Fire-fighting facility detection system and method based on Internet of Things and artificial intelligence

Abstract

The invention relates to the technical field of fire-fighting facility detection, and provides a fire-fighting facility detection system based on Internet of Things and artificial intelligence, a data area module is used for inputting basic data and classifying fire-fighting facilities according to positions and systems to generate different system data areas; fire-fighting facility operation data and linkage data are collected, and picture data and video data are uploaded to system data areas corresponding to the data; the data analysis module is used for matching the recorded or collected data of each sub-item technical requirement with a judgment index of a technical standard through a fire-fighting facility detection multi-mode large language model, and judging whether the data is qualified or not; the data judgment module is used for comprehensively judging whether the technical requirements are qualified or not according to judgment results of the technical requirements of all the sub-items; the detection report module is used for generating and printing a report. Data can be automatically called, an analysis model is entered to classify, process and convert the data, relevant technical requirements of standards are matched, and each item of data is intelligently judged and configured to automatically generate an analysis report.

Description

Technical Field

[0001] The present invention relates to the technical field of fire protection facility detection, and in particular to a fire protection facility detection system and method based on the Internet of Things and artificial intelligence. Background Art

[0002] Building fire protection facilities refer to the general term for equipment and facilities such as automatic fire alarm systems, automatic water sprinkler systems, and fire hydrant systems installed in buildings (structures) for preventing and fighting fires in buildings (structures). They are divided into fire power supply and distribution facilities, automatic fire alarm systems, fire water supply systems, fire hydrant systems, automatic water sprinkler systems, gas fire extinguishing systems, mechanical pressurized air supply systems, mechanical smoke exhaust systems, emergency lighting and evacuation signs, emergency broadcasts, fire-dedicated telephones, fire separation facilities, fire elevator systems, and fire extinguisher systems. It is an important facility to ensure the fire safety of buildings and the safety of personnel evacuation, and is an important part of modern buildings. It plays an important role in protecting buildings and effectively protects the lives of citizens and the safety of national property.

[0003] At present, there are different types of technical standards for building fire protection facility inspection, such as national standards, industry standards, and local standards, which all make technical requirements for the functions of fire protection facilities. Fire protection facility inspection is to inspect each system of building fire protection facilities using specified methods to determine whether the building fire protection facilities meet the indicators specified in the specifications and requirements, and ensure that the building fire protection facilities can achieve the expected use effect. The purpose of fire protection facility inspection is to ensure that the fire protection facilities in the building can operate normally, thereby ensuring the fire safety of the building. The inspection work is usually undertaken by professional fire protection inspection agencies or personnel with relevant qualifications and experience. They need to operate in accordance with relevant standards and specifications to ensure the accuracy and scientificity of the inspection results.

[0004] Traditional building fire protection facility inspection and fire protection maintenance are disconnected, and are generally undertaken by the inspection unit and the maintenance unit respectively. Fire protection facility inspection generally occurs during building acceptance and annual inspection, while fire protection maintenance occurs during annual maintenance. Although the objects are both fire protection facilities, the data of the two units are not interoperable, resulting in repeated waste of time, manpower and energy.

[0005] Currently, when it is necessary to issue a fire protection facility inspection report, the inspection personnel need to clarify the technical standards to be based on, and then go to the site to conduct inspections and collect data. After returning, they need to manually organize the data and fill in the inspection data, match it with the technical standards to determine whether each sub-item is qualified, and then make a comprehensive judgment to issue a report. The amount of information that the staff needs to review is huge, the labor intensity is high, and there are many human factors, making it difficult to ensure the scientificity and fairness of the data. In addition, the traditional fire protection facility inspection can only issue a report that matches the technical requirements at one time, and it cannot verify whether the fire protection facilities of the building meet other technical standards. The data utilization rate is low.

[0006] The following technical solutions have been disclosed in the prior art: (1) An optical fiber-based building fire protection facility detection network system Abstract: The utility model discloses a building fire protection facility detection network system based on optical fiber, including at least one fire protection facility detection terminal installed on the site where the fire protection facility needs to be detected, a concentrator and a monitoring center. The fire protection facility detection terminal and the concentrator, and the concentrator and the monitoring center are respectively connected by transmission optical fiber to achieve communication. The utility model can complete the remote data information transmission between the fire protection facility detection terminal and the concentrator, and between the concentrator and the monitoring center by using transmission optical fiber. The system structure is simple, the one-time investment is small, and it is easy to install, operate and maintain. It can be widely used in the field of fire protection. The advantages of the utility model can be more clearly shown in some occasions where it is impossible to wire between the monitoring center and the fire protection facility detection terminal. One centralized point can accommodate the centralized and external transmission of information of more than 200 fire protection facility detection terminals, realizing the centralized management and collection of fire protection facility detection terminal information.

[0007] (2) A method for directly exporting measurement data into a shipping report Description: This invention relates to a method for directly exporting measurement data into a shipping report, including a measurement data sorting module and a measurement data output shipping report module. The measurement data output shipping report module is used based on the measurement data sorting module. The measurement data sorting module sorts the original data obtained from three-coordinate measurement into easy-to-understand, neatly arranged, numbered dimension data, and exports it into a WORD format document; the measurement data output shipping report module identifies the WORD data document exported by the measurement data sorting module, and then searches in WORD according to the dimension number in the shipping report. After finding the corresponding number, the data of the corresponding number is copied to the shipping report; this method is convenient and fast, has a low error rate, and improves work efficiency.

[0008] (3) A universal MRI information acquisition device, method and process based on the Internet of Things Description: The present invention discloses a general MRI information acquisition device and method based on the Internet of Things. A general MRI information acquisition device based on the Internet of Things includes a data acquisition module, a data storage module, a data processing module, a data transmission module, and a cloud server; a method of a general MRI information acquisition device based on the Internet of Things includes the following steps: I. Configure the information acquisition device and parameters on the control system of each MRI device; II. The data acquisition module collects data; III. The collected log data is cleaned and translated in the data processing module; IV. After that, the system automatically triggers the upload service, and the data transmission module transmits the processed data to the cloud server. The present invention can realize the real-time collection of MRI data across regions and manufacturers, and transmit the collected data to the cloud server, which is convenient for unified classification analysis and management of data, and greatly improves the reliability and convenience of MRI data analysis.

[0009] In summary, the disadvantages of the prior art include: (1) The existing fire protection facility inspection data collection is that the inspection agency dispatches inspection personnel to the site to record data. This is a manual operation, and a large amount of manual data processing is required later, including data classification, analysis, and integration. The amount of information that the staff needs to review is huge, the labor intensity is high, and there are many human factors, making it difficult to ensure the scientificity and fairness of the test data.

[0010] (2) There are many technical standards for building fire protection facility inspection, including national standards, industry standards, and local standards, which are updated every year. These standards have different test methods and inconsistent judgment indicators. Traditional fire protection facility inspection can only issue a report that matches the technical requirements in one inspection, and cannot verify whether the building's fire protection facilities meet other technical standards. The data utilization rate is low. It is impossible to match multiple technical standards in one inspection to issue a report and determine whether the building's fire protection facilities meet other standards.

[0011] (3) The current fire protection facility inspection report generation system only has the function of report generation. It simply manually inputs the manually processed data, written descriptions, and manually determined results, and automatically issues a report by arranging and organizing the data and descriptions. It does not include the system's automatic collection and processing of inspection data and automatic determination and conclusion. It is a semi-manual operation. Summary of the invention

[0012] In response to the above problems, the purpose of the present invention is to provide a fire protection facility detection system and method based on the Internet of Things and artificial intelligence, which can automatically retrieve data, enter the analysis model to classify, process and convert the data, match the relevant technical requirements of the standard, make intelligent judgments on each sub-item data, and configure automatic generation of analysis reports to reduce labor costs, reduce repetitive work, and improve data accuracy.

[0013] The above-mentioned object of the present invention is achieved through the following technical solutions: A fire protection facility detection system based on the Internet of Things and artificial intelligence, comprising: a data area module, a data analysis module, a technical standard module, a data determination module and a detection report module; The data area module is used to input basic data including order data and project information, and classify fire protection facilities according to location and system according to fire protection drawings and fire protection facility location maps to generate different system data areas, collect fire protection facility operation data and linkage data, and upload image data and video data to the system data area corresponding to the current data; The data analysis module is used to match the data of each sub-item technical requirement entered or collected in the data area module with the determination index of the technical standard through the multimodal large language model of fire protection facility detection according to the technical standard selected for issuing the report, and determine whether it is qualified; The technical standard module is a database of technical indicators of different fire protection facility systems in the technical standards, which is sorted according to different technical standard contents; The data determination module comprehensively determines whether the requirements are qualified based on the determination results of each sub-item technical requirement; The test report module is used to generate and print a report.

