Fire source accurate positioning method and device based on deep learning

Abstract

The invention relates to the technical field of fire tracing and positioning, in particular to a fire source accurate positioning method and device based on deep learning, and the method comprises the steps: obtaining a sensor data set and a fire source information data set of a fire region, generating a feature vector according to the sensor data set to obtain a feature matrix, and based on the distribution condition of the sensor data, acquiring fire source information corresponding to each feature matrix in the fire source information data set to generate a training set, a verification set and a test set, and based on the training set, the verification set and the test set, acquiring fire source information of the predicted fire area by using a preset deep learning strategy. And fire source positioning information is obtained according to the predicted fire source information. Therefore, the problems that in the related technology, under the conditions that light is insufficient, other heat sources exist or flames are shielded, a sensor is prone to misinformation or missed information, the fire source positioning precision is unstable, the fire source position is difficult to position accurately, the fire extinguishing and rescue efficiency is seriously affected, and the optimal rescue opportunity is delayed are solved.

Description

Technical Field

[0001] The present application relates to the technical field of fire source tracing and positioning, and in particular to a method and device for accurately positioning a fire source based on deep learning. Background Art

[0002] In related technologies, fire alarm systems mainly rely on smoke sensors, temperature sensors, etc. for detection. Once the sensor triggers the alarm, the area where the fire occurs can be preliminarily determined.

[0003] However, in the related technology, in conditions of insufficient light, the presence of other heat sources, or when the flame is obstructed, the sensor is prone to false alarms or missed alarms, resulting in unstable fire source positioning accuracy, making it difficult to accurately locate the fire source, seriously affecting the efficiency of fire fighting and rescue, and delaying the best rescue time, which urgently needs to be improved. Summary of the invention

[0004] The present application provides a method and device for accurately locating a fire source based on deep learning to solve the problem in the related art that in situations such as insufficient light, the presence of other heat sources, or when the flame is blocked, the sensor is prone to false alarms or missed alarms, resulting in unstable fire source positioning accuracy, making it difficult to accurately locate the fire source, seriously affecting the efficiency of fire fighting and rescue, and delaying the best rescue time.

[0005] The first aspect of the present application provides a method for accurately locating a fire source based on deep learning, comprising the following steps: obtaining a sensor data set and a fire source information data set of a fire area; generating a feature vector based on the sensor data set to obtain a feature matrix of the sensor data, and based on the distribution of the sensor data, obtaining the fire source information corresponding to each feature matrix in the fire source information data set to generate a training set, a verification set and a test set; based on the training set, the verification set and the test set, using a preset deep learning strategy to obtain the predicted fire source information of the fire area, and obtaining fire source location information based on the predicted fire source information.

[0006] Through the above technical scheme, the embodiment of the present application can obtain the sensor data set and the fire source information data set of the fire area, and can integrate various data for analysis; generate feature vectors according to the sensor data set to obtain feature matrices, and obtain corresponding fire source information to construct training sets, verification sets and test sets, so that data processing is more systematic and organized, which helps to improve the accuracy and reliability of the model; based on these data sets, deep learning strategies are used to predict fire source information and obtain positioning information, and with the help of the powerful learning and analysis capabilities of deep learning, the fire source can be accurately located, providing accurate information support for fire response, which helps to improve the efficiency and success rate of fire rescue.

[0007] Optionally, in one embodiment of the present application, generating a feature vector based on the sensor data set to obtain a feature matrix of the sensor data includes: specifying a window width and a sliding step size based on the size of the data set to obtain multiple feature vectors using sliding window segmentation; and stacking the multiple feature vectors in a preset time order to generate the feature matrix.

[0008] Through the above technical scheme, the embodiment of the present application can specify the window width and sliding step size according to the scale of the data set to obtain multiple feature vectors by sliding window segmentation, and can flexibly process the data according to the actual situation, so as to better adapt to data sets of different sizes and ensure that representative feature vectors are effectively extracted from the data set; multiple feature vectors are stacked in a certain time sequence to generate a feature matrix, which takes into account the factor of time sequence, helps to reflect the temporal characteristics of the data in subsequent analysis, and lays a good foundation for accurately analyzing sensor data and obtaining fire source related information, thereby improving the accuracy and reliability of the entire fire source positioning method.

