Crew fatigue monitoring method and system, electronic equipment and storage medium

Abstract

The present invention proposes a crew fatigue monitoring method and system, electronic device and storage medium. The crew fatigue monitoring method includes: obtaining the first characteristic information of the crew, wherein the first characteristic information includes: behavioral information, physiological signals and environmental information; determining the name information of the crew; determining a risk prediction model according to the name information, wherein the risk prediction model is a risk prediction model corrected by the cognitive bias of the crew; determining the probability of a risk event according to the first characteristic information and the risk prediction model; determining an early warning prompt strategy according to the probability. In the present invention, the crew's behavioral information, physiological signals and environmental information are obtained, and then the above information is input into the risk prediction model after the cognitive bias is corrected, so as to obtain the probability of the risk event, that is, the fatigue state of the crew. The present invention solves the problems in the related art that crew fatigue monitoring is incomplete and cognitive bias is difficult to correct.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a crew fatigue monitoring method and system, electronic equipment and storage medium. Background Art

[0002] According to relevant data, 75% of water traffic accidents caused mainly by human factors are directly or indirectly related to crew fatigue driving. In recent years, crew fatigue has become a key focus area in the shipping industry, and the monitoring and management of ship cabs has been listed as an area that urgently needs to be improved in shipping safety. However, relevant technical solutions have certain limitations in crew fatigue monitoring, such as only monitoring fatigue status through one data source, namely video / photo images, without considering the integration of multiple data sources, and without considering that the crew's cognitive bias may lead to misjudgment of risks, improper operation and other problems. Summary of the invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art or related art.

[0004] To this end, a first aspect of the present invention provides a crew fatigue monitoring method.

[0005] A second aspect of the present invention provides a crew fatigue monitoring system.

[0006] A third aspect of the present invention provides an electronic device.

[0007] A fourth aspect of the present invention provides a storage medium.

[0008] In view of this, according to a first aspect of the present invention, a crew fatigue monitoring method is proposed, comprising: obtaining first characteristic information of the crew, wherein the first characteristic information includes: behavioral information, physiological signals and environmental information; determining the name information of the crew; determining a risk prediction model based on the name information, wherein the risk prediction model is a risk prediction model corrected by using the cognitive bias of the crew; determining the probability of a risk event occurring based on the first characteristic information and the risk prediction model; and determining an early warning prompt strategy based on the probability.

[0009] The crew fatigue monitoring method provided by the present invention mainly includes: first obtaining the first characteristic information of the crew, wherein the first characteristic information includes the behavioral information, physiological signals and environmental information of the crew, wherein the behavioral information refers to the operating behavior of the crew on the ship, which can be captured by a high-definition camera. Physiological signals refer to the physical information of the crew, such as heart rate, blood pressure, brain waves, etc., and the physiological signals of the crew can be collected by wearable devices (such as heart rate monitors). By obtaining the physiological signals of the crew, the fatigue state and psychological pressure of the crew can be reflected. Environmental information refers to the environmental information of the ship where the crew is located during the voyage, such as wind speed, wave height and visibility, etc., and the impact of the current environment on the crew's operation can be evaluated by obtaining environmental information. Further, the time of the first characteristic information of the crew can be obtained in real time or at preset intervals. Then determine the name information of the crew, wherein the name information refers to the unique identity of the crew. Then determine the risk prediction model according to the name information of the crew, wherein the risk prediction model is a risk prediction model corrected by the cognitive bias of the crew, and the accuracy of the risk prediction model is ensured by correcting the risk prediction model by using the cognitive bias of the crew. At the same time, it can be understood that the cognitive bias of each crew member is different, so it is necessary to determine the corrected risk prediction model corresponding to the crew member according to the crew member's name information. After determining the risk prediction model, the first characteristic information is input into the risk prediction model to obtain the probability of a risk event. The probability of a risk event can reflect the fatigue level of the crew. Finally, the corresponding early warning prompt strategy is determined according to the probability of a risk event, that is, the fatigue level of the crew. For example, when the probability of a risk event is high, that is, when the fatigue level of the crew is serious, the early warning prompt strategy can be to notify the crew to take a rest. When the probability of a risk event is moderate, that is, when the fatigue level of the crew is moderate, the early warning prompt strategy can be to send a prompt message to the crew to remind the crew to concentrate. By timely triggering appropriate early warning signals and emergency response measures based on the probability of a risk event, the safety level of shipping is improved. In the present invention, the crew's behavioral information, physiological signals and environmental information are obtained, and then the above information is input into the risk prediction model after cognitive bias correction to obtain the probability of risk events. Then, different early warning prompt strategies are executed according to the probability, thereby solving the problems of incomplete crew fatigue monitoring and difficulty in correcting cognitive bias in related technologies. At the same time, the safety level of shipping is improved and the risk of accidents is reduced.

[0010] The crew fatigue monitoring method according to the present invention may also have the following technical features:

[0011] In some technical schemes, optionally, before the step of determining the risk prediction model according to the name information, the crew fatigue monitoring method includes: obtaining the crew's historical information; obtaining the crew's cognitive biases, the cognitive biases including: the crew's confidence in successfully avoiding risk events in the past and the crew's attention to low-probability risk events; determining a first eigenvector based on the historical information; determining a second eigenvector based on the cognitive bias; and constructing a risk prediction model based on the first eigenvector and the second eigenvector.

[0012] In the technical scheme, before the step of determining the risk prediction model according to the name information, the crew fatigue monitoring method includes: first, obtaining the crew's historical information, wherein the historical information refers to the crew's previous behavioral information, physiological signals, environmental information and risk events corresponding to the above information. Then obtain the crew's cognitive bias, wherein the cognitive bias includes the crew's confidence in successfully avoiding risk events in the past and the crew's attention to low-probability risk events. Then determine the first feature vector according to the historical information, that is, extract the features of the historical information and then convert it into a feature vector. Determine the second feature vector according to the cognitive bias, that is, extract the features of the cognitive bias and then convert it into a feature vector. Finally, construct a risk prediction model according to the first feature vector and the second feature vector, that is, learn the risk prediction model together with the first feature vector and the second feature vector to adjust the parameters in the risk prediction model, so as to construct a risk prediction model that meets expectations. By using the crew's cognitive bias to correct the risk prediction model, the accuracy of the risk prediction model is guaranteed, and shipping safety accidents caused by human factors are effectively prevented.

[0013] In some technical schemes, optionally, the step of constructing a risk prediction model based on the first eigenvector and the second eigenvector includes: generating a third eigenvector based on the first eigenvector and the second eigenvector; determining the weight and bias value of the third eigenvector based on the third eigenvector and the maximized likelihood function; and constructing a risk prediction model based on the weight and bias value.

[0014] In this technical solution, the step of constructing a risk prediction model based on the first eigenvector and the second eigenvector includes: generating a third eigenvector based on the first eigenvector and the second eigenvector, that is, integrating the first eigenvector and the second eigenvector to obtain the third eigenvector. Then, according to the third eigenvector, the weight and bias value corresponding to the third eigenvector are determined by maximizing the likelihood function, that is, the weight and bias value are the parameters of the risk prediction model. Finally, the risk prediction model is obtained according to the weight and bias value. The third eigenvector is obtained by integrating the first eigenvector and the second eigenvector, and then the weight and bias value are estimated by maximizing the likelihood function, thereby ensuring the accuracy of the risk prediction model.

