A method and system for intelligent disaster monitoring and decision-making based on target detection

Abstract

The present application discloses a method and system for intelligent disaster monitoring and decision-making based on target detection, which relates to the field of disaster prevention and control technology. The method includes: obtaining the current environmental parameters of the monitoring area; obtaining the monitoring image of the monitoring area, and detecting the flame and smoke in the monitoring image; projecting the detected flame and smoke into a virtual environment for visual display; updating the virtual state parameters of the virtual environment according to the current environmental parameters; generating a fire extinguishing strategy according to the current virtual state parameters of the virtual environment; and assigning the fire extinguishing strategy to each intelligent agent so that each intelligent agent can execute the fire extinguishing strategy. The virtual state parameters of the virtual environment are updated according to the current environmental parameters of the monitoring environment, and the state of the visualization of the virtual environment is updated, so that the current state of the monitoring area can be understood in real time, thereby improving the monitoring efficiency; the fire extinguishing strategy is generated in a targeted manner according to the current state parameters of the virtual environment, and the intelligent agent can quickly respond and extinguish the fire.

Description

Technical Field

[0001] The present application relates to the field of disaster prevention and control technology, and in particular to a method and system for intelligent disaster monitoring and decision-making based on target detection. Background Art

[0002] Outdoor fires are a widespread type of disaster, including grassland fires, farmland fires, and park fires. These fires not only cause serious damage to the natural environment, but also pose a threat to human life and property. With climate change and increased human activities, the frequency of outdoor fires around the world is on the rise, posing a huge challenge to the management and protection of various outdoor resources. Traditional outdoor fire monitoring methods mainly rely on human patrols, observation towers, and satellite remote sensing. Traditional methods have problems such as blind spots in monitoring, untimely response, and inaccurate information acquisition in vast and complex outdoor environments, making it difficult to achieve effective early warning and timely response. Summary of the invention

[0003] The main purpose of the embodiments of the present application is to propose a method and system for intelligent disaster monitoring and decision-making based on target detection to improve the efficiency of disaster prevention and control.

[0004] To achieve the above objectives, an embodiment of the present application provides a method for intelligent disaster monitoring and decision-making based on target detection, the method comprising the following steps:

[0005] Get the current environmental parameters of the monitoring area;

[0006] Acquire a monitoring image of the monitoring area, and detect flames and smoke in the monitoring image;

[0007] Project the detected flames and smoke into a virtual environment for visualization;

[0008] Updating virtual state parameters of the virtual environment according to the current environment parameters;

[0009] generating a fire extinguishing strategy according to current virtual state parameters of the virtual environment;

[0010] The fire extinguishing strategy is distributed to each intelligent agent so that each intelligent agent executes the fire extinguishing strategy.

[0011] In some embodiments, the step of obtaining the current environmental parameters of the monitoring area includes the following steps:

[0012] Acquire current environmental data of the monitoring area through sensors;

[0013] Using a machine learning algorithm model to combine the current environmental data with the historical environmental data of the monitored area to obtain the current environmental parameters;

[0014] The expression of the current environment parameter is:

[0015] ;

[0016] in, is the current environmental parameter, is the current environment data, The historical environmental data; Representing the machine learning algorithm model;

[0017] ;

[0018] in, and is the weight matrix, is the bias term.

[0019] In some embodiments, the method further comprises the following steps:

[0020] Determining the fire risk level of the monitored area according to the current environmental parameters and the current environmental data using a fire risk assessment model;

[0021] The expression of the fire risk assessment model is:

[0022] ;

[0023] in, is the fire risk assessment model; is a nonlinear function based on the historical environmental data, used to describe the monitoring area Historical risk trends; is a function based on the current environmental data, used to reflect the current fire risk; is the weight coefficient; is the activation function;

[0024] The execution frequency of detecting the flame and smoke in the monitoring image is dynamically adjusted according to the fire risk level.

[0025] In some embodiments, the detecting of flames and smoke in the surveillance image comprises the following steps:

[0026] The monitoring image is enhanced by using a histogram equalization method to improve the brightness and contrast of the monitoring image to obtain an enhanced image;

[0027] The expression of the histogram equalization method is:

[0028] ;

[0029] in, q represents the pixel intensity value, is the integration variable; is the pixel intensity distribution of the monitoring image; is the pixel intensity distribution after being processed by the histogram equalization method;

[0030] Using median filtering or Gaussian filtering to filter the enhanced image to obtain a filtered image;

[0031] The expression of the Gaussian filter is:

[0032] ;

[0033] in, is the filtered image; is the pixel coordinate; is the Gaussian kernel function, is the size of the filter window; is the enhanced image;

[0034] Adjusting the color distribution of the filtered image using a white balance correction method to obtain a color-corrected image;

[0035] The expression of the white balance correction method is:

[0036] ;

[0037] in, For the color-corrected image, is the reference white value, is the white value in the filtered image;

[0038] The flame and smoke in the color-corrected image are detected using an object detection model to obtain a set of bounding box parameters.

[0039] In some embodiments, projecting the detected flame and smoke into a virtual environment for visual display comprises the following steps:

[0040] Projecting the detected flame and smoke bounding box parameter set to the virtual environment through a projection transformation matrix;

[0041] The fire source location and fire range are visualized in the virtual environment according to the bounding box parameter set.

