Fire rescue method and system based on rescue helmet

Abstract

The embodiment of the invention discloses a fire rescue method and system based on a rescue helmet, and relates to the technical field of data processing. The method comprises the following steps: acquiring field data collected by a rescue helmet; performing anomaly detection on the field data to obtain an abnormal point; generating an abnormal parameter set according to the abnormal point; determining a risk coefficient corresponding to each type of field data based on the abnormal parameter set corresponding to all the field data, and determining a risk assessment weight corresponding to each type of field data according to the risk coefficient; determining a data capture range based on the fluctuation characteristics of the field data, and determining target characteristic data corresponding to each type of field data at each moment according to the data capture range; according to the target feature data and the risk assessment weight, obtaining a feature vector corresponding to each type of field data; and inputting the feature vectors corresponding to all the field data into a rescue risk assessment model for processing to obtain a fire rescue risk assessment result.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a firefighting and rescue method and system based on a rescue helmet. Background Art

[0002] When rescuing at a fire scene, real-time acquisition of fire scene data and firefighters' relevant status data is crucial for firefighters' safety and rescue efficiency. Rescue helmets are a special device designed for firefighting and rescue. They can not only provide basic head protection functions, but also collect and transmit on-site environmental data (such as fire scene temperature data, gas concentration data, etc.) and firefighters' status data (such as motion status data, physiological status data, etc.) in real time, thereby providing the command center with comprehensive situational awareness and decision support.

[0003] However, traditional methods for processing on-site data collected by rescue helmets often lack a global perspective and only focus on the fluctuations of local data, ignoring the sporadic nature of data and the overall trend of global data. For example, the data fluctuation of a single sensor may be caused by an occasional situation and is not representative; for another example, the data fluctuation of a single sensor may not be significant, but the comprehensive analysis of data from multiple sensors can reveal larger trends or hidden dangers.

[0004] Based on this, it is necessary to improve the processing process of the on-site data collected by the rescue helmets to better provide rescue risk monitoring for the command center. Summary of the invention

[0005] To achieve the above objectives, one aspect of an embodiment of this specification provides a fire rescue method based on a rescue helmet, the method comprising:

[0006] Acquire on-site data collected by the rescue helmet, wherein the on-site data includes fire scene temperature data, firefighter movement status data, firefighter physiological data, gas concentration data, and obstacle detection data;

[0007] Performing anomaly detection on the field data to obtain anomaly points in each type of the field data;

[0008] Generate a set of abnormal parameters corresponding to each type of the field data according to the abnormal points, wherein the abnormal parameters include a timestamp and an abnormal degree corresponding to each abnormal point;

[0009] Determine a risk coefficient corresponding to each type of the field data based on a set of abnormal parameters corresponding to all the field data, and determine a risk assessment weight corresponding to each type of the field data according to the risk coefficient;

[0010] Determine a data capture range corresponding to each type of the field data based on the fluctuation characteristics of the field data, and determine target feature data corresponding to each type of the field data at each moment according to the data capture range;

[0011] According to the target feature data and the risk assessment weight, a feature vector corresponding to each type of the field data at each moment is obtained;

[0012] The characteristic vectors corresponding to all the on-site data are input into the rescue risk assessment model for processing to obtain a fire rescue risk assessment result.

[0013] In some embodiments, performing anomaly detection on the field data to obtain anomaly points in each type of the field data includes:

[0014] Calculate the difference between all adjacent data in the time series data corresponding to each type of field data to obtain a continuous difference sequence;

[0015] Based on the standard score detection, the mutation point in the continuous differential sequence is determined, and the abnormal point in the corresponding field data is determined according to the mutation point.

[0016] In some embodiments, determining the risk coefficient corresponding to each type of the field data based on the abnormal parameter set corresponding to all the field data includes:

[0017] Obtain the timestamps and abnormality levels corresponding to all abnormal points in each type of field data, and obtain the corresponding abnormal time series and abnormality level series;

[0018] Calculate the degree of dispersion and the influence coefficient corresponding to each abnormal point in each type of field data based on the abnormal time series and the abnormal degree series;

[0019] The risk coefficient corresponding to each type of the field data is determined according to the degree of dispersion, the influence coefficient and the total number of abnormal points appearing in the field data.

[0020] In some embodiments, the calculation of the discrete degree and influence coefficient corresponding to each abnormal point in each type of field data based on the abnormal time series and the abnormal degree series includes:

[0021] Determine a first ratio of the total number of timestamps in the abnormal time series corresponding to the target field data to the total amount of data of the target field data from the initial statistical time point to the target time point corresponding to the target abnormal point, and a first difference between a first data disorder degree corresponding to all data in the target field data up to the target abnormal point and a second data disorder degree corresponding to all non-abnormal data up to the target abnormal point, and then determine the discrete degree corresponding to the target abnormal point in the target field data based on the first ratio and the first difference;

[0022] Determine a first dynamic time warping distance between an abnormal time series corresponding to a target abnormal point in the target field data and an abnormal time series corresponding to abnormal points with the same sequence number in other field data, and a second dynamic time warping distance between an abnormal degree series corresponding to the target abnormal point in the target field data and an abnormal degree series corresponding to abnormal points with the same sequence number in other field data, and then obtain an influence coefficient corresponding to the target abnormal point in the target field data based on an average value of the sum of the first dynamic time warping distance and the second dynamic time warping distance corresponding to the target abnormal point in the target field data and abnormal points with the same sequence number in all other field data.

[0023] In some embodiments, determining the risk coefficient corresponding to each type of the field data according to the discrete degree, the influence coefficient and the total number of abnormal points appearing in the field data includes:

[0024] Calculate a second ratio of the influence coefficient and the degree of dispersion corresponding to each abnormal point in the target field data;

[0025] An average value of the second ratios corresponding to all abnormal points in the target field data is used as the risk coefficient corresponding to the target field data.

[0026] In some embodiments, determining the risk assessment weight corresponding to each type of the on-site data according to the risk coefficient includes:

[0027] A third ratio of the risk coefficient corresponding to the target field data to the sum of the risk coefficients corresponding to all field data is calculated, and the third ratio is used as the risk assessment weight corresponding to the target field data.

[0028] In some embodiments, determining the data capture range corresponding to each type of the field data based on the fluctuation characteristics of the field data includes:

[0029] For a target moment corresponding to the target field data, calculating a fluctuation regularity index corresponding to the target moment and a fluctuation influence index corresponding to the target field data;

[0030] Based on the fluctuation regularity index, the fluctuation influence index and the oscillation maintenance time in the target field data within a time period before the target time, a data capture range corresponding to the target time in the target field data is determined.

[0031] In some embodiments, the calculating of the fluctuation regularity index corresponding to the target time and the fluctuation impact index corresponding to the target field data includes:

[0032] Obtaining a third data disorder degree of a fluctuation level corresponding to data in the target field data in a time period before the target moment and a fourth data disorder degree of a steady-state level corresponding to data in the target field data in a time period before the target moment, and determining a fluctuation regularity index corresponding to the target moment based on a second difference between the third data disorder degree and the fourth data disorder degree and an average value of steady-state levels corresponding to all data in the target field data in a time period before the target moment;

[0033] The average value of the oscillation maintenance time of the abnormal points within the local statistical time in the target field data, the average value of the steady-state maintenance time within the local statistical time, and the distribution variance corresponding to adjacent moments within the local statistical time are obtained, and based on the distribution variance corresponding to the adjacent moments and the third difference between the average value of the oscillation maintenance time and the average value of the steady-state maintenance time, the fluctuation impact index corresponding to the target field data is determined.

[0034] In some embodiments, determining the target feature data corresponding to each type of the field data at each moment according to the data capture range includes:

[0035] Determine the data capture quantity S in each type of the field data according to the data capture range;

[0036] For the target moment, the data collected at S consecutive moments before the target moment in each type of field data is taken as the target data, and the mean value, variance and entropy value of the target data are calculated as the target feature data corresponding to each type of field data at the target moment.

