A motion capture method for security guard skill examination

Abstract

This invention relates to the field of motion capture technology, specifically disclosing a motion capture method for assessing safety officer skills, comprising the following steps: S1: Constructing a coordinate system and a three-dimensional model, setting capture points in the three-dimensional model, obtaining the coordinates of the capture points when they are at a detection node, calculating the average of the horizontal, vertical, and triangular coordinates, and obtaining the initial coordinates; S2: Calculating the completion time, obtaining the motion curve of the capture point, and setting detection sub-nodes; S3: Obtaining the detection sub-nodes and generating capture vectors, calculating the cosine similarity and the average of the cosine similarity; S4: Calculating the average action score and time score of the safety officer, calculating the comprehensive score of the safety officer, and quantifying the upper limit of the number of vehicles a safety officer can be responsible for based on the comprehensive score. This invention comprehensively judges the skill level by capturing the standardization of the safety officer's actions and the duration of the actions, thereby quantifying the upper limit of the number of vehicles a safety officer can safely be responsible for with data, thus optimizing labor costs and realizing the feasibility of the business model.

Description

Technical Field

[0001] This invention relates to the field of motion capture technology, and more specifically to a motion capture method for assessing the skills of safety officers. Background Technology

[0002] Safety operators are an important safeguard for ensuring vehicle safety. Among them, remote safety operators are a new profession that has emerged with the development of autonomous driving technology. The core role of remote safety operators is no longer to sit in the cockpit and directly control the vehicle, but to act as a remote, centralized controller to ensure that autonomous vehicles can operate safely and efficiently when they encounter complex situations that they cannot handle.

[0003] The business model based on remote control requires an efficient management system where one safety officer is responsible for multiple vehicles simultaneously. The success of this model hinges directly on the safety officer's skill level. If the officer's skills are inadequate, their accuracy in remote vehicle control, risk assessment, and emergency response capabilities will fail to meet practical needs. In such a scenario, to mitigate the extremely high risks of vehicle loss of control and accidents, the company will be hesitant to assign the same safety officer to manage too many vehicles. This leads to low management efficiency for a single safety officer, hinders effective optimization of labor costs, and ultimately undermines the core premise of "one safety officer managing multiple vehicles," severely restricting the feasibility of the entire remote control business model. Summary of the Invention

[0004] The purpose of this invention is to provide a motion capture method for safety officer skills assessment, thereby solving the above-mentioned technical problems.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A motion capture method for safety officer skills assessment includes the following steps:

[0007] S1: Construct a spatial rectangular coordinate system and a 3D model corresponding to the safety officer. Pre-set several capture points in the 3D model, set detection nodes with an interval of T between adjacent detection nodes, where T is the preset detection cycle.

[0008] When the device is at a detection node, obtain the coordinates P(x, y, z) of the capture point and calculate the mean of the x-coordinate. I represents the number of detection nodes, x i This represents the x-coordinate of the capture point when it is at the i-th detection node. Repeat the above operation to obtain the mean y-coordinate. ave and the mean of the vertical coordinates z ave Generate initial coordinates P0 = (x ave y ave , z ave );

[0009] S2: Send a skills assessment command, whereby the skills assessment refers to a command used to test the safety officer's skills. Record the time t when this skills assessment command is sent. start The time t for the safety officer to complete this skills assessment instruction end Calculate the completion time t = t end -t start ;

[0010] Obtain the motion curve of the capture point within the time interval (t). start , t end Within the range, N detection child nodes are set up, and the time interval between the detection child nodes is equal, where N is the preset number of nodes;

[0011] S3: Obtain the coordinates P of the capture point when it is in the nth detection child node. n (X n Y n Z n Generate capture vector Calculate cosine similarity in, This represents the reference vector corresponding to the preset nth detection child node;

[0012] Generate vectors according to the above method. And calculate the cosine similarity S 0，1 Calculate the mean S of the cosine similarity. ave ;

[0013] S4: Calculate the average score of the safety officer's actions Among them, S ave_j represents the mean cosine similarity corresponding to the j-th capture point, and J represents the total number of preset capture points;

[0014] Calculate the time score F2 = λ(t) sta -t) 2 +M, where λ represents the preset time coefficient and λ < 0, t sta M represents the preset reference time, M represents the preset complete time score, and the safety officer's comprehensive score F = γ1×F1 + γ2×F2 is calculated, where γ1 and γ2 represent the preset first and second weight coefficients, γ1 > γ2 > 0. Based on the comprehensive score F, the upper limit of the number of vehicles a safety officer is responsible for is quantified.

