Forest fire source detection method and system based on multi-source information fusion and storage medium

Abstract

The invention discloses a forest fire source detection method and system based on multi-source information fusion and a storage medium, and the method comprises the steps: carrying out the real-time data collection of a detection region, and enabling the real-time data to at least comprise infrared image data, visible light image data and temperature data; carrying out consistency and fusibility preprocessing on the real-time data; carrying out fire source key feature vector extraction on the preprocessed real-time data, wherein fire source key feature vectors comprise a heat radiation feature vector, a visual feature vector and a temperature abnormal point feature vector of a fire source; fusing the key feature vectors of the fire source, and generating a comprehensive feature vector; and fire source detection and positioning are carried out through the comprehensive feature vector. According to the invention, based on optical image data, infrared image data and temperature data, different data sources are fused by using a deep learning model; detecting and positioning a fire source through the comprehensive characteristics; therefore, efficient and accurate target detection is realized.

Description

Technical Field

[0001] The invention relates to forest fire source detection technology, in particular to a forest fire source detection method based on multi-source information fusion. Background Art

[0002] In the field of fire source detection technology, existing technologies mainly rely on a single data source for fire source identification and detection. These data sources usually include infrared images or visible light images. However, this detection method that relies on a single data source has significant limitations in complex environments.

[0003] First, complex environments (such as smoke, drastic changes in lighting, or multiple interference factors) may cause a significant drop in the detection accuracy of a single data source (such as infrared images or visible light images). Due to the interference of these environmental factors, the fire source signal may become blurred or difficult to distinguish, increasing the risk of false alarms and missed alarms.

[0004] Secondly, methods that process a single data source often require a lot of calculations to extract effective fire source features in complex environments. This large amount of computing demand not only increases the resource consumption of the system, but may also lead to slower response speeds and fail to meet the needs of real-time fire source detection. This delay can have serious consequences in terms of early warning and rapid response to fires, as timely fire source detection and alarm are crucial to reducing fire losses.

[0005] In summary, the existing fire source detection technology has the problems of insufficient detection accuracy and slow response speed in complex environments due to its reliance on a single data source, which poses a threat to the early warning and rapid response of fires. Therefore, it is necessary to develop a new fire source detection technology that can overcome these limitations. Summary of the invention

[0006] Therefore, in order to solve the above-mentioned shortcomings, the present invention provides a forest fire source detection method, system and storage medium based on multi-source information fusion. The present invention is based on optical image data, infrared image data and temperature data, and uses a deep learning model to fuse different data sources; finally, the fire source is detected and located through comprehensive features; the information of multiple data sources is comprehensively utilized to improve the detection accuracy and robustness, so as to achieve efficient and accurate target detection.

[0007] In a first aspect, the present invention provides a forest fire source detection method based on multi-source information fusion, comprising: Collecting real-time data of the detection area, wherein the real-time data includes at least infrared image data, visible light image data and temperature data; Preprocessing the real-time data for consistency and fusion; Extracting key feature vectors of fire sources from preprocessed real-time data, wherein the key feature vectors of fire sources include thermal radiation feature vectors, visual feature vectors, and temperature anomaly point feature vectors of fire sources; Fuse the key feature vectors of the fire source and generate a comprehensive feature vector; Fire source detection and location through comprehensive feature vectors.

[0008] Optionally, a method for real-time data collection of the detection area is: Use an infrared camera to obtain an infrared image of the detection area; Using a visible light camera to obtain a visible light image of the detection area; The temperature sensor is used to obtain the temperature data of the detection area.

[0009] Optionally, the preprocessing method includes denoising, image enhancement, and data alignment and normalization for real-time data.

[0010] Optionally, the method for extracting the key feature vector of the fire source includes extracting features of infrared images, and the method for extracting features of infrared images is: After preprocessing, the infrared image data is input into the deep residual network, and the front layer of the deep residual network is used to extract low-level features, including edge and texture features. The high-level features are extracted through the middle layer of the deep residual network, which include the state of mind and intensity of the fire source; The low-level features and high-level features are integrated through the deep residual network and finally through the fully connected layer to output the thermal radiation feature vector of the fire source.