[0037] Another aspect of the present specification also provides a fire rescue system based on a rescue helmet, the system comprising:

[0038] An acquisition module is used to acquire on-site data collected by a rescue helmet, wherein the on-site data includes fire scene temperature data, firefighter movement status data, firefighter physiological data, gas concentration data, and obstacle detection data;

[0039] An anomaly detection module, used to perform anomaly detection on the field data and obtain anomaly points in each type of the field data;

[0040] An abnormal parameter set generation module, used to generate an abnormal parameter set corresponding to each type of field data according to the abnormal point, wherein the abnormal parameters include a timestamp and an abnormal degree corresponding to each abnormal point;

[0041] A risk assessment weight determination module, used to determine the risk coefficient corresponding to each type of the field data based on the abnormal parameter set corresponding to all the field data, and determine the risk assessment weight corresponding to each type of the field data according to the risk coefficient;

[0042] A target characteristic data determination module, used to determine a data capture range corresponding to each type of the field data based on the fluctuation characteristics of the field data, and determine the target characteristic data corresponding to each type of the field data at each moment according to the data capture range;

[0043] A feature vector generation module, used for obtaining a feature vector corresponding to each type of the field data at each moment according to the target feature data and the risk assessment weight;

[0044] The risk assessment result determination module is used to input the characteristic vectors corresponding to all the on-site data into the rescue risk assessment model for processing to obtain the fire rescue risk assessment result.

[0045] The firefighting and rescue method and system based on the rescue helmet provided in the embodiments of this specification may bring at least the following beneficial effects:

[0046] (1) In the firefighting and rescue method and system based on rescue helmets provided in some embodiments of the present specification, by considering the distribution and change law characteristics of different on-site data, and then combining the correlation between different on-site data to determine the corresponding risk coefficient to measure the reliability of each on-site data for assessing the firefighting and rescue risk, and by calculating the risk assessment weight corresponding to each on-site data according to the risk coefficient, and finally determining the feature vector corresponding to each on-site data by the risk assessment weight and determining the firefighting and rescue risk assessment result based on the feature vector, the accuracy of the final firefighting and rescue risk assessment result can be improved to a certain extent, and the interference of occasional data fluctuations on the risk assessment result can be avoided;

[0047] (2) In the firefighting and rescue method and system based on the rescue helmet provided in some embodiments of the present specification, the corresponding data capture range is determined by considering the fluctuation characteristics of each type of on-site data, and the target feature data corresponding to each type of on-site data at each moment is determined based on the data capture range. Redundant data or irrelevant data can be filtered out from each type of on-site data, thereby reducing the probability of using it for subsequent processing, reducing the subsequent computing load to a certain extent, and avoiding interference with the final firefighting and rescue risk assessment results.

[0048] Additional features will be partially described in the following description. For those skilled in the art, the technical solutions provided by this application will become apparent by referring to the following description and drawings, or can be understood by the generation or operation of examples. The features of this specification can be implemented and obtained by practicing or using various aspects of the methods, tools and combinations described in the following detailed examples. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] This specification will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0050] Figure 1 is a schematic diagram of an exemplary application scenario of a fire rescue system based on a rescue helmet according to some embodiments of this specification;

[0051] Figure 2 is an exemplary module diagram of a fire rescue system based on a rescue helmet according to some embodiments of this specification;

[0052] Figure 3 It is an exemplary flow chart of a firefighting and rescue method based on a rescue helmet according to some embodiments of this specification. DETAILED DESCRIPTION

[0053] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following is a brief introduction to the drawings required for the description of the embodiments. Obviously, the drawings described below are only some examples or embodiments of this specification. For ordinary technicians in this field, this specification can also be applied to other similar scenarios based on these drawings without creative work. Unless it is obvious from the language environment or otherwise explained, the same reference numerals in the figures represent the same structure or operation.

[0054] It should be understood that the "system", "device", "unit" and / or "module" used in this specification is a method for distinguishing different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.

[0055] As shown in this specification and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "comprise" and "include" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0056] Flowcharts are used in this specification to illustrate the operations performed by the system according to the embodiments of this specification. It should be understood that the preceding or following operations are not necessarily performed precisely in order. Instead, the steps may be processed in reverse order or simultaneously. At the same time, other operations may also be added to these processes, or one or more operations may be removed from these processes.

[0057] The firefighting and rescue method and system based on the rescue helmet provided in the embodiments of this specification are described in detail below in conjunction with the accompanying drawings.

[0058] Figure 1 It is a schematic diagram of an exemplary application scenario of a fire rescue system based on a rescue helmet as shown in some embodiments of this specification.

[0059] Reference Figure 1 In some embodiments, the application scenario 100 of the fire rescue system based on the rescue helmet may include a rescue helmet 110, a storage device 120, a processing device 130, a monitoring terminal 140 and a network 150. The various components in the application scenario 100 can be connected in a variety of ways. For example, the rescue helmet 110 can be connected to the storage device 120 and / or the processing device 130 through the network 150, or directly connected to the storage device 120 and / or the processing device 130. For another example, the storage device 120 can be directly connected to the processing device 130 or connected through the network 150. For another example, the monitoring terminal 140 can be connected to the storage device 120 and / or the processing device 130 through the network 150, or directly connected to the storage device 120 and / or the processing device 130.

[0060] The rescue helmet 110 can be worn on the head of the firefighter during the firefighting and rescue process to protect the firefighter's head and collect on-site data during the rescue process. In some embodiments, the rescue helmet 110 may include a temperature sensor, a gas concentration sensor, an ultrasonic ranging sensor, a motion sensor, a physiological monitoring sensor, etc. The on-site data may include fire scene temperature data (obtained through a temperature sensor), gas concentration data (such as carbon monoxide concentration, carbon dioxide concentration, oxygen concentration, etc., obtained through a gas concentration sensor), obstacle detection data (through an ultrasonic ranging sensor), firefighter motion state data (such as acceleration, angular velocity, etc., obtained through a motion sensor), and firefighter physiological data (such as heart rate, blood pressure, etc., obtained through a physiological monitoring sensor). In some embodiments, the rescue helmet 110 can collect on-site data during the firefighting and rescue process according to a set data collection frequency (for example, once every 1 second). In some embodiments, the rescue helmet 110 may have an independent power supply, which can send the collected on-site data to other components in the application scenario 100 (for example, storage device 120, processing device 130, monitoring terminal 140) by wired or wireless (such as Bluetooth, WiFi, etc.).

[0061] In some embodiments, the rescue helmet 110 can send the collected field data to the storage device 120, the processing device 130, the monitoring terminal 140, etc. through the network 150. In some embodiments, the field data collected by the rescue helmet 110 can be processed by the processing device 130. For example, the processing device 130 can determine the feature vector corresponding to each type of field data at each moment, and input the feature vectors corresponding to all field data into the rescue risk assessment model for processing to obtain the fire rescue risk assessment result. In some embodiments, the feature vector and / or the fire rescue risk assessment result can be sent to the storage device 120 for recording, or sent to the monitoring terminal 140 for feedback to the command center.

[0062] The network 150 can facilitate the exchange of information and / or data. The network 150 may include any suitable network that can facilitate the exchange of information and / or data of the application scenario 100. In some embodiments, at least one component of the application scenario 100 (e.g., the rescue helmet 110, the storage device 120, the processing device 130, the monitoring terminal 140) can exchange information and / or data with at least one other component in the application scenario 100 through the network 150. For example, the processing device 130 can obtain field data from the rescue helmet 110 and / or the storage device 120 through the network 150. For another example, the processing device 130 can obtain user operation instructions from the monitoring terminal 140 through the network 150. Exemplary operation instructions may include, but are not limited to, consulting field data, reading the fire rescue risk assessment results corresponding to each moment determined based on the field data, etc.

[0063] In some embodiments, the network 150 may be any form of wired or wireless network, or any combination thereof. By way of example only, the network 150 may include a cable network, a wired network, an optical fiber network, a telecommunications network, an intranet, the Internet, a local area network (LAN), a wide area network (WAN), a wireless local area network (WLAN), a metropolitan area network (MAN), a public switched telephone network (PSTN), a Bluetooth network, a ZigBee network, a near field communication (NFC) network, etc., or any combination thereof. In some embodiments, the network 150 may include at least one network access point, and at least one component of the application scenario 100 may be connected to the network 150 through the access point to exchange data and / or information.

