Fire detection method, device and system

Abstract

The present invention discloses a fire detection method, device, and system. The fire detection method comprises: obtaining an infrared video image to be identified; inputting the infrared video image to be identified into a fire detection model pre-trained using a transfer learning method, and outputting a recognition result; determining whether the recognition result triggers an alarm mechanism; if the alarm mechanism is triggered, establishing a fire alarm state and issuing a fire alarm signal to provide a fire alarm status check reminder; if the alarm mechanism is not triggered, performing detection of the infrared video image to be identified next week. The fire detection method of the embodiments of the present invention can avoid false alarms caused by numerous interference sources under visible light in complex scenes, and is unaffected by changes in lighting and background environment, allowing normal detection even at night.

Description

Technical Field

[0001] The present invention relates to the technical field of fire detection, and in particular to a fire detection method, a fire detection device and a fire detection system. Background Art

[0002] Fire detection equipment is a device used in the automatic fire alarm system to detect fires and provide timely alarms at the early stages of a fire to reduce personal and property losses.

[0003] Currently, fire detection equipment in related technologies primarily uses two methods to detect flames: 1. Sensor-based detection methods, primarily including photoelectric detectors, particle detectors, and thermal detectors. Photoelectric and particle detectors primarily detect smoke particles produced during a fire; thermal detectors primarily use temperature sensors to detect changes in temperature to determine fire status. However, these three types of detectors have simple fire detection principles and are prone to false alarms and missed alarms. Furthermore, there is a certain delay in flame detection; the detection device must wait until the concentration of smoke particles or the temperature change caused by the fire reaches a predetermined level before it can respond. 2. Computer vision-based detection methods primarily utilize machine learning to analyze the color, texture, shape, and flame pulse frequency of suspected flame areas for comprehensive judgment to determine whether a fire has occurred. However, when used in complex outdoor environments, excessive interference sources (such as the shimmering surface of a lake or reflections from metal and glass) can lead to false alarms. Summary of the Invention

[0004] The present invention aims to solve one of the technical problems in the related art at least to a certain extent. To this end, the present invention aims to provide a fire detection method, device and system to achieve accurate and timely detection of fire.

[0005] To achieve the above-mentioned objectives, an embodiment of the first aspect of the present invention proposes a fire detection method, which includes: obtaining an infrared video image to be identified; inputting the infrared video image to be identified into a fire detection model that has been pre-trained using a transfer learning method, and outputting a recognition result; judging whether the recognition result triggers an alarm mechanism; if the alarm mechanism is triggered, establishing a fire alarm state, and issuing a fire alarm signal to provide a fire alarm state check reminder; if the alarm mechanism is not triggered, performing detection of the infrared video image expected to be identified next week.

[0006] In addition, the fire detection method according to the embodiment of the present invention may also have the following additional technical features:

[0007] According to one embodiment of the present invention, the training process of the fire detection model includes: obtaining an open source data set and a self-built data set, wherein the open source data set includes multiple visible light flame images, and the self-built data set includes multiple pairs of visible light flame images and infrared flame images; training a first detection model through the visible light flame images in the open source data set and the visible light flame images in the self-built data set, wherein the first detection model adopts a YOLO network; constructing a second detection model, wherein the second detection model includes an input layer network and the YOLO network connected in sequence, and the input layer network is used to process a single-channel infrared flame image; using a transfer learning method to copy the parameters in the trained first detection model to the YOLO network in the second detection model; using the infrared flame images in the self-built data set to train the second detection model after the parameters are copied to obtain the trained fire detection model.

[0008] According to one embodiment of the present invention, the images in the self-built dataset are obtained under at least one of the following conditions: multiple scale fire pots, multiple ignition scenes, and multiple infrared cameras, wherein the physical parameters of each infrared camera are different.

[0009] According to one embodiment of the present invention, the method is used for fire detection equipment, and the method further includes: when the fire detection equipment is turned on, the equipment is initialized to determine whether the fire detection equipment has a fault, and a fault alarm signal is issued when the fire detection equipment has a fault; wherein, when the fire detection equipment has not a fault, the step of obtaining the infrared video image to be identified is performed.

[0010] According to one embodiment of the present invention, the input layer network includes: multiple groups of slice layers, tensor splicing layers and convolutional layers, wherein the input end of each group of slice layers is used to input infrared images, the output end of each group of slice layers is connected to the input end of the tensor splicing layer, the output end of the tensor splicing layer is connected to the input end of the convolutional layer, and the output end of the convolutional layer is connected to the input end of the YOLO network in the second detection model.

