Integrated fire alarm prediction method and system based on artificial intelligence

Abstract

The invention discloses an artificial intelligence-based integrated fire alarm prediction method and system, and relates to the technical field of fire alarm, and the method comprises the steps: determining a fire management range, and constructing a management model based on the fire management range; collecting a sensitive target based on the fire management range, formulating an ecological circle corresponding to the sensitive target, constructing the ecological circle of the sensitive target into a management model, and tracking the sensitive target; fire information is collected corresponding to the ecosphere of the sensitive target; prediction information is calculated according to the fire information, the prediction information comprises a prediction index and a corresponding alarm scheme, and alarm information is generated according to the prediction information and sent to an alarm platform. According to the invention, prediction and positioning of a fire in a small range can be carried out, a more serious consequence caused by the fact that the fire is found after developing to a certain degree is avoided, a more accurate fire geographic position can be provided during alarm, and more accurate fire alarm is completed.

Description

Technical Field

[0001] The present invention relates to the field of fire alarm technology, and in particular to an artificial intelligence-based integrated fire alarm prediction method and system. Background Art

[0002] Fires are sudden and cause huge casualties, property losses and social impacts around the world every year. It not only threatens people's lives and safety, but may also lead to ecological environmental damage and social and economic stagnation. Therefore, predicting the risk of fire and taking effective prevention and response measures in a timely manner are of great significance to protecting people's lives and property, maintaining social stability and sustainable development. However, the current fire warning and prediction system is difficult to quickly and accurately determine the small-scale location of the fire. Therefore, it is difficult to achieve a rapid warning response when a fire breaks out. Generally, the fire can only be discovered when it spreads to a certain extent. Therefore, there is a lack of a fire alarm prediction method and system that can predict and locate small-scale fires and provide a more accurate fire geographical location when an alarm is triggered. Summary of the invention

[0003] The purpose of the present invention is to provide a fire alarm prediction method and system based on artificial intelligence integration to solve the shortcomings of the background technology.

[0004] In order to achieve the above object, the present invention provides the following technical solution: a fire alarm prediction method based on artificial intelligence integration, comprising the following steps:

[0005] Determine the fire management scope and build a management model based on the fire management scope;

[0006] Collect sensitive targets based on the fire management scope, formulate ecosystems for the sensitive targets, build the ecosystems of sensitive targets into the management model, and track the sensitive targets;

[0007] Collect fire information corresponding to the ecosystem of sensitive targets;

[0008] Prediction information is calculated based on the fire information, wherein the prediction information includes a prediction index and a corresponding alarm scheme, and alarm information is generated based on the prediction information and sent to the alarm platform.

[0009] In a preferred embodiment, the step of determining the fire management scope and constructing a management model based on the fire management scope includes:

[0010] Determine the fire management scope and collect characteristic information within the fire management scope, wherein the characteristic information includes geographic environment information, infrastructure information and dynamic information;

[0011] A three-dimensional model of the fire management range is constructed according to the characteristic information, and a palace layer is set corresponding to the three-dimensional model;

[0012] The management model is obtained by matching the Tiangong layer with the three-dimensional model.

[0013] In a preferred embodiment, the step of constructing a three-dimensional model of the fire management range according to the characteristic information and setting a palace layer corresponding to the three-dimensional model includes:

[0014] Three-dimensional modeling of features within the fire management range is performed according to the feature information to obtain a feature model, and the feature models are arranged according to positional relationships within the fire management range to obtain a three-dimensional model of the fire management range;

[0015] An upper control surface is set corresponding to the three-dimensional model, and a corresponding cloud server is configured corresponding to the upper control surface;

[0016] The upper reference surface is matched with the geographical surface of the fire management range, and a viscous relationship between the upper reference surface and the geographical surface is established, wherein the viscous relationship is a comprehensive upper and lower position correspondence between the upper reference surface and the geographical surface;

[0017] A plurality of landmarks are set corresponding to the upper control surface, the plurality of landmarks are communicated and connected with each other, and the plurality of landmarks are connected with the upper control surface;

[0018] The upper control surface with multiple landmarks and the upper control surface that establishes a viscous relationship with the geographical surface are regarded as the Tiangong layer.