[0064] The storage device 120 may store data, instructions and / or any other information. In some embodiments, the storage device 120 may store data obtained from the rescue helmet 110, the processing device 130 and / or the monitoring terminal 140. For example, the storage device 120 may store the field data collected by the rescue helmet 110; for another example, the storage device 120 may store the feature vector corresponding to each field data calculated by the processing device 130 at each time, and the fire rescue risk assessment result corresponding to each time determined based on the feature vector. In some embodiments, the storage device 120 may store data and / or instructions used by the processing device 130 to execute or use to complete the exemplary method described in this specification. In some embodiments, the storage device 120 may include a large capacity memory, a removable memory, a volatile read-write memory, a read-only memory (ROM), etc., or any combination thereof. Exemplary large capacity memory may include a disk, an optical disk, a solid-state disk, etc. In some embodiments, the storage device 120 may be implemented on a cloud platform. By way of example only, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-layer cloud, etc., or any combination thereof.

[0065] In some embodiments, the storage device 120 may be connected to the network 150 to communicate with at least one other component in the application scenario 100 (e.g., the rescue helmet 110, the processing device 130, the monitoring terminal 140). At least one component in the application scenario 100 may access data, instructions, or other information stored in the storage device 120 through the network 150. In some embodiments, the storage device 120 may be directly connected or communicate with one or more components in the application scenario 100 (e.g., the rescue helmet 110, the monitoring terminal 140). In some embodiments, the storage device 120 may be part of the rescue helmet 110 and / or the processing device 130.

[0066] The processing device 130 can process data and / or information obtained from the rescue helmet 110, the storage device 120, the monitoring terminal 140 and / or other components of the application scenario 100. In some embodiments, the processing device 130 can obtain field data from any one or more of the rescue helmet 110, the storage device 120 or the monitoring terminal 140, process the field data to determine the feature vector corresponding to each type of field data at each moment, and determine the fire rescue risk assessment result corresponding to each moment based on the feature vectors corresponding to all field data. In some embodiments, the processing device 130 can obtain pre-stored computer instructions from the storage device 120, and execute the computer instructions to implement the fire rescue method based on the rescue helmet described in this specification.

[0067] In some embodiments, the processing device 130 may be a single server or a server group. The server group may be centralized or distributed. In some embodiments, the processing device 130 may be local or remote. For example, the processing device 130 may access information and / or data from the rescue helmet 110, the storage device 120, and / or the monitoring terminal 140 via the network 150. For another example, the processing device 130 may be directly connected to the rescue helmet 110, the storage device 120, and / or the monitoring terminal 140 to access information and / or data. In some embodiments, the processing device 130 may be implemented on a cloud platform. For example, a cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an inter-cloud cloud, a multi-cloud, etc., or any combination thereof.

[0068] The monitoring terminal 140 can receive, send and / or display data. The received data may include field data collected by the rescue helmet 110, data stored in the storage device 120, data processed by the processing device 130, etc. Among them, the sent data may include input data and / or voice commands of the user (such as the person in charge of the command center). For example, the monitoring terminal 140 can send the voice command input by the command center to the rescue helmet 110 through the network 150, and output it to the firefighter through the voice output device (such as a speaker) of the rescue helmet 110. For another example, the monitoring terminal 140 can send the data processing command input by the user to the processing device 130 through the network 150 to control the processing device 130 to perform corresponding data processing.

[0069] In some embodiments, the monitoring terminal 140 may include a mobile device 141, a tablet computer 142, a laptop computer 143, etc., or any combination thereof. For example, the mobile device 141 may include a mobile phone, a personal digital assistant (PDA), a dedicated mobile terminal (e.g., a fire command terminal), etc., or any combination thereof. In some embodiments, the monitoring terminal 140 may include an input device (e.g., a keyboard, a touch screen), an output device (e.g., a display, a speaker), etc. In some embodiments, the processing device 130 may be a part of the monitoring terminal 140.

[0070] It should be noted that the above description of the application scenario 100 is only for example and explanation, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to the application scenario 100 under the guidance of this specification. However, these modifications and changes are still within the scope of this specification. In some embodiments, the rescue helmet 110 may include more or fewer functional components, such as a microphone, a speaker, etc.

[0071] Figure 2 Schematic diagram of a fire rescue system based on a rescue helmet according to some embodiments of this specification. Figure 2 The firefighting and rescue system 200 based on the rescue helmet shown can be applied to Figure 1 The application scenario 100 shown, for example, can be configured in the form of software and / or hardware to the processing device 130 and / or the monitoring terminal 140 to process the field data collected by the rescue helmet 110, and determine the feature vector corresponding to each type of field data at each moment based on the field data, and then determine the fire rescue risk assessment result corresponding to each moment based on the feature vectors corresponding to all the field data.

[0072] Reference Figure 2In some embodiments, the fire rescue system 200 based on the rescue helmet may include an acquisition module 210, an anomaly detection module 220, an abnormal parameter set generation module 230, a risk assessment weight determination module 240, a target feature data determination module 250, a feature vector generation module 260 and a risk assessment result determination module 270.

[0073] The acquisition module 210 can be used to acquire on-site data collected by the rescue helmet, where the on-site data includes fire scene temperature data, firefighter movement status data, firefighter physiological data, gas concentration data, and obstacle detection data.

[0074] The anomaly detection module 220 may be used to perform anomaly detection on the field data and obtain anomaly points in each type of the field data.

[0075] The abnormal parameter set generating module 230 may be used to generate an abnormal parameter set corresponding to each type of the field data according to the abnormal point, wherein the abnormal parameters include a timestamp and an abnormal degree corresponding to each abnormal point.

[0076] The risk assessment weight determination module 240 may be configured to determine a risk coefficient corresponding to each type of the field data based on a set of abnormal parameters corresponding to all of the field data, and determine a risk assessment weight corresponding to each type of the field data according to the risk coefficient.

[0077] The target characteristic data determination module 250 can be used to determine the data capture range corresponding to each type of the field data based on the fluctuation characteristics of the field data, and determine the target characteristic data corresponding to each type of the field data at each moment according to the data capture range.

[0078] The feature vector generation module 260 can be used to obtain a feature vector corresponding to each type of the field data at each moment according to the target feature data and the risk assessment weight.

[0079] The risk assessment result determination module 270 can be used to input the characteristic vectors corresponding to all the on-site data into the rescue risk assessment model for processing to obtain the fire rescue risk assessment result.

[0080] For more details about the above modules, please refer to other places in this manual (for example Figure 3 part and its related description), which will not be repeated here.

[0081] It should be understood that Figure 2The fire rescue system 200 based on the rescue helmet and its modules shown can be implemented in various ways. For example, in some embodiments, the system and its modules can be implemented by hardware, software or a combination of software and hardware. Among them, the hardware part can be implemented by using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or a dedicated design hardware. Those skilled in the art will understand that the above methods and systems can be implemented using computer executable instructions and / or included in processor control codes, such as such codes provided on carrier media such as disks, CDs or DVD-ROMs, programmable memories such as read-only memories (firmware), or data carriers such as optical or electronic signal carriers. The system and its modules of this specification can be implemented not only by hardware circuits such as ultra-large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., but also by software implemented by, for example, various types of processors, and can also be implemented by a combination of the above hardware circuits and software (e.g., firmware).

[0082] It should be noted that the above description of the fire rescue system 200 based on the rescue helmet is provided for illustrative purposes only and is not intended to limit the scope of this specification. It is understandable that for those skilled in the art, according to the description of this specification, without departing from this principle, the various modules can be arbitrarily combined, or a subsystem can be formed to connect with other modules. For example, Figure 2 The acquisition module 210, the anomaly detection module 220, the anomaly parameter set generation module 230, the risk assessment weight determination module 240, the target feature data determination module 250, the feature vector generation module 260 and the risk assessment result determination module 270 described in the embodiment can be different modules in a system, or a module can realize the functions of two or more of the above modules. Such variations are within the scope of protection of this specification. In some embodiments, the aforementioned modules can be part of the processing device 130 and / or the monitoring terminal 140.

[0083] Figure 3 is an exemplary flow chart of a firefighting and rescue method based on a rescue helmet according to some embodiments of this specification. In some embodiments, the firefighting and rescue method 300 based on a rescue helmet can be executed by processing logic, which can include hardware (e.g., circuits, dedicated logic, programmable logic, microcode, etc.), software (instructions running on a processing device to perform hardware simulation), etc. or any combination thereof. In some embodiments, Figure 3 One or more operations in the flowchart of the fire rescue method 300 based on the rescue helmet can be performed by Figure 1The processing device 130 and / or the monitoring terminal 140 shown are implemented. For example, the firefighting and rescue method 300 based on the rescue helmet can be stored in the storage device 120 in the form of instructions, and called and / or executed by the processing device 130 and / or the monitoring terminal 140. The specific execution process of the firefighting and rescue method 300 based on the rescue helmet is described below by taking the processing device 130 as an example.

