Multi-spectrum fire point intelligent detection and linkage alarm system based on Internet platform

Abstract

The invention relates to the technical field of intelligent alarm, and discloses a multi-spectrum fire point intelligent detection and linkage alarm system based on an internet platform. The system comprises a thermal radiation feature extraction module, a spatial distribution feature analysis module, a heterogeneous feature segmentation module, a multi-dimensional risk factor evaluation module, a linkage strategy mapping module and a linkage strategy response detection module, thermal radiation feature analysis is performed on multi-spectrum image data of a target area, and a thermal radiation feature vector is obtained; performing spatial distribution feature analysis on the multi-spectrum image data to obtain a spatial distribution feature vector; carrying out heterogeneous feature segmentation on the multi-spectrum image data to obtain a fire point region; performing multi-dimensional risk factor assessment on the environmental parameters of the fire point area to obtain a risk level; performing strategy mapping on the risk level to obtain a linkage strategy; performing response detection on the linkage strategy to obtain a linkage alarm report; according to the invention, the efficiency of multi-spectrum fire point intelligent detection and linkage alarm of the Internet platform can be improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent alarm technology, and in particular to a multi-spectral fire point intelligent detection and linkage alarm system based on an Internet platform. Background Technology

[0002] Existing fire detection technologies mostly rely on single-spectrum data for analysis, which cannot simultaneously and completely capture the thermal radiation characteristics and spatial distribution characteristics of the target area. This leads to omissions or misjudgments in the identification of potential fire areas, making it difficult to accurately separate the true fire areas.

[0003] Existing technologies do not fully integrate key environmental factors in the fire risk assessment process, resulting in a lack of comprehensiveness and accuracy in risk level determination. Furthermore, the linkage alarm strategy is not precisely adapted to the risk level and actual regional characteristics, leading to a lack of targeted alarm response, an inability to coordinate and adapt resources in a timely manner to execute effective countermeasures, and difficulty in achieving efficient collaboration between fire detection and alarms. Summary of the Invention

[0004] This invention provides a multi-spectral fire point intelligent detection and linkage alarm system based on an Internet platform to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a multi-spectral fire point intelligent detection and linkage alarm system based on an internet platform, characterized in that the system includes a thermal radiation feature extraction module, a spatial distribution feature analysis module, a heterogeneous feature segmentation module, a multi-dimensional risk factor assessment module, a linkage strategy mapping module, and a linkage strategy response detection module, wherein:

[0006] The thermal radiation feature extraction module is used to perform thermal radiation feature analysis on the multi-spectral image data of the target area to obtain the thermal radiation feature vector of the target area.

[0007] The spatial distribution feature parsing module is used to perform spatial distribution feature parsing on the multi-spectral image data to obtain the spatial distribution feature vector of the target region;

[0008] The heterogeneous feature segmentation module is used to perform heterogeneous feature segmentation on multi-spectral image data based on the thermal radiation feature vector and the spatial distribution feature vector to obtain the fire point region of the target region.

[0009] The multidimensional risk factor assessment module is used to perform multidimensional risk factor assessment on the environmental parameters of the fire point area to obtain the risk level of the target area.

[0010] The linkage strategy mapping module is used to map the risk level to a preset linkage strategy library to obtain the linkage strategy for the target area.

[0011] The linkage strategy response detection module is used to detect the response of the linkage strategy and obtain the linkage alarm report of the target area.

[0012] In a preferred embodiment, when the thermal radiation feature extraction module performs thermal radiation feature analysis on the multispectral image data of the target area to obtain the thermal radiation feature vector of the target area, it is specifically used for:

[0013] Infrared band separation is performed on the multispectral image data of the target area to obtain the infrared radiation data of the target area;

[0014] The infrared radiation data is calibrated to obtain standardized radiation intensity information for the target area;

[0015] The standardized radiation intensity information is used to construct feature primitives to obtain the radiation feature primitive set of the target region;

[0016] Multi-scale feature integration is performed on the radiation feature primitive set to obtain the thermal radiation feature vector of the target region.

[0017] In a preferred embodiment, when the spatial distribution feature parsing module performs spatial distribution feature parsing on the multispectral image data to obtain the spatial distribution feature vector of the target region, it is specifically used for:

[0018] The multi-spectral image data is spectrally decoupled to obtain visible light image data of the target region;

[0019] Texture feature recognition is performed on the visible light image data to obtain texture pattern information of the target region;

[0020] The texture pattern information is subjected to morphological topological characterization to obtain the spatial distribution feature vector of the target region.

[0021] In a preferred embodiment, when the spatial distribution feature parsing module performs texture feature recognition on the visible light image data to obtain texture pattern information of the target region, it is specifically used for:

[0022] The visible light image data is analyzed for texture orientation to obtain the texture orientation distribution of the target region;

[0023] The spatial density features of the texture direction distribution are integrated to obtain the texture density distribution features of the target region.

[0024] Multi-scale feature fusion is performed on the texture direction distribution and texture density distribution features to obtain the texture pattern information of the target region.

[0025] In a preferred embodiment, when the heterogeneous feature segmentation module performs heterogeneous feature segmentation on the multispectral image data based on the thermal radiation feature vector and the spatial distribution feature vector to obtain the fire point region of the target region, it is specifically used for:

[0026] By performing heterogeneous feature association mapping on the thermal radiation feature vector and the spatial distribution feature vector, the feature mapping relationship between thermal radiation and spatial distribution corresponding to the target region is obtained;

[0027] Based on the feature mapping relationship, the multi-spectral image data is subjected to spectral-spatial feature co-analysis to obtain the potential fire point region of the target area;

[0028] The potential fire point region is optimized by thermal radiation profile to obtain the fire point region of the target region.

[0029] In a preferred embodiment, when the multidimensional risk factor assessment module performs a multidimensional risk factor assessment of the environmental parameters of the fire point area to obtain the risk level of the target area, it is specifically used for:

[0030] The environmental parameters include: wind speed and direction factor, atmospheric humidity factor, and vegetation cover factor of the target area;

[0031] Based on the wind speed and direction factors, the fire spread trend of the fire point area is analyzed to obtain a dynamic spread description of the target area.

[0032] Based on the atmospheric humidity factor and the vegetation cover factor, the flammability level of the fire point area is determined to obtain a static fire hazard description of the target area.

[0033] The risk level of the target area is obtained by fusing the dynamic spread description and the static fire hazard description.

[0034] In a preferred embodiment, the risk level is calculated using the following formula:

[0035] ;

[0036] In the formula, For the aforementioned risk level, This is the fire spread intensity coefficient describing the dynamic spread. The wind field enhancement factor described by the dynamic spread, This refers to the terrain suppression factor describing the dynamic spread. This is the combustible material loading factor for the static fire hazard description. The dryness index is used to describe the static fire hazard.

[0037] In a preferred embodiment, the step of obtaining the risk level further includes:

[0038] Based on preset multi-level risk thresholds, the fire zone is mapped to a risk range to obtain the preliminary risk level of the target area.

[0039] When the initial risk level is within the boundary region of the multi-level risk threshold, trend feature analysis is performed on the dynamic spread description to obtain the risk evolution direction feature;

[0040] Based on the static fire risk description, the risk evolution direction characteristics are arbitrated to obtain the risk level arbitration result of the target area.

[0041] The risk level arbitration result is output as the risk level of the target area.

[0042] In a preferred embodiment, when the linkage strategy mapping module performs strategy mapping on the risk level based on a preset linkage strategy library to obtain the linkage strategy for the target area, it is specifically used for:

[0043] Based on a preset linkage strategy library, the basic strategy corresponding to the risk level is matched;

[0044] Based on the regional topological features of the fire point area, the basic strategy is parsed to obtain the strategy set of the target area;

[0045] Based on the intensity characteristics of the thermal radiation feature vector, the strategy set is configured with strategy parameters to obtain the linkage strategy for the target area.

[0046] In a preferred embodiment, when the linkage strategy response detection module performs response detection on the linkage strategy and obtains a linkage alarm report for the target area, it is specifically used for:

[0047] The linkage strategy is decomposed into strategy elements to obtain the alarm object set, action command set and resource allocation scheme of the target area;

[0048] The alarm object set is authenticated to obtain the alarm object list of the target area;

[0049] The feasibility of the action instruction set is determined to obtain the action instruction sequence for the target area;

[0050] The resource allocation plan is checked for resource status to obtain a list of available resources in the target area;

[0051] Perform multi-dimensional element integration on the list of alarm objects, the sequence of action instructions, and the list of available resources to obtain the linkage alarm report for the target area.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] 1. Through the synergistic effect of multi-spectrum data fusion and heterogeneous feature segmentation, the present invention significantly improves the accuracy and reliability of fire point recognition. By deeply integrating the thermal radiation feature vector and the spatial distribution feature vector, and combining the multi-dimensional risk factor assessment of dynamic spread description and static fire risk description, it realizes the full-chain precision processing of the fire point area from feature extraction to risk determination. This method effectively enhances the sensitivity and specificity of fire point recognition in complex environments through spatio-temporal continuity analysis and adaptive weight allocation mechanism. At the same time, through feature-level fusion and boundary topology reconstruction, it ensures the accurate definition of the contour of the fire point area.