[0015] As a further aspect of the present invention: in step S4, the method for quantifying the upper limit of the number of vehicles a safety officer can be responsible for based on a comprehensive score F includes:

[0016] If the overall score F≤0, the safety officer's skill level is low and no assignment will be made;

[0017] If 0 < F < 1, the safety officer's skill level is passable, and the safety officer is assigned to be responsible for one vehicle;

[0018] If F > 1, then the safety officer is responsible. Vehicles, of which Round up log2F.

[0019] As a further aspect of the present invention: a maximum number R of vehicles is preset, when... At that time, the number of vehicles under the responsibility of the safety officer is equal to R.

[0020] As a further aspect of the present invention: in step S1, the standard deviation s of the abscissa is calculated. x And set the horizontal axis fluctuation range [x ave -2s x x ave +2s x ], placing x within the horizontal axis fluctuation range [x ave -2s x x ave +2s x Detection nodes other than those specified are removed and will not participate in subsequent calculations;

[0021] Repeat the above steps to complete the check of the vertical and ordinate axes.

[0022] As a further aspect of the present invention: in step S2, an execution time threshold T is preset. max When the completion time t≥T max If the test fails, all subsequent operations will be stopped, and a message will be displayed indicating that the test has failed and no vehicle will be assigned.

[0023] As a further aspect of the present invention: in step S3, during the time interval (t) start , t end When setting N detection child nodes, ensure that the value of N is greater than or equal to N0. sta , where N sta This represents the preset number of detection child nodes.

[0024] As a further aspect of the present invention: in step S3, if the mean value of the cosine similarity S ave If the value is less than 0, the corresponding capture point is recorded as an anomaly. The number of anomalies YC is obtained, and the anomaly percentage B = YC / J is calculated. When the anomaly percentage B ≥ 20%, subsequent operations are stopped.

[0025] As a further aspect of the present invention: in step S4, K skill assessment instructions are sent, and the comprehensive scores corresponding to the three skill assessment instructions are obtained. The average of the K comprehensive scores is calculated as the value of F, where K represents the preset number of assessments.

[0026] The beneficial effects of this invention are as follows: First, the purpose of establishing a spatial rectangular coordinate system is to better collect and describe the safety officer's actions. At the same time, using the same coordinate system is conducive to the standardized analysis of the safety officer's actions. Then, the purpose of setting capture points is to simplify the actions, thereby facilitating subsequent analysis. By setting capture points, the safety officer's action trajectory can be further obtained. Furthermore, the advantages of constructing a three-dimensional model can also be seen at this time. Setting capture points at the joints of the human limbs can better capture the actions and simplify the analysis difficulty.

[0027] The purpose of dividing the detection nodes is to understand the coordinates of each capture point when the safety officer does not react, so as to facilitate the subsequent sending of skills assessment instructions and avoid the problem of issuing instructions too quickly. The purpose of calculating the average of the horizontal, vertical and vertical coordinates is to make the calculation results more accurate.

[0028] Then, the time taken for the safety officer to execute the skills assessment instructions is recorded, starting from the time the instructions are sent. This is because the safety officer's reaction speed is also an important reference value; the faster the safety officer's reaction time, the more time is available for handling emergencies. The motion curve of the capture point is obtained, and the motion curve is divided into multiple shorter curves based on the detection sub-nodes. This serves two purposes: first, it simplifies the comparison and analysis process; second, dividing the curves into shorter ones also helps to standardize the safety officer's actions and reduces the difficulty of quantification.

[0029] The main method for quantifying safety officer actions is to calculate the cosine similarity of adjacent vectors. It should be noted that the reference vectors used for comparison are sourced from the assessment system's database. The reference vectors are obtained by operating on the above method based on the personnel in the reference video. If the reference video is changed, the corresponding reference vectors need to be changed accordingly. This invention does not impose too many requirements on the reference vectors; everything is based on the actual situation.