[0084] Reference Figure 3 In some embodiments, the firefighting and rescue method 300 based on the rescue helmet may include:

[0085] S310, acquiring on-site data collected by the rescue helmet. In some embodiments, S310 may be executed by the acquisition module 210.

[0086] In some embodiments, the acquisition module 210 may be in communication connection with the aforementioned rescue helmet 110, and the acquisition module 110 may directly acquire from the rescue helmet 110 the on-site data collected by various sensors configured in the rescue helmet 110. In the embodiments of this specification, the on-site data may include fire scene temperature data, firefighter motion state data, firefighter physiological data, gas concentration data, and obstacle detection data.

[0087] Among them, the fire scene temperature data can be collected by a temperature sensor configured in the rescue helmet 110. In some embodiments, the temperature sensor may include a DHT1 1 module, which can record the temperature of the fire scene at all times under extreme temperatures, helping firefighters to understand the changes in the fire scene in real time and detect the danger of the fire scene.

[0088] The firefighter's motion state data can be collected by the motion sensor configured in the rescue helmet 110, which may include the firefighter's motion acceleration, angular velocity, etc. In some embodiments, the motion state (such as walking, running, stillness, falling, etc.) can be identified based on the acceleration and angular velocity. Specifically, in some embodiments, the MPU6050 module (three-axis gyroscope + three-axis acceleration) can be used to collect the firefighter's motion state data, which can accurately record the posture of the device and judge the firefighter's current body posture through the corresponding code, thereby realizing the perception of the firefighter's motion state. When the firefighter falls or falls to the ground, rescue can be carried out in the first time, achieving early detection and early assistance, thereby ensuring the life safety of the firefighter during the rescue process.

[0089] The firefighter's physiological data may be collected by the physiological monitoring sensor configured in the rescue helmet 110. Exemplary firefighter's physiological data may include heart rate, blood pressure, etc. It can be understood that the firefighter's physiological data may reflect whether the firefighter's current physical condition is suitable for continuing to perform firefighting and rescue tasks, thereby preventing the firefighter from entering the fire scene with physical overdrawn or poor physical condition, thereby increasing their own risk factor.

[0090] The gas concentration data can be collected by the gas concentration sensor configured in the rescue helmet 110, which may include but is not limited to carbon monoxide concentration, carbon dioxide concentration, oxygen concentration, etc. This data can reflect the difficulty of the rescue conditions at the fire scene. Generally, when the carbon monoxide concentration exceeds 800ppm, it will cause coma within 45 minutes and will be fatal within 2-3 hours; when the carbon dioxide concentration exceeds 30000ppm (3%), it will cause headaches, dizziness and hearing impairment; when the oxygen concentration is less than 17%, it will cause a decrease in judgment and make physical activities difficult.

[0091] The obstacle detection data can be collected by the ultrasonic ranging sensor configured in the rescue helmet 110. The obstacle detection data can reflect the distance between the obstacles at the rescue site and the firefighters, thereby improving the firefighters' ability to perceive and be perceived in a small space. In some embodiments, the obstacle detection data can reflect phenomena such as the collapse of buildings and the dumping of objects at the fire scene.

[0092] In some embodiments, the on-site data collected by the rescue helmet 110 can be stored in the storage device 120 , and the acquisition module 210 can acquire the above on-site data from the storage device 120 .

[0093] It should be pointed out that due to the complex and changeable situation at the rescue site, the on-site data collected by the rescue helmet 110 may also exist in various situations. It is difficult to accurately analyze the actual situation of the fire scene and the firefighter's danger index at the current moment through a single sensor data.

[0094] For example, when the obstacle detection data changes suddenly, it may be due to the collapse of the building or the dumping of objects at the fire scene, but the data alone cannot directly determine whether the firefighters are in danger at the rescue scene and whether they need assistance. It is understandable that when firefighters encounter dangers such as building collapse or object dumping at the fire scene, they will usually dodge or avoid due to conditioned reflexes, and when the building collapse or object dumping does not pose a danger to the firefighters, the body usually does not have a large movement reaction. Therefore, at this time, if the firefighters' motion state data is combined to know that the firefighters have similar motion state data fluctuations (even if the firefighters are hit by collapsed objects, similar motion state data fluctuations will be generated), it indicates that the firefighters may be in danger. On the contrary, if the firefighters' motion state data is combined to know that the firefighters did not dodge or avoid, it means that the building collapse or object dumping did not pose a threat to the firefighters, and the firefighters are currently in a safe state.

[0095] For another example, when the fire scene temperature data and / or gas concentration data collected by the rescue helmet 110 suddenly changes, it may be because the firefighter suddenly enters another space with more intense combustion (for example, opening a fire door). At this time, if the firefighter's heart rate, blood pressure and other physiological data reflect a large fluctuation, it means that the fire scene may be more dangerous, and the firefighter's heart rate and blood pressure increase due to fear. On the contrary, if the firefighter's heart rate, blood pressure and other physiological data do not have a large fluctuation at this time, it means that the fire scene situation may still be within the controllable range, and the degree of danger is still in a relatively low state.

[0096] Based on the above problems, in some embodiments of the present specification, in order to accurately obtain the fire rescue risk assessment result, various sensors configured in the fire helmet 110 can be used to collect different rescue scene data, and then the fire rescue risk assessment result at the current moment can be determined through the correlation and change characteristics between various scene data. The specific process is detailed in steps S320 to S370.

[0097] In some embodiments of the present specification, in order to avoid the problem that the subsequent calculation cannot be accurately performed due to the data loss phenomenon in the process of the rescue helmet 110 sending the collected field data to the monitoring terminal 140, the rescue helmet 110 can use the interpolation method or the mean filling method to pre-process the field data corresponding to each moment after sending the collected field data to the monitoring terminal 140, so as to fill the missing data. Specifically, after the detection terminal 140 receives the field data collected by the rescue helmet, the field data can be transmitted to the pre-processing system configured in the detection terminal 140, and then the field data at each moment is checked for missing data. If there is a missing, the mean of the data within a certain time range (for example, 5 sampling cycles) before and after the moment can be calculated, and the mean can be used as the data at the moment, so as to ensure that the continuity and integrity of the data are fully guaranteed, and the core features of the original data are restored to the maximum extent, so as to prevent the inaccurate impact on the subsequent data analysis results due to the loss of data during the data transmission process. It should be pointed out that by pre-processing the field data as above, the accuracy and reliability of the data can be ensured, providing a more solid foundation for subsequent data analysis.

[0098] S320 , performing anomaly detection on the field data to obtain anomaly points in each type of the field data. In some embodiments, S320 may be performed by the anomaly detection module 220 .

[0099] After obtaining the above-mentioned field data through the aforementioned process, the anomaly detection module 220 can perform anomaly detection on each type of field data respectively, so as to obtain anomalies in each type of field data. Specifically, in some embodiments of the present specification, the difference between all adjacent data in the time series data (i.e., the data column recorded in time order, also called time series data) corresponding to each type of field data can be calculated to obtain a continuous difference sequence, and then the mutation point in the continuous difference sequence is determined based on the standard score (z-score) detection, and the anomaly in the corresponding field data is determined according to the mutation point.

[0100] For example, suppose the time series data corresponding to a certain field data is {x t1 , x t2 , x t3 , ..., x tn}, then the corresponding continuous difference sequence is {d1, d2, d3, ..., d n-1}, where d i =x t(i+1) -x ti ,i=1,2,3,…,n-1.

[0101] Furthermore, after obtaining the continuous difference sequence, the z-score can be used to detect the mutation point in the continuous difference sequence. The z-score is a commonly used statistical method that can be used to measure the degree to which a data point deviates from its mean value. Specifically, for each data point d in the continuous difference sequence i , we can calculate the corresponding standard score z i :

[0102]

[0103] Among them, μ is the mean of the aforementioned continuous difference sequence, and σ1 is the standard deviation of the continuous difference sequence.

[0104] By calculating the standard score, we can identify points that deviate far from the mean, i.e., mutation points. Usually, we can set a threshold (e.g. |z i |>3), when the standard score exceeds the threshold, the point is considered to be a mutation point.

[0105] Furthermore, the anomaly detection module 220 can determine the anomaly points in the corresponding field data according to the mutation points determined by the above process. i When l>3, x i+1 Considered as an outlier.

