Dangerous area identification method and device based on machine vision

Abstract

The invention provides a dangerous area identification method and device based on machine vision, and relates to the technical field of dangerous area identification, and the method comprises the steps: obtaining the monitoring image data of a dangerous source in a historical monitoring area, forming a training image set through preprocessing, constructing a danger identification model, and training the danger identification model. Performing danger source identification on the acquired real-time monitoring image according to the trained danger identification model, delimiting a basic danger area by setting a basic safety distance according to an identification result, calculating a danger coefficient of each type according to historical data of a static area of each type, increasing an early warning distance, and performing early warning according to the early warning distance. According to historical data of each type of dynamic equipment, a dynamic danger coefficient is calculated, a buffer distance before a moving direction is calculated, historical dangerous accident data of a to-be-identified area is acquired, an equipment feature set is extracted, and a Bayesian prediction model is established, so that a future safety state of a sub-area is identified.

Description

Technical Field

[0001] The present invention relates to the technical field of dangerous area identification, and in particular to a dangerous area identification method and device based on machine vision. Background Art

[0002] With the rapid development of industrial automation and intelligent manufacturing, the production process in the factory is becoming more complex and running at high speed. There are many types of processing machinery, more and more frequent handling equipment, and more complex working environments. There are more and more dangerous areas in the factory, which increases the risk of safety accidents. Therefore, accurate identification of dangerous areas is a key factor in ensuring the safety of workers, efficient operation of equipment, and smooth progress of production processes.

[0003] At present, the technology for identifying dangerous areas in factories mainly relies on traditional safety monitoring methods, such as manual inspections and fixed surveillance cameras, which often have problems such as poor real-time performance, limited coverage, and human factors. In addition, with the development of automation and intelligent technology, there is an urgent need to improve the ability to identify dangerous areas through advanced technical means.

[0004] In the prior art, the publication number CN 114973140 A discloses a method and system for monitoring personnel intrusion into dangerous areas based on machine vision. By acquiring a monitoring image of the target construction area and inputting the monitoring image of the target construction area into a preset dangerous target detection model, it is determined whether there is a dangerous target in the target construction area: if it is determined that there is a dangerous target in the target construction area, the coordinates of the dangerous target are acquired, and the coordinates of the dangerous area are determined according to the coordinates of the dangerous target; the coordinates of the personnel in the monitoring image are identified, and the coordinates of the personnel are compared with the coordinates of the dangerous area to determine whether there is a person in the dangerous area. If it is determined that there is a person in the dangerous area, an alarm signal is sent to the user. However, this solution is only for static areas. In the process of time production, dynamic mobile devices often cause more harm to people, and it also lacks the ability to predict dangerous areas.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0006] The purpose of the present invention is to provide a method and device for identifying dangerous areas based on machine vision to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions: A method for identifying dangerous areas based on machine vision, the specific steps include: Step 1: Establish a spatial coordinate system, divide the area to be identified into sub-areas, install cameras in the area to be identified, and obtain monitoring image data of hazardous sources in the historical monitoring area. The types of hazardous sources include static areas, dynamic equipment, and workers. Form a training image set through preprocessing, build a hazard identification model, and train the hazard identification model; Step 2: Obtain real-time monitoring images, identify the hazard source through the trained hazard identification model, obtain the coordinate position and area size of the hazard source based on the identification results, delineate the basic hazard area by setting the basic safety distance, and alarm the dynamic equipment and staff that enter the basic hazard area; Step 3: Based on the historical data of each type of static area, obtain the number of times danger occurs in each type of static area, the area size, and the existence time of the dangerous area, calculate the danger coefficient of each type, and increase the warning distance according to the basic safety distance and the danger coefficient; Step 4: Based on the historical data of each type of dynamic equipment, through continuous identification of dynamic equipment, obtain the historical moving speed, moving distance and area size of the dynamic equipment, calculate the dynamic risk coefficient, and calculate the buffer distance before the moving direction based on the basic safety distance and the dynamic risk coefficient; Step 5: Obtain historical dangerous accident data of the area to be identified, including the number of dangerous accidents and the coordinates of the center of dangerous accidents. Obtain the operating time, dynamic equipment status, coordinate position of the dynamic equipment identified in the historical data, and the coordinate position and working hours of the equipment operator to form an equipment feature set. By establishing a Bayesian prediction model, the future safety status of the sub-area is identified.