[0054] 2. At the level of linkage alarm, based on the risk level mapping and policy parsing mechanism of the preset policy library, the present invention realizes the seamless connection from risk perception to emergency response. Through the spatial adaptability analysis and parameter dynamic configuration of policy elements, highly operable linkage instructions are generated; and then through the response detection link to systematically verify the alarm objects, action instructions and resource list, it ensures that the generated linkage alarm report has completeness and executability. The whole set of methods forms a closed-loop technical system from fire point detection to risk warning and then to linkage response, greatly improving the timeliness of warning response and the effectiveness of emergency disposal. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 FIG. is a system architecture diagram of a multi-spectrum fire point intelligent detection and linkage alarm system based on an Internet platform provided by an embodiment of the present invention;

[0056] The realization, functional characteristics and advantages of the object of the present invention will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments belong to a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0058] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “said” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0059] Depending on the context, the word "if" or "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0060] Furthermore, the timing of the steps in the following method embodiments is merely an example and not a strict limitation.

[0061] In practice, the server-side equipment deployed in a multi-spectral fire detection and linkage alarm system based on an internet platform may consist of one or more devices. This system can be implemented as a business instance, a virtual machine, or hardware devices. For example, it can be implemented as a business instance deployed on one or more devices in a cloud node. Simply put, it can be understood as software deployed on a cloud node to provide multi-spectral fire detection and linkage alarm services to various user terminals. Alternatively, it can be implemented as a virtual machine deployed on one or more devices in a cloud node, with application software installed to manage various user terminals. Or, it can also be implemented as a server-side system composed of numerous identical or different types of hardware devices, with one or more devices configured to provide multi-spectral fire detection and linkage alarm services to various user terminals.

[0062] In terms of implementation, the multi-spectral fire detection and linkage alarm system based on the internet platform and the user terminal are mutually compatible. That is, if the multi-spectral fire detection and linkage alarm system based on the internet platform is implemented as an application installed on a cloud service platform, then the user terminal is a client that establishes a communication connection with the application; or if the multi-spectral fire detection and linkage alarm system based on the internet platform is implemented as a website, then the user terminal is implemented as a webpage; or if the multi-spectral fire detection and linkage alarm system based on the internet platform is implemented as a cloud service platform, then the user terminal is implemented as a mini-program in an instant messaging application.

[0063] like Figure 1 The diagram shown is a system architecture diagram of a multi-spectral fire point intelligent detection and linkage alarm system based on an Internet platform provided in an embodiment of the present invention.

[0064] The multi-spectral fire detection and linkage alarm system 100 based on an internet platform described in this invention can be installed on a cloud server. In terms of implementation, it can be used as one or more service devices, or as an application installed on the cloud (e.g., a mobile service operator's server, server cluster, etc.), or it can be developed into a website. Depending on the functions implemented, the multi-spectral fire detection and linkage alarm system 100 based on an internet platform may include a thermal radiation feature extraction module 101, a spatial distribution feature analysis module 102, a heterogeneous feature segmentation module 103, a multi-dimensional risk factor assessment module 104, a linkage strategy mapping module 105, and a linkage strategy response detection module 106. The modules described in this invention can also be called units, referring to a series of computer program segments that can be executed by the processor of an electronic device and perform a fixed function, stored in the memory of the electronic device.

[0065] In this embodiment of the invention, in the multi-spectral fire detection and linkage alarm system based on an internet platform, each of the above modules can be implemented independently and can call other modules. Here, "calling" can be understood as one module connecting to multiple modules of another type and providing corresponding services to those connected modules. In the multi-spectral fire detection and linkage alarm system based on an internet platform provided by this embodiment of the invention, the applicable scope of the system architecture can be adjusted by adding modules and directly calling them without modifying the program code, achieving cluster-based horizontal expansion to quickly and flexibly expand the multi-spectral fire detection and linkage alarm system based on an internet platform. In practical applications, the above modules can be set in the same device or different devices, or they can be set in virtual devices, such as service instances in a cloud server.

[0066] The following describes, with reference to specific embodiments, each component and its specific workflow of the multi-spectral fire detection and linkage alarm system based on an internet platform:

[0067] The thermal radiation feature extraction module 101 is used to perform thermal radiation feature analysis on the multi-spectral image data of the target area to obtain the thermal radiation feature vector of the target area.

[0068] In this embodiment of the invention, when the thermal radiation feature extraction module performs thermal radiation feature analysis on the multispectral image data of the target area to obtain the thermal radiation feature vector of the target area, it is specifically used for:

[0069] Infrared band separation is performed on the multispectral image data of the target area to obtain the infrared radiation data of the target area;

[0070] The infrared radiation data is calibrated to obtain standardized radiation intensity information for the target area;

[0071] The standardized radiation intensity information is used to construct feature primitives to obtain the radiation feature primitive set of the target region;

[0072] Multi-scale feature integration is performed on the radiation feature primitive set to obtain the thermal radiation feature vector of the target region.

[0073] Specifically, when performing infrared band separation on multispectral image data of the target area, a band filtering device is required. This device pre-sets the wavelength range corresponding to the infrared band and filters out all non-infrared signals such as visible light and ultraviolet light contained in the multispectral image data, retaining only the signals whose wavelengths are within the preset infrared band range. These retained signals are the infrared radiation data of the target area. This data is the core initial data for subsequent thermal radiation characteristic analysis, laying the foundation for the alarm device to accurately capture the thermal radiation information of the fire point.

[0074] Furthermore, when calibrating the intensity of infrared radiation data, a standard radiation source calibrated by an authoritative institution is first selected. The radiation intensity value of this standard radiation source is known and fixed under specific environmental conditions. The radiation intensity corresponding to the infrared radiation data is compared point by point with the known radiation intensity of the standard radiation source, and the intensity difference of each data point is recorded. Then, each intensity data in the infrared radiation data is adjusted according to these differences to ensure that the intensity of the adjusted infrared radiation data is consistent with the intensity benchmark of the standard radiation source. The adjusted infrared radiation data is the standardized radiation intensity information of the target area. This information can eliminate the interference of different measurement environments on the radiation intensity detection results, ensure the accuracy of subsequent feature extraction, and support the alarm device to reliably judge the thermal radiation intensity of the fire point.

[0075] Furthermore, when constructing feature primitives for standardized radiation intensity information, the image region corresponding to the standardized radiation intensity information is first divided into multiple sub-regions of the same size according to a fixed spatial grid rule. For each sub-region, the highest, lowest, and average values ​​of the standardized radiation intensity within that sub-region, as well as the trend of intensity variation with spatial location, are extracted. These information corresponding to each sub-region together form an independent feature unit. All feature units corresponding to all sub-regions are collected and organized to form the radiation feature primitive set of the target region. This primitive set contains the specific feature details of standardized radiation intensity at different spatial locations, providing refined feature materials for subsequent multi-scale feature integration, and helping the alarm device to identify the distribution pattern of fire point thermal radiation from a local level.

[0076] Furthermore, when performing multi-scale feature integration on the radiation feature primitive set, two observation scales, local and global, are first determined. At the local scale, the intensity correlation and commonalities in distribution among radiation feature primitives corresponding to adjacent sub-regions are analyzed. At the global scale, the intensity distribution trend of all radiation feature primitives in the entire target area and the concentration range of high-radiation-intensity primitives are analyzed. Then, the correlation and commonalities information obtained at the local scale and the distribution trend and concentration range information obtained at the global scale are fused. During the fusion process, it is ensured that local details are not lost and the overall macroscopic information is complete. The fused result is the thermal radiation feature vector of the target area. This vector can comprehensively reflect the overall and local characteristics of thermal radiation in the target area, providing key feature basis for subsequent fire point area identification and risk assessment, and directly serving the realization of fire point detection and linkage alarm functions of the alarm device.

[0077] In summary, when performing infrared band separation on multispectral image data of the target area, non-infrared band signals are removed by band filtering, and infrared band signals are retained to finally obtain infrared radiation data. This data carries the initial information of thermal radiation, provides core material for subsequent thermal radiation feature extraction, and supports the accurate implementation of fire detection and linkage alarm functions.