[0106] In some other embodiments, the BG (Bernao1a Ga1van) sequence segmentation algorithm (a sequence segmentation algorithm based on Bayesian inference) can also be used to process the time series data corresponding to each field data, so as to identify the mutation points in the time series data. It should be noted that the BG sequence segmentation algorithm can be regarded as a well-known technology, and its specific process will not be repeated in this specification.

[0107] S330, generating an abnormal parameter set corresponding to each type of the field data according to the abnormal point, wherein the abnormal parameter includes a timestamp and an abnormal degree corresponding to each abnormal point. In some embodiments, S330 may be executed by the abnormal parameter set generation module 230.

[0108] In the embodiment of the present specification, the abnormal parameter set generation module 230 can generate an abnormal parameter set corresponding to each type of field data according to the abnormal points determined in the above process, wherein the abnormal parameters include a timestamp and abnormality degree corresponding to each abnormal point.

[0109] Specifically, in some embodiments, the abnormal parameter set generation module 230 can use the sequence of timestamps corresponding to all abnormal points in each type of field data as the abnormal time series corresponding to the field data. At the same time, the abnormal parameter set generation module 230 can obtain the difference between the data corresponding to each abnormal point and the data at the previous moment, and use the ratio of the difference to the data of the corresponding abnormal point as the abnormal degree corresponding to the abnormal point; then, according to the abnormal degree corresponding to all abnormal points in each type of field data, the abnormal degree sequence corresponding to the field data is obtained.

[0110] Specifically, it is assumed that the time corresponding to a certain target abnormal point (for example, the kth abnormal point) in the target scene data (the target scene data may be any one of the fire scene temperature data, the firefighter motion state data, the firefighter physiological data, the gas concentration data and the obstacle detection data) collected by the rescue helmet 110 is t 20 The k-1 outlier points before the target outlier point correspond to the times t5, t9, t 19 , then the abnormal time series corresponding to the target field data can be expressed as X(k)=(t5, t9, t 19 , t 20 Specifically, if the data corresponding to time t4 (for example, fire scene temperature data, in degrees Celsius) is 19, the data corresponding to time t5 is 27, the data corresponding to time t8 is 31, and the data corresponding to time t9 is 29, then 18 The data corresponding to the time is 49, t 19 The data corresponding to the time is 54, t 20 The data corresponding to the time is 53, then the difference between the data corresponding to time t5 and time t4 is 8, the difference between the data corresponding to time t9 and time t8 is 2, t 19 Time and t z8 The difference between the data corresponding to the time is 5, t 20 Time and t 19 The difference between the data corresponding to the time instants is 1. At this time, the abnormality degree sequence corresponding to the target field data (keeping two decimal places) can be expressed as Y(k)=(0.29, 0.07, 0.09, 0.02).

[0111] In summary, if a certain field data contains 4 abnormal points, their corresponding timestamps are t5, t9, t 19 , t 20 , and their corresponding abnormal degrees are 0.29, 0.07, 0.09, and 0.02 respectively. The abnormal parameter set corresponding to the field data can be expressed as {(t5, 0.29), (t9, 0.07), (t 19 , 0.09), (t 20, 0.02)}, each element in the abnormal parameter set can be understood as a two-dimensional array, which represents the timestamp and abnormal degree corresponding to each abnormal point.

[0112] S340, based on the abnormal parameter set corresponding to all the field data, determine the risk coefficient corresponding to each type of the field data, and determine the risk assessment weight corresponding to each type of the field data according to the risk coefficient. In some embodiments, S340 can be performed by the risk assessment weight determination module 240.

[0113] For the field data collected by the rescue helmet 110, if the change patterns of the two field data are relatively similar, then it means that the two field data are closely associated with the rescue status of the firefighters. In this case, the abnormal time series and abnormal degree series corresponding to the two field data will also have a certain degree of similarity. At this time, the higher the reliability of using the two field data to assess fire rescue risks. Based on this, in some embodiments of the present specification, it is considered to determine the risk assessment weights of various types of field data used to assess fire rescue risks in the future by the degree of correlation between different field data.

[0114] According to the above analysis, in some embodiments of the present specification, the risk coefficient corresponding to each type of field data can be determined based on the abnormal parameter set corresponding to all field data, and the risk coefficient can be used to characterize the reliability of each type of field data in evaluating fire rescue risks.

[0115] As an example only, in some embodiments, the risk factor corresponding to each type of field data may be calculated as follows:

[0116]

[0117] Among them, F i represents the risk coefficient corresponding to the i-th field data; p i，k represents the influence coefficient corresponding to the kth abnormal point in the i-th field data; v i，k It represents the discrete degree corresponding to the kth abnormal point in the i-th field data; K represents the total number of abnormal points appearing in the i-th field data as of the current calculation time.

[0118] Specifically, in the embodiment of this specification, the aforementioned discrete degree v i，k It can represent the distribution characteristics of each type of on-site data, which can reflect the dispersion of each data value around the average value or the center value. It can be understood that the greater the degree of dispersion, the greater the difference between the data points, the less obvious the data fluctuation law characteristics, and the lower the reliability of fire rescue risk assessment based on the data. Therefore, it is closely related to the risk coefficient F. iIn some embodiments, the discrete degree v i，k The calculation process can be expressed as follows:

[0119]

[0120] Among them, n i，k is the total number of timestamps in the abnormal time series corresponding to the target field data (such as the aforementioned i-th field data); N k H represents the total amount of data from the initial statistical time point to the target time point corresponding to the target abnormal point (such as the aforementioned kth abnormal point); i，k Indicates the degree of data confusion corresponding to all data in the target field data up to the target abnormal point (i.e., N in the target field data). k The degree of data confusion corresponding to each data); h i，k Indicates the degree of data confusion corresponding to all non-abnormal data up to the target abnormal point (that is, the data after removing the abnormal point). i，k The smaller the value, the more stable the overall distribution of the target site data is, and the higher the reliability of using its fluctuation characteristics to assess fire rescue risks. It should be noted that in this specification, the aforementioned data disorder degree refers to the disorder or uncertainty of the data. In some embodiments, the data disorder degree may refer to the entropy value corresponding to the data. The specific calculation method of the entropy value can be regarded as the prior art, and this specification will not describe it in detail.

[0121] Furthermore, the aforementioned influence coefficient p i，k It can indicate the mutual influence between different on-site data, and it can characterize the synchronization degree of different on-site data with the rescue status of firefighters. The higher the synchronization degree, the closer the correlation with the rescue status of firefighters, and the higher the reliability of fire rescue risk assessment based on this data. Therefore, it is closely related to the risk coefficient F. i In some embodiments, the influence coefficient p i，k The calculation process can be expressed as follows:

[0122]

[0123] Where m represents the total number of types of field data; X i (k), X j (k) are the abnormal time series corresponding to the target abnormal point (i.e., the kth abnormal point) in the i-th field data and the j-th field data respectively; Y i (k), Y j (k) are the abnormality degree sequences corresponding to the target abnormal points in the i-th field data and the j-th field data; DTW[X i (k), Xj (k)] represents the abnormal time series X i (k) and X j (k) Dynamic Time Warping Distance between DTW[(Y i (k), Y j (k) represents the abnormality degree sequence Y i (k), Y j (k) is the dynamic time warping distance between The smaller the p i，k The larger the value of is, the more similar the data change characteristics of the i-th on-site data are to those of other on-site data, the higher the degree of fluctuation synchronization is, and the higher the reliability of assessing fire rescue risks based on the i-th on-site data is.

[0124] In summary, in the embodiment of the present specification, the risk assessment weight determination module 240 can obtain the timestamps and abnormality levels corresponding to all abnormal points in each type of field data, and obtain the corresponding abnormal time series X i (k) and the abnormality degree sequence Y i (k) Then, based on the abnormal time series and the abnormal degree series, the discrete degree v corresponding to each abnormal point in each field data is calculated. i，k and the influence coefficient p i,k Finally, according to the discrete degree v i，k , influence coefficient p i,k And the total number of abnormal points K appearing in the target field data, determine the risk factor F corresponding to each type of target field data i .