[0015] In some technical solutions, optionally, the step of obtaining the first characteristic information of the crew includes: respectively collecting image information of the crew, distance information between the crew and the equipment, physiological signals of the crew, and environmental information of the ship's navigation; performing feature extraction on the image information to obtain second characteristic information; performing feature extraction on the distance information to obtain third characteristic information; performing feature extraction on the physiological signal to obtain fourth characteristic information; performing feature extraction on the environmental information to obtain fifth characteristic information; and determining the first characteristic information based on the second characteristic information, the third characteristic information, the fourth characteristic information, and the fifth characteristic information.

[0016] In this embodiment, the step of obtaining the first characteristic information of the crew includes: first, respectively collecting the image information of the crew, the distance information between the crew and the equipment, the physiological signals of the crew, and the environmental information of the ship's navigation. Among them, the image information is the image information of the crew's operating behavior on the ship captured by a high-definition camera. The distance information between the crew and the equipment can be obtained by installing a laser rangefinder sensor, and the distance between the crew and the key equipment or area can be monitored in real time to evaluate the safety of the operation. The physiological signals of the crew can be collected by a wearable device (heart rate monitor). Then, feature extraction is performed on the image information, and the key behavioral features of the crew can be extracted from the image information through image processing and computer vision algorithms, such as trajectory features (speed, acceleration), posture features (gesture type and its duration), interaction features (gesture category, interactive object ID (Identification), interactive behavior type), facial expression features, etc., and then the key behavior features are used as the second feature information. Feature extraction is performed on the distance information, and the difference between the real-time distance and the minimum safe distance, the distance state (whether it is safe), the relative position of the crew and the equipment / area, etc. are extracted, and then these features are used as the third feature information. Then, according to the collected physiological signals, physiological indicators such as heart rate variability, blood pressure variability, stability of blood oxygen saturation, and fatigue index and psychological stress index based on EEG (Electroencephalogram) are calculated, and then the above physiological indicators and / or indexes are used as the fourth feature information. Then, the environmental information is feature extracted to extract environmental factors related to the crew's operating behavior and psychological state, such as wind speed, wave height, visibility and other characteristics, and the above characteristics are used as the fifth feature information. Finally, the first feature information is determined according to the second feature information, the third feature information, the fourth feature information and the fifth feature information, that is, the second feature information, the third feature information, the fourth feature information and the fifth feature information are integrated to obtain the first feature information. By integrating the crew's image information, the distance information between the crew and the equipment, the crew's physiological signals and the environmental information of the ship's navigation into a unified first feature information, the comprehensiveness of crew fatigue monitoring is guaranteed, thereby improving the level of shipping safety and reducing the risk of accidents.

[0017] In some technical schemes, optionally, the step of determining the first characteristic information based on the second characteristic information, the third characteristic information, the fourth characteristic information and the fifth characteristic information includes: setting a time window of the same length in the second characteristic information, the third characteristic information, the fourth characteristic information and the fifth characteristic information, respectively; extracting the sixth characteristic information, the seventh characteristic information, the eighth characteristic information and the ninth characteristic information from the second characteristic information, the third characteristic information, the fourth characteristic information and the fifth characteristic information according to the time window; and splicing the sixth characteristic information, the seventh characteristic information, the eighth characteristic information and the ninth characteristic information to obtain the first characteristic information.

[0018] In the technical solution, the step of determining the first characteristic information according to the second characteristic information, the third characteristic information, the fourth characteristic information and the fifth characteristic information includes: firstly, setting a time window of the same time length in the second characteristic information, the third characteristic information, the fourth characteristic information and the fifth characteristic information respectively, and then extracting all relevant data in the time window in the second characteristic information, the third characteristic information, the fourth characteristic information and the fifth characteristic information respectively, that is, extracting the sixth characteristic information in the time window in the second characteristic information, extracting the seventh characteristic information in the time window in the third characteristic information, extracting the eighth characteristic information in the time window in the fourth characteristic information, and extracting the ninth characteristic information in the time window in the fifth characteristic information. Finally, a time window of the same time length is established, and then the sixth characteristic information, the seventh characteristic information, the eighth characteristic information and the ninth characteristic information are spliced ​​in the time window to obtain the first characteristic information. By splicing the characteristic information of multiple different data sources into one characteristic information in the same time window, the accuracy of the information input into the risk prediction model is guaranteed.

[0019] In some technical solutions, optionally, the step of determining the early warning prompt strategy based on the probability includes: matching the probability with a preset threshold to obtain a matching result; and determining the early warning prompt strategy based on the matching result.

[0020] In this technical solution, the step of determining the early warning prompt strategy according to the probability includes: matching the probability with the preset threshold to obtain a matching result, that is, comparing the probability of the occurrence of the risk event output by the risk prediction model with the preset threshold to obtain a comparison result. Then, the early warning prompt strategy is determined according to the matching result, that is, there can be multiple early warning prompt strategies, and the most appropriate early warning prompt strategy can be selected from multiple early warning prompt strategies according to the matching result. By triggering appropriate early warning signals and emergency response measures when necessary, the level of shipping safety is improved.

[0021] In some technical solutions, optionally, the step of determining the early warning prompt strategy based on the matching result includes: when the probability is greater than or equal to a first preset threshold, the early warning prompt strategy is to start the emergency evacuation procedure; when the probability is less than the first preset threshold and greater than or equal to a second preset threshold, the early warning prompt strategy is to send a prompt message to the crew, wherein the second preset threshold is less than the first preset threshold; when the probability is less than the second preset threshold, the early warning prompt strategy is to record the risk event.

[0022] In this technical solution, the step of determining the early warning prompt strategy according to the matching result includes: when the probability of a risk event is greater than or equal to the first preset threshold, the early warning prompt strategy can be to immediately notify the captain and safety officers and start the emergency evacuation procedure. When the probability is greater than or equal to the first preset threshold, it means that the probability of the current risk event is high, so for safety reasons, the emergency evacuation procedure needs to be started immediately. When the probability of a risk event is less than the first preset threshold, when it is greater than or equal to the second threshold, it means that the probability of the risk event is moderate at this time, so the early warning prompt strategy can be to issue a warning and remind the crew to pay attention to safety, and it may be necessary to adjust the operating or environmental conditions. When the probability of a risk event is less than the second preset threshold, it means that the probability of the risk event is low at this time, so the early warning prompt strategy can be to continue monitoring and record the risk event for subsequent analysis to find the cause.

[0023] According to a second aspect of the present invention, a crew fatigue monitoring system is proposed, comprising a first acquisition module, the first acquisition module is used to acquire first characteristic information of the crew, wherein the first characteristic information includes: behavioral information, physiological signals and environmental information; a first determination module, the first determination module is used to determine the name information of the crew; a second determination module, the second determination module is used to determine a risk prediction model according to the name information, wherein the risk prediction model is a risk prediction model corrected by using the cognitive bias of the crew; a third determination module, the third determination module is used to determine the probability of a risk event occurring according to the first characteristic information and the risk prediction model; and a fourth determination module, the fourth determination module is used to determine an early warning prompt strategy according to the probability.