[0011] According to an embodiment of the present invention, each group of slice layers includes a first slice layer and a second slice layer connected in sequence.

[0012] According to one embodiment of the present invention, when training the second detection model after parameter replication, the parameters of the input layer network are assigned values ​​by random initialization.

[0013] To achieve the above objectives, a second embodiment of the present invention provides a fire detection device comprising: an acquisition module, a recognition module, and a judgment module. The acquisition module is configured to acquire an infrared video image to be recognized; the recognition module is configured to input the infrared video image to be recognized into a fire detection model pre-trained using a transfer learning method, and output a recognition result; the judgment module is configured to determine whether the recognition result triggers an alarm mechanism; if the alarm mechanism is triggered, a fire alarm state is established and a fire alarm signal is issued to provide a fire alarm status check reminder; if the alarm mechanism is not triggered, the device proceeds to detect the infrared video image to be recognized next week.

[0014] In addition, the fire detection device according to the embodiment of the present invention may also have the following additional technical features:

[0015] According to one embodiment of the present invention, the device is used in a fire detection device, and further includes an initialization module. The initialization module is configured to initialize the device when the fire detection device is powered on, to determine whether the fire detection device has a fault, and to issue a fault alarm signal if the fire detection device has a fault.

[0016] To achieve the above objectives, a third embodiment of the present invention provides a fire detection system, comprising an infrared camera and a main control board. The infrared camera is configured to capture infrared video images; the main control board is electrically connected to the infrared camera and includes a memory, a processor, and a computer program stored in the memory. When the computer program is executed by the processor, it implements the fire detection method described above.

[0017] The fire detection method, device, and system of the present invention address the issues of limited infrared sample size, which hinder convergence and generalization performance in fire detection model training. The trained fire detection model exhibits strong anti-interference capabilities and excellent robustness. The fire detection method, device, and system of the present invention also avoid false alarms caused by numerous interference sources under visible light in complex scenes. Furthermore, the system is unaffected by changes in lighting and background environments, enabling normal detection even at night. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 1 is a flow chart of a fire detection method according to an embodiment of the present invention;

[0019] Figure 2 1 is a flow chart of a training process of a fire detection model according to an embodiment of the present invention;

[0020] Figure 3 1 is a schematic diagram of the structure of the input layer network according to an embodiment of the present invention;

[0021] Figure 41 is a schematic diagram of mAP indicator curves of a detection model trained using a transfer learning method and a detection model not trained using a transfer learning method according to an embodiment of the present invention;

[0022] Figure 5 1 is a structural diagram of a fire detection device according to an embodiment of the present invention;

[0023] Figure 6 is a schematic diagram of an alarm output and status update process according to an embodiment of the present invention;

[0024] Figure 7 1 is a schematic structural diagram of a fire detection system according to an embodiment of the present invention;

[0025] Figure 8 This is a schematic diagram comparing images captured by an infrared camera and a color camera according to an embodiment of the present invention;

[0026] Figure 9 FIG. 1 is a schematic diagram of the flame detection effect according to an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The following describes embodiments of the present invention in detail, examples of which 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 invention, and are not to be construed as limiting the present invention.

[0028] Please refer to the following Figure 1 -Attached Figure 9 The fire detection method, device and system according to the embodiments of the present invention are described.

[0029] Figure 1 FIG. 4 is a flow chart of a fire detection method according to an embodiment of the present invention.

[0030] like Figure 1 As shown, the fire detection method includes:

[0031] S1, obtaining the infrared video image to be identified.

[0032] Specifically, the infrared video can be collected by the infrared video collection module, and the infrared video image is taken from images corresponding to several frames in the infrared video.

[0033] S2, inputs the infrared video image to be identified into the fire detection model that has been pre-trained using the transfer learning method, and outputs the recognition result.

[0034] S3, determining whether the recognition result triggers an alarm mechanism.

[0035] S4, if the alarm mechanism is triggered, a fire alarm state is established and a fire alarm signal is issued to provide a fire alarm state check reminder.

[0036] S5, if the alarm mechanism is not triggered, then the detection of the infrared video image expected to be recognized in the next week is carried out.

[0037] The fire detection method of the embodiment of the present invention can avoid false alarms caused by a large number of interference sources under visible light in complex scenes, and is not affected by changes in lighting and background environment, and can also perform normal detection at night.

[0038] In some embodiments of the present invention, Figure 2 As shown in Figure 2, the training process of the fire detection model includes:

[0039] S21, obtaining an open source dataset and a self-built dataset, wherein the open source dataset includes a plurality of visible light flame images, and the self-built dataset includes a plurality of pairs of visible light flame images and infrared flame images.

