Computer vision and ontology-based fire operation safety monitoring method and system

Abstract

The embodiment of the application provides a kind of based on computer vision and ontology's hot work safety monitoring method and system, the method comprises: obtaining the video data of construction site;Video data is identified to target, and the work state information of construction personnel, the personal protective equipment information of construction personnel and the line of sight direction of construction personnel are obtained;Worker activity state identification is carried out to work state information, and the working state of construction personnel is determined;Based on personal protective equipment information, wear state identification is carried out, and the protective equipment wearing information of construction personnel is determined;Based on work state information, the personal protective equipment information of construction personnel, the line of sight direction of construction personnel, working state and protective equipment wearing information, safety compliance reasoning is carried out through the pre-determined hot work safety inspection ontology model, and safety inspection result is obtained;Based on safety inspection result, safety inspection report is generated.The above scheme can improve the accuracy and response timeliness of safety hazard identification.

Description

Technical Field

[0001] This application relates to the technical field of civil engineering, specifically to a method and system for monitoring the safety of hot work operations based on computer vision and ontology. Background Technology

[0002] The construction industry is a typical high-risk industry, and safety hazards at construction sites are characterized by complexity, concealment, and dynamism. Hot work is widespread in critical processes such as rebar cage welding and component bolting. Especially in steel structure construction, the continuous large volume of welding work makes it highly susceptible to serious production safety accidents if hazard identification is not timely or safety management is inadequate. Numerous accident analyses indicate that improper use of personal protective equipment (PPE) is a significant contributing factor to accidents. Therefore, automated compliance checks on worker PPE use during hot work at construction sites are essential.

[0003] Computer vision technology has shown promise in the compliance inspection of personal protective equipment. Related research continues to improve the adaptability and recognition accuracy of visual algorithms in complex construction environments, laying the foundation for technology implementation. However, safety management at construction sites requires more than just identifying object categories and locations; it demands an understanding of semantic information within the work environment to support deeper safety judgments and decisions. Traditional image recognition methods struggle to achieve deep semantic understanding and real-time response to dynamically changing and complex interactive construction scenarios. Summary of the Invention

[0004] This application aims to provide a method and system for hot work safety monitoring based on computer vision and ontology, which can improve the accuracy and response time of safety hazard identification.

[0005] The technical solution of this application is implemented as follows:

[0006] In a first aspect, embodiments of this application provide a method for monitoring the safety of hot work operations based on computer vision and ontology, the method comprising:

[0007] Acquire video data from the construction site; wherein the video data includes construction personnel;

[0008] Target recognition is performed on the video data to obtain the work status information of the construction workers, the personal protective equipment information of the construction workers, and the direction of the construction workers' gaze;

[0009] The work status information is used to identify the worker's activity status and determine the work status of the construction worker; the personal protective equipment information is used to identify the wearing status and determine the protective equipment wearing information of the construction worker.

[0010] Based on the work status information, the personal protective equipment information of the construction personnel, the line of sight of the construction personnel, the work status, and the protective equipment wearing information, safety compliance reasoning is performed through a pre-determined hot work safety inspection ontology model to obtain the safety inspection results;

[0011] Based on the security check results, a security check report is generated.

[0012] In the above scheme, the step of performing target recognition on the video data to obtain the construction worker's work status information, the construction worker's personal protective equipment information, and the construction worker's line of sight direction includes:

[0013] The video data is subjected to target detection and target tracking to determine the target area containing the construction personnel; and the target area is identified to obtain the work status information of the construction personnel.

[0014] The protective equipment in the target area is tested to obtain the personal protective equipment information of the construction personnel.

[0015] Target gaze estimation is performed on the target area to obtain the line of sight direction of the construction worker.

[0016] In the above scheme, the step of identifying the worker's activity status from the work status information to determine the work status of the construction personnel includes:

[0017] By using a gaze target estimation model, the work status information is identified, and the real-time gaze coordinates of the construction personnel are determined.

[0018] The target detection model is used to identify the work status information and determine the bounding box area where the welding area corresponding to the construction worker is located.

[0019] Based on the real-time gaze coordinates and the bounding box region, visual attention constraints and work space constraints are determined to ascertain the working status of the construction personnel.

[0020] In the above scheme, the step of identifying the wearing status based on the personal protective equipment information to determine the protective equipment wearing information of the construction worker includes:

[0021] By using a target detection model, feature extraction is performed on the personal protective equipment information to determine the first bounding box corresponding to the construction worker, the second bounding box corresponding to the head, the third bounding box corresponding to the safety helmet, the fourth bounding box corresponding to the protective gloves, the fifth bounding box corresponding to the welding mask, and the sixth bounding box corresponding to the reflective vest.

[0022] Calculate the intersection-union ratio (IUGR) between the first bounding box and the second, third, fourth, and sixth bounding boxes respectively; and based on the IUGR between the first bounding box and the second, third, and fourth bounding boxes, determine the wearing information of the first sub-protective equipment corresponding to the direct-wear type.

[0023] Based on the fifth bounding box and the real-time gaze coordinate point, the line-of-sight intersection is determined to obtain the line-of-sight intersection result; and based on the line-of-sight intersection result and the intersection-union comparison between the first bounding box and the sixth bounding box, the wearing information of the second sub-protective equipment corresponding to the non-direct wearing type is determined.

[0024] Based on the first sub-protective equipment wearing information and the second sub-protective equipment wearing information, the protective equipment wearing information of the construction personnel is determined.

[0025] In the above scheme, the safety inspection results are obtained by performing safety compliance reasoning based on the work status information, the personal protective equipment information of the construction personnel, the line of sight of the construction personnel, the work status, and the protective equipment wearing information, using a pre-determined hot work safety inspection ontology model. These results include:

[0026] Based on the work status information, the personal protective equipment information of the construction workers, and the direction of the construction workers' line of sight, class instances are determined;

[0027] Based on the working status and the protective equipment wearing information, attribute instances are determined; and an attribute matrix is ​​established based on the attribute instances.

