An electrical equipment fire extinguishing method based on image recognition and related equipment

Abstract

The application provides an electrical equipment fire extinguishing method based on image recognition, which comprises the following steps: acquiring a monitoring image inside electrical equipment; performing fire source identification on the monitoring image to obtain a fire source identification result; determining a fire source position in the internal space of the electrical equipment based on the fire source identification result; determining a target fire extinguishing piece from a plurality of fire extinguishing pieces based on the fire source position and the fire extinguishing range of the fire extinguishing piece; and controlling the target fire extinguishing piece to trigger to extinguish the fire source position. The fire source identification result is obtained by performing fire source identification on the monitoring image inside the electrical equipment, the fire source position is determined in the internal space of the electrical equipment based on the fire source identification result, the target fire extinguishing piece is determined from the plurality of fire extinguishing pieces based on the fire source position and the fire extinguishing range of the fire extinguishing piece, and the target fire extinguishing piece is controlled to trigger to extinguish the fire source position, so that the fire fighting work intensity of the electrical equipment is effectively improved, and the work risk of the fire fighters is reduced.

Description

A fire extinguishing method for electrical equipment based on image recognition and related equipment. Technical Field

[0001] This invention relates to the field of image processing, and more particularly to a fire extinguishing method and related equipment for electrical equipment based on image recognition. Background Technology

[0002] Currently, the common firefighting method in my country involves firefighters entering the fire scene and spraying water from hoses onto a specific area requiring firefighting. This method has many drawbacks: 1. In some extremely large fires, firefighters cannot enter to extinguish them, leading to the fire spreading and escalating further; 2. It takes time for firefighters to reach the scene, yet fires need to be detected and extinguished promptly. In many cases, small fires fail to be extinguished in time and develop into major fires; 3. Fire scenes are complex and unpredictable, posing a risk to firefighters' lives. This is because the smoke and toxic gases at fire scenes are very harmful to firefighters, and they may lose their lives due to gas poisoning. Summary of the Invention

[0003] This invention provides an image recognition-based fire extinguishing method for electrical equipment, aiming to offer an efficient and intelligent approach to improve the firefighting capabilities of electrical equipment. By identifying fire sources in monitoring images inside the electrical equipment, the method determines the location of the fire source within the equipment's interior space. Using the fire source location and the extinguishing range of fire extinguishing discs, a target fire extinguishing disc is identified from among several discs. This target disc is then triggered to extinguish the fire at the source, effectively increasing the firefighting capabilities of electrical equipment and reducing the operational risks for firefighters.

[0004] In a first aspect, embodiments of the present invention provide a fire extinguishing method for electrical equipment based on image recognition. The electrical equipment is equipped with an image device and a plurality of distributed fire extinguishing discs. The image device is used to acquire images of the interior of the electrical equipment. The method includes the following steps:

[0005] Acquire monitoring images of the interior of the electrical equipment;

[0006] Fire source identification is performed on the monitoring image to obtain the fire source identification result;

[0007] Based on the fire source identification results, the location of the fire source is determined in the internal space of the electrical equipment;

[0008] Based on the location of the fire source and the extinguishing range of the fire extinguishing disc, the target fire extinguishing disc is determined from among the fire extinguishing discs.

[0009] The target fire extinguishing disc is triggered to extinguish the fire at the fire source location.

[0010] Optionally, the step of performing fire source identification on the monitored image to obtain the fire source identification result includes:

[0011] The monitored image is processed by a trained fire source recognition model to obtain a target recognition result. The target recognition result includes a target detection box and the confidence level corresponding to the target detection box. The target recognition result includes at least one of the following: electric spark recognition result, flame recognition result, and smoke recognition result.

[0012] Based on the target identification results, the fire source identification results are determined.

[0013] Optionally, determining the acquisition of the recognition result based on the target recognition result includes:

[0014] If the target recognition result includes multiple target detection boxes;

[0015] The target detection boxes with a confidence level greater than or equal to the preset confidence threshold are then determined as fire source target boxes;

[0016] The fire source identification result is determined based on the fire source target box.

[0017] Optionally, determining the fire source identification result based on the fire source target box includes:

[0018] If the number of fire source target boxes is one, the fire source identification result is directly determined based on the fire source target box, and the fire source identification result includes the fire source target box;

[0019] If there are two or more fire source target boxes, then the two or more fire source target boxes are merged and fused to obtain a fused target box;

[0020] The fire source identification result is determined based on the fused target box, and the fire source identification result includes the fused target box.

[0021] Optionally, determining the location of the fire source within the internal space of the electrical equipment based on the fire source identification result includes:

[0022] If the fire source identification result includes the fire source target box, then a first image region is determined in the monitoring image based on the fire source target box;

[0023] Based on the first image region and the spatial mapping relationship between the monitoring image and the internal space of the electrical equipment, the location of the fire source is determined in the internal space of the electrical equipment.