[0078] In summary, when calibrating the intensity of infrared radiation data, the standard radiation intensity calibrated by an authoritative authority is used as the benchmark. The intensity values ​​of each point in the data are compared with the benchmark, and adjustments are made to eliminate environmental and equipment deviations to ensure that the data meets the standard. Finally, standardized radiation intensity information is obtained, avoiding interference from non-objective factors, ensuring the accuracy of subsequent feature extraction, and providing reliable intensity data support for fire point identification.

[0079] In summary, when constructing feature primitives for standardized radiation intensity information, the corresponding image region is divided into independent sub-regions according to fixed spatial rules. Key information such as the intensity extremes, average changes, and spatial distribution of each sub-region is extracted, integrated into independent feature units, and summarized to obtain a set of radiation feature primitives. This refines the details of intensity features, provides specific units for multi-scale feature analysis, and helps to accurately capture the thermal radiation features of fire points.

[0080] In summary, when performing multi-scale feature integration on the radiation feature primitive set, analysis is conducted at both local and global scales: local analysis examines the intensity correlation, commonalities, and variation patterns of adjacent primitives, while global analysis examines the overall intensity distribution trend, high radiation concentration range, and variation direction of the primitives. The fusion of the two-scale results yields a thermal radiation feature vector, comprehensively reflecting the thermal radiation characteristics. This provides a core basis for heterogeneous feature segmentation, multi-dimensional risk assessment, and the formulation of linkage strategies, thus serving the realization of system functions.

[0081] The spatial distribution feature parsing module 102 is used to perform spatial distribution feature parsing on the multi-spectral image data to obtain the spatial distribution feature vector of the target region.

[0082] In this embodiment of the invention, when the spatial distribution feature parsing module performs spatial distribution feature parsing on the multispectral image data to obtain the spatial distribution feature vector of the target region, it is specifically used for:

[0083] The multi-spectral image data is spectrally decoupled to obtain visible light image data of the target region;

[0084] Texture feature recognition is performed on the visible light image data to obtain texture pattern information of the target region;

[0085] The texture pattern information is subjected to morphological topological characterization to obtain the spatial distribution feature vector of the target region.

[0086] When the spatial distribution feature parsing module performs texture feature recognition on the visible light image data to obtain the texture pattern information of the target region, it is specifically used for:

[0087] The visible light image data is analyzed for texture orientation to obtain the texture orientation distribution of the target region;

[0088] The spatial density features of the texture direction distribution are integrated to obtain the texture density distribution features of the target region.

[0089] Multi-scale feature fusion is performed on the texture direction distribution and texture density distribution features to obtain the texture pattern information of the target region.

[0090] Specifically, when performing spectral decoupling on the multi-spectral image data, a spectral separation device is required. This device pre-sets the wavelength range corresponding to each spectral component, distinguishing the non-visible light spectral components such as infrared and ultraviolet light contained in the multi-spectral image data from the visible light spectral components. The device's internal filtering components filter out the signals corresponding to the non-visible light spectrum, retaining only the signals with wavelengths within the visible light range. These retained visible light signals are converted into an image form that can present the scene of the target area by a signal conversion component, ultimately obtaining visible light image data of the target area. This data can clearly display the spatial basic information such as the shape of objects and the division of areas within the target area, providing intuitive image basis for subsequent spatial distribution feature analysis. At the same time, it supports the multi-spectral fire point intelligent detection and linkage alarm system to accurately judge the spatial environment around the fire point, ensuring the spatial correlation of subsequent alarm actions.

[0091] Furthermore, when performing texture feature recognition on the visible light image data, the entire image corresponding to the visible light image data is first divided into multiple equal image sub-units according to fixed spatial units. For each image sub-unit, the arrangement of pixels within the sub-unit is observed through the image recognition component, and the variation pattern of pixel color depth, the arrangement density between adjacent pixels, and the clustering range of pixels of the same or similar color are recorded. Then, the differences and commonalities in pixel arrangement between different image sub-units are compared. The information such as pixel arrangement patterns, density characteristics, and clustering range of all sub-units is summarized and organized to form a set that can reflect the surface texture characteristics of objects within the target area. Finally, the texture pattern information of the target area is obtained. This information can reflect the spatial distribution differences of different objects within the target area, providing texture-level feature support for subsequent morphological topology representation, helping the system to accurately identify the spatial path of possible fire spread, and serving the targeted formulation of linkage alarm strategies.

[0092] Furthermore, when performing morphological topological characterization on the texture pattern information, the region range corresponding to different texture features in the texture pattern information is first analyzed to determine the specific location of each texture region within the target region. Then, the connection relationship between different texture regions is determined through the topological analysis component, such as the contact boundary length of adjacent texture regions and the interval distance of non-adjacent texture regions. At the same time, the shape characteristics of each texture region are recorded, such as whether it has a regular geometric shape and whether the edges are continuous. The position information, connection relationship and shape characteristics of these texture regions are transformed into a feature set that can be quantified to describe the spatial distribution. This set can comprehensively reflect the spatial arrangement law of texture features within the target region, and finally obtain the spatial distribution feature vector of the target region. This vector provides a key basis for the spatial dimension of subsequent multi-dimensional risk factor assessment, ensuring that the multi-spectral fire point intelligent detection and linkage alarm system can fully consider the spatial distribution characteristics of the target region when formulating linkage alarm strategies, thereby improving the accuracy and effectiveness of alarm response.

[0093] Specifically, when performing texture direction analysis on the visible light image data, a texture direction detection component is required. First, the entire image corresponding to the visible light image data is divided into multiple independent image sub-blocks at uniform spatial intervals. For each image sub-block, the gray value change trend of the pixels within the sub-block is observed through the gray value change sensing unit within the component. The direction of the most obvious gray value change is determined and marked as the texture direction of the corresponding sub-block. After the texture directions of all image sub-blocks are marked, the texture directions of all sub-blocks are arranged sequentially according to their spatial positions in the image, forming a set that reflects the texture direction of each part within the target area. Finally, the texture direction distribution of the target area is obtained. This distribution can clearly present the direction pattern of the object texture within the target area, providing a directional basis for subsequent judgment of the possible spread path of fire points, and supporting the multi-spectral fire point intelligent detection and linkage alarm system in making a preliminary judgment on the fire risk spread trend.

[0094] Furthermore, when integrating the spatial density features of the texture direction distribution, the target area needs to be divided into multiple statistical units of uniform size according to a fixed spatial grid. For each statistical unit, the number of all texture directions contained in the unit is counted by the density statistics component, and then the proportion of the same texture direction in the unit is calculated. If the proportion of a certain texture direction in the unit exceeds that of other directions, the ratio of the number corresponding to that direction to the unit area is determined as the texture density value of the unit. After the texture density values ​​of all statistical units have been calculated, the texture density values ​​of each unit are integrated according to their spatial location to form an information set that reflects the texture density at different locations in the target area. Finally, the texture density distribution feature of the target area is obtained. This feature can reflect the density of the objects in the target area, providing a reference for the subsequent assessment of the distribution of fuel required for fire ignition, and helping the linkage alarm system to more accurately determine the fire hazard level.

[0095] Furthermore, when performing multi-scale feature fusion on the texture direction distribution and texture density distribution features, it is necessary to operate through a multi-scale fusion component. First, starting from the local scale, a small area within the target region is selected, and the correlation between the consistency of the texture direction distribution and the density of the texture density distribution features within this small area is analyzed to determine the matching relationship between texture direction and density within the local area. Then, starting from the overall scale, the overall trend of texture direction distribution and the overall difference of texture density distribution features within the entire target region are observed. Subsequently, the direction and density matching relationship obtained at the local scale and the trend and difference information obtained at the overall scale are integrated. During the integration process, it is ensured that local detail information is not covered by overall information, while ensuring that overall regularity information is fully presented. Finally, a comprehensive information set that can fully reflect the texture direction and density of the target region is formed, and the texture pattern information of the target region is obtained. This information can completely characterize the texture features of the target region, providing a comprehensive texture basis for subsequent morphological topology representation. This ensures that the multi-spectral fire point intelligent detection and linkage alarm system can fully consider the texture-related spatial characteristics of the target region when formulating linkage strategies, thereby improving the pertinence and effectiveness of alarm response.

[0096] In summary, when performing spectral decoupling on the multi-spectral image data, the non-visible light spectral components are separated, the visible light signal is retained and converted into an image, and the visible light image data of the target area is obtained. This data clearly presents the object shape and spatial layout, providing an intuitive basis for the analysis of spatial distribution characteristics, supporting the system's judgment of the spatial environment around the fire point, and ensuring the spatial correlation of alarm actions.