[0008] Furthermore, the static area includes an oil stain area, a water stain area, a maintenance area, and a heat source area; The preprocessing method is: select the danger source in the monitoring image through a bounding box, the bounding box includes the coordinate position and the area size, and then input label information into the bounding box, the label information includes the type of danger source and the state of the danger source; The hazard source status includes dynamic equipment status and staff status. The dynamic equipment status includes operation, failure, and shutdown. The staff status includes working and non-working.

[0009] Furthermore, the hazard identification model is based on a CNN neural network, specifically including a convolutional layer, an activation layer, a pooling layer and an output layer; Convolutional layer: extracts features from the bounding box and label of the input image through convolution operation, and the calculation formula is: in, is the input matrix, is the convolution kernel, are the coordinates of the output matrix, The convolution kernel is Row and The value of the column; Activate layer: in, The coordinates of the output matrix of the convolutional layer output; Pooling layer: in, is the input matrix, is the maximum pooling output; Output layer: in, is the output of the pooling layer, is the result of image recognition, is the weight, is the bias parameter.

[0010] Furthermore, the specific method for demarcating the basic danger area is: When a static area is identified, the basic danger area is delineated according to the coordinate position and area size of the static area and marked as a danger area; When a dynamic device is identified, the basic danger zone is delineated and marked as a danger zone according to the coordinate position, area size and status of the dynamic device, and when the dynamic device status is in operation or failure.

[0011] Furthermore, the calculation step of increasing the warning distance according to the basic safety distance and the risk factor is: The calculation formula for calculating the risk factor of each type of static area is: in, is the hazard factor for each type of static area, is the number of times the danger occurs, For the The size of the area, For the The duration of time the danger zone exists.

[0012] The calculation formula of the warning distance is: in, For the warning distance, As the basic safety distance, Warning distance adjustment factor, , The hazard factor for each type of static area.

[0013] Furthermore, the calculation methods of the historical moving speed and moving distance are respectively: in, For the The moving distance, For the The moving speed of times, For the Before the move The coordinate position at the time, For the Before the move The coordinate position at the moment.

[0014] Furthermore, the specific steps of calculating the buffer distance before the moving direction according to the basic safety distance and the dynamic risk coefficient are: The calculation formula of dynamic risk factor is: in, is the dynamic risk factor, For the The moving speed of For the The moving distance, The area size of the dynamic device.

[0015] The calculation formula for the buffer distance is: in, is the buffer distance, is the dynamic distance adjustment coefficient, .

[0016] Furthermore, the specific steps of establishing the Bayesian-based comprehensive recognition model are: The safety status of each sub-area is calculated based on the coordinates of the dangerous events in the historical data, that is, the prior probability of whether a dangerous event occurs: in, The security status of the sub-area The prior probability of is the number of dangerous events in historical data, is the indicative function, For the historical data Type of security status, , is the security status of the sub-area in the historical data, Indicates the danger of a dangerous incident. It means no dangerous incident occurred. is the smoothing coefficient; The formula of the indicative function is: ; Calculate the likelihood probability of the device feature set in each safety state: in, For a given security status The device feature set is observed below The probability of For a given security status The device characteristics are observed below The probability of is the number of equipment feature types; The posterior probability of the safe state of the sub-region is calculated according to Bayes' theorem: in, For safe status The posterior probability of is a normalizing constant; Establish a security status decision model: in, The final identified safety status.

