Fire-fighting inspection method and system for unmanned chemical plant by using fire-fighting humanoid robot

Abstract

The invention discloses a fire-fighting inspection method and system for an unmanned plant by a fire-fighting humanoid robot, and relates to the technical field of the fire-fighting humanoid robot, and the method comprises the steps: determining a plurality of fire risk nodes according to the detection of each fire risk region; the sub-risk level of the fire risk area is determined based on the position and temperature data of the plurality of fire risk nodes, and the accuracy of the sub-risk level of the fire risk area is improved. Determining a plurality of fire-fighting inspection areas of the fire-fighting humanoid robot according to the sub-risk level of each fire risk area and the current position of the fire-fighting humanoid robot relative to the unmanned plant; based on the positions of the multiple fire-fighting inspection areas, the inspection sequence of the multiple sub fire-fighting inspection paths is determined, and according to the inspection sequence of the multiple sub fire-fighting inspection paths, the current position of the fire-fighting humanoid robot and the real-time events of the fire-fighting inspection areas, the real-time fire-fighting inspection path of the fire-fighting humanoid robot is determined. And the dynamic effectiveness of the real-time fire-fighting routing inspection path of the fire-fighting humanoid machine is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of fire-fighting humanoid robots, and in particular to a fire inspection method and system for unmanned factories by fire-fighting humanoid robots. Background Art

[0002] With the development of science and technology, fire-fighting humanoid robots are gradually applied to people's lives and perform fire inspections in unmanned factories. In the existing technology, a preset fire inspection path is marked based on the unmanned factory, and the preset fire inspection path is input into the fire-fighting humanoid robot. The fire-fighting humanoid robot performs fire inspections along the preset fire inspection path and collects real-time images of the surrounding area. However, the preset fire inspection path does not take into account the sub-risk level of the fire risk area in the unmanned factory, but only considers the distribution map of the unmanned factory, which affects the dynamic effectiveness of the real-time fire inspection path of the fire-fighting humanoid machine. Summary of the Invention

[0003] The purpose of the present invention is to overcome the deficiencies of the prior art. The present invention provides a fire inspection method and system for unmanned chemical factories using a fire-fighting humanoid robot.

[0004] An embodiment of the present invention provides a fire inspection method for an unmanned factory by a fire-fighting humanoid robot, comprising: collecting a distribution map of the unmanned factory, and determining a plurality of unmanned operation areas according to the division of the distribution map of the unmanned factory; in each unmanned operation area, determining a plurality of fire risk areas based on the operation data, real-time operation images and environmental images of the unmanned operation area; in each fire risk area, determining a plurality of fire risk nodes according to the detection of each fire risk area; determining a sub-risk level of the fire risk area based on the position and temperature data of the plurality of fire risk nodes; determining a plurality of fire inspection areas of the fire-fighting humanoid robot according to the sub-risk level of each fire risk area and the current position of the fire-fighting humanoid robot relative to the unmanned factory, and marking a sub-fire inspection path of each fire inspection area; determining an inspection order of the plurality of sub-fire inspection paths based on the positions of the plurality of fire inspection areas, and determining a real-time fire inspection path of the fire-fighting humanoid robot according to the inspection order of the plurality of sub-fire inspection paths, the current position of the fire-fighting humanoid robot and the real-time events of each fire inspection area.

[0005] An embodiment of the present invention provides a fire inspection system for an unmanned factory using a fire-fighting humanoid robot. The fire inspection system for an unmanned factory using a fire-fighting humanoid robot is applied to the fire inspection method for an unmanned factory using a fire-fighting humanoid robot. The fire inspection system for an unmanned factory using a fire-fighting humanoid robot includes: The unmanned operation area module is used to collect the distribution map of the unmanned factory and determine multiple unmanned operation areas according to the division of the distribution map of the unmanned factory; A fire risk area module is used to determine multiple fire risk areas in each unmanned operation area based on the operation data, real-time operation images and environmental images of the unmanned operation area; A sub-risk level module is used to determine a plurality of fire risk nodes in each fire risk area based on the detection of each fire risk area; and determine the sub-risk level of the fire risk area based on the location and temperature data of the plurality of fire risk nodes; The sub-fire inspection path module is used to determine multiple fire inspection areas of the fire humanoid robot based on the sub-risk level of each fire risk area and the current position of the fire humanoid robot relative to the unmanned chemical plant, and mark the sub-fire inspection paths of each fire inspection area; The real-time fire inspection path module is used to determine the inspection order of multiple sub-fire inspection paths based on the locations of multiple fire inspection areas, and to determine the real-time fire inspection path of the fire humanoid machine according to the inspection order of multiple sub-fire inspection paths, the current location of the fire humanoid robot and the real-time events of each fire inspection area.

[0006] Compared with the prior art, the present invention has the following beneficial effects: In an embodiment of the present invention, through the method in the embodiment of the present invention, in each fire risk area, multiple fire risk nodes are determined based on the detection of each fire risk area; the sub-risk level of the fire risk area is determined based on the position and temperature data of the multiple fire risk nodes, and the multiple fire risk nodes are targetedly managed and controlled, which is compatible with the overall consideration of the position and temperature data of the multiple fire risk nodes, thereby improving the accuracy of the sub-risk level of the fire risk area.

[0007] Therefore, according to the sub-risk level of each fire risk area and the current position of the fire humanoid robot relative to the unmanned factory, multiple fire inspection areas of the fire humanoid robot are determined, and the sub-fire inspection paths of each fire inspection area are marked; based on the positions of the multiple fire inspection areas, the inspection order of the multiple sub-fire inspection paths is determined, and the real-time fire inspection path of the fire humanoid machine is determined according to the inspection order of the multiple sub-fire inspection paths, the current position of the fire humanoid robot and the real-time events of each fire inspection area. The sub-fire inspection path of each fire inspection area is introduced, and the overall consideration of the inspection order of multiple sub-fire inspection paths, the current position of the fire humanoid robot and the real-time events of each fire inspection area is realized, thereby ensuring the dynamic effectiveness of the real-time fire inspection path of the fire humanoid machine. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 11 is a flow chart of a method for fire inspection in an unmanned factory by a fire-fighting humanoid robot according to an embodiment of the present invention; Figure 2 1 is a flow chart of step S11 in a fire inspection method for an unmanned factory by a fire-fighting humanoid robot according to an embodiment of the present invention; Figure 3 1 is a flow chart of step S12 in the fire inspection method for an unmanned factory by a fire-fighting humanoid robot in an embodiment of the present invention; Figure 4 1 is a flow chart of step S13 in the fire inspection method for an unmanned factory by a fire-fighting humanoid robot in an embodiment of the present invention; Figure 5 1 is a flow chart of step S14 in the fire inspection method for an unmanned factory by a fire-fighting humanoid robot in an embodiment of the present invention; Figure 6 1 is a flow chart of step S15 in the fire inspection method for an unmanned factory by a fire-fighting humanoid robot in an embodiment of the present invention; Figure 7 The figure is a schematic diagram of the structure of a fire inspection system for an unmanned factory using a fire-fighting humanoid robot in an embodiment of the present invention. DETAILED DESCRIPTION

[0009] The technical solutions in the embodiments of the present invention will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present invention.

[0010] See also Figures 1 to 7 A fire inspection method for an unmanned chemical factory using a fire-fighting humanoid robot is applied to a fire inspection scenario of an unmanned chemical factory using a fire-fighting humanoid robot. The fire inspection method for an unmanned chemical factory using a fire-fighting humanoid robot includes: Step S11: collecting a distribution map of the unmanned factory, and determining a plurality of unmanned operation areas according to the division of the distribution map of the unmanned factory; Step S12: In each unmanned operation area, a plurality of fire risk areas are determined based on the operation data, real-time operation images, and environmental images of the unmanned operation area; Step S13: In each fire risk area, a plurality of fire risk nodes are determined based on the detection of each fire risk area; and a sub-risk level of the fire risk area is determined based on the position and temperature data of the plurality of fire risk nodes; Step S14: determining multiple fire inspection areas of the fire humanoid robot based on the sub-risk level of each fire risk area and the current position of the fire humanoid robot relative to the unmanned factory, and marking the sub-fire inspection paths of each fire inspection area; Step S15: determining an inspection order of the plurality of sub-fire inspection paths based on the positions of the plurality of fire inspection areas, and determining a real-time fire inspection path of the fire humanoid robot based on the inspection order of the plurality of sub-fire inspection paths, the current position of the fire humanoid robot, and real-time events of each fire inspection area; refer to Figure 2 In step S11, a distribution map of the unmanned factory is collected, and a plurality of unmanned operation areas are determined according to the division of the distribution map of the unmanned factory; In the specific implementation process of the present invention, the specific steps are: S111: collecting a database of unmanned factories, determining a distribution map of the unmanned factories based on the database of the unmanned factories and the names and locations of the unmanned factories, and determining working areas and activity areas based on the detection of the distribution map of the unmanned factories; S112: Determine the current task signal of the unmanned factory based on the detection of the database of the unmanned factory, determine the current work content of the unmanned factory based on the analysis of the current task signal, determine multiple unmanned operation areas based on the matching of the current work content and the work area of ​​the unmanned factory, and mark the operation environment and operation content of the multiple unmanned operation areas.