[0024] The crew fatigue monitoring system provided by the present invention mainly includes: a first acquisition module, a first determination module, a second determination module, a third determination module and a fourth determination module. Among them, the first acquisition module can obtain the first characteristic information of the crew, wherein the first characteristic information includes the behavior information, physiological signals and environmental information of the crew, wherein the behavior information refers to the operation behavior of the crew on the ship, which can be captured by a high-definition camera. Physiological signals refer to the physical information of the crew, such as heart rate, blood pressure, brain waves, etc., and the physiological signals of the crew can be collected by wearable devices (such as heart rate monitors). By obtaining the physiological signals of the crew, the fatigue state and psychological pressure of the crew can be reflected. Environmental information refers to the environmental information of the ship where the crew is located during the voyage, such as wind speed, wave height and visibility, etc., and the impact of the current environment on the crew's operation can be evaluated by obtaining environmental information. Further, the time of the first characteristic information of the crew can be obtained in real time, or it can be obtained at preset intervals. Then the first determination module determines the name information of the crew, wherein the name information refers to the unique identity of the crew. Then the second determination module determines the risk prediction model according to the crew's name information, wherein the risk prediction model is a risk prediction model corrected by the crew's cognitive bias. By correcting the risk prediction model by using the crew's cognitive bias, the accuracy of the risk prediction model is ensured. At the same time, it can be understood that the cognitive bias of each crew member is different, so it is necessary to determine the corrected risk prediction model corresponding to the crew member according to the crew's name information. After determining the risk prediction model, the third determination module inputs the first feature information into the risk prediction model to obtain the probability of a risk event. The probability of a risk event can reflect the fatigue level of the crew. Finally, the fourth determination module determines the corresponding early warning prompt strategy according to the probability of a risk event, that is, the fatigue level of the crew. For example, when the probability of a risk event is large, that is, the fatigue level of the crew is serious, the early warning prompt strategy can be to notify the crew to take a rest. When the probability of a risk event is moderate, that is, the fatigue level of the crew is moderate, the early warning prompt strategy can be to send a prompt message to the crew to remind the crew to concentrate. By timely triggering appropriate early warning signals and emergency response measures based on the probability of a risk event, the safety level of shipping is improved. In the present invention, the crew's behavioral information, physiological signals and environmental information are obtained, and then the above information is input into the risk prediction model after cognitive bias correction to obtain the probability of risk events. Then, different early warning prompt strategies are executed according to the probability, thereby solving the problems of incomplete crew fatigue monitoring and difficulty in correcting cognitive bias in related technologies. At the same time, the safety level of shipping is improved and the risk of accidents is reduced.

[0025] According to a third aspect of the present invention, an electronic device is proposed, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of any of the above-mentioned crew fatigue monitoring methods are implemented.

[0026] The electronic device provided by the present invention implements the steps of the above-mentioned crew fatigue monitoring method when the processor executes the computer program, which can achieve the technical effects of any of the above-mentioned technical solutions and will not be repeated here.

[0027] According to a fourth aspect of the present invention, a storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of any of the above-mentioned crew fatigue monitoring methods are implemented.

[0028] The storage medium provided by the present invention implements the steps of the above-mentioned crew fatigue monitoring method when the computer program is executed by the processor, and can achieve the technical effects of any of the above-mentioned technical solutions, which will not be repeated here.

[0029] Additional aspects and advantages of the present invention will become apparent from the following description or may be learned by practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0031] Figure 1 A schematic diagram showing a process flow of a crew fatigue monitoring method according to an embodiment of the present invention;

[0032] Figure 2 A schematic diagram of a process flow before the step of determining a risk prediction model according to name information in a crew fatigue monitoring method according to an embodiment of the present invention is shown;

[0033] Figure 3 A schematic flow chart showing the steps of constructing a risk prediction model according to a first eigenvector and a second eigenvector in a crew fatigue monitoring method according to an embodiment of the present invention;

[0034] Figure 4 A schematic flow chart showing the steps of obtaining first characteristic information of a crew member in a crew member fatigue monitoring method according to an embodiment of the present invention;

[0035] Figure 5 A schematic flow chart showing a step of determining first characteristic information according to second characteristic information, third characteristic information, fourth characteristic information and fifth characteristic information in a crew fatigue monitoring method according to an embodiment of the present invention;

[0036] Figure 6A schematic flow chart showing the steps of determining an early warning strategy according to probability in a crew fatigue monitoring method according to an embodiment of the present invention;

[0037] Figure 7 A schematic flow chart showing the steps of determining an early warning strategy according to a matching result in a crew fatigue monitoring method according to an embodiment of the present invention;

[0038] Figure 8 A schematic block diagram showing a crew fatigue monitoring system according to an embodiment of the present invention;

[0039] Fig. 9 A schematic block diagram of an electronic device according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0040] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict.

[0041] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0042] Figure 1 A schematic flow chart of a crew fatigue monitoring method according to an embodiment of the present invention is shown.

[0043] Among them, the method includes:

[0044] Step 102: Acquire first characteristic information of the crew, wherein the first characteristic information includes: behavior information, physiological signals, and environmental information;

[0045] Step 104: Determine the name information of the crew;

[0046] Step 106: determining a risk prediction model according to the name information, wherein the risk prediction model is a risk prediction model corrected by using the cognitive bias of the crew;

[0047] Step 108: Determine the probability of a risk event occurring according to the first feature information and the risk prediction model;

[0048] Step 110: Determine an early warning strategy based on the probability.

[0049] The crew fatigue monitoring method provided by the present invention mainly includes: first obtaining the first characteristic information of the crew, wherein the first characteristic information includes the behavioral information, physiological signals and environmental information of the crew, wherein the behavioral information refers to the operating behavior of the crew on the ship, which can be captured by a high-definition camera. Physiological signals refer to the physical information of the crew, such as heart rate, blood pressure, brain waves, etc., and the physiological signals of the crew can be collected by wearable devices (such as heart rate monitors). By obtaining the physiological signals of the crew, the fatigue state and psychological pressure of the crew can be reflected. Environmental information refers to the environmental information of the ship where the crew is located during the voyage, such as wind speed, wave height and visibility, etc., and the impact of the current environment on the crew's operation can be evaluated by obtaining environmental information. Further, the time of the first characteristic information of the crew can be obtained in real time or at preset intervals. Then determine the name information of the crew, wherein the name information refers to the unique identity of the crew. Then determine the risk prediction model according to the name information of the crew, wherein the risk prediction model is a risk prediction model corrected by the cognitive bias of the crew, and the accuracy of the risk prediction model is ensured by correcting the risk prediction model by using the cognitive bias of the crew. At the same time, it can be understood that the cognitive bias of each crew member is different, so it is necessary to determine the corrected risk prediction model corresponding to the crew member according to the crew member's name information. After determining the risk prediction model, the first characteristic information is input into the risk prediction model to obtain the probability of a risk event. The probability of a risk event can reflect the fatigue level of the crew. Finally, the corresponding early warning prompt strategy is determined according to the probability of a risk event, that is, the fatigue level of the crew. For example, when the probability of a risk event is high, that is, when the fatigue level of the crew is serious, the early warning prompt strategy can be to notify the crew to take a rest. When the probability of a risk event is moderate, that is, when the fatigue level of the crew is moderate, the early warning prompt strategy can be to send a prompt message to the crew to remind the crew to concentrate. By timely triggering appropriate early warning signals and emergency response measures based on the probability of a risk event, the safety level of shipping is improved. In the present invention, the crew's behavioral information, physiological signals and environmental information are obtained, and then the above information is input into the risk prediction model after cognitive bias correction to obtain the probability of risk events. Then, different early warning prompt strategies are executed according to the probability, thereby solving the problems of incomplete crew fatigue monitoring and difficulty in correcting cognitive bias in related technologies. At the same time, the safety level of shipping is improved and the risk of accidents is reduced.

