Fire target detection system and method

Abstract

The invention provides a fire target detection system and method, and the system comprises a plurality of fire target detection groups and a unified management platform, and each fire target detection group comprises a data collection device and a brain-like calculation device. The data acquisition device acquires environment temperature data and color video data in a fire detection area, and transmits the acquired environment temperature data and color video data to the brain-like computing device; the brain-like computing device senses the environment temperature data and the color video data based on an environment sensing technology, and determines whether a fire occurs in the fire detection area; and when a fire occurs in the fire detection area, detecting the color video data based on a fire target detection technology, and determining fire condition information and a fire position, and the brain-like computing device is also used for transmitting the fire condition information and the fire position to the unified management platform. According to the scheme of the invention, the accuracy and real-time performance of fire detection are improved.

Description

Technical Field

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

[0002] Against the backdrop of continuous technological advancement and stagnant urbanization, the demand for fire target detection continues to grow. Traditional fire detection systems have many defects in data transmission and information processing due to insufficient computing power and model performance, making it difficult to meet the needs of rapid response, posing a threat to people's lives, health and property safety.

[0003] Traditional fire target detection technology has some shortcomings, mainly including the following aspects: On the one hand, in the early stage of a fire, the smoke state changes a lot and the flame size may be very small. Traditional target detection algorithms may have slow detection speeds in the face of such complex situations. On the other hand, due to the complex background of smoke and fire sources, detection problems of targets of different scales, etc., the accuracy of traditional algorithms in fire target detection may be low. On the other hand, traditional fire detection algorithms collect relevant data through temperature and smoke sensors to predict fire probability, but the real-time detection of this method cannot be guaranteed.

[0004] Therefore, the accuracy and real-time performance of fire detection in the prior art are relatively poor. Summary of the invention

[0005] The present invention provides a fire target detection system and method, which are used to solve the defects of poor accuracy and real-time performance of fire detection in the prior art, realize accurate and real-time fire target detection, and improve the accuracy and real-time performance of fire target detection.

[0006] The present invention provides a fire target detection system, comprising: a plurality of fire target detection groups and a unified management platform, each of the fire target detection groups being arranged in a fire detection area, each of the fire target detection groups comprising a data acquisition device and a brain-like computing device; The data acquisition device in each of the fire target detection groups is connected to the brain-like computing device, and the data acquisition device is used to collect ambient temperature data and color video data in the fire detection area, and transmit the collected ambient temperature data and color video data to the brain-like computing device; The brain-like computing device is used to sense the ambient temperature data and the color video data based on the environmental sensing technology to determine whether a fire occurs in the fire detection area; and when a fire occurs in the fire detection area, the color video data is detected based on the fire target detection technology to determine the fire situation information and the fire location; Each of the brain-like computing devices is connected to the unified management platform, and the brain-like computing device is also used to transmit the fire situation information and the fire location to the unified management platform.

[0007] According to a fire target detection system provided by the present invention, the brain-like computing device comprises: an environment perception module and a fire target detection module; The environment sensing module is connected to the data acquisition device, and is used to perform temperature mutation analysis based on the ambient temperature data, and flame perception, flame segmentation detection and smoke perception based on the color video data to determine whether a fire occurs in the fire detection area, and output alarm information when a fire occurs in the fire detection area; The fire target detection module is connected to the environment perception module and the data acquisition device. The fire target detection module is used to detect the color video data in response to the alarm information to determine the fire situation information and the fire location.

[0008] According to a fire target detection system provided by the present invention, the brain-like computing device further includes: an analysis module, an image integration module, a temperature integration module and a fire spread prediction module; The analysis module is connected to the data acquisition device, and is used to analyze the color video data to obtain images of different positions in the fire detection area, and to analyze the ambient temperature data to obtain temperatures of different positions in the fire detection area; The image integration module is connected to the analysis module, and the image integration module is used to stitch images at different positions within the fire detection area to obtain a panoramic image within the fire detection area; The temperature integration module is connected to the analysis module, and is used to combine the temperatures at different locations within the fire detection area to obtain the temperature distribution within the fire detection area; The fire spread prediction module is connected to the image integration module and the temperature integration module. The fire spread prediction module is used to perform flame segmentation detection and temperature mutation analysis based on the panoramic image within the fire detection area and the temperature distribution within the fire detection area to obtain the predicted direction and predicted position of fire spread.