[0030] The scheme calculates the cosine similarity S. 0，1The purpose is to fill in the gaps between the safety officer's ready state and the first detection sub-node, thus making the calculation results more accurate. Then, the accuracy of the safety officer's actions is analyzed, calculating the average cosine similarity of each capture point to obtain the average action score—a score judged based on action standardization. Next, a time-based assessment is taken. As the formula shows, faster isn't always better; it needs to be within a reasonable range. Too short a time indicates a more aggressive operation, leading to a poor passenger experience, while too long a time increases the probability of accidents. Therefore, a comprehensive analysis is needed. Finally, a comprehensive score is obtained by further analyzing both action standardization and time. Based on this comprehensive score, the upper limit of a safety officer's safe responsibility for a given number of vehicles is quantified. This quantifies the safety officer's skill level, determining the upper limit of vehicles a safety officer can safely manage, thereby optimizing labor costs and ensuring the feasibility of the business model. Attached Figure Description

[0031] The invention will now be further described with reference to the accompanying drawings.

[0032] Figure 1 This is a flowchart illustrating a motion capture method for assessing the skills of safety officers according to the present invention. Detailed Implementation

[0033] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] Please see Figure 1 As shown, the present invention is a motion capture method for safety officer skills assessment, comprising the following steps:

[0035] S1: Construct a spatial rectangular coordinate system and a 3D model corresponding to the safety officer. Pre-set several capture points in the 3D model, set detection nodes with an interval of T between adjacent detection nodes, where T is the preset detection cycle.

[0036] When the device is at a detection node, obtain the coordinates P(x, y, z) of the capture point and calculate the mean of the x-coordinate. I represents the number of detection nodes, x i This represents the x-coordinate of the capture point when it is at the i-th detection node. Repeat the above operation to obtain the mean y-coordinate. ave and the mean of the vertical coordinates z ave Generate initial coordinates P0 = (x ave y ave, z ave );

[0037] S2: Send a skills assessment command, whereby the skills assessment refers to a command used to test the safety officer's skills. Record the time t when this skills assessment command is sent. start The time t for the safety officer to complete this skills assessment instruction end Calculate the completion time t = t end -t start ;

[0038] Obtain the motion curve of the capture point within the time interval (t). start , t end Within the range, N detection child nodes are set up, and the time interval between the detection child nodes is equal, where N is the preset number of nodes;

[0039] S3: Obtain the coordinates P of the capture point when it is in the nth detection child node. n (X n Y n Z n Generate capture vector Calculate cosine similarity in, This represents the reference vector corresponding to the preset nth detection child node;

[0040] Generate vectors according to the above method. And calculate the cosine similarity S 0，1 Calculate the mean S of the cosine similarity. ave ;

[0041] S4: Calculate the average score of the safety officer's actions Among them, S ave_j represents the mean cosine similarity corresponding to the j-th capture point, and J represents the total number of preset capture points;

[0042] Calculate the time score F2 = λ(t) sta -t) 2 +M, where λ represents the preset time coefficient and λ < 0, t sta M represents the preset reference time, M represents the preset complete time score, and the safety officer's comprehensive score F = γ1×F1 + γ2×F2 is calculated, where γ1 and γ2 represent the preset first and second weight coefficients, γ1 > γ2 > 0. Based on the comprehensive score F, the upper limit of the number of vehicles a safety officer is responsible for is quantified.

[0043] It should be noted that the primary purpose of establishing a spatial rectangular coordinate system is to better achieve the accurate collection and description of safety officer movements. Every subtle movement of a safety officer can affect the overall safety situation, and the spatial rectangular coordinate system provides the basic framework for recording these movements. Through this unified coordinate system, the changes in their position in three-dimensional space can be represented with precise numerical values.

[0044] Meanwhile, using a unified coordinate system for standardized analysis of safety officer movements offers significant advantages. In practice, different observers or analysis systems using their own independent reference standards can easily lead to data confusion and inconsistencies, affecting the accuracy and reliability of the analysis results. This standardized environment allows for the study of individual safety officer movements and the analysis of collaborative work among multiple safety officers, resulting in more universal and instructive conclusions.