[0125] Specifically, in the embodiment of the present specification, the risk assessment weight determination module 240 may determine the total number n of timestamps in the abnormal time series corresponding to the target field data. i，k The total amount of data N between the target time point corresponding to the target field data from the initial statistical time point to the target abnormal point k The first ratio And the first data chaos degree H corresponding to all data in the target field data up to the target abnormal point i , k is the second data confusion degree h corresponding to all non-abnormal data up to the target abnormal point i，k The first difference of |H i，k -h i,k |, and then determine the discreteness corresponding to the target abnormal point in the target field data based on the first ratio and the first difference. For example, in some embodiments, the product of the first ratio and the first difference can be used as the discreteness corresponding to the target abnormal point in the target field data.

[0126] Furthermore, the risk assessment weight determination module 240 can also determine a first dynamic time warping distance (DTW[X]) between the abnormal time series corresponding to the target abnormal point in the target field data and the abnormal time series corresponding to the abnormal points with the same sequence number in other field data (i.e., the abnormal points with the same sequence number in the abnormal time series). i (k), X j (k)]), and the second dynamic time warping distance (DTW[(Y i (k), Y j (k)]), then based on the average of the sum of the first dynamic time warping distance and the second dynamic time warping distance corresponding to the target anomaly point in the target field data and the anomaly points with the same sequence number in all other field data, The influence coefficient corresponding to the target abnormal point in the target field data is obtained. Specifically, in some embodiments, the reciprocal of the average value of the sum of the first dynamic time warping distance and the second dynamic time warping distance corresponding to the target abnormal point in the target field data and the abnormal points with the same sequence number in all other field data can be used as the influence coefficient corresponding to the target abnormal point in the target field data.

[0127] Furthermore, the risk assessment weight determination module 240 can calculate the influence coefficient P corresponding to each abnormal point in the target field data. i，k and the degree of dispersion v i，k The second ratio Then the average value of the second ratio corresponding to all abnormal points in the target field data is calculated As the risk factor corresponding to the target field data.

[0128] It should be noted that in the embodiments of this specification, the risk coefficient takes into account the distribution and change characteristics of different on-site data, and can measure the reliability of each type of on-site data for assessing fire rescue risks in combination with the correlation between different on-site data. In the subsequent process, by calculating the risk assessment weight corresponding to each type of on-site data according to the risk coefficient, the accuracy of the final fire rescue risk assessment results can be improved to a certain extent, avoiding the interference of occasional data fluctuations on the risk assessment results.

[0129] After determining the risk coefficient corresponding to each type of field data through the above steps, the risk assessment weight determination module 240 can obtain the risk coefficients corresponding to all m types of field data, and then determine the risk assessment weight corresponding to each type of field data based on the risk coefficient.

[0130] It can be understood that the larger the risk coefficient, the more reliable such on-site data is in assessing fire rescue risks, and it should account for a larger proportion in the subsequent analysis process, so that the final fire rescue risk assessment results can more accurately express the current fire rescue status.

[0131] Specifically, in some embodiments, the risk assessment weight corresponding to each type of field data can be expressed as follows:

[0132]

[0133] Among them, w i represents the risk assessment weight corresponding to the i-th field data; F i is the risk coefficient corresponding to the i-th type of field data; m is the total number of types of all field data.

[0134] In other words, in the embodiment of the present specification, the risk assessment weight determination module 240 can calculate the risk coefficient F corresponding to the target field data. i The sum of the risk factors corresponding to all field data The third ratio The third ratio is then used as the risk assessment weight corresponding to the target field data.

[0135] In some embodiments, an initial risk assessment weight may be configured for each type of field data. When no abnormal points are detected in the field data and the risk coefficient and risk assessment weight cannot be determined through the abnormal parameter set corresponding to the field data, subsequent calculations may be performed based on the initial assessment weight.

[0136] S350, based on the fluctuation characteristics of the field data, determine the data capture range corresponding to each type of the field data, and determine the target feature data corresponding to each type of the field data at each moment according to the data capture range. In some embodiments, S350 can be performed by the target feature data determination module 250.

[0137] In the embodiments of this specification, the purpose of determining the data capture range corresponding to each type of field data by the fluctuation characteristics of the field data is to obtain the target data that can accurately express the current fire rescue status in the corresponding time series data by analyzing the fluctuation characteristics of each type of field data. However, not all the time series data corresponding to various field data are of reference significance. Therefore, it is necessary to determine the corresponding data capture range according to the distribution of each type of field data, so as to select the key data in each type of field data for subsequent analysis.

[0138] Specifically, in some embodiments, for the target moment corresponding to the target field data, the target characteristic data determination module 250 can calculate the fluctuation law index corresponding to the target moment and the fluctuation influence index corresponding to the target field data, and then determine the data capture range corresponding to the target moment in the target field data based on the fluctuation law index, the fluctuation influence index and the oscillation maintenance time in the time period before the target moment in the target field data. The target moment can be understood as the current calculation moment or the latest moment corresponding to the target field data.

[0139] In the embodiments of this specification, the fluctuation regularity index corresponding to the target moment is an indicator for measuring the regularity of data fluctuation and the degree of disorder of data distribution in the time period before the target moment, which can characterize the regularity of data fluctuation and the degree of disorder of data distribution before the target moment. The fluctuation influence index corresponding to the target field data can be understood as an indicator for measuring the degree of influence of fluctuations in a local time period on the data fluctuation regularity in the target field data.

[0140] Specifically, in some embodiments, for the i-th type of field data, the total number K of abnormal points up to the current moment (including the current moment) can be obtained, and the oscillation maintenance time F of consecutive abnormal points can be counted. The oscillation maintenance time can be understood as the data points corresponding to consecutive F moments are all abnormal points. For example, for field data such as obstacle detection data, t4, t5, t6, t7, t8, t9, t10, t11, t12, t13, t14, t15, t16, t17, t18, t19, t20, t21, t22, t23, t24, t25, t26, t27, t28, t29, t30, t31, t32, t33, t34, t35, t36, t37 14 ,t 20 The data detected at all times are abnormal points, so t4 corresponds to the oscillation maintenance time of the abnormal point. The oscillation maintenance time of the abnormal point corresponding to t5 is 3 consecutive moments from t5 to t7. For the field data of non-abnormal points, the steady-state maintenance time f at each moment can be counted. Assuming that the obstacle detection data from t0 to t3 is a non-abnormal point, the steady-state maintenance time of the non-abnormal point corresponding to t3 is The oscillation maintenance time corresponding to the abnormal points in the remaining field data and the steady-state maintenance time corresponding to the non-abnormal points can be determined based on a similar method, which will not be described in detail in this specification.

[0141] In some embodiments, the data capture range used to characterize the current data characteristics of each type of field data can be determined according to the oscillation maintenance time and steady-state maintenance time of each type of field data at each moment. Specifically, in some embodiments, the actual values ​​of the oscillation maintenance time and the steady-state maintenance time can be counted respectively, and then the same oscillation maintenance time can be divided into the same fluctuation level, and the fluctuation level can be recorded from small to large as B0 to B1. maxSimilarly, the same steady-state maintenance time can be divided into the same steady-state level, and then the steady-state levels are recorded from small to large as W0 to W max .

[0142] Based on the above analysis, in some embodiments of this specification, the corresponding data capture range can be determined according to the oscillation maintenance time and steady-state maintenance time of each type of field data at each moment, which is used to characterize the size of the data area when extracting data features for each type of field data at each moment.

[0143] In some embodiments, the data capture range ZQ corresponding to each type of field data at the target time (eg, the ath time) is a The calculation process can be expressed as follows:

[0144]

[0145] Among them, ZQ a represents the data capture range corresponding to the target field data (such as the aforementioned i-th field data) at the a-th time; F min is the minimum oscillation maintenance time of the target field data in the previous a moments; n1 is the number of local statistical time, which refers to the quotient of the total duration of the previous a moments and the local statistical time g, where the size of the local statistical time g is the mean value of the steady-state maintenance time f in the previous a moments; B g Indicates the fluctuation impact index corresponding to the target field data; G ia It represents the fluctuation regularity index corresponding to the ath moment in the i-th field data (i.e., the target field data).

[0146] In some embodiments, the aforementioned fluctuation regularity index G ia The calculation process can be expressed as follows:

[0147]

[0148] Among them, E W and E B They respectively represent the data disorder degree of the fluctuation level (e.g., entropy value) and the data disorder degree of the steady-state level in the i-th field data within the previous a moments; It represents the mean of all steady-state levels contained in the previous a moments. It can be understood that G ia The larger the value of is, the more regular the fluctuation of the data corresponding to the previous a moments in the field data is.