[0042] In some embodiments, generating a fire extinguishing strategy according to the current virtual state parameters of the virtual environment comprises the following steps:

[0043] Generate the fire extinguishing strategy according to the current virtual state parameters of the virtual environment using a reinforcement learning model;

[0044] The expression of the reinforcement learning model is:

[0045] ;

[0046] in, For the fire fighting strategy, is the current state of the agent, is a set of actions that the agent can perform; represents the reinforcement learning model, is the current virtual state parameter;

[0047] Using the reinforcement learning model to update the fire fighting strategy through a policy gradient method and maximizing cumulative rewards;

[0048] The expression of the policy gradient method is:

[0049] ;

[0050] in, are parameters of the reinforcement learning model; To maximize the cumulative reward, ; for Instant rewards at any time, is the discount factor; is the total time taken to extinguish the fire.

[0051] In some embodiments, after allocating the fire extinguishing strategy to each intelligent agent so that each intelligent agent executes the fire extinguishing strategy, the method further comprises the following steps:

[0052] Acquire current monitoring data of the monitoring area; wherein the monitoring data includes environmental status and agent status;

[0053] Dynamically adjust the fire extinguishing strategy according to the current monitoring data;

[0054] The expression for adjusting the fire extinguishing strategy is:

[0055] ;

[0056] in, for the adjusted fire extinguishing strategy; is the monitoring data, ; is the environmental state, For the N The intelligent agent The state of the agent at time .

[0057] To achieve the above objectives, another aspect of the embodiments of the present application provides a system for intelligent disaster monitoring and decision-making based on target detection, the system comprising:

[0058] A parameter acquisition unit, used to acquire current environmental parameters of the monitoring area;

[0059] an object detection unit, configured to acquire a monitoring image of the monitoring area and detect flames and smoke in the monitoring image;

[0060] A target monitoring unit is used to project the detected flames and smoke into a virtual environment for visual display;

[0061] A parameter updating unit, configured to update a virtual state parameter of the virtual environment according to the current environment parameter;

[0062] A strategy generating unit, configured to generate a fire extinguishing strategy according to current virtual state parameters of the virtual environment;

[0063] The strategy allocation unit is used to allocate the fire extinguishing strategy to each intelligent agent so that each intelligent agent executes the fire extinguishing strategy.

[0064] To achieve the above objective, another aspect of an embodiment of the present application provides an electronic device, the electronic device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the above method when executing the computer program.

[0065] To achieve the above objective, another aspect of an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program implements the above method when executed by a processor.

[0066] The embodiments of the present application include at least the following beneficial effects:

[0067] The present application can obtain the current environmental parameters of the monitoring area; obtain the monitoring image of the monitoring area, and detect the flame and smoke in the monitoring image; project the detected flame and smoke into the virtual environment for visual display; update the virtual state parameters of the virtual environment according to the current environmental parameters; generate a fire extinguishing strategy according to the current virtual state parameters of the virtual environment; and assign the fire extinguishing strategy to each intelligent agent so that each intelligent agent can execute the fire extinguishing strategy. The virtual state parameters of the virtual environment are updated according to the current environmental parameters of the monitoring environment, and the state of the visualization of the virtual environment is updated, so that the current state of the monitoring area can be understood in real time, which improves the monitoring efficiency. Moreover, when flames and smoke are detected, a fire extinguishing strategy can be generated specifically according to the current state parameters of the virtual environment. After the intelligent agent is assigned to the fire extinguishing strategy, it can respond quickly and extinguish the fire, which improves the extinguishing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0069] Figure 1 A flowchart of a method for intelligent disaster monitoring and decision-making based on target detection provided in an embodiment of the present application;

[0070] Figure 2 A reference flow chart of a method for intelligent disaster monitoring and decision-making based on target detection provided in an embodiment of the present application;

[0071] Figure 3 Provided in the embodiments of this application Figure 2 Reference flow chart of S1~S2;

[0072] Figure 4 Provided in the embodiments of this application Figure 2 Reference flow chart of S3~S4;

[0073] Figure 5 A schematic diagram of the structure of a system for intelligent disaster monitoring and decision-making based on target detection provided in an embodiment of the present application;

[0074] Figure 6 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0075] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below in conjunction with the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present application. They are only examples of systems and methods consistent with some aspects of the embodiments of the present application as detailed in the attached claims.

[0076] It is understood that the terms "first", "second", etc. used in this application can be used to describe various concepts in this article, but unless otherwise specified, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another concept. For example, without departing from the scope of the embodiment of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if" and "if" as used herein can be interpreted as "at the time of" or "when" or "in response to determination".

[0077] The terms "at least one", "multiple", "each", "any", etc. used in this application, at least one includes one, two or more, multiple includes two or more, each refers to each of the corresponding multiple, and any refers to any one of the multiple.