[0017] The present invention also provides a dangerous area identification device based on machine vision, and the dangerous area identification device based on machine vision is used to execute the above-mentioned dangerous area identification method based on machine vision, including: An image acquisition module is used to acquire real-time monitoring images of the area to be identified, establish a spatial coordinate system, and obtain the number of monitoring images of the types of dangerous sources in the historical monitoring area. The types of dangerous sources include static areas, dynamic equipment, and staff, and form a training image set through preprocessing; A hazard identification module, which is used to construct a hazard identification model and train the hazard identification model. The hazard source in the real-time monitoring image is identified by the trained hazard identification model. According to the identification result, the coordinate position and area size corresponding to the hazard source are obtained. The basic hazard area is delineated by setting a basic safety distance, and an alarm is issued to dynamic equipment and staff that enter the basic hazard area. A safety warning module is used to obtain the number of times danger occurs in each static area, the area size, and the existence time of the dangerous area according to the historical data of each type of static area, calculate the danger coefficient of each type, and increase the warning distance according to the basic safety distance and the danger coefficient; A dynamic warning module, which is used to obtain the historical moving speed, moving distance and area size of the dynamic device according to the historical data of each type of dynamic device through continuous identification of the dynamic device, calculate the dynamic risk coefficient, and calculate the buffer distance in the moving direction according to the basic safety distance and the current moving speed; The regional prediction module is used to obtain historical dangerous accident data of the area to be identified, including the number of dangerous accidents, the coordinates of the center location of dangerous accidents, the operating time, operating status, coordinate position of the dynamic equipment identified in the historical data, and the coordinate position and working hours of the equipment operator to form an equipment feature set, and identify the area where dangerous events will occur in the future by establishing a Bayesian prediction model.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention obtains monitoring image data of hazardous sources in the historical monitoring area, forms a training image set through preprocessing, constructs a hazard identification model, and trains the hazard identification model; the real-time monitoring images are obtained to identify hazardous sources according to the trained hazard identification model; according to the identification results, the basic hazardous area is delineated by setting a basic safety distance; according to the historical data of each type of static area, the hazard coefficient of each type is calculated to increase the warning distance; according to the historical data of each type of dynamic equipment, the dynamic hazard coefficient is calculated; the buffer distance before the moving direction is calculated; by obtaining the historical hazardous accident data of the area to be identified, the equipment feature set is extracted; and by establishing a Bayesian prediction model, the future safety status of the sub-area is identified.

[0019] The present invention calculates the risk factor of each type according to the historical data of each type of static area and increases the warning distance, which can provide earlier warnings for personnel and reduce the probability of accidents. The present invention calculates the risk factor of each type according to the historical data of each type of static area and increases the warning distance, which can not only effectively identify the dangers in the area where the dynamic equipment is located, but also reasonably increase the buffer distance according to its movement trajectory, which can effectively reduce the probability of accidents. By establishing a Bayesian prediction model, the future safety status of the sub-area is identified, and potential danger areas are effectively and accurately identified. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of the overall method flow of the present invention; Figure 2 It is a schematic diagram of the overall structure of the device of the present invention. DETAILED DESCRIPTION

[0021] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0022] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0023] Example

[0024] See also Figure 1 , the present invention provides a technical solution: A method for identifying dangerous areas based on machine vision, the specific steps include: Step 1: Establish a spatial coordinate system, divide the area to be identified into sub-areas, install cameras in the area to be identified, and obtain monitoring image data of hazardous sources in the historical monitoring area. The types of hazardous sources include static areas, dynamic equipment, and staff. Form a training image set through preprocessing, build a hazard identification model, and train the hazard identification model.

[0025] In the factory area to be identified, the coordinate system is more intuitive and easy to understand, with the center point of the area or the landmark building as the origin, the east-west direction as the horizontal coordinate, the north-south direction as the vertical coordinate, and the center point or landmark equipment as the origin. People are more sensitive and clear about the spatial location of the center point or landmark equipment, which helps workers and managers quickly understand the location of the coordinate point. Defining the coordinate axis by geographic orientation (east, west, south, and north) conforms to people's daily orientation habits and reduces learning and adaptation costs. Such a coordinate system is closer to natural orientation perception, and workers and managers can identify directions more quickly and improve efficiency.

