A detection method, device, equipment, and medium based on a production scenario.

Abstract

This application provides a detection method, apparatus, equipment, and medium based on a production scenario, comprising: acquiring image information in a production scenario; employing a safety detection model to acquire the scenario type of the production scenario, and the type of the operating environment or operation behavior to be detected in the image information under the scenario type, and acquiring a safety threshold corresponding to the type of operating environment or operation behavior under the scenario type based on the type of operating environment or operation behavior and the safety parameters of the production scenario; and responding to the fact that the operating environment or operation behavior to be detected meets the alarm conditions according to the safety threshold, performing corresponding alarm processing. This solves the problems of unnecessary losses such as work stoppages and production shutdowns caused by excessive alarms or larger safety accidents caused by missed alarms in the existing detection process, thereby improving the effectiveness of detection.

Description

Technical Field

[0001] This application relates to the industrial field, and more particularly to a detection method, apparatus, equipment, and medium based on a production scenario. Background Technology

[0002] Industrial production safety is not only related to property safety, but also to personnel safety. A large number of industrial production safety accidents are caused by unsafe and non-compliant behaviors of employees. On the one hand, employee training can be strengthened to reduce production safety accidents; on the other hand, artificial intelligence detection models can be used to issue alerts for unsafe and non-compliant behaviors in industrial production processes, thereby reducing production safety accidents.

[0003] In existing technologies, artificial intelligence is used to detect unsafe and non-compliant behaviors in industrial production processes. When the detection results exceed the default safety limits, it indicates that unsafe and non-compliant behaviors exist in industrial production. In this case, alarm devices can be activated to issue alerts or the relevant production lines can be shut down to prevent accidents from occurring.

[0004] However, in the process of using artificial intelligence for detection, regardless of the industrial production scenario, the safety regulations are set based on human experience. This can lead to over-alarms or under-alarms during the detection process, resulting in unnecessary losses such as work stoppages and production shutdowns due to alarms, or even greater safety accidents due to under-alarms. Summary of the Invention

[0005] This application provides a detection method, device, equipment, and medium based on a production scenario, which addresses the problems of unnecessary losses such as work stoppages and production shutdowns caused by over-alarms during the detection process or larger safety accidents caused by under-alarms due to the use of human experience to set safety regulations.

[0006] Firstly, this application provides a detection method based on a production scenario, comprising:

[0007] Acquire image information in a production setting;

[0008] A security detection model is used to obtain the scene type of the production scene, and the type of the operation environment or operation behavior to be detected in the image information under the scene type. Based on the type of operation environment or operation behavior and the security parameters of the production scene, the security threshold corresponding to the type of operation environment or operation behavior under the scene type is obtained.

[0009] In response to the detection of an operating environment or operating behavior meeting alarm conditions according to the safety threshold, corresponding alarm processing is performed.

[0010] In one specific implementation, it further includes:

[0011] Based on the scenario type of the production scenario, the operating environment or operating behavior under the production scenario is classified into types, and based on the type of the operating environment or operating behavior and the safety parameters of the production scenario, the safety threshold corresponding to the type of the operating environment or operating behavior under the scenario type is obtained;

[0012] The security detection model to be trained is trained according to the security threshold to obtain the trained security detection model.

[0013] In one specific implementation, the safety parameters of the production scenario include:

[0014] The weight of personal safety corresponding to the production scenario, the weight of property loss corresponding to the production scenario, the average number of personnel casualties per year under the production scenario, the average annual property loss cost when the safety detection model is not used under the production scenario, and the average annual safety production cost under the production scenario.

[0015] In one specific implementation, obtaining the security threshold corresponding to the type of operating environment or operating behavior under the scenario type based on the type of operating environment or operating behavior and the security parameters of the production scenario includes:

[0016] For standard types, the following formula is used:

[0017] f = λ1 * P + λ2 * e1 / e2

[0018]

[0019] Obtain the security threshold T corresponding to the specified specification type;

[0020] Where f represents the intermediate result; λ1 and λ2 represent the personal safety weight and property loss weight corresponding to the standard type, respectively; P represents the average number of personnel casualties per year in the production scenario; e1 represents the average annual property loss cost when the safety detection model is not used in the production scenario; and e2 represents the average annual safety production cost in the production scenario.