[0097] In summary, when performing texture feature recognition on the visible light image data, the image is divided into sub-units, and the pixel arrangement patterns, density, and aggregation characteristics are analyzed and summarized to obtain the texture pattern information of the target area. This information reflects the spatial distribution differences of objects, provides texture support for morphological topology representation, helps identify the spatial path of fire spread, and serves the formulation of linkage alarm strategies.

[0098] In summary, when performing morphological topological characterization on the texture pattern information, the location, connectivity, and shape features of the texture regions are analyzed and transformed into a set of spatial distribution features to obtain the spatial distribution feature vector of the target region. This provides a spatial dimension basis for multidimensional risk factor assessment, ensuring that the system considers spatial distribution characteristics when formulating strategies and improving the accuracy of alarm response.

[0099] In summary, when performing texture direction analysis on the visible light image data, the data is divided into independent sub-blocks. The direction in which the pixel grayscale change is most obvious in each sub-block is detected and marked. The direction information is arranged according to spatial position to obtain the texture direction distribution of the target area. This distribution presents a regular texture direction, providing a directional basis for the preliminary judgment of the fire spread path and supporting subsequent texture feature recognition.

[0100] In summary, when integrating the spatial density features of the texture direction distribution, statistical units are divided according to a fixed spatial grid. The number and proportion of texture directions are counted to determine the density value and integrate them to obtain the texture density distribution features of the target area. This feature reflects the density of the object distribution, provides a reference for assessing fuel distribution, and helps the system grasp the spatial characteristics of fire hazards.

[0101] In summary, when performing multi-scale feature fusion on the texture direction distribution and texture density distribution features, the matching relationship between the two is analyzed locally, and the trends and differences are analyzed overall. By integrating information from all levels, the texture pattern information of the target area is obtained. This information fully characterizes the texture features, provides a comprehensive basis for morphological topology representation, ensures that the system considers texture-related characteristics when formulating strategies, and improves the pertinence of alarm response.

[0102] The heterogeneous feature segmentation module 103 is used to perform heterogeneous feature segmentation on multi-spectral image data based on the thermal radiation feature vector and the spatial distribution feature vector to obtain the fire point region of the target region.

[0103] In this embodiment of the invention, when the heterogeneous feature segmentation module performs heterogeneous feature segmentation on the multispectral image data based on the thermal radiation feature vector and the spatial distribution feature vector to obtain the fire point region of the target region, it is specifically used for:

[0104] By performing heterogeneous feature association mapping on the thermal radiation feature vector and the spatial distribution feature vector, the feature mapping relationship between thermal radiation and spatial distribution corresponding to the target region is obtained;

[0105] Based on the feature mapping relationship, the multi-spectral image data is subjected to spectral-spatial feature co-analysis to obtain the potential fire point region of the target area;

[0106] The potential fire point region is optimized by thermal radiation profile to obtain the fire point region of the target region.

[0107] Specifically, when performing heterogeneous feature association mapping between the thermal radiation feature vector and the spatial distribution feature vector, a feature association component is required. First, feature segments representing high radiation intensity are extracted from the thermal radiation feature vector. Then, feature segments representing specific texture distributions are extracted from the spatial distribution feature vector. A one-to-one correspondence is established between the high radiation intensity feature segments and the specific texture distribution feature segments at the corresponding spatial locations. The texture distribution feature attributes corresponding to each high radiation feature segment in space are recorded. At the same time, the overlapping areas and association strengths of the two in spatial locations are marked. All correspondences, overlapping areas, and association strengths are integrated into a structured set of association information. Finally, the feature mapping relationship between thermal radiation and spatial distribution corresponding to the target area is obtained. This relationship can establish a direct correlation between thermal radiation characteristics and spatial texture characteristics, providing a feature association basis for subsequent accurate identification of fire point areas and supporting the comprehensive judgment of fire point characteristics by the multi-spectral fire point intelligent detection and linkage alarm system.

[0108] Furthermore, based on the aforementioned feature mapping relationship, when performing spectral-spatial feature collaborative analysis on the multi-spectral image data, a collaborative analysis component is required. First, spectral signals related to thermal radiation features and spatial image signals related to spatial distribution features are separated from the multi-spectral image data. According to the association rules between high radiation and specific textures in the feature mapping relationship, spectral regions that meet the high radiation intensity threshold are selected from the spectral signals, and spatial regions that meet the specific texture distribution are selected from the spatial image signals. The selected high radiation spectral regions and specific texture spatial regions are compared in spatial position, and the regions where their spatial positions overlap are retained. These overlapping regions are the regions where fire points may exist. All overlapping regions are integrated and marked to finally obtain the potential fire point region of the target area. This region provides a preliminary positioning range for fire point identification, avoids misjudgment caused by single feature judgment, and narrows the fire point investigation range for the linkage alarm system.

[0109] Furthermore, when optimizing the thermal radiation profile of the potential fire point area, a thermal radiation profile calibration component is required. First, real-time thermal radiation intensity information of each location within the potential fire point area is collected. The differences in thermal radiation intensity at different locations within the area are compared, and edge locations with thermal radiation intensity lower than the minimum radiation standard for fire points are excluded. Core locations with stable thermal radiation intensity that conform to the radiation characteristics of fire points are retained. Then, the boundary of the core location is adjusted by the profile correction unit to eliminate irregular parts of the boundary caused by image noise, so that the profile of the core area presents a continuous and clear shape. At the same time, the boundary coordinates and internal area range of the optimized profile are marked. The optimized core area and boundary information are integrated to finally obtain the fire point area of ​​the target area. This area can accurately lock the actual range of the fire point, ensuring that the multi-spectral fire point intelligent detection and linkage alarm system can formulate targeted linkage strategies based on the accurate fire point location, thereby improving the accuracy of alarm response.

[0110] In summary, when performing heterogeneous feature association mapping on the thermal radiation feature vector and the spatial distribution feature vector, the high-intensity thermal radiation feature fragments and the specific texture distribution feature fragments of the spatial distribution are extracted, and the correlation and overlap information are recorded according to their spatial positions to obtain the feature mapping relationship between thermal radiation and spatial distribution in the target area. This relationship establishes a direct association between the two, providing a core basis for subsequent feature collaborative analysis and supporting the system's comprehensive judgment of fire point features.

[0111] In summary, when performing spectral-spatial feature co-analysis on multi-spectral image data based on the aforementioned feature mapping relationship, thermal radiation-related spectral signals and spatial distribution-related image signals are separated. High-radiation spectral regions and specific texture spatial regions are screened according to the mapping relationship. Spatial positions are compared and overlapping regions are retained to obtain potential fire point regions of the target area. This delineates the initial range for fire point identification, avoids misjudgment based on a single feature, and narrows the fire point investigation focus range of the linkage alarm system.

[0112] In summary, when optimizing the thermal radiation profile of the potential fire point area, thermal radiation intensity information at each location within the area is collected, edge locations below the minimum radiation standard for fire points are excluded, the boundary of the core area is corrected to eliminate irregular parts, and the optimized profile boundary and range are marked to obtain the fire point area of ​​the target area. This accurately locks the actual range of the fire point, ensuring that the system formulates targeted linkage strategies based on accurate locations, thereby improving the accuracy and effectiveness of alarm response.

[0113] The multidimensional risk factor assessment module 104 is used to perform multidimensional risk factor assessment on the environmental parameters of the fire point area to obtain the risk level of the target area.

[0114] In this embodiment of the invention, when the multidimensional risk factor assessment module performs a multidimensional risk factor assessment of the environmental parameters of the fire point area to obtain the risk level of the target area, it is specifically used for:

[0115] The environmental parameters include: wind speed and direction factor, atmospheric humidity factor, and vegetation cover factor of the target area;

[0116] Based on the wind speed and direction factors, the fire spread trend of the fire point area is analyzed to obtain a dynamic spread description of the target area.

[0117] Based on the atmospheric humidity factor and the vegetation cover factor, the flammability level of the fire point area is determined to obtain a static fire hazard description of the target area.

[0118] The risk level of the target area is obtained by fusing the dynamic spread description and the static fire hazard description.

[0119] The formula for calculating the risk level is as follows:

[0120] ;

[0121] In the formula, For the aforementioned risk level, This is the fire spread intensity coefficient describing the dynamic spread. The wind field enhancement factor described by the dynamic spread, This refers to the terrain suppression factor describing the dynamic spread. This is the combustible material loading factor for the static fire hazard description. The dryness index is used to describe the static fire hazard.

[0122] The steps for obtaining the risk level also include:

[0123] Based on preset multi-level risk thresholds, the fire zone is mapped to a risk range to obtain the preliminary risk level of the target area.