[0078] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0079] Before describing the embodiments of the present application in detail, the related technologies involved in the embodiments of the present application are first described as follows:

[0080] In recent years, with the development of artificial intelligence and image processing technology, target detection technology has gradually been applied to the field of outdoor fire monitoring. Flame recognition technology based on target detection can accurately locate the fire source in the image and provide key support for the early warning of fire. However, traditional flame detection methods are often limited to the detection and alarm of fire sources, lacking comprehensive assessment of fire risks and intelligent decision-making capabilities, resulting in difficulty in achieving effective emergency response and resource allocation in the actual fire fighting process. With the increasing demand for the application of multi-agents such as drones and robots in fire fighting, how to coordinate and command these devices through intelligent means to achieve the goal of quickly and efficiently extinguishing fires has become an urgent problem to be solved. Most of the current research focuses on the application of a single technology, such as flame detection and environmental monitoring, but lacks a solution that can integrate multiple technologies for comprehensive fire monitoring and intelligent decision-making. Therefore, designing a comprehensive outdoor fire intelligent monitoring and decision-making solution that combines target detection, reinforcement learning, and multi-agent decision-making can effectively improve the accuracy of fire monitoring, the intelligence of decision-making, and the efficiency of fire fighting, and has important research value and application prospects.

[0081] In order to solve the shortcomings of existing outdoor fire monitoring and extinguishing, the present application provides a method and system for intelligent disaster monitoring and decision-making based on target detection, which realizes the comprehensive solution of early warning, fire monitoring, risk assessment and multi-agent collaborative fire extinguishing through intelligent means. The technical solution of the present application includes: obtaining the current environmental parameters of the monitoring area; obtaining the monitoring image of the monitoring area, and detecting the flame and smoke in the monitoring image; projecting the detected flame and smoke into the virtual environment for visual display; updating the virtual state parameters of the virtual environment according to the current environmental parameters; generating a fire extinguishing strategy according to the current virtual state parameters of the virtual environment; and assigning the fire extinguishing strategy to each intelligent agent for each intelligent agent to execute the fire extinguishing strategy. According to the current environmental parameters of the monitoring environment, the virtual state parameters of the virtual environment are updated, and the state of the virtual environment visualization is updated, so that the current state of the monitoring area can be understood in real time, which improves the monitoring efficiency. Moreover, when flames and smoke are detected, the fire extinguishing strategy can be generated according to the current state parameters of the virtual environment. After the intelligent agent is assigned to the fire extinguishing strategy, it can quickly respond and extinguish the fire, which improves the extinguishing efficiency.

[0082] The embodiment of the present application provides a method for intelligent monitoring and decision-making of disasters based on target detection, which relates to the field of disaster prevention and control technology. The method for intelligent monitoring and decision-making of disasters based on target detection provided in the embodiment of the present application can be applied to a terminal, can also be applied to a server, and can also be software running in a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, and a car terminal, etc., but is not limited to this; the server side can be configured as an independent physical server, or it can be configured as a server cluster or distributed system composed of multiple physical servers, and can also be configured as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network; the software can be an application that implements a knowledge extraction method, etc., but is not limited to the above forms.

[0083] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0084] Reference Figure 1 The embodiment of the present application provides a method for intelligent disaster monitoring and decision-making based on target detection. The method may include but is not limited to S100 to S150, which are as follows:

[0085] S100: Acquire current environmental parameters of the monitoring area.

[0086] Specifically, in this embodiment, various sensors may be set in the monitoring area to obtain current environmental parameters through the sensors. There may be one or more monitoring areas in this embodiment, and each monitoring area may be divided into multiple sub-areas for easy monitoring.

[0087] Furthermore, S100 may include S101-S102:

[0088] S101: Acquire current environmental data of the monitoring area through a sensor;

[0089] S102: using a machine learning algorithm model to combine the current environmental data and the historical environmental data of the monitored area to obtain the current environmental parameters;

[0090] The expression of the current environment parameter is:

[0091] ;

[0092] in, is the current environmental parameter, is the current environment data, The historical environmental data; Representing the machine learning algorithm model;

[0093] ;

[0094] in, and is the weight matrix, is the bias term.

[0095] Exemplarily, the current environmental parameters of this embodiment may include the temperature, humidity, wind speed, etc. currently monitored, and the historical environmental data may include the temperature, humidity, wind speed, etc. monitored in the monitoring area at historical moments.

[0096] Among them, the machine learning algorithm model can be a linear regression model, a nonlinear regression model or a neural network model.

[0097] S110: Acquire a monitoring image of the monitoring area, and detect flames and smoke in the monitoring image.

[0098] Specifically, in this embodiment, an image of the monitoring area may be captured as a monitoring image, and target detection may be performed on the monitoring image to detect whether flames and smoke exist in the monitoring image.

[0099] Considering that fire is an emergency event with a relatively low frequency, when the environmental parameters are normal, target detection may not be performed on the surveillance image to reduce the consumption of computing resources. Therefore, this embodiment can dynamically adjust the frequency of executing the step of "detecting flames and smoke in the surveillance image" according to the current environmental parameters, specifically including the following steps:

[0100] Determining the fire risk level of the monitored area according to the current environmental parameters and the current environmental data using a fire risk assessment model;

[0101] The expression of the fire risk assessment model is:

[0102] ;

[0103] in, is the fire risk assessment model; is a nonlinear function based on the historical environmental data, used to describe the monitoring area Historical risk trends; is a function based on the current environmental data, used to reflect the current fire risk; is the weight coefficient; is the activation function;

[0104] The execution frequency of detecting the flame and smoke in the monitoring image is dynamically adjusted according to the fire risk level.