[0019] In a preferred embodiment, the step of obtaining a management model by matching the Tiangong layer with the three-dimensional model includes:

[0020] Configure the corresponding external source space on the cloud server corresponding to the Tiangong layer, and connect the external source space to the cloud server;

[0021] Mark the storage points in the external space and establish the corresponding relationship between the storage points and the Tiangong layer in the cloud server;

[0022] The three-dimensional model is put into the external space. According to the viscous relationship between the geographical surface of the three-dimensional model and the Tiangong layer and the corresponding relationship between the storage point and the Tiangong layer in the cloud server, the geographical surface of the three-dimensional model is stored in the external space corresponding to the storage point, and the correspondence between the Tiangong layer and the three-dimensional model is completed to obtain the management model.

[0023] In a preferred embodiment, the steps of collecting sensitive targets based on the fire management scope, formulating ecosystems corresponding to the sensitive targets, building the ecosystems of the sensitive targets into the management model, and tracking the sensitive targets include:

[0024] Setting a sensitive target database, wherein the sensitive target database includes sensitive targets, information on triggering conditions of fires on sensitive targets, and danger levels of sensitive targets combined with the triggering condition information;

[0025] Based on the sensitive target database, multiple targets within the fire management range are matched respectively to obtain multiple sensitive targets;

[0026] According to the fire inducing conditions of sensitive targets, the safety range of sensitive targets is determined as the ecosystem of sensitive targets, and the ecosystem of sensitive targets is constructed into a three-dimensional model. The ecosystem always takes the corresponding sensitive targets as the center and follows the sensitive targets.

[0027] According to the multiple landmarks in the Tiangong layer, multiple sensitive targets are matched and bound one by one. The landmarks are always connected to the upper-level control surface, and the corresponding sensitive targets are tracked through the landmarks.

[0028] In a preferred embodiment, the step of collecting fire information from the ecosystem corresponding to the sensitive target includes:

[0029] Obtain environmental information in the sensitive target ecosystem, where environmental information includes target objects, basic environmental information, and behavioral information;

[0030] Matching the fire-inducing condition information of the corresponding sensitive target and the danger level of the sensitive target combined with the fire-inducing condition information in the sensitive target database according to the environmental information;

[0031] Obtaining the geographic location of sensitive targets based on landmarks;

[0032] The sensitive target, the geographical location of the sensitive target, the fire inducing condition information of the corresponding sensitive target matched in the sensitive target database, and the danger level of the sensitive target combined with the fire inducing condition information are used as fire information.

[0033] In a preferred embodiment, the step of generating alarm information according to the prediction information and sending it to the alarm platform includes:

[0034] The prediction index is calculated based on the fire information, where the calculation formula of the prediction index is: Among them, W is the prediction index, μ is the risk assessment coefficient of the sensitive target, i is the number of the induced condition information, k is the total number of induced condition information, β i is the danger level of the i-th inducing condition information, L i is the distance between the triggering condition information of the i-th sensitive target fire and the sensitive target in the ecosystem;

[0035] Formulate alarm rules, wherein the alarm rules include alarm schemes corresponding to the numerical range of the prediction index, determine the numerical range in the alarm rule according to the prediction index, match the corresponding alarm scheme, and use the prediction index and the corresponding alarm scheme as prediction information;

[0036] The prediction index is combined with the fire information and sent to the alarm platform through the alarm plan for alarm.

[0037] The present invention also provides an integrated fire alarm prediction system based on artificial intelligence, comprising:

[0038] A construction module is used to determine the fire management scope and build a management model based on the fire management scope;

[0039] The formulation module is connected with the construction module and is used to collect sensitive targets based on the fire management scope, formulate an ecosystem corresponding to the sensitive targets, build the ecosystem of the sensitive targets into the management model, and track the sensitive targets;

[0040] The collection module is used to formulate module connections and collect fire information corresponding to the ecosystem of sensitive targets;

[0041] The alarm module is connected to the acquisition module and is used to calculate prediction information based on the fire information, wherein the prediction information includes a prediction index and a corresponding alarm scheme, and generates alarm information based on the prediction information and sends it to the alarm platform.