[0124] When the initial risk level is within the boundary region of the multi-level risk threshold, trend feature analysis is performed on the dynamic spread description to obtain the risk evolution direction feature;

[0125] Based on the static fire risk description, the risk evolution direction characteristics are arbitrated to obtain the risk level arbitration result of the target area.

[0126] The risk level arbitration result is output as the risk level of the target area.

[0127] Specifically, when defining environmental parameters including wind speed and direction factors, atmospheric humidity factors, and vegetation cover factors for the target area, information on each factor needs to be acquired step-by-step using specialized detection equipment. Wind speed and direction factors are collected in real-time by wind speed and direction sensors deployed in the target area. These sensors continuously capture the speed and direction of airflow within the target area and convert the captured airflow information into clearly describable wind speed and direction data, integrating them into the wind speed and direction factor. Atmospheric humidity factors are collected by atmospheric humidity sensors located in the target area. These sensors can sense the water vapor content in the air and convert the water vapor content information into corresponding humidity data, defined as the atmospheric humidity factor. Vegetation cover factors are obtained by performing vegetation identification on visible light image data of the target area. Using image recognition components, pixel regions in the image are analyzed one by one to distinguish between vegetation-covered and non-vegetation-covered areas, determining the coverage and density of vegetation within the target area. This vegetation-related information is then integrated into the vegetation cover factor, ensuring that the environmental parameters accurately include the above three types of factors.

[0128] Furthermore, when analyzing the fire spread trend in the fire area based on wind speed and direction factors to obtain a dynamic spread description of the target area, the fire trend analysis component needs to be used. First, wind direction data is extracted from the wind speed and direction factors. Based on the wind direction data, the direction of airflow is determined, thereby identifying the main direction in which the fire may spread under the influence of wind within the fire area. For example, if the wind direction is a certain direction, the fire will preferentially spread in that direction. Next, wind speed data is extracted from the wind speed and direction factors. Wind speed data reflects the speed of airflow. The faster the wind speed, the stronger the wind's driving force on the fire, and the faster the fire spreads. The slower the wind speed, the slower the fire spreads. The determined main direction of fire spread, the spread speed, and the area that the fire may affect based on both are integrated into a complete description of the fire's spread trend over time. This information is the dynamic spread description of the target area. This description can reflect the dynamic development of fire risk in real time, providing a dynamic basis for subsequent risk assessment and supporting the linkage alarm system in timely control of fire risk spread.

[0129] Furthermore, when determining the flammability level of a fire point area based on atmospheric humidity factors and vegetation cover factors to obtain a static fire hazard description of the target area, a flammability analysis component is required. First, the humidity data corresponding to the atmospheric humidity factor is analyzed. Lower atmospheric humidity indicates less water vapor content in the air of the target area, resulting in lower moisture content in objects within the fire point area, making them easier to ignite and causing more intense combustion. Higher atmospheric humidity indicates higher moisture content in objects, making them less ignitable and reducing combustion intensity. Next, the vegetation information corresponding to the vegetation cover factor is analyzed. Wider and denser vegetation cover indicates a greater amount of vegetation within the fire point area that can serve as combustible material, providing a more abundant material basis for fire combustion. Insufficient vegetation cover and low density result in a smaller amount of combustible vegetation, weakening the material support for sustained fire combustion. The flammability level of the fire zone is determined by combining atmospheric humidity (reflecting the flammability of objects) and vegetation cover (reflecting the amount of combustible material). For example, a high flammability level is indicated by low atmospheric humidity and dense vegetation cover, while a low flammability level is indicated by high atmospheric humidity and sparse vegetation cover. The determined flammability level, along with the atmospheric humidity and vegetation cover information used for the determination, are integrated into a description of the inherent fire hazard conditions of the fire zone. This information constitutes the static fire hazard description of the target area, reflecting its inherent fire hazard potential and providing static support for risk level assessment.

[0130] Furthermore, when determining the risk level by fusing dynamic spread descriptions and static fire hazard descriptions to obtain the risk level of the target area, an integrated analysis needs to be conducted through a risk fusion component. First, the fire spread speed and potential impact range are extracted from the dynamic spread description. If the fire spread speed is fast and the impact range is large, it indicates that the fire hazard is likely to spread in a short period of time, and the risk level at the dynamic level is high; if the fire spread speed is slow and the impact range is small, the risk level at the dynamic level is low.

[0131] Furthermore, the flammability rating is extracted from the static fire hazard description. A high flammability rating means that objects within the fire area are easily combustible and there is sufficient combustible material, indicating a high level of risk at the static level; a low flammability rating indicates a low level of risk at the static level. Subsequently, the dynamic risk level and the static risk level are comprehensively considered. If the dynamic spread speed is fast and the static flammability rating is high, it indicates that the fire not only spreads quickly but also has a good combustion foundation, indicating a high overall risk level; if the dynamic spread speed is slow and the static flammability rating is low, the overall risk level is low; if there is a situation where the dynamic spread is fast but the static flammability is low, or the dynamic spread is slow but the static flammability is high, then the risk level is determined by combining the comprehensive impact of both on the development of the fire hazard. For example, when the dynamic spread is fast but the static flammability is low, although there is insufficient combustible material, the rapid spread still needs to be judged as medium to high risk; when the dynamic spread is slow but the static flammability is high, although the spread is slow, the prolonged combustion also needs to be judged as medium to high risk.

[0132] Furthermore, based on the comprehensive assessment of the risk level, a specific risk level is determined, which is the risk level of the target area. This risk level can comprehensively reflect the severity of the fire hazard in the target area, providing an accurate basis for subsequent linkage strategy mapping, ensuring that the multi-spectral fire point intelligent detection and linkage alarm system can trigger appropriate alarms and response actions, and improving the pertinence and effectiveness of fire hazard handling.

[0133] Specifically, the fire spread intensity coefficient described in the dynamic spread description originates from the multi-dimensional risk factor assessment module. Based on the wind speed and direction factors of the target area, the module analyzes the fire spread trend in the fire point area to obtain the dynamic spread description of the target area. Information that can characterize the intensity of fire spread is extracted from the dynamic spread description. During the extraction process, it is necessary to analyze the fire spread speed, the range of areas that may be affected, and the spread intensity of the fire at different locations recorded in the dynamic spread description. Based on this information, the corresponding fire spread intensity coefficient is determined. This coefficient is directly related to the dynamic spread description and is a quantitative representation of the core characteristics of fire spread in the dynamic spread description.

[0134] Furthermore, the wind field enhancement factor in the dynamic spread description also originates from the dynamic spread description. When the multidimensional risk factor assessment module analyzes the fire spread trend based on wind speed and direction factors, wind speed data directly affects the speed of fire spread. The obtained dynamic spread description reflects the magnitude of the wind's driving effect on the fire. Extracting the information corresponding to this driving effect from the dynamic spread description is the wind field enhancement factor. When extracting, it is necessary to combine the wind speed information in the wind speed and direction factors to determine the degree of wind enhancement on fire spread and convert this degree into the wind field enhancement factor to ensure that the factor can accurately reflect the influence of the wind field on the fire.

[0135] Furthermore, the terrain inhibition factor in the dynamic spread description originates from the dynamic spread description itself. When the multidimensional risk factor assessment module analyzes the fire spread trend based on wind speed and direction factors, it also considers the impact of terrain factors on fire spread in the target area. For example, whether the terrain has features that hinder fire spread will be reflected in the dynamic spread description. Information that reflects the inhibitory effect of terrain on fire spread is extracted from the dynamic spread description, which is the terrain inhibition factor. During the extraction process, it is necessary to analyze the extent to which the fire spread is hindered at different terrain locations in the dynamic spread description, and determine the specific content of the terrain inhibition factor based on the degree of obstruction, ensuring that the factor is consistent with the terrain influence characteristics in the dynamic spread description.

[0136] Furthermore, the combustible material load factor in the static fire hazard description originates from the static fire hazard description obtained by the multi-dimensional risk factor assessment module based on atmospheric humidity factor and vegetation cover factor to determine the combustibility level of the fire point area. Information that can characterize the amount of combustible material in the target area is extracted from the static fire hazard description, which is the combustible material load factor. During the extraction process, it is necessary to combine the vegetation coverage range and growth density information in the vegetation cover factor, as well as the combustibility level determined in the static fire hazard description, to determine the corresponding total combustible material quantity characteristics. This characteristic is then converted into the combustible material load factor to ensure that the coefficient is directly related to the combustible material quantity characteristics in the static fire hazard description.