[0009] Optionally, in one embodiment of the present application, the mapping relationship of the preset deep learning strategy is:

[0010] L′=M(F;θ)

[0011] Among them, L′ is the predicted fire source information, M is the deep learning method, F is the feature matrix used when using the deep learning method, and θ is the parameter set using the deep learning method.

[0012] Through the above technical scheme, the embodiment of the present application can clarify the mapping relationship of the deep learning strategy so that the entire fire source positioning process has a clear mathematical expression framework. This mapping relationship organically combines various key elements, which helps to accurately utilize the feature matrix and related parameters for deep learning, thereby more accurately predicting fire source information and providing strong technical support for accurate positioning of fire sources.

[0013] Optionally, in an embodiment of the present application, obtaining the fire source location information according to the predicted fire source information includes: determining specific location information within the building of the fire area according to the predicted fire source information.

[0014] Through the above technical scheme, the embodiment of the present application can use the predicted fire source information to determine the specific location information within the building in the fire area. The conversion from the predicted fire source information to the specific location information within the building makes the output result of the entire fire source positioning system highly practical, and provides precise guidance for firefighters to quickly reach the fire source and take effective fire-fighting measures, greatly improving the efficiency of fire rescue and reducing the losses caused by fire.

[0015] The second aspect of the present application provides a fire source accurate positioning device based on deep learning, including: an acquisition module, used to acquire a sensor data set and a fire source information data set of a fire area; a generation module, used to generate a feature vector according to the sensor data set to obtain a feature matrix of the sensor data, and based on the distribution of the sensor data, obtain the fire source information corresponding to each feature matrix in the fire source information data set to generate a training set, a verification set and a test set; a positioning module, used to obtain the predicted fire source information of the fire area based on the training set, the verification set and the test set using a preset deep learning strategy, and obtain fire source positioning information according to the predicted fire source information.

[0016] Through the above technical scheme, the embodiment of the present application can obtain the sensor data set and the fire source information data set of the fire area, and can integrate various data for analysis; generate feature vectors according to the sensor data set to obtain feature matrices, and obtain corresponding fire source information to construct training sets, verification sets and test sets, so that data processing is more systematic and organized, which helps to improve the accuracy and reliability of the model; based on these data sets, deep learning strategies are used to predict fire source information and obtain positioning information, and with the help of the powerful learning and analysis capabilities of deep learning, the fire source can be accurately located, providing accurate information support for fire response, which helps to improve the efficiency and success rate of fire rescue.

[0017] Optionally, in one embodiment of the present application, the generation module includes: a segmentation unit, used to specify a window width and a sliding step size according to the size of the data set, so as to obtain multiple feature vectors by sliding window segmentation; and a generation unit, used to stack the multiple feature vectors in a preset time order to generate the feature matrix.

[0018] Through the above technical scheme, the embodiment of the present application can specify the window width and sliding step size according to the scale of the data set to obtain multiple feature vectors by sliding window segmentation, and can flexibly process the data according to the actual situation, so as to better adapt to data sets of different sizes and ensure that representative feature vectors are effectively extracted from the data set; multiple feature vectors are stacked in a certain time sequence to generate a feature matrix, which takes into account the factor of time sequence, helps to reflect the temporal characteristics of the data in subsequent analysis, and lays a good foundation for accurately analyzing sensor data and obtaining fire source related information, thereby improving the accuracy and reliability of the entire fire source positioning method.

[0019] Optionally, in one embodiment of the present application, the mapping relationship of the preset deep learning strategy is:

[0020] L′=M(F;θ)

[0021] Among them, L′ is the predicted fire source information, M is the deep learning method, F is the feature matrix used when using the deep learning method, and θ is the parameter set using the deep learning method.

[0022] Through the above technical scheme, the embodiment of the present application can clarify the mapping relationship of the deep learning strategy so that the entire fire source positioning process has a clear mathematical expression framework. This mapping relationship organically combines various key elements, which helps to accurately utilize the feature matrix and related parameters for deep learning, thereby more accurately predicting fire source information and providing strong technical support for accurate positioning of fire sources.

[0023] Optionally, in one embodiment of the present application, the positioning module is further used to determine specific location information within the building of the fire area based on the predicted fire source information.