[0026] Classifying the types of hazardous sources into static areas, dynamic equipment, and workers can help identify and manage different safety risks more clearly. For example, oily and water-stained areas in static areas are not only prone to slip and fall accidents, increasing the risk of injury in the workplace, but also interfere with the normal use of equipment. There are scattered tools and equipment in the maintenance area, which leads to the risk of tripping or scratching, and may also be exposed to electrical or chemical hazards. Heat source areas may cause burns or fire risks, especially when close to flammable materials. Therefore, these areas need to be identified for hazards. Dynamic equipment such as conveyor belts, robots, and transport vehicles may cause collisions, pinching, and other injuries due to failures, misoperation, or improper maintenance. Operators need to pay special attention to the interaction with the equipment. The risks of human factors faced by workers include operational errors, lack of safety training, psychological pressure, etc., and attention should be paid to the behavior and safety awareness of personnel to prevent accidents. By classifying hazardous sources, safety risks can be identified and assessed more effectively, and by demarcating corresponding hazardous areas, corresponding management and emergency measures can be effectively formulated.

[0027] In this embodiment, the static area includes an oil stain area, a water stain area, a maintenance area, and a heat source area; The preprocessing method is: select the danger source in the monitoring image through a bounding box, the bounding box includes the coordinate position and the area size, and then input label information into the bounding box, the label information includes the type of danger source and the state of the danger source; The hazard source status includes dynamic equipment status and staff status. The dynamic equipment status includes operation, failure, and shutdown. The staff status includes working and non-working.

[0028] CNN neural network is a commonly used algorithm in deep learning image classification and recognition. It can effectively capture local feature information of images through convolution operations and can gradually reduce the dimension of images through pooling operations. Due to the design of convolution operations and pooling layers, CNN can identify the position changes of objects in images, thereby enhancing the ability to detect dangerous areas in dynamic environments. CNN shows stronger feature extraction capabilities and computational efficiency when processing image data, making it an ideal choice for dangerous area identification.

[0029] In this embodiment, the risk identification model is based on a CNN neural network, which specifically includes a convolutional layer, an activation layer, a pooling layer, and an output layer; Convolutional layer: extracts features from the bounding box and label of the input image through convolution operation, and the calculation formula is: in, is the input matrix, is the convolution kernel, are the coordinates of the output matrix, The convolution kernel is Row and The value of the column; Activate layer: in, The coordinates of the output matrix of the convolutional layer output; Pooling layer: in, is the input matrix, is the maximum pooling output; Output layer: in, is the output of the pooling layer, is the result of image recognition, is the weight, is the bias parameter.

[0030] The training image set is divided into two parts, training and testing, in a ratio of 7:3. The image data in the training part is used as the input of the model, and the bounding box and label information in the image is used as the output of the model. After the training is completed, the model is tested and evaluated using the image set in the testing part to evaluate the performance of the model on new data other than the training data it has seen.

[0031] Step 2: Obtain real-time monitoring images, identify the hazard source through the trained hazard identification model, obtain the coordinate position and area size of the hazard source based on the identification results, define the basic hazard area by setting the basic safety distance, and alarm the dynamic equipment and personnel that enter the basic hazard area.

[0032] When the equipment is running, there are moving parts, which may pose a risk of physical injury such as pinching or collision to people nearby. During the fault repair process, maintenance personnel may need to come into close contact with the equipment. If the equipment is not completely powered off or residual energy is not released, there will be safety hazards. At the same time, the fault may cause the control system to fail, causing the equipment to move unexpectedly or stop normal monitoring of dangerous conditions.

[0033] In this embodiment, the specific method of demarcating the basic danger area is: When a static area is identified, the basic danger area is delineated according to the coordinate position and area size of the static area and marked as a danger area.

[0034] When a dynamic device is identified, the basic danger zone is delineated and marked as a danger zone according to the coordinate position, area size and status of the dynamic device, and when the dynamic device status is in operation or failure.