[0105] It is understandable that based on the calculated fire risk level, the flame and smoke detection frequency of each monitoring area is dynamically adjusted so that high-risk areas are fully monitored and the monitoring frequency of low-risk areas is appropriately reduced to save resources.

[0106] Further, detecting the flame and smoke in the monitoring image includes the following steps S111-S114:

[0107] S111: enhancing the monitoring image by using a histogram equalization method to improve the brightness and contrast of the monitoring image to obtain an enhanced image;

[0108] The expression of the histogram equalization method is:

[0109] ;

[0110] in, q represents the pixel intensity value, is the integration variable; is the pixel intensity distribution of the monitoring image; is the pixel intensity distribution after being processed by the histogram equalization method;

[0111] S112: filtering the enhanced image using median filtering or Gaussian filtering to obtain a filtered image;

[0112] The expression of the Gaussian filter is:

[0113] ;

[0114] in, is the filtered image; is the pixel coordinate; is the Gaussian kernel function, is the size of the filter window; is the enhanced image;

[0115] S113: adjusting the color distribution of the filtered image using a white balance correction method to obtain a color-corrected image;

[0116] The expression of the white balance correction method is:

[0117] ;

[0118] in, For the color-corrected image, is the reference white value, is the white value in the filtered image;

[0119] S114: Detect flames and smoke in the color-corrected image using a target detection model to obtain a bounding box parameter set.

[0120] Specifically, the present embodiment can first perform preprocessing such as enhancement, filtering, and white balance correction on the monitoring image, thereby improving the quality of the monitoring image, so that the features therein can be efficiently and accurately extracted during subsequent target detection, and then the target detection model is used to detect the flame and smoke in the preprocessed image to obtain a set of bounding box parameters. The bounding box parameters may include the coordinates of the four vertices of the bounding box and the probability of the category of the target in the bounding box, so the position and category of the detected target can be determined based on the bounding box parameters.

[0121] S120: Projecting the detected flame and smoke into a virtual environment for visual display.

[0122] Specifically, by projecting and displaying flames and smoke in the monitoring area in a virtual environment, a large monitoring area can be monitored in a short period of time, solving problems such as low efficiency of manual inspections and the existence of monitoring blind spots.

[0123] Furthermore, S120 may include S121-S122:

[0124] S121: Projecting the detected flame and smoke boundary box parameter set to the virtual environment through a projection transformation matrix;

[0125] S122: Visually display the fire source location and fire range in the virtual environment according to the bounding box parameter set.

[0126] It is understandable that in this embodiment, the fire source position can be determined according to the vertex coordinates of the bounding box, and the fire range can be determined according to the type of the target in the bounding box, and then visualized in a virtual environment.

[0127] S130: Update virtual state parameters of the virtual environment according to the current environment parameters.

[0128] It is understandable that in order to display the environmental status of the monitored area in real time, this embodiment can dynamically synchronize the current environmental parameters obtained through monitoring to the virtual environment, so that the virtual environment can display the current status of the monitored area in real time.

[0129] S140: Generate a fire extinguishing strategy according to the current virtual state parameters of the virtual environment.

[0130] Specifically, the virtual state parameters are synchronized with the state of the monitoring area in real time, so the present embodiment can generate a corresponding fire extinguishing strategy according to the current virtual state parameters of the virtual environment, so as to extinguish the fire efficiently.

[0131] Further, S140 may include S141-S142:

[0132] S141: Generate the fire extinguishing strategy according to the current virtual state parameters of the virtual environment using a reinforcement learning model;

[0133] The expression of the reinforcement learning model is:

[0134] ;

[0135] in, For the fire fighting strategy, is the current state of the agent, is a set of actions that the agent can perform; represents the reinforcement learning model, is the current virtual state parameter;

[0136] S142: Using the reinforcement learning model to update the fire extinguishing strategy through a policy gradient method and maximizing cumulative rewards;

[0137] The expression of the policy gradient method is:

[0138] ;

[0139] in, are parameters of the reinforcement learning model; To maximize the cumulative reward, ; for Instant rewards at any time, is the discount factor; is the total time taken to extinguish the fire.

[0140] S150: Distribute the fire extinguishing strategy to each intelligent agent, so that each intelligent agent executes the fire extinguishing strategy.

[0141] Exemplarily, intelligent agents may include drones, firefighting robots, etc. Fire extinguishing strategies may be assigned to appropriate intelligent agents, and then each intelligent agent may execute the fire extinguishing strategy in a targeted manner, thereby improving the extinguishing efficiency.

[0142] Considering that the environmental state of the fire scene may change randomly, in order to adapt to the changing disaster scene, the embodiment of the present application may also include the following steps:

[0143] Acquire current monitoring data of the monitoring area; wherein the monitoring data includes environmental status and agent status;

[0144] Dynamically adjust the fire extinguishing strategy according to the current monitoring data;

[0145] The expression for adjusting the fire extinguishing strategy is:

[0146] ;

[0147] in, for the adjusted fire extinguishing strategy; is the monitoring data, ; is the environmental state, For the N The intelligent agent The state of the agent at time .