[0024] Through the above technical scheme, the embodiment of the present application can use the predicted fire source information to determine the specific location information within the building in the fire area. The conversion from the predicted fire source information to the specific location information within the building makes the output result of the entire fire source positioning system highly practical, and provides precise guidance for firefighters to quickly reach the fire source and take effective fire-fighting measures, greatly improving the efficiency of fire rescue and reducing the losses caused by fire.

[0025] The third aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for accurately locating a fire source based on deep learning as described in the above embodiment.

[0026] The fourth aspect of the present application provides a computer-readable storage medium, which stores a computer program. When the program is executed by a processor, it implements the above-mentioned method for accurately locating a fire source based on deep learning.

[0027] The fifth aspect of the present application provides a computer program product, including a computer program, which, when executed, is used to implement the above-mentioned method for accurately locating a fire source based on deep learning.

[0028] The embodiments of the present application can obtain sensor data sets and fire source information data sets to comprehensively analyze various data, generate feature matrices through sensor data sets and construct related data sets, improve model accuracy and reliability, predict fire source location information with the help of deep learning, and improve fire rescue efficiency and success rate; determine the window width and sliding step size according to the size of the data set to obtain feature vectors and stack them into feature matrices in time series, which can adapt to different data sets and reflect time series, thereby improving the accuracy of fire source location; clarify the deep learning strategy mapping relationship to construct a clear mathematical framework, which is conducive to accurate deep learning to accurately predict fire source information; convert the predicted fire source information into specific location information in the building, so that the output results are practical, provide accurate guidance for firefighters, greatly improve fire rescue efficiency and reduce losses.

[0029] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0031] Figure 1 A flowchart of a method for accurately locating a fire source based on deep learning according to an embodiment of the present application;

[0032] Figure 2 A flowchart of a characteristic relationship of a method for accurately locating a fire source based on deep learning according to a specific embodiment of the present application;

[0033] Figure 3 A schematic diagram of the structure of a fire source accurate positioning device based on deep learning provided according to an embodiment of the present application;

[0034] Figure 4 The figure is a structural example diagram of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0035] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0036] The following describes the method and device for accurately locating the fire source based on deep learning in the embodiment of the present application with reference to the accompanying drawings. In view of the related technologies mentioned in the above background technology center, in the case of insufficient light, the presence of other heat sources or the flame being blocked, the sensor is prone to false alarms or missed alarms, resulting in unstable fire source positioning accuracy, making it difficult to accurately locate the fire source, seriously affecting the efficiency of fire fighting and rescue, and delaying the best rescue time. The present application provides a method for accurately locating the fire source based on deep learning, in which a sensor data set and a fire source information data set of the fire area can be obtained, and multiple data can be integrated for analysis; feature vectors are generated according to the sensor data set to obtain feature matrices, and corresponding fire source information is obtained to construct training sets, verification sets and test sets, so that data processing is more systematic and organized, which helps to improve the accuracy and reliability of the model; based on these data sets, deep learning strategies are used to predict fire source information and obtain positioning information, and with the help of the powerful learning and analysis capabilities of deep learning, the fire source is accurately located, providing accurate information support for fire response, which helps to improve the efficiency and success rate of fire rescue. This solves the problem in the related art that in conditions of insufficient light, the presence of other heat sources or when the flame is blocked, the sensor is prone to false alarms or missed alarms, resulting in unstable fire source positioning accuracy, making it difficult to accurately locate the fire source, seriously affecting the efficiency of fire fighting and rescue, and delaying the best time for rescue.

[0037] Specifically, Figure 1 A flowchart of a method for accurately locating a fire source based on deep learning provided in an embodiment of the present application.

[0038] like Figure 1 As shown, the fire source accurate positioning method based on deep learning includes the following steps:

[0039] In step S101, a sensor data set and a fire source information data set of a fire area are obtained.

[0040] It should be noted that when using the existing deep learning-based fire source location technology to locate sensor data in real time, the severity and potential danger of the fire situation place extremely high demands on the accuracy of the model. Incorrect positioning may lead to serious consequences. Therefore, the algorithm model needs to be well trained through historical examples, adapt to various complex fire scenarios and different environmental conditions, and achieve accurate positioning of the fire source.