[0137] Furthermore, the dryness index of the static fire hazard description originates from the static fire hazard description itself. When the multidimensional risk factor assessment module determines the flammability level based on atmospheric humidity factor and vegetation cover factor, the atmospheric humidity factor directly affects the dryness of the target area. This dryness is reflected in the obtained static fire hazard description. The information that reflects the dryness of the target area is extracted from the static fire hazard description, which is the dryness index. During the extraction process, it is necessary to combine the humidity data in the atmospheric humidity factor, analyze the judgment result of the dryness of the area in the static fire hazard description, and convert the result into the dryness index to ensure that the index can accurately reflect the dryness characteristics in the static fire hazard description.

[0138] Furthermore, this formula is used to calculate the risk level of a target area. The calculation integrates fire spread intensity coefficient, wind field enhancement factor, and terrain inhibition factor related to dynamic spread, as well as combustible material load coefficient and dryness index related to static fire hazard. This fuses the dynamic fire spread trend characteristics with the static fire hazard basic conditions. Specifically, the fire spread intensity coefficient and wind field enhancement factor reflect the fire's diffusion capacity and the driving force of wind during dynamic spread; the terrain inhibition factor reflects the constraint of terrain on dynamic spread; and the combustible material load coefficient and dryness index reflect the quantity of combustible material and the dryness of the area during static fire hazard. Through this fusion calculation, a risk level that comprehensively reflects the severity of fire hazard in the target area is obtained. This risk level can be directly provided to the linkage strategy mapping module, serving as the core basis for the module to perform strategy mapping based on a preset linkage strategy library. This ensures that the linkage strategy mapping module can accurately match the linkage strategy adapted to the risk level of the target area, thereby supporting the multi-spectral intelligent fire detection and linkage alarm system to achieve accurate linkage alarm functions.

[0139] Specifically, when mapping the fire point area to risk intervals based on preset multi-level risk thresholds to obtain the preliminary risk level of the target area, it is necessary to first clarify that the preset multi-level risk thresholds are formulated in advance based on historical fire hazard handling data and safety protection standards under different scenarios. Each threshold corresponds to a specific risk interval, such as the defined ranges for low risk, medium risk, and high risk. Then, the comprehensive risk characteristics formed in the early multi-dimensional risk factor assessment of the fire point area are extracted, including the core information of dynamic spread description and static fire hazard description. The comprehensive risk characteristics are compared with the preset multi-level risk thresholds one by one to determine which risk interval the content corresponding to the comprehensive risk characteristics falls into. The risk level corresponding to this interval is the preliminary risk level of the target area. This ensures that the preliminary risk level can preliminarily define the degree of fire hazard based on a unified standard, providing a basis for possible subsequent level arbitration and supporting the risk judgment process of the multi-spectral fire point intelligent detection and linkage alarm system.

[0140] Furthermore, when the preliminary risk level is located in the boundary region of the multi-level risk thresholds, trend feature analysis is performed on the dynamic spread description to obtain the risk evolution direction characteristics. First, it is clarified that the boundary region refers to the content corresponding to the preliminary risk level being exactly at the critical position of two adjacent risk thresholds, such as between the upper limit of the medium risk threshold and the lower limit of the high risk threshold. Then, core information such as fire spread speed, stability of spread direction, and rate of expansion of influence area are extracted from the dynamic spread description. It is analyzed whether the fire spread speed is continuously accelerating, remaining stable, or gradually slowing down, whether the spread direction is shifting towards areas with dense combustibles, and whether the rate of expansion of influence area shows an upward trend. These analyzed fire change trends are integrated into an information set that can clearly describe the future development direction of the risk. This information set is the risk evolution direction characteristic of the target area. This characteristic can clearly reflect the dynamic development trend of fire risk and provide dynamic trend basis for subsequent level arbitration.

[0141] Furthermore, based on the static fire hazard description, when arbitrating the risk evolution direction characteristics to obtain the risk level arbitration result for the target area, core information such as flammability rating, combustible material load characteristics, and dryness status are first extracted from the static fire hazard description. This information reflects the inherent fire hazard basis of the fire point area. If the static fire hazard description shows a high flammability rating, sufficient combustible material load, and high dryness, indicating a strong inherent fire hazard basis, and the risk evolution direction characteristics show accelerated fire spread and expanded impact range, indicating an upward trend in risk, then the preliminary risk level is arbitrated to a higher level. The risk level is determined by the following criteria: If the static fire hazard description indicates a low flammability rating, low combustible material load, and low dryness, meaning the inherent fire hazard basis is weak, the initial risk level will be maintained or slightly adjusted even if the risk evolution direction shows an upward trend. If the static fire hazard basis is strong but the risk evolution direction shows a downward trend, the initial risk level will be arbitrated down to a lower risk level. By combining the inherent fire hazard basis with the dynamic trend in this comprehensive judgment, the final determined risk level is the risk level arbitration result for the target area. This arbitration result can avoid the deviation in the determination of the initial risk level of the boundary area.

[0142] Furthermore, when outputting the risk level arbitration result as the risk level of the target area, the risk level arbitration result is transmitted from the multi-dimensional risk factor assessment module to the linkage strategy mapping module through the information transmission component inside the system. During the transmission process, the integrity of the arbitration result is ensured, including the final risk level and the core points of the arbitration basis. At the same time, the risk level arbitration result and the corresponding judgment process are recorded in the system storage unit for subsequent traceability and verification. The risk level arbitration result transmitted to the linkage strategy mapping module at this time serves as the final risk level of the target area, providing an accurate basis for the linkage strategy mapping module to match and adapt linkage strategies based on the preset linkage strategy library.

[0143] In summary, when the environmental parameters are clearly defined as wind speed and direction, atmospheric humidity, and vegetation cover factor, core information is collected through dedicated equipment. Wind speed and direction are captured by corresponding sensors, atmospheric humidity is sensed by humidity sensors, and vegetation cover is determined through visible light image recognition. This ensures that the parameters fully cover the three types of factors, providing accurate basic data for multidimensional risk assessment and supporting system risk judgment.

[0144] In summary, when analyzing the fire spread trend based on the wind speed and direction factors, the main direction of spread is determined by the wind direction, the speed of spread is judged by the wind speed, and relevant information is integrated to obtain a dynamic spread description of the target area. This description reflects the dynamic spread trend of fire risk, provides a dynamic basis for risk level assessment, and helps the system to grasp the development of fire risk.

[0145] In summary, when determining the flammability level based on the atmospheric humidity and vegetation cover factor, the flammability of objects is judged by humidity data, the amount of combustible material is determined by combining the vegetation conditions, the level is comprehensively determined and the information is integrated to obtain a static fire risk description of the target area, reflecting the inherent fire risk basis and providing static support for risk assessment.

[0146] In summary, when determining the risk level by integrating the dynamic and static fire hazard descriptions, the core information of both is extracted to determine the dynamic and static risk levels. After comprehensive consideration, the risk level of the target area is obtained, which fully reflects the severity of the fire hazard, provides an accurate basis for mapping linkage strategies, and improves the targeting of fire hazard response.

[0147] In summary, when mapping the initial risk level based on preset multi-level risk thresholds, the preset thresholds are formulated based on historical data of fire hazard handling and scenario safety standards. Each threshold corresponds to a clear interval. The comprehensive risk characteristics of the early stage are extracted and compared with the thresholds to determine the initial risk level corresponding to the interval, providing a basis for subsequent level arbitration and supporting the system risk assessment.

[0148] In summary, when the initial risk level is in the threshold boundary area, the dynamic spread description is analyzed to obtain the risk evolution direction characteristics, clarify the boundary area attributes, extract information such as fire spread speed and directional stability, and integrate the fire change trend to form an information set. This set reflects the dynamic trend of fire risk and provides a basis for level arbitration.

[0149] In summary, when arbitrating the risk evolution direction based on static fire risk descriptions, core information reflecting the inherent fire risk basis is extracted. If the inherent fire risk is strong and the risk is rising, the arbitration is upward; if the basis is weak, the arbitration is maintained or slightly adjusted; if the basis is strong but the risk is declining, the arbitration is downward. The comprehensive judgment yields the risk level arbitration result, avoiding bias in boundary area judgments.

[0150] In summary, when the risk level arbitration result is output as the risk level of the target area, it is transmitted from the multi-dimensional risk factor assessment module to the linkage strategy mapping module through internal system components. The result and judgment process are recorded for traceability. The arbitration result is the final risk level, providing an accurate basis for strategy matching.

[0151] The linkage strategy mapping module 105 is used to perform strategy mapping on the risk level based on a preset linkage strategy library to obtain the linkage strategy for the target area.