[0148] Next, the solution of the embodiment of the present application will be introduced and explained in detail with reference to specific application examples.

[0149] For example, Figure 2 This is a reference flow chart of the method of this embodiment.

[0150] Specifically, this embodiment may include the following steps:

[0151] S1, environmental monitoring and fire risk assessment, collect current environmental parameters in real time through the sensor network, and evaluate the fire risk level of each area based on the current environmental parameters;

[0152] S2, fire detection and environment visualization, using cameras to detect flames and smoke, and display the fire situation in real time in a virtual environment;

[0153] S3, virtual environment update and decision generation, updates the virtual environment in real time based on fire detection and environmental monitoring data, and uses pre-trained reinforcement learning models to generate the optimal fire extinguishing strategy;

[0154] S4. Construct emergency response and fire fighting implementation. After generating fire fighting strategies, assign tasks and use intelligent agents (such as drones, fire fighting robots, etc.) to perform fire fighting tasks to ensure the effective implementation of strategies.

[0155] Reference Figure 3 , Figure 3 This is a reference flow chart of the above S1~S2.

[0156] Specifically, in S1, the environmental monitoring module is started to continuously monitor key environmental parameters in the outdoor area, such as humidity, temperature and wind speed. After data processing and analysis, a summary of environmental parameters in each area is generated, and the fire risk level of each area is calculated using the fire risk assessment model, so as to dynamically adjust the frequency and duration of flame and smoke detection to ensure accurate monitoring of key areas. Specifically, S1 may include the following steps:

[0157] S11, Environmental parameter collection and processing, first deploy a distributed sensor network, and use sensors to collect current environmental data of outdoor areas in real time, including but not limited to humidity ,temperature , wind speed Each node in the sensor network The current environment parameters are transmitted to the server, and the server performs The collected environmental data is preliminarily processed to form a vector of the current environmental data , specifically expressed as:

[0158] ;

[0159] in, , , Respectively indicate at time humidity, temperature and wind speed parameters.

[0160] S12, Environmental data analysis and parameter summary generation, in order to conduct fire risk assessment in outdoor areas, the collected environmental data is processed and analyzed using machine learning algorithm models. Then, using the machine learning algorithm model Combined with current environmental data and historical environmental data Get the current environment parameters , specifically expressed as:

[0161] ;

[0162] Machine Learning Algorithm Model It can be further expressed as a linear or nonlinear regression model, or a neural network model, as follows:

[0163] ;

[0164] in, and is the weight matrix, is the bias term. The above parameters can be determined through model training. Current environment parameters Can describe outdoor areas A multidimensional vector of comprehensive environmental conditions.

[0165] S13, Fire risk level assessment, generating current environmental parameters Afterwards, the fire risk assessment model For each region Fire risk level By analyzing the current environmental parameters of the area and its historical environmental data, using the following formula:

[0166] ;

[0167] in, is a nonlinear function based on historical data, describing the region Historical risk trends; is a function based on current environmental data that reflects instantaneous fire risk; is the weight coefficient used to balance the impact of historical data and current data. Function is the activation function, usually a sigmoid function or a ReLU function, so that Within reasonable limits:

[0168] ;

[0169] Represents the input of the activation function.

[0170] Finally, based on the calculated fire risk level, the flame and smoke detection frequency of each area is dynamically adjusted to achieve adequate monitoring of high-risk areas, while the monitoring frequency of low-risk areas is appropriately reduced to save resources.

[0171] Specifically, in S2, based on the previous environmental monitoring and fire risk assessment, the camera is started for real-time monitoring to obtain high-quality image data, and the flame and smoke are accurately identified and located in combination with target detection technology. By visualizing the detected fire source location, this embodiment can intuitively display the current fire development in a virtual environment, providing key support for subsequent decision-making. Specifically, S2 may include:

[0172] S21. In the process of fire detection, the quality of image data directly affects the accuracy and reliability of target detection. Therefore, the monitoring images obtained by the camera are first Preprocessing is performed to remove noise from surveillance images and enhance image details. The preprocessing process includes several main steps:

[0173] Image enhancement, through the histogram equalization method, enhances the monitoring image, so that the brightness and contrast of the monitoring image can be improved, thereby improving the recognizability of the target area. Specifically, suppose the pixel intensity distribution of the monitoring image is , the image intensity distribution after histogram equalization is :

[0174] ;

[0175] in, q represents the pixel intensity value, is the integration variable.

[0176] Denoising: Since noise is likely to appear in images taken by cameras under low light conditions, the present invention uses denoising techniques such as median filtering or Gaussian filtering to eliminate random noise in the image and enhance the clarity of the target edge. It can be expressed as:

[0177] ;

[0178] in, is the Gaussian kernel function, is the size of the filter window; is the pixel coordinate.

[0179] Color correction: In order to eliminate the color cast of the camera under different lighting conditions, the white balance correction method is used to adjust the color distribution of the image to make it more consistent with the color characteristics of the natural environment. Image after color correction It can be expressed as:

[0180] ;

[0181] in, is the reference white value, is the white value in the image.