[0011] In the embodiment of the present application, a database of an unmanned factory is collected. The database typically contains the factory's layout diagram, equipment list, production line configuration, safety specifications, historical maintenance records, etc. This information is crucial for determining the factory layout and dividing work areas and activity areas in subsequent steps. In this case, the database is stored on the factory's internal server and is also accessed through a cloud computing platform. The data exists in various forms such as CAD drawings, Excel spreadsheets, and database files. Specific software tools or API interfaces are required to extract the data. The distribution map should accurately reflect the actual layout of the factory, including the location, size, and interrelationships of each area. The map should also include the factory's name and location information to facilitate area division and route planning in subsequent steps. At this point, GIS (Geographic Information System) technology is used to convert the CAD drawings into a digital distribution map. A unified coordinate system is established on the distribution map to facilitate the precise positioning of each area and equipment. Key information such as the factory's name, location, main entrances and exits, and safety passages should be marked on the distribution map. Optionally, assume that there is an unmanned factory called "Smart Manufacturing XX Company", whose database is stored on the company's internal server; the database contains a detailed CAD drawing that shows the overall layout of the factory, including production lines, warehouses, equipment areas, office areas, etc.; in addition, there is an equipment inventory form in the database that lists the model, location, operating status and other information of all key equipment, and uses GIS technology to convert the CAD drawing into a digital distribution map; on the distribution map, the factory name "Smart Manufacturing XX Company", location (assuming it is a certain road and number in a certain district in a certain city), main entrances and exits, safety passages and other information are marked; in addition, according to the equipment inventory form, the location and model information of all key equipment are marked on the distribution map.

[0012] Based on the distribution map, work areas and activity areas are divided according to the factory's actual operations and safety regulations. Work areas generally refer to areas directly involved in production activities, such as production lines and processing areas. Activity areas refer to areas supporting production activities, such as material handling corridors, personnel rest areas, and equipment maintenance areas. At this time, the work and activity areas should be reasonably divided based on factors such as the factory's production process, equipment layout, and personnel flow. When dividing areas, full consideration should be given to safety regulations to ensure safe isolation and unobstructed access between areas. Key information such as the name, boundaries, and main equipment of each area should be clearly marked on the distribution map. Optionally, the factory can be divided into multiple work areas and activity areas based on the factory's production process and equipment layout. For example, the production line can be divided into different work areas, each responsible for specific production tasks; material handling corridors can be divided into activity areas to ensure smooth material transportation from the warehouse to the production line; and equipment maintenance areas can be divided into dedicated activity areas for regular maintenance and overhaul of equipment. The name, boundaries, and main equipment information of each area are clearly marked on the distribution map to facilitate fire risk analysis and inspection route planning in subsequent steps.

[0013] Furthermore, the current task signal of the unmanned factory is determined based on the detection of the database of the unmanned factory, the current work content of the unmanned factory is determined based on the analysis of the current task signal, and multiple unmanned operation areas are determined based on the matching of the current work content and the work area of ​​the unmanned factory, and the operating environment and operation content of multiple unmanned operation areas are marked, which is compatible with the overall consideration of the matching of the current work content and the work area of ​​the unmanned factory, and ensures the accuracy of multiple unmanned operation areas.

[0014] At this time, the current task signal of the unmanned factory is determined based on the detection of the database of the unmanned factory. The current task signal usually refers to the various instructions and data generated by the factory during the production process, which reflect the factory's current production plan, equipment status, material requirements, etc.; this information is crucial for determining the current work content and dividing the unmanned operation area in subsequent steps; at this time, the current task signal comes from the factory's production management system, equipment monitoring system, material management system, etc.; the task signal exists in various forms such as real-time data streams, database records, API interface responses, etc.; specific software tools or programming interfaces are required to extract and analyze the task signal.

[0015] refer to Figure 3 In step S12, in each unmanned operation area, multiple fire risk areas are determined based on the operation data, real-time operation images, and environmental images of the unmanned operation area; In the specific implementation process of the present invention, the specific steps are: S121: Real-time monitoring of each unmanned operation area, and determining the operation data space corresponding to each unmanned operation area based on matching between each unmanned operation area and the database of the unmanned factory; S122: In the operation data space corresponding to each unmanned operation area, determining the operation data of the corresponding unmanned operation area based on detection of the operation data space, and determining the corresponding operation event based on analysis of the operation data and the operation content of the unmanned operation area; S123: Monitor the unmanned operation area in real time, and collect real-time operation images and environmental images corresponding to the unmanned operation area, determine a first sub-fire risk range based on the operation events and the real-time operation images, determine a second sub-fire risk range based on the operation events and the environmental images, and determine multiple fire risk areas based on the synthesis of the first sub-fire risk range and the second sub-fire risk range. The multiple fire risk areas are located in the same unmanned factory.

[0016] In an embodiment of the present application, each unmanned operation area is monitored in real time to ensure a real-time and comprehensive understanding of the status of all unmanned operation areas in the factory. At this time, various sensors (such as temperature sensors, humidity sensors, smoke detectors, infrared cameras, etc.) are integrated to monitor environmental parameters; RFID technology or Internet of Things (IoT) devices are used to track material flow and equipment status; and a video surveillance system is used to capture images of the operation area in real time.

[0017] Real-time monitoring data is matched with historical data and configuration information in the database of the unmanned factory to obtain more comprehensive information on the status of the operating area. At this time, a database management system (DBMS) is used to store and manage various types of data in the unmanned factory. Real-time monitoring data is matched with records in the database through data matching algorithms (such as rule-based matching and fuzzy matching).

[0018] Create a virtual space containing all relevant data for each unmanned operation area for subsequent analysis and processing; at this time, build a model of the operation data space based on real-time monitoring data and database matching results; use data visualization tools or platforms to display the status of the operation data space.

[0019] Specifically, the unmanned factory has multiple unmanned operation areas, each equipped with advanced monitoring equipment and sensors. To optimize production processes and improve safety, the factory decided to implement a real-time monitoring and data analysis system. Temperature sensors, humidity sensors, and smoke detectors are installed in each unmanned operation area to monitor environmental parameters in real time. RFID tags are used to track the flow of materials on the production line, and IoT devices are used to monitor the operating status of equipment (such as motor speed, heater temperature, etc.). Cameras are installed to capture real-time images of the operation area for video analysis.

[0020] The MySQL database is used to store historical data, equipment configuration information, and production processes of unmanned factories. By writing SQL query statements, real-time monitoring data is matched with records in the database. For example, when a temperature sensor detects a temperature increase in a certain operating area, the system will search the database for the historical temperature data, equipment configuration information, and related production processes for that area to gain a more comprehensive understanding of the causes and impacts of the temperature increase.

[0021] Based on real-time monitoring data and database matching results, an operation data space is created for each unmanned operation area; this space contains all relevant data of the area, such as environmental parameters, material flow, equipment status, and real-time images; a data visualization platform (such as Tableau or Power BI) is used to display the status of the operation data space; for example, a dashboard is created to display key indicators such as temperature, humidity, smoke concentration, etc. of each operation area in real time, as well as the operating status of the equipment and the flow of materials; by implementing this real-time monitoring and data analysis system, the unmanned factory of "Smart Manufacturing XX Company" can more effectively monitor and manage the status of unmanned operation areas, promptly identify potential safety hazards, and optimize production processes to improve efficiency and safety.

[0022] Furthermore, relevant data from the unmanned operation area is extracted and analyzed from the established operation data space. Data is retrieved from the database using a data query language (such as SQL) or a data access API. Data preprocessing techniques (such as data cleaning and data conversion) are applied to ensure data quality and consistency. Key indicators and parameters within the operation data, such as equipment status, material flow, and environmental conditions, are analyzed. The specific work content of the operation area is then analyzed in conjunction with the unmanned factory's production plan and process flow.

[0023] Based on the parsed job content and real-time job data, specific job events are identified. At this point, the types and standards of job events are defined, such as equipment failure, material shortage, environmental changes, etc. A rule engine or event detection algorithm is applied to match real-time job data with predefined job events.

[0024] Specifically, the unmanned factory has multiple unmanned operation areas, each equipped with advanced monitoring equipment and sensors. To optimize production processes and improve safety, the factory decided to implement a real-time monitoring and data analysis system that can automatically detect and analyze operation events. Each unmanned operation area has a corresponding operation data space, which contains all relevant data for the area, such as equipment status, material flow, environmental conditions, etc. The system uses SQL query statements to retrieve real-time data for each operation area from the database and performs data preprocessing to ensure data quality and consistency.