[0009] According to a fire target detection system provided by the present invention, the color video data includes a plurality of sub-video data and a plurality of sub-temperature data; the data acquisition device includes: a plurality of dual-light pan-tilt cameras; Each dual-light PTZ camera is used to shoot different positions in the fire detection area to obtain the multiple sub-video data and the multiple sub-temperature data; or, The data acquisition device comprises: a dual-light PTZ camera; The dual-light pan-tilt camera is used to adjust the shooting angle and focal length according to preset shooting requirements, shoot different positions in the fire detection area, and obtain the color video data and the ambient temperature data.

[0010] According to a fire target detection system provided by the present invention, the brain-like computing device further includes: a preprocessing module; The preprocessing module is connected to the data acquisition device, the environmental perception module and the fire target detection module. The preprocessing module is used to preprocess the ambient temperature data and the color video data to obtain the preprocessed ambient temperature data and color video data, and transmit the preprocessed ambient temperature data and color video data to the environmental perception module and the fire target detection module.

[0011] According to a fire target detection system provided by the present invention, the brain-like computing device further includes: a dual-light positioning module; The dual-light calibration module is connected to the data acquisition device, the environmental perception module and the fire target detection module. The dual-light calibration module is used to perform dual-light calibration on the ambient temperature data and the color video data to obtain calibrated ambient temperature data and color video data, and transmit the calibrated ambient temperature data and color video data to the environmental perception module and the fire target detection module.

[0012] According to a fire target detection system provided by the present invention, the brain-like computing device further includes: an interference suppression module; The interference suppression module is connected to the data acquisition device, the environmental perception module and the fire target detection module. The interference suppression module is used to identify and filter interference data in the ambient temperature data and the color video data, and transmit the filtered ambient temperature data and color video data to the environmental perception module and the fire target detection module.

[0013] According to a fire target detection system provided by the present invention, the brain-like computing device further includes: a display module; The display module is connected to the environmental perception module, the fire target detection module and the fire spread prediction module, and the display module is used to obtain the alarm information, the fire situation information, the fire location, the predicted direction of the fire spread and the predicted position of the fire spread, and to visualize the alarm information, the fire situation information, the fire location, the predicted direction of the fire spread and the predicted position of the fire spread.

[0014] According to a fire target detection system provided by the present invention, the unified management platform includes: an integrated management server.

[0015] The present invention also provides a fire target detection method, which is applied to a fire target detection system, wherein the fire target detection system comprises: a plurality of fire target detection groups and a unified management platform, each of the fire target detection groups is arranged in a fire detection area, and each of the fire target detection groups comprises a data acquisition device and a brain-like computing device; the method comprises: The data acquisition device in each of the fire target detection groups collects ambient temperature data and color video data within the fire detection area, and transmits the collected ambient temperature data and color video data to the brain-like computing device; The brain-like computing device senses the ambient temperature data and the color video data based on the environmental sensing technology to determine whether a fire occurs in the fire detection area; and when a fire occurs in the fire detection area, the brain-like computing device detects the color video data based on the fire target detection technology to determine the fire situation information and the fire location; The brain-like computing device transmits the fire situation information and the fire location to the unified management platform.

[0016] The fire target detection system and method provided by the present invention, by setting a data acquisition device and a brain-like computing device in each fire detection area, the data acquisition device can timely collect the ambient temperature data and color video data in the fire detection area, and the data acquisition device and the brain-like computing device are deployed in each fire detection area, which can reduce the data transmission time, and the brain-like computing device can process the ambient temperature data and color video data collected by the data acquisition device in real time, thereby realizing a rapid response to fire target detection. Furthermore, the brain-like computing device performs fire target detection based on the ambient temperature data and color video data, can determine the fire situation information and the fire location, and can realize accurate detection of the fire; further, the brain-like computing device transmits the fire situation information and the fire location to a unified management platform to realize centralized monitoring and management of the fire. Therefore, the scheme of the present application improves the accuracy and real-time performance of fire target detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 This is one of the structural schematic diagrams of the fire target detection system provided by the present invention.