[0011] Optionally, the method for extracting the key feature vector of the fire source also includes feature extraction of a visible light image, and the method for extracting the feature of the visible light image is: The preprocessed visible light image data is input into the YOLO model, which divides the image into multiple grids. Each grid is responsible for detecting whether there is a fire source inside it and generating a corresponding bounding box and confidence level. The features extracted by the YOLO model include the color, shape, size, and edge information of the fire source. These features are then encoded into feature vectors and output as visual feature vectors.

[0012] Optionally, the method for extracting the key feature vector of the fire source also includes extracting the feature of temperature data. The method for extracting the feature of temperature data is: The preprocessed temperature data is input into the LSTM network. The LSTM network analyzes the change pattern in the temperature data through its memory unit and gating mechanism. Based on the sudden temperature rise phenomenon, the LSTM network outputs the temperature anomaly point feature vector, which contains the temperature trend and fluctuation amplitude of the fire source location.

[0013] Optionally, the method of fusing the key feature vectors of the fire source and generating a comprehensive feature vector is: The multimodal fusion network using the attention mechanism first focuses on the thermal radiation feature vector in the infrared image, and then generates a comprehensive feature vector by combining the temporal features of the visual feature vector and the temperature anomaly point feature vector.

[0014] Optionally, the method for detecting and locating a fire source by using a comprehensive feature vector includes: a fire source detection method and a locating method; The fire source detection method is: Using a multi-layer perceptron model or a support vector machine-based classifier, the fused comprehensive feature vector is classified to determine whether a fire source exists; The positioning method is: After the fire source is determined, the specific location and range of the fire source are calculated through regression networks or geometric algorithms; the positioning process combines the position information in the visible light image and the heat source distribution in the infrared image to generate accurate fire source location coordinates.

[0015] In a second aspect, the present invention provides a forest fire source detection system based on multi-source information fusion, comprising: An acquisition module, used for acquiring real-time data of the detection area, wherein the real-time data at least includes infrared image data, visible light image data and temperature data; A preprocessing module, used for preprocessing the real-time data for consistency and fusion; A feature extraction module extracts key feature vectors of fire sources from preprocessed real-time data, wherein the key feature vectors of fire sources include thermal radiation feature vectors, visual feature vectors, and temperature anomaly point feature vectors of fire sources; A fusion module, used to fuse the key feature vectors of the fire source and generate a comprehensive feature vector; and The judgment and positioning module is used to detect and locate the fire source through comprehensive feature vectors.

[0016] In a third aspect, the present invention provides a storage medium having a computer program stored thereon, the program being executed by a processor to implement the forest fire source detection method based on multi-source information fusion.

[0017] The present invention has the following advantages: The present invention combines data obtained by multiple sensors such as infrared cameras, visible light cameras and temperature sensors to achieve comprehensive utilization of information from multiple data sources, improve detection accuracy and robustness, and achieve efficient and accurate target detection.

[0018] The present invention introduces a deep learning model for data preprocessing and feature extraction to fuse different data sources. This not only improves the efficiency of data processing, but also ensures the accuracy and reliability of the data. Through the powerful learning ability of the deep learning model, it is possible to fully explore and utilize the complementary information between various data sources, further improving the accuracy of fire source detection.

[0019] At the same time, the present invention detects and locates the fire source through comprehensive features, achieving efficient and accurate target detection. Compared with the traditional single data source detection method, the detection accuracy of the present invention is significantly improved. Especially in complex environments, such as smoke, light changes, etc., the present invention can still maintain high-precision detection, effectively reducing the false alarm and missed alarm problems caused by insufficient single data source. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a flow chart of the forest fire source detection method based on multi-source information fusion according to the present invention; Figure 2 It is a systematic structural block diagram of the forest fire source detection system based on multi-source information fusion according to the present invention; In the figure: 100, acquisition module; 200, preprocessing module; 300, feature extraction module; 400, fusion module; 500, judgment and positioning module. DETAILED DESCRIPTION

[0021] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0022] In this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises", or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or apparatus that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article, or apparatus.

