters.

[0101] For example, in a certain work scenario, during a single eye-tracking data acquisition session, there are 11 fixation points, of which 6 fixation points fall within the sensitive area and 5 fixation points fall outside the sensitive area. Using the formula above, the sensitive area fixation point parameter for this drilling operator can be calculated to be 54.5%.

[0102] Sensitive area fixation duration parameters:

[0103]

[0104] in, —Gaze duration in each sensitive area within the scene; t dc —Time to collect eye movement data in a single session; —Sensitive area fixation duration parameter.

[0105] For example, in a certain work scenario, during a single eye-tracking data acquisition session (20 seconds), the fixation duration in the sensitive area was 8960 milliseconds. Using the formula for the sensitive fixation duration parameter, the sensitive fixation duration parameter for this drilling worker was calculated to be 44.8%.

[0106] Time parameters for the first fixation on the sensitive area:

[0107]

[0108] in, - First fixation time in each sensitive area within the scene; t dc -Time to acquire eye movement data in a single session; - The time parameter for the first gaze of the sensitive area.

[0109] For example, in a certain work scenario, within a single eye-tracking data acquisition time (20 seconds), the minimum first fixation time for each sensitive area is 7228 milliseconds. Using the formula for the first fixation time of the sensitive area, it can be calculated that the drilling operator's first fixation time of the sensitive area is 36.1%.

[0110] The time parameter for the second fixation on the sensitive area:

[0111]

[0112] Among them, t ssa -Second fixation time in sensitive areas within the scene; t dc -Time to acquire eye movement data in a single session; - The time parameter for the second gaze to the sensitive area.

[0113] For example, in a certain work scenario, within a single eye-tracking data acquisition time (20 seconds), the minimum second fixation time for each sensitive area is 7695 milliseconds. Using the formula for the time parameter of the second fixation on the sensitive area, it can be calculated that the drilling operator's first fixation time on the sensitive area is 36.1%.

[0114] Mean fixation duration parameter:

[0115]

[0116] in, - Fixation duration at each fixation point throughout the entire work scenario (including fixation duration at each fixation point within sensitive areas); t dc -Time to acquire eye movement data in a single session; -Mean gaze duration parameter.

[0117] For example, in a certain work scenario, the average fixation duration during a single eye-tracking data acquisition session (20 seconds) is 1735.875 milliseconds. The average fixation duration parameter for this drilling worker can be calculated using the formula for the average fixation duration ratio. It is 86.8%.

[0118] The server processes the filtered eye movement data to obtain the above five eye movement feature parameters, and can then execute the steps shown in S504.

[0119] S504. Input the eye-tracking feature parameters into the recognition model and obtain the output results of the recognition model.

[0120] The recognition model is a pre-trained model, and the training process of the recognition model is as follows:

[0121] Obtain various types of training data.

[0122] In this embodiment, the XGBoost model is used as the training object for the recognition model, wherein the recognition model includes a 4-seed classification model. Figure 6 As shown, the specific classification model can be categorized into four types: normal, violation, error, and safety hazard.

[0123] The server obtains the eye movement feature parameters of drilling workers corresponding to the four seed classification models as training data.

[0124] If the training data corresponds to the first sub-classification model, then the label value of the first sub-classification model is 1, and the label value of the other second sub-classification models is 0.

[0125] Multiple sub-classification models in the recognition model are trained to obtain the trained recognition model.

[0126] For example, the normal sub-classification model in the recognition model is trained. The eye movement data information of the drilling operator's normal work behavior is processed into the corresponding eye movement feature parameters and then input into the four sub-classification models respectively. The label value of the normal sub-classification model is set to 1, and the label value of the other three sub-classification models is set to 0.

[0127] The training methods for the other three sub-classification models are similar to those for the normal sub-classification model. After multiple training sessions, the trained recognition model is obtained.

[0128] The server will acquire real-time eye-tracking data from drilling workers, process it into corresponding eye-tracking feature parameters, and input them into four sub-classification models in the trained recognition model to obtain four classification values ​​output by the recognition model.

[0129] Each category value ranges from 0 to 1.

[0130] For example, the four classification values ​​obtained by inputting eye-tracking feature parameters into the recognition model can be shown in Table 1 below:

[0131] Table 1 shows the eye movement data collected from four drilling workers, processed into corresponding eye movement feature parameters, and the four classification values ​​obtained by inputting the eye movement feature parameters into the recognition model.