[0021] In one specific implementation, obtaining the security threshold corresponding to the type of operating environment or operating behavior under the scenario type based on the type of operating environment or operating behavior and the security parameters of the production scenario includes:

[0022] For exception types, the following formula is used:

[0023] f = λ3*P + λ4*e1 / e2

[0024]

[0025] Obtain the security threshold T corresponding to the anomaly type;

[0026] Where f represents the intermediate result; λ3 and λ4 represent the personal safety weight and property loss weight corresponding to the anomaly type, respectively; P represents the average number of casualties per year in the production scenario; e1 represents the average annual property loss cost when the safety detection model is not used in the production scenario; and e2 represents the average annual safety production cost in the production scenario.

[0027] In one specific implementation, training the security detection model to be trained based on the security threshold to obtain the trained security detection model includes:

[0028] Based on the security threshold, the security detection model to be trained is trained, and the probability value is calculated based on the softmax layer of the security detection model to be trained.

[0029] Based on the probability value, the safety threshold, the weight, and the true value, calculate and obtain the loss function value of the safety threshold;

[0030] If the loss function value is not within the preset range, the security detection model to be trained is trained repeatedly according to the security threshold until the obtained loss function value is within the preset range.

[0031] In one specific implementation, the alarm processing includes:

[0032] Control the alarm device to issue an alarm; or,

[0033] This triggered a shutdown of the relevant production lines.

[0034] Secondly, this application provides a detection device based on a production scenario, comprising:

[0035] The acquisition module is used to acquire image information in production scenarios;

[0036] The processing module is used to use a security detection model to obtain the scene type of the production scene, and the type of the operation environment or operation behavior to be detected in the image information under the scene type, and to obtain the security threshold corresponding to the type of operation environment or operation behavior under the scene type based on the type of operation environment or operation behavior and the security parameters of the production scene.

[0037] An alarm module is used to perform corresponding alarm processing in response to the detection environment or operation behavior meeting the alarm conditions according to the safety threshold.

[0038] In one specific embodiment, the processing module is further configured to:

[0039] Based on the scenario type of the production scenario, the operating environment or operating behavior under the production scenario is classified into types, and based on the type of the operating environment or operating behavior and the safety parameters of the production scenario, the safety threshold corresponding to the type of the operating environment or operating behavior under the scenario type is obtained;

[0040] The security detection model to be trained is trained according to the security threshold to obtain the trained security detection model.

[0041] In one specific implementation, the safety parameters of the production scenario include:

[0042] The weight of personal safety corresponding to the production scenario, the weight of property loss corresponding to the production scenario, the average number of personnel casualties per year in the production scenario, the average annual property loss cost when the safety detection model is not used in the production scenario, and the average annual safety production cost in the production scenario.

[0043] In one specific embodiment, the processing module is specifically used for:

[0044] For standard types, the following formula is used:

[0045] f = λ1 * P + λ2 * e1 / e2

[0046]

[0047] Obtain the security threshold T corresponding to the specified specification type;

[0048] Where f represents the intermediate result; λ1 and λ2 represent the personal safety weight and property loss weight corresponding to the standard type, respectively; P represents the average number of personnel casualties per year in the production scenario; e1 represents the average annual property loss cost when the safety detection model is not used in the production scenario; and e2 represents the average annual safety production cost in the production scenario.

[0049] In one specific embodiment, the processing module is specifically used for:

[0050] For exception types, the following formula is used:

[0051] f = λ3*P + λ4*e1 / e2

[0052]

[0053] Obtain the security threshold T corresponding to the anomaly type;

[0054] Where f represents the intermediate result; λ3 and λ4 represent the personal safety weight and property loss weight corresponding to the anomaly type, respectively; P represents the average number of casualties per year in the production scenario; e1 represents the average annual property loss cost when the safety detection model is not used in the production scenario; and e2 represents the average annual safety production cost in the production scenario.

[0055] In one specific embodiment, the processing module is specifically used for:

[0056] Based on the security threshold, the security detection model to be trained is trained, and the probability value is calculated based on the softmax layer of the security detection model to be trained.