[0028] Based on the class instance and the attribute matrix, the safety inspection result is obtained by performing hierarchical safety compliance reasoning through the hot work safety inspection ontology model.

[0029] In the above scheme, generating a security inspection report based on the security inspection results includes:

[0030] If the security check results indicate a violation, a warning message is generated and played via voice.

[0031] In response to the warning message, obtain images of the site after rectification;

[0032] The rectified on-site images are subjected to safety inspection. If the inspection results indicate that the construction is safe and normal, a warning cancellation message is generated.

[0033] The safety inspection report is generated based on the safety inspection results, the on-site images of the violations corresponding to the safety inspection results, and the on-site images after rectification.

[0034] The above scheme is characterized in that, before obtaining the safety inspection result by performing safety compliance reasoning through a pre-determined hot work safety inspection ontology model based on the work status information, the personal protective equipment information of the construction personnel, the line of sight of the construction personnel, the work status, and the protective equipment wearing information, the method further includes:

[0035] Acquire standard and specification data for the construction industry; and extract category and sub-category information based on the standard and specification data;

[0036] Based on the category information and the subclass information, information is extracted to determine object attributes and data attributes;

[0037] Based on the object attributes, the data attributes, and the pre-determined SWRL rules, an initial hot work safety inspection ontology model is determined.

[0038] Based on the acquired historical construction site video data, the initial hot work safety inspection ontology model is trained to determine the hot work safety inspection ontology model.

[0039] Secondly, embodiments of this application provide a hot work safety monitoring system based on computer vision and ontology. The hot work safety monitoring system includes: an acquisition module, a detection module, a determination module, an inference module, and a generation module.

[0040] The acquisition module is used to acquire video data from the construction site; wherein the video data includes construction personnel.

[0041] The detection module is used to perform target recognition on the video data to obtain the work status information of the construction workers, the personal protective equipment information of the construction workers, and the direction of the construction workers' gaze.

[0042] The determining module is used to identify the worker's activity status based on the work status information to determine the work status of the construction worker; and to identify the wearing status based on the personal protective equipment information to determine the protective equipment wearing information of the construction worker.

[0043] The reasoning module is used to perform safety compliance reasoning based on the work status information, the personal protective equipment information of the construction personnel, the line of sight of the construction personnel, the work status, and the protective equipment wearing information, through a pre-determined hot work safety inspection ontology model, and obtain the safety inspection results.

[0044] The generation module is used to generate a security inspection report based on the security inspection results.

[0045] Thirdly, embodiments of this application provide a hot work safety monitoring device based on computer vision and ontology, comprising: a processor and a memory; wherein,

[0046] The memory is used to store computer programs;

[0047] The processor is configured to call and run the computer program from the memory to perform the method as described in the first aspect.

[0048] Fourthly, embodiments of this application provide a computer-readable storage medium storing executable instructions for causing a processor to perform the method described in the first aspect.

[0049] This application provides a method and system for hot work safety monitoring based on computer vision and ontology. The method includes: acquiring video data from a construction site; wherein the video data includes construction workers; performing target recognition on the video data to obtain the workers' work status information, personal protective equipment information, and line-of-sight information; identifying worker activity status based on the work status information to determine the workers' work status; identifying the wearing status of personal protective equipment based on the personal protective equipment information to determine the workers' protective equipment wearing information; performing safety compliance reasoning based on the work status information, the workers' personal protective equipment information, the workers' line-of-sight information, the work status, and the protective equipment wearing information, obtaining a safety inspection result; and generating a safety inspection report based on the safety inspection result. In this solution, by integrating computer vision and ontology technologies, the system can perceive the work environment in real time, accurately identify worker activity status and personal protective equipment wearing status, and perform semantic reasoning and compliance judgment based on a pre-determined hot work safety inspection ontology model, significantly improving the accuracy and timeliness of safety hazard identification. Attached Figure Description

[0050] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the technical solutions of this application. Obviously, the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0051] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0052] Figure 1 This is an optional flowchart illustrating a hot work safety monitoring method based on computer vision and ontology, provided as an embodiment of this application.

[0053] Figure 2 A schematic diagram illustrating gaze behavior determination in a hot work safety monitoring method based on computer vision and ontology, provided as an embodiment of this application;

[0054] Figure 3 A schematic diagram of the welding operation space range and maximum working distance of a hot work operation safety monitoring method based on computer vision and ontology provided in an embodiment of this application;

[0055] Figure 4 A schematic diagram of a welding mask wearing status determination method provided in this application embodiment for a hot work safety monitoring method based on computer vision and ontology;

[0056] Figure 5 A schematic diagram of a hot work safety monitoring system based on computer vision and ontology is provided for an embodiment of this application;

[0057] Figure 6 This is a schematic diagram of the structure of a hot work safety monitoring device based on computer vision and ontology, provided as an embodiment of this application. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the specific technical solutions of this application will be further described in detail below with reference to the accompanying drawings of the embodiments of this application. The following embodiments are used to illustrate this application, but are not intended to limit the scope of this application.

[0059] Unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used in this application is for the purpose of describing embodiments of this application only and is not intended to be limiting of this application.

[0060] In the following description, references to "some embodiments," "this embodiment," "this application embodiment," and examples, etc., describe a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same subset or different subset of all possible embodiments and may be combined with each other without conflict.

[0061] If the application documents contain similar descriptions such as "first / second", the following explanation shall be added: In the following description, the terms "first / second / third" are used only to distinguish similar objects and do not represent a specific order of objects. It is understood that "first / second / third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0062] Currently, safety inspections at construction sites primarily rely on manual patrols. This method has numerous limitations, including limited monitoring coverage, response delays, lack of continuity in supervision, and inefficient record management. Although existing research has explored the application of computer vision technology in areas such as violation identification, intrusion detection in hazardous areas, and personal protective equipment (PPE) wearing recognition, most current methods still focus on the hazardous condition detection stage and have not yet achieved a closed-loop management system covering the entire process from violation identification and voice alerts to rectification tracking and report generation.