[0035] Step 3: Based on the historical data of each type of static area, obtain the number of times danger occurs in each type of static area, the area size, and the existence time of the dangerous area, calculate the danger factor of each type, and increase the warning distance based on the basic safety distance and the danger factor.

[0036] In this embodiment, the calculation steps for increasing the warning distance according to the basic safety distance and the risk factor are as follows: The calculation formula for calculating the risk factor of each type of static area is: in, is the hazard factor for each type of static area, is the number of times danger occurs, For the The size of the area, For the The duration of time the danger zone exists.

[0037] The hazard coefficient of a static area reflects the risk level of dangerous incidents occurring in each static area. The higher the hazard coefficient, the more likely dangerous incidents will occur in that type of static area. By counting the frequency of dangers occurring in static areas, high-risk areas can be identified. Larger areas of dangerous areas affect the movement of more people and equipment, and are more likely to be affected by normal work. The greater the danger and the longer the static area exists, the higher the potential risk.

[0038] The calculation formula for increasing the warning distance according to the basic safety distance and the risk factor is: in, For the warning distance, As the basic safety distance, Warning distance adjustment factor, , is the hazard factor for each type of static area.

[0039] Reasonably increasing the warning distance can provide personnel with earlier warnings and reduce the probability of accidents. By setting the warning distance adjustment coefficient, the warning distance can be reasonably adjusted according to actual needs and the type of static area to avoid excessive distance, excessive occupation of space and frequent alarms, while avoiding too small a warning distance and inability to effectively warn.

[0040] Step 4: Based on the historical data of each type of dynamic equipment, through continuous identification of dynamic equipment, obtain the historical moving speed, moving distance and area size of the dynamic equipment, calculate the dynamic hazard coefficient, and calculate the buffer distance before the moving direction based on the basic safety distance and the dynamic hazard coefficient.

[0041] The speed of a dynamic device directly affects its relative position to surrounding dynamic devices, static devices, and workers. A higher speed means a shorter reaction time, which increases the risk of potential collisions. The distance a dynamic device moves within a specific period of time can help predict its future position and help determine whether it will enter a dangerous area. At the same time, the size of the dynamic device and the space it occupies determine the extent of its impact on surrounding dynamic devices, static devices, and workers. Larger devices may pose a greater threat to the surrounding area when moving. Therefore, the combination of moving speed, moving distance, and the size of the area of ​​the dynamic device can form a more comprehensive dynamic hazard coefficient, which helps to more accurately assess the risk of the device in a specific environment.

[0042] In this embodiment, the calculation methods of the historical moving speed and moving distance are respectively: in, For the The moving distance, For the The moving speed of times, For the Before the move The coordinate position at the time, For the Before the move The coordinate position at the moment.

[0043] In this embodiment, the specific steps of calculating the buffer distance before the moving direction according to the basic safety distance and the dynamic risk coefficient are: The calculation formula of dynamic risk factor is: in, is the dynamic risk factor, For the The moving speed of times, For the The moving distance, The area size of the dynamic device.

[0044] The calculation formula for the buffer distance is: in, is the buffer distance, is the dynamic distance adjustment coefficient, .

[0045] The dynamic hazard coefficient reflects the degree of danger of dynamic equipment. The higher the dynamic hazard coefficient, the more likely this type of dynamic equipment is to have dangerous incidents. By reasonably setting the buffer distance, when the equipment approaches the dangerous area or affects other equipment, sufficient safety distance can be reserved to protect the safety of workers and dynamic equipment. Reasonable increase of the buffer distance can effectively reduce the probability of accidents. By setting the dynamic distance adjustment coefficient, the buffer distance can be reasonably adjusted according to actual needs and the type of dynamic equipment to avoid excessive distance, excessive occupation of space and frequent alarms, and at the same time avoid the warning distance being too small and unable to effectively perform buffer warnings.