[0025] Taking one of the work areas (such as the "material handling area") as an example, the system analyzed the real-time data of this area and found that the operating speed of the material handling robot suddenly slowed down, and the readings of the material sensor showed a decrease in material inventory; combined with the factory's production plan and process flow, the system analyzed that the specific work content of this work area is "moving materials from the warehouse to the production line" and identified the work event as "material handling robot failure" or "material shortage."

[0026] The system defines the types and standards of operational events, including equipment failure, material shortages, and environmental changes. Applying a rules engine, the system matches real-time data with predefined operational events. In this example, due to the simultaneous slowdown of the material handling robot and the reduction in material inventory, the system identified a specific operational event: "Material shortage caused by material handling robot failure." By implementing this real-time monitoring and data analysis system, unmanned factories can automatically detect and analyze operational events in unmanned operating areas, promptly identify potential problems, and take appropriate measures, thereby optimizing production processes and improving safety.

[0027] Therefore, the unmanned operation area is monitored in real time, and the real-time operation images and environmental images corresponding to the unmanned operation area are collected. The first sub-fire risk range is determined according to the operation events and the real-time operation images, and the second sub-fire risk range is determined according to the operation events and the environmental images. Based on the synthesis of the first sub-fire risk range and the second sub-fire risk range, multiple fire risk areas are determined. The locations of multiple fire risk areas are in the same unmanned factory, which is compatible with the overall consideration of the synthesis of the first sub-fire risk range and the second sub-fire risk range, thereby ensuring the accuracy of multiple fire risk areas.

[0028] At this time, the unmanned operation area is monitored in real time and images are collected. The operation images and environmental images of the unmanned operation area are obtained in real time through monitoring cameras and other image acquisition equipment. At this time, cameras and infrared cameras are installed to capture visible light and infrared images respectively. Cameras are configured to cover key operation areas and environmental areas to ensure monitoring without blind spots. The video surveillance system is used to collect and store image data in real time.

[0029] By combining operational events and real-time operational images, the specific operational areas that caused the fire are identified, and operational events such as equipment overheating and material leakage are analyzed, as these events increase the fire risk. Image recognition technology is used to detect abnormal phenomena in real-time operational images, such as flames, smoke, and high-temperature areas. The first sub-fire risk range is determined by integrating operational events and image recognition results.

[0030] Combine operational events and environmental images to assess the environmental fire risk of the entire unmanned operation area. At this time, analyze the environmental images to detect fire hazards, such as accumulation of flammable materials and poor ventilation. Combined with operational events, assess the impact of these hazards on the fire risk. Determine a second sub-fire risk range, where the first and second sub-fire risk ranges include multiple operational areas or the entire factory environment.

[0031] The first sub-fire risk range and the second sub-fire risk range are synthesized to form a more comprehensive fire risk assessment map. At this time, the two risk ranges are superimposed using a geographic information system (GIS) or similar spatial analysis tools. Based on the superposition results, multiple specific fire risk areas are determined. Each risk area is classified into risk levels, such as high risk, medium risk, low risk, etc.

[0032] Specifically, the unmanned factory has multiple unmanned operation areas, each equipped with cameras and infrared cameras for real-time monitoring. To improve fire warning capabilities, the factory decided to implement a fire risk assessment system based on real-time monitoring and image recognition. In the "unmanned factory," cameras and infrared cameras are installed in key operation and environmental areas, such as production lines, warehouses, and material handling areas. The video surveillance system collects real-time operation and environmental images of these areas and stores them in a central database.

[0033] Determining the first sub-fire risk range: The system detected that a material conveyor belt in the material handling area was overheating, and the infrared camera captured localized high temperatures in the area. Combining the operation event (the material conveyor belt was overheating) and the real-time operation image (the localized high temperature phenomenon), the system determined that a specific area in the material handling area was the first sub-fire risk range.

[0034] Determining the second sub-fire risk range: Environmental images showed that a large amount of flammable materials were piled up in the warehouse area and the ventilation conditions were poor. Combined with operational events (no directly related events, but considering that the accumulation of flammable materials and poor ventilation are potential hazards), the system assessed the warehouse area as the second sub-fire risk range.

[0035] The system superimposes the first and second sub-fire risk ranges to form a more comprehensive fire risk assessment map. Based on the superposition results, the system identifies multiple specific fire risk areas, including overheated areas in material handling areas and areas where flammable materials are accumulated in warehouses. Each risk area is classified into a risk level, such as overheated areas in material handling areas as high-risk areas and areas where flammable materials are accumulated in warehouses as medium-risk areas. By implementing this fire risk assessment system based on real-time monitoring and image recognition, "unmanned factories can more accurately identify fire risk areas and take corresponding preventive measures, thereby improving factory safety and production efficiency."

[0036] In some embodiments of the present application, a fire risk range matching table preset for an unmanned factory is collected, and the fire risk range matching table is shown in Table 1: Table 1 Fire risk range matching table

[0037] In this fire risk range matching table, each row represents an operation event and its corresponding image features and fire risk range. For example, when an equipment overheating alarm sounds, the system identifies the local high-temperature area through real-time operation images and, combined with environmental images (no obvious abnormalities), determines that the first sub-fire risk range is the area where the equipment is located. At the same time, considering the chain reaction caused by overheating, the system will also include the affected adjacent areas in the second sub-fire risk range.

[0038] refer to Figure 4 In step S13, in each fire risk area, a plurality of fire risk nodes are determined according to the detection of each fire risk area; and a sub-risk level of the fire risk area is determined based on the position and temperature data of the plurality of fire risk nodes; In the specific implementation process of the present invention, the specific steps are: S131: Monitor each fire risk area in real time, determine a corresponding regional detection method based on the regional location, regional shape, and surrounding environment of the fire risk area, and determine multiple fire risk marker positions based on the fire risk area and the corresponding regional detection method; S132: Determine multiple fire risk nodes based on the screening of multiple fire risk mark positions, mark the positions of the multiple fire risk nodes, and trigger corresponding temperature detection based on the multiple fire risk nodes to collect temperature data of the multiple fire risk nodes; S133: Determine the temperature difference range based on the comparison of temperature data of multiple fire risk nodes, mark the risk temperature data in the temperature data of multiple fire risk nodes, determine the first risk level coefficient based on the temperature difference range and the position of the fire risk node, determine the second risk level coefficient based on the risk temperature data and the position of the fire risk node, and determine the sub-risk level of the fire risk area based on the mapping relationship between the first risk level coefficient, the second risk level coefficient and the sub-risk level.

[0039] In an embodiment of the present application, the identified fire risk areas are continuously monitored to ensure that any abnormal situation is discovered in time. At this time, real-time monitoring is carried out using monitoring equipment (such as cameras, infrared thermal imagers, smoke detectors, etc.) installed in each fire risk area; the monitoring software is configured and the alarm threshold is set to automatically trigger the alarm when an abnormal situation is detected; the monitoring center should have a dedicated person on duty to monitor the screen and respond to the alarm in a timely manner.

[0040] Select the most appropriate detection method based on the characteristics of the fire risk area (location, shape, and surrounding environment). Analyze the geographical location of the fire risk area and consider whether it is close to flammable materials, crowded areas, or important equipment. Evaluate the area shape, such as area size, height, and whether there are obstructions, as these factors affect the selection and layout of monitoring equipment. Consider the surrounding environment, such as meteorological conditions such as wind direction, wind speed, and humidity, as well as whether there are other potential fire sources or flammable materials. Based on the above analysis, select an appropriate detection method, such as video surveillance, infrared thermal imaging, gas detection, etc.

[0041] Mark key locations within the fire risk area for focused monitoring and risk assessment; determine blind spots or key monitoring areas based on the layout of the fire risk area and the coverage of monitoring equipment; install or mark monitoring equipment at these key locations to ensure full coverage of the fire risk area; marked locations should include potential fire sources, areas where flammable materials accumulate, key nodes of evacuation routes, etc.

[0042] Specifically, suppose there is an unmanned chemical factory whose production area is divided into multiple fire risk areas, including raw material storage areas, finished product warehouses, etc.; in order to ensure safe production, cameras and infrared thermal imagers are installed in key areas such as raw material storage areas and finished product warehouses for 24-hour uninterrupted monitoring; monitoring software is configured and alarm thresholds are set. For example, if the temperature exceeds a certain range or the smoke concentration exceeds the standard, the alarm will be automatically triggered; there is a dedicated person on duty at the monitoring center who is responsible for monitoring the screen and immediately notifies relevant personnel to handle the alarm once an alarm message is received.