[0050] Figure 2 A flow chart of a crew fatigue monitoring method according to an embodiment of the present invention is shown before the step of determining a risk prediction model according to name information; wherein, before the step of determining a risk prediction model according to name information, the crew fatigue monitoring method includes:

[0051] Step 202: Obtain historical information of the crew;

[0052] Step 204: Obtaining the crew's cognitive bias, which includes: the crew's confidence in successfully avoiding risk events in the past and the crew's attention to low-probability risk events;

[0053] Step 206: Determine a first feature vector according to historical information;

[0054] Step 208: determining a second eigenvector according to cognitive bias;

[0055] Step 210: Construct a risk prediction model based on the first eigenvector and the second eigenvector.

[0056] In this embodiment, before the step of determining the risk prediction model according to the name information, the crew fatigue monitoring method includes: first, obtaining the crew's historical information, wherein the historical information refers to the crew's previous behavioral information, physiological signals, environmental information and risk events corresponding to the above information. Then, the crew's cognitive bias is obtained, wherein the cognitive bias includes the crew's confidence in successfully avoiding risk events in the past and the crew's attention to low-probability risk events. Then, the first feature vector is determined according to the historical information, that is, the historical information is feature extracted and then converted into a feature vector. The second feature vector is determined according to the cognitive bias, that is, the cognitive bias is feature extracted and then converted into a feature vector. Finally, a risk prediction model is constructed according to the first feature vector and the second feature vector, that is, the risk prediction model is learned together with the first feature vector and the second feature vector to adjust the parameters in the risk prediction model, thereby constructing a risk prediction model that meets expectations. By using the crew's cognitive bias to correct the risk prediction model, the accuracy of the risk prediction model is guaranteed, and shipping safety accidents caused by human factors are effectively prevented.

[0057] Figure 3 A schematic flow chart of the step of constructing a risk prediction model according to the first eigenvector and the second eigenvector in a crew fatigue monitoring method according to an embodiment of the present invention is shown; wherein the step of constructing a risk prediction model according to the first eigenvector and the second eigenvector includes:

[0058] Step 302: Generate a third eigenvector based on the first eigenvector and the second eigenvector;

[0059] Step 304: determining a weight and a bias value of the third eigenvector according to the third eigenvector and the maximized likelihood function;

[0060] Step 306: Construct a risk prediction model based on the weights and bias values.

[0061] In this embodiment, the step of constructing a risk prediction model based on the first eigenvector and the second eigenvector includes: generating a third eigenvector based on the first eigenvector and the second eigenvector, that is, integrating the first eigenvector and the second eigenvector to obtain the third eigenvector. Then, according to the third eigenvector, the weight and bias value corresponding to the third eigenvector are determined by maximizing the likelihood function, that is, the weight and bias value are the parameters of the risk prediction model. Finally, the risk prediction model is obtained according to the weight and bias value. The third eigenvector is obtained by integrating the first eigenvector and the second eigenvector, and then the weight and bias value are estimated by maximizing the likelihood function, thereby ensuring the accuracy of the risk prediction model.

[0062] Figure 4 A schematic flow chart of the step of obtaining the first characteristic information of a crew member in a crew fatigue monitoring method according to an embodiment of the present invention is shown; wherein the step of obtaining the first characteristic information of a crew member includes:

[0063] Step 402: collecting image information of the crew, distance information between the crew and the equipment, physiological signals of the crew, and environmental information of the ship's navigation respectively;

[0064] Step 404: extracting features from the image information to obtain second feature information;

[0065] Step 406: extracting features from the distance information to obtain third feature information;

[0066] Step 408: extracting features from the physiological signal to obtain fourth feature information;

[0067] Step 410: extracting features from the environmental information to obtain fifth feature information;

[0068] Step 412: Determine the first characteristic information according to the second characteristic information, the third characteristic information, the fourth characteristic information, and the fifth characteristic information.

[0069] In this embodiment, the step of obtaining the first characteristic information of the crew includes: first, respectively collecting the image information of the crew, the distance information between the crew and the equipment, the physiological signals of the crew, and the environmental information of the ship's navigation. Among them, the image information is the image information of the crew's operating behavior on the ship captured by a high-definition camera. The distance information between the crew and the equipment can be obtained by installing a laser rangefinder sensor, and the distance between the crew and the key equipment or area can be monitored in real time to evaluate the safety of the operation. The physiological signals of the crew can be collected by a wearable device (heart rate monitor). Then, feature extraction is performed on the image information, and the key behavioral features of the crew can be extracted from the image information through image processing and computer vision algorithms, such as trajectory features (speed, acceleration), posture features (gesture type and its duration), interaction features (gesture category, interactive object ID (Identification), interactive behavior type), facial expression features, etc., and then the key behavior features are used as the second feature information. Feature extraction is performed on the distance information, and the difference between the real-time distance and the minimum safe distance, the distance state (whether it is safe), the relative position of the crew and the equipment / area, etc. are extracted, and then these features are used as the third feature information. Then, according to the collected physiological signals, physiological indicators such as heart rate variability, blood pressure variability, stability of blood oxygen saturation, and fatigue index and psychological stress index based on EEG (Electroencephalogram) are calculated, and then the above physiological indicators and / or indexes are used as the fourth feature information. Then, the environmental information is feature extracted to extract environmental factors related to the crew's operating behavior and psychological state, such as wind speed, wave height, visibility and other characteristics, and the above characteristics are used as the fifth feature information. Finally, the first feature information is determined according to the second feature information, the third feature information, the fourth feature information and the fifth feature information, that is, the second feature information, the third feature information, the fourth feature information and the fifth feature information are integrated to obtain the first feature information. By integrating the crew's image information, the distance information between the crew and the equipment, the crew's physiological signals and the environmental information of the ship's navigation into a unified first feature information, the comprehensiveness of crew fatigue monitoring is guaranteed, thereby improving the level of shipping safety and reducing the risk of accidents.

[0070] Figure 5 A flow chart showing a step of determining first characteristic information according to second characteristic information, third characteristic information, fourth characteristic information and fifth characteristic information in a crew fatigue monitoring method according to an embodiment of the present invention is shown; wherein the step of determining first characteristic information according to second characteristic information, third characteristic information, fourth characteristic information and fifth characteristic information comprises:

[0071] Step 502: setting a time window of the same length in each of the second characteristic information, the third characteristic information, the fourth characteristic information and the fifth characteristic information;

[0072] Step 504: extracting sixth feature information, seventh feature information, eighth feature information and ninth feature information from the second feature information, the third feature information, the fourth feature information and the fifth feature information respectively according to the time window;

[0073] Step 506: concatenate the sixth characteristic information, the seventh characteristic information, the eighth characteristic information and the ninth characteristic information to obtain the first characteristic information.

[0074] In this embodiment, the step of determining the first characteristic information according to the second characteristic information, the third characteristic information, the fourth characteristic information and the fifth characteristic information includes: firstly, setting a time window of the same time length in the second characteristic information, the third characteristic information, the fourth characteristic information and the fifth characteristic information respectively, and then extracting all relevant data in the time window in the second characteristic information, the third characteristic information, the fourth characteristic information and the fifth characteristic information respectively, that is, extracting the sixth characteristic information in the time window in the second characteristic information, extracting the seventh characteristic information in the time window in the third characteristic information, extracting the eighth characteristic information in the time window in the fourth characteristic information, and extracting the ninth characteristic information in the time window in the fifth characteristic information. Finally, a time window of the same time length is established, and then the sixth characteristic information, the seventh characteristic information, the eighth characteristic information and the ninth characteristic information are spliced ​​in the time window to obtain the first characteristic information. By splicing the characteristic information of multiple different data sources into one characteristic information in the same time window, the accuracy of the information input into the risk prediction model is guaranteed.