[0182] In summary, the above preprocessing will obtain As a preprocessed image ,The preprocessed image has a higher signal-to-noise ratio and contrast, making it suitable for subsequent target detection tasks.

[0183] S22, Application of flame and smoke target detection model, preprocessed image is input into the object detection model In order to identify and locate the flame and smoke areas in the image. This object detection model is based on the deep learning architecture of convolutional neural network. The output of the model includes the bounding boxes of the flame area and the smoke area:

[0184] ;

[0185] in, is a collection of multiple bounding boxes, each of which A target detected in the corresponding image is defined as:

[0186] ;

[0187] in, are the vertex coordinates of the bounding box, is the classification probability of the target (such as fire or smoke).

[0188] S23, environmental visualization of the fire source location and fire range. After the flame and smoke are detected and located, the detection results are further projected into a two-dimensional virtual environment for visualization. Assume that the coordinate system in the virtual environment is , a collection of bounding boxes for flames and smoke Through the projection transformation matrix Mapped to the corresponding location in the virtual environment :

[0189] ;

[0190] in, is the projection transformation matrix, defined as:

[0191] ;

[0192] in, and is the scaling factor, which is used to adjust the ratio between the real coordinates and the virtual environment coordinates; and is the translation factor, which is used to adjust the starting position of the coordinates. Through projection transformation, the detection results of flame and smoke can be intuitively displayed in the virtual environment.

[0193] The fire expansion situation is displayed. The fire visualization is not limited to the current fire source location, but also includes the real-time display of the fire range. The fire source location is the center point of the mapped bounding box. Indicates that the fire range is a radius of Circle express:

[0194] ;

[0195] Fire Range Calculated according to the combustion model, it reflects the expansion of the fire within a certain period of time.

[0196] Reference Figure 4 , Figure 4 This is a reference flow chart for the above S3~S4.

[0197] Specifically, in S3, in order to generate the optimal fire extinguishing strategy, real-time data in the virtual environment and the pre-trained reinforcement learning model are used to make decisions. First, the state data in the virtual environment is updated to ensure that it accurately reflects the actual situation. Then, based on the updated state data, the pre-trained reinforcement learning model is called to generate the optimal fire extinguishing strategy to provide guidance for the agent's actions.

[0198] S31. The state data in the virtual environment is updated in real time. In order to make the virtual environment accurately reflect the fire situation in reality, the data provided by the fire detection module and the environmental monitoring module are synchronized with the grid state in the virtual environment. Indicates that each grid Corresponding to a small part of the outdoor monitoring area. In the initial state, the state of all grids is set to normal state , indicating no fire and low risk:

[0199] ;

[0200] in, and are the number of rows and columns of the grid, respectively.

[0201] In order to make the virtual environment accurately reflect the current fire situation, the grid status of different areas is dynamically adjusted according to the data provided by the fire detection module and the environmental monitoring module. The grid status can be divided into the following types:

[0202]

[0203] The state matrix G(t) of the virtual environment is updated according to the following rules:

[0204] ;

[0205] The state matrix of the virtual environment over time It will be continuously updated to ensure that the system can reflect the expansion trend of the fire and changes in the environment in real time.

[0206] S32, based on the reinforcement learning model, the optimal extinguishing strategy is generated. After completing the preprocessing of the environmental data, the present invention uses the pre-trained reinforcement learning model Analyze the current state of the environment and generate the optimal fire extinguishing strategy. The reinforcement learning model optimizes the behavior of the agent through the policy gradient method, with the goal of maximizing the cumulative reward 𝑅. The formula is as follows:

[0207] ;

[0208] in, For time Instant rewards at any time, is the discount factor. The model is based on the environmental state Input, output each agent's action strategy in the current state :

[0209] ;

[0210] in, is the current state, is the set of actions that the agent can perform. The reinforcement learning model updates the strategy through the policy gradient method, specifically:

[0211] ;

[0212] in, are model parameters, which are optimized through back propagation to obtain the optimal fire extinguishing strategy.

[0213] Specifically, in S4, after the optimal extinguishing strategy is generated, the extinguishing task is executed by multiple agents in a collaborative manner, and the execution effect is verified. The strategy is assigned to each agent, and the agent operates according to the strategy instructions, while the execution of the task is monitored in real time, and necessary adjustments are made based on feedback to improve the effectiveness of the extinguishing action. Specifically, S4 may include the following steps:

[0214] S41. Assign tasks to each agent based on the optimal fire extinguishing strategy generated by the pre-trained reinforcement learning model. Assume that Agents, each agent The status includes the location , remaining resources , and the current task status . This state is described by the following equation:

[0215] ;

[0216] Based on the optimal strategy generated Each intelligent agent Assigned to different extinguishing tasks, including but not limited to locating the fire source, controlling the spread of the fire, and monitoring the re-ignition of the extinguished area. The goal of task allocation is to minimize the area and time of fire spread while maximizing the efficiency of resource use.

[0217] At each time step , for example, the operator manually operates each agent according to the task instructions , in order to execute the instructions in the strategy. The operator's decision continuously affects task adjustment and optimization through the feedback mechanism.