[0132]

[0133] S505. Determine whether the operator's work behavior is unsafe based on the output results.

[0134] This application's embodiments categorize unsafe work behaviors into violations, errors, and safety hazards.

[0135] Violation of regulations refers to an action that has already occurred, but is a deliberate error; incorrect behavior refers to an action that has already occurred, but is an unintentional error; safety hazard refers to an action that has not yet occurred, but an error may occur in the future.

[0136] The server obtains four classification values ​​output by the recognition model. Among the four classification values, the largest classification value is selected, and the classification corresponding to the largest classification value is determined as the operator's work behavior.

[0137] For example, the server obtains four classification values ​​output by the recognition model: 0.005 from the normal sub-classification model, 0.984 from the violation sub-classification model, 0.005 from the error sub-classification model, and 0.006 from the safety hazard sub-classification model. The highest classification value is 0.984 from the violation sub-classification model; therefore, the corresponding worker's work behavior is considered a violation.

[0138] If it is not an unsafe work activity, the steps shown in S501 can be performed.

[0139] If the work is unsafe, the steps shown in S506 can be followed.

[0140] S506. Generate warning messages and issue reminders.

[0141] Warning messages can be text messages, used to record abnormal work behavior of drilling personnel at a certain time, and can be stored on a server.

[0142] For example, when the server determines that the drilling worker A's work behavior is a violation of regulations, the server records in the background that A committed the violation at xx year xx month xx hour.

[0143] At the same time, the server can alert drilling workers through the wristbands they wear, for example, by providing vibration or sound reminders, to correct unsafe behaviors.

[0144] Actual testing has shown that the embodiments of this application can achieve 100% accuracy in identifying unsafe behaviors of drilling personnel, which meets the engineering application requirements for unsafe behaviors in drilling operations.

[0145] The importance of the eye-tracking data feature parameters proposed in this application is shown in Table 2:

[0146]

[0147] As shown in Table 2, the five eye-tracking data feature parameters proposed in this application all play an important role in identifying unsafe behaviors in drilling operations. Among them, the average fixation duration parameter and the time parameter of the first fixation on the sensitive area are the most important, while the fixation point parameter of the sensitive area and the time parameter of the second fixation on the sensitive area are relatively less important. This indicates that the location of the sensitive area, the degree of attention concentration, the eye habits of different operators, and the position of the fixation point at the end of the previous round of eye-tracking data acquisition have a significant impact on the identification of unsafe behaviors. However, numerically, the proposed identification solution has a good generalization effect on the identification of unsafe behaviors in multiple scenarios.

[0148] S507. Determine whether unsafe work practices of drilling personnel have been corrected.

[0149] The server uses the eye-tracking data of the drilling workers collected in the next cycle to determine whether the workers' behavior is normal.

[0150] If this is normal work behavior, it means that the abnormal work behavior of the operator has been corrected. The steps shown in S501 can be executed.

[0151] If unsafe work practices continue, the steps shown in S508 can be performed.

[0152] S508. Notify the on-site safety management personnel.

[0153] On-site safety management personnel are responsible for correcting unsafe work practices by drilling workers.

[0154] In the above method, eye-tracking data sensors are used to collect eye-tracking data from drilling workers. This data is processed into corresponding eye-tracking feature parameters, which are then input into a trained recognition model. Multiple classification values ​​are obtained from the model's output. The maximum classification value determines the work behavior, and if the behavior is classified as abnormal, the drilling worker is alerted to correct the abnormal behavior. This method achieves 24 / 7 monitoring and identification of unsafe work behaviors by drilling workers, overcoming the shortcomings of traditional methods such as blind spots and missing monitoring points. Furthermore, by calculating the corresponding eye-tracking feature parameters, the method improves the model's insufficient generalization ability, enabling the identification of unsafe behaviors by drilling workers in multiple scenarios.

[0155] like Figure 7 As shown in the figure, this application embodiment also provides an unsafe behavior identification device 70 for drilling operators, including a data acquisition module 701, a data processing module 702, an analysis module 703, and an alarm module 704.

[0156] The data acquisition module 701 is used to collect eye movement data of drilling personnel within the work scene information.