[0075] Figure 6 A flow chart showing the steps of determining the early warning prompt strategy according to probability in a crew fatigue monitoring method according to an embodiment of the present invention is shown; wherein the steps of determining the early warning prompt strategy according to probability include:

[0076] Step 602: Match the probability with a preset threshold to obtain a matching result;

[0077] Step 604: Determine an early warning prompt strategy based on the matching result.

[0078] In this embodiment, the step of determining the early warning prompt strategy according to the probability includes: matching the probability with the preset threshold value to obtain a matching result, that is, comparing the probability of the occurrence of the risk event output by the risk prediction model with the preset threshold value to obtain a comparison result. Then, the early warning prompt strategy is determined according to the matching result, that is, there can be multiple early warning prompt strategies, and the most appropriate early warning prompt strategy can be selected from the multiple early warning prompt strategies according to the matching result. By triggering appropriate early warning signals and emergency response measures when necessary, the level of shipping safety is improved.

[0079] Figure 7 A schematic flow chart of the step of determining a warning prompt strategy according to a matching result in a crew fatigue monitoring method according to an embodiment of the present invention is shown; wherein the step of determining a warning prompt strategy according to a matching result includes:

[0080] Step 702: When the probability is greater than or equal to a first preset threshold, the early warning prompt strategy is to start an emergency evacuation procedure;

[0081] Step 704: when the probability is less than the first preset threshold and greater than or equal to the second preset threshold, the early warning prompt strategy is to issue a prompt message to the crew, wherein the second preset threshold is less than the first preset threshold;

[0082] Step 706: When the probability is less than the second preset threshold, the early warning prompt strategy is to record the risk event.

[0083] In this embodiment, the step of determining the early warning prompt strategy according to the matching result includes: when the probability of a risk event is greater than or equal to the first preset threshold, the early warning prompt strategy can be to immediately notify the captain and the safety officer and start the emergency evacuation procedure. When the probability is greater than or equal to the first preset threshold, it means that the probability of the current risk event is high, so for safety reasons, the emergency evacuation procedure needs to be started immediately. When the probability of a risk event is less than the first preset threshold, when it is greater than or equal to the second threshold, it means that the probability of the risk event is moderate at this time, so the early warning prompt strategy can be to issue a warning and remind the crew to pay attention to safety, and it may be necessary to adjust the operating or environmental conditions. When the probability of a risk event is less than the second preset threshold, it means that the probability of the risk event is low at this time, so the early warning prompt strategy can be to continue monitoring and record the risk event for subsequent analysis to find the cause.

[0084] The crew fatigue monitoring method provided by the present invention mainly includes: firstly, multi-dimensional data collection and fusion. Specifically, by integrating multiple technical means such as video recognition, laser ranging and physiological signal monitoring, and combining shipping environment data, various key information of the crew during the operation process can be fully captured. Among them, the specific technologies and collected data are as follows: The first is video recognition technology, which uses high-definition cameras to capture the operating behavior of the crew on board, and extracts key behavioral features through image processing and computer vision algorithms. The specific indicators are shown in Table 1:

[0085] Table 1

[0086]

[0087] The second is laser ranging technology, which can monitor the distance between the crew and key equipment or areas in real time by installing laser ranging sensors to assess the safety of operations. Its specific indicators are shown in Table 2:

[0088] Table 2

[0089]

[0090] The third is physiological signal monitoring: wearable devices (such as heart rate monitors, brain wave sensors (optional)) are used to collect physiological signals of crew members, such as heart rate, blood pressure, brain waves, etc., to assess their fatigue status and psychological stress. The specific indicators are shown in Table 3:

[0091] Table 3

[0092]

[0093]

[0094] The fourth is environmental data collection: collecting environmental data during the ship's voyage, such as wind speed, wave height, visibility, etc., to assess the impact of the environment on crew operations. The specific indicators are shown in Table 4:

[0095] Table 4

[0096]

[0097]

[0098] After obtaining the above data, feature extraction is performed on the above indicators. Specifically, for video recognition data, key features such as trajectory features (speed, acceleration), posture features (posture type and duration), interaction features (gesture category, interaction object ID, interaction behavior type), facial expression features, etc. can be extracted. For laser ranging data, the difference between the real-time distance and the minimum safe distance, the distance status (whether it is safe), the relative position of the crew and the equipment / area, etc. can be extracted. For physiological signal data, physiological indicators such as heart rate variability, blood pressure variability, stability of blood oxygen saturation, and fatigue index and psychological stress index based on EEG can be calculated. For environmental information data, environmental factors related to the crew's operating behavior and psychological state, such as wind speed, wave height, visibility, etc., can be extracted. After extracting the above features, the data is fused based on the time window, a time window of the same length is set for each data type, and all relevant data in the window are extracted. Features from different data sources are spliced ​​into a feature vector X in the same time window.

[0099] Then, the risk prediction model is constructed. Specifically, it can be based on the fused data and feature vector X in the previous step. Since X includes features extracted from different data sources, we can get:

[0100] X=[x 1 , x 2 , …, x n ];

[0101] Among them, x n Represents the nth feature, including: x 1 , x 2 , x 3 Can be trajectory characteristics (speed, acceleration, direction change rate); x 4 , x 5 Can be posture features (posture type, posture duration); x 6 , x 7 , x 8 Can be interaction features (gesture category, interaction object ID, interaction behavior type); x 9 Can be facial expression features; x 10 , x 11 , x 12 It can be physiological signal characteristics (heart rate variability, blood pressure variability, blood oxygen saturation stability); x 13 , x 14 It can be a fatigue index and psychological stress index based on EEG; 15 , x 16 , x 17It can be laser ranging feature (difference between real-time distance and minimum safety distance, distance status, relative position); x 18 , x 19 , x 20 It can be environmental information characteristics (wind speed, wave height, visibility).

[0102] Then, based on the above feature vectors, a logistic regression model is constructed to predict the probability of discovering risks in the future based on historical data:

[0103]

[0104] Among them, P(y=1|x) is the probability that the crew is in a high-risk state when a given feature vector X is given, w is a weight vector containing the weight of each feature, and b is a bias term.

[0105] After building the model, it is necessary to learn the historical data for the parameters and adjust the parameters. In this patent, we use maximum likelihood estimation to learn the weight w and bias b. The goal is to maximize the likelihood function of the data:

[0106]

[0107] Where m is the number of samples, y i is the actual label of the i-th sample, P(y=1|x i ) is the probability that the crew is in a high-risk state given the feature vector X.

[0108] Then use the gradient ascent (or gradient descent) algorithm to optimize w and b to maximize the likelihood function. The specific formula can be:

[0109]

[0110] Where m is the number of samples, y i is the actual label of the i-th sample, P(y=1|x i ) is the probability that the crew is in a high-risk state when the feature vector X is given, and L represents the maximum likelihood function.