[0057] Based on the probability value, the safety threshold, the weight, and the true value, calculate and obtain the loss function value of the safety threshold;

[0058] If the loss function value is not within the preset range, the security detection model to be trained is trained repeatedly according to the security threshold until the obtained loss function value is within the preset range.

[0059] In one specific embodiment, the alarm module is specifically used for:

[0060] Control the alarm device to issue an alarm; or,

[0061] This triggered a shutdown of the relevant production lines.

[0062] Thirdly, this application discloses an electronic device comprising:

[0063] Processor, memory, communication interface;

[0064] The memory is used to store executable instructions that can be executed by the processor;

[0065] The processor is configured to execute the production scenario-based detection method described in the first aspect by executing the executable instructions.

[0066] Fourthly, this application provides a readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the production scenario-based detection method described in the first aspect.

[0067] This application provides a detection method, apparatus, equipment, and medium based on a production scenario. It acquires image information from the production scenario and uses a safety detection model to determine the scenario type and the type of operating environment or behavior to be detected within the image information under that scenario type. Based on the type of operating environment or behavior and the safety parameters of the production scenario, it obtains the safety threshold corresponding to that type of operating environment or behavior. In response to the detection of the operating environment or behavior meeting alarm conditions according to the safety threshold, corresponding alarm processing is performed. Compared to existing technologies where safety regulations for any production scenario are set based on human experience, the safety detection model used in this application considers different scenario types during processing and detects and judges the operating environment or behavior to be detected based on the safety thresholds corresponding to different types of operating environments or behaviors under each scenario type. This solves the problems of unnecessary losses such as work stoppages and production shutdowns caused by excessive alarms or larger safety accidents caused by missed alarms in existing technologies, thereby improving the effectiveness of detection. Attached Figure Description

[0068] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0069] Figure 1 A system structure diagram of a detection method based on a production scenario provided in this application;

[0070] Figure 2 A flowchart illustrating an embodiment of a detection method based on a production scenario provided in this application;

[0071] Figure 3 A schematic flowchart of a second embodiment of a detection method based on a production scenario provided in this application;

[0072] Figure 4 This is a flowchart of a specific implementation method for obtaining the safety threshold corresponding to the type of operating environment or operating behavior under the scenario type in step S301 based on the type of operating environment or operating behavior and the safety parameters of the production scenario.

[0073] Figure 5 A flowchart illustrating a third embodiment of a detection method based on a production scenario provided in this application;

[0074] Figure 6A schematic diagram of the structure of an embodiment of a detection device based on a production scenario provided in this application;

[0075] Figure 7 This is a schematic diagram of the structure of an electronic device provided in this application. Detailed Implementation

[0076] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments made by those skilled in the art under the guidance of these embodiments are within the scope of protection of this application.

[0077] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0078] In industrial production, a large number of safety accidents are caused by employees' non-standard operating procedures or hidden dangers in the environment. Therefore, safety inspection is of paramount importance. Currently, with the development and improvement of artificial intelligence, safety inspection is gradually adopting AI methods to replace manual inspection. However, the inspection process does not consider the changes in the industrial production scenario and uses default safety regulations for judgment. This leads to unnecessary losses such as work stoppages due to over-alarms or larger safety accidents due to under-alarms. Based on this technical problem, the technical concept of this application is to improve the effectiveness of safety inspection by adapting the inspection judgment to different production scenarios.

[0079] Specifically, Figure 1 The system architecture diagram of a detection method based on a production scenario provided in this application is as follows: Figure 1 As shown, the system may include: an image acquisition device 101, a detection device 102, and an alarm device 103. These three devices can be configured independently or integrated into a single device.

[0080] For example, the system can be installed in various industrial production scenarios that require safety testing, such as factories used to produce fireworks and firecrackers.

[0081] It should be noted that, Figure 1 This is merely a schematic diagram of the system structure of a detection method based on a production scenario provided in this application embodiment. This application embodiment does not represent... Figure 1 The document does not limit the actual form of the various devices included, nor does it specify the form of the devices. Figure 1 The interaction methods between devices are limited, and can be set according to actual needs in the specific application of the solution.