[0045] After the coordinate system is established, the first step is to set capture points at the joints of the human limbs and simultaneously divide the detection nodes. Among them, the joints of the limbs are the key parts for the execution of human movements and are the core layout area of ​​the capture points; the division of detection nodes is carried out around these joint capture points, forming a coordinate monitoring system that corresponds one-to-one with the capture points, providing accurate positioning basis for subsequent data collection.

[0046] The significance of setting up capture points in this way is mainly reflected in two aspects: First, it can specifically simplify the continuous and complex movement patterns of safety officers. The actual movements of safety officers contain a large number of non-essential minor changes. Too many details will obscure the key movement characteristics. However, focusing on the capture points of the limb joints can directly lock the main stage and core parts of the movement, extract the most representative movement patterns, and remove interference for subsequent in-depth analysis. Second, dividing the detection nodes can realize the pre-positioning function. Its core purpose is to accurately obtain the specific coordinate information of each joint capture point before the safety officer has reacted, providing a key reference for the instruction sending in the skills assessment stage.

[0047] From the perspective of motion capture itself, this setup offers several advantages. Leveraging the strengths of 3D models, it allows for real-time tracking of parameters such as displacement and rotation angles at each joint capture point, presenting the data in an intuitive manner. This significantly improves the accuracy and efficiency of motion capture while substantially reducing the difficulty of subsequent analysis. From the perspective of skills assessment, the initial coordinate information obtained from the detection nodes clearly indicates the safety officer's readiness status, effectively avoiding the problem of blindly and rapidly sending assessment instructions. If instructions are issued too quickly without understanding the readiness status, the safety officer may not have enough time to respond correctly. This setup ensures a reasonable pace of instruction issuance, thereby maintaining the fairness and effectiveness of the assessment.

[0048] Meanwhile, to further improve data quality and the accuracy of analysis results, calculating the mean of the horizontal, vertical, and axial coordinates becomes an essential step. By summarizing multiple relevant data points and calculating the mean of their horizontal, vertical, and axial coordinates, the impact of these random errors can be effectively offset.

[0049] Recording the time spent executing skills assessment instructions is crucial. The reason for choosing to start timing from the exact moment the skills assessment instruction is sent is that a safety officer's reaction speed is not merely a simple time indicator; it is actually one of the core elements measuring their emergency response capabilities. In the complex and ever-changing real-world work environment, unexpected situations can occur at any time, and the safety officer's ability to react quickly often determines the course of the event and the severity of its consequences. When encountering emergencies, the faster a safety officer reacts, the more valuable operational time they gain. This extra time means they have a greater opportunity to take effective countermeasures, such as accurately assessing the situation, rationally allocating resources, and implementing appropriate solutions, thereby greatly improving the efficiency and success rate of crisis resolution.

[0050] To comprehensively and deeply analyze the safety officer's motion characteristics, obtaining the motion curves of the capture points is an indispensable step. However, directly analyzing and processing a long and complex complete motion curve undoubtedly faces many difficulties and challenges. Therefore, scientifically and rationally dividing these motion curves according to the detection sub-nodes has the following two important implications. First, it greatly simplifies the comparison and analysis process. Through the fine division of the detection sub-nodes, each short curve segment has clear start and end boundaries and relatively independent motion characteristics, making the comparison between different segments clearer and more intuitive.

[0051] Secondly, dividing the movement into shorter curves significantly facilitates the accurate assessment of the standardization of safety officer actions and effectively reduces the difficulty of quantification. By breaking down the movement into multiple short-range curves, the actions within each segment are relatively simple and more representative, making it easier to establish a unified and clear quantitative indicator system.

[0052] In the process of quantitatively evaluating safety officer actions, the cosine similarity of adjacent vectors is used as a primary method. Cosine similarity serves as an indicator of the consistency of the directions of two vectors. Specifically, it determines their similarity by calculating the cosine of the angle between two vectors. The closer the value is to 1, the more consistent their directions are, indicating more standardized actions; conversely, the smaller the value, the greater the deviation in actions.

[0053] It's important to note that these reference vectors are all derived from the assessment system's vast database, and their construction process uses the professionals in the reference videos as a benchmark. It's crucial to understand that these reference vectors are not static, unchanging data. When different reference videos are used, the corresponding reference vectors must be updated accordingly, as the presenters' movement styles, rhythms, and details may differ in the new videos.

