Oil depot operation behavior intelligent identification method and system

Abstract

The invention discloses an oil depot operation behavior intelligent identification method and system, and relates to the field of oil depot safety monitoring, and the method comprises the steps: obtaining motion image data and operation environment parameter data of an oil depot operator, and the operation environment parameter data comprises temperature, humidity and oil gas concentration; action mode recognition is executed based on the action image data, an electrostatic action risk index is evaluated based on the action of the operator, and the electrostatic action risk index is used for measuring the risk index of static electricity possibly generated by the action of the operator; evaluating an electrostatic environment risk index based on the operation environment parameter data, wherein the electrostatic environment risk index is inversely proportional to humidity, temperature and oil gas concentration; and comprehensively weighting the electrostatic action risk index and the electrostatic environment risk index to obtain a total electrostatic risk index, and triggering an alarm when the total electrostatic risk index exceeds a threshold value. The problem that it is difficult to evaluate and early warn static accumulation risks in oil depot operation in time is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of oil depot safety monitoring, and more particularly, to an oil depot operation behavior intelligent identification method and system. BACKGROUND

[0002] Oil depot operation belongs to a typical high-risk industrial scene. During the processes of oil unloading, oil loading, pipeline connection, and valve operation, friction actions and environmental risk factors inevitably exist. Since the oil gas concentration in the storage environment of finished oil or crude oil may be within the explosive limit range, a spark can cause a serious accident, and static electricity is one of the main sources of sparks. Static electricity may come from the friction between the clothes, gloves, etc. of the operator and the metal surface, or from the friction between the hose, container, and other equipment.

[0003] Under winter or dry conditions, air humidity decreases, and static electricity accumulation is more likely to occur. If the operator frequently performs actions such as dragging and carrying, static electricity charges are continuously accumulated, and may discharge instantaneously upon contact with metal parts, thereby igniting the surrounding oil gas. At the same time, the oil depot operation area is usually divided into different regions such as unloading area, pump area, and fire-prohibited area, and environmental factors such as temperature, humidity, and oil gas concentration also affect the size of the static electricity risk. Therefore, how to jointly identify and quantitatively evaluate the actions and environment during the operation process, timely discover high-risk behaviors that may generate static electricity, and trigger an alarm when the risk exceeds the safety threshold is a problem that needs to be solved in oil depot safety management. SUMMARY

[0004] The technical problem solved by the present application is to provide an oil depot operation behavior intelligent identification method and system to solve the problems mentioned in the background.

[0005] In order to achieve the above-mentioned purpose, the present application adopts the following technical scheme: An oil depot operation behavior intelligent identification method, comprising: obtaining action image data of an oil depot operator and operation environment parameter data, the operation environment parameter data including temperature, humidity, and oil gas concentration; performing action pattern recognition based on the action image data, and evaluating a static electricity action risk index based on the action of the operator, the static electricity action risk index being used to measure the risk index of static electricity that may be generated by the action of the operator; evaluating a static electricity environment risk index based on the operation environment parameter data, the static electricity environment risk index being inversely proportional to humidity, inversely proportional to temperature, and proportional to oil gas concentration; comprehensively weighting the static electricity action risk index and the static electricity environment risk index to obtain a static electricity total risk index, and triggering an alarm when the static electricity total risk index exceeds a threshold.

[0006] Specifically, the method for obtaining the static action risk index comprises: performing action pattern recognition based on the action image data, identifying two friction objects involved in the action and their materials, calculating the speed, duration and contact area of the action, and obtaining the static action risk index based on the following formula: ; wherein M represents a friction object factor, A represents the contact area, V represents the action speed, D represents the action duration, and k is a coefficient; wherein the friction object factor is used to express the electrification ability of mutual friction of two materials.