[0043] Raw material storage area: Since a large amount of flammable and explosive raw materials are stored there, a combination of infrared thermal imaging and gas detection was chosen. Infrared thermal imaging is used to monitor temperature changes, and gas detection is used to monitor the concentration of combustible gases. Finished product warehouse: Since most finished products are solid and non-flammable, cameras were selected for monitoring, and smoke detectors were installed at the warehouse entrance.

[0044] In the raw material storage area, the surrounding areas and tops of the raw material stacks are marked as key monitoring locations to ensure that fire hazards can be discovered in a timely manner; in the finished product warehouse, the key nodes of the evacuation passages, the surrounding areas of the cargo stacks, and the entrances and exits of the warehouse are marked; through such implementation, the unmanned factory can monitor each fire risk area in real time, select appropriate detection methods based on regional characteristics, and mark key locations for key monitoring, thereby effectively improving the ability to prevent fires and respond to emergencies.

[0045] Furthermore, key nodes where fire occurs or spreads are screened out from the marked fire risk locations; at this point, the characteristics of each marked location are analyzed, such as whether it is close to flammable materials, whether there are electrical lines crossing, whether there are historical fire records, etc.; combined with the fire risk assessment model or expert experience, the probability and consequences of fire at each location are evaluated; based on the assessment results, nodes with higher fire risks are screened out as key monitoring targets.

[0046] Ensure that monitoring and emergency response personnel can quickly locate these key nodes; at this time, place obvious signs or marks on site to indicate the location and number of fire risk nodes; set up corresponding monitoring screens and alarm information for each risk node in the monitoring system to facilitate real-time monitoring and rapid response; update the fire emergency plan to ensure that fires can be quickly located and measures can be taken when they occur.

[0047] Real-time temperature data from fire risk nodes is collected to promptly identify potential fire hazards. Temperature sensors are installed at each fire risk node to ensure accurate temperature measurement at the node. A temperature detection system is configured, connecting the sensors to the monitoring center to transmit temperature data in real time. Temperature alarm thresholds are set, and when the temperature exceeds the preset value, an alarm is automatically triggered, notifying relevant personnel to handle the situation. Furthermore, the monitoring center continuously receives and stores temperature data from each fire risk node. This temperature data is regularly analyzed to assess changing trends in fire risk. In the event of a fire, temperature data from risk nodes is quickly acquired to provide a basis for emergency decision-making.

[0048] Specifically, an unmanned chemical plant has marked multiple fire risk locations in step S131. It now proceeds to step S132 to identify fire risk nodes and trigger temperature detection. The unmanned chemical plant analyzes the marked risk locations and finds that a stack in the raw material storage area is close to electrical wiring and has a history of fires. Therefore, it identifies it as fire risk node A. A stack in the finished product warehouse is identified as fire risk node B because it is stacked too high and close to a wall, resulting in poor ventilation. Clear signboards are set up in the raw material storage area, reactor area and finished product warehouse to indicate the location and number of fire risk nodes A and B; a corresponding monitoring screen is set up for each risk node in the monitoring system to ensure that monitoring personnel can monitor the status of these nodes in real time; the fire emergency plan has been updated to clarify how to quickly locate and take measures to deal with these risk nodes in the event of a fire.

[0049] Temperature sensors are installed at fire risk nodes A and B and connected to the monitoring center; a temperature detection system is configured to transmit temperature data to the monitoring center in real time, and a temperature alarm threshold is set; when the temperature exceeds the preset value, the system automatically triggers an alarm, and the monitoring personnel can immediately see the alarm information and take measures; the monitoring center continuously receives and stores temperature data from fire risk nodes A and B; the temperature data is regularly analyzed to evaluate the changing trend of fire risk. If the temperature of a node is found to continue to rise, timely measures are taken to deal with it; when a fire occurs, the temperature data of the risk node is quickly obtained to provide a basis for emergency decision-making, such as deciding whether to initiate emergency evacuation procedures or call fire-fighting equipment for fire extinguishing; through such implementation, unmanned chemical plants can accurately determine fire risk nodes and collect temperature data in real time, providing strong support for fire prevention and emergency response.

[0050] Therefore, the temperature difference range is determined based on the comparison of temperature data of multiple fire risk nodes, and the risk temperature data in the temperature data of multiple fire risk nodes are marked. The first risk level coefficient is determined according to the temperature difference range and the position of the fire risk node, and the second risk level coefficient is determined according to the risk temperature data and the position of the fire risk node. The sub-risk level of the fire risk area is determined based on the mapping relationship between the first risk level coefficient, the second risk level coefficient and the sub-risk level, which is compatible with the overall consideration of the first risk level coefficient, the second risk level coefficient and the sub-risk level mapping relationship, thereby ensuring the accuracy of the sub-risk level of the fire risk area. At the same time, targeted management and control of multiple fire risk nodes are carried out, which is compatible with the overall consideration of the position and temperature data of multiple fire risk nodes, thereby improving the accuracy of the sub-risk level of the fire risk area.

[0051] At this time, by comparing the temperature data of different fire risk nodes, the range of temperature anomalies, that is, the temperature difference range, is identified; at this time, the temperature data of all fire risk nodes are collected to ensure the accuracy and timeliness of the data; the data is statistically analyzed to calculate the temperature mean, standard deviation and other statistical quantities; based on the statistical results and the preset temperature threshold, the temperature difference range is determined, that is, which nodes have temperature data outside the normal range.

[0052] Identify data indicating fire risk, i.e., risk temperature data, from temperature data; compare the temperature data with the preset alarm threshold; mark temperature data exceeding the alarm threshold as risk temperature data; and record information such as the time and node location of the risk temperature data for subsequent analysis.

[0053] Determine the first risk level coefficient based on the temperature difference range and the location of the fire risk node, evaluate the impact of the temperature difference range on the fire risk, and determine the first risk level coefficient accordingly; at this time, analyze the relationship between the temperature difference range and the fire risk, such as the greater the temperature difference, the higher the fire risk; consider the location of the fire risk node, such as nodes close to flammable materials, electrical equipment or crowded areas have higher risks; combine the temperature difference range and node location, and determine the first risk level coefficient based on the preset risk assessment model or expert experience.

[0054] Determine the second risk level coefficient based on the risk temperature data and the location of the fire risk node, evaluate the specific impact of the risk temperature data on the fire risk, and determine the second risk level coefficient accordingly; at this time, analyze the value, duration and change trend of the risk temperature data; consider the node position corresponding to the risk temperature data, such as the more critical the node position, the higher the risk level; combine the risk temperature data and node position, and determine the second risk level coefficient based on the preset risk assessment model or expert experience.

[0055] The sub-risk level of the fire risk area is determined by combining the first risk level coefficient and the second risk level coefficient. At this time, the sub-risk level mapping relationship is pre-defined, such as combining the first risk level coefficient and the second risk level coefficient into different level intervals. According to the actual values ​​of the first risk level coefficient and the second risk level coefficient, the corresponding sub-risk level is found in the sub-risk level mapping relationship. The sub-risk level of the fire risk area is recorded and updated for subsequent risk management and emergency response.

[0056] Specifically, an unmanned operation has identified multiple fire risk nodes in step S132 and collected temperature data in real time; now enters step S133 to determine the sub-risk level of the fire risk area based on the temperature data; the unmanned operation has collected the temperature data of all fire risk nodes and conducted statistical analysis; it is found that the node temperature in a certain raw material storage area is generally high, and the temperature difference with other nodes is large, exceeding the preset temperature threshold range; the temperature data of the raw material storage area node is compared with the alarm threshold, and the temperature data exceeding the threshold is marked as risk temperature data; the time, node location and other information of the risk temperature data are recorded.

[0057] Analyze the relationship between the temperature difference range and the fire risk. Considering that the temperature of the nodes in the raw material storage area is generally high and the difference is large, and the area is close to flammable materials, the first risk level coefficient is determined to be high; analyze the value, duration and change trend of the risk temperature data, and find that the temperature of a certain node continues to rise, and the node is located in the center of the raw material storage area, so the second risk level coefficient is also determined to be high; according to the pre-defined sub-risk level mapping relationship, the first risk level coefficient and the second risk level coefficient are combined into a high-level interval; the sub-risk level of the raw material storage area is determined to be extremely high, and immediate measures need to be taken to intervene, such as strengthening monitoring, adding fire-fighting facilities, and restricting personnel entry; through such implementation, no one can accurately assess the fire risk based on temperature data and determine the sub-risk level of the fire risk area, providing strong support for subsequent risk management and emergency response.