[0046] Step 5: Obtain historical dangerous accident data of the area to be identified, including the number of dangerous accidents and the coordinates of the center of dangerous accidents. Obtain the operating time, dynamic equipment status, coordinate position of the dynamic equipment identified in the historical data, and the coordinate position and working hours of the equipment operator to form an equipment feature set. By establishing a Bayesian prediction model, the future safety status of the sub-area is identified.

[0047] Bayesian models are particularly good at dealing with uncertainty and noise, which is particularly important for dynamic and complex industrial environments. Prior knowledge, such as historical accident data, is used to improve predictions. This means that even when data is insufficient, the model can make more accurate judgments based on existing knowledge, and the Bayesian model can simultaneously consider the impact of multiple variables on the results, reflecting the complex relationship between equipment, operators and the environment. This comprehensiveness helps to more accurately identify future hazardous areas.

[0048] In the plant area, the Bayesian prediction model is used to identify future safety status by integrating historical dangerous accident data with dynamic equipment information. First, the historical dangerous accident data provides prior knowledge about the frequency and location of accidents, including the number of accidents and the coordinates of their center locations. This information can help the model understand which areas have higher risks.

[0049] The data of dynamic equipment's running time, status, coordinate position and operator's working hours constitute the equipment feature set. These features not only reflect the dangers that may be caused by the equipment during operation, but also provide the spatial relationship between the equipment and the operator, thus enhancing the ability to identify dangerous factors.

[0050] Using these features, the Bayesian model calculates the prior probability of an accident and estimates the likelihood of future potential dangers based on the current state of the dynamic equipment and its historical operating data. The model combines the prior probability with the new observation data and uses the Bayesian theorem to update the posterior probability, thereby generating a predicted value of the safe state for each sub-area.

[0051] The advantage of this method is that it can handle uncertainty and combine multiple influencing factors, such as equipment status and operator behavior, to achieve dynamic prediction of dangerous areas. Finally, by outputting the posterior probability of each sub-area, decision makers can intuitively evaluate the safety risks of different areas, thereby realizing the identification of future dangerous areas.

[0052] In this embodiment, the specific steps of establishing the Bayesian-based comprehensive recognition model are: The safety status of each sub-area is calculated based on the coordinates of the dangerous events in the historical data, that is, the prior probability of whether a dangerous event occurs: in, The security status of the sub-area The prior probability of is the number of dangerous events in historical data, is the indicative function, For the historical data Type of security status, , is the security status of the sub-area in the historical data, Indicates the danger of a dangerous incident. It means no dangerous incident occurred. is the smoothing coefficient; The formula of the indicative function is: ; Calculate the likelihood probability of the device feature set in each safety state: in, For a given security status The device feature set is observed below The probability of For a given security status The device characteristics are observed below The probability of is the number of equipment feature types; The posterior probability of the safe state of the sub-region is calculated according to Bayes' theorem: in, For safe status The posterior probability of is a normalizing constant; Establish a security status decision model: in, The final identified safety status.