[0024] Optionally, determining the location of the fire source within the internal space of the electrical equipment based on the fire source identification result includes:

[0025] If the fire source identification result includes the fused target box, then based on the fused target box, the heat map corresponding to the fused target box is determined in the monitoring image;

[0026] In the heat map, regions with heat values ​​greater than or equal to a heat value threshold are selected to obtain the second image region;

[0027] Based on the second image region and the spatial mapping relationship between the monitoring image and the internal space of the electrical equipment, the location of the fire source is determined in the internal space of the electrical equipment.

[0028] Optionally, determining the target fire extinguishing pad from among the plurality of fire extinguishing pads based on the location of the fire source and the extinguishing range of the fire extinguishing pad includes:

[0029] Calculate the distance between the fire source location and each of the fire extinguishing discs;

[0030] The fire extinguishing discs located at a distance smaller than the fire extinguishing range are identified as candidate fire extinguishing discs;

[0031] Among the candidate fire extinguishing discs, the target fire extinguishing disc is determined by selecting the one with the largest coverage area and the fewest fire extinguishing discs.

[0032] Secondly, embodiments of the present invention also provide an image recognition-based fire extinguishing device for electrical equipment, the image recognition-based fire extinguishing device for electrical equipment comprising:

[0033] The acquisition module is used to acquire monitoring images of the interior of the electrical equipment;

[0034] The fire source identification module is used to identify the source of the fire source in the monitoring image and obtain the fire source identification result.

[0035] The first determining module is used to determine the location of the fire source in the internal space of the electrical equipment based on the fire source identification result;

[0036] The second determining module is used to determine the target fire extinguishing disc among a plurality of fire extinguishing discs based on the location of the fire source and the extinguishing range of the fire extinguishing disc.

[0037] The control module is used to control the target fire extinguishing disc to trigger, thereby extinguishing the fire at the fire source location.

[0038] Thirdly, embodiments of the present invention provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps in the image recognition-based electrical equipment fire extinguishing method provided in embodiments of the present invention.

[0039] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps in the image recognition-based electrical equipment fire extinguishing method provided in the embodiments of the present invention.

[0040] In this embodiment of the invention, monitoring images inside electrical equipment are acquired; fire source identification is performed on the monitoring images to obtain fire source identification results; based on the fire source identification results, the location of the fire source is determined within the internal space of the electrical equipment; based on the fire source location and the extinguishing range of the fire extinguishing discs, a target fire extinguishing disc is identified from among several fire extinguishing discs; the target fire extinguishing disc is controlled to trigger, thereby extinguishing the fire at the fire source location. By performing fire source identification on monitoring images inside electrical equipment, obtaining fire source identification results, and determining the location of the fire source within the internal space of the electrical equipment based on the fire source identification results, and using the fire source location and the extinguishing range of the fire extinguishing discs to identify a target fire extinguishing disc from among several fire extinguishing discs, the target fire extinguishing disc is controlled to trigger, thereby extinguishing the fire at the fire source location. This effectively improves the fire protection intensity of electrical equipment and reduces the working risks for firefighters. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 is a flowchart of an image recognition-based fire extinguishing method for electrical equipment provided in an embodiment of the present invention;

[0043] Figure 2 is a schematic diagram of the structure of an electrical equipment fire extinguishing device based on image recognition provided in an embodiment of the present invention;

[0044] Figure 3 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] As shown in Figure 1, Figure 1 is a schematic flowchart of an image recognition-based fire extinguishing method for electrical equipment provided in an embodiment of the present invention.

[0047] In this embodiment of the invention, an image device and a plurality of distributed fire extinguishing discs are installed inside the aforementioned electrical equipment. The image device is used to acquire images of the interior of the electrical equipment. The image recognition-based fire extinguishing method for electrical equipment includes the following steps:

[0048] 101. Obtain monitoring images of the interior of electrical equipment.

[0049] In this embodiment of the invention, the above-described image recognition-based fire extinguishing method for electrical equipment can be applied to a task scheduling system. The task scheduling system includes at least one electrical device, multiple fire extinguishing disc devices, and a task scheduling module. The task scheduling module can be a task scheduling platform or an electronic device. The electronic device can be a smart gateway, a smart network device, or a server, etc. The task scheduling platform can be built based on a server or a distributed server. The task scheduling platform or electronic device can include an electrical device interface, a fire extinguishing disc device interface, a task processing program, etc. The task scheduling platform or electronic device can communicate with the electrical device through the electrical device interface, thereby receiving status data of the electrical device and sending tasks to the fire extinguishing discs. The task scheduling platform can also communicate with the fire extinguishing disc devices through the electrical device interface, thereby receiving status data of the electrical device and sending tasks to the fire extinguishing discs.