[0007] Specifically, the acquisition of the friction object factor is based on the experimental calibration results of materials in the oil depot, and specifically comprises: Selecting common material combinations in oil depot operations in a controlled environment, including metal pipes, rubber hoses, chemical fiber protective clothing, insulating gloves, plastic barrels and ground coverings, and performing friction tests on any two materials; In each friction test, the contact area, friction speed and friction time are kept consistent, and the static electricity generated after the friction of the two materials is recorded by a static electricity measuring device; The measured static electricity or voltage is normalized to obtain the corresponding friction object factor, which is used to reflect the electrification ability between two materials.

[0008] Specifically, the static action risk index is a cumulative value of multiple actions, and is calculated using the following formula: ; wherein, represents the static action risk index of the i-th action, represents the time when the i-th action ends, represents the decay coefficient of static electricity natural dissipation, represents the current time, represents the total static action risk index after superposition, represents the number of actions. Specifically, when a grounding action corresponding to a friction object is identified, the static cumulative risk index of the friction object is reset to zero.

[0009] Specifically, the friction objects involved in the action and their materials, as well as the speed, duration and contact area of the action, are determined by the following method:

[0010] Specifically, the friction objects involved in the action and their materials, as well as the speed, duration and contact area of the action, are determined by the following method: ​Based on target detection and image segmentation algorithms, the friction objects of the workers are identified, and the material type of the friction objects is determined by combining the material identification model or the preset material database. The speed and duration of an action are calculated through motion key point detection and trajectory tracking. The contour of the contact area is obtained by image segmentation, and the contact area is calculated by combining depth estimation or multi-view imaging results.

[0011] Specifically, the electrostatic environment risk index is calculated based on the work environment parameter data according to the following formula: ; in, This indicates the electrostatic environmental risk index. Indicates oil and gas concentration. Indicates humidity. Indicates temperature. is a coefficient.

[0012] Specifically, the calculation of the total electrostatic risk index includes: The electrostatic action risk index and the electrostatic environment risk index are normalized to map them to a dimensionless standardized interval. Based on the normalization results, the total electrostatic risk index is calculated using the following formula: ; in, This indicates the overall electrostatic risk index. This represents the normalized electrostatic discharge risk index. This represents the normalized electrostatic environmental risk index. and This is a weighting factor used to balance the contributions of action factors and environmental factors to the total risk.

[0013] Specifically, the normalization method is Min-Max normalization or Z-score standardization.

[0014] This invention also discloses an intelligent recognition system for oil depot operation behavior, comprising: The motion and environment data acquisition module is used to acquire motion image data of oil depot workers and work environment parameter data, including temperature, humidity and oil and gas concentration. The motion pattern recognition and risk assessment module is used to perform motion pattern recognition based on the motion image data and assess the electrostatic motion risk index based on the worker's motion. The electrostatic motion risk index is used to measure the risk index of electrostatic discharge that may be generated by the worker's motion. The environmental risk assessment module is used to determine the electrostatic environmental risk index based on the work environment parameter data. The electrostatic environmental risk index is inversely proportional to humidity, inversely proportional to temperature, and directly proportional to oil and gas concentration. The risk integration and alarm module is used to comprehensively weight the electrostatic action risk index and the electrostatic environment risk index to obtain the total electrostatic risk index, and to trigger an alarm when the total electrostatic risk index exceeds a threshold.