[0082] The technical solution of this application will now be described in detail through specific embodiments. It should be noted that the following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0083] Figure 2 This is a schematic flowchart illustrating an embodiment of a detection method based on a production scenario provided in this application. See also... Figure 2 The detection method based on the production scenario specifically includes the following steps:

[0084] Step S201: Obtain image information in the production scene.

[0085] In this embodiment, in conjunction with the above Figure 1 The system structure shown allows image acquisition devices 101 to be installed inside industrial production facilities, enabling comprehensive monitoring of the entire facility through the cooperation of these devices. Then, detection device 102 can effectively acquire various image information from the production scene using the image acquisition devices 101.

[0086] Step S202: Using a security detection model, obtain the scene type of the production scene, and the type of the operating environment or operating behavior to be detected in the image information under the production scene type. Based on the type of operating environment or operating behavior and the security parameters of the production scene, obtain the security threshold corresponding to the type of operating environment or operating behavior under the scene type.

[0087] In this embodiment, industrial production facilities can be categorized into safe and compliant scenarios based on production safety requirements. For example, fireworks and firecracker factories have high safety requirements, so their scenario type is a safe scenario; flour mills have lower safety requirements, so their scenario type is a compliant scenario.

[0088] Furthermore, both the operating environment and the operational behaviors to be tested are closely related to the safety of the production scenario. For example, if the production scenario is a safety scenario, meaning that alarms can be triggered but not missed, the operating environment to be tested is related to water accumulation, fire, area intrusion, smoke, and equipment malfunctions. The operational behaviors to be tested are related to workers wearing protective equipment and compliance with high-risk operations. Protective equipment can refer to safety helmets, reflective vests, anti-static caps, and insulated shoes. Conversely, if the production scenario is a compliance scenario, meaning that alarms can be missed but not triggered, the operational behaviors to be tested are related to on-duty status, operating procedures, smoking, and using mobile phones. The operating environment to be tested is related to human-machine safety and the separation of personnel and vehicles.

[0089] More specifically, if the detection device 102 acquires a production scene type of a safe scene, then under this scene type, if the image information reveals operational environments related to water accumulation, fire, area intrusion, smoke, and equipment malfunctions, its type is determined to be an abnormal type; if the image information reveals operational behaviors related to workers wearing protective equipment and compliance with high-risk operations, its type is determined to be a compliant type. Furthermore, if the detection device 102 acquires a production scene type of a compliant scene, then under this scene type, if the image information reveals operational behaviors related to on-duty status and operational procedures, or operational environments related to human-machine safety, its type is determined to be a compliant type; if the image information reveals operational behaviors related to smoking or using mobile phones, or operational environments related to human-machine separation, its type is determined to be an abnormal type.

[0090] After determining the type of operating environment or behavior to be detected, the corresponding safety threshold can be obtained based on that type and the safety parameters of the production scenario. Optionally, the safety parameters of the production scenario include: the personal safety weight corresponding to the production scenario, the property loss weight corresponding to the production scenario, the average number of personnel injuries and fatalities per year in the production scenario, the average annual property loss cost when the safety detection model is not used in the production scenario, and the average annual safety production cost in the production scenario.

[0091] Step S203: In response to the alarm conditions being met by the operating environment or operating behavior to be detected according to the safety threshold, the corresponding alarm processing is performed.

[0092] In this embodiment, if it is determined that the operating environment or operating behavior to be detected obtained from the image information meets the alarm conditions, the alarm device 103 can perform corresponding alarm processing. Specifically, the alarm device can be controlled to issue an alarm; or, the relevant production line can be triggered to stop.

[0093] In this embodiment, image information from a production scenario is acquired, and a safety detection model is used to determine the scenario type and the type of operational environment or behavior to be detected within the image information. Based on the type of operational environment or behavior and the safety parameters of the production scenario, a safety threshold corresponding to that type of operational environment or behavior is obtained. If the operational environment or behavior to be detected meets the alarm conditions according to the safety threshold, corresponding alarm processing is performed. Compared to existing technologies where safety regulations for any production scenario are set based on human experience, the safety detection model used in this application considers different scenario types during processing. Based on different types of operational environments or behaviors within each scenario type, and the safety parameters of that production scenario, a corresponding safety threshold is obtained. This threshold is then used to detect and judge the operational environment or behavior to be detected, thus solving the problems of unnecessary losses such as work stoppages due to excessive alarms or larger safety accidents due to missed alarms in existing technologies, thereby improving the effectiveness of detection.