[0014] Further, the data area module includes a conventional data submodule and a fire data submodule; The conventional data submodule includes an order data unit and a project data unit; Wherein, the order data unit is used to obtain and store the order data, and the project data unit is used to obtain and store the project information; The fire data submodule includes a data partition unit, a data acquisition unit and a data upload unit; Wherein, the data partitioning unit is used to classify the fire protection facilities according to the location and system according to the fire protection drawings and the fire protection facility point map uploaded by the user to generate different system data areas, so as to facilitate the classified transmission of the data acquired in the data acquisition unit and the data uploading unit to the fire protection facility system information cabinets corresponding to the different system data areas; The data acquisition unit is used to connect with the fire host through the fire information transmission device to realize the sharing of information in the fire host, obtain the operation data and linkage data of the fire facilities, collect the operation data of the fire facilities of the building through the fire Internet of Things device including the pressure sensor, the water level sensor and the flow sensor, and transmit the collected different system data to the corresponding system data area respectively according to the corresponding relationship between the fire host and the fire facility point map; The data uploading unit is used by the fire protection technical service agency to upload picture data and video data to the system data area of ​​the corresponding fire protection facility.

[0015] Furthermore, in the data partitioning unit, the fire protection facilities are classified according to the location and system according to the fire protection drawings and the fire protection facility location map uploaded by the user to generate different system data areas, specifically: Perform image recognition or data analysis on the fire protection drawings and the fire protection facility point map to identify key information of the fire protection facilities including icons, names, and annotations, and create a system template corresponding to the current system according to different system types, wherein the system template includes the types of fire protection facilities that each system should have; According to the type of fire-fighting facilities included in each of the system templates, a search is performed in the fire-fighting drawings and the fire-fighting facility point map to obtain the locations and related parameters of the fire-fighting facilities in the searched system templates, and according to the distribution on the fire-fighting drawings and the fire-fighting facility point map, the system data area of ​​the system corresponding to the current system template is determined, and the relevant data of these fire-fighting facilities are sorted into the system data area.

[0016] Furthermore, in the data acquisition unit, the fire information transmission device is connected to the fire host to realize the sharing of information in the fire host, obtain the operation data and linkage data of the fire facilities, and collect the operation data of the fire facilities of the building through the fire Internet of Things device including the pressure sensor, the water level sensor and the flow sensor. According to the corresponding relationship between the fire host and the fire facility point map, the collected different system data are respectively transmitted to the corresponding system data area, specifically: The data acquisition unit acquires the information in the fire host through the fire information transmission device in a polling or interrupt-driven manner, wherein in the polling mode, the data acquisition unit sends a query request to the fire host at a preset time interval to acquire the operation data and the linkage data during the operation of the fire protection facilities, and in the interrupt-driven mode, the data acquisition unit is actively notified when the fire host generates new data or the data changes; After the data acquisition unit acquires the information in the fire host, the acquired data is transmitted to the corresponding system data area according to a pre-planned path, wherein the pre-planned path is planned using a method including a shortest path algorithm; After the collected data is transferred to the corresponding system data area, a transaction processing mechanism is used to ensure the integrity and consistency of the data writing, wherein the transaction processing mechanism is used to handle failures that occur during the writing process.

[0017] Furthermore, in the data uploading unit, the fire protection technical service agency uploads the picture data and the video data to the system data area of ​​the corresponding fire protection facility, and further includes: In the process of uploading the image data, encryption technologies including SSL / TLS encryption protocols are used to encrypt the transmitted image data. At the same time, for large image data, a block uploading technology is used to divide the large image data into multiple small blocks for uploading; In the process of uploading the video data, the streaming media uploading technology is used to buffer and upload the video data. At the same time, when the network status is not good, the resolution and frame rate of the video data are adjusted to reduce the data volume of the video data.

[0018] Furthermore, before the data analysis module, it is necessary to train the multimodal large language model for fire protection facility detection, which is used to determine whether each sub-item technical requirement meets the technical standard, specifically: Collect a large amount of sensor data during the detection process of the fire-fighting facilities in advance, and perform preprocessing on the sensor data including noise removal and data standardization, collect a large amount of uploaded image data and video data, and annotate the image data with information including key parts of the fire-fighting facilities, component status, and key nameplate information, and annotate the video data with information including the action start time, key nodes, and whether the action is normal; Select an open-source multimodal large language training model, and fine-tune the model according to the particularity of the fire protection facility field, including modifying the input layer of the model to adapt to the sensor data, the image data, and the video data, defining mission objectives suitable for fire protection facilities, and adjusting parameters in the model; The sensor data is numerically encoded so that it can be accepted by the model, the image data is subjected to a feature vector extracted using a convolutional neural network (CNN) technique and converted into a feature representation, the video data is first decomposed into frames, and the feature of each frame is extracted using a convolutional neural network (CNN) technique, and the feature representation of the video is constructed in combination with the event sequence information of the video, and the encoded sensor data, image feature vector and video feature vector are fused to form multimodal input data; The model is trained for multiple rounds using the fused multimodal input data. In each round of training, the multimodal input data is fed into the model, and the model makes predictions based on predefined mission objectives suitable for firefighting facilities. The error between the prediction results and the labeled results is calculated, and the model parameters are adjusted based on the error. The model is continuously optimized until the model performance meets the preset requirements.

[0019] Furthermore, in the data analysis module, according to the technical standard selected as the basis for issuing the report, the data of each sub-item technical requirement entered or collected in the data area module are matched with the judgment index of the technical standard through the multimodal large language model for fire protection facility detection to determine whether it is qualified, specifically: Conducting an in-depth analysis of the selected technical standards, converting the clauses in the technical standards into quantifiable judgment indicators, and performing structured storage of the quantified judgment indicators to facilitate model query; After the data stored in each of the system data areas are passed through the fire-fighting facility detection multimodal large language model, the model comprehensively processes the input sensor data, the image data and the video data, extracts text information related to the status of the fire-fighting facilities, and formats the text information output by the model so that it matches the stored quantifiable judgment indicators in data format and expression; Each of the text information output by the model is matched one-to-one with the quantifiable judgment index. At the same time, for fire protection facilities that require comprehensive judgment based on multiple indicators, an overall judgment is performed in accordance with the comprehensive judgment requirements in the technical standards.

[0020] Furthermore, in the detection report module, the process of generating a report is specifically as follows: Referring to the technical standard document, a basic structure template of the report is determined, including a general data part and a fire protection facility data part, and a data placeholder is set in the basic structure template for subsequent import of data of the general data part and the fire protection facility data part, wherein the general data part includes information including a report title, a report number, information of a testing unit, information of a tested building, and a testing date, and the fire protection facility data part is divided according to different system data areas, and each system contains information including testing indicators, testing results, and judgment conclusions; Acquire the data processed by the data analysis module and regular data from each of the system data areas, clean the data, remove invalid data, duplicate data and erroneous data, and convert the acquired data according to the data type and format requirements of the data placeholders in the basic structure template; Develop a data import program to identify the data placeholders in the basic structure template and fill the sorted data into the corresponding positions; The final report is generated according to the basic structure template filled with data.

[0021] A firefighting facility detection method based on the Internet of Things and artificial intelligence using the firefighting facility detection system based on the Internet of Things and artificial intelligence as described above comprises the following steps: S1: Enter basic data including order data and project information, and classify fire protection facilities by location and system according to fire protection drawings and fire protection facility location maps to generate different system data areas, collect fire protection facility operation data and linkage data, and upload image data and video data to the system data area corresponding to the current data; S2: According to the selected technical standard for issuing the report, the data of each sub-item technical requirement entered or collected in the data area module is matched with the determination index of the technical standard through the multimodal large language model for fire protection facility detection to determine whether it is qualified; S3: Comprehensively determine whether the product is qualified based on the results of the determination of each sub-item technical requirement; S4: Generate a report and print it.

[0022] A computer-readable storage medium stores computer codes. When the computer codes are executed, the above method is executed.