[0149] The aforementioned volatility impact index B g The calculation process can be expressed as follows:

[0150]

[0151] in, It is the average value of the oscillation maintenance time F of the abnormal point within the local statistical time; is the average value of all steady-state maintenance time within the local statistical time; (C q，q+1 ) is the adjacent moment in the local statistical time, and its size is equal to the difference between the durations of two adjacent moments. For example, if the qth moment is an abnormal point and the q+1th moment is a non-abnormal point, then the duration corresponding to the qth moment is the oscillation maintenance time F q , Steady state maintenance time f q+1 , at this time (C q，q+1 ) is |F q -f q+1 |;σ2(C q，q+1 ) represents the distribution variance of adjacent moments; is a fixed parameter used to prevent the denominator from being 0 (in this specification can be 0.01).

[0152] In summary, in the embodiment of the present specification, for the ath moment of the i-th type of field data, the target feature data determination module 250 can calculate the fluctuation regularity index G corresponding to the ath moment ia , and determine the fluctuation impact index B corresponding to the i-th field data g , then, based on the fluctuation law index G ia , Volatility Impact Index B g and the minimum value E of the oscillation maintenance time in the time period before the ath moment (i.e., the target moment) in the i-th field data min , determine the data capture range ZQ corresponding to the ath moment in the i-th field data d .

[0153] Specifically, in calculating the aforementioned fluctuation law index G ia , the target feature data determination module 250 may obtain the third data disorder degree E of the fluctuation level corresponding to the data in the time period before the target moment (for example, the ath moment) in the target field data (for example, the i-th field data mentioned above). W and the fourth data disorder degree E of the steady-state level corresponding to the data in the time period before the target time in the target field data B , and based on the second difference between the third data disorder degree and the fourth data disorder degree (|E W -E B |), and the average steady-state level of all data in the target field data in the time period before the target time Determine the fluctuation law index G corresponding to the target moment ia .

[0154] Furthermore, in calculating the aforementioned volatility impact index B g When the target characteristic data determination module 250 can obtain the average value of the oscillation maintenance time of the abnormal point within the local statistical time The average value of all steady-state maintenance times within the local statistical time And the distribution variance of adjacent moments in the local statistical time σ2(C q，q+1 ), then, based on the distribution variance of adjacent moments and the third difference between the average value of the oscillation maintenance time and the average value of all steady-state maintenance times Determine the fluctuation impact index B corresponding to the target field data g Specifically, in some embodiments, the target feature data determination module 250 may determine the target feature data based on the distribution variance σ2(C q，q+1 ) and the third difference The ratio of determines the fluctuation impact index B corresponding to the target field data g .

[0155] After the above steps are performed to determine the data capture range ZQ corresponding to each type of field data at each moment, a Afterwards, the target feature data determination module 250 can capture the range ZQ according to the data. a Determine the data capture quantity S in each type of field data. In some embodiments, the data capture range ZQ can be mapped by a preconfigured mapping function. a Mapping is performed to obtain the data capture quantity S (S is an integer) in each type of field data. In some embodiments, the data capture range ZQ a Round up to the nearest integer to get the corresponding data capture quantity S.

[0156] Furthermore, for the target moment, the target feature data determination module 250 may use the data collected at S consecutive moments before the target moment in each type of field data as the target data, and calculate the average value, variance and entropy value of the target data as the target feature data corresponding to each type of field data at the target moment. It should be noted that in the embodiment of the present specification, the target feature data may cover the average value, variance and entropy value of multiple data within the aforementioned data capture range, which may characterize the information characteristics reflected by one or more data corresponding to each type of field data at the current moment, thereby being able to more accurately reflect the relationship between it and the current fire rescue status. In the subsequent process, by using the target feature data to evaluate the fire rescue risk, a more accurate and reliable fire rescue risk assessment result may be obtained.

[0157] S360, obtaining a feature vector corresponding to each type of the field data at each moment according to the target feature data and the risk assessment weight. In some embodiments, S360 may be executed by the feature vector generation module 260.

[0158] In some embodiments, the feature vector generation module 26 can obtain the risk assessment weight corresponding to the target feature data, and the target feature data corresponding to the target field data at the target moment, and then obtain the feature vector corresponding to the target moment based on the risk assessment weight and the target feature data.

[0159] For example, if the risk assessment weight corresponding to the i-th field data is w i , the mean, variance and entropy of the target feature data corresponding to the ath moment are e i1 、e i2 、e i3 , then the feature vector corresponding to the i-th field data at the a-th time can be expressed as T ai =(w i , e i1 , e i2 , e i3 ).

[0160] It can be understood that by processing each type of on-site data acquired by the rescue helmet 110 through the above steps, the feature vector corresponding to each type of on-site data at each moment can be obtained.

[0161] S370, inputting the feature vectors corresponding to all the on-site data into the rescue risk assessment model for processing to obtain a fire rescue risk assessment result. In some embodiments, S370 can be executed by the risk assessment result determination module 270.

[0162] After obtaining the feature vector corresponding to each type of on-site data collected by the rescue helmet 110 at each moment through the above steps, the feature vector corresponding to all on-site data at the target moment can be input into the rescue risk assessment model for processing, and the corresponding fire rescue risk assessment result can be obtained through the rescue risk assessment model.

[0163] In some embodiments, the feature vectors corresponding to all on-site data can be fused (for example, serial fusion, parallel fusion or Attention feature fusion) to obtain a fusion vector corresponding to each moment, which can contain the feature information of the fire scene temperature data, firefighter motion state data, firefighter physiological data, gas concentration data and obstacle detection data at each moment. The specific implementation process of feature vector fusion can be regarded as prior art and will not be discussed in detail in this specification. In some embodiments, the feature vector fusion process can be performed by a sub-model in the above-mentioned rescue risk assessment model.

[0164] After obtaining the fusion vector, the rescue risk assessment model can process the fusion vector to obtain the fire rescue risk assessment result corresponding to the current moment. Specifically, in some embodiments, the fire rescue risk assessment result can be reflected as a fire rescue risk index or level. When the fire rescue risk index or level is greater than a preset threshold (for example, when the fire rescue risk index ranges from 0 to 1, the preset threshold can be set to 0.7), it means that the firefighter is in a dangerous state and needs to pay more attention or provide assistance.

[0165] It should be noted that the above rescue risk assessment model can be constructed based on historical data and existing statistical models (such as Logistic regression or Cox regression), and then trained and optimized through machine learning algorithms (such as random forest, support vector machine or AdaBoost, etc.). As an example only, in some embodiments, several fusion vectors can be obtained as training samples, and the label corresponding to each fusion vector is determined according to the actual risk state of the original on-site data corresponding to each fusion vector (the label can be a floating point number between 0.0 and 1.0, used to characterize its corresponding fire rescue risk index), and then the training sample and its corresponding label are input into the initial model for training, until the loss function reaches a preset value (the preset value can be set according to actual needs) or the number of iterations reaches a preset value (for example, 2000 times) to stop training, and obtain a trained rescue risk assessment model.

[0166] In some embodiments, the aforementioned loss function can be constructed using mean square error loss or cross entropy loss. Then, the output value of the loss function is gradually reduced by a gradient descent algorithm, thereby optimizing the rescue risk assessment model. The training process of the machine learning model can be regarded as a known technology, and therefore, it will not be discussed in detail in this specification.

[0167] It should be pointed out that in the embodiments of the present specification, the fusion vector is obtained based on the fusion of feature vectors corresponding to all the on-site data, and the feature vector corresponding to each type of on-site data is obtained based on its corresponding adaptive risk assessment weight and data capture range. Since the data characteristics of the data itself and the correlation characteristics with other on-site data are taken into account in determining the risk assessment weight and data capture range corresponding to each type of on-site data, the fusion vector obtained by the method provided in the embodiments of the present specification can more accurately reflect the relationship between it and the fire rescue status, thereby improving the accuracy and reliability of the fire rescue risk assessment results to a certain extent.