[0023] As described in the background technology, in the field of fire source detection technology, existing technologies mainly rely on a single data source for fire source identification and detection. These data sources usually include infrared images or visible light images. However, this detection method that relies on a single data source has significant limitations in complex environments.

[0024] First, complex environments (such as smoke, drastic changes in lighting, or multiple interference factors) may cause a significant drop in the detection accuracy of a single data source (such as infrared images or visible light images). Due to the interference of these environmental factors, the fire source signal may become blurred or difficult to distinguish, increasing the risk of false alarms and missed alarms.

[0025] Secondly, methods that process a single data source often require a lot of calculations to extract effective fire source features in complex environments. This large amount of computing demand not only increases the resource consumption of the system, but may also lead to slower response speeds and fail to meet the needs of real-time fire source detection. This delay can have serious consequences in terms of early warning and rapid response to fires, as timely fire source detection and alarm are crucial to reducing fire losses.

[0026] Therefore, in order to overcome the above shortcomings, this embodiment provides a forest fire source detection method based on multi-source information fusion, such as Figure 1 As shown, the method comprises the following steps: Step S100, collecting real-time data of the detection area, wherein the real-time data at least includes infrared image data, visible light image data and temperature data; Step S200, pre-processing the real-time data for consistency and fusion; Step S300, extracting key feature vectors of fire sources from the preprocessed real-time data, wherein the key feature vectors of fire sources include thermal radiation feature vectors, visual feature vectors, and temperature abnormal point feature vectors of fire sources; Step S400, fusing the key feature vectors of the fire source and generating a comprehensive feature vector; Step S500: Detect and locate the fire source through comprehensive feature vectors.

[0027] The above technical features are based on optical image data, infrared image data and temperature data, and use deep learning models to fuse different data sources. Finally, the fire source is detected and located through comprehensive features. The information of multiple data sources is comprehensively utilized to improve detection accuracy and robustness, so as to achieve efficient and accurate target detection.

[0028] In one embodiment, the method for collecting real-time data of the detection area in step S100 is: Use the following multiple sensors to synchronously obtain multi-source data around the fire source: Infrared camera: Obtain infrared images of the fire source and form infrared image data, which reflect the thermal radiation characteristics of the fire source and its surrounding objects. Infrared images are particularly important in low-light environments or smoke conditions because they do not rely on visible light.

[0029] Visible light camera: Obtain visible light images of the fire source and form visible light image data. These images can provide information such as the color, shape, and texture of the fire source. Under normal lighting conditions, visible light images can effectively supplement the missing information in infrared images.

[0030] Temperature sensor: measures the temperature data of the fire source and its surrounding environment in real time to form temperature data. This data is used to monitor the temperature change trend and abnormal points, and assist in identifying the location of the fire source.

[0031] The frequency and accuracy of data collection depend on the specific application scenario, and the sensor parameter settings can be adjusted as needed.

[0032] Data preprocessing is a key step to ensure that subsequent feature extraction and information fusion proceed smoothly. In one embodiment, the preprocessing method in step S200 includes the following operations: Denoising: Perform denoising processing on the collected infrared and visible light images, such as using median filtering or Gaussian filtering, to reduce noise interference in the image and improve image quality.

[0033] Image enhancement: Enhance the contrast and adjust the brightness of infrared and visible light images to ensure that the fire source features in the image are clearer and more visible.

[0034] Data alignment: Since the time and space coordinates of data collected by different sensors may be inconsistent, time and space alignment is required. Image registration technology is used to align infrared and visible light images into a unified coordinate system, and temperature data is synchronized with image data through interpolation methods.

[0035] Normalization: Normalize the temperature data so that it is in the same range as the image features to facilitate subsequent feature fusion.