[0023] Compared with the prior art, the present invention has at least one of the following beneficial effects: (1) By providing a fire protection facility detection system based on the Internet of Things and artificial intelligence, the system uses the operation data, maintenance pictures, videos and other information of each fire protection facility collected by the system to process the information using a multimodal large language model, and compares it with the judgment indicators in the built-in technical standard database. The system automatically determines whether each sub-item is qualified, and then makes a general judgment to draw a detection conclusion. The data is identified and sorted, and a report is generated according to the report format template.

[0024] (2) Convenient and fast: There is no need to look up the technical requirements and standards of each fire protection facility. The system can automatically match the corresponding data and automatically judge the results. It can solve the problem of manual sorting of experimental data and automatically issue a report that meets the standard requirements.

[0025] (3) High accuracy: It greatly avoids human errors and ensures the scientificity and fairness of the test data.

[0026] (4) Save time: Each annual inspection requires inspectors to arrive at the site, manually collect data, and then go back to organize and edit the report, which takes at least 2 days. Using this system, the report can be issued within 10 minutes. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is the overall structure diagram of the fire protection facility detection system based on the Internet of Things and artificial intelligence of the present invention; Figure 2 This is an overall flow chart of the fire protection facility detection system based on the Internet of Things and artificial intelligence of the present invention; Figure 3 The figure is an overall flow chart of the fire protection facility detection method based on the Internet of Things and artificial intelligence of the present invention. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0029] Those skilled in the art will appreciate that, unless otherwise stated, the singular forms "a", "an", "said" and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0030] Technical explanation of the present invention: Internet of Things data collection technology: Internet of Things, as the name implies, all data collection is collected from devices. There are many types of devices, most of which are collected through sensors. Internet of Things data collection technology uses processors, control devices, and corresponding sensor devices, such as temperature and humidity sensors, current, voltage sensors, RFID identification, switch quantity / analog quantity collection devices, etc., to transmit the collected operation data to the server and cloud platform, obtain valuable device data, and realize the control and analysis of data.

[0031] Traditional building fire protection facility inspections rely on regular manual processing. Each time, the inspection unit dispatches inspection personnel to the site to collect different types of data for different fire protection devices. After returning to the unit, they process the data and prepare reports.

[0032] The operating data of fire protection facilities can often be obtained through the fire protection facility maintenance process. Traditionally, the inspection and maintenance of building fire protection facilities are carried out separately. Each fire protection facility inspection requires on-site collection of data, resulting in a waste of time, manpower and energy.

[0033] In this process, data is recorded manually instead of through the Internet, which can lead to data loss, omission or errors. Manual analysis of test results also relies on the operator's experience and judgment and lacks accuracy. Reports rely on manual editing and are prone to omissions and false alarms.

[0034] This process relies heavily on regular manual inspections and is easily affected by human factors. It requires a lot of human resources and has high labor costs. The data utilization rate is low, and there is a lack of systematic data analysis and utilization. It is only used for recording and filing and cannot provide valuable decision support. This traditional method of inspecting building fire protection facilities is generally a random inspection that is difficult to cover all time periods and all areas of the building. There may be blind spots in the inspection and cannot fully guarantee the building fire safety.

[0035] The technical problems to be solved by the present invention are as follows: (1) Traditional data collection cannot cover all fire protection facilities in a building, and the data is not representative; (2) Manual data analysis depends on experience and personnel capabilities; (3) Manual recording and reporting is slow and prone to errors.

[0036] The present invention provides a fire protection facility detection system based on the Internet of Things and artificial intelligence. First, the system classifies fire protection facilities through the basic building information (project category, usage nature, etc.) and the fire protection facility location map provided by the user.

[0037] The operating data of the building's fire protection facilities are collected through fire protection information transmission devices, pressure sensors, water level sensors, flow sensors and other fire protection Internet of Things devices. These data are stored in the corresponding fire protection facility information cabinets. The maintenance units upload pictures and video information of the maintenance process of different fire protection facilities, and these data are stored in the corresponding fire protection facility information cabinets.

[0038] By using artificial intelligence and an open-source multimodal language training model, the model is fine-tuned according to the particularities of the fire protection facility detection field, and a large amount of sensor data, images, and video training in the actual fire protection facility detection process are used to derive the "fire protection facility detection multimodal language model". The model can comprehensively process sensor information, image information, and video information, extract text information, and match it with the technical standard judgment indicators of the background to determine whether it is qualified, issue a fire protection facility inspection report, and can be downloaded and saved. This information-based method ensures the authenticity of the data and the efficiency of the detection.

[0039] The invention of the present invention is described in detail below through specific embodiments: First embodiment like Figure 1 and Figure 2 As shown, a fire protection facility detection system based on the Internet of Things and artificial intelligence is characterized by comprising: a data area module 1, a data analysis module 2, a technical standard module 3, a data determination module 4 and a detection report module 5; The data area module 1 is used to input basic data including order data and project information, and classify fire protection facilities according to location and system according to fire protection drawings and fire protection facility location maps to generate different system data areas, and collect fire protection facility operation data and linkage data, and upload picture data and video data to the system data area corresponding to the current data; The data analysis module 2 is used to match the data of each sub-item technical requirement entered or collected in the data area module with the determination index of the technical standard through the multimodal large language model of fire protection facility detection according to the technical standard selected for issuing the report, and determine whether it is qualified; The technical standard module 3 is a database of technical indicators of different fire protection facility systems in the technical standards sorted out according to different technical standard contents; The data determination module 4 is used to comprehensively determine whether the product is qualified based on the determination results of each sub-item technical requirement; The detection report module 5 is used to generate and print a report.

[0040] The data area module 1 is connected to the data analysis module 2, the test report module 5, and the technical standard module 3 respectively; the data area module 1 and the technical standard module 3 are connected to the data determination module 4 respectively; the data determination module 4 and the test report module 5 are connected in sequence; the data area module 1, the technical standard module 3 and the data determination module 4 are connected to the test report module 5 respectively.

[0041] The following is a detailed description of each of the above modules: (1) Data area module 1 The data area module 1 includes a conventional data submodule 11 and a fire data submodule 12; The conventional data submodule 11 includes an order data unit 111 and a project data unit 112; The order data unit 111 is used to obtain and store the order data, including but not limited to the contract number, commission number, report number, commissioning unit and inspection date. The project data unit is used to obtain and store the project information, including but not limited to the project name, project address, construction permit number, construction unit, use nature, inspection area, and project category.

[0042] The fire data submodule 12 includes a data partition unit 121, a data acquisition unit 122 and a data upload unit 123; Among them, the data partitioning unit 121 is used to classify the fire protection facilities according to the location and system according to the fire protection drawings and the fire protection facility location map uploaded by the user to generate different system data areas, which is convenient for classifying and transmitting the data acquired in the data acquisition unit and the data upload unit to the fire protection facility system information cabinets corresponding to different system data areas, as well as the classified uploading of data, which is conducive to the subsequent information partition processing.

[0043] According to the system, it is divided into the data area of ​​fire power supply and distribution facilities, the data area of ​​automatic fire alarm system, the data area of ​​fire water supply system, the data area of ​​fire hydrant system, the data area of ​​automatic sprinkler fire extinguishing system, the data area of ​​gas fire extinguishing system, the data area of ​​mechanical pressurized air supply system, the data area of ​​mechanical smoke exhaust system, the data area of ​​emergency lighting and evacuation signs, the data area of ​​emergency broadcasting, the data area of ​​fire-fighting telephone, the data area of ​​fire separation facilities, the data area of ​​fire elevator system, and the data area of ​​fire extinguisher system. It is normal for some buildings to lack several of these systems. According to the point map, there are several system data areas according to the systems.

[0044] The data acquisition unit 122 is used to connect with the fire host through the fire information transmission device, realize the sharing of information in the fire host, obtain the operation data and linkage data of the fire facilities, and collect the operation data of the fire facilities of the building through the fire Internet of Things device including pressure sensors, water level sensors and flow sensors, such as the pressure value of the main network of the fire water supply system and the terminal pressure value. According to the corresponding relationship between the fire host and the fire facility point map, the collected different system data are respectively transmitted to the corresponding system data area, for example, the operation data of the automatic fire alarm system is stored in the automatic fire alarm system data area.