[0053] See also Figure 2 The present invention also provides a dangerous area identification device based on machine vision, and the dangerous area identification device based on machine vision is used to execute the above-mentioned dangerous area identification method based on machine vision, including: An image acquisition module is used to acquire real-time monitoring images of the area to be identified, establish a spatial coordinate system, and obtain the number of monitoring images of the types of dangerous sources in the historical monitoring area. The types of dangerous sources include static areas, dynamic equipment, and staff. A training image set is formed through preprocessing. A hazard identification module, which is used to construct a hazard identification model and train the hazard identification model. The hazard source in the real-time monitoring image is identified by the trained hazard identification model. According to the identification result, the coordinate position and area size corresponding to the hazard source are obtained. The basic hazard area is delineated by setting a basic safety distance, and an alarm is issued to dynamic equipment and staff that enter the basic hazard area. A safety warning module is used to obtain the number of times danger occurs in each static area, the area size, and the existence time of the dangerous area according to the historical data of each type of static area, calculate the danger coefficient of each type, and increase the warning distance according to the basic safety distance and the danger coefficient; A dynamic warning module, which is used to obtain the historical moving speed, moving distance and area size of the dynamic device according to the historical data of each type of dynamic device through continuous identification of the dynamic device, calculate the dynamic risk coefficient, and calculate the buffer distance in the moving direction according to the basic safety distance and the current moving speed; A regional prediction module, which is used to obtain historical dangerous accident data of the area to be identified, including the number of dangerous accidents and the coordinates of the center location of dangerous accidents, and obtain the operating time, operating status, and coordinate location of the dynamic equipment identified in the historical data, and the coordinate location and working hours of the equipment operator, to form an equipment feature set, and identify the area where dangerous events will occur in the future by establishing a Bayesian prediction model; The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0054] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0055] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0056] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for identifying dangerous areas based on machine vision, characterized in that: The specific steps include: Step 1: Establish a spatial coordinate system, divide the area to be identified into sub-areas, install cameras in the area to be identified, and obtain monitoring image data of hazardous sources in the historical monitoring area. The types of hazardous sources include static areas, dynamic equipment, and workers. Form a training image set through preprocessing, build a hazard identification model, and train the hazard identification model; Step 2: Obtain real-time monitoring images, identify the hazard source through the trained hazard identification model, obtain the coordinate position and area size of the hazard source based on the identification results, delineate the basic hazard area by setting the basic safety distance, and alarm the dynamic equipment and staff that enter the basic hazard area; Step 3: Based on the historical data of each type of static area, obtain the number of times danger occurs in each type of static area, the area size, and the existence time of the dangerous area, calculate the danger coefficient of each type, and increase the warning distance according to the basic safety distance and the danger coefficient; Step 4: Based on the historical data of each type of dynamic equipment, through continuous identification of dynamic equipment, obtain the historical moving speed, moving distance and area size of the dynamic equipment, calculate the dynamic risk coefficient, and calculate the buffer distance before the moving direction based on the basic safety distance and the dynamic risk coefficient; Step 5: Obtain historical dangerous accident data of the area to be identified, including the number of dangerous accidents and the coordinates of the center of dangerous accidents. Obtain the operating time, dynamic equipment status, coordinate position of the dynamic equipment identified in the historical data, and the coordinate position and working hours of the equipment operator to form an equipment feature set. By establishing a Bayesian prediction model, the future safety status of the sub-area is identified.

2. The method for identifying dangerous areas based on machine vision according to claim 1, characterized in that: The static area includes an oil stain area, a water stain area, a maintenance area, and a heat source area; The preprocessing method is: select the danger source in the monitoring image through a bounding box, the bounding box includes the coordinate position and the area size, and then input label information into the bounding box, the label information includes the type of danger source and the state of the danger source; The hazard source status includes dynamic equipment status and staff status. The dynamic equipment status includes operation, failure, and shutdown. The staff status includes working and non-working.

3. The method for identifying dangerous areas based on machine vision according to claim 1, characterized in that: The hazard identification model is based on a CNN neural network, which specifically includes a convolutional layer, an activation layer, a pooling layer, and an output layer; Convolutional layer: extracts features from the bounding box and label of the input image through convolution operation, and the calculation formula is: in, is the input matrix, is the convolution kernel, are the coordinates of the output matrix, The convolution kernel is Row and The value of the column; Activate layer: in, The coordinates of the output matrix of the convolutional layer output; Pooling layer: in, is the input matrix, is the maximum pooling output; Output layer: in, is the output of the pooling layer, is the result of image recognition, is the weight, is the bias parameter.

4. The method for identifying dangerous areas based on machine vision according to claim 1, characterized in that: The specific method for delineating the basic danger area is: When a static area is identified, the basic danger area is delineated according to the coordinate position and area size of the static area and marked as a danger area; When a dynamic device is identified, the basic danger zone is delineated and marked as a danger zone according to the coordinate position, area size and status of the dynamic device, and when the dynamic device status is in operation or failure.