[0041] In the actual implementation process, multiple fires have occurred inside the building and D sensors are installed. The data file corresponding to each fire includes D columns of sensor data. The data file and sensor information corresponding to each fire are used as the sensor data set of the fire area. For any fire, the recorded fire source information is obtained, and the fire source information of the fires that have occurred inside the building is recorded to obtain the fire source information data set of the fire area.

[0042] The embodiments of the present application can obtain a sensor data set containing multiple fire data files and sensor information, as well as a fire source information data set that records multiple fire source information, and use historical examples to train the algorithm model well. This helps the model adapt to complex fire scenarios and different environmental conditions, thereby achieving accurate positioning of the fire source and reducing the risk of serious consequences due to positioning errors.

[0043] In step S102, a feature vector is generated according to the sensor data set to obtain a feature matrix of the sensor data, and based on the distribution of the sensor data, the fire source information corresponding to each feature matrix in the fire source information data set is obtained to generate a training set, a verification set and a test set.

[0044] It can be understood that the feature matrix of sensor data contains key feature information of sensor data in different time windows, which can be used as input of deep learning model for fire source location task. Fire source information includes but is not limited to fire source location, fire source intensity and other information.

[0045] Optionally, in one embodiment of the present application, generating a feature vector according to a sensor data set to obtain a feature matrix of the sensor data includes: specifying a window width and a sliding step size according to the size of the data set to obtain a plurality of feature vectors by using a sliding window segmentation; and stacking the plurality of feature vectors in a preset time sequence to generate a feature matrix, wherein the formula is:

[0046]

[0047] Among them, F is the feature matrix of the sensor data set, N is the number of feature vectors obtained by using sliding window segmentation of the data set, w is the width of the sliding window, D is the number of sensors inside the building, and each row in the feature matrix F is a feature vector of a sliding window, containing a total of wD elements.

[0048] Furthermore, for each feature matrix, if the corresponding fire source location information in the fire source information dataset is paired with the feature matrix, the feature matrix and the corresponding fire source location information form a training sample, and all training samples are combined to form a dataset for training. The description of the dataset for training is:

[0049] Q={(F i ,L i )|F i ∈F,L i ∈L,i=1,2,...,m}

[0050] Among them, Q is the final dataset used for training, F i The feature matrix F = {F1, F2, ..., F m}, L iThe fire source information dataset L = {L1, L2, ..., L m}, m is the total number of data sets used for training.

[0051] In the actual implementation process, the data set used for training refers to the data set used by the deep learning algorithm model in the training process, and the training set, validation set, and test set are divided from the data set used for training according to the required proportion. Specifically, a random sampling method is used to extract samples from the data set used for training, and the training set, validation set, and test set are obtained according to the data set division ratio.

[0052] The embodiment of the present application can generate feature vectors by specifying window width and sliding step size according to the scale of the data set, and then stack them into feature matrices in a certain time sequence, which can effectively extract key feature information of sensor data in different time windows; according to the distribution of sensor data, the feature matrix is ​​paired with the corresponding fire source information to form training samples, and then a data set for training is constructed to make full use of the association between sensor data and fire source information; the training set, validation set and test set are divided into training set, validation set and test set according to proportion from the data set used for training, and samples are extracted by random sampling method, which helps to reasonably distribute data and provide scientific data support for the training, validation and testing of deep learning algorithm models, thereby improving the accuracy of the model in the fire source positioning task.

[0053] In step S103, based on the training set, the validation set and the test set, the fire source information of the predicted fire area is obtained using a preset deep learning strategy, and the fire source location information is obtained according to the predicted fire source information.

[0054] It should be noted that the fire source location information is the specific location information of the fire source in a specific space (such as but not limited to the interior of a building), usually expressed in a form that is easy to understand and apply, such as a specific room number, floor location, coordinates relative to a reference point of the building, etc. The fire source location information is the final output of the entire fire source location method, directly providing key guidance for fire emergency rescue, helping firefighters to quickly determine the specific location of the fire, so as to take effective fire extinguishing and rescue measures, reduce the losses caused by the fire, and protect the safety of life and property.