[0042] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0043] The upper comparison surface of the present invention is a virtual surface that is set in advance and accurately corresponds to the geographical surface position of the fire management scope. There is no need to perform independent position calculation or acquisition based on the geographical surface of the fire management scope. Accurate position tracking can be performed directly based on the upper comparison surface set in advance. The geographical location is acquired efficiently and accurately, and small-scale fires can be predicted and located to avoid more serious consequences caused by fires that are not discovered until they develop to a certain extent. A more accurate fire geographical location can be provided when an alarm is triggered, thereby completing a more accurate fire alarm. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments are briefly introduced below.

[0045] Figure 1 The figure is a flow chart of the method of the present invention.

[0046] Figure 2 It is a system block diagram of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] Example 1, please refer to Figure 1 As shown, the fire alarm prediction method based on artificial intelligence integration described in this embodiment includes the following steps:

[0049] S1. Determine the fire management scope and build a management model based on the fire management scope;

[0050] S2. Collect sensitive targets based on the fire management scope, formulate ecosystems corresponding to sensitive targets, build the ecosystems of sensitive targets into the management model, and track sensitive targets;

[0051] S3, collect fire information corresponding to the ecosystem of sensitive targets;

[0052] S4. Calculate prediction information based on the fire information, wherein the prediction information includes a prediction index and a corresponding alarm scheme, and generate alarm information based on the prediction information and send it to the alarm platform;

[0053] As described in the above steps S1-S4, fire is a sudden, extremely high-risk social and economic problem, which causes huge casualties, property losses and social impacts around the world every year. It not only threatens people's lives and safety, but also may lead to ecological environmental damage and social and economic stagnation. Therefore, predicting the risk of fire and taking effective prevention and response measures in time are of great significance for protecting people's lives and property, maintaining social stability and sustainable development. However, the current early warning and prediction system is difficult to quickly and accurately determine the small-scale location of the fire. Therefore, it is difficult to achieve a rapid early warning response during a fire outbreak. Generally, the fire can only be discovered when it spreads to a certain extent. Once a sudden fire occurs, it will cause serious damage to people and property. There is no good fire alarm prediction ability. The upper-level control surface of the present application is a virtual surface that is set in advance and accurately corresponds to the geographical surface position of the fire management range. There is no need to perform independent position calculation or acquisition based on the geographical surface of the fire management range. Accurate position tracking can be performed directly based on the upper-level control surface set in advance. The acquisition of geographical location is efficient and accurate, and small-scale fires can be predicted and located to avoid more serious consequences caused by the fire developing to a certain extent before being discovered. It can provide a more accurate fire geographical location when an alarm is issued, and complete a more accurate fire alarm.

[0054] In one embodiment, the step S1 of determining the fire management scope and building a management model based on the fire management scope includes:

[0055] S11, determining the fire management scope, and collecting characteristic information within the fire management scope, wherein the characteristic information includes geographical environment information, infrastructure information, and dynamic information;

[0056] S12, constructing a three-dimensional model of the fire management range according to the characteristic information, and setting a palace layer corresponding to the three-dimensional model;

[0057] S13, matching the Tiangong layer with the three-dimensional model to obtain a management model;

[0058] As described in the above steps S11-S13, the fire management scope is determined. The fire management scope here is the fire alarm prediction scope covered by this system. Then, for subsequent fire alarm prediction, it is necessary to collect geographical environment information and infrastructure information within the fire management scope as feature information. The geographical environment information is the geographical environment characteristics within the fire management scope, such as rivers, trees and other information. The infrastructure information is the characteristics of buildings, roads and man-made facilities. The dynamic information includes the mobile characteristics of moving vehicles and people. Then, a three-dimensional model within the fire management scope is constructed according to the feature information. The three-dimensional model monitors the characteristics within the fire management scope in real time and updates it in real time. Then, a palace layer is set corresponding to the three-dimensional model. The palace layer here is for fire alarm prediction within the fire management scope. Then, the palace layer is matched with the three-dimensional model to obtain a management model, which can better perform fire alarm prediction and fire alarm.