[0063] Disadvantages of existing technology:

[0064] A. Lack of adaptability to dynamic work scenarios and differentiated inspection rules. Construction site environments are complex and ever-changing, and workers' tasks often change in real time. Different work behaviors correspond to different safety compliance requirements. For example, ordinary workers need to be checked for wearing safety helmets and reflective vests, while personnel engaged in professional operations such as welding also need to be checked for the proper wearing of welding masks and protective gloves. Existing methods often fail to dynamically adjust inspection rules according to the actual type of work, resulting in a lack of flexibility and accuracy in compliance judgments.

[0065] B. Personal protective equipment (PPE) inspection has not yet fully achieved a deep semantic understanding of "wearing status." Most existing methods can only determine whether a certain type of PPE exists in an image, but ignore whether it is correctly worn on the corresponding part of the body and whether it meets the requirements of the current work situation. Even if the presence of a safety helmet or welding mask is identified, it is still impossible to effectively determine whether the equipment is in normal working condition, which limits the practical significance of the inspection results.

[0066] C. An automated closed loop covering the entire safety management process has not yet been formed. Although the application of computer vision in the field of construction safety has made some progress, most research is still at the stage of hazard identification and has failed to effectively integrate a series of links such as real-time early warning, rectification tracking and report generation, making it difficult to truly achieve an efficient, coherent and traceable automated safety management process.

[0067] Based on this, embodiments of this application provide a method for hot work safety monitoring based on computer vision and ontology. Figure 1This is an optional flowchart illustrating a hot work safety monitoring method based on computer vision and ontology, provided as an embodiment of this application. Figure 1 The steps shown are explained.

[0068] S101. Obtain video data from the construction site; the video data includes construction personnel.

[0069] In some embodiments of this application, video data of the construction site is acquired using an image acquisition device, and the video data includes construction workers. The image acquisition device may be a drone-based acquisition device; however, this application does not specifically limit its use in this regard.

[0070] In some embodiments of this application, the hot work safety monitoring method based on computer vision and ontology is adapted to building construction scenarios.

[0071] In some embodiments of this application, the hot work safety monitoring method based on computer vision and ontology is adapted to the hot work safety monitoring system based on computer vision and ontology.

[0072] S102. Perform target recognition on the video data to obtain information on the work status of the construction workers, information on their personal protective equipment, and the direction of their line of sight.

[0073] In some embodiments of this application, target detection and tracking are performed on video data to determine the target area containing construction workers; the target area is identified to obtain the work status information of the construction workers; protective equipment detection is performed on the target area to obtain the personal protective equipment information of the construction workers; and target gaze estimation is performed on the target area to obtain the gaze direction of the construction workers.

[0074] S103. Identify the worker's activity status based on the work status information to determine the work status of the construction personnel; identify the wearing status based on the personal protective equipment information to determine the protective equipment wearing information of the construction personnel.

[0075] In some embodiments of this application, a gaze target estimation model is used to identify the work status information and determine the real-time gaze coordinates of the construction worker; a target detection model is used to identify the work status information and determine the bounding box region where the welding area corresponding to the construction worker is located; based on the real-time gaze coordinates and the bounding box region, visual attention constraints and work space constraints are determined to determine the work status of the construction worker.

[0076] In some embodiments of this application, a target detection model is used to extract features from personal protective equipment (PPE) information to determine the first bounding box corresponding to the construction worker, the second bounding box corresponding to the head, the third bounding box corresponding to the safety helmet, the fourth bounding box corresponding to the protective gloves, the fifth bounding box corresponding to the welding mask, and the sixth bounding box corresponding to the reflective vest. The intersection-union ratio (IUGR) between the first bounding box and the second, third, fourth, and sixth bounding boxes is calculated respectively. Based on the IUGR between the first bounding box and the second, third, and fourth bounding boxes, the first sub-protective equipment wearing information corresponding to the direct-wearing type is determined. The line-of-sight intersection is judged based on the fifth bounding box and the real-time gaze coordinate point to obtain the line-of-sight intersection result. Based on the line-of-sight intersection result and the IUGR between the first and sixth bounding boxes, the second sub-protective equipment wearing information corresponding to the indirect-wearing type is determined. Based on the first and second sub-protective equipment wearing information, the PPE wearing information of the construction worker is determined.

[0077] S104. Based on the work status information, the personal protective equipment information of the construction personnel, the line of sight of the construction personnel, the work status and the wearing information of protective equipment, the safety compliance reasoning is performed through the pre-determined hot work safety inspection ontology model to obtain the safety inspection results.

[0078] In some embodiments of this application, class instances are determined based on work status information, personal protective equipment information of construction workers, and the direction of the construction workers' line of sight; attribute instances are determined based on work status and protective equipment wearing information; and an attribute matrix is ​​established based on the attribute instances; based on the class instances and the attribute matrix, hierarchical safety compliance reasoning is performed through the hot work safety inspection ontology model to obtain the safety inspection results.

[0079] S105. Generate a security inspection report based on the security inspection results.

[0080] In some embodiments of this application, if the safety inspection results indicate a violation, a warning message is generated; the warning message is played via voice; in response to the warning message, a rectified on-site image is acquired; a safety inspection is performed on the rectified on-site image, and if the obtained inspection results indicate that the construction safety is normal, a warning cancellation message is generated; a safety inspection report is generated based on the safety inspection results, the on-site image corresponding to the violation, and the rectified on-site image.

[0081] Understandably, by integrating computer vision and ontology technology, this application enables the system to perceive the working environment in real time, accurately identify the worker's activity status and personal protective equipment wearing status, and perform semantic reasoning and compliance judgment based on a pre-determined ontology model for hot work safety inspection, which significantly improves the accuracy and timeliness of safety hazard identification.

[0082] In some embodiments of this application, the worker activity status identification in step S103 to determine the work status of construction workers can be achieved through steps S201-S203, as follows:

[0083] S201. By using the gaze target estimation model, the operation status information is identified, and the real-time gaze coordinates of the construction personnel are determined.