[0040] Specifically, the images in the self-built dataset are obtained under at least one of the following conditions: multiple scales of fire pots, multiple ignition scenes, and multiple infrared cameras, wherein the physical parameters of each infrared camera are different.

[0041] More specifically, the visible light flame images in the open source dataset can be ordinary flame images under visible light that are publicly available on the Internet. The self-built dataset can be collected through field ignition tests to collect samples with one-to-one correspondence between color and infrared.

[0042] S22: Training a first detection model using visible light flame images in an open source dataset and visible light flame images in a self-built dataset, wherein the first detection model adopts a YOLO network.

[0043] Specifically, the first detection model can be defined as a convolutional neural network M1, which can use the YOLO network. The convolutional neural network M1 can be fine-tuned using a homologous-different-target learning approach. That is, based on the existing YOLO model, the model training is completed using visible light flame images from the above-mentioned open source dataset and the self-built dataset. After training and fine-tuning the internal parameters, the model automatically learns flame-related features in visible light images and has good detection performance for flame targets. Among them, the homologous-different-target learning approach is an inductive isomorphic transfer learning method. Homologous means that the source and target domain feature spaces of the YOLO official model and the convolutional neural network M1 are the same, and the input ends are both RGB three-channel color images of the same size; different targets means that the YOLO official model is trained based on the COCO dataset, and its detection targets are the 80 categories of targets in the COCO dataset, while the convolutional neural network M1 detects flame targets.

[0044] It should be noted that before training the first detection model, the parameters in the first detection model can be fine-tuned so that the first detection model can automatically learn the relevant characteristics of the flame during the training process, so that the trained first detection model can have better detection performance for flame targets.

[0045] S23, constructing a second detection model, wherein the second detection model includes an input layer network and a YOLO network connected in sequence, and the input layer network is used to process a single-channel infrared flame image.

[0046] As an example, the model data loader, image data preprocessing, and data enhancement parts in the YOLO network can be reconfigured so that the YOLO network can support the processing of single-channel infrared flame images.

[0047] It should be noted that the first detection model and the second detection model have different structures in the input layer network, and the backbone network and the neck network have the same structure.

[0048] Specifically, if Figure 3 As shown, the input layer network includes multiple slice layers, tensor splicing layers, and convolutional layers. The input of each slice layer is used to input an infrared image, and the output of each slice layer is connected to the input of the tensor splicing layer. The output of the tensor splicing layer is connected to the input of the convolutional layer, and the output of the convolutional layer is connected to the input of the YOLO network in the second detection model.

[0049] More specifically, each group of slice layers includes a first slice layer and a second slice layer that are sequentially connected.

[0050] As an example, the parameter axis in the tensor concatenation layer is set to 1. The relevant parameters in the convolution layer are set to: dilation=1,1; group=1; kernel_shape=3,3; pads=1,1,1,1; strides=1,1.

[0051] S24, using a transfer learning method to copy the parameters in the trained first detection model to the YOLO network in the second detection model.

[0052] S25, using the infrared flame image in the self-built data set to train the second detection model after the parameter copying, to obtain a trained fire detection model.

[0053] Specifically, when training the second detection model after parameter replication, the parameters of the input layer network are assigned values ​​using a random initialization method.

[0054] Optionally, before training the second detection model after parameter replication, the parameters in the second detection model after parameter replication may be fine-tuned to enable the trained second detection model to better detect flame targets.

[0055] As an example, after randomly initialized parameters are input into the input end of the second detection model after parameter replication, all network model structure levels are activated, and training is performed using infrared flame images in a self-built dataset. The model parameter update is completed through the BP algorithm and gradient descent related optimizer.

[0056] In this embodiment, the second detection model draws on the feature extractor of the first detection model through a heterogeneous-source, same-target learning approach. Therefore, even when infrared flame videos and images are insufficient, a fire detection model with good convergence and generalization can be trained, achieving ideal results and improving fire detection accuracy. The heterogeneous-source, same-target learning approach is a form of direct heterogeneous transfer learning. Heterogeneous sources refer to different feature spaces between the source and target domains, with the first and second detection models taking as input an RGB three-channel color image and a single-channel grayscale image, respectively. Same-target refers to the fact that the first and second detection models have the same detection target: both are used to detect flame targets.