[0059] In one embodiment, the step S12 of constructing a three-dimensional model of the fire management range according to the characteristic information and setting a palace layer corresponding to the three-dimensional model includes:

[0060] S121, performing three-dimensional modeling on the features within the fire management range according to the feature information to obtain a feature model, and arranging the feature models according to their positional relationships within the fire management range to obtain a three-dimensional model within the fire management range;

[0061] S122, setting an upper comparison surface corresponding to the three-dimensional model, and configuring a corresponding cloud server corresponding to the upper comparison surface;

[0062] S123, making a vertical correspondence between the upper-layer reference surface and the geographical surface of the fire management range, and establishing a viscous relationship between the upper-layer reference surface and the geographical surface, wherein the viscous relationship is a comprehensive vertical correspondence between the upper-layer reference surface and the geographical surface;

[0063] S124, a plurality of landmarks are set corresponding to the upper-layer control surface, the plurality of landmarks are communicatively connected to each other, and the plurality of landmarks are connected to the upper-layer control surface;

[0064] S125, setting the upper comparison surface with multiple landmarks and the upper comparison surface with a viscous relationship with the geographic surface as the Tiangong layer;

[0065] As described in the above steps S121-S125, the characteristic information collected within the fire management scope is modeled. For example, rivers, artificial facilities and various dynamic information are three-dimensionally modeled to obtain a characteristic model. The characteristic model exists within the fire management scope. Therefore, the characteristic model is arranged in the fire management scope according to the position of the collected characteristic information, thereby obtaining a three-dimensional model. The fire management scope represents the scope involved in a geographical surface. The three-dimensional model is a model for displaying three-dimensional characteristic information established on the plane of the fire management scope. Then, an upper control surface is set corresponding to the three-dimensional model, and the three-dimensional model is displayed on a computer. The upper control surface set here is located at the upper layer of the three-dimensional model in terms of the physical relationship of the display. The upper control surface is corresponded to the geographical surface of the fire management scope from top to bottom, and a viscous relationship between the upper control surface and the geographical surface is established, which means that any position on the upper control surface can correspond to the geographical surface position. Then, the position of the subsequent landmark corresponding to the upper control surface can directly represent the position on the geographical surface, and has a complete and accurate position correspondence, which is convenient for the subsequent alarm of the fire location predicted by the fire alarm to the alarm platform, and has better subsequent rescue work and subsequent evacuation work. Multiple geographical surfaces are set on the upper control surface. The communication connection between multiple landmarks is used for subsequent correction. The correction is through the two farthest edge points on the upper comparison surface. Here it is mainly used as a reference. The specific facility is the two farthest edge points on the upper comparison surface or the two closest edge points (recording the distance between the two edge points) or directly setting two fixed points on the upper comparison surface to determine the position and distance of the two fixed points. The multiple landmarks are used for subsequent tracking of sensitive targets. There is a one-to-one correspondence between sensitive targets and landmarks. For example, there are ten landmarks (the number of landmarks in actual use is several times more, which can be used for enough sensitive targets) and five sensitive targets. Then, five landmarks are required to be one-to-one according to the five sensitive targets respectively, and then the fire alarm prediction is carried out for the sensitive targets according to the subsequent ecosystem of the sensitive targets. Then, the corresponding alarm platform is alarmed, which can accurately alarm the fire location according to the Tiangong layer and facilitate the subsequent evacuation and alarm prompt work. Here, five landmarks are required to track the five sensitive targets, and the remaining five landmarks are on standby. When additional sensitive targets appear in the three-dimensional model, the sensitive targets will be directly bound and tracked by the landmarks until the sensitive targets leave the three-dimensional model, that is, the sensitive targets leave the fire management scope.

[0066] In one embodiment, the step S13 of obtaining the management model by matching the Tiangong layer with the three-dimensional model includes:

[0067] S131, configuring the cloud server corresponding to the Tiangong layer with the corresponding external source space, and connecting the external source space to the cloud server;

[0068] S132, marking the storage point in the external source space, and establishing a corresponding relationship between the storage point and the Tiangong layer in the cloud server;

[0069] S133, putting the three-dimensional model into the external source space, according to the viscosity relationship between the geographical surface of the three-dimensional model and the Tiangong layer and the corresponding relationship between the storage point and the Tiangong layer in the cloud server, the geographical surface of the three-dimensional model is stored in the external source space corresponding to the storage point, and the correspondence between the Tiangong layer and the three-dimensional model is completed to obtain the management model;