[0050] The aforementioned imaging device is used to capture real-time images of the internal components of the imaging equipment, providing visual information to promptly detect any anomalies and aiding in diagnosis or maintenance. The fire extinguishing disc interface connects to the aforementioned fire extinguishing disc, which may contain an electric heating element (such as a heating wire or heating element). Heating of the electric heating element triggers the fire extinguishing disc. The fire extinguishing disc may be composed of a silicone substrate and perfluorohexanone microcapsules, designed for rapid response in the event of a fire or other emergency. Upon receiving a fire extinguishing command, the electric heating element is energized and heated. When the temperature reaches a certain level or an open flame is encountered, the perfluorohexanone microcapsules rupture, releasing perfluorohexanone, which rapidly absorbs heat and isolates the air, lowering the temperature below the flame point temperature to automatically extinguish the fire, controlling its spread and preventing its further spread.

[0051] The aforementioned monitoring images are visual information monitored in real time inside the equipment. These monitoring images can be understood as image data that needs to be analyzed to find out if there is a fire source, including the equipment's operating status, temperature change status, smoke generation, fire signs, or other possible indications of abnormality.

[0052] It should be noted that by acquiring monitoring images of electrical equipment, fire sources or other potential problems can be detected in a timely manner, and corresponding fire extinguishing measures can be taken.

[0053] 102. Perform fire source identification on the monitoring images to obtain the fire source identification results.

[0054] In this embodiment of the invention, a fire source identification model can be used to identify fire sources in the monitoring image to obtain the fire source identification result.

[0055] The above-mentioned fire source identification is a process of detecting fire sources from monitoring images.

[0056] The aforementioned fire source identification model can be a fire source identification model built based on deep learning or machine learning, such as a convolutional neural network (CNN) or a recurrent neural network (RNN). This fire source identification model is trained using a sample image training dataset, which includes image data of fire sources and corresponding fire source label data. The training can be supervised training, which can be understood as adjusting the model's parameters using a set of samples of known categories to achieve the required performance.

[0057] The above fire source identification results include fire source target boxes, which can be understood as detection boxes marked in images or videos to identify potential fire source areas, including information such as the location and size of the fire source.

[0058] 103. Based on the fire source identification results, determine the location of the fire source in the internal space of the electrical equipment.

[0059] In this embodiment of the invention, the location of the fire source can be determined in the internal space of the electrical equipment based on the fire source identification result.

[0060] The aforementioned fire source location can be understood as the place where the flame originated.

[0061] In one possible embodiment, the location of the fire source can be determined in the internal space of the electrical equipment based on the image area of ​​the fire source target frame in the monitoring image and the spatial mapping relationship between the monitoring image and the internal space of the electrical equipment.

[0062] The spatial mapping relationship described above can be understood as the process of mapping elements in one space to another. Similarly, the spatial mapping relationship between the monitoring image and the internal space of the electrical equipment can be understood as associating the monitoring image with the internal space of the electrical equipment, establishing a mapping relationship between the two.

[0063] It should be noted that by monitoring the spatial mapping relationship between the image and the internal space of the electrical equipment, the location of the fire source can be determined in the internal space of the electrical equipment based on the image area of ​​the fire source target frame in the monitoring image.

[0064] 104. Based on the location of the fire source and the extinguishing range of the fire extinguishing disc, identify the target fire extinguishing disc from among several fire extinguishing discs.

[0065] In this embodiment of the invention, the extinguishing range of the fire extinguishing disc can be understood as the spray distance and coverage area of ​​the fire extinguishing disc, and the extinguishing range of the fire extinguishing disc determines the size of the space it can extinguish.

[0066] Specifically, the distance between the fire source and each fire extinguishing disc can be calculated based on the location of the fire and the extinguishing range of the fire extinguishing disc. Then, based on the extinguishing range of the fire extinguishing disc, the fire extinguishing discs that are less than their extinguishing range from the fire source are identified as candidate fire extinguishing discs. Among the candidate fire extinguishing discs, the most suitable fire extinguishing disc is selected as the target fire extinguishing disc with the goal of maximizing the coverage area and minimizing the number of fire extinguishing discs used. This ensures that the fewest fire extinguishing discs are used during fire extinguishing while maximizing the fire extinguishing effect, effectively covering the area that needs to be extinguished without wasting resources.

[0067] 105. Control the target fire extinguishing disc to trigger, thereby extinguishing the fire at the fire source location.

[0068] In this embodiment of the invention, after the target fire extinguishing disc is identified, the controller of the electrical equipment will trigger and activate these target fire extinguishing discs. The internal electric heating component of the fire extinguishing disc is energized and heated. When the temperature reaches a certain level or when it encounters an open flame, the perfluorohexanone microcapsules of the fire extinguishing disc rupture, releasing perfluorohexanone, which quickly absorbs heat and isolates the air, thereby reducing the temperature below the flame point temperature and achieving automatic fire extinguishing.

[0069] In this embodiment of the invention, the invention effectively improves the fire protection intensity of electrical equipment and reduces the working risks of firefighters by accurately locating the fire source, intelligently selecting appropriate fire extinguishing resources, and efficiently executing fire extinguishing operations.