[0057] As an example, the training method of the fire detection model of the present invention is described by comparing the mAP index curves of the detection model trained using the transfer learning method and the detection model not trained using the transfer learning method. Figure 4 As shown in the figure, using transfer learning accelerates model training convergence, allowing model parameters to quickly reach the global optimal solution. The mAP value after training using transfer learning reaches 0.92, an improvement of 0.13 compared to the mAP value of 0.79 without transfer learning. Therefore, the model trained using transfer learning demonstrates excellent flame detection performance in infrared images.

[0058] In some embodiments of the present invention, the fire detection method is used for a fire detection device, and the fire detection method further includes: when the fire detection device is turned on, the device is initialized to determine whether the fire detection device has a fault, and a fault alarm signal is issued when the fire detection device has a fault; wherein, when the fire detection device has not a fault, the step of obtaining an infrared video image to be identified is executed.

[0059] Specifically, device initialization includes: reading device parameters, loading the fire detection model, and performing a self-test of device functions. Device initialization can be automatically initiated when the device is powered on.

[0060] Corresponding to the above embodiment, the present invention further provides a fire detection device.

[0061] Figure 5It is a structural schematic diagram of a fire detection device according to an embodiment of the present invention.

[0062] like Figure 5 As shown, the fire detection device includes: an acquisition module 100 , an identification module 200 and a judgment module 300 .

[0063] The acquisition module 100 is used to acquire the infrared video image to be identified.

[0064] The recognition module 200 is used to input the infrared video image to be recognized into the fire detection model that has been pre-trained using the transfer learning method, and output the recognition result.

[0065] The judgment module 300 is used to judge whether the recognition result triggers the alarm mechanism; if the alarm mechanism is triggered, a fire alarm state is established and a fire alarm signal is issued to provide a fire alarm state check reminder; if the alarm mechanism is not triggered, the detection of the infrared video image expected to be recognized in the next week is carried out.

[0066] As an example, Figure 6 As shown, the judgment module 300 can push an alarm signal to the host computer or related control system, which can activate an audible and visual alarm to remind relevant personnel to check the alarm information. Relevant personnel can modify parameters on the host computer, and complete the update of internal parameters and status changes through the communication module, parameter management module, algorithm module and video module.

[0067] It should be noted that, for other specific implementations of the fire detection device according to the embodiment of the present invention, reference may be made to other specific implementations of the fire detection method according to the above embodiment of the present invention.

[0068] The fire detection device of the embodiment of the present invention can avoid false alarms caused by a large number of interference sources under visible light in complex scenes, and is not affected by changes in lighting and background environment, and can also perform normal detection at night.

[0069] Corresponding to the above embodiment, the present invention also provides a fire detection system.

[0070] Figure 7 It is a structural diagram of a fire detection system according to an embodiment of the present invention.

[0071] like Figure 7 As shown, the fire detection system includes: an infrared camera 10 and a main control board 20.

[0072] The infrared camera 10 is used to collect infrared video images.

[0073] The main control board 20 is electrically connected to the infrared camera 10. The main control board 20 includes a memory 201, a processor 202 and a computer program stored in the memory 201. When the computer program is executed by the processor 202, the above-mentioned fire detection method is implemented.

[0074] For example, if Figure 8 As shown, the use of an infrared camera 10 can reduce the problem of false alarms caused by excessive interference sources in ordinary visible light images. This is particularly true in complex outdoor scenes. The infrared image only retains information about actual flames and objects that scatter infrared light, thus filtering flame detection. Furthermore, infrared images are single-channel pixel matrices. For infrared and visible light images of the same size, infrared image preprocessing is simpler, consumes less memory, and achieves faster detection speeds.

[0075] In this embodiment, the processor 202 and the memory 201 are connected, for example, via a bus 203. Optionally, the main control board 20 may further include a transceiver 204. It should be noted that in actual applications, the number of transceivers 204 is not limited to one, and the structure of the main control board 20 does not constitute a limitation on the embodiment of the present invention.

[0076] The processor 202 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the present disclosure. The processor 202 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like.

[0077] The bus 203 may include a path for transmitting information between the above components. The bus 203 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 203 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 7 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0078] The memory 201 is used to store a computer program corresponding to the fire detection method of the above embodiment of the present invention, and the computer program is controlled and executed by the processor 202. The processor 202 is used to execute the computer program stored in the memory 201 to implement the content shown in the above method embodiment.

[0079] In this embodiment, Figure 9 As shown, by using the fire detection system of the present invention, interference sources in complex scenarios can be avoided and the location of the fire source can be accurately determined.

[0080] The fire detection system of the embodiment of the present invention can avoid false alarms caused by a large number of interference sources under visible light in complex scenes, and is not affected by changes in lighting and background environment, and can also perform normal detection at night.