[0070] As described in the above steps S131-S133, the cloud server corresponding to the Tiangong layer is configured with a corresponding external space, and the external space is communicatively connected with the cloud server. Since the storage of the Tiangong layer in the cloud server involves different storage locations, multiple storage locations of the external space are established as multiple storage points. The external space is used to store and operate the three-dimensional model. There is a viscous relationship between the three-dimensional model and the Tiangong layer in the geographical surface. Therefore, it is necessary to respectively correspond the multiple storage points in the external space to the storage locations where the Tiangong layer is stored in the cloud server. In order to maintain the viscous relationship between the upper-layer reference surface and the geographical surface, the three-dimensional model needs to be matched. The corresponding storage points of the geographic surface of the type are stored in the external source space. For example, the Tiangong layer has storage operations of related information such as the upper layer, the reference surface, and the landmarks, and the viscous relationship is between the upper reference surface and the geographic surface. Therefore, the three-dimensional model needs to be stored in the storage point of the upper reference surface of the Tiangong layer in the external source space, and then the correspondence between the Tiangong layer and the three-dimensional model can be completed to obtain a management model, which can greatly improve the corresponding operation management between the three-dimensional model and the Tiangong layer, facilitate the subsequent tracking of sensitive targets and fire alarm prediction, assist in the subsequent fire location information to the alarm platform, and can accurately assist in the subsequent police dispatch and evacuation of personnel;

[0071] In one embodiment, the step S2 of collecting sensitive targets based on the fire management scope, formulating an ecosystem corresponding to the sensitive targets, building the ecosystem of the sensitive targets into the management model, and tracking the sensitive targets includes:

[0072] S21, setting a sensitive target database, wherein the sensitive target database includes sensitive targets, information on triggering conditions of fires on sensitive targets, and danger levels of sensitive targets combined with the triggering condition information;

[0073] S22, matching multiple targets within the fire management range based on the sensitive target database to obtain multiple sensitive targets;

[0074] S23, determining the safety range of the sensitive target as the ecosystem of the sensitive target according to the inducing condition information of the fire of the sensitive target, and constructing the ecosystem of the sensitive target into a three-dimensional model, wherein the ecosystem always takes the corresponding sensitive target as the center and follows the sensitive target;

[0075] S24. Multiple sensitive targets are matched and bound one by one according to multiple landmarks in the Tiangong layer. The landmarks are always connected to the upper-layer control surface, and the corresponding sensitive targets are tracked through the landmarks.

[0076] As described in the above steps S21-S24, in order to quickly determine whether the target object within the fire management scope is a sensitive target, it is necessary to set in advance the sensitive target and the inducing condition information of the sensitive target fire and the danger level of the sensitive target combined with the inducing condition information (for example, there are high-temperature objects around the gas tank, or there is a gas tank leak, etc., the target object, basic environmental information and behavior information, all belong to the inducing condition information) (here is based on the evaluation coefficient of the destructive power of the fire caused by the sensitive target under certain inducing condition information) to obtain the sensitive target database. For example, the target object within the fire management scope here is the item within the fire management scope, which can be a water cup, computer, lighter, mobile phone , gas tanks, cars and other targets. Here, lighters, gas tanks and cars are all sensitive targets. The safety range of sensitive targets is determined according to the triggering conditions of fires of sensitive targets as the ecosystem of sensitive targets. The ecosystem of sensitive targets is constructed into the three-dimensional model. The ecosystem is always centered on the corresponding sensitive targets and follows the sensitive targets. The significance of the ecosystem lies in the safety range of the triggering conditions of fires of corresponding sensitive targets. Therefore, it is necessary to bind these sensitive targets to each other through landmarks. The landmarks will focus on tracking the corresponding bound sensitive targets. Since the landmarks are always connected to the upper-level control surface for communication, they will move according to the sensitive targets during the tracking process. During the movement The corresponding connection position between the landmark and the upper control surface will also move. The landmark is bound to the sensitive target, but the landmark is a target corresponding tracking. Its information connection is always the upper control surface, and it is set to move with the sensitive target. At the same time, the corresponding position on the upper control surface will also change. The landmark here is a target tracker, and the data is loaded on the upper control surface. On the upper control surface, there is a completely corresponding viscous relationship between the upper control surface and the geographical surface of the fire management range (three-dimensional model). For example, in the physical sense, two surfaces of the same shape and area are arranged in a spatial relationship up and down. There is a spatially comprehensive corresponding relationship between the surfaces of the same shape and area up and down. This is due to There is a viscous relationship between the upper reference surface and the geographic surface of the three-dimensional model. Therefore, the position of the landmark on the upper reference surface represents the position on the geographic surface, and the sensitive target can be accurately located. The data calculation will not be placed on the three-dimensional model here. The three-dimensional model is only used to collect various targets within the scope of fire management and update its own model. Other data calculations are performed through the Tiangong layer, which can ensure that the three-dimensional model will not cause data confusion due to excessive data calculation pressure. It has a better data processing division of labor, and its purpose is to obtain the location of sensitive targets more accurately, so as to facilitate the subsequent sending of the fire location to the alarm platform for rescue, and also facilitate subsequent fire evacuation.