[0045] The data uploading unit 123 is used by the fire protection technical service agency to upload picture data and video data to the system data area of ​​the corresponding fire protection facility. Picture data refers to the appearance pictures of fire protection facilities taken by the fire protection technical service agency during the maintenance process, including nameplate information, related accessories, etc. Video data refers to the action videos of fire protection facilities taken by the fire protection technical service agency during the maintenance process, including: action videos of fireproof roller shutters, opening and closing videos of fireproof doors, opening and closing videos of fireproof windows, and other fire protection facility opening and closing videos; panel and display function inspection videos, linkage videos between fire protection facilities, etc.

[0046] Furthermore, in the data partitioning unit 121, the fire protection facilities are classified according to the location and system according to the fire protection drawings and the fire protection facility location map uploaded by the user to generate different system data areas, specifically: A: Perform image recognition or data analysis on the fire protection drawings and the fire protection facility location map to identify key information of the fire protection facilities including icons, names, and annotations, and create a system template corresponding to the current system based on different system types. The system template includes the types of fire protection facilities that each system should have.

[0047] If the fire protection drawings and point maps are in the form of pictures, use image recognition technology to convert the graphics and text information in the drawings into computer-recognizable data. For electronic formats (such as PDF, CAD files, etc.), data parsing is performed. According to different types of fire protection systems, corresponding templates are created, including the types of fire protection facilities that each system should have. For example, the template of the automatic fire alarm system includes facilities such as smoke detectors, temperature detectors, and alarm buttons.

[0048] B: According to the type of fire-fighting facilities included in each of the system templates, a search is performed in the fire-fighting drawings and the fire-fighting facility point map to obtain the locations and related parameters of the fire-fighting facilities in the searched system templates, and according to the distribution on the fire-fighting drawings and the fire-fighting facility point map, the system data area of ​​the system corresponding to the current system template is determined, and the relevant data of these fire-fighting facilities are sorted into the system data area.

[0049] For example, the following fire protection system types are classified by system: Fire power supply and distribution facilities Search for the locations and related parameters of equipment marked as distribution boxes, distribution cabinets (especially those clearly marked as fire power supply and distribution related), emergency power supply devices, etc. in the drawings and point maps. According to their distribution on the drawings, determine the data area belonging to the fire power supply and distribution facilities, and organize the relevant data of these equipment (such as power, connection lines, etc.) into this data area.

[0050] Automatic fire alarm system Find the identification of smoke detectors, temperature detectors, flame detectors, manual alarm buttons, fire alarm controllers and other equipment. According to their location information, the data of these facilities are classified into the fire automatic alarm system data area, including the model of the equipment, detection range and other information.

[0051] Fire water supply system Determine the location and marking of water supply equipment such as fire water pools, fire water tanks, pressure-stabilizing pumps, and variable-frequency pumps. Assign data related to these equipment (such as volume, flow, head, etc.) to the fire water supply system data area.

[0052] Fire hydrant system Identify the locations of fire hydrant boxes (including indoor and outdoor fire hydrants), fire hoses, water guns, fire hydrant pumps and other facilities on the drawings. Classify the relevant data of each facility in the fire hydrant system (such as hydrant port pressure, hose length, etc.) into the fire hydrant system data area.

[0053] Automatic sprinkler system Find out the identification of sprinklers (such as upright sprinklers, pendant sprinklers, etc.), alarm valve groups (wet, dry, pre-action alarm valves, etc.), water flow indicators, etc. According to their layout, divide the data of each component in the automatic sprinkler system (such as the flow coefficient of the sprinkler, the rated working pressure of the alarm valve, etc.) into the corresponding data area.

[0054] Gas fire extinguishing system Determine the location and parameter marking of gas fire extinguishing cylinders, selector valves, nozzles, gas fire extinguishing controllers and other equipment. Organize the data of gas fire extinguishing system equipment (such as extinguishing agent type, cylinder capacity, etc.) into the gas fire extinguishing system data area.

[0055] Mechanical pressurized air supply system Find the identification and location information of the blower, air duct, air outlet, etc. Assign the data related to the mechanical pressurized air supply system (such as air supply volume, fan power, etc.) to the data area of ​​the system.

[0056] Mechanical smoke exhaust system Identify the locations of smoke exhaust fans, smoke exhaust ducts, smoke exhaust outlets and other facilities on the drawings. Assign the data of mechanical smoke exhaust system equipment (such as smoke exhaust volume, fan speed, etc.) to the mechanical smoke exhaust system data area.

[0057] Emergency lighting and evacuation signs Find the location and related labels (such as brightness, emergency time, etc.) of emergency lighting fixtures and evacuation signs. Classify the data of these facilities into the emergency lighting and evacuation sign data area.

[0058] Emergency Broadcast System Determine the location and related parameters (such as power, coverage, etc.) of the broadcast loudspeakers, broadcast host and other equipment. Assign the data of the emergency broadcast system equipment to the emergency broadcast data area.

[0059] Fire-fighting telephone system Identify the location and phone number of fire-fighting telephone extensions, main units and other equipment, and organize their data into the fire-fighting telephone data area.

[0060] Fire separation facilities Find the location and specification of fire walls, fire shutters, fire doors and other facilities on the drawings. Assign the data of fire separation facilities (such as the fire resistance limit of fire shutters, the grade of fire doors, etc.) to the fire separation facility data area.

[0061] Fire Elevator System Determine the location of the fire elevator, the size of the elevator shaft, the load capacity, etc. Divide the data of the fire elevator system (such as lifting speed, stop floors, etc.) into the fire elevator system data area.

[0062] Fire extinguisher system Identify the location, type (such as dry powder fire extinguisher, carbon dioxide fire extinguisher, etc.), capacity and other labels of the fire extinguisher. Classify the data of the fire extinguisher system (such as the expiration date and maintenance date of the fire extinguisher) into the fire extinguisher system data area.

[0063] In addition, the system partitions can be adjusted based on the location relationship of the fire protection facilities. According to the different areas of the building (such as floors, fire protection zones, etc.), check whether the distribution of each fire protection facility is reasonable. For facilities located in the same area and belonging to the same system, it may be necessary to further refine the organization of the data to better reflect the layout of the fire protection facilities in the area. For example, the fire hydrant system facility data of a certain floor is listed separately to facilitate the management of the fire protection facilities on that floor. Analyze the adjacent relationship between different fire protection facilities, such as whether the placement distance between fire hydrants and fire extinguishers meets the regulations. According to this positional relationship, the classified data is necessary for association annotation, so that the collaborative working ability of fire protection facilities can be better evaluated in the later information processing.

[0064] Furthermore, in the data acquisition unit 122, the fire information transmission device is connected to the fire host to realize the sharing of information in the fire host, obtain the operation data and linkage data of the fire facilities, and collect the operation data of the fire facilities of the building through the fire Internet of Things device including the pressure sensor, the water level sensor and the flow sensor. According to the corresponding relationship between the fire host and the fire facility point map, the collected different system data are respectively transmitted to the corresponding system data area, specifically: A: The data acquisition unit obtains the information in the fire host through the fire information transmission device in a polling or interrupt-driven manner. In the polling mode, the data acquisition unit sends a query request to the fire host at a preset time interval to obtain the operation data and the linkage data during the operation of the fire protection facilities. In the interrupt-driven mode, the data acquisition unit is actively notified when the fire host generates new data or the data changes.

[0065] The data acquisition module is connected to the fire host through the fire information transmission device. This connection involves specific interface standards. For example, common industrial communication interface standards such as RS-485 interface or Ethernet interface may be used. RS-485 interface has the characteristics of strong anti-interference ability and long transmission distance (up to 1200 meters or more), which is suitable for stable communication with the fire host in complex building environments. Ethernet interface can provide higher transmission rate, meet the information sharing needs of large-scale fire host, and facilitate integration with network facilities in the building. In terms of communication protocol, it may adopt custom protocols or follow some industry standard protocols, such as Modbus protocol. Custom protocols can be optimized according to the special needs of the fire system to ensure that the information in the fire host can be accurately shared with the data acquisition module.

[0066] To achieve information sharing within the host, an effective data interaction mechanism needs to be established. The data acquisition module may use polling or interrupt-driven methods to obtain information within the fire host. In the polling method, the data acquisition module sends query requests to the fire host at a certain time interval to obtain fire facility operation data and linkage data. This method is simple and reliable, but there may be a certain delay when the data is updated frequently. The interrupt-driven method is that the fire host actively notifies the data acquisition module when new data is generated or the data changes, which can improve the real-time performance of data acquisition, but it has high requirements for the hardware and software design of the fire host and the data acquisition module.