[0055] Optionally, in one embodiment of the present application, the mapping relationship of the preset deep learning strategy is:

[0056] L′=M(F;θ)

[0057] Among them, L′ is the predicted fire source information, M is the deep learning method, F is the feature matrix used when using the deep learning method, and θ is the parameter set using the deep learning method.

[0058] Specifically, the mapping relationship of the deep learning strategy is the core mathematical expression of the entire deep learning fire source location model. It clearly explains the relationship between the model input and output, as well as the impact of the model's internal parameters on the prediction results.

[0059] In the actual execution process, when the feature matrix F is input into the deep learning model M, the model performs a series of linear and nonlinear transformation operations on the feature matrix according to its internal network structure and the current parameter set (including but not limited to convolution operations in the convolution layer, downsampling operations in the pooling layer, weighted summation of neurons in the fully connected layer, etc.), and gradually extracts and abstracts the information in the feature matrix through these operations, and finally obtains the predicted fire source information. In the training stage, the model can continuously adjust the parameter set according to the data of the training set and the validation set to optimize the mapping relationship so that the predicted fire source information is as close to the real fire source information as possible.

[0060] Furthermore, the specific location information in the building of the fire area is determined based on the predicted fire source information. For example, if a three-dimensional coordinate system with a fixed point as the origin is established in the building, then the predicted fire source location coordinates (x ′ ,y ′ ,z ′ )(Assuming L ′ =(x ′ ,y ′ ,z ′ )) can be directly corresponded to the specific spatial location in the building. ′ For other forms of representation, such as relative position information relative to a reference area or sensor, it is necessary to convert it into specific room numbers, floor locations, and other location information that is easy to understand and apply through pre-established conversion rules and building layout information.

[0061] In some cases, the predicted fire source information may have a certain error range. In this case, a possible fire source location range can be delineated based on the error distribution, and the prediction results can be continuously updated over time, and the positioning range can be further narrowed in combination with newly collected data to improve the accuracy of fire source positioning. For example, algorithms including but not limited to Kalman filtering can be used to fuse and optimize continuous prediction results to gradually approach the actual fire source location. At the same time, the fire source positioning information can also be comprehensively evaluated and adjusted in combination with other auxiliary information in the building, such as personnel evacuation routes, distribution of fire-fighting facilities, etc., to ensure that fire rescue personnel can accurately reach the fire source location in the shortest time and implement effective fire extinguishing and rescue operations.

[0062] The embodiment of the present application can use a deep learning strategy to obtain fire source information and obtain positioning information based on a training set, a validation set, and a test set. Among them, the use of a training set, a validation set, and a test set provides a comprehensive data foundation for the deep learning strategy, so that the model can fully learn different types of data features during the training process, and improve the accuracy and generalization ability of the model. Secondly, the deep learning strategy can efficiently analyze and process the data, accurately predict the fire source information of the fire area, and then obtain the fire source positioning information, which helps to quickly and accurately determine the location of the fire source in a fire scene, and provide an important basis for the formulation of fire response measures, thereby improving the efficiency and effectiveness of fire rescue.

[0063] The following is a detailed description of the fire source accurate positioning method based on deep learning of the present application in conjunction with a specific embodiment. Figure 2 As shown, Figure 2 This is a flow chart of the characteristic relationships involved in this application. Figure 2 It can be seen that:

[0064] Step S201: Fire area data.

[0065] D different types of sensors are reasonably arranged inside the building. These sensors serve as key equipment for data collection. Their types include but are not limited to temperature sensors, smoke sensors, carbon monoxide sensors, etc. They can sense changes in environmental parameters when a fire occurs from multiple dimensions. Every time a fire occurs, the corresponding sensors will continue to record data.

[0066] Step S202: sensor data set.

[0067] The changes of various environmental parameters at different sensor locations during the fire process are recorded, including but not limited to temperature, smoke concentration, etc.

[0068] Step S203: Fire source information data set.

[0069] The fire source information dataset records in detail the fire source related information of each fire, including the location of the fire source (such as including but not limited to the specific room number, floor location or coordinates relative to the building reference point), the intensity of the fire source (such as including but not limited to the quantitative representation of the size of the fire), etc.

[0070] Step S204: Data set for training.