[0070] In this embodiment of the invention, monitoring images inside electrical equipment are acquired; fire source identification is performed on the monitoring images to obtain fire source identification results; based on the fire source identification results, the location of the fire source is determined within the internal space of the electrical equipment; based on the fire source location and the extinguishing range of the fire extinguishing discs, a target fire extinguishing disc is identified from among several fire extinguishing discs; the target fire extinguishing disc is controlled to trigger, thereby extinguishing the fire at the fire source location. By performing fire source identification on monitoring images inside electrical equipment, obtaining fire source identification results, and determining the fire source location within the internal space of the electrical equipment based on the fire source identification results, and using the fire source location and the extinguishing range of the fire extinguishing discs to identify a target fire extinguishing disc from among several fire extinguishing discs, the target fire extinguishing disc is controlled to trigger, thereby extinguishing the fire at the fire source location. This method can quickly and accurately locate the fire source and effectively extinguish the fire, effectively improving the fire protection intensity of electrical equipment and reducing the working risks for firefighters.

[0071] It is understood that in the specific implementation of this application, data related to electrical equipment, image data, task data, etc. are involved. When the embodiments in this application are applied to specific products or technologies, user permission or consent is required. Furthermore, the collection, use and processing of related data, as well as the training, deployment and invocation of large language models, must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0072] Optionally, in the step of identifying fire sources in the monitoring image and obtaining the fire source identification result, the monitoring image can be processed by a trained fire source identification model to obtain the target identification result; based on the target identification result, the fire source identification result is determined.

[0073] In this embodiment of the invention, the target recognition result includes a target detection box and a corresponding confidence score. The target recognition result includes at least one of the following: spark recognition result, flame recognition result, and smoke recognition result. The target can be understood as a fire-related phenomenon such as an electric spark, flame, or smoke. The target recognition result includes information on the locations of all potential fire sources detected in the image and their corresponding confidence scores.

[0074] The aforementioned monitoring images are image data that need to be analyzed to determine if a fire source exists.

[0075] The trained fire source recognition model described above can be a fire source recognition model built based on deep learning or machine learning, such as a convolutional neural network (CNN) or a recurrent neural network (RNN). The trained fire source recognition model is obtained by training the model on a sample image training dataset, which includes image data of fire sources and corresponding fire source label data.

[0076] Specifically, the aforementioned fire source labeling data may include electric spark labels, flame labels, or smoke labels. Sample images are input into the fire source recognition model to be trained. The model outputs the fire source recognition result corresponding to the sample image. A loss function is used to calculate the error loss between the fire source recognition result and the corresponding fire source label data. The model parameters are adjusted to minimize this error loss. This adjustment process is iterated until the number of iterations or the error loss converges, at which point training stops, resulting in a trained fire source recognition model. The aforementioned fire source labeling data may include electric spark labels, flame labels, and smoke labels. Electric spark labels can be electric spark bounding boxes, flame labels are flame bounding boxes, and smoke labels are smoke bounding boxes. The fire source recognition results output by the model to be trained may include electric spark recognition results, flame recognition results, and smoke recognition results. Electric spark recognition results are electric spark bounding boxes, flame recognition results are flame bounding boxes, and smoke recognition results are smoke bounding boxes.

[0077] The loss function described above can be calculated as follows:

[0078]

[0079] LOSS=γ1LOSS1+γ2LOSS2+γ3LOSS3

[0080] Wherein, LOSS represents the error loss between the fire source identification result corresponding to the sample image and the fire source label data corresponding to the sample image; LOSS1 represents the error loss between the electric spark identification result and the electric spark label; LOSS2 represents the error loss between the flame identification result and the flame label; LOSS3 represents the error loss between the smoke identification result and the smoke label; and N A M represents the number of electrical spark bounding boxes corresponding to the sample image. A This indicates the number of electric spark bounding boxes corresponding to the fire source identification result of the sample image. This represents the i-th electrical discharge annotation box. N represents the j-th electrical discharge target box. B M represents the number of flame annotation boxes corresponding to the sample images. B This indicates the number of fire source identification results (flame target boxes) corresponding to the sample images. This represents the i-th flame label box. N represents the j-th flame target box. C M represents the number of smoke-marked boxes corresponding to the sample image. C This indicates the number of smoke target boxes representing the fire source identification results corresponding to the sample image. This represents the i-th smoke label box. Let γ1, γ2, and γ3 represent the j-th smoke target box, where γ1, γ2, and γ3 are empirical coefficients set manually, and the sum of γ1, γ2, and γ3 is 1.

[0081] The target recognition process described above can be understood as analyzing monitoring images to identify fire source characteristics. These fire source characteristics include features such as electrical sparks, flames, or smoke.

[0082] Optionally, in the step of determining the fire source identification result based on the target identification result, if the target identification result includes multiple target detection boxes, then the target detection boxes with a confidence level greater than or equal to a preset confidence threshold are identified as fire source target boxes; and the fire source identification result is determined based on the fire source target boxes.

[0083] In this embodiment of the invention, the aforementioned preset confidence threshold is a confidence threshold preset by the system.

[0084] The aforementioned fire source target bounding box can be understood as a detection box marked in an image or video to identify potential fire source areas.