[0019] Figure 2 This is the second structural schematic diagram of the fire target detection system provided by the present invention.

[0020] Figure 3 It is a flow chart of the fire target detection method provided by the present invention.

[0021] Reference numerals: 10: Fire target detection group; 110: Data acquisition device; 120: Brain-like computing device; 1201: Environmental perception module; 1202: Fire target detection module; 1203: Analysis module; 1204: Image integration module; 1205: Temperature integration module; 1206: Fire spread prediction module; 1207: Preprocessing module; 1208: Dual-light positioning module; 1209: Interference suppression module; 1210: Display module; 30: Unified management platform. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] It should be noted that the brief description of terms in this application is only for the convenience of understanding the embodiments described below, and is not intended to limit the embodiments of this application. Unless otherwise specified, these terms should be understood according to their common and usual meanings.

[0024] The term "brain-inspired computing" used in the embodiments of this application is a general term for computing theories, architectures, chip designs, and application models and algorithms that draw on the information processing patterns and structures of biological nervous systems. Brain-inspired computing is an emerging computing paradigm that is inspired by the neural network structure and information processing mechanisms of the human brain and aims to develop new computing models, algorithms, and hardware systems. The core idea of ​​brain-inspired computing is to imitate the working method of the human brain to achieve more efficient and intelligent information processing.

[0025] Against the backdrop of continuous technological advancement and stagnant urbanization, the demand for fire target detection continues to grow. Traditional fire detection systems have many defects in data transmission and information processing due to insufficient computing power and model performance, making it difficult to meet the needs of rapid response, posing a threat to people's lives, health and property safety.

[0026] Traditional fire target detection technology has some shortcomings, mainly including the following aspects: On the one hand, in the early stage of a fire, the smoke state changes a lot and the flame size may be very small. Traditional target detection algorithms may have slow detection speeds in the face of such complex situations. On the other hand, due to the complex background of smoke and fire sources, detection problems of targets of different scales, etc., the accuracy of traditional algorithms in fire target detection may be low. On the other hand, traditional fire detection algorithms collect relevant data through temperature and smoke sensors to predict fire probability, but the real-time detection of this method cannot be guaranteed.

[0027] Therefore, how to improve the accuracy and real-time performance of fire detection is a technical problem that needs to be solved urgently.

[0028] The technical solution of the present invention is used to solve the above technical problems. The technical solution of the present application and how the technical solution of the present application solves the above technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. Figure 1 and Figure 2 The fire target detection system of the present invention is described.

[0029] Figure 1 FIG. 1 is one of the structural diagrams of the fire target detection system provided by the present invention. Figure 1 As shown, the fire target detection system includes: multiple fire target detection groups 10 and a unified management platform 30, each fire target detection group 10 is set in a fire detection area, and each fire target detection group 10 includes a data acquisition device 110 and a brain-like computing device 120.

[0030] like Figure 1As shown, the data acquisition device 110 in each fire target detection group 10 is connected to the brain-like computing device 120. The data acquisition device 110 is used to collect ambient temperature data and color video data in the fire detection area, and transmit the collected ambient temperature data and color video data to the brain-like computing device 120.

[0031] The brain-like computing device 120 is used to sense the ambient temperature data and color video data based on the environmental sensing technology to determine whether a fire occurs in the fire detection area; and when a fire occurs in the fire detection area, detect the color video data based on the fire target detection technology to determine the fire situation information and the fire location.

[0032] Each brain-like computing device 120 is connected to the unified management platform 30 , and the brain-like computing device 120 is also used to transmit fire situation information and fire location to the unified management platform 30 .