[0036] In the process of fire source detection, feature extraction is a crucial step, which directly affects the accuracy of fire source identification and the effect of subsequent processing. In one embodiment, in step S300, for different data sources, a deep learning model is used to extract key features related to the fire source from infrared images, visible light images and temperature data. Exemplarily, the specific feature extraction process is as follows: Feature extraction of infrared images: The model selected is ResNet. Infrared images are mainly used to capture the thermal radiation characteristics of the fire source and its surrounding objects. In order to effectively extract these features, a deep residual network (ResNet) is used. By introducing a residual learning mechanism, ResNet can maintain the stability of the gradient in a very deep network structure, thereby extracting richer thermal radiation information. First, the infrared image is input into the ResNet network after preprocessing steps (such as denoising and enhancement). The first few layers of the network are responsible for extracting low-level features, such as edges and textures, which are essential for distinguishing the fire source from other heat sources. The middle layers extract higher-level features, such as the shape and intensity distribution of the fire source. The final fully connected layer combines these features and outputs a vector representing the thermal radiation characteristics of the fire source. This vector will be used for subsequent multi-source information fusion.

[0037] Feature extraction of visible light images: The model selected is YOLOv6. Visible light images provide information such as the color, shape, and texture of the fire source, which is very important under normal lighting conditions. The present invention uses the YOLO model for real-time target detection and feature extraction. The YOLO model is known for its efficient end-to-end detection capabilities and can simultaneously output the category and location of the target in a single forward propagation. First, the preprocessed visible light image is input into the YOLO model. The YOLO model divides the image into multiple grids, each of which is responsible for detecting whether there is a fire source inside it and generating a corresponding bounding box and confidence. The features extracted by the model include the color, shape, size, and edge information of the fire source, which are encoded as feature vectors. Unlike the thermal radiation features extracted by infrared images, the features of visible light images focus more on visual appearance information, which helps to improve the accuracy of fire source detection under good lighting conditions.

[0038] Feature extraction of temperature data: The model selected is LSTM (Long Short-Term Memory Network). The data from the temperature sensor provides information on temperature changes at various locations in the environment. These data are of time series nature, so the present invention uses the LSTM model for feature extraction. LSTM is known for its superior performance in processing time series data, and can effectively capture long-term dependencies and anomalies in temperature changes. First, the temperature data is preprocessed, such as denoising and normalization, to ensure data consistency. Then, the preprocessed temperature data is input into the LSTM network. LSTM analyzes the change pattern in the temperature data through its memory unit and gating mechanism, especially the sudden temperature rise phenomenon, which is usually a sign of the occurrence of a fire source. The output of LSTM is a temperature feature vector, which contains the temperature trend, fluctuation amplitude and other relevant information at the location of the fire source. This vector will be used in the subsequent fusion step together with the features of other data sources.

[0039] In one embodiment, the method of fusing the key feature vectors of the fire source in step S400 and generating a comprehensive feature vector is to use advanced multimodal information fusion technology to fuse feature vectors from different data sources to generate a comprehensive feature. The present invention uses the following method to achieve information fusion: (1) Multimodal fusion network based on attention mechanism: This network can automatically assign weights according to the importance of features from different data sources to achieve optimal feature fusion. Specifically, the attention mechanism focuses on the thermal signal features in infrared images, while combining the spatial information of visible light images and the temporal features of temperature data to generate a comprehensive feature vector.

[0040] (2) Feature-level fusion: In the feature extraction stage, the feature vectors extracted by each model are concatenated together and then processed through a fully connected layer for dimensionality reduction to generate a unified feature representation. This feature representation covers the multi-dimensional information of the fire source, which helps to improve the accuracy and robustness of detection.

[0041] In one embodiment, the method for detecting and locating a fire source by using a comprehensive feature vector in step S500 is: Based on the fused comprehensive feature vector, the fire source is detected and located in the last step. The fused feature vector is classified using a multi-layer perceptron (MLP) model or a classifier based on a support vector machine (SVM) to determine whether a fire source exists. The specific location and range of the fire source are further calculated using a regression network or a geometric algorithm. The positioning process combines the location information in the visible light image and the heat source distribution in the infrared image to generate accurate fire source location coordinates.