[0054] This invention explicitly states that it does not impose excessively strict requirements on the specific form, content, or other details of the reference vector. This is because the definition of "ideal" action may vary depending on different application scenarios, industry sectors, and specific work tasks. Everything is based on the actual situation, aiming to fully leverage the universality and practicality of this quantitative method. Whether in high-risk industrial production environments or complex public safety fields, appropriate reference videos can be flexibly selected to generate corresponding reference vectors based on specific work characteristics and safety requirements, thereby achieving effective quantitative assessment of safety officer actions.

[0055] Calculate cosine similarity S 0，1 Aimed at accurately supplementing the safety officer's state information during the transition from the preparation state to the first detection sub-node, this effectively fills potential data gaps. From the perspective of action standardization, an in-depth analysis is conducted, calculating the average cosine similarity of each capture point to arrive at an average action score. Each capture point records the movement of key parts of the safety officer's body, and the cosine similarity quantifies the degree of matching between these actual actions and the ideal template. The averaging process further smooths out individual fluctuations, highlighting the overall trend in action quality. A high score indicates that the safety officer's actions are smooth, precise, and strictly adhere to operating procedures, demonstrating a high degree of professionalism and stability; conversely, a lower score suggests potential issues such as action distortion or incomplete execution, requiring timely correction and improvement.

[0056] Meanwhile, time is also a crucial factor in evaluating safety officer performance. It's important to note that faster speed isn't always better; there's a delicate balance to be struck. A specific formula shows that excessively short operation times often involve overly abrupt actions, potentially causing discomfort or even panic for passengers and creating safety hazards due to haste. Conversely, excessively long response times significantly increase the probability of accidents, as prolonged response cycles may lead to missed opportunities for optimal intervention. Therefore, time must be controlled within a reasonable range, testing the safety officer's responsiveness, as well as their ability to make calm judgments and execute efficiently. It's important to note that in this invention, "responsible for the number of vehicles" indicates that their duties extend beyond monitoring, including remote rescue and remote voice communication. Remote monitoring is only one aspect of their responsibilities; therefore, all these duties are collectively referred to as "responsible."

[0057] Based on the above dual considerations—standardized actions and reasonable timing—a comprehensive evaluation system was constructed. Through weighted allocation and comprehensive analysis of these two aspects, a final comprehensive score was obtained. Based on this comprehensive score, the upper limit of a safety officer's safe responsibility for a vehicle can be scientifically quantified. On the one hand, it provides precise guidance for human resource allocation, helping companies rationally allocate the number of safety officers according to their own business scale and operational characteristics, avoiding waste or shortage of manpower; on the other hand, it ensures the stability and reliability of service quality, achieving economic feasibility and sustainable development of the business model while guaranteeing passenger safety.

[0058] In another preferred embodiment of the present invention, the method for quantifying the upper limit of the number of vehicles a safety officer is responsible for based on a comprehensive score F includes:

[0059] If the overall score F≤0, the safety officer's skill level is low and no assignment will be made;

[0060] If 0 < F < 1, the safety officer's skill level is passable, and the safety officer is assigned to be responsible for one vehicle;

[0061] If F > 1, then the safety officer is responsible. Vehicles, of which Round up log2F.

[0062] It is worth noting the management strategy of quantifying safety officers' work capabilities based on a comprehensive score F and dynamically assigning tasks. Its core significance lies in establishing clear quantitative standards to precisely link safety officers' skill levels with their actual workload, thereby achieving efficient utilization of human resources and controllable operational safety. When the comprehensive score F ≤ 0, the safety officer is deemed insufficiently skilled and unable to perform any duties, therefore no tasks are assigned. If 0 < F < 1, it indicates that the officer is at a passing level but has limited ability, only capable of managing a single vehicle to ensure basic safety. When F > 1, the number of vehicles the officer can manage simultaneously is determined by rounding up log2F, acknowledging the multitasking capabilities of highly skilled individuals while avoiding increased risk of errors due to overload.

[0063] In a preferred embodiment, a maximum number of responsible vehicles R is preset, when... At that time, the number of vehicles under the responsibility of the safety officer is equal to R.