[0157] The data processing module 702 is used to filter and process the eye movement data of drilling personnel to obtain eye movement feature parameters, including: sensitive area fixation point parameters, sensitive area fixation duration parameters, time parameters of the first fixation on the sensitive area, time parameters of the second fixation on the sensitive area, and average fixation duration parameters.

[0158] Analysis module 703 is used to analyze eye movement feature parameters and obtain analysis results.

[0159] Specifically, the analysis module 703 inputs eye-tracking feature parameters into the recognition model and obtains the results output by the recognition model.

[0160] The warning module 704 is used to determine the drilling operator's work behavior based on the analysis results, and to issue a reminder to the operator when the work behavior is a preset category.

[0161] The unsafe behavior identification device for drilling workers provided in this embodiment can perform... Figure 2 The technical solutions of the method embodiments shown are similar in principle and in effect, and will not be described again here.

[0162] Furthermore, embodiments of this application also provide a device for identifying unsafe behaviors of drilling workers. Figure 7 Based on the unsafe behavior identification device 70 for drilling workers provided in the embodiment, it also includes a model training module 705.

[0163] The data processing module 702 is also used to filter the following 7 types of eye movement data information based on the collected eye movement data information:

[0164] Total number of fixation points, time of single eye movement data acquisition, average fixation duration, time to first fixation on the sensitive area, time to second fixation on the sensitive area, number of fixation points in the sensitive area, and fixation time in the sensitive area.

[0165] The analysis module 703 is also used to obtain multiple classification values ​​based on eye movement feature parameters.

[0166] Specifically, the eye-tracking feature parameters are input into the recognition model to obtain the classification values ​​corresponding to the four sub-classification models in the recognition model. Among the four classification values, the largest classification value is obtained, where the classification value ranges from 0 to 1.

[0167] The identification module 704 is also used to determine the category corresponding to the largest classification value as the operator's work behavior.

[0168] Model training module 705 is used to train the recognition model.

[0169] Specifically, this involves acquiring various types of training data.

[0170] If the training data corresponds to the first sub-classification model, then the label value of the first sub-classification model is 1, and the label value of the other second sub-classification models is 0.

[0171] Multiple sub-classification models in the recognition model are trained to obtain the trained recognition model.

[0172] The unsafe behavior identification device for drilling workers provided in this embodiment can perform... Figure 5 The technical solutions of the method embodiments shown are similar in principle and in effect, and will not be described again here.

[0173] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 8 As shown, the device 80 provided in this embodiment may include:

[0174] Processor 801.

[0175] Memory 802 is used to store executable instructions for electronic devices.

[0176] The processor is configured to execute the technical solution of the above-described method for identifying unsafe behaviors of drilling workers by executing executable instructions. Its implementation principle and technical effect are similar, and will not be described again here.

[0177] This application also provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements the technical solution of the above-described method for identifying unsafe behaviors of drilling workers. Its implementation principle and technical effects are similar, and will not be repeated here.

[0178] In one possible implementation, a computer-readable medium may include random access memory (RAM), read-only memory (ROM), compact discread-only memory (CD-ROM) or other optical disc storage, disk storage or other magnetic storage devices, or any other medium targeted to carry or to store the required program code in the form of instructions or data structures, and accessible by a computer. Furthermore, any connection is appropriately referred to as a computer-readable medium. For example, if software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave, then coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of medium. As used herein, disks and optical discs include optical discs, laser discs, optical discs, Digital Versatile Discs (DVDs), floppy disks, and Blu-ray discs, where disks typically reproduce data magnetically, while optical discs optically reproduce data using lasers. The above combinations should also be included within the scope of computer-readable media.

[0179] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the technical solution of the above-described method for identifying unsafe behaviors of drilling workers. Its implementation principle and technical effects are similar and will not be repeated here.

[0180] In the specific implementation of the aforementioned terminal device or server, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the embodiments of this application can be directly manifested as being executed by a hardware processor, or executed by a combination of hardware and software modules within the processor.

[0181] Those skilled in the art will understand that all or part of the steps in any of the above method embodiments can be implemented by hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium, and when the program is executed, all or part of the steps in the above method embodiments are performed.

[0182] If the technical solution of this application is implemented in software form and sold or used as a product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the technical solution of this application can be embodied in the form of a software product, which is stored in a storage medium and includes a computer program or several instructions. This computer software product causes a computer device (which may be a personal computer, server, network device, or similar electronic device) to execute all or part of the steps of the method described in the embodiments of this application.