[0085] Understandably, when multiple target detection boxes are detected, those with a confidence level greater than or equal to a preset threshold that contain a fire source are selected and identified as fire source target boxes. Only when the confidence level of a detection box is higher than the preset confidence threshold will it be considered a valid fire source target box.

[0086] Optionally, in the step of determining the fire source identification result based on the fire source target box, if there is only one fire source target box, the fire source identification result is directly determined based on the fire source target box; if there are two or more fire source target boxes, the two or more fire source target boxes are merged and fused to obtain a fused target box; and the fire source identification result is determined based on the fused target box.

[0087] In this embodiment of the invention, the above-mentioned fire source identification result includes a fire source target box.

[0088] The above fire source identification results include fused target bounding boxes.

[0089] The aforementioned fire source target box is a rectangular box marked in an image or video to identify potential fire source areas.

[0090] The above merging and fusion can be understood as a process of combining multiple similar objects or fire source target boxes into a whole to improve the accuracy of recognition.

[0091] The above fire source identification results include confirmation that a fire source does exist in a certain area, as well as information such as the specific location and size of the fire source.

[0092] It should be noted that when only one fire source target box is detected, the fire source identification result can be directly determined based on the fire source target box. When there are two or more fire source target boxes, these target boxes need to be merged and fused to form a fused target box. Finally, the fire source identification result is determined based on the fused target box. The above merging and fusion can be achieved by connecting the boundaries of multiple target boxes to form a minimum rectangle, thereby obtaining the fused target box.

[0093] Optionally, in the step of determining the location of the fire source in the internal space of the electrical equipment based on the fire source identification result, if the fire source identification result includes a fire source target box, then a first image region is determined in the monitoring image based on the fire source target box; and the location of the fire source is determined in the internal space of the electrical equipment based on the first image region and the spatial mapping relationship between the monitoring image and the internal space of the electrical equipment.

[0094] In this embodiment of the invention, the above-mentioned fire source target box is used as a label box to represent the location and size of an existing fire source.

[0095] The above spatial mapping relationship can be understood as the process of mapping elements in one space to another space.

[0096] The spatial mapping relationship between the aforementioned monitoring images and the internal space of electrical equipment can be established by associating the monitored image data with the internal structure or spatial layout of the electrical equipment. Specifically, it is necessary to extract key feature points, edges, textures, and other information from the monitoring images. A three-dimensional model can be predefined within the internal space of the electrical equipment, describing the equipment's geometry and component layout. By matching and aligning feature points in the monitoring images with corresponding points or regions in the internal space model of the electrical equipment, a mapping relationship between the image space and the equipment's internal space can be established. Through this spatial mapping relationship, the location of a fire source or other abnormal conditions within the electrical equipment's internal space can be pinpointed. For example, if the monitoring image shows an abnormally high temperature in a certain area, the exact location of this hotspot inside the equipment can be determined through the mapping relationship.

[0097] It should be noted that after identifying the fire source target box, the first image area can be marked using the fire source target box, and the specific location of the fire source can be accurately determined in the internal space of the electrical equipment by using the first image area and the spatial mapping relationship between the monitoring image and the internal space of the electrical equipment.

[0098] Optionally, in the step of determining the location of the fire source in the internal space of the electrical equipment based on the fire source identification result, if the fire source identification result includes a fusion target box, then based on the fusion target box, a heat map corresponding to the fusion target box is determined in the monitoring image; in the heat map, areas with heat values ​​greater than or equal to a heat value threshold are selected to obtain a second image region; based on the second image region and the spatial mapping relationship between the monitoring image and the internal space of the electrical equipment, the location of the fire source is determined in the internal space of the electrical equipment.

[0099] In this embodiment of the invention, the aforementioned fused target box improves the accuracy of target detection by combining information from multiple detection boxes. The aforementioned imaging device may have an infrared module to acquire infrared images of the interior of the electrical equipment, thereby obtaining a monitoring heat map of the electrical equipment. Since the monitoring heat map is the same size as the monitoring image, the heat map corresponding to the fused target box can be found within the monitoring heat map.

[0100] The heatmap above uses color variations to represent the density or magnitude of data. In the heatmap, areas with higher values ​​are represented by warm colors (such as red), and lower values ​​by cool colors (such as blue). This heatmap can reflect information such as the temperature distribution or flame intensity of the target frame being fused.

[0101] The above-mentioned heat value threshold is a pre-set heat value threshold.

[0102] Specifically, each pixel or region in the heat map is compared based on its heat value. If the heat value of a pixel or region is greater than or equal to a heat value threshold, then the region that is greater than or equal to the heat value threshold is taken as the second image region. This allows for a more intuitive identification of which parts of the image are hotspots or high-energy regions, such as the location of the fire source in fire monitoring.

[0103] By monitoring the spatial mapping relationship between images and the internal space of electrical equipment, the location of fire sources or other abnormal conditions within the equipment's interior can be pinpointed. For example, if a monitoring image shows an abnormally high temperature in a certain area, the exact location of this hotspot inside the equipment can be determined through the mapping relationship.