[0067] In addition, for different types of sensors provided to the fire host to obtain data, special data acquisition circuits can be designed. For sensors with weak signal output, such as some high-precision pressure sensors, the data acquisition circuit needs to include an amplifier circuit to amplify the weak signal to a recognizable range. At the same time, a filter circuit needs to be designed to remove noise interference in the sensor signal. For example, a low-pass filter is used to remove high-frequency noise, and a band-pass filter is used to filter sensor signals within a specific frequency range. In addition, the analog-to-digital conversion (ADC) circuit is also an important part of the data acquisition circuit, which converts the analog sensor signal into a digital signal for subsequent processing by the data acquisition module.

[0068] B: After the data acquisition unit acquires the information in the fire host, the acquired data is transmitted to the corresponding system data area according to a pre-planned path, wherein the pre-planned path is planned using a method including a shortest path algorithm.

[0069] For example, there is a fire facility monitoring system, which contains multiple data acquisition nodes and multiple system data area nodes, which are connected through a network. Each connection has a weight, which represents the delay or cost of data transmission.

[0070] I. Data Structure Definition Define a node set V, which includes all data acquisition nodes and system data area nodes.

[0071] Define an edge set E, which represents the connections between nodes. Each edge includes a source node, a target node, and a weight.

[0072] Define a distance array dist[] to store the current shortest distance from the starting node to each node.

[0073] Define a predecessor node array prev[] to store the predecessor nodes of each node in the shortest path.

[0074] II. Algorithm Steps Initialization Set the distance of the starting node (data acquisition module) to 0, i.e., dist[starting node] = 0. Set the distances of all other nodes to infinity, i.e., dist[node]= ∞ (except the starting node). Mark all nodes as having undetermined shortest distances.

[0075] Main Loop Select the node u with the minimum distance from the nodes with undetermined shortest distances.

[0076] Mark node u as having a determined shortest distance.

[0077] For each node v adjacent to node u: Calculate the distance alt to reach node v through node u as alt = dist[u] + weight(u, v), where weight(u, v) is the weight of the edge between node u and node v. If alt < dist[v], then update dist[v]= alt and prev[v] = u.

[0078] Determine the Shortest Path Once all nodes are marked as having determined shortest distances, the shortest path can be determined by backtracking from the target node (system data area node) to the starting node using the predecessor node array prev[].

[0079] III. Improvements and Redundant Design Dynamic Weight Adjustment Considering the real-time nature of fire data, the weights of the edges can be dynamically adjusted according to factors such as network congestion and link quality. For example, if an increase in the transmission delay of a certain link is detected, the weight of that link can be increased so that the algorithm is more likely to select other paths in subsequent calculations.

[0080] Redundant Path Calculation While calculating the shortest path, multiple suboptimal paths can be calculated as redundant paths. For example, in the main loop, for each node v, in addition to recording the distance and predecessor node to v through the current shortest path, the distance and predecessor node to v through the second shortest path can also be recorded. In this way, when the main path fails, it can quickly switch to the redundant path.

[0081] Fault detection and switching Establish a real-time network monitoring mechanism to detect transmission link failures. When a link failure is detected, immediately switch to a pre-calculated redundant path. At the same time, the algorithm can be triggered to recalculate the shortest path to adapt to the new network status.

[0082] As a more concrete example, consider the following network of nodes and edges: Nodes: A (data acquisition module), B, C, D (system data area nodes) Sides: (A,B,5), (A,C,3), (B,D,4), (C,B,2), (C,D,6) Use the improved Dijkstra algorithm to calculate the shortest path from node A to node D: initialization: dist[A] = 0, dist[B]= ∞, dist[C] = ∞, dist[D]= ∞.

[0083] Select node A and update the adjacent nodes: dist[B] = 5 (reach B via A), prev[B] = A.

[0084] dist[C] = 3 (reach C via A), prev[C] = A.

[0085] Select node C (the undetermined node with the smallest current distance) and update the adjacent nodes: dist[B] = 3 + 2 = 5 (reaching B via C, the distance remains unchanged).

[0086] dist[D] = 3 + 6 = 9 (reaching D via C), prev[D] = C.

[0087] Select node B and update the adjacent nodes: dist[D] = 5 + 4 = 9 (reaching D via B, the distance remains unchanged).

[0088] The shortest path finally determined is A ->C ->D, with a total distance of 9. If a failure of link A ->C is detected during transmission, the suboptimal path A ->B ->D can be immediately switched.

[0089] C: After the collected data is transferred to the corresponding system data area, a transaction processing mechanism is used to ensure the integrity and consistency of the data writing, wherein the transaction processing mechanism is used to handle failures that occur during the writing process.

[0090] Each system data area has a specific storage structure to accommodate different types of fire data. For example, the data area of ​​the automatic fire alarm system may use a tree structure to store data, where the root node represents the entire system, the child nodes represent the alarm devices in different areas or floors, and the leaf nodes store the operating data and status information of specific devices. The data area of ​​the fire water supply system may use a relational database structure to store data, associating data such as pressure values, water level values, and flow values ​​with corresponding water supply facilities (such as fire water tanks, water pumps, etc.) for easy query and management.

[0091] When the collected data is transferred to the corresponding system data area, a reliable data writing mechanism is required. The transaction processing mechanism is used to ensure the integrity and consistency of data writing. For example, when writing new firefighting facility operation data to the system data area, if a failure occurs during the writing process (such as sudden power outage, storage device failure, etc.), the transaction processing mechanism can ensure that the data is either written completely or not written, avoiding data damage or loss.

[0092] Furthermore, in the data uploading unit 123, the fire protection technical service agency uploads the picture data and the video data to the system data area of ​​the corresponding fire protection facility, and further includes: A: In the process of uploading the image data, encryption technologies including SSL / TLS encryption protocol are used to encrypt the transmitted image data. At the same time, for large image data, block uploading technology is adopted to divide the large image data into multiple small blocks for uploading.

[0093] When fire protection technical service agencies take pictures of the exterior of fire protection facilities, the shooting equipment (such as digital cameras or smart phones) used must meet certain parameter requirements. In order to clearly obtain details such as nameplate information and related accessories, the shooting equipment should have a high resolution, such as at least 8 million pixels. In areas with darker light, the shooting equipment should support manual adjustment of parameters such as exposure and focus to ensure that clear and accurate pictures reflecting the status of fire protection facilities are taken.

[0094] The images you take may have different formats (such as JPEG, PNG, etc.). In order to facilitate uploading and storage, the image format needs to be standardized. Generally speaking, the JPEG format is a commonly used lossy compression format suitable for storing color photos. It can effectively reduce the file size while maintaining high image quality, which is conducive to increasing the upload speed. If the image contains important text information (such as detailed parameters on a nameplate), it may need to be converted to a lossless compression format, such as PNG, to ensure that the text information is clear and legible.

[0095] The data upload unit needs to provide a specific upload interface, which can be based on common network protocols such as HTTP or FTP. The HTTP protocol is a hypertext transfer protocol widely used on the Internet. It is simple, easy to use, and highly versatile, and is suitable for uploading images in various network environments. During the upload process, in order to ensure the security and integrity of the data, encryption technology such as SSL / TLS encryption protocol may be used to encrypt the transmitted image data. At the same time, in order to improve the upload efficiency, the block upload technology may be used to split large images into multiple small blocks for uploading, reducing the risk of upload failures caused by network interruptions and other reasons.

[0096] In order to accurately upload pictures to the corresponding fire protection facility system data area, an effective association mechanism needs to be established. The metadata of the picture can contain identification information related to the fire protection facility, such as the facility name, location, system type, etc. When the picture is uploaded, the data upload unit will store the picture in the correct fire protection facility system data area based on the identification information in these metadata and the pre-established system data area mapping relationship. For example, if a picture is taken of a fire hydrant nameplate on a certain floor, and the picture metadata contains identifications such as "fire hydrant" and "certain floor", then it will be uploaded to the fire hydrant system data area.

[0097] B: In the process of uploading the video data, a streaming media uploading technology is used to buffer and upload the video data. When the network status is poor, the resolution and frame rate of the video data are adjusted to reduce the data volume of the video data.