[0077] In one embodiment, the step S3 of collecting fire information from the ecosystem corresponding to the sensitive target includes:

[0078] S31, obtaining environmental information in the sensitive target ecosystem, wherein the environmental information includes target objects, basic environmental information, and behavior information;

[0079] S32, matching the fire-inducing condition information of the corresponding sensitive target and the danger level of the sensitive target combined with the fire-inducing condition information in the sensitive target database according to the environmental information;

[0080] S33, obtaining the geographic location of sensitive targets based on landmarks;

[0081] S34: The sensitive target, the geographical location of the sensitive target, the fire inducing condition information of the corresponding sensitive target matched in the sensitive target database, and the danger level of the sensitive target combined with the fire inducing condition information are taken as fire information.

[0082] As described in the above steps S31-S34, the target object, basic environmental information and behavior information in the sensitive target ecosystem are obtained as environmental information, wherein the basic environmental information is the temperature, smoke and air content information within the sensitive target ecosystem (there may be a situation where the gas or dust density is too high) and the scope involved in the ecosystem; the behavior information includes any moving behavior, for example, a person places a high-temperature object in the sensitive target ecosystem, or a person repeatedly turns on and off the gas switch (which will cause gas leakage). The above specific data to be monitored are set according to the actual situation and the use environment, which will not be repeated here. Then, the corresponding sensitive target fire inducing condition information and the danger level of the sensitive target combined with the inducing condition information are matched in the sensitive target database according to the environmental information. The data coverage in the sensitive target database is relatively comprehensive, so it is possible to match the appropriate information. The geographical location of the sensitive target is obtained based on the landmark. The landmark here is a part of the Tiangong layer, which can accurately obtain the geographical location of the sensitive target, and multiple landmarks that have been bound and tracked with the sensitive target have an information connection relationship through the upper-layer comparison surface, and can be automatically corrected according to the actual position between the landmarks and the position of the tracked sensitive target to avoid errors in the tracked position. For example, the distance between two landmarks on the upper comparison surface is 10 meters, but the distance between the corresponding sensitive targets collected by the two landmarks is 7 meters, which means that there is confusion in the corresponding landmark positions on the upper comparison surface; further explained as the distance between the landmarks on the upper comparison surface is 10 meters, but the geographical location of one of the sensitive targets collected by the digital display is A, and the geographical location of the other sensitive target is B, and the distance between AB is 7, which is a mismatch. This means that the geographical location of the sensitive target is not accurately obtained. When this happens, the edge position of the upper comparison surface (for example, the farthest The distance between the two edge points is T, and the landmark is tracked and corrected through the two farthest edge points). The correction is performed in stages or when there is a mismatch. The upper control surface is a virtual surface that is set in advance and accurately corresponds to the geographical surface position of the fire management range. Therefore, there is no need to perform independent position calculation or acquisition based on the geographical surface of the fire management range. Here, accurate position tracking can be performed directly based on the upper control surface set in advance. The acquisition of geographic location is efficient and accurate, and small-scale fires can be predicted and located, avoiding the fire from being discovered until it develops to a certain extent, causing more serious consequences.

[0083] In one embodiment, the step S4 of generating alarm information according to the prediction information and sending it to the alarm platform includes:

[0084] S41. Calculate a prediction index based on the fire information, wherein the calculation formula of the prediction index is: Among them, W is the prediction index, μ is the risk assessment coefficient of the sensitive target, i is the number of the induced condition information, k is the total number of induced condition information, β i is the danger level of the i-th inducing condition information, L i is the distance between the triggering condition information of the i-th sensitive target fire and the sensitive target in the ecosystem. It should be noted that β i The larger the value, the i The smaller the value of is, the larger the value of μ is, which means the more dangerous the fire is. This algorithm is only applicable to the situation where there is inducing condition information of sensitive target fire in the sensitive target ecosystem and the dangerous level of sensitive target combined with inducing condition information;