[0168] In summary, the beneficial effects that may be brought about by the embodiments of the present specification include but are not limited to: (1) In the firefighting and rescue method and system based on the rescue helmet provided in some embodiments of the present specification, by considering the distribution and change law characteristics of different field data, and then combining the correlation between different field data to determine the corresponding risk coefficient to measure the reliability of each field data for assessing the firefighting and rescue risk, and by calculating the risk assessment weight corresponding to each field data according to the risk coefficient, and finally determining the feature vector corresponding to each field data according to the risk assessment weight and determining the firefighting and rescue risk assessment result based on the feature vector, the accuracy of the final firefighting and rescue risk assessment result can be improved to a certain extent, and the interference of occasional data fluctuations on the risk assessment result can be avoided; (2) In the firefighting and rescue method and system based on the rescue helmet provided in some embodiments of the present specification, by considering the own fluctuation characteristics of each field data to determine its corresponding data capture range, and determining the target feature data corresponding to each field data at each moment according to the data capture range, redundant data or irrelevant data can be filtered out from each field data, thereby reducing the probability of its use for subsequent processing, reducing the subsequent calculation load to a certain extent, and avoiding its interference with the final firefighting and rescue risk assessment result.

[0169] It should be noted that different embodiments may produce different beneficial effects. In different embodiments, the beneficial effects that may be produced may be any one or a combination of the above, or may be any other beneficial effects that may be obtained.

[0170] The basic concepts have been described above. Obviously, for those skilled in the art, the above detailed disclosure is only for example and does not constitute a limitation of this specification. Although not explicitly stated here, those skilled in the art may make various modifications, improvements and corrections to this specification. Such modifications, improvements and corrections are suggested in this specification, so such modifications, improvements and corrections still belong to the spirit and scope of the exemplary embodiments of this specification.

Claims

1. A firefighting and rescue method based on a rescue helmet, characterized in that: include: Acquire on-site data collected by the rescue helmet, wherein the on-site data includes fire scene temperature data, firefighter movement status data, firefighter physiological data, gas concentration data, and obstacle detection data; Performing anomaly detection on the field data to obtain anomaly points in each type of the field data; Generate a set of abnormal parameters corresponding to each type of the field data according to the abnormal points, wherein the abnormal parameters include a timestamp and an abnormal degree corresponding to each abnormal point; Determine a risk coefficient corresponding to each type of the field data based on a set of abnormal parameters corresponding to all the field data, and determine a risk assessment weight corresponding to each type of the field data according to the risk coefficient; Determine a data capture range corresponding to each type of the field data based on the fluctuation characteristics of the field data, and determine target feature data corresponding to each type of the field data at each moment according to the data capture range; According to the target feature data and the risk assessment weight, a feature vector corresponding to each type of the field data at each moment is obtained; The characteristic vectors corresponding to all the on-site data are input into the rescue risk assessment model for processing to obtain a fire rescue risk assessment result.

2. The method according to claim 1, characterized in that Performing anomaly detection on the field data to obtain anomaly points in each type of the field data includes: Calculate the difference between all adjacent data in the time series data corresponding to each type of field data to obtain a continuous difference sequence; Based on the standard score detection, a mutation point in the continuous differential sequence is determined, and an abnormal point in the corresponding field data is determined according to the mutation point.

3. The method according to claim 1 or 2, characterized in that The determining the risk coefficient corresponding to each type of the field data based on the abnormal parameter set corresponding to all the field data includes: Obtain the timestamps and abnormality levels corresponding to all abnormal points in each type of field data, and obtain the corresponding abnormal time series and abnormality level series; Calculate the degree of dispersion and the influence coefficient corresponding to each abnormal point in each type of field data based on the abnormal time series and the abnormal degree series; The risk coefficient corresponding to each type of the field data is determined according to the degree of dispersion, the influence coefficient and the total number of abnormal points appearing in the field data.

4. The method according to claim 3, characterized in that The calculating the discrete degree and influence coefficient corresponding to each abnormal point in each type of field data based on the abnormal time series and the abnormal degree series includes: Determine a first ratio of the total number of timestamps in the abnormal time series corresponding to the target field data to the total amount of data of the target field data from the initial statistical time point to the target time point corresponding to the target abnormal point, and a first difference between a first data disorder degree corresponding to all data in the target field data up to the target abnormal point and a second data disorder degree corresponding to all non-abnormal data up to the target abnormal point, and then determine the discrete degree corresponding to the target abnormal point in the target field data based on the first ratio and the first difference; Determine a first dynamic time warping distance between an abnormal time series corresponding to a target abnormal point in the target field data and an abnormal time series corresponding to abnormal points with the same sequence number in other field data, and a second dynamic time warping distance between an abnormal degree series corresponding to the target abnormal point in the target field data and an abnormal degree series corresponding to abnormal points with the same sequence number in other field data, and then obtain an influence coefficient corresponding to the target abnormal point in the target field data based on an average value of the sum of the first dynamic time warping distance and the second dynamic time warping distance corresponding to the target abnormal point in the target field data and abnormal points with the same sequence number in all other field data.

5. The method according to claim 3, characterized in that Determining the risk coefficient corresponding to each type of the field data according to the degree of dispersion, the influence coefficient and the total number of abnormal points appearing in the field data includes: Calculate a second ratio of the influence coefficient and the degree of dispersion corresponding to each abnormal point in the target field data; An average value of the second ratios corresponding to all abnormal points in the target field data is used as the risk coefficient corresponding to the target field data.

6. The method according to claim 1, characterized in that Determining the risk assessment weight corresponding to each type of the on-site data according to the risk coefficient includes: A third ratio of the risk coefficient corresponding to the target field data to the sum of the risk coefficients corresponding to all field data is calculated, and the third ratio is used as the risk assessment weight corresponding to the target field data.

7. The method according to claim 1, characterized in that The determining of the data capture range corresponding to each type of the field data based on the fluctuation characteristics of the field data includes: For a target moment corresponding to the target field data, calculating a fluctuation regularity index corresponding to the target moment and a fluctuation influence index corresponding to the target field data; Based on the fluctuation regularity index, the fluctuation influence index and the oscillation maintenance time in the target field data within a time period before the target time, a data capture range corresponding to the target time in the target field data is determined.

8. The method according to claim 7, characterized in that The calculating of the fluctuation regularity index corresponding to the target time and the fluctuation influence index corresponding to the target field data includes: Obtaining a third data disorder degree of a fluctuation level corresponding to data in the target field data in a time period before the target moment and a fourth data disorder degree of a steady-state level corresponding to data in the target field data in a time period before the target moment, and determining a fluctuation regularity index corresponding to the target moment based on a second difference between the third data disorder degree and the fourth data disorder degree and an average value of steady-state levels corresponding to all data in the target field data in a time period before the target moment; The average value of the oscillation maintenance time of the abnormal points in the local statistical time in the target field data, the average value of the steady-state maintenance time in the local statistical time, and the distribution variance corresponding to adjacent moments in the local statistical time are obtained, and the fluctuation impact index corresponding to the target field data is determined based on the distribution variance corresponding to the adjacent moments and the third difference between the average value of the oscillation maintenance time and the average value of the steady-state maintenance time.

9. The method according to claim 8, characterized in that The step of determining the target feature data corresponding to each type of the field data at each moment according to the data capture range includes: Determine the data capture quantity S in each type of the field data according to the data capture range; For the target moment, the data collected at S consecutive moments before the target moment in each type of field data is taken as the target data, and the mean value, variance and entropy value of the target data are calculated as the target feature data corresponding to each type of field data at the target moment.

10. A fire rescue system based on a rescue helmet, characterized in that: include: An acquisition module is used to acquire on-site data collected by a rescue helmet, wherein the on-site data includes fire scene temperature data, firefighter movement status data, firefighter physiological data, gas concentration data, and obstacle detection data; An anomaly detection module is used to perform anomaly detection on the field data and obtain anomalies in each field data. An abnormal parameter set generation module, used to generate an abnormal parameter set corresponding to each type of field data according to the abnormal point, wherein the abnormal parameters include a timestamp and an abnormal degree corresponding to each abnormal point; A risk assessment weight determination module is used to determine the risk coefficient corresponding to each type of field data based on the abnormal parameter set corresponding to all the field data, and determine the risk assessment weight corresponding to each type of field data according to the risk coefficient. A target characteristic data determination module, used to determine a data capture range corresponding to each type of the field data based on the fluctuation characteristics of the field data, and determine the target characteristic data corresponding to each type of the field data at each moment according to the data capture range; A feature vector generation module, used for obtaining a feature vector corresponding to each type of the field data at each moment according to the target feature data and the risk assessment weight; The risk assessment result determination module is used to input the characteristic vectors corresponding to all the on-site data into the rescue risk assessment model for processing to obtain the fire rescue risk assessment result.

Images