[0094] Figure 3 This is a flowchart illustrating a second embodiment of a detection method based on a production scenario provided in this application. In the above... Figure 2 Based on the illustrated embodiment, see also Figure 3 This production-scenario-based detection method can further include:

[0095] Step S301: Based on the scenario type of the production scenario, classify the operating environment or operating behavior in the production scenario, and obtain the safety threshold corresponding to the type of operating environment or operating behavior under the scenario type based on the type of operating environment or operating behavior and the safety parameters of the production scenario.

[0096] In this example, during the training of the security detection model, it is still necessary to consider the scenario type of different production scenarios, the type of operating environment or operation line under the scenario type, and the security parameters of the production scenario in order to obtain the security threshold.

[0097] Optionally, Figure 4 This is a flowchart illustrating a specific implementation method for obtaining the security threshold corresponding to the type of operating environment or operating behavior under a scenario type in step S301, based on the type of operating environment or operating behavior and the security parameters of the production scenario. Figure 4 As shown, the specific implementation method is as follows:

[0098] Step S401: Determine the type of operating environment or operating behavior.

[0099] If the type of operating environment or operating behavior is a standard type, proceed to step S402; if the type of operating environment or operating behavior is an abnormal type, proceed to step S403.

[0100] Step S402: Obtain the security threshold corresponding to the specification type.

[0101] The following formulas (1) and (2) are used:

[0102] f=λ1*P+λ2*e1 / e2 (1)

[0103]

[0104] Obtain its corresponding security threshold T;

[0105] Where λ1 and λ2 represent the personal safety weight and property loss weight corresponding to the standard type, respectively; P represents the average number of personnel injuries and deaths per year in the production scenario; e1 represents the average annual property loss cost when no safety detection model is used in the production scenario; and e2 represents the average annual safety production cost in the production scenario.

[0106] Optionally, λ1 and λ2 can each be 0.5, and λ1 and λ2 can be any value between 0 and 1. This application does not limit the specific values ​​of λ1 and λ2. It is understood that the values ​​of λ1 and λ2 can be determined based on historical data and the manufacturer's emphasis on personal and property safety.

[0107] Step S403: Obtain the security threshold corresponding to the anomaly type.

[0108] The following formulas (3) and (4) are used:

[0109] f=λ3*P+λ4*e1 / e2 (3)

[0110]

[0111] Obtain its corresponding security threshold T;

[0112] Wherein, λ3 and λ4 represent the personal safety weight and property loss weight corresponding to the production scenario, respectively; P represents the average number of personnel injuries and deaths per year in the production scenario; e1 represents the average annual property loss cost when no safety detection model is used in the production scenario; and e2 represents the average annual safety production cost in the production scenario.

[0113] Optionally, λ3 and λ4 can each be 0.7, and λ3 and λ4 can be any value between 0 and 1. This application does not limit the specific values ​​of λ3 and λ4. It is understood that the values ​​of λ3 and λ4 can be determined based on historical data and the manufacturer's emphasis on personal and property safety.

[0114] Step S302: Train the security detection model to be trained according to the security threshold to obtain the trained security detection model.

[0115] In this embodiment, the security detection model used in step S202 can be either the security detection model obtained after training based on steps S301 and S302 at the initial stage, or an updated security detection model that has been retrained based on steps S301 and S302 after a period of detection.

[0116] In this embodiment, safety thresholds corresponding to different operational behaviors or operating environments under different scenario types can be obtained through safety parameters such as personal safety weight and property loss weight. The safety detection model is then trained based on these safety thresholds, enabling the model to have different alarm sensitivities in different application scenarios, thus providing users with a better experience. Simultaneously, it can further reduce the time technical personnel spend on-site testing, thereby improving production efficiency.

[0117] Figure 5 This is a flowchart illustrating a third embodiment of a detection method based on a production scenario provided in this application. In the above... Figure 3 Based on the illustrated embodiment, see also Figure 5 One specific implementation of step S302 is as follows:

[0118] Step S501: Train the security detection model to be trained according to the security threshold, and calculate and obtain the probability value based on the softmax layer of the security detection model to be trained.