[0015] The advantages of this invention compared to existing technologies lie in its novel intelligent recognition scheme for operational behaviors. This scheme integrates action and environmental factors in oil depot operations to achieve dynamic quantitative assessment of electrostatic risks. By introducing the recognition and analysis of action image data, it directly correlates worker actions with potential friction conditions and calculates an electrostatic action risk index, thus solving the problem that traditional monitoring methods cannot accurately quantify the relationship between actions and electrostatic risks. Combined with operational environment parameter data, an electrostatic environmental risk index is established, clearly demonstrating its inverse relationship with humidity, inverse relationship with temperature, and direct relationship with oil and gas concentration, making the risk assessment results closer to physical laws. Finally, by weighted fusion of action risk and environmental risk, a total electrostatic risk index is obtained. When the risk exceeds a threshold, an alarm is automatically triggered, effectively improving the timeliness and accuracy of risk detection. This invention further provides a method for calculating action risk based on the identification of friction objects and materials. It utilizes triboelectric sequence lists or experimental calibration methods to obtain friction object factors, making the action risk index more interpretable. By introducing quantitative parameters such as action speed, duration, and contact area, the electrostatic risk assessment not only depends on the type of action but also reflects the intensity of the action. On the other hand, the cumulative risk model and time decay factor proposed in this invention can simulate the gradual accumulation and natural dissipation of electrostatic charge, more closely resembling actual physical phenomena. Simultaneously, when a grounding action is detected, the risk can be reduced to zero, thus demonstrating the effect of electrostatic discharge. For the quantification of environmental risk, this invention employs formulaic modeling, ensuring the consistency and repeatability of risk calculations. This invention not only achieves integrated identification of actions and the environment but also indexes, dynamically adjusts, and controls electrostatic risk, significantly improving the safety level of oil depot operations. Attached Figure Description

[0016] Figure 1 This is an overall schematic diagram of the invention; Figure 2 This is a schematic diagram of the action pattern recognition and risk assessment module of the present invention; Figure 3 This is a schematic diagram of the environmental risk assessment module of the present invention. Detailed Implementation

[0017] The specific embodiments of the present invention will now be described with reference to the accompanying drawings.

[0018] like Figure 1 As shown, the method of the present invention includes the following steps: Acquire motion image data of oil depot workers and work environment parameter data, including temperature, humidity, and oil and gas concentration; Action pattern recognition is performed based on the motion image data, and an electrostatic action risk index is evaluated based on the worker's actions. The electrostatic action risk index is used to measure the risk index of electrostatic discharge that may be generated by the worker's actions. The electrostatic environment risk index is evaluated based on the aforementioned work environment parameter data. The electrostatic environment risk index is inversely proportional to humidity and temperature, and directly proportional to oil and gas concentration. The electrostatic activity risk index and the electrostatic environment risk index are combined to obtain the total electrostatic risk index, and an alarm is triggered when the total electrostatic risk index exceeds a threshold.

[0019] In specific embodiments, motion image data is collected by high-resolution cameras deployed in key areas of the oil depot. Industrial-grade cameras with a resolution of at least 1920x1080 pixels and a frame rate of at least 30 frames per second can be used to capture subtle movements of workers, such as arm swings and tool operations. The cameras cover the main operating areas of the oil depot, such as the oil tank loading and unloading area, pipeline connection area, and oil storage area, and are equipped with wide-angle lenses to ensure a complete field of view. To cope with complex lighting conditions, such as nighttime or strong light reflection, some embodiments may also use cameras supporting infrared imaging or high dynamic range, and these cameras are regularly calibrated to ensure image clarity. Data transmission uses low-latency 5G or Wi-Fi networks to ensure real-time performance.

[0020] In a specific embodiment, environmental parameter data, including temperature, humidity, and oil and gas concentration, are collected by dedicated sensors.

[0021] High-precision temperature and humidity sensors, such as the DHT22, can be used to measure temperature from -40°C to 80°C and humidity from 0 to 100%RH, with accuracies of ±0.5°C and ±2%RH, respectively.

[0022] Oil and gas concentrations are measured using volatile organic compound (VOC) sensors, such as the MQ-135, with a measurement range of 10 to 1000 ppm and an accuracy of ±5 ppm. To improve data reliability, in some specific embodiments, the sensors are calibrated regularly, at least once a month, and deployed in key locations within the work area, such as near oil tanks and pipeline interfaces. Sensor data is transmitted to the central processing unit via wireless communication modules, such as ZigBee or LoRa, at a frequency of once per second, reflecting environmental changes in real time.