[0084] S202. Using the target detection model, the operation status information is identified to determine the bounding box area where the welding area corresponding to the construction worker is located.

[0085] S203. Based on real-time gaze coordinates and bounding box regions, determine visual attention constraints and work space constraints to ascertain the working status of construction personnel.

[0086] For example, during welding operations, operators need to maintain continuous observation of the welding point, and visual attention features have significant behavioral characteristics. By combining gaze point estimation and target detection technology, a method for quantifying visual attention was established. The specific implementation method is shown in Figure 2. In the two-dimensional image coordinate system, (1) the real-time gaze coordinate point G(x) of the operator is obtained using the gaze target estimation model. gaze ,y gaze (2) Identify the bounding box region W of the welding area using the target detection model; (3) Establish the inclusion relationship judgment criteria between the gaze point and the welding area:

[0087] (1)

[0088] (2)

[0089] When the gaze point G falls within the bounding box W of the welding area, it is determined that the operator is currently focused on the welding task area, and the visual attention constraint is considered satisfied. Figure 2 The real-time gaze coordinates of operator 1, G1(x), are obtained in the process. gaze ,y gaze ),judge 1 This indicates that operator 1's gaze state is "yes". Obtain operator 2's real-time gaze coordinates G2(x). gaze ,y gaze ),judge 2 If so, it means that operator 2's gaze state is no.

[0090] In real-world work scenarios, besides welders, other personnel (such as support staff and safety supervisors) may also briefly glance at the welding area. To distinguish these temporary gaze behaviors, this study introduces ergonomically based workspace constraints. For example... Figure 3 As shown, when an operator performs welding work, their working space must be controlled within the maximum reachable area.

[0091] Based on the GB / T 10000-2023 standard "Anthropometric Dimensions of Chinese Adults" and combined with the structural characteristics of workers in the "Big Data Analysis Report on the Current Status of Labor Employment in the Construction Industry (2021)" (males account for 89%, and the 30-60 age group accounts for 88.70%), this study selected the 95th percentile anthropometric dimensions of males aged 36-60 as the baseline parameter. This selection strategy can cover the physiological characteristics of the main workforce in the construction industry. The specific parameters used include: upper limb forward extension length (886mm), head length (200mm), and head height (253mm). The maximum reachable range (686mm) was calculated, which is the maximum working distance Dmax.

[0092] The working distance D is defined as the distance between the operator's eye and the point of fixation. To simplify the mapping between physical coordinates and image coordinates, a scaling factor α is introduced, defined as follows:

[0093] (3)

[0094] In image coordinates, the height of the head bounding box obtained through object detection. Represented as:

[0095] (4)

[0096] This allows us to estimate the maximum working distance in the image coordinate system:

[0097] (5)

[0098] The operator's eye coordinates (x, y) are predicted using a gaze target estimation model. eye , y eye ) and gaze point coordinates (x gaze , y gaze ), thereby calculating the actual working distance (D′) in image coordinates:

[0099] (6)

[0100] To ensure that the operator's actual working distance is within the constraints of an ergonomic workspace, the following must be met:

[0101] (7)

[0102] The system identifies a worker as being in a welding state only if the gaze point falls within the welding area (visual attention constraint) and the actual working distance does not exceed the maximum working range (workspace constraint).

[0103] Understandably, introducing visual attention constraints and workspace constraints in the process of determining the working status of construction workers can improve the accuracy of work status recognition.

[0104] In some embodiments of this application, the identification of the wearing status based on personal protective equipment information in S103 to determine the protective equipment wearing information of construction workers can be achieved through S301-S304, as follows:

[0105] S301. Using a target detection model, feature extraction is performed on personal protective equipment information to determine the first bounding box corresponding to the construction worker, the second bounding box corresponding to the head, the third bounding box corresponding to the safety helmet, the fourth bounding box corresponding to the protective gloves, the fifth bounding box corresponding to the welding mask, and the sixth bounding box corresponding to the reflective vest.

[0106] S302. Calculate the intersection-union ratio between the first bounding box and the second, third, fourth, and sixth bounding boxes respectively; and based on the intersection-union ratio between the first bounding box and the second, third, and fourth bounding boxes, determine the wearing information of the first sub-protective equipment corresponding to the direct-wear type.

[0107] S303. Based on the fifth bounding box and the real-time gaze coordinates, the line-of-sight intersection is determined to obtain the line-of-sight intersection result; and based on the line-of-sight intersection result and the intersection-union comparison of the first bounding box and the sixth bounding box, the wearing information of the second sub-protective equipment corresponding to the non-direct wearing type is determined.

[0108] S304. Based on the first sub-protective equipment wearing information and the second sub-protective equipment wearing information, determine the protective equipment wearing information of the construction personnel.

[0109] For example, the first step for all workers entering a construction site is to confirm that they are wearing safety helmets and reflective vests correctly. In hot work scenarios, such as welding, in addition to safety helmets and reflective vests, it is also necessary to check whether welding masks and protective gloves are worn. This study focuses on the testing of personal protective equipment (PPE) for hot work, and will discuss both directly worn PPE (such as safety helmets, welding masks, and protective gloves) and indirectly worn PPE (such as reflective vests).

[0110] For directly worn PPEs such as safety helmets, reflective vests, and protective gloves, this application adopts a method combining object detection and Intersection over Union (IoU) analysis for identification. The bounding boxes of the worker, head, helmet, vest, and gloves are extracted using an object detection model, and the IoU values ​​between the worker and each type of PPE are calculated, as shown in formula (9). Specifically, the determination of helmet wearing status requires two steps: first, the IoU between "person" and "head" is calculated to confirm the rationality of head positioning; then, it is determined whether the IoU between "head" and "helmet" meets the threshold requirement. The logic for determining the wearing status is as follows:

[0111] (8)

[0112] (9)

[0113] Welding masks are typically designed with a transparent viewing window through which the worker focuses their gaze on the welding area. Specifically, the worker's line of sight (i.e., the line connecting their eyes to the point of focus) needs to pass through the transparent portion of the welding mask to ensure a clear view of the weld point. Simplifying this relationship, as... Figure 4 As shown, in a given two-dimensional planar image, the bounding box of the welding mask is composed of an ordered sequence of vertices. Definitions, including , , and ,edge (agreement) The edge normal vector is The endpoint of the line segment of the line of sight, P0(x) eye , y eye ), P1(x gaze , y gaze ).