[0042] The forest fire source detection method based on multi-source information fusion of the present invention can realize efficient and accurate detection and positioning of fire sources in complex environments through the above steps. Compared with the prior art, the present invention has obvious advantages in detection accuracy and environmental adaptability.

[0043] At the same time, the present invention innovatively applies multi-source information fusion technology, innovatively combines infrared images, visible light images and temperature data, and uses the advantages of each data source through multi-source information fusion technology to achieve higher fire source detection accuracy and robustness. Especially in complex environments, multi-source fusion effectively reduces false alarms and missed alarms, and significantly improves the reliability of fire source detection.

[0044] At the same time, the present invention adopts the optimization and application of deep learning models, and adopts better deep learning models for feature extraction according to the characteristics of different data sources. Specifically, the ResNet model is used to extract thermal radiation features in infrared images, the YOLOv6 model is used to detect and extract fire source features in visible light images in real time, and the LSTM model is used to analyze the time series changes in temperature data. This highly targeted model selection and optimization significantly improves the accuracy and efficiency of fire source detection.

[0045] At the same time, the multimodal fusion network based on the attention mechanism of the present invention is used in the multi-source information fusion, which can dynamically adjust the weight according to the importance of different data sources to achieve the optimal feature fusion. This innovation ensures that the fire source detection system can play the best performance under various environmental conditions, and further improves the accuracy and real-time performance of the detection.

[0046] In another embodiment, if Figure 2 As shown, this embodiment provides a forest fire source detection system based on multi-source information fusion, including The acquisition module 100 is used to acquire real-time data of the detection area, wherein the real-time data at least includes infrared image data, visible light image data and temperature data; A preprocessing module 200 is used to preprocess the real-time data for consistency and fusion; The feature extraction module 300 extracts the key feature vector of the fire source from the preprocessed real-time data, wherein the key feature vector of the fire source includes the heat radiation feature vector, the visual feature vector and the temperature abnormal point feature vector of the fire source; A fusion module 400 is used to fuse the key feature vectors of the fire source and generate a comprehensive feature vector; and The judgment and positioning module 500 is used to detect and locate the fire source through the comprehensive feature vector.

[0047] The forest fire source detection system based on multi-source information fusion is based on the above-mentioned forest fire source detection method based on multi-source information fusion, and uses a deep learning model to fuse different data sources based on optical image data, infrared image data and temperature data; finally, the fire source is detected and located through comprehensive features; the information of multiple data sources is comprehensively utilized to improve the detection accuracy and robustness, so as to achieve efficient and accurate target detection.

[0048] It should be noted that, in the above embodiments, the description of each embodiment has its own focus, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0049] It should be understood by those skilled in the art that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may take the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-readable program codes.

[0050] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0051] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0052] These computer program instructions may also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0053] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement the above-mentioned forest fire source detection method based on multi-source information fusion.

[0054] If the module / unit integrated in the multi-source information fusion forest fire source detection system / terminal device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of each of the above-mentioned method embodiments can be implemented.

[0055] The computer program includes computer program code, which may be in source code form, object code form, executable file or some intermediate form, etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer readable media do not include electric carrier signals and telecommunication signals.

[0056] In the several embodiments provided in this application, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0057] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0058] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0059] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present application. The above embodiments are only used to illustrate the technical solution of the present application, not to limit it; although the present application is described in detail with reference to the above embodiments, ordinary technicians in this field should understand that: it is still possible to modify the technical solutions recorded in the above embodiments, or to replace some of the technical features therein by equivalent; and these modifications or replacements do not make the essence of the corresponding technical solution deviate from the spirit and scope of the technical solutions of each embodiment of the present application.

[0060] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A forest fire source detection method based on multi-source information fusion, characterized in that: include Collecting real-time data of the detection area, wherein the real-time data includes at least infrared image data, visible light image data and temperature data; Preprocessing the real-time data for consistency and fusion; Extracting key feature vectors of fire sources from preprocessed real-time data, wherein the key feature vectors of fire sources include thermal radiation feature vectors, visual feature vectors, and temperature anomaly point feature vectors of fire sources; Fuse the key feature vectors of the fire source and generate a comprehensive feature vector; Fire source detection and location through comprehensive feature vectors.