[0098] For the shooting of firefighting facility action videos, certain shooting specifications need to be followed. For example, when shooting the action video of a fireproof roller shutter, it should be ensured that the entire process from opening to closing (or vice versa) can be fully recorded, and the shooting angle should be able to clearly show the running track of the roller shutter, its relationship with the surrounding building structure, etc. The shooting equipment should have a stable shooting function to prevent the image from shaking and affecting the video quality. At the same time, in order to clearly record the action details of the firefighting facilities, the shooting equipment should support high-definition video shooting, such as 1080p or higher resolution video shooting.

[0099] The videos you shoot may have different formats, such as MP4, AVI, etc. For unified management and uploading, you need to convert the video format. MP4 is a video format widely used on the Internet. It has a good compression ratio and compatibility, and is suitable as a standard format for uploading. During the conversion process, you need to pay attention to the video encoding parameters, such as the adjustment of frame rate, bit rate, etc., to reduce the file size and increase the upload speed while ensuring the video quality.

[0100] The process of uploading video data is similar to uploading pictures, but because video files are usually large, more optimization measures are needed. Streaming media upload technology can be used to upload video data while buffering, rather than waiting for the entire video file to be generated before uploading. At the same time, in order to improve the success rate and speed of uploading, when the network condition is poor, the resolution or frame rate of the video can be automatically adjusted to reduce the amount of video data. For example, when the network bandwidth is detected to be low, the resolution of a 1080p video can be reduced to 720p for uploading.

[0101] The metadata of the video should contain detailed information about firefighting facilities, such as facility name, action type (such as fire curtain opening action), and location. The data upload module accurately uploads the video to the corresponding firefighting facility system data area based on these metadata and pre-established association rules. For example, if a fire door opening and closing video contains labels such as "fire door" and "certain area" in its metadata, it will be uploaded to the fire separation facility data area to facilitate the query and analysis of the maintenance of the facility in the future.

[0102] (2) Data Analysis Module 2 The "Multimodal Large Language Model for Fire Facility Detection" is an open source multimodal large language training model. The model is fine-tuned according to the particularity of the fire facility detection field, and is trained using a large amount of sensor data, images, and videos from actual fire facility detection processes to derive a model that can meet the needs of fire facility detection data analysis. The model can comprehensively process sensor information, image information, and video information, extract text information, and match it with the technical standard judgment indicators of the background to determine whether it is qualified.

[0103] Before the data analysis module 2, it is necessary to train the multimodal large language model for fire protection facility detection used to determine whether each sub-item technical requirement meets the technical standard, specifically: A: Collect a large amount of sensor data from the fire-fighting facility detection process in advance, and perform preprocessing on the sensor data including noise removal and data standardization; collect a large amount of uploaded image data and video data, and annotate the image data with information including key parts of the fire-fighting facilities, component status, and nameplate key information; annotate the video data with information including the start time of the action, key nodes, and whether the action is normal.

[0104] Sensor data is collected from many actual fire protection facilities inspection processes, such as data generated by pressure sensors, water level sensors, flow sensors, etc. These data reflect various parameters of fire protection facilities operation. Sensor data is preprocessed, including removing noise data (for example, abnormal pressure spikes may be caused by sensor failure) and data standardization, unifying data in different units into standard units for model processing.

[0105] A large amount of image and video data is collected during the inspection of fire-fighting facilities. Images may include the appearance, internal structure, nameplate and other information of fire-fighting facilities; videos cover the action process of fire-fighting facilities, such as the lifting and lowering of fire-proof roller shutters and the opening and closing of fire doors. Professionals annotate images and videos. For images, key components of fire-fighting facilities, component status (such as whether they are damaged), key information on nameplates, etc. are annotated; for videos, the start time of the action, key nodes (such as the time it takes for fire-proof roller shutters to descend to a specific height), whether the action is normal, and other information are annotated. The annotated image and video data provide a basis for supervised learning for model training.

[0106] B: Select an open source multimodal large language training model and fine-tune the model according to the particularities of the firefighting facility field, including modifying the input layer of the model to adapt to the sensor data, the image data, and the video data, defining mission objectives suitable for firefighting facilities, and adjusting parameters in the model.

[0107] Evaluate a variety of open source multimodal large language models, considering factors such as the model architecture, the applicability of pre-trained data, and the ability to process multimodal data. For example, some models may have unique architectural advantages in processing the fusion of images and text, while others may perform better in processing long sequence data (such as time series in video data). According to the needs of the fire facility detection field, select a most suitable open source multimodal large language model as the basic model.

[0108] Adjust the model according to the specificity of the firefighting facility detection field. For example, modify the input layer of the model to adapt to the input format of sensor data, image data, and video data. Define task objectives suitable for firefighting facility detection, such as determining the status of firefighting facilities (qualified / unqualified). Then use the collected sensor, image, and video data to fine-tune the model and adjust the model's parameters (such as weights in the neural network) so that the model can better adapt to the firefighting facility detection task. During the fine-tuning process, optimization algorithms such as mini-batch gradient descent may be used to minimize the model's prediction error.

[0109] C: Numerically encode the sensor data so that it can be accepted by the model, extract feature vectors from the image data using convolutional neural network (CNN) technology, and convert them into feature representations. For the video data, first decompose the video into frames, extract features from each frame using convolutional neural network (CNN) technology, and construct feature representations of the video in combination with the event sequence information of the video. The encoded sensor data, image feature vectors, and video feature vectors are fused to form multimodal input data.

[0110] D: Use the fused multimodal input data to train the model for multiple rounds. In each round of training, the multimodal input data is fed into the model, and the model makes predictions based on predefined mission objectives suitable for firefighting facilities. The error between the prediction results and the labeled results is calculated, and the model parameters are adjusted based on the error. The model is continuously optimized until the model performance meets the preset requirements.

[0111] Furthermore, in the data analysis module, according to the technical standard selected as the basis for issuing the report, the data of each sub-item technical requirement entered or collected in the data area module are matched with the judgment index of the technical standard through the multimodal large language model for fire protection facility detection to determine whether it is qualified, specifically: A: Conduct an in-depth analysis of the selected technical standards, convert the various clauses in the technical standards into quantifiable judgment indicators, and store the quantified judgment indicators in a structured manner to facilitate model query.

[0112] Conduct an in-depth analysis of selected technical standards (such as XF503 "Technical Regulations for Inspection of Building Fire Protection Facilities"). Convert the various clauses in the standard into quantifiable judgment indicators. For example, for the inspection standards of fire water supply systems, convert text descriptions such as main network pressure and pool water level into specific numerical ranges. These digital judgment indicators are structured and stored, for example, in a relational database. Each judgment indicator corresponds to a record in the database, which contains fields such as indicator name, indicator numerical range, and applicable fire protection facility type, so as to facilitate model query.

[0113] B: After the data stored in each of the system data areas passes through the fire-fighting facility detection multimodal large language model, the model comprehensively processes the input sensor data, the image data and the video data, extracts text information related to the status of the fire-fighting facilities, and formats the text information output by the model so that it matches the stored quantifiable judgment indicators in data format and expression.

[0114] When the data in the data area of ​​each fire protection facility system passes through the "Fire Protection Facility Detection Multimodal Large Language Model", the model will comprehensively process the input sensor, image and video data to extract text information related to the status of the fire protection facility. For example, the model may extract information about whether the appearance of the fire hydrant is damaged from the image, and extract the pressure value of the fire protection water supply system from the sensor data. The text information output by the model is formatted so that it matches the stored technical standard judgment indicators in data format and expression. For example, the pressure value output by the model is converted into the same unit and data type as in the standard.

[0115] C: Each piece of text information output by the model is matched one-to-one with the quantifiable judgment index. At the same time, for fire protection facilities that require comprehensive judgment based on multiple indicators, an overall judgment is performed in accordance with the comprehensive judgment requirements in the technical standards.

[0116] For each key information output by the model, a one-to-one match is made with the judgment indicators in the background technical standard database. For example, the pressure value of the main network of the fire water supply system output by the model is compared with the qualified range of the main network pressure of the system stored in the database. If the pressure value is within the qualified range, the indicator is judged to be qualified; otherwise, it is judged to be unqualified. For some fire protection facilities that require comprehensive judgment of multiple indicators, such as fire doors, the model needs to comprehensively obtain the opening and closing action information obtained from the video, the appearance information obtained from the image, and the door closer status that may be obtained from the sensor data, and make an overall qualified or unqualified judgment in accordance with the comprehensive judgment requirements for fire doors in the technical standards.