[0058] In some embodiments of the present application, the temperature difference range has a weight of 0.5; the risk temperature data (value, duration, and change trend) has a weight of 0.3; and the node location has a weight of 0.2. Assuming the temperature difference range is 12°C, according to the preset scoring standard (e.g., 0.1 points for each 1°C increase), the score is 1.2 (12×0.1). If the risk temperature data value exceeds the preset threshold by 20°C for 1 hour, and the change trend continues to rise, according to the scoring standard (e.g., 0.5 points for each 10°C increase above the threshold, 0.2 points for each hour of duration, and 0.3 points for an upward trend), the score is 1.7 (0.5×2+0.2×1+0.3). If the node is located near flammable materials, according to the scoring standard (e.g., 0.6 points for being near flammable materials), the score is 0.6. Total score = 1.2×0.5 + 1.7×0.3 + 0.6×0.2 = 1.31; Based on the preset correspondence between the score range and the risk level (e.g., a total score > 1.2 is high risk), the sub-risk level of this node is determined to be "high risk".

[0059] refer to Figure 5 In step S14, multiple fire inspection areas of the fire humanoid robot are determined according to the sub-risk level of each fire risk area and the current position of the fire humanoid robot relative to the unmanned factory, and the sub-fire inspection paths of each fire inspection area are marked; In the specific implementation process of the present invention, the specific steps are: S141: When the firefighting humanoid robot is in the unmanned factory, determine the current position of the firefighting humanoid robot relative to the unmanned factory based on the position of the firefighting humanoid robot and a distribution map of the unmanned factory; and determine a first sub-firefighting inspection range based on the sub-risk level of each fire risk area and the distribution map of the unmanned factory. S142: Determine a second sub-fire inspection range based on the location of each fire risk area and the current location of the firefighting humanoid robot relative to the unmanned factory; and determine multiple fire inspection areas for the firefighting humanoid robot based on a combination of the first sub-fire inspection range and the second sub-fire inspection range. S143: In each fire inspection area, multiple operation images of the fire inspection area are monitored in real time, the operation range of the fire inspection area is determined based on the recognition of the multiple operation images of the fire inspection area, and the sub-fire inspection path of the fire inspection area is determined based on the operation range of the fire inspection area, the area morphology of the fire inspection area and the degree of freedom of movement of the fire humanoid robot, so as to mark the sub-fire inspection path of the fire inspection area.

[0060] In an embodiment of the present application, the fire-fighting humanoid robot needs to collect its own location data; this is usually achieved through a built-in GPS module, an inertial navigation system (INS) or other positioning technologies (such as RFID, UWB, etc.); at the same time, the robot also needs to obtain a distribution map of the unmanned factory, which is an electronic map, a CAD drawing or a specially designed factory layout map; the collected location data is matched with the distribution map of the unmanned factory; this usually involves converting the robot's coordinates (longitude, latitude or relative position) into specific points on the distribution map; for systems using relative positioning technology, a known reference point or landmark is required to calibrate the position; once the location data is successfully matched with the distribution map, the fire-fighting humanoid robot can determine its specific location within the factory; this is usually presented to the operator in the form of a graphical interface or displayed on the robot's display screen.

[0061] Optionally, assume that a fire-fighting humanoid robot is performing an inspection task in an unmanned factory; the robot obtains its own precise location coordinates (for example, latitude 34.0522°, longitude -118.2437°) through its built-in GPS module; at the same time, the robot also loads a distribution map of the factory, which shows the locations of each workshop, warehouse, and aisle in detail; through a position matching algorithm, the robot converts its location coordinates into a specific point on the distribution map and finds that it is located at the entrance of the main production workshop; this location information is displayed in real time on the robot's display screen and is also transmitted to the operator of the remote monitoring center.

[0062] It is necessary to obtain the sub-risk level of each fire risk area; these levels are usually derived from a comprehensive assessment of factors such as historical fire data, equipment type, and stored materials; the sub-risk level is qualitative (such as high, medium, and low) or quantitative (such as a specific numerical score); based on the sub-risk level and the distribution map of the unmanned factory, the first sub-fire inspection range is determined; this usually involves marking high-risk areas as priority inspection objects and drawing a reasonable boundary around these areas to form the inspection range; the demarcation of the boundary should take into account factors such as the robot's mobility, factory layout, and obstacle location; after determining the inspection range, further optimization and adjustment are required; for example, if a high-risk area is far away from the robot's current position, it needs to be removed from the first sub-inspection range, or the boundary needs to be adjusted to ensure that the robot can reach and inspect the area efficiently.

[0063] Optionally, the fire-fighting humanoid robot has determined its own position in the unmanned chemical factory; next, it determines the first sub-fire inspection range based on the sub-risk level of each fire risk area; assuming that there are two main fire risk areas in the factory: one is the warehouse where flammable chemicals are stored (sub-risk level is high), and the other is some old equipment on the production line (sub-risk level is medium); based on these levels and distribution maps, the robot marks the warehouse where flammable chemicals are stored as a priority inspection object, and draws a reasonable boundary around the warehouse as the first sub-fire inspection range; when determining the range, the robot also takes into account the width of the aisles around the warehouse, the location of obstacles, and its own mobility; finally, it draws a boundary range that includes the warehouse itself and is convenient for the robot to enter and inspect; this range is displayed in real time on the robot's display screen and transmitted to the operator of the remote monitoring center for further monitoring and guidance.

[0064] Furthermore, the precise location information of each fire risk area is obtained; this information usually comes from the layout diagram of the unmanned factory, CAD drawings or special fire risk assessment reports; the location of each risk area should be represented by its center point or a representative point, and accompanied by corresponding coordinates or relative position descriptions; in step S141, the current position of the fire-fighting humanoid robot relative to the unmanned factory has been determined; this step needs to ensure the accuracy and real-time nature of the location information for subsequent path planning and range determination.

[0065] Based on the robot's mobility (such as speed, endurance, steering flexibility, etc.) and current position, the range of areas that it can efficiently reach and inspect is calculated; this range is usually a circular or elliptical area centered on the robot's current position and with a certain distance as the radius, but it is also adjusted according to the factory layout and obstacle location; the regional location of each fire risk area is compared with the robot's reachable range to determine which areas are within the reachable range; these areas will constitute the second sub-fire inspection range; when determining the range, the priority of the risk areas, the adjacent relationship and the inspection order also need to be considered.

[0066] Specifically, assume that the fire-fighting humanoid robot is located in the main production area of ​​an unmanned chemical factory, and its current location is known; based on the factory layout map, three main fire risk areas are identified: A (chemical warehouse), B (old production line) and C (high-voltage power room); the robot calculates a reachable range based on its mobility and current location, which covers most of the main production area but does not extend to the remote corners of the factory; by comparing the locations of the risk areas and the robot's reachable range, it is found that areas A (chemical warehouse) and B (old production line) are both within the reachable range, while area C (high-voltage power room) is relatively far away; therefore, the robot identifies areas A and B as the second sub-fire inspection range; at the same time, it also takes into account that area A has a higher priority (because flammable chemicals are stored there), so it will be inspected first.

[0067] The first sub-fire inspection range and the second sub-fire inspection range are synthesized to form a comprehensive inspection area map; during the synthesis process, attention should be paid to the overlapping parts and uncovered areas of the range to ensure that all important fire risk areas are included in the inspection range; based on the synthesized inspection area map, the inspection area is divided into multiple specific small areas that are convenient for robots to inspect; these small areas are based on the natural divisions of the factory layout (such as workshops, warehouses, etc.), and are also artificially divided according to the robot's inspection path and efficiency; each inspection area is prioritized to guide the robot's inspection order; the priority is usually determined based on a combination of factors such as risk level, urgency, the robot's current location and accessibility.

[0068] Specifically, the first sub-fire inspection range (mainly around chemical warehouse A) and the second sub-fire inspection range (including chemical warehouse A and old production line B) have been determined; next, the two ranges are synthesized to form a comprehensive inspection area map; during the synthesis process, it is found that area A is the overlapping part of the two ranges, and is therefore particularly important; at the same time, it is also noted that although area B is located within the second sub-inspection range, it is relatively close to the robot's current position, and therefore also has a higher inspection priority; based on these factors, the inspection area is divided into two specific small areas: Area 1 (chemical warehouse A and its surroundings) and Area 2 (old production line B and its nearby areas); at the same time, the inspection priority order is determined to be Area 1 first and then Area 2; in this way, the fire humanoid robot has a clear inspection target and path planning, and performs inspection tasks efficiently.

[0069] Therefore, in each fire inspection area, multiple operation images of the fire inspection area are monitored in real time, the operation range of the fire inspection area is determined based on the identification of multiple operation images of the fire inspection area, and the sub-fire inspection path of the fire inspection area is determined based on the operation range of the fire inspection area, the regional form of the fire inspection area and the freedom of movement of the fire humanoid robot, so as to mark the sub-fire inspection path of the fire inspection area, which is compatible with the overall consideration of the operation range of the fire inspection area, the regional form of the fire inspection area and the freedom of movement of the fire humanoid robot, and ensures the accuracy of the sub-fire inspection path of the fire inspection area.