2. A forest fire source detection method based on multi-source information fusion according to claim 1, characterized in that: The method for real-time data collection of the detection area is: Use an infrared camera to obtain an infrared image of the detection area; Using a visible light camera to obtain a visible light image of the detection area; The temperature sensor is used to obtain the temperature data of the detection area.

3. A forest fire source detection method based on multi-source information fusion according to claim 1, characterized in that: The preprocessing method includes denoising, image enhancement, data alignment and normalization for real-time data.

4. A forest fire source detection method based on multi-source information fusion according to claim 1, characterized in that: The method for extracting the key feature vector of the fire source includes feature extraction of infrared images, and the feature extraction method of the infrared image is: After preprocessing, the infrared image data is input into the deep residual network, and the front layer of the deep residual network is used to extract low-level features, including edge and texture features. The middle layer of the deep residual network extracts high-level features, including the state and intensity of the fire source; The low-level features and high-level features are integrated through the deep residual network and finally through the fully connected layer to output the thermal radiation feature vector of the fire source.

5. A forest fire source detection method based on multi-source information fusion according to claim 4, characterized in that: Another method for extracting the key feature vector of the fire source includes feature extraction of visible light images. The method for extracting the feature of visible light images is: The preprocessed visible light image data is input into the YOLO model, which divides the image into multiple grids. Each grid is responsible for detecting whether there is a fire source inside it and generating a corresponding bounding box and confidence level. The features extracted by the YOLO model include the color, shape, size, and edge information of the fire source. These features are then encoded into feature vectors and output as visual feature vectors.

6. A forest fire source detection method based on multi-source information fusion according to claim 4, characterized in that: The method for extracting the key feature vector of the fire source also includes the feature extraction of temperature data. The method for extracting the feature of temperature data is: The preprocessed temperature data is input into the LSTM network. The LSTM network analyzes the change pattern in the temperature data through its memory unit and gating mechanism. Based on the sudden temperature rise phenomenon, the LSTM network outputs the temperature anomaly point feature vector, which contains the temperature trend and fluctuation amplitude of the fire source location.

7. A forest fire source detection method based on multi-source information fusion according to claim 1, characterized in that: The method for fusing the key feature vectors of the fire source and generating a comprehensive feature vector is: The multimodal fusion network using the attention mechanism first focuses on the thermal radiation feature vector in the infrared image, and then generates a comprehensive feature vector by combining the temporal features of the visual feature vector and the temperature anomaly point feature vector.

8. The forest fire source detection method based on multi-source information fusion according to claim 1, characterized in that: Methods for detecting and locating fire sources by using comprehensive feature vectors include: fire source detection methods and locating methods; The fire source detection method is: Using a multi-layer perceptron model or a support vector machine-based classifier, the fused comprehensive feature vector is classified to determine whether a fire source exists; The positioning method is: After the fire source is determined, the specific location and range of the fire source are calculated through regression networks or geometric algorithms; the positioning process combines the position information in the visible light image and the heat source distribution in the infrared image to generate accurate fire source location coordinates.

9. A forest fire source detection system based on multi-source information fusion, characterized in that: include An acquisition module, used for acquiring real-time data of the detection area, wherein the real-time data at least includes infrared image data, visible light image data and temperature data; A preprocessing module, used for preprocessing the real-time data for consistency and fusion; A feature extraction module extracts key feature vectors of fire sources from preprocessed real-time data, wherein the key feature vectors of fire sources include thermal radiation feature vectors, visual feature vectors, and temperature anomaly point feature vectors of fire sources; A fusion module, used to fuse the key feature vectors of the fire source and generate a comprehensive feature vector; and The judgment and positioning module is used to detect and locate the fire source through comprehensive feature vectors.

10. A storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the forest fire source detection method based on multi-source information fusion as described in any one of claims 1 to 8.

Images