[0114] The line-of-sight vector is:

[0115] (10)

[0116] The parametric equation for the line segment is:

[0117] (11)

[0118] P0(x eye , y eyeWithin the boundary frame of the welding mask, the line of sight segment It must intersect the boundary of the welding mask at least once:

[0119] (12)

[0120] It is understandable that in the process of determining the protective equipment worn by construction workers, direct-wear personal protective equipment (such as safety helmets, welding masks, and protective gloves) and indirect-wear personal protective equipment (such as reflective vests) are detected and identified separately to make the determined protective equipment wearing information more accurate.

[0121] In some embodiments of this application, S104 can be implemented by S401-S403, as follows:

[0122] S401. Based on the work status information, the personal protective equipment information of the construction personnel, and the line of sight of the construction personnel, determine the class instance.

[0123] S402. Based on the working status and protective equipment wearing information, determine the attribute instances; and establish an attribute matrix based on the attribute instances.

[0124] S403. Based on class instances and attribute matrices, hierarchical safety compliance reasoning is performed through the hot work safety inspection ontology model to obtain safety inspection results.

[0125] For example, after completing target detection, gaze target estimation and clothing relationship recognition, the entity categories and attribute states extracted by the computer vision module need to be mapped to the hot work safety inspection ontology model to establish class instances and property instances.

[0126] Specifically, class instance mapping dynamically maps the categories (such as Person, Helmet, WeldingArea) and their unique identifiers (IDs) output by the target detection and tracking module to concrete instances (such as person_2, helmet_3, weldingarea_4) in the ontology. The category-ID mapping relationship is defined as follows:

[0127] (13)

[0128] The instantiated collection is defined as follows:

[0129] } (14)

[0130] Attribute instance mapping, based on gaze target estimation and wear relationship recognition, extracts the worker's work status (e.g., isWelding) and PPE wearing status (e.g., withHelmet, withSafetyVest), and assigns them as data attributes to the corresponding instances. The attribute matrix of the instances is as follows:

[0131] (15)

[0132] In actual inspection processes, due to factors such as the non-100% accuracy of the ontology model for hot work safety inspections and temporary obstructions, directly using the recognition results of a single frame for safety inspections may result in missed or false detections. Therefore, a state smoothing method based on a time window is introduced. Specifically, a state queue Q is maintained, with a window length of N=20 frames:

[0133] (16)

[0134] To smooth out state changes, a majority voting decision function is used to determine stable attribute states:

[0135] (17)

[0136] This means that an instance's `personID` attribute is considered stable as `True` only if its state is `True` for more than 50% of the frames within the time window. Otherwise, the attribute state is considered `False`. This approach effectively reduces errors caused by misjudgments in a single frame.

[0137] First, the worker's activity status and PPE wearing status are instantiated using a visual-ontology mapping. Then, based on ontology reasoning, the system outputs the risk factors, potential risks, and corrective measures for the current scenario. The system employs a tiered inspection strategy: 1) When the worker is welding, it needs to detect that the worker is wearing a safety helmet, reflective vest, welding mask, and protective gloves; 2) If the worker is not welding, only the safety helmet and reflective vest are identified to obtain the safety inspection result.

[0138] In some embodiments of this application, S105 can be implemented by S501-S503, as follows:

[0139] S501. If the safety inspection results indicate that there is a violation, a warning message will be generated and played via voice.

[0140] S502. In response to the warning message, perform safety rectification, acquire on-site images after rectification, and generate warning cancellation information.

[0141] S503. Generate a safety inspection report based on the safety inspection results, the on-site images of the violations corresponding to the safety inspection results, and the on-site images after rectification.

[0142] For example, when the system detects a violation, the voice alarm module will broadcast specific warning information in real time (e.g., "Warning! No protective gloves are being worn, posing a risk of burns. Please wear protective gloves immediately."). After rectification is completed, the system will automatically send a cancellation notification (e.g., "Protective gloves are being worn correctly, warning cancelled. Thank you for your cooperation."). The final generated safety inspection report includes the following five categories of information: 1) Basic project information (from the ontology model); 2) Hot work information (from the ontology model); 3) Safety inspection results (based on visual analysis and ontology reasoning); 4) On-site images at the time of the violation (based on visual analysis); 5) Verification images after rectification (based on visual analysis). This solution realizes a complete closed-loop management process from status recognition and violation warning to rectification verification.

[0143] In some embodiments of this application, S106-S109 are executed before S104, as follows:

[0144] S106. Obtain standard and specification data for the construction industry; and extract category and sub-category information based on the standard and specification data.

[0145] S107. Extract information based on category information and subclass information to determine object attributes and data attributes.

[0146] S108. Based on object attributes, data attributes, and pre-determined SWRL rules, determine the initial hot work safety inspection ontology model.

[0147] S109. Based on the acquired historical construction site video data, train the initial hot work safety inspection ontology model and determine the hot work safety inspection ontology model.

[0148] For example, by standardizing the workflow for safety inspections related to hot work and systematically organizing domain knowledge, the computer's ability to understand and apply this domain can be improved. The ontology model for hot work safety inspections in this application is implemented using OWL (Web Ontology Language), and mainly includes three key components: core class hierarchy definition, attribute definition, and SWRL rule design.

[0149] The knowledge sources for this hot work safety inspection ontology model are mainly existing laws, regulations, and standards. It also references ontology models in the construction field and combines them with the unique characteristics of hot work safety inspection to design a top-level conceptual framework for the hot work safety inspection ontology model.