5. The method for identifying dangerous areas based on machine vision according to claim 1, characterized in that: The calculation steps for increasing the warning distance according to the basic safety distance and the risk factor are as follows: The calculation formula for calculating the risk factor of each type of static area is: in, is the hazard factor for each type of static area, is the number of times danger occurs, For the The size of the area, For the How long the secondary danger zone exists; The calculation formula of the warning distance is: in, For the warning distance, As the basic safety distance, Warning distance adjustment factor, , is the hazard factor for each type of static area.

6. The method for identifying dangerous areas based on machine vision according to claim 1, characterized in that: The calculation methods of the historical moving speed and moving distance are respectively: in, For the The moving distance, For the The moving speed of For the Before the move The coordinate position at the time, For the Before the move The coordinate position at the moment.

7. The method for identifying dangerous areas based on machine vision according to claim 6, characterized in that: The specific steps of calculating the buffer distance before the moving direction according to the basic safety distance and the dynamic risk coefficient are as follows: The calculation formula of dynamic risk factor is: in, is the dynamic risk factor, For the The moving speed of times, For the The moving distance, is the area size of the dynamic device; The calculation formula for the buffer distance is: in, is the buffer distance, is the dynamic distance adjustment coefficient, .

8. The method for identifying dangerous areas based on machine vision according to claim 1, characterized in that: The specific steps of establishing the Bayesian-based comprehensive recognition model are: The safety status of each sub-area is calculated based on the coordinates of the dangerous events in the historical data, that is, the prior probability of whether a dangerous event occurs: in, The security status of the sub-area The prior probability of is the number of dangerous events in historical data, is the indicative function, For the historical data Type of security status, , is the security status of the sub-area in the historical data, Indicates the danger of a dangerous incident. It means no dangerous incident occurred. is the smoothing coefficient; The formula of the indicative function is: ; Calculate the likelihood probability of the device feature set in each safety state: in, For a given security status The device feature set is observed below The probability of For a given security status The device characteristics are observed below The probability of is the number of equipment feature types; The posterior probability of the safe state of the sub-region is calculated according to Bayes' theorem: in, For safe status The posterior probability of is a normalizing constant; Establish a security status decision model: in, The final identified safety status.

9. A dangerous area identification device based on machine vision, characterized in that: The dangerous area identification device based on machine vision is used to execute the dangerous area identification method based on machine vision according to any one of claims 1 to 8, comprising: An image acquisition module is used to acquire real-time monitoring images of the area to be identified, establish a spatial coordinate system, and obtain the number of monitoring images of the types of dangerous sources in the historical monitoring area. The types of dangerous sources include static areas, dynamic equipment, and staff, and form a training image set through preprocessing; A hazard identification module, which is used to construct a hazard identification model and train the hazard identification model. The hazard source in the real-time monitoring image is identified by the trained hazard identification model. According to the identification result, the coordinate position and area size corresponding to the hazard source are obtained. The basic hazard area is delineated by setting a basic safety distance, and an alarm is issued to dynamic equipment and staff that enter the basic hazard area. A safety warning module is used to obtain the number of times danger occurs in each static area, the area size, and the existence time of the dangerous area according to the historical data of each type of static area, calculate the danger coefficient of each type, and increase the warning distance according to the basic safety distance and the danger coefficient; A dynamic warning module, which is used to obtain the historical moving speed, moving distance and area size of the dynamic device according to the historical data of each type of dynamic device through continuous identification of the dynamic device, calculate the dynamic risk coefficient, and calculate the buffer distance in the moving direction according to the basic safety distance and the current moving speed; The regional prediction module is used to obtain historical dangerous accident data of the area to be identified, including the number of dangerous accidents, the coordinates of the center location of dangerous accidents, the operating time, operating status, coordinate position of the dynamic equipment identified in the historical data, and the coordinate position and working hours of the equipment operator to form an equipment feature set, and identify the area where dangerous events will occur in the future by establishing a Bayesian prediction model.

Images