[0111] The performance of the model can be evaluated by setting relevant indicators, such as accuracy, recall, F1 score, etc. The model training method is common and will not be described in detail in this patent. Furthermore, based on the historical data we have, through learning, we can give our calculation results for reference: w = [w 1 ,w ％ ,…,w ％' ]=

[0112] [0.1,-0.2,0.05,0.3,-0.15,0.25,0.2,-0.1,0.05,0.15,-0.05,0.1,0.05,-0.2,0.3,-0.1,0.2,-0.05,0.1];

[0113] b = 0.05;

[0114] Among them, w represents the weight, b represents the bias value, and w 1 = 0.1 indicates that the trajectory characteristics (speed) have a slight positive impact on the model output. ％ =-0.2 means that the trajectory characteristics (acceleration) have a greater negative impact on the model output. - = 0.05 means that the posture feature (posture type) has a slight positive impact on the model output... The others are the same and will not be elaborated. The bias term is usually used to adjust the baseline probability of the model output. b = 0.05 means that the model predicts that the baseline probability of the crew being in a high-risk state is 5%.

[0115] Furthermore, cognitive biases are identified and corrected. Based on the model constructed in the previous step, the cognitive biases (such as optimism bias and neglect of low-probability events) of crew members in the risk perception process are taken into account. To explain further, in the shipping field, optimism bias refers to the fact that crew members are often overconfident when assessing the risks they face and tend to underestimate the possibility of potential dangers or accidents. This psychological phenomenon may cause crew members to ignore safety regulations, delay taking preventive measures, or make inappropriate decisions in emergency situations. For example, they may believe that their sailing skills are sufficient to cope with bad weather and ignore the unpredictability of weather changes; or after working for a long time, they may underestimate the impact of fatigue on operational capabilities and continue to perform tasks without taking necessary rest. The existence of this bias increases the risk of maritime accidents and poses a threat to the safety of crew members and the reliability of shipping.

[0116] Ignoring low-probability events means that when making risk assessments and decisions, crew members tend to ignore those events that have a low probability of occurrence but may lead to serious consequences once they occur. This phenomenon is common among crew members because they may think that these low-probability events are unlikely to happen to them, or that existing safety measures are sufficient to prevent these risks. For example, crew members may not be fully prepared to deal with extreme weather conditions because they believe that such weather does not occur often, or that the design and construction of the ship can withstand most severe conditions. Similarly, they may not regularly check life-saving equipment or conduct emergency evacuation drills because they believe that these emergencies rarely occur and there is no need to prepare frequently. In addition, crew members may neglect regular maintenance of key ship systems because they believe that these systems are unlikely to fail.

[0117] Therefore, it is necessary to introduce cognitive bias factors into the model based on the model built in the previous step, and expand the original feature vector X to include features related to cognitive bias:

[0118] x .e" =[x,c 1 ,c ％ ,…,c 0 ];

[0119] Among them, x new represents the new vector feature, x is the original feature vector, c 1 ,c ％ ,…,c 0 is a newly introduced cognitive bias-related feature, c 1 Indicates the crew's confidence in successfully avoiding risk events in the past, c ％ Indicates the crew's attention to low-probability risk events, c 0 Indicates other factors related to cognitive biases.

[0120] Finally, the model after considering cognitive bias is:

[0121]

[0122] Among them, x new Represents the new vector feature, P(D=1|x new ) is a given eigenvector x new When , the probability that the crew is in a high-risk state, w is the weight vector, which contains the weight of each feature, b is the bias value, and the subsequent method of model adjustment and training is the same as the previous step, and the weight matrix w will be obtained in the end.

[0123] Finally, the intelligent early warning and emergency response system is integrated, that is, the intelligent early warning prompt strategy is set based on the risk prediction model obtained in the previous step. According to the output of the risk assessment model, it will be decided whether to trigger the early warning signal. Among them, different thresholds can be set to correspond to different risk warning situations. The formula is:

[0124]

[0125] Among them, T high is the high risk threshold, T medium is the medium risk threshold, Trigger Alert is the level of the triggered warning signal, P(R=1|x new) indicates the probability, otherwise indicates no probability. Then, according to the level of the early warning signal, the corresponding emergency response measures will be automatically executed. Specifically, the emergency response measures for Level 1 (high risk) are to immediately notify the captain and safety officers and initiate emergency evacuation procedures. The emergency response measures for Level 2 (medium risk) are to issue warnings and remind the crew to pay attention to safety, which may require adjustments to operating or environmental conditions. The emergency response measures for Level 3 (low risk) are to record risk events, conduct routine monitoring and subsequent analysis.

[0126] The risk prediction model and intelligent early warning and emergency response system are integrated into the intelligent shipping system, and tested and optimized in the actual environment. According to the test results and user feedback, the model parameters and thresholds are adjusted to improve the accuracy and response efficiency of the model. Through such a design, the behavior and psychological state of the crew can be monitored in real time, and then the risks can be assessed, and appropriate early warning signals and emergency response measures can be triggered when necessary, thereby improving the level of shipping safety.

[0127] Figure 8 A schematic block diagram of a crew fatigue monitoring system according to an embodiment of the present invention is shown; wherein the crew fatigue monitoring system 80 comprises:

[0128] A first acquisition module 802 is used to acquire first characteristic information of the crew, wherein the first characteristic information includes: behavior information, physiological signals and environmental information;

[0129] The first determination module 804 is used to determine the name information of the crew member;

[0130] A second determination module 806 is used to determine a risk prediction model according to the name information, wherein the risk prediction model is a risk prediction model corrected by using the cognitive bias of the crew;

[0131] The third determination module 808 is used to determine the probability of a risk event occurring according to the first feature information and the risk prediction model;

[0132] The fourth determination module 810 is used to determine the early warning prompt strategy according to the probability.

[0133] The crew fatigue monitoring system 80 provided by the present invention mainly includes: a first acquisition module 802, a first determination module 804, a second determination module 806, a third determination module 808 and a fourth determination module 810. Among them, the first acquisition module 802 can obtain the first characteristic information of the crew, wherein the first characteristic information includes the behavior information, physiological signals and environmental information of the crew, wherein the behavior information refers to the operation behavior of the crew on the ship, which can be captured by a high-definition camera. Physiological signals refer to the physical information of the crew, such as heart rate, blood pressure, brain waves, etc., and the physiological signals of the crew can be collected by wearable devices (such as heart rate monitors). By obtaining the physiological signals of the crew, the fatigue state and psychological pressure of the crew can be reflected. Environmental information refers to the environmental information of the ship where the crew is located during the voyage, such as wind speed, wave height and visibility, etc., and the impact of the current environment on the crew's operation can be evaluated by obtaining environmental information. Further, the time of the first characteristic information of the crew can be obtained in real time or at preset intervals. Then the first determination module 804 determines the name information of the crew, wherein the name information refers to the unique identity of the crew. Then the second determination module 806 determines the risk prediction model according to the crew's name information, wherein the risk prediction model is a risk prediction model corrected by the crew's cognitive bias. By correcting the risk prediction model by using the crew's cognitive bias, the accuracy of the risk prediction model is ensured. At the same time, it can be understood that the cognitive bias of each crew member is different, so it is necessary to determine the corrected risk prediction model corresponding to the crew member according to the crew's name information. After determining the risk prediction model, the third determination module 808 inputs the first feature information into the risk prediction model to obtain the probability of a risk event. The probability of a risk event can reflect the fatigue level of the crew. Finally, the fourth determination module 810 determines the corresponding early warning prompt strategy according to the probability of a risk event, that is, the fatigue level of the crew. For example, when the probability of a risk event is large, that is, the fatigue level of the crew is serious, the early warning prompt strategy can be to notify the crew to take a rest. When the probability of a risk event is moderate, that is, the fatigue level of the crew is moderate, the early warning prompt strategy can be to send a prompt message to the crew to remind the crew to concentrate. By timely triggering appropriate early warning signals and emergency response measures based on the probability of a risk event, the safety level of shipping is improved. In the present invention, the crew's behavioral information, physiological signals and environmental information are obtained, and then the above information is input into the risk prediction model after cognitive bias correction to obtain the probability of risk events. Then, different early warning prompt strategies are executed according to the probability, thereby solving the problems of incomplete crew fatigue monitoring and difficulty in correcting cognitive bias in related technologies. At the same time, the safety level of shipping is improved and the risk of accidents is reduced.