[0150] Table 1

[0151]

[0152] Table 2

[0153]

[0154] Table 1 shows the categories described in this ontology, their definitions, and some subclasses. Table 2 shows the object attributes and some data attributes described in the ontology model for hot work safety inspection. By predefining the logical relationships between antecedents and consequents, SWRL rules enable the system to derive new knowledge from existing facts. Sixteen SWRL rules were developed for the safety inspection of personal protective equipment (PPE) for hot work, as follows:

[0155] Rule 1: If a worker is not wearing a safety helmet, then that action is a risk factor. Person(?a)^with_Helmet(?a,false)->hasPrecursors(?a,failure_to_wear_helmet)

[0156] Rule 2: If a worker is not wearing reflective clothing, then that behavior is a risk factor.

[0157] Person(?a)^with_Safety_Vest(?a,false)->hasPrecursors(?a,failure_to_wear_safety_vest)

[0158] Rule 3: If a welder is not wearing a welding mask, then that behavior is a risk factor.

[0159] Person(?a)^is_Welding(?a,True)^with_Weld_Mask(?a,false)->hasPrecursors(?a,failure_to_wear_weld_mask)

[0160] Rule 4: If a welder is not wearing gloves, then that behavior is a risk factor.

[0161] Person(?a)^is_Welding(?a,True)^with_Glove(?a,false)->hasPrecursors(?a,failure_to_wear_glove)

[0162] Rule 5: If a worker is not wearing a safety helmet, he is at risk of being struck by an object.

[0163] Person(?a) ^hasPrecursors(?a,failure_to_wear_helmet)

[0164] ->causeAccident(failure_to_wear_helmet, object_strike)

[0165] Rule 6: If a worker is not wearing a reflective vest, he is at risk of being struck by an object.

[0166] Person(?a)^ hasPrecursors(?a,failure_to_wear_safety_vest)

[0167] ->causeAccident(failure_to_wear_safety_vest, object_strike)

[0168] Rule 7: If a worker is welding without wearing a welding mask, he is at risk of burns.

[0169] Person(?a)^is_Welding^(?a,true)^ hasPrecursors(?a,failure_to_wear_weld_mask)

[0170] ->causeAccident(failure_to_wear_weld_mask, burn)

[0171] Rule 8: If a worker is welding without wearing gloves, he is at risk of burns.

[0172] Person(?a)^is_Welding^(?a,true)^ hasPrecursors(?a,failure_to_wear_glove)

[0173] ->causeAccident(failure_to_wear_glove, burn)

[0174] Rule 9: If a worker is not wearing a hard hat, he or she should be reminded to wear one.

[0175] Person(?a) ^hasPrecursors(?a,failure_to_wear_helmet)

[0176] ->controlledBy(failure_to_wear_helmet, remind_to_wear_helmet)

[0177] Rule 10: If a worker is not wearing a reflective vest, he or she should be reminded to wear one.

[0178] Person(?a)^ hasPrecursors(?a,failure_to_wear_safety_vest)

[0179] ->controlledBy(failure_to_wear_safety_vest, remind_to_wear_safety_vest)

[0180] Rule 11: If a worker is welding without wearing a welding mask, he or she should be reminded to wear one.

[0181] Person(?a)^is_Welding^(?a,true)^ hasPrecursors(?a,failure_to_wear_weld_mask)

[0182] ->controlledBy(failure_to_wear_weld_mask, remind_to_wear_Weld_Mask)

[0183] Rule 12: If a worker is welding without wearing a welding mask, he / she should be reminded to wear one.

[0184] Person(?a)^is_Welding^(?a,true)^ hasPrecursors(?a,failure_to_wear_glove)

[0185] ->controlledBy(failure_to_wear_glove,remind_to_wear_ Glove)

[0186] Rule 13: If a worker is not wearing a safety helmet, then the inspection standard "Standard for Safety Inspection of Building Construction" (JGJ 59-2011) 3.13.3 should be referred to.

[0187] Person(?a) ^hasPrecursors(?a,failure_to_wear_helmet)

[0188] ->hasReference(?a,failure_to_wear_helmet, “Standards for SafetyInspection in Construction” (JGJ 59-2011) 3.13.3.

[0189] Rule 14: If a worker is not wearing a reflective vest, then the inspection standard Article 33 of the "Regulations on the Safety Management of Construction Projects" should be referred to.

[0190] Person(?a) ^hasPrecursors(?a,failure_to_wear_helmet)

[0191] ->hasReference(?a,failure_to_wear_helmet, “"Regulations on SafetyProduction in Construction Projects" article 33)

[0192] Rule 15: If a worker is not wearing a welding mask, the inspection standard "General Specifications for Safety, Hygiene and Occupational Health at Construction Sites" 6.0.4.1 should be consulted.

[0193] Person(?a)^is_Welding^(?a,true)^ hasPrecursors(?a,failure_to_wear_weld_mask)

[0194] ->hasReference(?a,failure_to_wear_helmet, “General Specifications for Safety, Hygiene and Occupational Health at Construction and MunicipalConstruction Sites” 6.0.4.1)

[0195] Rule 16: If a worker is not wearing protective gloves, then the inspection standard “Safety of Welding and Cutting” 4.2.2.2 should be consulted.

[0196] Person(?a)^is_Welding^(?a,true)^ hasPrecursors(?a,failure_to_wear_glove)

[0197] ->hasReference(?a,failure_to_wear_helmet, "Welding and CuttingSafety" 4.2.2.2)

[0198] These rules deduce potential safety hazards or risks based on worker activity and PPE wearing status. For example, when the system detects a worker not wearing a safety helmet, it flags this as a safety hazard and triggers necessary safety warnings and corrective actions. These rules are integrated into the inference layer of the hot work safety inspection ontology model, using the Pellet inference engine for dynamic reasoning. The inference engine processes visual data from the construction site (such as whether workers are wearing safety helmets) and maps this data to predefined rules. Based on the logic of the rules, the system deduces new instances (such as potential risks or corrective actions) and updates the instances in the hot work safety inspection ontology model. Worker instances (e.g., person_1) continuously update their attributes based on activity and PPE wearing status. These statuses are provided by visual detection and gaze target estimation techniques. SWRL rules utilize these updated instances for safety inspection inference, and the safety inspection system triggers warnings and corrective actions based on the inference results. Through automated inference, the system can promptly identify hazards and take corrective actions in dynamic and complex environments.