[0152] In this embodiment of the invention, when the linkage strategy mapping module performs strategy mapping on the risk level based on a preset linkage strategy library to obtain the linkage strategy for the target area, it is specifically used for:

[0153] Based on a preset linkage strategy library, the basic strategy corresponding to the risk level is matched;

[0154] Based on the regional topological features of the fire point area, the basic strategy is parsed to obtain the strategy set of the target area;

[0155] Based on the intensity characteristics of the thermal radiation feature vector, the strategy set is configured with strategy parameters to obtain the linkage strategy for the target area.

[0156] Specifically, when matching the basic strategy corresponding to the risk level based on the preset linkage strategy library, the preset linkage strategy library needs to be formulated in advance according to the fire hazard handling needs, fire safety regulations and historical fire hazard response cases under different risk levels. Each risk level in the library corresponds to a set of clear general fire hazard response measures, i.e., basic strategies. When performing matching, the risk level of the target area is first extracted, and then the risk level is compared with each risk level recorded in the preset linkage strategy library one by one to find the general fire hazard response measures that completely correspond to the risk level of the target area. These measures are the basic strategies of the target area, ensuring that the basic strategies can initially adapt to the fire hazard severity of the target area, providing a general framework for subsequent strategy optimization, and supporting the linkage response function of the multi-spectral fire point intelligent detection and linkage alarm system.

[0157] Furthermore, based on the regional topological features of the fire point area, when performing strategy analysis on the basic strategy to obtain the strategy set for the target area, the regional topological features of the fire point area are first obtained through the regional topology acquisition component. These features include information such as the distribution of buildings around the fire point, road directions and locations, the layout and quantity of fire-fighting facilities, and the range of densely populated areas. Subsequently, the general response measures in the basic strategy are correlated with these regional topological features. For example, the general measure of "organizing personnel evacuation" in the basic strategy is combined with the building distribution to determine the specific buildings to be evacuated and the evacuation routes are determined by combining the road directions. The general measure of "dispatching fire-fighting resources" in the basic strategy is combined with the layout of fire-fighting facilities to determine the locations of fire-fighting facilities to be called first. These specific response measures adapted to the regional topological features are compiled and summarized, and the resulting set of measures is the strategy set for the target area, ensuring that the strategy set can fit the actual spatial layout of the fire point area and improve the pertinence of the coordinated response.

[0158] Furthermore, based on the intensity characteristics of the thermal radiation feature vector, when configuring the strategy parameters of the strategy set to obtain the linkage strategy for the target area, the intensity characteristics are first extracted from the thermal radiation feature vector. These characteristics reflect the intensity of thermal radiation from the fire point, including information such as the thermal radiation coverage area and the amount of radiation energy. Subsequently, detailed parameters are configured for each specific measure in the strategy set according to the intensity characteristics. For example, for the measure "activate fire extinguishing equipment" in the strategy set, if the thermal radiation intensity is high, i.e., the radiation energy is large and the coverage area is wide, it is configured to activate fire extinguishing equipment with higher power and wider coverage area in the area; if the thermal radiation intensity is low, it is configured to activate fire extinguishing equipment with conventional power. For the measure "set warning range" in the strategy set, if the thermal radiation intensity is high, the warning range is configured to be larger; if the thermal radiation intensity is low, the warning range is configured to be smaller. After the parameters of all measures in the strategy set are configured according to the thermal radiation intensity characteristics, the resulting complete and parameter-clear response plan is the linkage strategy for the target area. This linkage strategy can accurately adapt to the thermal radiation state of the fire point and the actual situation of the area.

[0159] In summary, when matching basic strategies corresponding to risk levels based on a pre-set linkage strategy library, the strategy library is formulated according to the handling needs of different risk levels, fire safety regulations, and historical cases. Each level corresponds to a set of general response measures. By extracting the risk level of the target area and comparing it with the library, the corresponding general measures are found to form the basic strategy. This can initially adapt to the degree of fire risk, provide a framework for subsequent optimization, and support the system's linkage response.

[0160] In summary, when obtaining a strategy set based on the topological features of the fire point area, the topological features such as the distribution of surrounding buildings, road directions, layout of fire-fighting facilities, and densely populated areas are obtained through the acquisition component. The general measures of the basic strategy are correlated and analyzed with these features, and the specific measures after adaptation are summarized to form a strategy set that fits the actual spatial layout and improves the targeting of the linkage.

[0161] In summary, when configuring strategy set parameters based on the intensity features of thermal radiation feature vectors to obtain linkage strategies, intensity features reflecting the strength, coverage, and energy of thermal radiation from fire points are extracted and used as strategy set configuration parameters. The complete response plan formed after all parameters are configured is the linkage strategy, which accurately adapts to the fire point and regional conditions, ensuring that the system triggers effective alarms and response actions.

[0162] The linkage strategy response detection module 106 is used to detect the response of the linkage strategy and obtain the linkage alarm report of the target area.

[0163] When the linkage strategy response detection module performs response detection on the linkage strategy and obtains the linkage alarm report for the target area, it is specifically used for:

[0164] The linkage strategy is decomposed into strategy elements to obtain the alarm object set, action command set and resource allocation scheme of the target area;

[0165] The alarm object set is authenticated to obtain the alarm object list of the target area;

[0166] The feasibility of the action instruction set is determined to obtain the action instruction sequence for the target area;

[0167] The resource allocation plan is checked for resource status to obtain a list of available resources in the target area;

[0168] The alarm object list, the action command sequence, and the available resource list are compiled into a multi-dimensional data set to obtain a linkage alarm report for the target area.

[0169] Specifically, when decomposing the linkage strategy into its elements to obtain the alarm object set, action instruction set, and resource allocation plan for the target area, it is necessary to first identify the core components of the linkage strategy. This involves identifying the relevant clauses regarding alarm objects, action execution, and resource mobilization within the strategy. All objects in the alarm object clauses that need to receive fire alarm information are organized into a set, which constitutes the alarm object set for the target area. All response actions to fire points in the action execution clauses are organized into a set, which constitutes the action instruction set for the target area. All resource types, quantities, and allocation paths to be allocated in the resource mobilization clauses are organized into a structured plan, which constitutes the resource allocation plan for the target area. This ensures that the three types of elements after decomposition fully cover the core content of the linkage strategy, providing a clear elemental basis for subsequent response detection.

[0170] Furthermore, when performing response authentication on the alarm object set to obtain the alarm object list for the target area, the permission verification component is used to verify the identity information of each object in the alarm object set against the preset alarm receiving permission list. The permission list clearly records the range of alarm information that different types of objects can receive and the validity period of the permissions. If the object's identity information completely matches the list item and the permission is within the validity period, it is determined that it has the permission and is retained; if the identity does not match or the permission has expired, it is determined that it has no permission and is removed. All retained authorized objects are sorted according to notification priority, and the resulting ordered object list is the alarm object list for the target area, ensuring that alarm information is accurately transmitted to the valid objects, avoiding mistransmission or omission, and supporting the accuracy of linkage alarms.

[0171] Furthermore, when determining the feasibility of the action instruction set and obtaining the action instruction sequence for the target area, it is necessary to analyze the actual situation of the fire point area and determine whether the execution conditions of each instruction in the action instruction set are met. The execution conditions include the terrain conditions of the fire point area, the status of existing facilities, and the personnel capability configuration. If all execution conditions of an instruction are met, it is deemed feasible; if any condition is not met, it is deemed infeasible and is removed. All feasible instructions are arranged in the order of execution logic, and the resulting ordered instruction sequence is the action instruction sequence for the target area. This ensures that the action instructions are executed in a reasonable order and avoids reduced response efficiency due to disordered instructions.

[0172] Furthermore, when verifying the resource status of the resource allocation plan to obtain the list of available resources in the target area, it is necessary to confirm the current status of each resource in the plan through resource inspection or by connecting with the resource management system. The verification includes whether the resource is working normally, whether the quantity meets the requirements, and whether the storage location is convenient for allocation. If the resource meets the conditions of "working normally, sufficient quantity, and accessible location", it is determined to be available and the resource name, specifications, quantity, storage location, and contact person for calling are recorded. If the resource is damaged, insufficient in quantity, or located in a remote area, it is determined to be unavailable and excluded. All available resource information is classified and organized by type, and the resulting resource information list is the list of available resources in the target area, which provides accurate resource support for the execution of subsequent action instructions and avoids interruption of response actions due to resource unavailability.