[0104] Optionally, in the step of determining the target fire extinguishing disc from several fire extinguishing discs based on the location of the fire source and the extinguishing range of the fire extinguishing disc, the distance between the fire source location and each fire extinguishing disc can be calculated; fire extinguishing discs with a distance smaller than the extinguishing range can be identified as candidate fire extinguishing discs; among the candidate fire extinguishing discs, the one with the largest coverage area and the fewest fire extinguishing discs can be selected as the target fire extinguishing disc.

[0105] In this embodiment of the invention, when determining which fire extinguishing pads to use to extinguish a fire source, it is necessary to calculate the distance between the location of each fire extinguishing pad and the fire source. If the distance between a fire extinguishing pad and the fire source is less than its effective extinguishing range, then this fire extinguishing pad is considered an effective candidate fire extinguishing pad and can be considered for use to extinguish the fire.

[0106] Furthermore, among multiple candidate fire extinguishing discs, the one with the largest coverage area should be selected, and the number used should be minimized. Specifically, a balance needs to be found that effectively covers the area requiring fire suppression without wasting resources.

[0107] In this embodiment of the invention, the distance between the fire source and each fire extinguishing disc is calculated. Then, based on the extinguishing range of the fire extinguishing disc, fire extinguishing discs that are less than their extinguishing range from the fire source are identified as candidate fire extinguishing discs. Among the candidate fire extinguishing discs, the most suitable fire extinguishing disc is selected as the target fire extinguishing disc with the largest coverage area and the fewest number of fire extinguishing discs used. This is to ensure that the fewest fire extinguishing discs are used during fire extinguishing and to maximize the fire extinguishing effect, so as to effectively cover the area that needs to be extinguished without wasting resources.

[0108] As shown in Figure 2, this embodiment of the invention provides an image recognition-based fire extinguishing device for electrical equipment, which includes:

[0109] The acquisition module 201 is used to acquire monitoring images inside the electrical equipment;

[0110] Fire source identification module 202 is used to identify fire sources in the monitoring image and obtain fire source identification results;

[0111] The first determining module 203 is used to determine the location of the fire source in the internal space of the electrical equipment based on the fire source identification result;

[0112] The second determining module 204 is used to determine the target fire extinguishing disc among a plurality of fire extinguishing discs based on the location of the fire source and the fire extinguishing range of the fire extinguishing disc.

[0113] The control module 205 is used to control the target fire extinguishing disc to trigger, thereby extinguishing the fire at the fire source location.

[0114] Optionally, the fire source identification module 202 is further configured to perform target identification processing on the monitoring image using a trained fire source identification model to obtain a target identification result. The target identification result includes a target detection box and the confidence level corresponding to the target detection box. The target identification result includes at least one of an electric spark identification result, a flame identification result, and a smoke identification result. Based on the target identification result, a fire source identification result is determined.

[0115] Optionally, the fire source identification module 202 is further configured to: if the target identification result includes multiple target detection boxes; determine the target detection boxes with a confidence level greater than or equal to a preset confidence threshold as fire source target boxes; and determine the fire source identification result based on the fire source target boxes.

[0116] Optionally, the fire source identification module 202 is further configured to: if the number of fire source target boxes is one, directly determine the fire source identification result based on the fire source target box, wherein the fire source identification result includes the fire source target box; if the number of fire source target boxes is two or more, merge and fuse the two or more fire source target boxes to obtain a fused target box; determine the fire source identification result based on the fused target box, wherein the fire source identification result includes the fused target box.

[0117] Optionally, the first determining module 203 is further configured to, if the fire source identification result includes the fire source target frame, determine a first image region in the monitoring image based on the fire source target frame; and determine the fire source location in the internal space of the electrical equipment based on the first image region and the spatial mapping relationship between the monitoring image and the internal space of the electrical equipment.

[0118] Optionally, the first determining module 203 is further configured to, if the fire source identification result includes the fusion target box, determine the heat map corresponding to the fusion target box in the monitoring image based on the fusion target box; in the heat map, filter out areas with heat values ​​greater than or equal to a heat value threshold to obtain the second image area; and based on the second image area and the spatial mapping relationship between the monitoring image and the internal space of the electrical equipment, determine the fire source location in the internal space of the electrical equipment.

[0119] Optionally, the second determining module 204 is further configured to calculate the distance between the fire source location and each of the fire extinguishing discs; determine the fire extinguishing discs whose distance is less than the fire extinguishing range as candidate fire extinguishing discs; and determine the target fire extinguishing disc by selecting the one with the largest coverage area and the fewest fire extinguishing discs among the candidate fire extinguishing discs.

[0120] As shown in Figure 3, this embodiment of the invention also provides an electronic device, including a processor, which can execute any of the above-mentioned image recognition-based fire extinguishing methods for electrical equipment.