[0023] For example, in an oil depot loading and unloading scenario, cameras recorded the actions of workers connecting oil pipes at a frequency of 30 frames per second. Simultaneously, temperature and humidity sensors recorded an ambient temperature of 26 degrees Celsius and a humidity of 55% RH, while an oil and gas concentration sensor detected a concentration of 60 ppm. This data provides accurate input for subsequent risk assessments.

[0024] In a further embodiment, such as Figure 2 As shown, action pattern recognition is based on acquired image data and implemented using advanced computer vision technology. In a specific embodiment, a target detection algorithm, such as YOLOv5, is used to identify workers and their interactions with friction objects, such as friction between clothing. YOLOv5 is suitable for real-time scenarios due to its high speed and high accuracy; the model input is an image captured by a camera, and the output is a target bounding box.

[0025] Once the target is detected, an image segmentation algorithm, such as Mask R-CNN, is used to extract the precise contour of the friction object in order to calculate the contact area. The segmentation algorithm needs to be trained to recognize common objects in oil depots, such as metal pipes, plastic buckets, and protective clothing. The training data includes a large number of labeled images covering different angles and lighting conditions.

[0026] In a further embodiment, the material of the friction object is identified by combining a material database and a deep learning model. The material database stores the visual features and physical properties of common materials in oil depots, such as the metallic luster of stainless steel, the flexible surface of polyethylene, and the texture of cotton protective clothing. The material recognition model uses a convolutional neural network, such as the ResNet-50 architecture, with the segmented target region image as input and the material category probability as output. The training dataset contains a large number of images covering common materials in oil depots. The training process uses the Adam optimizer with a learning rate of 0.001, a batch size of 32, and 60 epochs, with cross-entropy loss as the loss function. The model needs to achieve a classification accuracy of over 97% to ensure the reliability of material recognition. For example, if the model identifies a worker wearing synthetic fiber protective clothing contacting a metal pipe, the probabilities of synthetic fiber and stainless steel are 0.98 and 0.95, respectively.

[0027] Action speed and duration are calculated through keypoint detection and trajectory tracking. The OpenPose algorithm is used to detect keypoints such as the worker's arms and legs, recording their positional changes across consecutive frames. Speed ​​is calculated by inter-frame bit removal at time intervals, in meters per second. Duration is calculated using the timestamps of the start and end frames of the action, in seconds. For example, the speed of a worker swinging a wrench is 0.6 meters per second, lasting 2.5 seconds. Contact area is calculated by combining the contour obtained from image segmentation with depth estimation. The MiDaS monocular depth estimation model is used, outputting pixel area, which is then converted to the actual area in square meters using the camera's focal length and calibration parameters. For example, the segmentation result shows a contact area of ​​1200 pixels, corresponding to 0.012 square meters after calibration.

[0028] The electrostatic discharge (ESD) risk index is used to quantify the ESD risk generated by workers' actions. The calculation formula is as follows: ; Where R is the electrostatic discharge risk index, k is the adjustment coefficient, M is the friction object factor, A is the contact area in square meters, V is the action speed in meters per second, and D is the action duration in seconds. This formula is based on the physical principle of static electricity generation: the amount of charge is directly proportional to the charging ability of the friction material, the contact area, the action speed, and the duration. Considering that R needs to be normalized later, the value of the coefficient k can be freely chosen in the early stages, for example, choosing 1.

[0029] In one implementation, the friction factor M is obtained based on experimental calibration results of common materials used in oil depot operations. To ensure the repeatability and representativeness of the experimental results, the combination of materials most likely to experience friction during oil depot operations is first selected under controlled experimental conditions. These materials include metal pipes, rubber hoses, synthetic fiber protective clothing, insulating gloves, plastic buckets, and ground coverings. These materials cover the main friction interfaces between personnel, equipment, containers, and pipelines during oil depot operations and are typical sources of electrostatic accumulation risk.