[0117] The following is a specific example to illustrate the specific analysis process of the data analysis module: The data analysis unit can first click the smart analysis button to pop up the standard selection box, and select the technical standard on which the report is based, such as the industry standard XF503 "Technical Regulations for Inspection of Building Fire Protection Facilities". Then, the data stored in the data area of ​​each fire protection facility system is matched with the judgment indicators of the technical standard through the "Fire Protection Facility Detection Multimodal Large Language Model" to determine whether it is qualified.

[0118] If the selected technical standard is the industry standard XF503 "Technical Regulations for Inspection of Building Fire Protection Facilities", the technical requirements of XF503 for mechanical pressurized air supply systems are converted into technical indicators as shown in Table 1 below (only some are listed as examples): Table 1 System classification Single item name Technical requirements Mechanical pressurized air supply system Control cabinet The exterior has a logo indicating the system name and number The meter reading is within the normal range and the indicator light is on normally. With manual and automatic switching device System Features The wind speed at the air outlet should not be greater than 7m / s. The residual pressure value of the smoke-proof staircase should be 40Pa~50Pa The residual pressure value of the antechamber and shared antechamber should be 25Pa~30Pa. The mechanical pressurized air supply system linkage data, pictures, and videos obtained by the host can obtain the appearance information of the control cabinet or fan, whether the installation is firm, feedback signals and other technical information. Now, instead of manual judgment, the "fire protection facility detection multimodal large language model" can be used to analyze the information to determine whether each technical indicator of the system meets the requirements. As shown in Table 2.

[0119] Table 2 System classification Single item name Technical requirements Data collection method Post-analysis judgment Mechanical pressurized air supply system Control cabinet The exterior should have a sign indicating the system name and number. Image information Failure The meter reading is within the normal range and the indicator light is on normally. Video Information Failure There should be manual and automatic switching devices. Image information qualified System Features The wind speed at the air outlet should not be greater than 7m / s. Sensor information qualified The residual pressure value of the smoke-proof staircase should be 40Pa~50Pa Sensor information qualified The residual pressure value of the antechamber and shared antechamber should be 25Pa~30Pa. Sensor information qualified (3) Technical Standards Module 3 Technical standard module 3 is a database that organizes the technical indicators of different fire protection facilities systems in the technical standards according to different technical standards. When performing data analysis, by selecting the corresponding standard, the technical indicators of the technical standard in this module can be matched with the fire protection facilities system area one by one, which is the basis for the judgment of the data analysis module.

[0120] For example, the technical requirements for each fire protection facility system area in XF503 "Technical Code for Inspection of Building Fire Protection Facilities" can be understood by referring to Table 1 and Table 2.

[0121] (4) Data judgment module 4 The data determination module 4 comprehensively determines whether the product is qualified based on the determination results of each sub-item technical requirement. If all sub-items are qualified, the comprehensive determination is qualified; if there are items that are unqualified, the comprehensive determination is unqualified.

[0122] (5) Test report module 5 The detection report module 5 includes a report generation submodule 51 and a report printing submodule 52; the report generation submodule 51 is used to identify and organize the fire data and the conventional data according to the report template, import them into the position set by the report template, and organize and generate a report; this report template can refer to the report file. The report printing submodule 52 is used to download and save the generated report and print out the report.

[0123] Furthermore, in the detection report module 5, the process of generating a report is specifically as follows: A: Refer to the technical standard document to determine the basic structure template of the report, including a general data part and a fire protection facility data part. Set data placeholders in the basic structure template for subsequent import of data from the general data part and the fire protection facility data part. The general data part includes information such as report title, report number, inspection unit information, inspected building information, and inspection date. The fire protection facility data part is divided according to different system data areas, and each system contains information including inspection indicators, inspection results, and judgment conclusions.

[0124] For example, for the name of the detected building, you can set a placeholder called "building_name"; for the number of detectors in the automatic fire alarm system, you can set a placeholder like "fire_alarm_detector_number". These placeholders can be identified by specific markup languages ​​(such as XML tags or custom template tags) so that they can be accurately identified when data is imported.

[0125] The basic structure template can be stored in a variety of file formats, such as Word template files (.dotx or .dotm), HTML template files, or custom text template files. If you choose a Word template file, it can use the typesetting and formatting functions of Word software to easily generate a formatted and beautiful report; HTML template files are convenient for display and editing in a web page environment. For custom text template files, a set of strict grammatical rules needs to be defined to identify placeholders and template structures, and a corresponding parser needs to be developed to handle this template format.

[0126] Establish a template version management mechanism to record the modification history of the template and information about different versions. This helps to find the right version when you need to trace back the report format or update the template. You can use a version control system (such as Git) to manage the versions of template files. Each version has a corresponding version number, modification description, and update date.

[0127] B: Acquire the data processed by the data analysis module and regular data from each of the system data areas, clean the data, remove invalid data, duplicate data and erroneous data, and convert the acquired data according to the data type and format requirements of the data placeholders in the basic structure template.

[0128] C: Develop a data import program to identify the data placeholders in the basic structure template and fill the sorted data into the corresponding positions.

[0129] D: Generate a final report based on the basic structure template filled with data.

[0130] Generate the final report based on the template filled with data. If you are using a Word template file, you can use the automation interface of Microsoft Word (such as the COM interface) to generate the report and save the filled template as a complete Word document. For HTML template files, you can directly save the HTML file filled with data or further convert it to other formats such as PDF files. In the process of generating a report, you can add some additional formatting settings, such as adjusting the font, font size, paragraph spacing, etc., to ensure the beauty and readability of the report.

[0131] Provides multiple report output methods, such as saving to local disk (specify file path), printing directly or sending via email. If you choose to save to local disk, it can support multiple file formats (such as DOCX, PDF, etc.) to meet the needs of different users. For email sending, you can integrate mail clients (such as Microsoft Outlook or other SMTP mail servers) to send the generated report as an attachment to the specified recipient.

[0132] Second embodiment like Figure 3 As shown, this embodiment provides a firefighting facility detection method based on the Internet of Things and artificial intelligence using the firefighting facility detection system based on the Internet of Things and artificial intelligence as in the first embodiment, comprising the following steps: S1: Enter basic data including order data and project information, and classify fire protection facilities according to location and system based on fire protection drawings and fire protection facility location maps to generate different system data areas. At the same time, collect fire protection facility operation data and linkage data, and upload image data and video data to the system data area corresponding to the current data.

[0133] S2: Based on the technical standard selected for issuing the report, the data of each sub-item technical requirement entered or collected in the data area module are matched with the judgment indicators of the technical standard through the multimodal large language model of fire protection facility detection to determine whether it is qualified.

[0134] S3: Make a comprehensive determination of whether the product is qualified based on the results of each sub-item technical requirement.

[0135] S4: Generate a report and print it.

[0136] A computer-readable storage medium stores computer code. When the computer code is executed, the above method is executed. A person skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program. The program can be stored in a computer-readable storage medium. The storage medium can include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0137] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

[0138] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described 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.

[0139] It should be noted that the above embodiments can be freely combined as needed. The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention.

Claims

1. A fire protection facility detection system based on the Internet of Things and artificial intelligence, characterized in that: include: Data area module, data analysis module, technical standard module, data determination module and test report module; The data area module is used to input basic data including order data and project information, and classify fire protection facilities according to location and system according to fire protection drawings and fire protection facility location maps to generate different system data areas, collect fire protection facility operation data and linkage data, and upload image data and video data to the system data area corresponding to the current data; The data analysis module is used to match the data of each sub-item technical requirement entered or collected in the data area module with the determination index of the technical standard through the multimodal large language model of fire protection facility detection according to the technical standard selected for issuing the report, and determine whether it is qualified; The technical standard module is a database of technical indicators of different fire protection facility systems in the technical standards, which is sorted according to different technical standard contents; The data determination module comprehensively determines whether the requirements are qualified based on the determination results of each sub-item technical requirement; The test report module is used to generate and print a report.