[0121] Specifically, it includes a processor 301 and a memory 302, as well as a computer program stored in the memory 302 and capable of running on the processor 301, which executes an image recognition-based electrical equipment fire extinguishing method, wherein:

[0122] The processor 301 executes the calculator program for an image recognition-based fire extinguishing method for electrical equipment stored in the memory 302, and performs the following steps:

[0123] Acquire monitoring images of the interior of the electrical equipment;

[0124] Fire source identification is performed on the monitoring image to obtain the fire source identification result;

[0125] Based on the fire source identification results, the location of the fire source is determined in the internal space of the electrical equipment;

[0126] Based on the location of the fire source and the extinguishing range of the fire extinguishing disc, the target fire extinguishing disc is determined from among the fire extinguishing discs.

[0127] The target fire extinguishing disc is triggered to extinguish the fire at the fire source location.

[0128] Optionally, the process of processor 301 performing fire source identification on the monitored image to obtain a fire source identification result includes:

[0129] The monitored image is processed by a trained fire source recognition model to obtain a target recognition result. The target recognition result includes a target detection box and the confidence level corresponding to the target detection box. The target recognition result includes at least one of the following: electric spark recognition result, flame recognition result, and smoke recognition result.

[0130] Based on the target identification results, the fire source identification results are determined.

[0131] Optionally, the process executed by processor 301 to determine the acquisition of the recognition result based on the target recognition result includes:

[0132] If the target recognition result includes multiple target detection boxes;

[0133] The target detection boxes with a confidence level greater than or equal to the preset confidence threshold are then determined as fire source target boxes;

[0134] The fire source identification result is determined based on the fire source target box.

[0135] Optionally, the process of determining the fire source identification result based on the fire source target box executed by the processor 301 includes:

[0136] If the number of fire source target boxes is one, the fire source identification result is directly determined based on the fire source target box, and the fire source identification result includes the fire source target box;

[0137] If there are two or more fire source target boxes, then the two or more fire source target boxes are merged and fused to obtain a fused target box;

[0138] The fire source identification result is determined based on the fused target box, and the fire source identification result includes the fused target box.

[0139] Optionally, the process executed by processor 301 to determine the location of the fire source within the internal space of the electrical equipment based on the fire source identification result includes:

[0140] If the fire source identification result includes the fire source target box, then a first image region is determined in the monitoring image based on the fire source target box;

[0141] Based on the first image region and the spatial mapping relationship between the monitoring image and the internal space of the electrical equipment, the location of the fire source is determined in the internal space of the electrical equipment.

[0142] Optionally, the process executed by processor 301 to determine the location of the fire source within the internal space of the electrical equipment based on the fire source identification result includes:

[0143] If the fire source identification result includes the fused target box, then based on the fused target box, the heat map corresponding to the fused target box is determined in the monitoring image;

[0144] In the heat map, regions with heat values ​​greater than or equal to a heat value threshold are selected to obtain the second image region;

[0145] Based on the second image region and the spatial mapping relationship between the monitoring image and the internal space of the electrical equipment, the location of the fire source is determined in the internal space of the electrical equipment.

[0146] Optionally, the process executed by processor 301 to determine the target fire extinguishing disc among a plurality of fire extinguishing discs based on the location of the fire source and the extinguishing range of the fire extinguishing disc includes:

[0147] Calculate the distance between the fire source location and each of the fire extinguishing discs;

[0148] The fire extinguishing discs located at a distance smaller than the fire extinguishing range are identified as candidate fire extinguishing discs;

[0149] Among the candidate fire extinguishing discs, the target fire extinguishing disc is determined by selecting the one with the largest coverage area and the fewest fire extinguishing discs.

[0150] This invention also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the various processes of the image recognition-based electrical equipment fire extinguishing method or the application-side image recognition-based electrical equipment fire extinguishing method provided in this invention, and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0151] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0152] The above description discloses only preferred embodiments of the present invention and should not be construed as limiting the scope of the present invention. Therefore, equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A fire extinguishing method for electrical equipment based on image recognition, wherein the electrical equipment is equipped with an image device and a plurality of distributed fire extinguishing discs, the image device being used to acquire images of the interior of the electrical equipment, characterized in that, The method includes the following steps: acquiring monitoring images of the interior of the electrical equipment; performing fire source identification on the monitoring images to obtain fire source identification results; specifically, performing target identification processing on the monitoring images using a trained fire source identification model to obtain target identification results, the target identification results including target detection boxes and the confidence scores corresponding to the target detection boxes, the target identification results including at least one of electric spark identification results, flame identification results, and smoke identification results; determining the fire source identification result based on the target identification results; wherein, the trained fire source identification model is obtained by training a fire source identification model to be trained using a sample image training dataset, the sample image training dataset including image data of fire sources and corresponding fire source label data, the fire source label data including electric spark labels, flame labels, or smoke labels; inputting sample images into the fire source identification model to be trained, and outputting sample images through the identification model to be trained. The fire source identification result corresponding to this image is used to calculate the error loss between the fire source identification result corresponding to the sample image and the fire source label data corresponding to the sample image. Minimizing the error loss is the optimization objective. The model parameters of the identification model to be trained are adjusted, and the adjustment process is iterated until the number of iterations or the error loss converges. Training then stops, resulting in a trained fire source identification model. The fire source identification results output by the identification model to be trained include electric spark identification results, flame identification results, and smoke identification results. The electric spark identification result is an electric spark target box, the flame identification result is a flame target box, and the smoke identification result is a smoke target box. Based on the fire source identification results, the fire source location is determined within the internal space of the electrical equipment. Based on the fire source location and the extinguishing range of the fire extinguishing disc, a target fire extinguishing disc is determined from among several fire extinguishing discs. The target fire extinguishing disc is triggered to extinguish the fire at the fire source location.