[0030] During the experiment, the two test materials were subjected to contact and friction under constant pressure, temperature, and humidity conditions. The relative humidity of the laboratory environment was controlled between 40% and 60%, and the temperature between 20°C and 25°C, to closely approximate the common operating environment of oil depots. The same contact area and friction speed were maintained for each test, for example, a contact area of ​​100 cm², a friction speed of 0.2 m / s, and a friction time of 10 s, to avoid statistical deviations between different tests.

[0031] During the friction process, the electrostatic charge (in coulombs) or electrostatic voltage (in volts) generated on the friction surfaces per unit time is recorded using an electrostatic charge measuring device or an electrostatic voltage sensor. If the charge measurement method is used, a high-impedance charge amplifier can be connected to the two material surfaces to monitor the accumulated charge in real time; if the voltage method is used, the potential difference between the two surfaces can be measured immediately after friction. Each material combination requires at least three repeated tests, and the average value is taken to reduce random errors.

[0032] By comparing the average electrostatic charge or voltage generated by different material combinations, their relative electrostatic charging capabilities can be obtained. Material combinations with stronger electrostatic charging capabilities are more prone to electrostatic accumulation in oil depot operations. The measured average charge or voltage values ​​are normalized to map the results to a standardized range of 0–1, defined as the friction object factor M. For the material combination generating the highest charge in the experiment, its friction object factor is set to 1, the lowest to 0, and the rest distributed proportionally, thus achieving normalization. This friction object factor can be directly used to calculate the electrostatic action risk index, reflecting the degree of electrostatic charging differences between different friction objects.

[0033] In some embodiments, to improve the long-term adaptability of the system, experimental calibration can be performed periodically to update the database of friction object factors. When the surface condition of the material changes due to aging, contamination, or oil coverage, the electrostatic characteristics are recalibrated through new experiments, thereby maintaining the matching accuracy of the identification system to the actual oil depot environment. This experimental calibration and database update mechanism enables the electrostatic risk identification model to have high reliability and engineering usability.

[0034] In a further embodiment, to simulate the cumulative effect of static charge, the system calculates a cumulative risk index for multiple actions. The subscript sum indicates cumulative totals, and the formula is: ; in, For the first Electrostatic discharge risk index for this action. This is the end time of the action. Let λ be the current time, and λ be the decay coefficient for the natural dissipation of static charge. This formula is based on the principle that static charge decays exponentially over time. The value of λ is related to ambient humidity and temperature, and in some embodiments it ranges from 0.01 to 0.12, with a typical value of 0.06. For example, in a high-humidity environment, λ can be taken as 0.08 to reflect faster dissipation.

[0035] When a grounding action is detected, such as a worker's hand touching a grounding stake, the system identifies the grounding behavior using a target detection algorithm and assigns it to the worker's hand. Resetting to zero reflects the complete release of charge.

[0036] like Figure 3 As shown, the electrostatic environmental risk index E reflects the impact of environmental parameters on electrostatic risk, and the calculation formula is: ; Where C is the oil and gas concentration in ppm; H is the humidity in %RH (relative humidity); T is the temperature (preferably Kelvin in some regions where temperatures drop below zero Celsius, otherwise Celsius is acceptable). b is an adjustment coefficient. The formula is designed based on the following principles: higher oil and gas concentrations increase the risk of explosion caused by electrostatic sparks, therefore C is positively correlated with E; high humidity and high temperature promote charge dissipation, therefore H and T are inversely proportional to E. Considering the need for normalization of E later, the coefficient b can be freely chosen initially, for example, set to 1.

[0037] The total electrostatic risk index of this invention The formula for calculating combined action and environmental risks is as follows: ;in, and Normalized electrostatic motion risk index (which can also be the cumulative risk index of multiple motions) and environmental risk index. and These are weighting coefficients. This represents the total electrostatic risk index, with the subscript "to" indicating "total," meaning the sum. In a specific implementation, normalization can be achieved using the Min-Max method, mapping R and E to 0 to 1. X'=(XX min ) / (X max -X min ); where the parameter value X can represent either R or E, max represents the maximum value, and min represents the minimum value.