[0199] Understandably, this system achieves intelligent, closed-loop monitoring of the safety status of construction workers during hot work operations. By integrating computer vision and ontology technologies, the system can perceive the work environment in real time, accurately identify worker activity status and personal protective equipment wearing status, and perform semantic reasoning and compliance judgment based on dynamic and differentiated safety rules, significantly improving the accuracy and timeliness of safety hazard identification. It constructs an integrated safety monitoring architecture that combines visual perception and knowledge reasoning. This system not only achieves a leap from "image recognition" to "semantic understanding," but also formalizes and logically reasons for safety regulations through a hot work safety inspection ontology model, supporting adaptive inspections in complex work scenarios. It realizes a complete safety management closed loop from risk warning to rectification verification. The system has real-time voice alarms, automatic generation of structured reports, and rectification process tracking functions, effectively promoting the automation, standardization, and traceability of construction site safety management, and providing reliable technical support for improving the overall efficiency and intelligence level of hot work safety management.

[0200] Based on the above embodiments of the hot work safety monitoring method based on computer vision and ontology, this application also provides a hot work safety monitoring system based on computer vision and ontology, such as... Figure 5 As shown, Figure 5This application provides a schematic diagram of a hot work safety monitoring system based on computer vision and ontology, comprising: an acquisition module 501, a detection module 502, a determination module 503, an inference module 504, and a generation module 505.

[0201] The acquisition module 501 is used to acquire video data of the construction site; wherein, the video data includes construction personnel;

[0202] The detection module 502 is used to perform target recognition on the video data to obtain the work status information of the construction workers, the personal protective equipment information of the construction workers, and the direction of the construction workers' gaze.

[0203] The determining module 503 is used to identify the worker's activity status based on the work status information to determine the work status of the construction worker; and to identify the wearing status based on the personal protective equipment information to determine the protective equipment wearing information of the construction worker.

[0204] The reasoning module 504 is used to perform safety compliance reasoning based on the work status information, the personal protective equipment information of the construction personnel, the line of sight of the construction personnel, the work status and the protective equipment wearing information, through a pre-determined hot work safety inspection ontology model, and obtain the safety inspection result.

[0205] The generation module 505 is used to generate a security inspection report based on the security inspection results.

[0206] Based on the above embodiments of the hot work safety monitoring method based on computer vision and ontology, this application also provides a hot work safety monitoring device based on computer vision and ontology, such as... Figure 6 As shown, Figure 6 This is a schematic diagram of a hot work safety monitoring device based on computer vision and body recognition, provided in an embodiment of this application. The device 6 includes a processor 601 and a memory 602. The memory 602 stores computer programs; the processor 601 retrieves and runs the computer programs from the memory to execute the hot work safety monitoring method based on computer vision and body recognition as described in the above embodiment.

[0207] In the embodiments of this application, the processor 601 described above can be at least one of the following: Application-Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), Controller, Microcontroller, and Microprocessor. It is understood that for different devices, the electronic device used to implement the above processor function can also be other types, and the embodiments of this application do not specifically limit it.

[0208] This application provides a computer-readable storage medium storing a computer program for implementing, when executed by a processor, the hot work safety monitoring method based on computer vision and ontology as described in any of the above embodiments.

[0209] For example, the program instructions corresponding to the hot work safety monitoring method based on computer vision and body in this embodiment can be stored on storage media such as optical discs, hard disks, and USB flash drives. When the program instructions corresponding to the hot work safety monitoring method based on computer vision and body in the storage media are read or executed by an electronic device, the hot work safety monitoring method based on computer vision and body as described in any of the above embodiments can be realized.

[0210] Furthermore, in the embodiments of this application, the functional modules can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional module.

[0211] If the integrated unit is implemented as a software functional module and is not sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this embodiment, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the method of this embodiment. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0212] It should be understood that the phrases "one embodiment," "an embodiment," or "some embodiments" mentioned throughout the specification mean that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment," "in one embodiment," or "in some embodiments" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the embodiments in this application are merely for descriptive purposes and do not represent the superiority or inferiority of the embodiments. The descriptions of the various embodiments above tend to emphasize the differences between the various embodiments; their similarities or commonalities can be referred to mutually, and for the sake of brevity, these will not be repeated here.

[0213] The modules described above as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules. They may be located in one place or distributed across multiple network units. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0214] In addition, each functional module in the various embodiments of this application can be integrated into one processing unit, or each module can be a separate unit, or two or more modules can be integrated into one unit; the integrated modules can be implemented in hardware or in the form of hardware plus software functional units.

[0215] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory (ROM), magnetic disks, or optical disks.

[0216] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

[0217] The features disclosed in the several product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0218] The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0219] The above description is merely an embodiment of this application, but the protection scope of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the protection scope of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.