[0064] Understandably, the preset maximum number of vehicles a safety officer can be responsible for simultaneously, R, is used as the upper limit. When the calculated theoretical value exceeds R, the limit is forcibly set to R. The core significance of this mechanism lies in balancing the dual demands of efficiency and safety through rigid constraints. This avoids cost waste caused by inefficient redundancy and eliminates potential accident hazards caused by overloaded operation, ultimately building a sustainable management system that balances economic benefits and operational safety.

[0065] In another preferred embodiment of the present invention, the standard deviation s of the abscissa is calculated. x And set the horizontal axis fluctuation range [x ave -2s x x ave +2s x ], placing x within the horizontal axis fluctuation range [x ave -2s x x ave +2s x Detection nodes other than those specified are removed and will not participate in subsequent calculations;

[0066] Repeat the above steps to complete the check of the vertical and ordinate axes.

[0067] It is important to note that this is a data cleaning and outlier filtering mechanism based on statistical principles. Its core significance lies in using the characteristics of normal distribution to identify and exclude extreme measurement points that deviate from the normal motion trajectory. These outliers may originate from equipment errors, accidental interference, or atypical motion deformations; if left untreated, they will distort the overall analysis results. Through standardized screening using three-axis linkage, effective data consistent with the main trend is retained to ensure the accuracy of motion capture, while the impact of noise interference on subsequent quantitative assessments is eliminated. This allows key indicators such as cosine similarity calculation and duration statistics to more accurately reflect the safety officer's operational standardization and stability, thereby improving the robustness and decision reliability of the motion analysis model.

[0068] In another preferred embodiment of the present invention, a preset execution time threshold T is set. max When the completion time t≥T max If the test fails, all subsequent operations will be stopped, and a message will be displayed indicating that the test has failed and no vehicle will be assigned.

[0069] It should be noted that the preset execution time threshold T max As a rigid cutoff standard, the actual time t taken for the safety officer to complete the skills assessment must be greater than or equal to T. max The core significance of immediately terminating the process and determining failure lies in reinforcing the bottom-line thinking of safe operation through rigid time constraints. It prevents the distortion of actions, omission of details, or accumulation of potential risks due to excessive pursuit of speed, and ensures that all participants must complete standardized operating procedures within a specified reasonable time, thereby drawing an uncompromising safety boundary between efficiency and quality.

[0070] In another preferred embodiment of the present invention, in the time interval (t) start , t end When setting N detection child nodes, ensure that the value of N is greater than or equal to N0. sta , where N sta This represents the preset number of detection child nodes.

[0071] Understandably, increasing sampling density ensures the statistical validity of motion trajectory analysis and the granularity of motion decomposition. When N≥3, it can not only form basic time series data to support trend judgment, but also accurately capture the dynamic characteristics of the start, intermediate transition and end stages of the motion through multi-point positioning, avoiding linearization misjudgment or keyframe loss due to too few nodes.

[0072] In another preferred embodiment of the present invention, if the mean of the cosine similarity S ave If the value is less than 0, the corresponding capture point is recorded as an anomaly. The number of anomalies YC is obtained, and the anomaly percentage B = YC / J is calculated. When the anomaly percentage B ≥ 20%, subsequent operations are stopped.

[0073] In another preferred embodiment of the present invention, K skill assessment instructions are sent, and the comprehensive scores corresponding to the three skill assessment instructions are obtained. The average of the K comprehensive scores is calculated as the value of F, where K represents the preset number of assessments.

[0074] It is worth noting that repeated experiments eliminate the impact of random errors on the evaluation results, constructing a more stable and representative competence measurement system. Multiple independent tests can capture the safety officer's adaptability, concentration, and operational consistency in different situations, avoiding scoring bias caused by accidental factors in a single assessment. The mean calculation is based on the law of large numbers to smooth out extreme value fluctuations, making the F-score closer to the true level, reducing the interference of luck in the evaluation, and strengthening the consideration of consistent performance.