[0038] When performing Min-Max normalization on the electrostatic action risk index R and the electrostatic environment risk index E, the minimum and maximum values ​​can be derived from historical sample data or set based on theoretical boundaries. Specifically, the minimum value X... min The lower limit of the corresponding risk index can be taken from historical monitoring samples or experimental data, and the maximum value X is taken as the lower limit. maxThe upper limit value can be taken from historical samples or experimental data to ensure that the normalization result matches the actual working conditions. If sufficient historical samples cannot be obtained in a specific implementation environment, boundary conditions can be determined based on theoretical analysis. For example, the electrostatic action risk index is close to zero when the friction object factor is minimum, the contact area approaches zero, and the speed and duration are extremely low, and it is at its maximum when the friction object factor is maximum and the contact area, speed, and duration are all at preset upper limits; the electrostatic environment risk index can be taken as the minimum value under the conditions of highest humidity, highest temperature, and lowest oil and gas concentration, and as the maximum value under the conditions of lowest humidity, lowest temperature, and highest oil and gas concentration.

[0039] α and β satisfy α+β=1, for example, α=0.65 and β=0.35 can be selected, reflecting that the contribution of action factors to risk is slightly higher than that of environmental factors.

[0040] The threshold can be selected from 0.7 to 0.9. If the threshold is exceeded, the system triggers multi-level alarms, such as on-site audible and visual alarms, SMS notifications to management personnel, and warning messages displayed at the control center. In some embodiments, the alarm devices can use high-decibel buzzers and LED warning lights, and SMS messages are sent via a GSM module.

[0041] The present invention also discloses a corresponding system, comprising: The motion and environmental data acquisition module integrates high-definition cameras and sensors. The data is transmitted to a central server via a 5G network. The server is equipped with a high-performance CPU and GPU.

[0042] The action pattern recognition and risk assessment module runs models such as YOLOv5, Mask R-CNN, and OpenPose, and is deployed on an NVIDIA A100 GPU, supporting real-time processing of image streams at 30 frames per second.

[0043] The environmental risk assessment module processes sensor data and runs on an embedded device, such as a Raspberry Pi4, to calculate the E-value.

[0044] The risk integration and alarm module is used to perform normalization and weighted calculations. It runs on the central server and triggers alarms.

[0045] The modules communicate via a RESTful API, with data in JSON format. For example, the action pattern recognition module outputs action parameters in JSON format, the environment module outputs sensor data in JSON format, and the integration module updates every second. .

[0046] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for intelligent recognition of oil depot operational behavior, characterized in that, include: Acquire motion image data of oil depot workers and work environment parameter data, including temperature, humidity, and oil and gas concentration; Action pattern recognition is performed based on the motion image data, and an electrostatic action risk index is evaluated based on the worker's actions. The electrostatic action risk index is used to measure the risk index of electrostatic discharge that may be generated by the worker's actions. The electrostatic environment risk index is evaluated based on the aforementioned work environment parameter data. The electrostatic environment risk index is inversely proportional to humidity and temperature, and directly proportional to oil and gas concentration. The electrostatic activity risk index and the electrostatic environment risk index are combined to obtain the total electrostatic risk index, and an alarm is triggered when the total electrostatic risk index exceeds a threshold.

2. The intelligent recognition method for oil depot operation behavior according to claim 1, characterized in that, The method for obtaining the electrostatic discharge risk index includes: Based on the motion image data, motion pattern recognition is performed, and the two friction objects involved in the motion and their materials are identified. The speed, duration, and contact area of ​​the motion are calculated, and the electrostatic motion risk index is obtained based on the following formula: ； Where M represents the friction object factor, A represents the contact area, V represents the action speed, D represents the action duration, and k is a coefficient; wherein, the friction object factor is used to express the electrostatic ability of two materials rubbing against each other.