2. The firefighting facility detection system based on the Internet of Things and artificial intelligence according to claim 1 is characterized in that: The data area module includes a conventional data submodule and a fire data submodule; The conventional data submodule includes an order data unit and a project data unit; Wherein, the order data unit is used to obtain and store the order data, and the project data unit is used to obtain and store the project information; The fire data submodule includes a data partition unit, a data acquisition unit and a data upload unit; Wherein, the data partitioning unit is used to classify the fire protection facilities according to the location and system according to the fire protection drawings and the fire protection facility point map uploaded by the user to generate different system data areas, so as to facilitate the classified transmission of the data acquired in the data acquisition unit and the data uploading unit to the fire protection facility system information cabinets corresponding to the different system data areas; The data acquisition unit is used to connect with the fire host through the fire information transmission device to realize the sharing of information in the fire host, obtain the operation data and linkage data of the fire facilities, collect the operation data of the fire facilities of the building through the fire Internet of Things device including the pressure sensor, the water level sensor and the flow sensor, and transmit the collected different system data to the corresponding system data area respectively according to the corresponding relationship between the fire host and the fire facility point map; The data uploading unit is used by the fire protection technical service agency to upload picture data and video data to the system data area of ​​the corresponding fire protection facility.

3. The firefighting facility detection system based on the Internet of Things and artificial intelligence according to claim 2 is characterized in that: In the data partitioning unit, the fire protection facilities are classified according to the location and system according to the fire protection drawings and the fire protection facility point map uploaded by the user to generate different system data areas, specifically: Perform image recognition or data analysis on the fire protection drawings and the fire protection facility point map to identify key information of the fire protection facilities including icons, names, and annotations, and create a system template corresponding to the current system according to different system types, wherein the system template includes the types of fire protection facilities that each system should have; According to the type of fire-fighting facilities included in each of the system templates, a search is performed in the fire-fighting drawings and the fire-fighting facility point map to obtain the locations and related parameters of the fire-fighting facilities in the searched system templates, and according to the distribution on the fire-fighting drawings and the fire-fighting facility point map, the system data area of ​​the system corresponding to the current system template is determined, and the relevant data of these fire-fighting facilities are sorted into the system data area.

4. The fire protection facility detection system based on the Internet of Things and artificial intelligence according to claim 2 is characterized in that: In the data acquisition unit, the fire information transmission device is connected to the fire host to realize the sharing of information in the fire host, obtain the operation data and linkage data of the fire facilities, and collect the operation data of the fire facilities of the building through the fire Internet of Things device including the pressure sensor, the water level sensor and the flow sensor. According to the corresponding relationship between the fire host and the fire facility point map, the collected different system data are respectively transmitted to the corresponding system data area, specifically: The data acquisition unit acquires the information in the fire host through the fire information transmission device in a polling or interrupt-driven manner, wherein in the polling mode, the data acquisition unit sends a query request to the fire host at a preset time interval to acquire the operation data and the linkage data during the operation of the fire protection facilities, and in the interrupt-driven mode, the data acquisition unit is actively notified when the fire host generates new data or the data changes; After the data acquisition unit acquires the information in the fire host, the acquired data is transmitted to the corresponding system data area according to a pre-planned path, wherein the pre-planned path is planned using a method including a shortest path algorithm; After the collected data is transferred to the corresponding system data area, a transaction processing mechanism is used to ensure the integrity and consistency of the data writing, wherein the transaction processing mechanism is used to handle failures that occur during the writing process.

5. The firefighting facility detection system based on the Internet of Things and artificial intelligence according to claim 2 is characterized in that: In the data uploading unit, the fire protection technical service agency uploads the picture data and video data to the system data area of ​​the corresponding fire protection facility, and further includes: In the process of uploading the image data, encryption technologies including SSL / TLS encryption protocols are used to encrypt the transmitted image data. At the same time, for large image data, a block uploading technology is used to divide the large image data into multiple small blocks for uploading; In the process of uploading the video data, the streaming media uploading technology is used to buffer and upload the video data. At the same time, when the network status is not good, the resolution and frame rate of the video data are adjusted to reduce the data volume of the video data.

6. The firefighting facility detection system based on the Internet of Things and artificial intelligence according to claim 1 is characterized in that: Also includes: Before the data analysis module, it is necessary to train the multimodal large language model for fire protection facility detection used to determine whether each sub-item technical requirement meets the technical standard, specifically: Collect a large amount of sensor data during the detection process of the fire-fighting facilities in advance, and perform preprocessing on the sensor data including noise removal and data standardization, collect a large amount of uploaded image data and video data, and annotate the image data with information including key parts of the fire-fighting facilities, component status, and key nameplate information, and annotate the video data with information including the action start time, key nodes, and whether the action is normal; Select an open-source multimodal large language training model, and fine-tune the model according to the particularity of the fire protection facility field, including modifying the input layer of the model to adapt to the sensor data, the image data, and the video data, defining mission objectives suitable for fire protection facilities, and adjusting parameters in the model; The sensor data is numerically encoded so that it can be accepted by the model, the image data is subjected to a feature vector extracted using a convolutional neural network (CNN) technique and converted into a feature representation, the video data is first decomposed into frames, and the feature of each frame is extracted using a convolutional neural network (CNN) technique, and the feature representation of the video is constructed in combination with the event sequence information of the video, and the encoded sensor data, image feature vector and video feature vector are fused to form multimodal input data; The model is trained for multiple rounds using the fused multimodal input data. In each round of training, the multimodal input data is fed into the model, and the model makes predictions based on predefined mission objectives suitable for firefighting facilities. The error between the prediction results and the labeled results is calculated, and the model parameters are adjusted based on the error. The model is continuously optimized until the model performance meets the preset requirements.

7. The firefighting facility detection system based on the Internet of Things and artificial intelligence according to claim 1 is characterized in that: In the data analysis module, according to the technical standard selected as the basis for issuing the report, the data of each sub-item technical requirement entered or collected in the data area module is matched with the judgment index of the technical standard through the multimodal large language model for fire protection facility detection to determine whether it is qualified, specifically: Conducting an in-depth analysis of the selected technical standards, converting the clauses in the technical standards into quantifiable judgment indicators, and performing structured storage of the quantified judgment indicators to facilitate model query; After the data stored in each of the system data areas are passed through the fire-fighting facility detection multimodal large language model, the model comprehensively processes the input sensor data, the image data and the video data, extracts text information related to the status of the fire-fighting facilities, and formats the text information output by the model so that it matches the stored quantifiable judgment indicators in data format and expression; Each of the text information output by the model is matched one-to-one with the quantifiable judgment index. At the same time, for fire protection facilities that require comprehensive judgment based on multiple indicators, an overall judgment is performed in accordance with the comprehensive judgment requirements in the technical standards.

8. The firefighting facility detection system based on the Internet of Things and artificial intelligence according to claim 1 is characterized in that: In the detection report module, the process of generating a report is specifically as follows: Referring to the technical standard document, a basic structure template of the report is determined, including a general data part and a fire protection facility data part, and a data placeholder is set in the basic structure template for subsequent import of data of the general data part and the fire protection facility data part, wherein the general data part includes information including a report title, a report number, information of a testing unit, information of a tested building, and a testing date, and the fire protection facility data part is divided according to different system data areas, and each system contains information including testing indicators, testing results, and judgment conclusions; Acquire the data processed by the data analysis module and regular data from each of the system data areas, clean the data, remove invalid data, duplicate data and erroneous data, and convert the acquired data according to the data type and format requirements of the data placeholders in the basic structure template; Develop a data import program to identify the data placeholders in the basic structure template and fill the sorted data into the corresponding positions; The final report is generated according to the basic structure template filled with data.

9. A firefighting facility detection method based on the Internet of Things and artificial intelligence using the firefighting facility detection system based on the Internet of Things and artificial intelligence as claimed in any one of claims 1 to 8, characterized in that: The following steps are involved: S1: Enter basic data including order data and project information, and classify fire protection facilities by location and system according to fire protection drawings and fire protection facility location maps to generate different system data areas, collect fire protection facility operation data and linkage data, and upload image data and video data to the system data area corresponding to the current data; S2: According to the selected technical standard for issuing the report, the data of each sub-item technical requirement entered or collected in the data area module is matched with the determination index of the technical standard through the multimodal large language model for fire protection facility detection to determine whether it is qualified; S3: Comprehensively determine whether the product is qualified based on the results of the determination of each sub-item technical requirement; S4: Generate a report and print it. 10 . A computer-readable storage medium storing a computer code, wherein when the computer code is executed, the method according to claim 9 is executed.

Images