[0085] S42, formulating an alarm rule, wherein the alarm rule includes an alarm scheme corresponding to a numerical range of the prediction index, determining a numerical range in the alarm rule according to the prediction index, matching the corresponding alarm scheme, and using the prediction index and the corresponding alarm scheme as prediction information;

[0086] S43, sending the prediction index combined with the fire information to the alarm platform through the alarm plan for alarming;

[0087] As described in the above steps S41-S43, the prediction index is calculated through the collected fire information and matched with the corresponding alarm plan, wherein the alarm plan includes a fire emergency plan, for example, prompts to evacuate the crowd and prevent people in the surrounding area from moving to the accident area. The prompt method here can be a buzzer and a voice broadcast prompt, which can quickly respond to the evacuation of personnel, and also prevent uninformed personnel outside the evacuation area from entering the area that needs to be evacuated, thereby reducing the evacuation pressure. The numerical range in the alarm rule is determined according to the prediction index, and the corresponding alarm plan is matched. The prediction index and the corresponding alarm plan are used as prediction information, and the prediction index is combined with the fire information through the alarm plan to send an alarm to the alarm platform for alarm. The location of the fire can be accurately and quickly determined, and it has good fire response and prediction capabilities, and can predict the degree of harm to sensitive targets and evacuate people as soon as possible.

[0088] Example 2, please refer to Figure 2 As shown, the fire alarm prediction system based on artificial intelligence integration described in this embodiment includes:

[0089] A construction module is used to determine the fire management scope and build a management model based on the fire management scope;

[0090] The formulation module is connected with the construction module and is used to collect sensitive targets based on the fire management scope, formulate an ecosystem corresponding to the sensitive targets, build the ecosystem of the sensitive targets into the management model, and track the sensitive targets;

[0091] The collection module is used to formulate module connections and collect fire information corresponding to the ecosystem of sensitive targets;

[0092] The alarm module is connected to the acquisition module and is used to calculate prediction information based on the fire information, wherein the prediction information includes a prediction index and a corresponding alarm scheme, and generates alarm information based on the prediction information and sends it to the alarm platform.

[0093] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A fire alarm prediction method based on artificial intelligence integration, characterized in that: The following steps are involved: Determine the fire management scope and build a management model based on the fire management scope; Collect sensitive targets based on the fire management scope, formulate ecosystems for the sensitive targets, build the ecosystems of sensitive targets into the management model, and track the sensitive targets; Collect fire information corresponding to the ecosystem of sensitive targets; Prediction information is calculated based on the fire information, wherein the prediction information includes a prediction index and a corresponding alarm scheme, and alarm information is generated based on the prediction information and sent to the alarm platform.

2. The fire alarm prediction method based on artificial intelligence integration according to claim 1 is characterized in that: The steps of determining the fire management scope and building a management model based on the fire management scope include: Determine the fire management scope and collect characteristic information within the fire management scope, wherein the characteristic information includes geographic environment information, infrastructure information and dynamic information; A three-dimensional model of the fire management range is constructed according to the characteristic information, and a palace layer is set corresponding to the three-dimensional model; The management model is obtained by matching the Tiangong layer with the three-dimensional model.

3. The fire alarm prediction method based on artificial intelligence integration according to claim 2 is characterized in that: The step of constructing a three-dimensional model of the fire management range according to the characteristic information and setting a palace layer corresponding to the three-dimensional model includes: Three-dimensional modeling of features within the fire management range is performed according to the feature information to obtain a feature model, and the feature models are arranged according to positional relationships within the fire management range to obtain a three-dimensional model of the fire management range; An upper control surface is set corresponding to the three-dimensional model, and a corresponding cloud server is configured corresponding to the upper control surface; The upper reference surface is matched with the geographical surface of the fire management range, and a viscous relationship between the upper reference surface and the geographical surface is established, wherein the viscous relationship is a comprehensive upper and lower position correspondence between the upper reference surface and the geographical surface; A plurality of landmarks are set corresponding to the upper control surface, the plurality of landmarks are communicated and connected with each other, and the plurality of landmarks are connected with the upper control surface; The upper control surface with multiple landmarks and the upper control surface that establishes a viscous relationship with the geographical surface are regarded as the Tiangong layer.