[0134] In some embodiments, optionally, the crew fatigue monitoring system 80 also includes: a second acquisition module for acquiring historical information of the crew; a third acquisition module for acquiring cognitive biases of the crew, the cognitive biases including: the crew's confidence in successfully avoiding risk events in the past and the crew's attention to low-probability risk events; a fifth determination module for determining a first eigenvector based on historical information; a sixth determination module for determining a second eigenvector based on cognitive biases; and a construction module for constructing a risk prediction model based on the first eigenvector and the second eigenvector.

[0135] In this embodiment, the crew fatigue monitoring system 80 also includes: a second acquisition module, a third acquisition module, a fifth determination module, a sixth determination module and a construction module. Among them, the second acquisition module can obtain the crew's historical information, wherein the historical information refers to the crew's previous behavior information, physiological signals, environmental information and risk events corresponding to the above information. Then the third acquisition module obtains the crew's cognitive bias, wherein the cognitive bias includes the crew's confidence in successfully avoiding risk events in the past and the crew's attention to low-probability risk events. Then the fifth determination module determines the first feature vector based on the historical information, that is, extracts the features of the historical information and then converts it into a feature vector. The sixth determination module determines the second feature vector based on the cognitive bias, that is, extracts the features of the cognitive bias and then converts it into a feature vector. Finally, the construction module constructs a risk prediction model based on the first feature vector and the second feature vector, that is, the first feature vector and the second feature vector are used together to learn the risk prediction model to adjust the parameters in the risk prediction model, thereby constructing a risk prediction model that meets expectations. By using the crew's cognitive bias to correct the risk prediction model, the accuracy of the risk prediction model is guaranteed, and shipping safety accidents caused by human factors are effectively prevented.

[0136] In some embodiments, optionally, the construction module is specifically used to generate a third eigenvector based on the first eigenvector and the second eigenvector; determine the weight and bias value of the third eigenvector based on the third eigenvector and the maximized likelihood function; and construct a risk prediction model based on the weight and bias value.

[0137] In this embodiment, the construction module is specifically used to generate a third eigenvector based on the first eigenvector and the second eigenvector, that is, the first eigenvector and the second eigenvector are integrated to obtain the third eigenvector. Then, the weight and bias values ​​corresponding to the third eigenvector are determined based on the third eigenvector by maximizing the likelihood function, that is, the weight and bias values ​​are the parameters of the risk prediction model. Finally, the risk prediction model is obtained based on the weight and bias values. The third eigenvector is obtained by integrating the first eigenvector and the second eigenvector, and then the weight and bias values ​​are estimated by maximizing the likelihood function, thereby ensuring the accuracy of the risk prediction model.

[0138] In some embodiments, optionally, the first acquisition module 802 is specifically used to respectively collect image information of the crew, distance information between the crew and the equipment, physiological signals of the crew, and environmental information of the ship's navigation; perform feature extraction on the image information to obtain second feature information; perform feature extraction on the distance information to obtain third feature information; perform feature extraction on the physiological signal to obtain fourth feature information; perform feature extraction on the environmental information to obtain fifth feature information; and determine the first feature information based on the second feature information, the third feature information, the fourth feature information, and the fifth feature information.

[0139] In this embodiment, the first acquisition module 802 is specifically used to collect the image information of the crew, the distance information between the crew and the equipment, the physiological signals of the crew, and the environmental information of the ship's navigation. Among them, the image information is the image information of the crew's operating behavior on the ship captured by a high-definition camera. The distance information between the crew and the equipment can be obtained by installing a laser rangefinder sensor, and the distance between the crew and the key equipment or area can be monitored in real time to evaluate the safety of the operation. The physiological signals of the crew can be collected by a wearable device (heart rate monitor). Then the image information is feature extracted, and the key behavior features of the crew can be extracted from the image information through image processing and computer vision algorithms, such as trajectory features (speed, acceleration), posture features (gesture type and its duration), interaction features (gesture category, interactive object ID (Identification, identity representation number), interactive behavior type), facial expression features, etc., and then the key behavior features are used as the second feature information. Feature extraction is performed on the distance information to extract the difference between the real-time distance and the minimum safe distance, the distance state (whether it is safe), the relative position of the crew and the equipment / area, and other features, and then these features are used as the third feature information. Then, according to the collected physiological signals, physiological indicators such as heart rate variability, blood pressure variability, stability of blood oxygen saturation, and fatigue index and psychological stress index based on EEG (Electroencephalogram) are calculated, and then the above physiological indicators and / or indexes are used as the fourth feature information. Then, the environmental information is feature extracted to extract environmental factors related to the crew's operating behavior and psychological state, such as wind speed, wave height, visibility and other characteristics, and the above characteristics are used as the fifth feature information. Finally, the first feature information is determined according to the second feature information, the third feature information, the fourth feature information and the fifth feature information, that is, the second feature information, the third feature information, the fourth feature information and the fifth feature information are integrated to obtain the first feature information. By integrating the crew's image information, the distance information between the crew and the equipment, the crew's physiological signals and the environmental information of the ship's navigation into a unified first feature information, the comprehensiveness of crew fatigue monitoring is guaranteed, thereby improving the level of shipping safety and reducing the risk of accidents.

[0140] In some embodiments, optionally, the first acquisition module 802 is also specifically used to set a time window of the same length in the second characteristic information, the third characteristic information, the fourth characteristic information and the fifth characteristic information, respectively; extract the sixth characteristic information, the seventh characteristic information, the eighth characteristic information and the ninth characteristic information from the second characteristic information, the third characteristic information, the fourth characteristic information and the fifth characteristic information according to the time window; and concatenate the sixth characteristic information, the seventh characteristic information, the eighth characteristic information and the ninth characteristic information to obtain the first characteristic information.

[0141] In this embodiment, the first acquisition module 802 is also specifically used to set a time window with the same time length in the second characteristic information, the third characteristic information, the fourth characteristic information and the fifth characteristic information, and then extract all relevant data in the time window from the second characteristic information, the third characteristic information, the fourth characteristic information and the fifth characteristic information, that is, extract the sixth characteristic information in the time window in the second characteristic information, extract the seventh characteristic information in the time window in the third characteristic information, extract the eighth characteristic information in the time window in the fourth characteristic information, and extract the ninth characteristic information in the time window in the fifth characteristic information. Finally, a time window with the same time length is established, and then the sixth characteristic information, the seventh characteristic information, the eighth characteristic information and the ninth characteristic information are spliced ​​in the time window to obtain the first characteristic information. By splicing the characteristic information of multiple different data sources into one characteristic information in the same time window, the accuracy of the information input into the risk prediction model is guaranteed.

[0142] In some embodiments, optionally, the fourth determination module 810 is specifically configured to match the probability with a preset threshold to obtain a matching result; and determine a warning prompt strategy according to the matching result.