[0119] Step S502: Calculate the loss function value for obtaining the safety threshold.

[0120] Using formulas (5) and (6):

[0121]

[0122] LossT=λ*(RR g (6)

[0123] Calculate the loss function value LossT for obtaining the safety threshold;

[0124] Where prob represents the probability value; T represents the safety threshold; Rg λ represents the truth value, which is the expected output result corresponding to the operation environment or operation behavior to be detected in the image information; λ represents the weight, which is the weight value determined based on the difference between the actual output result obtained from the previous training of the security detection model and the expected output result.

[0125] Step S503: Determine whether the loss function value is within the preset range. If not, repeat step S501; if it is, the security detection model training is complete.

[0126] Understandably, probability values ​​can be calculated using other network layers, and this application does not limit the method of obtaining probability values.

[0127] This embodiment trains the security detection model to be trained based on a security threshold. It calculates probability values ​​using the softmax layer of the security detection model, and then calculates a loss function value for the security threshold based on the probability values, the security threshold, weights, and the ground truth. If the loss function value is outside a preset range, the training process is repeated using the security threshold until the obtained loss function value falls within the preset range, at which point the security detection model training is complete. The security detection model in this application includes a softmax layer, making the output of the security detection model a probability value. The loss function value for the security threshold is calculated based on the probability values, the security threshold, weights, and the ground truth. The loss function is compared with a preset range, and the comparison result determines whether the model training is complete. This achieves automatic training of the security detection model based on the security threshold, eliminating the need for long-term on-site modification of the security detection model's parameters and accelerating the deployment and use of the security detection model.

[0128] The following are embodiments of the apparatus described in this application, which can be used to execute the embodiments of the method described in this application. For details not disclosed in the apparatus embodiments of this application, please refer to the embodiments of the method described in this application.

[0129] Figure 6 This application provides a schematic diagram of the structure of an embodiment of a detection device based on a production scenario; as shown below. Figure 6 As shown, the production scenario-based detection device 60 includes: an acquisition module 61, a processing module 62, and an alarm module 63; wherein, the acquisition module 61 is used to acquire image information in the production scenario; the processing module 62 is used to use a safety detection model to acquire the scenario type of the production scenario, and the type of the operating environment or operating behavior to be detected in the image information under the scenario type, and to acquire the safety threshold corresponding to the type of operating environment or operating behavior under the scenario type according to the type of operating environment or operating behavior and the safety parameters of the production scenario; the alarm module 63 is used to perform corresponding alarm processing in response to the fact that the operating environment or operating behavior to be detected meets the alarm conditions according to the safety threshold.

[0130] The detection device based on the production scenario provided in this application embodiment can execute the technical solution shown in the above method embodiment. Its implementation principle and beneficial effects are similar, and will not be repeated here.

[0131] In one possible implementation, the processing module 62 is further configured to: classify the operating environment or operating behavior in the production scenario according to the scenario type of the production scenario, and obtain the security threshold corresponding to the type of operating environment or operating behavior under the scenario type according to the type of operating environment or operating behavior and the security parameters of the production scenario; and train the security detection model to be trained according to the security threshold to obtain the trained security detection model.

[0132] Optionally, the safety parameters of the production scenario include: the personal safety weight corresponding to the production scenario, the property loss weight corresponding to the production scenario, the average number of personnel injuries and deaths per year in the production scenario, the average annual property loss cost when no safety detection model is used in the production scenario, and the average annual safety production cost in the production scenario.

[0133] Specifically, in one possible implementation, for specification types, processing module 62 is specifically used for:

[0134] The following formulas (1) and (2) are used:

[0135] f=λ1*P+λ2*e1 / e2 (1)

[0136]

[0137] Obtain the security threshold T corresponding to the specification type;

[0138] Where f represents the intermediate result, λ1 and λ2 represent the personal safety weight and property loss weight corresponding to the standard type, respectively; P represents the average number of personnel injuries and deaths per year in the production scenario; e1 represents the average annual property loss cost when the safety detection model is not used in the production scenario; and e2 represents the average annual safety production cost in the production scenario.