[0033] In this embodiment, the fire target detection system can be applied to a wide range of fire detection scenarios, and the fire detection scenarios are divided into multiple fire detection areas. A data acquisition device 110 and a brain-like computing device 120 are respectively set in each fire detection area. It can be understood that by setting the data acquisition device 110 and the brain-like computing device 120 in the fire detection area, real-time fire detection can be achieved in each fire detection area, thereby improving the real-time and reliability of fire target detection.

[0034] The data acquisition device 110 refers to a device capable of acquiring ambient temperature data and color video data. In this embodiment, the data acquisition device 110 is not specifically limited. For example, the data acquisition device 110 may be a dual-light pan-tilt camera, a visible light camera, a multi-spectral camera, or the like.

[0035] As an example, the data acquisition device 110 is a dual-light pan-tilt camera, which is a camera device that integrates visible light and infrared light imaging technologies. It can simultaneously collect ambient temperature data and color (RGB) video data. Specifically, the visible light camera is responsible for shooting color video data, that is, RGB video data. The infrared camera is responsible for capturing thermal radiation in the scene, that is, infrared radiation. This radiation is directly related to the temperature of the object, so it can be used to measure and monitor the ambient temperature. The dual-light pan-tilt camera usually has dual-light fusion technology, which can superimpose visible light images and infrared images on the same screen, or switch the display as needed. This technology allows the operator to observe the visible light image and temperature distribution of the scene at the same time.

[0036] In practical applications, the data acquisition device 110 and the brain-like computing device 120 can be electrically connected or communicatively connected, and the data acquisition device 110 and the brain-like computing device 120 are arranged in the fire detection area. For example, the data acquisition device 110 can transmit the color video data and the ambient temperature data to the brain-like computing device 120 through the local area network. It can be understood that the data acquisition device 110 can transmit the collected ambient temperature data and color video data to the brain-like computing device 120 in real time, and the brain-like computing device 120 can process the ambient temperature data and the color video data in time to realize the detection of the fire. Therefore, the scheme of this embodiment can improve the real-time performance of the fire target detection.

[0037] Exemplarily, the data acquisition device 110 can transmit the color video data and the ambient temperature data to the brain-like computing device 120 in the form of a video stream and a data stream, respectively. The video stream refers to the continuous transmission process of the video data, and the data acquisition device 110 sends the color video data to the brain-like computing device 120 in the form of a continuous stream.

[0038] The brain-like computing device 120 is deployed in the fire detection area to identify the fire target in the fire detection area. Specifically, the brain-like computing device 120 can be a brain-like computer, a brain-like chip, a brain-like server, etc. For example, the brain-like computing device 120 can be a brain-like computer, which is a new type of computer model that simulates the operation of the brain neural network. It builds corresponding computing theories, chip architectures, and application models and algorithms by drawing on the information processing mode and structure of the biological nervous system.

[0039] In actual applications, a fire target detection model is deployed in the brain-like computing device 120. The fire target detection model is an algorithm model built and pre-trained based on brain-like computing and deep learning technology, which is used to identify and detect targets such as flames and smoke in fire images or videos. This model can analyze mutations in color video data and ambient temperature data to determine whether a fire has occurred, and when a fire occurs, determine the location of the fire and fire situation information. For different fire detection areas, deploying a brain-like computer (edge ​​computer) means setting up an edge computing node in each area where fire monitoring is required. Edge computing nodes are close to data sources and can process data quickly, reduce latency, and improve response speed.

[0040] As an example, the fire target detection model can be a model built and trained based on the YOLOv8s flame detection algorithm. Specifically, the brain-like computer is equipped with the KA200 brain-like computing chip. Based on the new storage-computing-in-one, multi-core parallel, heterogeneous fusion architecture, the AI ​​computing power reaches 32TOPS (INT8) and 16TFLOPS (FP16). It mainly receives the video stream and temperature data collected by the dual-light pan-tilt camera, calls LynDriver and LynSDK, and uses the YOLOv8s model for cross-compilation to achieve real-time detection of fire targets.