2. The fire extinguishing method for electrical equipment based on image recognition as described in claim 1, characterized in that, The step of determining the fire source identification result based on the target identification result includes: if the target identification result includes multiple target detection boxes; then the target detection boxes with a confidence level greater than or equal to a preset confidence threshold are identified as fire source target boxes; and the fire source identification result is determined based on the fire source target boxes.

3. The fire extinguishing method for electrical equipment based on image recognition as described in claim 2, characterized in that, The step of determining the fire source identification result based on the fire source target box includes: if the number of fire source target boxes is one, then the fire source identification result is directly determined based on the fire source target box, and the fire source identification result includes the fire source target box; if the number of fire source target boxes is two or more, then the two or more fire source target boxes are merged and fused to obtain a fused target box; the fire source identification result is determined based on the fused target box, and the fire source identification result includes the fused target box.

4. The fire extinguishing method for electrical equipment based on image recognition as described in claim 3, characterized in that, The step of determining the location of the fire source in the internal space of the electrical equipment based on the fire source identification result includes: if the fire source identification result includes the fire source target frame, then determining a first image region in the monitoring image based on the fire source target frame; and determining the location of the fire source in the internal space of the electrical equipment based on the first image region and the spatial mapping relationship between the monitoring image and the internal space of the electrical equipment.

5. The fire extinguishing method for electrical equipment based on image recognition as described in claim 3, characterized in that, The step of determining the location of the fire source in the internal space of the electrical equipment based on the fire source identification result includes: if the fire source identification result includes the fusion target box, then based on the fusion target box, determining a heat map corresponding to the fusion target box in the monitoring image; in the heat map, filtering out areas with heat values ​​greater than or equal to a heat value threshold to obtain a second image area; based on the second image area and the spatial mapping relationship between the monitoring image and the internal space of the electrical equipment, determining the location of the fire source in the internal space of the electrical equipment.

6. The image recognition-based fire extinguishing method for electrical equipment as described in any one of claims 1 to 5, characterized in that, The step of determining the target fire extinguishing pad from among several fire extinguishing pads based on the location of the fire source and the extinguishing range of the fire extinguishing pad includes: calculating the distance between the location of the fire source and each of the fire extinguishing pads; determining the fire extinguishing pads whose distance is less than the extinguishing range as candidate fire extinguishing pads; and determining the target fire extinguishing pad by selecting the one with the largest coverage area and the fewest fire extinguishing pads among the candidate fire extinguishing pads.

7. A fire extinguishing device for electrical equipment based on image recognition, characterized in that, The image recognition-based fire extinguishing device for electrical equipment includes: an acquisition module for acquiring monitoring images of the interior of the electrical equipment; a fire source identification module for identifying fire sources in the monitoring images to obtain fire source identification results. Specifically, the monitoring images are processed by a trained fire source identification model to obtain target identification results, which include target detection boxes and their corresponding confidence scores. The target identification results include at least one of electric spark identification results, flame identification results, and smoke identification results. Based on the target identification results, a fire source identification result is determined. The trained fire source identification model is obtained by training a sample image training dataset. The sample image training dataset includes image data of fire sources and corresponding fire source label data. The fire source label data includes electric spark labels, flame labels, or smoke labels. Sample images are input into the fire source identification model to be trained, and the model outputs sample images. The fire source identification result corresponding to this image is calculated using a loss function to determine the error loss between the fire source identification result corresponding to the sample image and the fire source label data corresponding to the sample image. Minimizing this error loss is the optimization objective. The model parameters of the identification model to be trained are adjusted iteratively until the number of iterations or the error loss converges, at which point training stops, resulting in a trained fire source identification model. The fire source identification results output by the identification model to be trained include spark identification results, flame identification results, and smoke identification results. The spark identification result is a spark target box, the flame identification result is a flame target box, and the smoke identification result is a smoke target box. A first determining module is used to determine the fire source location within the internal space of the electrical equipment based on the fire source identification results. A second determining module is used to determine a target fire extinguishing disc among several fire extinguishing discs based on the fire source location and the extinguishing range of the fire extinguishing disc. A control module is used to control the target fire extinguishing disc to trigger, thereby extinguishing the fire at the fire source location.

8. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the image recognition-based electrical equipment fire extinguishing method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the image recognition-based electrical equipment fire extinguishing method as described in any one of claims 1 to 6.

Images