[0081] As an example, the main control board 20 can be a Rockchip RV1126 core board. By using the RKNN internal model conversion, inference and other related script files in the Core Micro Software Development Kit, the optimization and deployment of the second detection model are completed. Among them, the Rockchip RV1126 core board contains two NPU units for accelerating inference of deep learning models.

[0082] It should be noted that the logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device), or in conjunction with such instruction execution system, apparatus, or device. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transmit a program for use by an instruction execution system, apparatus, or device, or in conjunction with such instruction execution system, apparatus, or device. More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection portion having one or more wires (electronic device), a portable computer disk cartridge (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and portable compact disc read-only memory (CDROM). Furthermore, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing it in another suitable manner if necessary, and then storing it in a computer memory.

[0083] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0084] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0085] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0086] Although the embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A fire detection method, characterized in that: The method comprises: Acquire the infrared video image to be identified; Inputting the infrared video image to be identified into a fire detection model that has been pre-trained using a transfer learning method, and outputting a recognition result; Determining whether the recognition result triggers an alarm mechanism; If the alarm mechanism is triggered, a fire alarm state is established and a fire alarm signal is issued to provide a fire alarm state check reminder; If the alarm mechanism is not triggered, the detection of the infrared video image is expected to be carried out in the next week; The training process of the fire detection model includes: Acquire an open source data set and a self-built data set, wherein the open source data set includes a plurality of visible light flame images, and the self-built data set includes a plurality of pairs of visible light flame images and infrared flame images; Training a first detection model using visible light flame images in the open source dataset and visible light flame images in the self-built dataset, wherein the first detection model adopts a YOLO network; Constructing a second detection model, wherein the second detection model includes an input layer network and the YOLO network connected in sequence, and the input layer network is used to process a single-channel infrared flame image; Use a transfer learning method to copy the parameters of the trained first detection model to the YOLO network in the second detection model; The second detection model after parameter replication is trained using the infrared flame image in the self-built data set to obtain the trained fire detection model.

2. The fire detection method according to claim 1, characterized in that: The images in the self-built dataset are obtained under at least one of the following conditions: Multiple scale braziers, multiple ignition scenes, and multiple infrared cameras, wherein the physical parameters of each infrared camera are different.

3. The fire detection method according to claim 1, characterized in that: The method is used for a fire detection device, and the method further comprises: When the fire detection device is turned on, the device is initialized to determine whether the fire detection device has a fault, and a fault alarm signal is issued when the fire detection device has a fault; Wherein, when the fire detection equipment is not faulty, the step of acquiring the infrared video image to be identified is performed.

4. The fire detection method according to claim 1, characterized in that: The input layer network includes: multiple groups of slice layers, tensor splicing layers and convolution layers, wherein the input end of each group of slice layers is used to input infrared images, the output end of each group of slice layers is connected to the input end of the tensor splicing layer, the output end of the tensor splicing layer is connected to the input end of the convolution layer, and the output end of the convolution layer is connected to the input end of the YOLO network in the second detection model.

5. The fire detection method according to claim 4, characterized in that: Each group of slice layers includes a first slice layer and a second slice layer connected in sequence.

6. The fire detection method according to claim 5, characterized in that: When training the second detection model after parameter replication, the parameters of the input layer network are assigned values ​​by random initialization.

7. A fire detection device, characterized in that: For implementing the fire detection method according to any one of claims 1 to 6, the device comprises: An acquisition module, used for acquiring infrared video images to be identified; A recognition module, configured to input the infrared video image to be recognized into a fire detection model that has been pre-trained using a transfer learning method, and output a recognition result; The judgment module is used to judge whether the recognition result triggers the alarm mechanism; if the alarm mechanism is triggered, a fire alarm state is established and a fire alarm signal is issued to provide a fire alarm status check reminder; if the alarm mechanism is not triggered, the detection of the infrared video image expected to be recognized in the next week is carried out.

8. The fire detection device according to claim 7, characterized in that: The device is used for fire detection equipment, and the device also includes: The initialization module is used to initialize the fire detection device when the fire detection device is turned on to determine whether the fire detection device has a fault and to send a fault alarm signal when the fire detection device has a fault.

9. A fire detection system, characterized in that: The system comprises: Infrared camera, used to collect infrared video images; A main control board is electrically connected to the infrared camera, and the main control board includes a memory, a processor, and a computer program stored in the memory. When the computer program is executed by the processor, the fire detection method according to any one of claims 1 to 6 is implemented.

Images