3. The intelligent recognition method for oil depot operation behavior according to claim 2, characterized in that, The friction factor was obtained based on the experimental calibration results of materials in the oil depot, specifically including: In a controlled environment, a combination of materials commonly used in oil depot operations was selected, including metal pipes, rubber hoses, synthetic fiber protective clothing, insulating gloves, plastic barrels, and ground coverings. Friction tests were conducted on any two of these materials. In each friction test, the contact area, friction speed and friction time are kept consistent, and the electrostatic charge or voltage generated after the two materials are rubbed is recorded by an electrostatic charge measuring device. The measured electrostatic charge or voltage is normalized to obtain the corresponding friction object factor, which reflects the electrostatic ability between the two materials.

4. The intelligent recognition method for oil depot operation behavior according to claim 2, characterized in that, The electrostatic discharge risk index is the cumulative value of multiple actions, and is calculated using the following formula: ； in, Indicates the first i Electrostatic discharge risk index for this action. Indicates the first i The time it takes for this action to end. This represents the attenuation coefficient of static electricity dissipation. Indicates the current time. This represents the overall electrostatic discharge risk index after superposition. n Indicates the number of actions.

5. The intelligent recognition method for oil depot operation behavior according to claim 4, characterized in that, When a grounding action corresponding to the friction object is detected, the electrostatic accumulation risk index of the friction object is reset to zero.

6. The intelligent recognition method for oil depot operation behavior according to claim 2, characterized in that, The friction object and its material involved in the action, as well as the speed, duration, and contact area of ​​the action, are determined in the following ways: Based on target detection and image segmentation algorithms, the friction objects of the workers are identified, and the material type of the friction objects is determined by combining the material identification model or the preset material database. The speed and duration of an action are calculated through motion key point detection and trajectory tracking. The contour of the contact area is obtained by image segmentation, and the contact area is calculated by combining depth estimation or multi-view imaging results.

7. The intelligent recognition method for oil depot operation behavior according to claim 1, characterized in that, The electrostatic environment risk index is calculated based on the work environment parameter data according to the following formula: ； in, This indicates the electrostatic environmental risk index. Indicates oil and gas concentration. Indicates humidity. Indicates temperature. is a coefficient.

8. The intelligent recognition method for oil depot operation behavior according to claim 1, characterized in that, The calculation of the total electrostatic risk index specifically includes: The electrostatic action risk index and the electrostatic environment risk index are normalized to map them to a dimensionless standardized interval. Based on the normalization results, the total electrostatic risk index is calculated using the following formula: ； in, This indicates the overall electrostatic risk index. This represents the normalized electrostatic discharge risk index. This represents the normalized electrostatic environmental risk index. and This is a weighting factor used to balance the contributions of action factors and environmental factors to the total risk.

9. The intelligent recognition method for oil depot operation behavior according to claim 8, characterized in that, The normalization method is either Min-Max normalization or Z-score standardization.

10. An intelligent recognition system for oil depot operational behavior, characterized in that, include: The motion and environment data acquisition module is used to acquire motion image data of oil depot workers and work environment parameter data, including temperature, humidity and oil and gas concentration. The motion pattern recognition and risk assessment module is used to perform motion pattern recognition based on the motion image data and assess the electrostatic motion risk index based on the worker's motion. The electrostatic motion risk index is used to measure the risk index of electrostatic discharge that may be generated by the worker's motion. The environmental risk assessment module is used to determine the electrostatic environmental risk index based on the work environment parameter data. The electrostatic environmental risk index is inversely proportional to humidity, inversely proportional to temperature, and directly proportional to oil and gas concentration. The risk integration and alarm module is used to comprehensively weight the electrostatic action risk index and the electrostatic environment risk index to obtain the total electrostatic risk index, and to trigger an alarm when the total electrostatic risk index exceeds a threshold.

Images