[0143] In this embodiment, the fourth determination module 810 is specifically used to match the probability with the preset threshold value to obtain a matching result, that is, to compare the probability of the occurrence of the risk event output by the risk prediction model with the preset threshold value to obtain a comparison result. Then, the early warning prompt strategy is determined according to the matching result, that is, there can be multiple early warning prompt strategies, and the most appropriate early warning prompt strategy can be selected from the multiple early warning prompt strategies according to the matching result. By triggering appropriate early warning signals and emergency response measures when necessary, the level of shipping safety is improved.

[0144] In some embodiments, optionally, the fourth determination module 810 is also specifically used to, when the probability is greater than or equal to a first preset threshold, the early warning prompt strategy is to start an emergency evacuation procedure; when the probability is less than the first preset threshold and greater than or equal to a second preset threshold, the early warning prompt strategy is to send a prompt message to the crew, wherein the second preset threshold is less than the first preset threshold; when the probability is less than the second preset threshold, the early warning prompt strategy is to record the risk event.

[0145] In this embodiment, the fourth determination module 810 is also specifically used for when the probability of a risk event is greater than or equal to the first preset threshold, the early warning prompt strategy can be to immediately notify the captain and the safety officer and start the emergency evacuation procedure. When the probability is greater than or equal to the first preset threshold, it means that the probability of the current risk event is relatively high, so for safety reasons, the emergency evacuation procedure needs to be started immediately. When the probability of a risk event is less than the first preset threshold, when it is greater than or equal to the second threshold, it means that the probability of the risk event is moderate at this time, so the early warning prompt strategy can be to issue a warning and remind the crew to pay attention to safety, and it may be necessary to adjust the operating or environmental conditions. When the probability of a risk event is less than the second preset threshold, it means that the probability of the risk event is low at this time, so the early warning prompt strategy can be to continue monitoring and record the risk event at the same time for subsequent analysis to find the cause.

[0146] Fig. 9 A schematic block diagram of an electronic device according to an embodiment of the present invention is shown; wherein the electronic device 90 includes a memory 902, a processor 904, and a computer program stored in the memory 902 and executable on the processor 904, and when the processor 904 executes the computer program, the steps of any of the above-mentioned crew fatigue monitoring methods are implemented.

[0147] The electronic device 90 provided by the present invention implements the steps of the above-mentioned crew fatigue monitoring method when the processor 904 executes the computer program, which can achieve the technical effects of any of the above-mentioned embodiments and will not be repeated here.

[0148] An embodiment of the present invention provides a storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of any of the above-mentioned crew fatigue monitoring methods are implemented.

[0149] The storage medium provided by the present invention implements the steps of the above-mentioned crew fatigue monitoring method when the computer program is executed by the processor, and can achieve the technical effects of any of the above-mentioned embodiments, which will not be repeated here.

[0150] In the description of this specification, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance unless otherwise clearly specified and limited; the terms "connection", "installation", "fixation" and the like should be understood in a broad sense, for example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be directly connected or indirectly connected through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0151] In the description of this specification, the description of the terms "one embodiment", "some embodiments", "specific embodiments", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0152] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A crew fatigue monitoring method, characterized in that: include: Acquire first characteristic information of the crew, wherein the first characteristic information includes: behavior information, physiological signals and environmental information; Determine the name information of the crew member; Determining a risk prediction model according to the name information, wherein the risk prediction model is a risk prediction model corrected by using the cognitive bias of the crew member; Determining the probability of a risk event occurring according to the first feature information and the risk prediction model; Determining an early warning strategy based on the probability; Before the step of determining the risk prediction model according to the name information, the crew fatigue monitoring method includes: obtaining historical information of said crew; Acquiring the cognitive bias of the crew member, wherein the cognitive bias includes: the crew member's confidence in successfully avoiding risk events in the past and the crew member's attention to low-probability risk events; Determine a first feature vector according to the historical information; determining a second eigenvector according to the cognitive bias; The risk prediction model is constructed based on the first eigenvector and the second eigenvector.

2. The crew fatigue monitoring method according to claim 1, characterized in that: The step of constructing the risk prediction model according to the first feature vector and the second feature vector comprises: generating a third eigenvector according to the first eigenvector and the second eigenvector; Determining a weight and a bias value of the third eigenvector according to the third eigenvector and a maximized likelihood function; The risk prediction model is constructed according to the weight and the bias value.

3. The crew fatigue monitoring method according to claim 1, characterized in that: The step of obtaining the first characteristic information of the crew member comprises: respectively collecting image information of the crew member, distance information between the crew member and the equipment, physiological signals of the crew member, and environmental information of the ship's navigation; Extracting features from the image information to obtain second feature information; Performing feature extraction on the distance information to obtain third feature information; Extracting features from the physiological signal to obtain fourth feature information; Extracting features from the environmental information to obtain fifth feature information; The first feature information is determined according to the second feature information, the third feature information, the fourth feature information, and the fifth feature information.

4. The crew fatigue monitoring method according to claim 3, characterized in that: The step of determining the first feature information according to the second feature information, the third feature information, the fourth feature information, and the fifth feature information includes: Setting a time window of the same length in each of the second characteristic information, the third characteristic information, the fourth characteristic information, and the fifth characteristic information; Extracting sixth feature information, seventh feature information, eighth feature information, and ninth feature information from the second feature information, the third feature information, the fourth feature information, and the fifth feature information respectively according to the time window; The sixth characteristic information, the seventh characteristic information, the eighth characteristic information and the ninth characteristic information are concatenated to obtain the first characteristic information.

5. The crew fatigue monitoring method according to any one of claims 1 to 4, characterized in that: The step of determining the early warning prompt strategy according to the probability includes: Matching the probability with a preset threshold to obtain a matching result; The early warning prompt strategy is determined according to the matching result.

6. The crew fatigue monitoring method according to claim 5, characterized in that: The step of determining the early warning prompt strategy according to the matching result includes: When the probability is greater than or equal to a first preset threshold, the early warning prompt strategy is to initiate an emergency evacuation procedure; Based on the fact that the probability is less than the first preset threshold and greater than or equal to a second preset threshold, the early warning prompt strategy is to issue a prompt message to the crew, wherein the second preset threshold is less than the first preset threshold; When the probability is less than the second preset threshold, the early warning strategy is to record the risk event.

7. A crew fatigue monitoring system, characterized in that: include: A first acquisition module, the first acquisition module is used to acquire first characteristic information of the crew, wherein the first characteristic information includes: behavior information, physiological signals and environmental information; A first determining module, the first determining module is used to determine the name information of the crew member; a second determination module, the second determination module being used to determine a risk prediction model according to the name information, wherein the risk prediction model is a risk prediction model corrected by using the cognitive bias of the crew member; a third determination module, the third determination module being used to determine the probability of a risk event occurring according to the first feature information and the risk prediction model; A fourth determination module, the fourth determination module is used to determine a warning prompt strategy according to the probability; The crew fatigue monitoring system further comprises: A second acquisition module is used to acquire the historical information of the crew member; A third acquisition module is used to acquire the cognitive bias of the crew member, wherein the cognitive bias includes: the crew member's confidence in successfully avoiding risk events in the past and the crew member's attention to low-probability risk events; a fifth determining module, configured to determine a first feature vector according to the historical information; a sixth determination module, configured to determine a second feature vector according to the cognitive bias; A construction module is used to construct the risk prediction model according to the first feature vector and the second feature vector.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the crew fatigue monitoring method according to any one of claims 1 to 6 are implemented.

9. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the crew fatigue monitoring method according to any one of claims 1 to 6 are implemented.

Images