4. The fire alarm prediction method based on artificial intelligence integration according to claim 3 is characterized in that: The step of obtaining a management model by matching the Tiangong layer with the three-dimensional model includes: Configure the corresponding external source space on the cloud server corresponding to the Tiangong layer, and connect the external source space to the cloud server; Mark the storage points in the external space and establish the corresponding relationship between the storage points and the Tiangong layer in the cloud server; The three-dimensional model is put into the external space. According to the viscous relationship between the geographical surface of the three-dimensional model and the Tiangong layer and the corresponding relationship between the storage point and the Tiangong layer in the cloud server, the geographical surface of the three-dimensional model is stored in the external space corresponding to the storage point, and the correspondence between the Tiangong layer and the three-dimensional model is completed to obtain the management model.

5. The fire alarm prediction method based on artificial intelligence integration according to claim 1 is characterized in that: The steps of collecting sensitive targets based on the fire management scope, formulating an ecosystem corresponding to the sensitive targets, building the ecosystem of the sensitive targets into the management model, and tracking the sensitive targets include: Setting a sensitive target database, wherein the sensitive target database includes sensitive targets, information on triggering conditions of fires on sensitive targets, and danger levels of sensitive targets combined with the triggering condition information; Based on the sensitive target database, multiple targets within the fire management range are matched respectively to obtain multiple sensitive targets; According to the fire inducing conditions of sensitive targets, the safety range of sensitive targets is determined as the ecosystem of sensitive targets, and the ecosystem of sensitive targets is constructed into a three-dimensional model. The ecosystem always takes the corresponding sensitive targets as the center and follows the sensitive targets. According to the multiple landmarks in the Tiangong layer, multiple sensitive targets are matched and bound one by one. The landmarks are always connected to the upper-level control surface, and the corresponding sensitive targets are tracked through the landmarks.

6. The fire alarm prediction method based on artificial intelligence integration according to claim 1 is characterized in that: The step of collecting fire information from the ecosystem corresponding to the sensitive target includes: Obtain environmental information in the sensitive target ecosystem, where environmental information includes target objects, basic environmental information, and behavioral information; Matching the fire-inducing condition information of the corresponding sensitive target and the danger level of the sensitive target combined with the fire-inducing condition information in the sensitive target database according to the environmental information; Obtaining the geographic location of sensitive targets based on landmarks; The sensitive target, the geographical location of the sensitive target, the fire inducing condition information of the corresponding sensitive target matched in the sensitive target database, and the danger level of the sensitive target combined with the fire inducing condition information are used as fire information.

7. The fire alarm prediction method based on artificial intelligence integration according to claim 1 is characterized by: The step of generating alarm information according to the prediction information and sending it to the alarm platform includes: The prediction index is calculated based on the fire information, where the calculation formula of the prediction index is: Among them, W is the prediction index, μ is the risk assessment coefficient of the sensitive target, i is the number of the induced condition information, k is the total number of induced condition information, β i is the danger level of the i-th inducing condition information, L i is the distance between the triggering condition information of the i-th sensitive target fire and the sensitive target in the ecosystem; Formulate alarm rules, wherein the alarm rules include alarm schemes corresponding to the numerical range of the prediction index, determine the numerical range in the alarm rule according to the prediction index, match the corresponding alarm scheme, and use the prediction index and the corresponding alarm scheme as prediction information; The prediction index is combined with the fire information and sent to the alarm platform through the alarm plan for alarm.

8. An integrated fire alarm prediction system based on artificial intelligence, used to implement an integrated fire alarm prediction method based on artificial intelligence as described in any one of claims 1 to 7, characterized in that: include: A construction module is used to determine the fire management scope and build a management model based on the fire management scope; The formulation module is connected with the construction module and is used to collect sensitive targets based on the fire management scope, formulate an ecosystem corresponding to the sensitive targets, build the ecosystem of the sensitive targets into the management model, and track the sensitive targets; The collection module is used to formulate module connections and collect fire information corresponding to the ecosystem of sensitive targets; The alarm module is connected to the acquisition module and is used to calculate prediction information based on the fire information, wherein the prediction information includes a prediction index and a corresponding alarm scheme, and generates alarm information based on the prediction information and sends it to the alarm platform.

Images