[0173] Furthermore, when compiling the alarm object list, action command sequence, and available resource list into a multi-dimensional alarm report for the target area, the three types of elements need to be integrated according to the logical relationship of "alarm object - action command - supporting resources". First, the basic situation of the fire point area is summarized in the report, and then the information is presented in modules: the first module lists the alarm objects, specifying the name, contact information, and alarm content to be notified for each object; the second module presents the action command sequence, explaining the execution time, execution subject, and operation requirements of each command; the third module marks the available resource list, corresponding to the resources to be called and the allocation method for each action command, ensuring that the information of the three types of elements matches without contradiction. Finally, all information is organized in a standardized format, and the resulting complete document is the target area's alarm report, providing a comprehensive action guide for fire response personnel and ensuring that the alarm system's response is orderly and efficient.

[0174] In summary, when decomposing the linkage strategy to obtain the alarm object set, action instruction set, and resource allocation plan, relevant clauses are identified and organized into three types of elements to ensure coverage of the core content of the strategy, provide a clear foundation for subsequent response detection, and meet the alarm system element decomposition requirements corresponding to the classification number.

[0175] In summary, when performing response authentication on the alarm object set, the permission verification component checks the identity information of each object against the preset permission list, retains the matching objects that are within the validity period, and sorts them by priority to form an alarm object list, ensuring accurate transmission of alarm information and supporting the accuracy of linkage alarms.

[0176] In summary, when making a feasibility determination on the action instruction set, the actual situation of the fire point area is considered to determine whether the execution conditions of each instruction are met. Feasible instructions are retained and arranged in logical order to form an action instruction sequence, ensuring that the instructions are executed in an orderly manner and in accordance with the alarm post-alarm action coordination requirements corresponding to the classification number.

[0177] In summary, when verifying the resource status of the resource allocation plan, the working status, quantity and storage location of each resource are confirmed through resource inspection or connection with the management system. Available resources are screened and classified to form a list of available resources, providing accurate resource support for the execution of action instructions.

[0178] In summary, when integrating the alarm object list, action command sequence, and available resource list, the elements are integrated according to the logic of "alarm object - action command - supporting resources", the basic situation of the fire point is summarized and the information is presented in modules. The information is organized into a linkage alarm report in a standardized format, providing action guidelines for response personnel, which meets the alarm report generation function requirements of the classification number.

[0179] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0180] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0181] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A multi-spectral fire detection and linkage alarm system based on an internet platform, characterized in that: The system includes a thermal radiation feature extraction module, a spatial distribution feature parsing module, a heterogeneous feature segmentation module, a multi-dimensional risk factor assessment module, a linkage strategy mapping module, and a linkage strategy response detection module, wherein: The thermal radiation feature extraction module is used to perform thermal radiation feature analysis on the multi-spectral image data of the target area to obtain the thermal radiation feature vector of the target area. The spatial distribution feature parsing module is used to perform spatial distribution feature parsing on the multi-spectral image data to obtain the spatial distribution feature vector of the target region; The heterogeneous feature segmentation module is used to perform heterogeneous feature segmentation on multi-spectral image data based on the thermal radiation feature vector and the spatial distribution feature vector to obtain the fire point region of the target region. The multidimensional risk factor assessment module is used to perform multidimensional risk factor assessment on the environmental parameters of the fire point area to obtain the risk level of the target area. The linkage strategy mapping module is used to map the risk level to a preset linkage strategy library to obtain the linkage strategy for the target area. The linkage strategy response detection module is used to detect the response of the linkage strategy and obtain the linkage alarm report of the target area.

2. The multi-spectral fire detection and linkage alarm system based on an internet platform as described in claim 1, characterized in that, When the thermal radiation feature extraction module performs thermal radiation feature analysis on the multispectral image data of the target area to obtain the thermal radiation feature vector of the target area, it is specifically used for: Infrared band separation is performed on the multispectral image data of the target area to obtain the infrared radiation data of the target area; The infrared radiation data is calibrated to obtain standardized radiation intensity information for the target area; The standardized radiation intensity information is used to construct feature primitives to obtain the radiation feature primitive set of the target region; Multi-scale feature integration is performed on the radiation feature primitive set to obtain the thermal radiation feature vector of the target region.

3. The multi-spectral fire detection and linkage alarm system based on an internet platform as described in claim 1, characterized in that, When the spatial distribution feature parsing module performs spatial distribution feature parsing on the multispectral image data to obtain the spatial distribution feature vector of the target region, it is specifically used for: The multi-spectral image data is spectrally decoupled to obtain visible light image data of the target region; Texture feature recognition is performed on the visible light image data to obtain texture pattern information of the target region; The texture pattern information is subjected to morphological topological characterization to obtain the spatial distribution feature vector of the target region.

4. The multi-spectral fire detection and linkage alarm system based on an internet platform as described in claim 3, characterized in that, When the spatial distribution feature parsing module performs texture feature recognition on the visible light image data to obtain the texture pattern information of the target region, it is specifically used for: The visible light image data is analyzed for texture orientation to obtain the texture orientation distribution of the target region; The spatial density features of the texture direction distribution are integrated to obtain the texture density distribution features of the target region. Multi-scale feature fusion is performed on the texture direction distribution and texture density distribution features to obtain the texture pattern information of the target region.

5. The multi-spectral fire detection and linkage alarm system based on an internet platform as described in claim 1, characterized in that, When the heterogeneous feature segmentation module performs heterogeneous feature segmentation on the multispectral image data based on the thermal radiation feature vector and the spatial distribution feature vector to obtain the fire point region of the target region, it is specifically used for: By performing heterogeneous feature association mapping on the thermal radiation feature vector and the spatial distribution feature vector, the feature mapping relationship between thermal radiation and spatial distribution corresponding to the target region is obtained; Based on the feature mapping relationship, the multi-spectral image data is subjected to spectral-spatial feature co-analysis to obtain the potential fire point region of the target area; The potential fire point region is optimized by thermal radiation profile to obtain the fire point region of the target region.

6. The multi-spectral fire detection and linkage alarm system based on an internet platform as described in claim 1, characterized in that, When the multidimensional risk factor assessment module performs a multidimensional risk factor assessment of the environmental parameters of the fire point area to obtain the risk level of the target area, it is specifically used for: The environmental parameters include: wind speed and direction factor, atmospheric humidity factor, and vegetation cover factor of the target area; Based on the wind speed and direction factors, the fire spread trend of the fire point area is analyzed to obtain a dynamic spread description of the target area. Based on the atmospheric humidity factor and the vegetation cover factor, the flammability level of the fire point area is determined to obtain a static fire hazard description of the target area. The risk level of the target area is obtained by fusing the dynamic spread description and the static fire hazard description.

7. The multi-spectral fire detection and linkage alarm system based on an internet platform as described in claim 6, characterized in that, The formula for calculating the risk level is as follows: ； In the formula, For the aforementioned risk level, This is the fire spread intensity coefficient describing the dynamic spread. The wind field enhancement factor described by the dynamic spread, This refers to the terrain suppression factor describing the dynamic spread. This refers to the combustible material loading coefficient used in the static fire hazard description. The dryness index is used to describe the static fire hazard.

8. The multi-spectral fire detection and linkage alarm system based on an internet platform as described in claim 6, characterized in that, The steps for obtaining the risk level also include: Based on preset multi-level risk thresholds, the fire zone is mapped to a risk range to obtain the preliminary risk level of the target area. When the initial risk level is within the boundary region of the multi-level risk threshold, trend feature analysis is performed on the dynamic spread description to obtain the risk evolution direction feature; Based on the static fire risk description, the risk evolution direction characteristics are arbitrated to obtain the risk level arbitration result of the target area. The risk level arbitration result is output as the risk level of the target area.

9. The multi-spectral fire detection and linkage alarm system based on an internet platform as described in claim 1, characterized in that, When the linkage strategy mapping module executes a strategy mapping based on a preset linkage strategy library to obtain the linkage strategy for the target area, it is specifically used for: Based on a preset linkage strategy library, the basic strategy corresponding to the risk level is matched; Based on the regional topological features of the fire point area, the basic strategy is parsed to obtain the strategy set of the target area; Based on the intensity characteristics of the thermal radiation feature vector, the strategy set is configured with strategy parameters to obtain the linkage strategy for the target area.

10. The multi-spectral fire detection and linkage alarm system based on an Internet platform as described in claim 1, characterized in that, When the linkage strategy response detection module performs response detection on the linkage strategy and obtains a linkage alarm report for the target area, it is specifically used for: The linkage strategy is decomposed into strategy elements to obtain the alarm object set, action command set and resource allocation scheme of the target area; The alarm object set is authenticated to obtain the alarm object list of the target area; The feasibility of the action instruction set is determined to obtain the action instruction sequence for the target area; The resource allocation plan is checked for resource status to obtain a list of available resources in the target area; The alarm object list, the action command sequence, and the available resource list are compiled into a multi-dimensional data set to obtain a linkage alarm report for the target area.

Images