[0071] Further deriving a training data set from the data of the fire area involves the integration and processing of the sensor data set and the fire source information data set. Specifically, for the sensor data set, the appropriate window width and sliding step size are specified according to its scale, and the feature vector is obtained by sliding window segmentation. These feature vectors are then stacked in chronological order to obtain the feature matrix of the sensor data set. Next, the feature matrix is ​​paired with the corresponding fire source location information in the fire source information data set. If the two match, the feature matrix and the corresponding fire source location information form a training sample. All such training samples are combined together to form a training data set.

[0072] Step S205: Deep learning method.

[0073] The deep learning method is the core technical means of the entire fire source location process. It takes the feature matrix and parameter set as input, and performs data processing and learning through the internal complex neural network structure and algorithm. During the training process, the deep learning method continuously adjusts the parameter set according to the data of the training set to optimize the prediction ability of the model.

[0074] Step S206: predicting fire source information.

[0075] When the feature matrix of the test set is input into the trained deep learning model, the model calculates the predicted fire source information through a mapping relationship based on the learned knowledge and parameter set.

[0076] Step S207: Locating the fire source.

[0077] Based on the predicted fire source information, the specific location information within the building of the fire area is further determined to obtain the fire source positioning information.

[0078] According to the method for accurately locating the fire source based on deep learning proposed in the embodiment of the present application, the sensor data set and the fire source information data set of the fire area can be obtained, and the data from various aspects can be integrated for analysis; the feature vector is generated according to the sensor data set to obtain the feature matrix, and the corresponding fire source information is obtained to construct the training set, verification set and test set, so that the data processing is more systematic and organized, which helps to improve the accuracy and reliability of the model; based on these data sets, the deep learning strategy is used to predict the fire source information and obtain the positioning information, and with the help of the powerful learning and analysis capabilities of deep learning, the fire source can be accurately located, providing accurate information support for fire response, which helps to improve the efficiency and success rate of fire rescue.

[0079] Next, a fire source accurate positioning device based on deep learning proposed in accordance with an embodiment of the present application is described with reference to the accompanying drawings.

[0080] Figure 3It is a block diagram of a fire source accurate positioning device based on deep learning according to an embodiment of the present application.

[0081] like Figure 3 As shown, the fire source accurate positioning device 10 based on deep learning includes: an acquisition module 100, a generation module 200 and a positioning module 300.

[0082] Specifically, the acquisition module 100 is used to acquire a sensor data set and a fire source information data set of a fire area.

[0083] The generation module 200 is used to generate a feature vector according to the sensor data set to obtain a feature matrix of the sensor data, and based on the distribution of the sensor data, obtain the fire source information corresponding to each feature matrix in the fire source information data set to generate a training set, a verification set and a test set.

[0084] The positioning module 300 is used to obtain the fire source information of the predicted fire area based on the training set, the validation set and the test set by using a preset deep learning strategy, and obtain the fire source positioning information according to the predicted fire source information.

[0085] Optionally, in one embodiment of the present application, the generation module 200 includes: a segmentation unit and a generation unit.

[0086] The segmentation unit is used to specify the window width and sliding step size according to the scale of the data set, so as to obtain multiple feature vectors by using the sliding window segmentation.

[0087] The generating unit is used for stacking a plurality of feature vectors in a preset time sequence to generate a feature matrix.

[0088] Optionally, in one embodiment of the present application, the mapping relationship of the preset deep learning strategy is:

[0089] L′=M(F;θ)

[0090] Among them, L′ is the predicted fire source information, M is the deep learning method, F is the feature matrix used when using the deep learning method, and θ is the parameter set using the deep learning method.

[0091] Optionally, in one embodiment of the present application, the positioning module 300 is further used to determine specific location information within the building of the fire area according to the predicted fire source information.

[0092] It should be noted that the aforementioned explanation of the embodiment of the method for accurately locating a fire source based on deep learning is also applicable to the device for accurately locating a fire source based on deep learning in this embodiment, and will not be repeated here.