[0070] At this time, the fire-fighting humanoid robot needs to be equipped with cameras or other image acquisition devices to monitor the dynamics within the inspection area in real time. These cameras should have a wide-angle field of view, night vision function (if inspections are required at night), and sufficient resolution to capture details. The collected image data needs to be transmitted in real time to the robot's control system or remote monitoring center. This is usually achieved through wireless networks (such as Wi-Fi, 4G / 5G) to ensure the timeliness and reliability of the data. In the control system or monitoring center, the operator views the images of the inspection area in real time to promptly detect abnormal situations or potential risks.

[0071] Optionally, in an unmanned chemical plant, a fire-fighting humanoid robot is performing inspection tasks; it is equipped with multiple cameras, which are distributed in different parts of the robot to ensure all-round field of view coverage; when the robot enters the chemical warehouse inspection area, the camera begins to collect images inside the warehouse in real time and transmits them to the remote monitoring center via the wireless network; the operator can clearly see the shelf layout, chemical storage conditions and any abnormal conditions (such as leakage, accumulation, etc.) in the warehouse on the monitoring screen.

[0072] Computer vision technology is used to identify and analyze the collected images; this includes identifying objects, people, equipment and their status (such as whether they are operating normally or whether there are any abnormal signs) in the images; based on the results of image recognition, the scope of work within the inspection area is determined; this usually involves dividing key areas (such as chemical storage areas, equipment operation areas, etc.) and non-key areas (such as corridors, rest areas, etc.); optionally, in the chemical warehouse, the images collected by the fire-fighting humanoid robot's camera are transmitted to the image recognition system; the system identifies key elements in the warehouse such as chemical shelves, safety exits, fire-fighting equipment, and analyzes their status; based on these recognition results, the system determines the scope of work in the warehouse, including key areas such as chemical storage areas, equipment operation areas and safety passages.

[0073] When determining the sub-fire inspection path, multiple factors need to be considered, including the size and shape of the operating range, the obstacles in the area, the robot's mobility (such as speed, steering flexibility, climbing ability, etc.) and freedom of movement (that is, the degree to which the robot can move freely in space); use path planning algorithms (such as Dijkstra algorithm, RRT algorithm, etc.) combined with the above considerations to generate optimal or suboptimal inspection paths; these algorithms can usually take into account multiple indicators such as path length, safety, and accessibility; mark the generated inspection path on the distribution map of the unmanned factory so that the robot can conduct inspections according to the planned path; these marks include the starting point, end point, key turning points of the path, and any obstacles or special areas that require attention.

[0074] Optionally, after determining the operating scope of the chemical warehouse, the fire-fighting humanoid robot begins to plan an inspection route; taking into account the shelf layout in the warehouse, the location of the safety passage and the robot's mobility, the robot uses a path planning method based on the RRT algorithm; this method takes into account multiple factors such as the length, safety and accessibility of the path, and generates an optimal inspection path from the warehouse entrance to each key area; this path is marked on the distribution map and transmitted to the robot; the robot begins inspection according to the planned path, ensuring comprehensive coverage of key areas such as chemical storage areas, equipment operation areas and safety passages.

[0075] In some embodiments of the present application, a sub-fire inspection path matching table is collected. The sub-fire inspection path matching table is used to illustrate how to determine the sub-fire inspection path according to the operating range, regional morphology, and robot activity freedom. The sub-fire inspection path matching table is shown in Table 2: Table 2 Sub-fire inspection path matching table

[0076] refer to Figure 6 In step S15, the inspection order of the multiple sub-fire inspection paths is determined based on the positions of the multiple fire inspection areas, and the real-time fire inspection path of the fire humanoid robot is determined according to the inspection order of the multiple sub-fire inspection paths, the current position of the fire humanoid robot, and the real-time events of each fire inspection area; In the specific implementation process of the present invention, the specific steps are: S151: monitoring the positions of multiple fire inspection areas, collecting real-time images of the multiple fire inspection areas, and determining the real-time status of the fire inspection areas based on recognition of the real-time images of the fire inspection areas; S152: Determine the inspection order of multiple sub-fire inspection paths according to the real-time status and location of each fire inspection area; determine the first fire inspection path based on the inspection order of the multiple sub-fire inspection paths and the current location of the fire humanoid robot, S153: Determine multiple dynamic features based on the detection of real-time images of the fire inspection area, and determine real-time events of the fire inspection area based on the synthesis of the multiple dynamic features, determine a second fire inspection path based on the inspection sequence of multiple sub-fire inspection paths and the real-time events of each fire inspection area, and determine the real-time fire inspection path of the fire humanoid machine based on the dynamic synthesis of the first fire inspection path and the second fire inspection path.

[0077] In an embodiment of the present application, the locations of multiple fire inspection areas are monitored, and real-time images of multiple fire inspection areas are collected. The real-time status of the fire inspection areas is determined based on the recognition of the real-time images of the fire inspection areas, which is compatible with the overall consideration of the recognition of real-time images of the fire inspection areas and ensures the accuracy of the real-time status of the fire inspection areas.

[0078] At this time, monitoring systems are deployed in multiple fire inspection areas. These systems include GPS locators, wireless signal transmitters, cameras or other equipment that can determine the location of the area; the monitoring system collects the location data of each inspection area in real time; this data includes latitude and longitude coordinates, distance and direction relative to a fixed point, etc.; the collected location data is transmitted to the central control room or the control system of the fire humanoid robot via wired or wireless means; the control system continuously updates the location information of each inspection area to ensure the real-time and accuracy of the data.

[0079] Specifically, in an unmanned chemical plant, multiple fire inspection areas are deployed; each area is equipped with a GPS locator and a wireless signal transmitter; these devices collect their respective location data in real time and transmit it to the central control room via a wireless network; the software system in the control room continuously updates this location information and displays the exact location of each inspection area on the map in the form of icons or marks.

[0080] Cameras are installed in each fire inspection area to ensure that real-time images of the area can be captured; the cameras collect images of the inspection area in real time and transmit this image data to the control system via wired or wireless means; the image data received by the control system is stored in a local server or cloud storage system for subsequent analysis and processing; to ensure the clarity and accuracy of the image data, the control system will regularly check the operating status and image quality of the camera.

[0081] Specifically, in unmanned factories, cameras are installed in each fire inspection area; these cameras have night vision capabilities and wide-angle fields of view, and can capture real-time images of the inspection area around the clock; the image data collected by the cameras is transmitted to the central control room via a wireless network and stored in a cloud storage system; operators in the control room can view these image data at any time to monitor the status of the inspection area.

[0082] Computer vision technology is used to identify and analyze the collected real-time images. This includes identifying potential risk elements such as objects, human activities, flames, smoke, etc. in the images. Based on the results of image recognition, combined with preset rules or algorithms, the real-time status of the inspection area is determined. These statuses include normal, warning, emergency, etc. The control system continuously updates the real-time status of each inspection area and presents this information in a visual manner to the operator or fire-fighting humanoid robot. If an emergency state (such as fire, personnel evacuation, etc.) is identified, the control system will immediately trigger the alarm mechanism and notify relevant personnel to respond.

[0083] Optionally, in an unmanned chemical plant, operators in the central control room view image data of each inspection area in real time through the monitoring system; using computer vision technology, the system recognizes smoke in the chemical warehouse area and determines that the area is in an emergency state; the control system immediately triggers the alarm mechanism, notifying the fire-fighting humanoid robot and other personnel in the unmanned chemical plant to go to the chemical warehouse for emergency treatment; at the same time, the control system also marks the location of the emergency in a striking manner on the map, so that operators can quickly understand the safety status of the entire park.

[0084] Furthermore, the inspection order of multiple sub-fire inspection paths is determined according to the real-time status and position of each fire inspection area; the first fire inspection path is determined based on the inspection order of multiple sub-fire inspection paths and the current position of the fire humanoid robot, which is compatible with the overall consideration of the inspection order of multiple sub-fire inspection paths and the current position of the fire humanoid robot, ensuring the accuracy of the first fire inspection path.

[0085] At this point, the system needs to collect and analyze the real-time status and location information of each fire inspection area; this includes whether there are abnormal conditions such as fire, smoke, and gathering of people in each area, as well as the positional relationship of these areas relative to each other and the fire humanoid robot; based on the real-time status and location information, the system conducts a priority assessment on each inspection area; generally, areas that are in an emergency or have high-risk hidden dangers are given higher priority; in addition, areas that are close to the robot's current location or easy to reach are also given priority; based on the priority assessment results, the system determines the inspection order of multiple sub-fire inspection paths; this order is designed to ensure that the robot can visit each area efficiently and orderly, giving priority to emergency and high-risk situations.