[0139] For exception types, the processing module 62 is also specifically used for:

[0140] The following formulas (3) and (4) are used:

[0141] f=λ3*P+λ4*e1 / e2 (3)

[0142]

[0143] Obtain the security threshold T corresponding to the anomaly type;

[0144] Where f represents the intermediate result, λ3 and λ4 represent the personal safety weight and property loss weight corresponding to the anomaly type, respectively; P represents the average number of personnel casualties per year in the production scenario; e1 represents the average annual property loss cost when no safety detection model is used in the production scenario; and e2 represents the average annual safety production cost in the production scenario.

[0145] The detection device based on the production scenario provided in this application embodiment can execute the technical solution shown in the above method embodiment. Its implementation principle and beneficial effects are similar, and will not be repeated here.

[0146] In one possible implementation, for training the security detection model, the processing module 62 is specifically used to: train the security detection model to be trained according to a security threshold, and calculate and obtain probability values ​​based on the softmax layer of the security detection model to be trained; calculate and obtain the loss function value of the security threshold based on the probability value, security threshold, weights, and ground truth; if the loss function value is not within a preset range, then repeat the training of the security detection model to be trained according to the security threshold until the obtained loss function value is within a preset range.

[0147] In one possible implementation, alarm module 63 is specifically used to: control alarm devices to issue alarms; or, trigger the shutdown of the relevant production line.

[0148] The detection device based on the production scenario provided in this application embodiment can execute the technical solution shown in the above method embodiment. Its implementation principle and beneficial effects are similar, and will not be repeated here.

[0149] Figure 7 This is a schematic diagram of the structure of an electronic device provided in this application. Figure 7 As shown, the electronic device 70 includes: a processor 71, a memory 72, and a communication interface 73; wherein, the memory 72 is used to store executable instructions that can be executed by the processor 71; the processor 71 is configured to execute the technical solutions in any of the foregoing method embodiments by executing the executable instructions.

[0150] Optionally, the memory 72 can be either standalone or integrated with the processor 71.

[0151] Optionally, when the memory 72 is a device independent of the processor 71, the electronic device 70 may further include a bus for connecting the aforementioned devices.

[0152] The electronic device is used to execute the technical solutions in any of the foregoing method embodiments. Its implementation principle and technical effect are similar, and will not be described again here.

[0153] This application also provides a readable storage medium storing a computer program thereon, which, when executed by a processor, implements the technical solutions provided in any of the foregoing embodiments.

[0154] This application also provides a computer program product, including a computer program, which, when executed by a processor, is used to implement the technical solutions provided in any of the foregoing method embodiments.

[0155] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0156] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A detection method based on a production scenario, characterized in that, include: Acquire image information in a production setting; A security detection model is used to obtain the scene type of the production scene, and the type of the operation environment or operation behavior to be detected in the image information under the scene type. Based on the type of operation environment or operation behavior and the security parameters of the production scene, the security threshold corresponding to the type of operation environment or operation behavior under the scene type is obtained. In response to the detection of the operating environment or operating behavior meeting the alarm conditions according to the security threshold, corresponding alarm processing is performed. Also includes: Based on the scenario type of the production scenario, the operating environment or operating behavior under the production scenario is classified into types, and based on the type of the operating environment or operating behavior and the safety parameters of the production scenario, the safety threshold corresponding to the type of the operating environment or operating behavior under the scenario type is obtained; Based on the security threshold, the security detection model to be trained is trained to obtain the trained security detection model; The safety parameters for the production scenario include: The personal safety weight corresponding to the production scenario, the property loss weight corresponding to the production scenario, the average number of personnel casualties per year under the production scenario, the average annual property loss cost when the safety detection model is not used under the production scenario, and the average annual safety production cost under the production scenario. The step of obtaining the security threshold corresponding to the type of operating environment or operating behavior under the scenario type based on the type of operating environment or operating behavior and the security parameters of the production scenario includes: For standard types, the following formula is used: Obtain the security threshold T corresponding to the specified specification type; in, Indicates intermediate results; and These represent the personal safety weight and property loss weight corresponding to the standard type, respectively; P represents the average number of personnel injuries and fatalities per year in the production scenario; e1 represents the average annual property loss cost when the safety detection model is not used in the production scenario; and e2 represents the average annual safety production cost in the production scenario. For exception types, the following formula is used: Obtain the security threshold T corresponding to the anomaly type; in, Indicates intermediate results; and These represent the personal safety weight and property loss weight corresponding to the anomaly type, respectively; P represents the average number of personnel injuries and fatalities per year in the production scenario; e1 represents the average annual property loss cost when the safety detection model is not used in the production scenario; and e2 represents the average annual safety production cost in the production scenario.