[0041] In combination with the above description, the fire target detection system can be applied to a wide range of fire detection scenarios, and the fire detection scenarios can be divided into multiple fire detection areas. A data acquisition device 110 and a brain-like computing device 120 are respectively set in each fire detection area.

[0042] The unified management platform 30 is used to uniformly manage multiple target detection areas. The unified management platform 30 relies on advanced information technology, such as the Internet of Things (IoT), big data analysis, artificial intelligence, and cloud computing, to achieve efficient and intelligent management of multiple target detection areas. For example, the unified management platform 30 can be a server, a central computer, a cloud platform (such as a smart fire protection cloud platform), etc.

[0043] In actual applications, when the brain-like computing device 120 determines that a fire has occurred in the fire detection area, it outputs alarm information and determines the fire situation information and the fire location. Furthermore, the brain-like computing device 120 combines the alarm information, the fire situation information, the fire location, the color video data, and the ambient temperature data into a file, names the file according to a pre-set naming rule, and saves the file locally.

[0044] Furthermore, the brain-like computing device 120 transmits the fire situation information and the fire location to the unified management platform 30. Optionally, the brain-like computing device 120 may transmit the alarm information, the fire situation information, the fire location, the color video data and the ambient temperature data, or transmit the locally stored files to the unified management platform 30.

[0045] In actual applications, when the brain-like computing device 120 in the fire target detection system detects a fire in the corresponding fire detection area, it will promptly notify the user through a real-time notification mechanism and synchronize the data to the unified management platform 30 to achieve centralized monitoring and management of the fire.

[0046] As an example, in a possible implementation, the unified management platform 30 includes an integrated management server.

[0047] Specifically, the integrated management server is a centralized computing platform that plays a core role in an organization or system, responsible for processing, analyzing and storing data from different sources and providing decision support. For example, the integrated management server can be a physical server, a cloud server, a virtualized server, etc.

[0048] The scheme of this embodiment, by setting a data acquisition device 110 and a brain-like computing device 120 in each fire detection area, the data acquisition device 110 can timely collect the ambient temperature data and color video data in the fire detection area. The data acquisition device 110 and the brain-like computing device 120 are deployed in each fire detection area, which can reduce the data transmission time. The brain-like computing device 120 can process the ambient temperature data and color video data collected by the data acquisition device 110 in real time, and realize the rapid response of fire target detection. Further, the brain-like computing device 120 performs fire target detection based on the ambient temperature data and color video data, can determine the fire situation information and the fire location, and can realize accurate detection of the fire; further, the brain-like computing device 120 transmits the fire situation information and the fire location to the unified management platform 30 to realize the centralized monitoring and management of the fire. Therefore, the scheme of this embodiment improves the accuracy and real-time performance of fire target detection.

[0049] In addition, in some possible implementations, the above-mentioned brain-inspired computing device 120 includes: an environment perception module 1201 and a fire target detection module 1202 .

[0050] The environmental sensing module 1201 is connected to the data acquisition device 110. The environmental sensing module 1201 is used to perform temperature mutation analysis based on ambient temperature data, and to perform flame perception, flame segmentation detection and smoke perception based on color video data to determine whether a fire occurs in the fire detection area, and output alarm information when a fire occurs in the fire detection area.

[0051] The fire target detection module 1202 is connected to the environment perception module 1201 and the data acquisition device 110. The fire target detection module 1202 is used to detect the color video data in response to the alarm information to determine the fire situation information and the fire location.

[0052] In this embodiment, whether a fire occurs in the fire detection area is comprehensively evaluated through temperature mutation analysis, flame perception, flame segmentation detection and smoke perception, which can effectively improve the accuracy of fire target detection.

[0053] Specifically, temperature mutation analysis is a key technology for detecting early signs of fire. It identifies possible fire events by monitoring sudden changes in temperature in the environment. In one example, a baseline ambient temperature distribution is pre-set. After the environmental perception module 1201 obtains the ambient temperature data collected by the data acquisition device 110, it compares the collected ambient temperature data with the baseline ambient temperature distribution to determine the temperature change value. When the monitored temperature change value exceeds the preset temperature change threshold, the environmental perception module 1201 determines that the temperature has changed suddenly. Further, temperature mutation events are identified, and it is evaluated whether these events may be caused by fire.