[0218] S42, real-time monitoring and dynamic adjustment. During the process of the intelligent agent performing the extinguishing task, the system monitors the changes of the fire and the task execution status of each intelligent agent in real time through the environmental monitoring module and the fire detection module. The monitoring data includes the flame position, fire range, and intelligent agent resource consumption. Assume that the environmental state at time 𝑡 is , the agent state is , the monitoring data can be expressed as:

[0219] ;

[0220] Based on real-time monitoring data , for strategy Dynamic adjustments are made to respond to changes in fire conditions and unexpected situations. Adjusted strategies The optimization will be performed according to the new environment and agent status, with the goal of continuing to maintain the best extinguishing effect:

[0221] ;

[0222] in, It is an instant reward based on the current monitoring data.

[0223] S43, data recording and effect verification, during the entire extinguishing process, each time step can be recorded Key data, including the action path of the agent , Resource consumption , the spread of fire etc. The recorded data is used to evaluate the effectiveness of the eradication strategy and generate a detailed summary report after the eradication mission is completed.

[0224] The summary report calculates the comprehensive effect of the extinguishing through the following formula:

[0225] Performance indicators ;

[0226] in, and is the weight coefficient, balancing the relationship between extinguishing effect and resource consumption.

[0227] The recorded data will not only be used for summary analysis of current tasks, but will also be used to train and optimize future reinforcement learning models to further improve the efficiency and accuracy of fire fighting.

[0228] The beneficial effects of the method of this embodiment include: by deploying an environmental monitoring module and a sensor network, the environmental parameters of the outdoor area are collected and analyzed in real time, and the fire risk level of each area can be accurately assessed; combined with the target detection method, accurate identification of flames and smoke is achieved, and the fire is visualized in a virtual environment, which helps to grasp the development of the fire in real time; the pre-trained reinforcement learning model is used to generate the optimal fire extinguishing strategy, and the dynamic decision-making strategy of multiple agents is combined to achieve efficient execution of the extinguishing task; finally, through real-time monitoring and data feedback, the extinguishing effect is evaluated and optimized, forming a closed-loop fire management system, which effectively improves the intelligence level of outdoor fire monitoring and fighting.

[0229] Reference Figure 5 The embodiment of the present application also provides a system for intelligent disaster monitoring and decision-making based on target detection, which can implement the above-mentioned method for intelligent disaster monitoring and decision-making based on target detection. The system includes:

[0230] A parameter acquisition unit, used to acquire current environmental parameters of the monitoring area;

[0231] an object detection unit, configured to acquire a monitoring image of the monitoring area and detect flames and smoke in the monitoring image;

[0232] A target monitoring unit is used to project the detected flames and smoke into a virtual environment for visual display;

[0233] A parameter updating unit, configured to update a virtual state parameter of the virtual environment according to the current environment parameter;

[0234] A strategy generating unit, configured to generate a fire extinguishing strategy according to current virtual state parameters of the virtual environment;

[0235] The strategy allocation unit is used to allocate the fire extinguishing strategy to each intelligent agent so that each intelligent agent executes the fire extinguishing strategy.

[0236] It can be understood that the contents of the above method embodiments are all applicable to the present system embodiments, the functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0237] The embodiment of the present application also provides an electronic device, the electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned method of intelligent disaster monitoring and decision-making based on target detection when executing the computer program. The electronic device can be any intelligent terminal including a tablet computer, a car computer, etc.

[0238] It can be understood that the contents of the above method embodiments are all applicable to the present device embodiments, the functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0239] See also Figure 6 , Figure 6 The hardware structure of an electronic device of another embodiment is illustrated, and the electronic device includes:

[0240] The processor 601 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0241] The memory 602 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 602 can store an operating system and other application programs. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 602, and the processor 601 calls and executes a method for intelligent disaster monitoring and decision-making based on target detection in the embodiment of this application;

[0242] Input / output interface 603, used to implement information input and output;

[0243] Communication interface 604, used to realize communication interaction between the device and other devices, which can be realized through wired mode (such as USB, network cable, etc.) or wireless mode (such as mobile network, WIFI, Bluetooth, etc.);

[0244] A bus 605 that transmits information between various components of the device (e.g., processor 601, memory 602, input / output interface 603, and communication interface 604);

[0245] The processor 601 , the memory 602 , the input / output interface 603 and the communication interface 604 are connected to each other in communication within the device via a bus 605 .

[0246] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the above-mentioned method of intelligent disaster monitoring and decision-making based on target detection.

[0247] It can be understood that the contents of the above method embodiments are all applicable to the present storage medium embodiments, the functions specifically implemented by the present storage medium embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0248] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0249] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0250] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0251] The system embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment.

[0252] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0253] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0254] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0255] In the several embodiments provided in the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of systems or units, which can be electrical, mechanical or other forms.

[0256] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0257] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0258] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0259] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention should be within the scope of the rights of the present invention.