2. The detection method based on a production scenario according to claim 1, characterized in that, The step of training the security detection model to be trained according to the security threshold to obtain the trained security detection model includes: Based on the security threshold, the security detection model to be trained is trained, and the probability value is calculated based on the softmax layer of the security detection model to be trained. Based on the probability value, the safety threshold, the weight, and the true value, calculate and obtain the loss function value of the safety threshold; If the loss function value is not within the preset range, the security detection model to be trained is trained repeatedly according to the security threshold until the obtained loss function value is within the preset range.

3. The detection method based on a production scenario according to claim 1, characterized in that, The corresponding alarm handling includes: Control the alarm device to issue an alarm; or, This triggered a shutdown of the relevant production lines.

4. A detection device based on a production scenario, characterized in that, include: The acquisition module is used to acquire image information in production scenarios; The processing module is used to use a security detection model to obtain the scene type of the production scene, and the type of the operation environment or operation behavior to be detected in the image information under the scene type, and to obtain the security threshold corresponding to the type of operation environment or operation behavior under the scene type based on the type of operation environment or operation behavior and the security parameters of the production scene. The alarm module is used to perform corresponding alarm processing in response to the fact that the operating environment or operating behavior to be detected meets the alarm conditions according to the safety threshold. The processing module is also used for: Based on the scenario type of the production scenario, the operating environment or operating behavior under the production scenario is classified into types, and based on the type of the operating environment or operating behavior and the safety parameters of the production scenario, the safety threshold corresponding to the type of the operating environment or operating behavior under the scenario type is obtained; Based on the security threshold, the security detection model to be trained is trained to obtain the trained security detection model; The safety parameters for the production scenario include: The personal safety weight corresponding to the production scenario, the property loss weight corresponding to the production scenario, the average number of personnel casualties per year in the production scenario, the average annual property loss cost when the safety detection model is not used in the production scenario, and the average annual safety production cost in the production scenario. The processing module is specifically used for: For standard types, the following formula is used: Obtain the security threshold T corresponding to the specified specification type; in, Indicates intermediate results; and These represent the personal safety weight and property loss weight corresponding to the standard type, respectively; P represents the average number of personnel injuries and fatalities per year in the production scenario; e1 represents the average annual property loss cost when the safety detection model is not used in the production scenario; and e2 represents the average annual safety production cost in the production scenario. For exception types, the following formula is used: Obtain the security threshold T corresponding to the anomaly type; in, Indicates intermediate results; and These represent the personal safety weight and property loss weight corresponding to the anomaly type, respectively; P represents the average number of personnel injuries and fatalities per year in the production scenario; e1 represents the average annual property loss cost when the safety detection model is not used in the production scenario; and e2 represents the average annual safety production cost in the production scenario.

5. The detection device based on a production scenario according to claim 4, characterized in that, The processing module is specifically used for: Based on the security threshold, the security detection model to be trained is trained, and the probability value is calculated based on the softmax layer of the security detection model to be trained. Based on the probability value, the safety threshold, the weight, and the true value, calculate and obtain the loss function value of the safety threshold; If the loss function value is not within the preset range, the security detection model to be trained is trained repeatedly according to the security threshold until the obtained loss function value is within the preset range.

6. The detection device based on a production scenario according to claim 4, characterized in that, The alarm module is specifically used for: Control the alarm device to issue an alarm; or, This triggered a shutdown of the relevant production lines.

7. An electronic device, characterized in that, include: Processor, memory, communication interface; The memory is used to store executable instructions that can be executed by the processor; The processor is configured to execute the detection method based on any one of claims 1 to 3 by executing the executable instructions.

8. A readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the detection method based on the production scenario as described in any one of claims 1 to 3.

Images