[0054] Specifically, flame sensing is a key link, the purpose of which is to determine whether there is a flame in the monitored area. It can help quickly identify fires and respond in a timely manner. Flame sensing usually involves real-time analysis of color video data to identify the characteristics of flames, such as color, brightness, shape, and motion. Flame detection algorithms search for these features to determine whether there is a flame. In practice, flames usually have specific color characteristics, especially in the infrared and near-infrared spectrum. Flames can be identified by analyzing the color information in the image. Flames move with the hot air flow and the combustion process, and the motion characteristics of the flames can be captured by motion detection algorithms. The edges of flames are usually obvious, and the outline of the flame can be identified using edge detection algorithms. The texture of flames is complex and dynamically changing, and texture analysis can be used to distinguish flames from other similar visual phenomena.

[0055] Further, after determining the presence of a flame, flame segmentation detection is performed. Specifically, flame segmentation detection can accurately identify and segment the flame area from the color video data. This involves more detailed image processing techniques such as color threshold-based segmentation, edge detection, region growing or deep learning methods to determine the boundaries and shape of the flame.

[0056] Specifically, smoke perception involves identifying and analyzing the presence and characteristics of smoke. In the early stages of a fire, smoke is often the first visible sign, so effective smoke perception is crucial for early fire warning. In practical applications, smoke feature extraction is performed on color video data, for example, the color, texture, shape, and motion characteristics of the smoke are extracted. Further, based on the characteristics of the smoke, such as color thresholds, texture patterns, and motion changes, candidate smoke regions in the image are identified. Further, smoke perception algorithms, such as threshold-based segmentation, region growing, edge detection, or machine learning methods, are applied to identify and confirm the presence of smoke.

[0057] In this implementation, the results of temperature mutation analysis, flame perception, flame segmentation detection and smoke perception are integrated to determine whether a fire occurs in the fire detection area, thereby improving the accuracy and robustness of fire target detection.

[0058] It can be understood that when the environment perception module 1201 determines that a fire has occurred in the fire detection area, timely output of alarm information can ensure that the fire is responded to immediately after it is discovered in the early stage, which is crucial for controlling the fire and preventing the fire from spreading. The alarm information can quickly notify the on-site personnel and the emergency response team, thereby initiating the evacuation procedure and protecting the safety of personnel.

[0059] In actual applications, after the environment perception module 1201 determines that a fire has occurred in the fire detection area, the fire target detection module 1202 further detects the color video data to determine the fire situation information and the fire location. The fire situation information can be preset when constructing the environment perception module 1201 according to actual needs. For example, the fire situation information includes the size, intensity, color and diffusion of the flame, etc.

[0060] In this embodiment, on the one hand, through the temperature mutation analysis of the environmental perception module 1201 and the flame and smoke perception of the video data, the system can detect the fire in the early stage of the fire and achieve early detection and early warning. The multimodal detection combining temperature data and video data improves the accuracy of fire detection and reduces the possibility of false alarms and missed alarms. On the other hand, the fire target detection module 1202 further analyzes the color video data to accurately determine the fire situation information and the fire location, providing accurate guidance for fire fighting and evacuation. Therefore, the scheme of this embodiment realizes the early detection, accurate positioning and rapid response of fire, improves the real-time and accuracy of fire target detection, improves the efficiency and effectiveness of fire management, reduces the losses caused by fire, and enhances the automation and intelligence level of the system.

[0061] Optionally, in a possible implementation, the brain-like computing device 120 further includes: an analysis module 1203 , an image integration module 1204 , a temperature integration module 1205 and a fire spread prediction module 1206 .

[0062] The analysis module 1203 is connected to the data acquisition device 110 and is used to analyze the color video data to obtain images of different positions in the fire detection area, and to analyze the ambient temperature data to obtain temperatures of different positions in the fire detection area.

[0063] The image integration module 1204 is connected to the analysis module 1203 and is used to stitch images at different positions in the fire detection area to obtain a panoramic image of the fire detection area.