[0086] Specifically, assume that there are three fire inspection areas in an unmanned chemical factory: Area A (chemical warehouse, smoke appears, emergency state), Area B (old production line, normal state), and Area C (distribution room, circuit aging risk, warning state); the fire humanoid robot is currently located near the entrance of the unmanned chemical factory; the system first analyzes the real-time status and location information of these three areas; due to the appearance of smoke in Area A, it is judged to be an emergency state and is therefore given the highest priority; Area C has the risk of circuit aging and is judged to be a warning state, with the second highest priority; Area B is in normal state and has the lowest priority; based on these priority evaluation results, the system determines the inspection order: first go to Area A to deal with the emergency, then go to Area C to check the circuit aging, and finally visit Area B for routine inspections.

[0087] After determining the inspection order, the system needs to plan a path for the robot starting from the current position and visiting each area in the inspection order; this usually involves path optimization algorithms, such as Dijkstra algorithm, A* algorithm, etc., to ensure the efficiency and safety of the path; when planning the path, the system also needs to consider factors such as obstacles, narrow passages, and potential risks in the unmanned factory; these factors affect the robot's movement speed and direction, so corresponding adjustments need to be made in path planning; taking into account factors such as the inspection order, current position, obstacles and risks, the system finally determines the first fire inspection path; this path should ensure that the robot can safely and efficiently visit each area in the inspection order.

[0088] Specifically, after determining the inspection sequence (Area A → Area C → Area B), the system begins to plan the first fire inspection route for the fire-fighting humanoid robot; considering the layout and obstacles in the unmanned factory, the system selects a route that avoids crowded areas and narrow passages; this route starts from the entrance of the unmanned factory and first goes to Area A to deal with emergencies; after dealing with the situation in Area A, the robot will follow the planned route to Area C to check for circuit aging; finally, the robot will visit Area B for routine inspections and eventually return to the entrance of the unmanned factory or a designated safe area.

[0089] Therefore, multiple dynamic features are determined based on the detection of real-time images of the fire inspection area, and real-time events of the fire inspection area are determined based on the synthesis of multiple dynamic features. The second fire inspection path is determined based on the inspection order of multiple sub-fire inspection paths and the real-time events of each fire inspection area. The real-time fire inspection path of the fire humanoid machine is determined based on the dynamic synthesis of the first fire patrol path and the second fire patrol path. The overall consideration of the dynamic synthesis of the first fire patrol path and the second fire patrol path is compatible, ensuring the accuracy of the real-time fire inspection path of the fire humanoid machine. At the same time, the sub-fire patrol paths of each fire inspection area are introduced, and the overall consideration of the inspection order of multiple sub-fire patrol paths, the current position of the fire humanoid robot and the real-time events of each fire patrol area is realized, ensuring the dynamic effectiveness of the real-time fire inspection path of the fire humanoid machine.

[0090] At this time, the fire-fighting humanoid robot uses its onboard camera to collect images of each inspection area in real time; these images are then transmitted to the robot's control system for pre-processing, such as denoising and contrast enhancement, to improve the accuracy of subsequent analysis; using computer vision technology, the system performs dynamic feature detection on the pre-processed images; these features include the diffusion trend of flames and smoke, the movement direction of people, and the abnormal state of objects (such as leakage, collapse), etc.; the system identifies these features through algorithms and extracts their key information, such as position, size, speed, etc.; the detected dynamic features are recorded in the system and continuously updated as the images are updated in real time; this ensures that the system can continuously track dynamic changes in the inspection area.

[0091] Specifically, in a large shopping mall, a fire-fighting humanoid robot is performing inspection tasks; the camera collects images of each floor of the shopping mall in real time; the system detects smoke in a shop area on the second floor and identifies the diffusion trend and speed of the smoke; at the same time, the system also detects a crowd gathering on the third floor and analyzes the movement direction and speed of the crowd; these dynamic features are recorded and updated in the system in real time.

[0092] Switch the scene and change the "large shopping mall" to an "unmanned factory".

[0093] The system synthesizes multiple detected dynamic features to form a comprehensive understanding of the overall situation in the inspection area; this involves analyzing the relationship between features, such as the relationship between smoke diffusion and fire sources, the relationship between crowd gathering and evacuation channels, etc.; based on the feature synthesis results, the system uses preset rules or algorithms to determine real-time events occurring in the inspection area; these events include fires, personnel evacuations, equipment failures, etc.; the system also classifies the events according to their urgency and scope of impact; the determined real-time events are recorded in the system and immediately notified to the fire humanoid robot and shopping center managers; this helps to take timely response measures to reduce potential risks and losses.

[0094] Specifically, in an unmanned factory, the system comprehensively analyzes the smoke diffusion characteristics of a certain production area and the personnel activities in another area (assuming the distribution of technicians or inspectors), determines that a fire or similar emergency has occurred in the factory, and determines that the main area of ​​smoke diffusion is the fire scene, and that there is a need for emergency evacuation in areas with dense personnel distribution; the system immediately records this emergency event and automatically notifies the fire-fighting humanoid robot and the factory's control center or safety management personnel.

[0095] Based on the identified real-time events, the system conducts an event priority assessment for each inspection area; emergency events (such as fires) are given the highest priority and require immediate attention; other events (such as personnel evacuation and equipment failures) are given corresponding priorities based on the degree of urgency and the scope of impact; based on the event priority assessment and the preset inspection sequence, the system adjusts multiple sub-fire inspection paths; this involves rearranging the order in which the inspection areas are visited to ensure that the robot can prioritize emergency events while taking into account the inspection of other important areas; after the adjustment, the system determines the second fire inspection path; this path is designed to ensure that the robot can respond to real-time events efficiently and orderly while maintaining monitoring of other important areas.

[0096] Specifically, in the unmanned factory, the system made immediate adjustments to the inspection route; in view of a fire incident in a certain production area, that area was automatically given the highest priority, so the fire-fighting humanoid robot was quickly instructed to go to that area to deal with the fire; at the same time, the system detected that technicians or staff were gathered in another area, and considering the need for emergency evacuation, the system planned that the robot would immediately move to that area after completing the fire handling task to assist in personnel evacuation; for routine inspection tasks in other production areas, the system temporarily shelved them until the emergency was effectively responded to and resolved.

[0097] The system dynamically synthesizes the first fire inspection path (a path determined based on the inspection sequence and location information) and the second fire inspection path (a path determined based on real-time events); this involves analyzing and processing the intersections, conflicts and overlapping parts of the two paths; during the synthesis process, the system optimizes and adjusts the path to ensure that the robot can efficiently and safely perform inspection tasks according to the synthesized path; this involves shortening the path length, adjusting the robot's movement speed, etc.; after synthesis and optimization, the system finally determines the real-time fire inspection path of the fire humanoid robot; this path combines the preset inspection sequence and real-time dynamic events, ensuring that the robot can flexibly respond to various situations in complex and changing environments.

[0098] Specifically, the system dynamically synthesized a primary fire inspection route (visiting each production area in a pre-set inspection order) and a secondary fire inspection route (prioritizing fire handling in Area 2 and personnel isolation in Area 3). After optimization and adjustment, the system determined the final real-time fire inspection route: the robot first departed from the unmanned factory entrance and immediately headed to Area 2 to handle the fire. After handling the fire, the robot quickly moved to Area 3 to assist in personnel isolation. Finally, depending on the actual situation, the robot returned to other production areas for inspection or returned to its starting point for standby.

[0099] See also Figure 7 , Figure 7 : This is a schematic diagram of the structure of a fire inspection system for an unmanned factory by a fire-fighting humanoid robot in an embodiment of the present invention; the fire inspection system for an unmanned factory by a fire-fighting humanoid robot includes: The unmanned operation area module 21 is used to collect a distribution map of the unmanned factory and determine a plurality of unmanned operation areas according to the division of the distribution map of the unmanned factory; A fire risk area module 22 is configured to determine, in each unmanned operation area, a plurality of fire risk areas based on the operation data, real-time operation images, and environmental images of the unmanned operation area; The sub-risk level module 23 is configured to determine, in each fire risk area, a plurality of fire risk nodes according to the detection of each fire risk area; and determine a sub-risk level of the fire risk area based on the location and temperature data of the plurality of fire risk nodes; The sub-fire inspection path module 24 is used to determine multiple fire inspection areas of the fire humanoid robot based on the sub-risk level of each fire risk area and the current position of the fire humanoid robot relative to the unmanned factory, and mark the sub-fire inspection path of each fire inspection area; The real-time fire inspection path module 25 is used to determine the inspection order of multiple sub-fire inspection paths based on the positions of multiple fire inspection areas, and determine the real-time fire inspection path of the fire humanoid machine according to the inspection order of multiple sub-fire inspection paths, the current position of the fire humanoid robot and the real-time events of each fire inspection area.