[0075] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A motion capture method for safety officer skills assessment, characterized in that, Includes the following steps: S1: Construct a spatial rectangular coordinate system and a 3D model corresponding to the safety officer. Pre-set several capture points in the 3D model, set detection nodes with an interval of T between adjacent detection nodes, where T is the preset detection cycle. When the device is at a detection node, obtain the coordinates P(x, y, z) of the capture point and calculate the mean of the x-coordinate. I represents the number of detection nodes, x i This represents the x-coordinate of the capture point when it is at the i-th detection node. Repeat the above operation to obtain the mean y-coordinate. ave and the mean of the vertical coordinates z ave Generate initial coordinates P0 = (x ave y ave , z ave ); S2: Send a skills assessment command, whereby the skills assessment refers to a command used to test the safety officer's skills. Record the time t when this skills assessment command is sent. start The time t for the safety officer to complete this skills assessment instruction end Calculate the completion time t = t end -t start ; Obtain the motion curve of the capture point within the time interval (t). start , t end Within the range, N detection child nodes are set up, and the time interval between the detection child nodes is equal, where N is the preset number of nodes; S3: Obtain the coordinates P of the capture point when it is in the nth detection child node. n (X n Y n Z n Generate capture vector Calculate cosine similarity in, This represents the reference vector corresponding to the preset nth detection child node; Generate vectors according to the above method. And calculate the cosine similarity S 0，1 Calculate the mean S of the cosine similarity. ave ; S4: Calculate the average score of the safety officer's actions Among them, S ave_j represents the mean cosine similarity corresponding to the j-th capture point, and J represents the total number of preset capture points; Calculate the time score F2 = λ(t) sta -t) 2 +M, where λ represents the preset time coefficient and λ < 0, t sta M represents the preset reference time, M represents the preset complete time score, and the safety officer's comprehensive score F = γ1×F1 + γ2×F2 is calculated, where γ1 and γ2 represent the preset first and second weight coefficients, γ1 > γ2 > 0. Based on the comprehensive score F, the upper limit of the number of vehicles a safety officer is responsible for is quantified.

2. The motion capture method for safety officer skills assessment according to claim 1, characterized in that, In step S4, the method for quantifying the upper limit of the number of vehicles a safety officer can be responsible for based on the comprehensive score F includes: If the overall score F≤0, the safety officer's skill level is low and no assignment will be made; If 0 < F < 1, the safety officer's skill level is passable, and the safety officer is assigned to be responsible for one vehicle; If F > 1, then the safety officer is responsible. Vehicles, of which Round up log2F.

3. The motion capture method for safety officer skills assessment according to claim 2, characterized in that, The maximum number of vehicles to be responsible is preset to R. At that time, the number of vehicles under the responsibility of the safety officer is equal to R.

4. The motion capture method for safety officer skills assessment according to claim 1, characterized in that, In step S1, the standard deviation s of the abscissa is calculated. x And set the horizontal axis fluctuation range [x ave -2s x x ave +2s x ], placing x within the horizontal axis fluctuation range [x ave -2s x x ave +2s x Detection nodes other than those specified are removed and will not participate in subsequent calculations; Repeat the above steps to complete the check of the vertical and ordinate axes.

5. The motion capture method for safety officer skills assessment according to claim 1, characterized in that, In step S2, a preset execution time threshold T is set. max When the completion time t≥T max If the test fails, all subsequent operations will be stopped, and a message will be displayed indicating that the test has failed and no vehicle will be assigned.

6. The motion capture method for safety officer skills assessment according to claim 1, characterized in that, In step S3, during the time interval (t) start , t end When setting N detection child nodes, ensure that the value of N is greater than or equal to N0. sta , where N sta This represents the preset number of detection child nodes.

7. The motion capture method for safety officer skills assessment according to claim 1, characterized in that, In step S3, if the mean of the cosine similarity S ave If the value is less than 0, the corresponding capture point is recorded as an anomaly. The number of anomalies YC is obtained, and the anomaly percentage B = YC / J is calculated. When the anomaly percentage B ≥ 20%, subsequent operations are stopped.

8. The motion capture method for safety officer skills assessment according to claim 1, characterized in that, In step S4, K skill assessment instructions are sent, and the comprehensive scores corresponding to the three skill assessment instructions are obtained. The average of the K comprehensive scores is calculated as the value of F, where K represents the preset number of assessments.

Images