[0064] The temperature integration module 1205 is connected to the analysis module 1203 , and is used to combine the temperatures at different locations within the fire detection area to obtain the temperature distribution within the fire detection area.

[0065] The fire spread prediction module 1206 is connected to the image integration module 1204 and the temperature integration module 1205. The fire spread prediction module 1206 is used to perform flame segmentation detection and temperature mutation analysis based on the panoramic image within the fire detection area and the temperature distribution within the fire detection area, so as to obtain the predicted direction and predicted position of the fire spread.

[0066] In actual applications, due to the limitation of the dual-light PTZ camera's own performance, the images captured by the dual-light PTZ camera can only cover a certain range, and the infrared temperature measurement results can only meet the accuracy requirements within a certain distance range. Therefore, multiple dual-light PTZ cameras are required or the shooting angle and focal length of the dual-light PTZ camera are continuously adjusted to obtain color video data and ambient temperature data.

[0067] Specifically, the parsing module 1203 is responsible for processing the raw data from the data acquisition device 110, including the color video data and the ambient temperature data. The parsing module 1203 parses these data into a format that can be used for further analysis, such as parsing the video stream (color video data) into images at different locations in the fire detection area, and parsing the ambient temperature data into temperatures at different locations in the fire detection area.

[0068] Furthermore, the image integration module 1204 receives images of different locations within the fire detection area provided by the analysis module 1203, and stitches these images into a panoramic image within the fire detection area, providing a comprehensive view of the fire detection area. The image integration module 1204 enables the system to go beyond the viewing angle limitation of a single camera, obtain a wider monitoring field of view, and help to fully understand the fire situation.

[0069] Furthermore, the temperature integration module 1205 receives the temperatures of different locations within the fire detection area provided by the analysis module 1203, and integrates the temperatures of different locations within the fire detection area into a temperature distribution within the fire detection area, and displays the temperature changes within the fire detection area. Through the temperature integration module 1205, the system can identify temperature abnormal areas, which is crucial for early fire detection and understanding of fire thermal characteristics.

[0070] Furthermore, the fire spread prediction module 1206 performs flame segmentation detection and temperature mutation analysis to predict the spread direction and location of the fire based on the panoramic image provided by the image integration module 1204 and the temperature distribution provided by the temperature integration module 1205. The fire spread prediction module 1206 enables the system to predict the development trend of the fire, provide a scientific basis for fire fighting decisions, and help deploy resources in advance to reduce fire losses.

[0071] In combination with the above description, multiple dual-light pan-tilt cameras or continuously adjusting the shooting angle and focal length of the dual-light pan-tilt camera are required to capture and obtain color video data and ambient temperature data.

[0072] Optionally, in an example, the color video data includes a plurality of sub-video data and a plurality of sub-temperature data; the data acquisition device 110 includes: a plurality of dual-light pan-tilt cameras.

[0073] Each dual-light PTZ camera is used to shoot different positions in the fire detection area to obtain multiple sub-video data and multiple sub-temperature data.

[0074] It is understandable that setting up multiple dual-light PTZ cameras in the fire detection area can cover a wider area and reduce monitoring blind spots. In practice, cameras in different locations can capture fire details from different angles, which helps to more accurately identify and analyze the fire. The sub-video data and sub-temperature data provided by the dual-light PTZ camera can be used for a variety of analyses, including flame detection, smoke identification, and temperature anomaly analysis. Furthermore, if one camera fails, other cameras can still continue to provide monitoring data, improving the reliability of the system.

[0075] Optionally, in one example, the data acquisition device 110 includes: a dual-light PTZ camera; The dual-light PTZ camera is used to adjust the shooting angle and focal length according to preset shooting requirements, shoot different positions in the fire detection area, and obtain color video data and ambient temperature data.

[0076] In practical applications, the PTZ refers to the camera's support and mobile platform, allowing the camera to rotate and tilt horizontally and vertically to cover a wider monitoring area. The PTZ of the dual-light PTZ camera can be remotely controlled to achieve precise monitoring of a specific area.