Claims

1. A method for monitoring the safety of hot work operations based on computer vision and ontology, characterized in that, The method includes: Acquire video data from the construction site; wherein the video data includes construction personnel; Target recognition is performed on the video data to obtain the work status information of the construction workers, the personal protective equipment information of the construction workers, and the direction of the construction workers' gaze; The work status information is used to identify the worker's activity status and determine the work status of the construction worker; the personal protective equipment information is used to identify the wearing status and determine the protective equipment wearing information of the construction worker. Based on the work status information, the personal protective equipment information of the construction personnel, the line of sight of the construction personnel, the work status, and the protective equipment wearing information, safety compliance reasoning is performed through a pre-determined hot work safety inspection ontology model to obtain the safety inspection results; Based on the security inspection results, a security inspection report is generated; The step of identifying worker activity status from the work status information to determine the work status of the construction personnel includes: By using a gaze target estimation model, the work status information is identified, and the real-time gaze coordinates of the construction personnel are determined. The target detection model is used to identify the work status information and determine the bounding box area where the welding area corresponding to the construction worker is located. Based on the real-time gaze coordinates and the bounding box region, visual attention constraints and work space constraints are determined to ascertain the working status of the construction personnel. Based on the work status information, the personal protective equipment information of the construction workers, the line of sight of the construction workers, the work status, and the protective equipment wearing information, the safety compliance reasoning is performed through a pre-determined hot work safety inspection ontology model to obtain the safety inspection results, including: Based on the work status information, the personal protective equipment information of the construction workers, and the direction of the construction workers' line of sight, class instances are determined; Based on the working status and the protective equipment wearing information, attribute instances are determined; and an attribute matrix is ​​established based on the attribute instances. Based on the class instance and the attribute matrix, the safety inspection result is obtained by performing hierarchical safety compliance reasoning through the hot work safety inspection ontology model.

2. The method according to claim 1, characterized in that, The step of performing target recognition on the video data to obtain the construction worker's work status information, the construction worker's personal protective equipment information, and the construction worker's line of sight direction includes: The video data is subjected to target detection and target tracking to determine the target area containing the construction personnel; and the target area is identified to obtain the work status information of the construction personnel. The protective equipment in the target area is tested to obtain the personal protective equipment information of the construction personnel. Target gaze estimation is performed on the target area to obtain the line of sight direction of the construction worker.

3. The method according to claim 1, characterized in that, The step of identifying the wearing status based on the personal protective equipment information to determine the protective equipment wearing information of the construction worker includes: By using a target detection model, feature extraction is performed on the personal protective equipment information to determine the first bounding box corresponding to the construction worker, the second bounding box corresponding to the head, the third bounding box corresponding to the safety helmet, the fourth bounding box corresponding to the protective gloves, the fifth bounding box corresponding to the welding mask, and the sixth bounding box corresponding to the reflective vest. Calculate the intersection-union ratio (IUGR) between the first bounding box and the second, third, fourth, and sixth bounding boxes respectively; and based on the IUGR between the first bounding box and the second, third, and fourth bounding boxes, determine the wearing information of the first sub-protective equipment corresponding to the direct-wear type. Based on the fifth bounding box and the real-time gaze coordinate point, the line-of-sight intersection is determined to obtain the line-of-sight intersection result; and based on the line-of-sight intersection result and the intersection-union comparison between the first bounding box and the sixth bounding box, the wearing information of the second sub-protective equipment corresponding to the non-direct wearing type is determined. Based on the first sub-protective equipment wearing information and the second sub-protective equipment wearing information, the protective equipment wearing information of the construction personnel is determined.

4. The method according to claim 1, characterized in that, The generation of a security inspection report based on the security inspection results includes: If the security check results indicate a violation, a warning message is generated and played via voice. In response to the warning message, obtain images of the site after rectification; The rectified on-site images are subjected to safety inspection. If the inspection results indicate that the construction is safe and normal, a warning cancellation message is generated. The safety inspection report is generated based on the safety inspection results, the on-site images of the violations corresponding to the safety inspection results, and the on-site images after rectification.

5. The method according to any one of claims 1-4, characterized in that, Before obtaining the safety inspection result by performing safety compliance reasoning through a pre-determined hot work safety inspection ontology model based on the work status information, the personal protective equipment information of the construction personnel, the line of sight of the construction personnel, the work status, and the protective equipment wearing information, the method further includes: Acquire standard and specification data for the construction industry; and extract category and sub-category information based on the standard and specification data; Based on the category information and the subclass information, information is extracted to determine object attributes and data attributes; Based on the object attributes, the data attributes, and the pre-determined SWRL rules, an initial hot work safety inspection ontology model is determined. Based on the acquired historical construction site video data, the initial hot work safety inspection ontology model is trained to determine the hot work safety inspection ontology model.

6. A hot work safety monitoring system based on computer vision and ontology, characterized in that, The hot work safety monitoring system based on computer vision and ontology includes: an acquisition module, a detection module, a determination module, an inference module, and a generation module, wherein... The acquisition module is used to acquire video data from the construction site; wherein the video data includes construction personnel. The detection module is used to perform target recognition on the video data to obtain the work status information of the construction workers, the personal protective equipment information of the construction workers, and the direction of the construction workers' gaze. The determining module is used to identify the worker's activity status based on the work status information to determine the work status of the construction worker; and to identify the wearing status based on the personal protective equipment information to determine the protective equipment wearing information of the construction worker. The reasoning module is used to perform safety compliance reasoning based on the work status information, the personal protective equipment information of the construction personnel, the line of sight of the construction personnel, the work status, and the protective equipment wearing information, through a pre-determined hot work safety inspection ontology model, and obtain the safety inspection results. The generation module is used to generate a security inspection report based on the security inspection results; The determining module is further configured to identify the work status information through a gaze target estimation model to determine the real-time gaze coordinates of the construction worker; identify the work status information through a target detection model to determine the bounding box region where the welding area corresponding to the construction worker is located; and determine the work status of the construction worker by judging visual attention constraints and work space constraints based on the real-time gaze coordinates and the bounding box region. The reasoning module is further configured to determine class instances based on the work status information, the personal protective equipment information of the construction personnel, and the line of sight of the construction personnel; determine attribute instances based on the work status and the protective equipment wearing information; establish an attribute matrix based on the attribute instances; and perform hierarchical safety compliance reasoning through the hot work safety inspection ontology model based on the class instances and the attribute matrix to obtain the safety inspection results.

7. A hot work safety monitoring device based on computer vision and ontology, characterized in that, include: Processor and memory, of which, The memory is used to store computer programs; The processor is configured to call and run the computer program from the memory to perform the method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, It stores executable instructions for causing a processor to execute, thereby implementing the method of any one of claims 1 to 5.

Images