[0093] According to the deep learning-based fire source accurate positioning device proposed in the embodiment of the present application, the sensor data set and the fire source information data set of the fire area can be obtained, and the data from various aspects can be integrated for analysis; the feature vector is generated according to the sensor data set to obtain the feature matrix, and the corresponding fire source information is obtained to construct the training set, verification set and test set, so that the data processing is more systematic and organized, which helps to improve the accuracy and reliability of the model; based on these data sets, the deep learning strategy is used to predict the fire source information and obtain the positioning information, and with the help of the powerful learning and analysis capabilities of deep learning, the fire source can be accurately located, providing accurate information support for fire response, which helps to improve the efficiency and success rate of fire rescue.

[0094] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0095] Memory 401 , processor 402 , and a computer program stored in the memory 401 and executable on the processor 402 .

[0096] When the processor 402 executes the program, the method for accurately locating a fire source based on deep learning provided in the above embodiment is implemented.

[0097] Furthermore, the electronic device further comprises:

[0098] The communication interface 403 is used for communication between the memory 401 and the processor 402 .

[0099] The memory 401 is used to store computer programs that can be executed on the processor 402 .

[0100] The memory 401 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0101] If the memory 401, the processor 402 and the communication interface 403 are implemented independently, the communication interface 403, the memory 401 and the processor 402 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0102] Optionally, in a specific implementation, if the memory 401, the processor 402 and the communication interface 403 are integrated on a chip, the memory 401, the processor 402 and the communication interface 403 can communicate with each other through an internal interface.

[0103] The processor 402 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0104] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for accurately locating a fire source based on deep learning.

[0105] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed, is used to implement the above-mentioned method for accurately locating a fire source based on deep learning.

[0106] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0107] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0108] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0109] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.

[0110] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented by hardware, as in another embodiment, it can be implemented by any one or a combination of multiple of the following technologies known in the art: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0111] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0112] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0113] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A method for accurately locating a fire source based on deep learning, characterized in that: The following steps are involved: Obtain sensor data sets and fire source information data sets for the fire area; Generate a feature vector according to the sensor data set to obtain a feature matrix of the sensor data, and obtain the fire source information corresponding to each feature matrix in the fire source information data set based on the distribution of the sensor data to generate a training set, a validation set and a test set; Based on the training set, the validation set and the test set, a preset deep learning strategy is used to obtain fire source information predicting the fire area, and fire source location information is obtained based on the predicted fire source information.

2. The method according to claim 1, characterized in that Generating a feature vector according to the sensor data set to obtain a feature matrix of the sensor data includes: Specify the window width and sliding step size according to the size of the data set to obtain multiple feature vectors using sliding window segmentation; The plurality of feature vectors are stacked in a preset time sequence to generate the feature matrix.

3. The method according to claim 1, characterized in that The mapping relationship of the preset deep learning strategy is: L′=M(F;θ) Among them, L′ is the predicted fire source information, M is the deep learning method, F is the feature matrix used when using the deep learning method, and θ is the parameter set using the deep learning method.

4. The method according to claim 1, characterized in that: The obtaining fire source location information according to the predicted fire source information includes: The specific location information within the building of the fire area is determined based on the predicted fire source information.

5. A fire source accurate positioning device based on deep learning, characterized in that: include: An acquisition module is used to acquire a sensor data set and a fire source information data set of a fire area; A generation module, used to generate a feature vector according to the sensor data set to obtain a feature matrix of the sensor data, and based on the distribution of the sensor data, obtain the fire source information corresponding to each feature matrix in the fire source information data set to generate a training set, a validation set and a test set; A positioning module is used to obtain fire source information of the fire area based on the training set, the validation set and the test set using a preset deep learning strategy, and obtain fire source positioning information based on the predicted fire source information.

6. The device according to claim 5, characterized in that The generating module comprises: A segmentation unit is used to specify a window width and a sliding step size according to the size of the data set so as to obtain multiple feature vectors by using a sliding window segmentation; A generating unit is used to stack the multiple feature vectors in a preset time sequence to generate the feature matrix.

7. The device according to claim 5, characterized in that The positioning module is further used to determine the specific location information in the building of the fire area according to the predicted fire source information.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for accurately locating a fire source based on deep learning as described in any one of claims 1 to 4.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the method for accurately locating a fire source based on deep learning as described in any one of claims 1 to 4.

10. A computer program product, comprising a computer program, characterized in that The computer program is executed to implement the method for accurately locating a fire source based on deep learning as described in any one of claims 1 to 4.

Images