Claims

1. A method for intelligent disaster monitoring and decision-making based on target detection, characterized in that: The method comprises the following steps: Get the current environmental parameters of the monitoring area; Acquire a monitoring image of the monitoring area, and detect flames and smoke in the monitoring image; Project the detected flames and smoke into a virtual environment for visualization; Updating virtual state parameters of the virtual environment according to the current environment parameters; generating a fire extinguishing strategy according to current virtual state parameters of the virtual environment; Distributing the fire extinguishing strategy to each intelligent agent so that each intelligent agent executes the fire extinguishing strategy; The generating of the fire extinguishing strategy according to the current virtual state parameters of the virtual environment comprises the following steps: Generate the fire extinguishing strategy according to the current virtual state parameters of the virtual environment using a reinforcement learning model; The expression of the reinforcement learning model is: ； in, For the fire fighting strategy, is the current state of the agent, is a set of actions that the agent can perform; represents the reinforcement learning model, is the current virtual state parameter; Using the reinforcement learning model to update the fire fighting strategy through a policy gradient method and maximizing cumulative rewards; The expression of the policy gradient method is: ； in, are parameters of the reinforcement learning model; To maximize the cumulative reward, ; for Instant rewards at any time, is the discount factor; The total time taken to extinguish the fire; After the fire extinguishing strategy is distributed to each intelligent agent so that each intelligent agent executes the fire extinguishing strategy, the method further comprises the following steps: Acquire current monitoring data of the monitoring area; wherein the monitoring data includes environmental status and agent status; Dynamically adjust the fire extinguishing strategy according to the current monitoring data; The expression for adjusting the fire extinguishing strategy is: ； in, for the adjusted fire extinguishing strategy; is the monitoring data, ; is the environmental state, For the N The intelligent agent The state of the agent at time .

2. The method for intelligent disaster monitoring and decision-making based on target detection according to claim 1 is characterized in that: The step of obtaining the current environmental parameters of the monitoring area includes the following steps: Acquire current environmental data of the monitoring area through sensors; Using a machine learning algorithm model to combine the current environmental data with the historical environmental data of the monitored area to obtain the current environmental parameters; The expression of the current environment parameter is: ； in, is the current environmental parameter, is the current environment data, The historical environmental data; Representing the machine learning algorithm model; ； in, and is the weight matrix, is the bias term.

3. The method for intelligent disaster monitoring and decision-making based on target detection according to claim 2 is characterized in that: The method further comprises the following steps: Determining the fire risk level of the monitored area according to the current environmental parameters and the current environmental data using a fire risk assessment model; The expression of the fire risk assessment model is: ； in, is the fire risk assessment model; is a nonlinear function based on the historical environmental data, used to describe the monitoring area Historical risk trends; is a function based on the current environmental data, used to reflect the current fire risk; is the weight coefficient; is the activation function; The execution frequency of detecting the flame and smoke in the monitoring image is dynamically adjusted according to the fire risk level.

4. The method for intelligent disaster monitoring and decision-making based on target detection according to claim 1, characterized in that: The detecting of flames and smoke in the monitoring image comprises the following steps: The monitoring image is enhanced by using a histogram equalization method to improve the brightness and contrast of the monitoring image to obtain an enhanced image; The expression of the histogram equalization method is: ； in, q represents the pixel intensity value, is the integration variable; is the pixel intensity distribution of the monitoring image; is the pixel intensity distribution after being processed by the histogram equalization method; Using median filtering or Gaussian filtering to filter the enhanced image to obtain a filtered image; The expression of the Gaussian filter is: ； in, is the filtered image; is the pixel coordinate; is the Gaussian kernel function, is the size of the filter window; is the enhanced image; Adjusting the color distribution of the filtered image using a white balance correction method to obtain a color-corrected image; The expression of the white balance correction method is: ； in, For the color-corrected image, is the reference white value, is the white value in the filtered image; The flame and smoke in the color-corrected image are detected using an object detection model to obtain a set of bounding box parameters.

5. The method for intelligent disaster monitoring and decision-making based on target detection according to claim 1 is characterized in that: The step of projecting the detected flame and smoke into a virtual environment for visual display comprises the following steps: Projecting the detected flame and smoke bounding box parameter set to the virtual environment through a projection transformation matrix; The fire source location and fire range are visualized in the virtual environment according to the bounding box parameter set.

6. A system for intelligent disaster monitoring and decision-making based on target detection, characterized in that: For implementing the method of disaster intelligent monitoring and decision-making based on target detection as claimed in claim 1, the system comprises: A parameter acquisition unit, used to acquire current environmental parameters of the monitoring area; A target detection unit, used to obtain a monitoring image of the monitoring area and detect flames and smoke in the monitoring image; A target monitoring unit is used to project the detected flames and smoke into a virtual environment for visual display; A parameter updating unit, configured to update a virtual state parameter of the virtual environment according to the current environment parameter; A strategy generating unit, configured to generate a fire extinguishing strategy according to current virtual state parameters of the virtual environment; The strategy allocation unit is used to allocate the fire extinguishing strategy to each intelligent agent so that each intelligent agent executes the fire extinguishing strategy.

7. An electronic device, characterized in that: The electronic device comprises a memory and a processor, the memory stores a computer program, and the processor implements the method according to any one of claims 1 to 5 when executing the computer program.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images