[0100] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

Claims

1. A fire inspection method for an unmanned factory using a fire-fighting humanoid robot, characterized in that: include: Collect a distribution map of unmanned factories, and determine multiple unmanned operation areas based on the division of the distribution map of unmanned factories; In each unmanned operation area, multiple fire risk areas are determined based on the operation data, real-time operation images, and environmental images of the unmanned operation area; In each fire risk area, a plurality of fire risk nodes are determined based on detection of each fire risk area; Determine the sub-risk level of the fire risk area based on the location and temperature data of multiple fire risk nodes; Determine multiple fire inspection areas for the fire humanoid robot based on the sub-risk level of each fire risk area and the current position of the fire humanoid robot relative to the unmanned factory, and mark the sub-fire inspection paths for each fire inspection area; The inspection order of multiple sub-fire inspection paths is determined based on the positions of multiple fire inspection areas, and the real-time fire inspection path of the fire humanoid machine is determined according to the inspection order of multiple sub-fire inspection paths, the current position of the fire humanoid robot and the real-time events of each fire inspection area.

2. The fire inspection method for unmanned factories using a fire-fighting humanoid robot according to claim 1, characterized in that: The collecting of the distribution map of the unmanned factory and determining a plurality of unmanned operation areas according to the division of the distribution map of the unmanned factory include: Collecting a database of unmanned factories, determining a distribution map of the unmanned factories based on the database of the unmanned factories and the names and locations of the unmanned factories, and determining working areas and activity areas based on the detection of the distribution map of the unmanned factories; Based on the detection of the database of the unmanned factory, the current task signal of the unmanned factory is determined, and the current work content of the unmanned factory is determined based on the analysis of the current task signal. Based on the matching of the current work content and the work area of ​​the unmanned factory, multiple unmanned operation areas are determined, and the operation environment and operation content of the multiple unmanned operation areas are marked.

3. The fire inspection method for unmanned factories using a fire-fighting humanoid robot according to claim 1, characterized in that: In each unmanned operation area, determining multiple fire risk areas based on the operation data, real-time operation images, and environmental images of the unmanned operation area includes: Monitor each unmanned operation area in real time, and determine the operation data space corresponding to each unmanned operation area based on the matching of each unmanned operation area with the database of the unmanned factory; In the operation data space corresponding to each unmanned operation area, the operation data of the corresponding unmanned operation area is determined based on the detection of the operation data space, and the corresponding operation event is determined based on the analysis of the operation data and the operation content of the unmanned operation area; Real-time monitoring of unmanned operation areas, and collection of real-time operation images and environmental images corresponding to the unmanned operation areas, determining a first sub-fire risk range based on operation events and real-time operation images, determining a second sub-fire risk range based on operation events and environmental images, and determining multiple fire risk areas based on the synthesis of the first sub-fire risk range and the second sub-fire risk range, where the multiple fire risk areas are located in the same unmanned factory.

4. The fire inspection method for an unmanned factory using a fire-fighting humanoid robot according to claim 1, characterized in that: Determining, in each fire risk area, a plurality of fire risk nodes based on detection of each fire risk area; The sub-risk level of the fire risk area is determined based on the location and temperature data of multiple fire risk nodes, including: Monitor each fire risk area in real time, determine the corresponding regional detection method based on the regional location, regional shape and surrounding environment of the fire risk area, and determine multiple fire risk mark locations based on the fire risk area and the corresponding regional detection method; Based on the screening of multiple fire risk mark positions, multiple fire risk nodes are determined, and the positions of the multiple fire risk nodes are marked. Corresponding temperature detection is triggered based on the multiple fire risk nodes to collect temperature data of the multiple fire risk nodes.

5. The fire inspection method for unmanned factories using a fire-fighting humanoid robot according to claim 4, characterized in that: Determining, in each fire risk area, a plurality of fire risk nodes based on detection of each fire risk area; Determine the sub-risk level of a fire risk area based on the location and temperature data of multiple fire risk nodes, including: The temperature difference range is determined based on the comparison of temperature data of multiple fire risk nodes, and the risk temperature data in the temperature data of multiple fire risk nodes are marked. The first risk level coefficient is determined according to the temperature difference range and the position of the fire risk node, and the second risk level coefficient is determined according to the risk temperature data and the position of the fire risk node. The sub-risk level of the fire risk area is determined based on the mapping relationship between the first risk level coefficient, the second risk level coefficient and the sub-risk level.

6. The fire inspection method for an unmanned factory using a fire-fighting humanoid robot according to claim 1, characterized in that: The method of determining multiple fire inspection areas of the fire humanoid robot according to the sub-risk level of each fire risk area and the current position of the fire humanoid robot relative to the unmanned factory, and marking the sub-fire inspection paths of each fire inspection area, includes: When the fire-fighting humanoid robot is in the unmanned factory, the current position of the fire-fighting humanoid robot relative to the unmanned factory is determined based on the position of the fire-fighting humanoid robot and the distribution map of the unmanned factory; and the first sub-fire inspection range is determined based on the sub-risk level of each fire risk area and the distribution map of the unmanned factory. The second sub-fire inspection range is determined according to the regional location of each fire risk area and the current position of the fire humanoid robot relative to the unmanned factory; and multiple fire inspection areas of the fire humanoid robot are determined based on the synthesis of the first sub-fire inspection range and the second sub-fire inspection range.

7. The fire inspection method for an unmanned factory using a fire-fighting humanoid robot according to claim 6, characterized in that: The method of determining multiple fire inspection areas of the fire humanoid robot according to the sub-risk level of each fire risk area and the current position of the fire humanoid robot relative to the unmanned factory, and marking the sub-fire inspection paths of each fire inspection area, further includes: In each fire inspection area, multiple operation images of the fire inspection area are monitored in real time, and the operation range of the fire inspection area is determined based on the identification of multiple operation images of the fire inspection area. The sub-fire inspection path of the fire inspection area is determined based on the operation range of the fire inspection area, the area morphology of the fire inspection area and the freedom of movement of the fire humanoid robot to mark the sub-fire inspection path of the fire inspection area.

8. The fire inspection method for an unmanned factory using a fire-fighting humanoid robot according to claim 1, characterized in that: The method of determining the inspection order of the plurality of sub-fire inspection paths based on the positions of the plurality of fire inspection areas, and determining the real-time fire inspection path of the fire humanoid robot according to the inspection order of the plurality of sub-fire inspection paths, the current position of the fire humanoid robot, and the real-time events of each fire inspection area, includes: The positions of multiple fire inspection areas are monitored, and real-time images of the multiple fire inspection areas are collected. The real-time status of the fire inspection areas is determined based on the recognition of the real-time images of the fire inspection areas.

9. The fire inspection method for an unmanned factory using a fire-fighting humanoid robot according to claim 8, characterized in that: The method further includes determining the inspection order of the plurality of sub-fire inspection paths based on the positions of the plurality of fire inspection areas, and determining the real-time fire inspection path of the fire humanoid robot according to the inspection order of the plurality of sub-fire inspection paths, the current position of the fire humanoid robot, and the real-time events of each fire inspection area. Determine the inspection order of multiple sub-fire inspection paths according to the real-time status and location of each fire inspection area; determine the first fire inspection path based on the inspection order of multiple sub-fire inspection paths and the current location of the fire humanoid robot, Based on the detection of real-time images of the fire inspection area, multiple dynamic features are determined, and real-time events in the fire inspection area are determined based on the synthesis of multiple dynamic features. A second fire inspection path is determined based on the inspection sequence of multiple sub-fire inspection paths and the real-time events of each fire inspection area. The real-time fire inspection path of the fire humanoid machine is determined based on the dynamic synthesis of the first fire inspection path and the second fire inspection path.

10. A fire inspection system for unmanned factories using a fire-fighting humanoid robot, characterized in that: The fire inspection system for unmanned factories by a fire-fighting humanoid robot is applied to the fire inspection method for unmanned factories by a fire-fighting humanoid robot as claimed in any one of claims 1 to 9. The fire inspection system for unmanned factories by a fire-fighting humanoid robot comprises: The unmanned operation area module is used to collect the distribution map of the unmanned factory and determine multiple unmanned operation areas according to the division of the distribution map of the unmanned factory; A fire risk area module is used to determine multiple fire risk areas in each unmanned operation area based on the operation data, real-time operation images and environmental images of the unmanned operation area; A sub-risk level module is used to determine a plurality of fire risk nodes in each fire risk area based on the detection of each fire risk area; and determine the sub-risk level of the fire risk area based on the location and temperature data of the plurality of fire risk nodes; The sub-fire inspection path module is used to determine multiple fire inspection areas of the fire humanoid robot based on the sub-risk level of each fire risk area and the current position of the fire humanoid robot relative to the unmanned chemical plant, and mark the sub-fire inspection paths of each fire inspection area; The real-time fire inspection path module is used to determine the inspection order of multiple sub-fire inspection paths based on the locations of multiple fire inspection areas, and to determine the real-time fire inspection path of the fire humanoid machine according to the inspection order of multiple sub-fire inspection paths, the current location of the fire humanoid robot and the real-time events of each fire inspection area.

Images