[0183] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for identifying unsafe behaviors of drilling workers, characterized in that, Applied to drilling areas, the method includes: Eye movement data of the operator's gaze at each area in the drilling operation scenario is acquired from an eye movement data sensor. The operation scenario includes multiple sensitive and non-sensitive areas. The eye movement data is then filtered to obtain filtered eye movement data, which includes the total number of fixation points, the time of a single eye movement data acquisition, the average fixation duration, the time of the first fixation on a sensitive area, the time of the second fixation on a sensitive area, the number of fixation points in the sensitive area, and the fixation time in the sensitive area. The filtered eye movement data is processed to obtain the eye movement feature parameters corresponding to the filtered eye movement data. The eye movement feature parameters include: sensitive area fixation point parameters, sensitive area fixation duration parameters, time parameters for the first fixation on the sensitive area, time parameters for the second fixation on the sensitive area, and average fixation duration parameters. The eye-tracking feature parameters are input into the recognition model to obtain multiple classification values, wherein the recognition model includes multiple sub-classification models, and each sub-classification model outputs a classification value; The operator's work behavior is determined based on the classification values ​​of the multiple sub-classification models, and a warning is issued to the operator when the work behavior falls into a preset category.

2. The method according to claim 1, characterized in that, The process of processing the filtered eye movement data to obtain the eye movement feature parameters corresponding to the filtered eye movement data includes: The sensitive area fixation point parameters are obtained by the ratio of the sum of the number of fixation points in each sensitive area within the work scenario to the total number of fixation points in the work scenario. The fixation duration parameter of the sensitive area is obtained by the ratio of the sum of the fixation duration of each sensitive area in the work scenario to the duration of a single eye movement data acquisition. The time parameter for the first fixation on the sensitive area is obtained by the ratio of the minimum value of the first fixation time of each sensitive area in the work scenario to the duration of a single eye movement data acquisition. The time parameter for the second fixation on the sensitive area is obtained by comparing the ratio of the second fixation time to the duration of a single eye movement data acquisition within the described work scenario. The average fixation duration parameter is determined based on the ratio of the sum of fixation durations at N fixation points within the work scenario to N times the duration of a single eye movement data acquisition.

3. The method according to claim 1, characterized in that, Determining the worker's work behavior based on the multiple classification values ​​includes: Among the plurality of classification values, the largest classification value is obtained, wherein the value of the classification value ranges from 0 to 1; The category corresponding to the largest category value is determined as the work behavior of the operator.

4. The method according to claim 1, characterized in that, The method further includes: Acquire various types of training data; If the training data corresponds to the first sub-classification model, then the label value of the first sub-classification model is 1, and the label value of the other second sub-classification models is 0; Multiple sub-classification models in the recognition model are trained to obtain the trained recognition model.

5. The method according to any one of claims 1, 3, or 4, characterized in that, The specific work behavior can be one of the following: normal behavior, violation of regulations, erroneous behavior, or safety hazard.

6. A device for identifying unsafe behaviors of drilling workers, characterized in that, include: The data acquisition module is used to collect eye-tracking data of drilling personnel within the operational scenario. The data processing module is used to filter and process the eye movement data of the operator to obtain eye movement feature parameters; and to filter the eye movement data to obtain filtered eye movement data, which includes the total number of fixation points, the time of a single eye movement data acquisition, the average fixation duration, the time of the first fixation to the sensitive area, the time of the second fixation to the sensitive area, the number of fixation points in the sensitive area, and the fixation time in the sensitive area. The analysis module is used to input the eye movement feature parameters into the recognition model and obtain the results output by the recognition model. The recognition model includes multiple sub-classification models, each of which outputs a classification value. The results output by the recognition model include the classification values ​​output by each of the multiple sub-classification models. The alert module is used to determine the operator's work behavior based on the classification values ​​output by the multiple sub-classification models, and to issue a reminder to the operator when the work behavior is a preset category.

7. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the method of any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, It stores a computer program, which is executed by a processor to implement the method as described in any one of claims 1-5.

9. A computer program product, characterized in that, Includes a computer program, which is executed by a processor to implement the method as described in any one of claims 1-5.

Images