[0077] Specifically, according to the preset shooting requirements, such as the size of the shooting range, the shooting position, etc., determine the locations that need to be shot within the fire detection area, adjust the shooting angle and focal length of the dual-light pan-tilt camera in real time, and shoot different locations within the fire detection area to ensure that the system can effectively monitor and respond to environmental changes.

[0078] In this embodiment, a dual-light PTZ camera is used to achieve comprehensive monitoring of the fire detection area while maintaining the flexibility and cost-effectiveness of the system.

[0079] In addition, as an example, in a possible implementation, the above-mentioned brain-inspired computing device 120 also includes: a preprocessing module 1207.

[0080] The preprocessing module 1207 is connected to the data acquisition device 110, the environmental perception module 1201 and the fire target detection module 1202. The preprocessing module 1207 is used to preprocess the ambient temperature data and the color video data, obtain the preprocessed ambient temperature data and the color video data, and transmit the preprocessed ambient temperature data and the color video data to the environmental perception module 1201 and the fire target detection module 1202.

[0081] For example, preprocessing includes but is not limited to: denoising, enhancing, adjusting the frame rate, etc. of the RGB video stream. In one example, the color video data includes three channels (red, green, and blue), each channel representing a color component of the image. The preprocessing module 1207 converts the color video data into single-channel video data to ensure that each frame of the image is consistent with the main bit rate.

[0082] It can be understood that the preprocessing module 1207 preprocesses the ambient temperature data and the color video data to effectively improve the data quality, and transmits the preprocessed ambient temperature data and the color video data to the environment perception module 1201 and the fire target detection module 1202 to improve the accuracy and efficiency of subsequent processing. Therefore, the solution of this embodiment improves the accuracy and reliability of fire target detection.

[0083] Optionally, in a possible implementation, the brain-inspired computing device 120 further includes: a dual-light positioning module 1208 .

[0084] The dual-light calibration module 1208 is connected to the data acquisition device 110, the environment perception module 1201 and the fire target detection module 1202. The dual-light calibration module 1208 is used to perform dual-light calibration on the ambient temperature data and the color video data, obtain calibrated ambient temperature data and color video data, and transmit the calibrated ambient temperature data and color video data to the environment perception module 1201 and the fire target detection module 1202.

[0085] Wherein, bi-modal calibration refers to calibrating the ambient temperature data and the color video data to ensure the consistency of the ambient temperature data and the color video data in terms of space and radiation characteristics. For example, the bi-modal calibration module 1208 performs bi-modal calibration on the ambient temperature data and the color video data, including but not limited to the following processing: spatial alignment, radiation calibration, geometric transformation, pixel mapping, color correction, etc.

[0086] In this embodiment, the dual-light calibration module 1208 performs dual-light calibration on the ambient temperature data and the color video data to ensure the consistency of the ambient temperature data and the color video data in terms of space and radiation characteristics, and transmits the dual-light calibrated ambient temperature data and the color video data to the environment perception module 1201 and the fire target detection module 1202 to improve the accuracy and efficiency of subsequent processing. Therefore, the solution of this embodiment improves the accuracy and reliability of fire target detection.

[0087] Optionally, in a possible implementation, the brain-inspired computing device 120 further includes: an interference suppression module 1209 .

[0088] The interference suppression module 1209 is connected to the data acquisition device 110, the environmental perception module 1201 and the fire target detection module 1202. The interference suppression module 1209 is used to identify and filter interference data in the ambient temperature data and the color video data, and transmit the filtered ambient temperature data and the color video data to the environmental perception module 1201 and the fire target detection module 1202.

[0089] In actual applications, there are changes between frames of flame images. For example, if some pixels of the flame do not change within 1 second, it can be considered that this is not a real flame, but a perception misjudgment, and the subsequent work of the system is terminated and restored to normal state.

[0090] In this embodiment, on the one hand, by effectively identifying and filtering interference caused by non-fire, interference suppression reduces the number of false fire alarms by the system, reduces the false alarm rate, reduces unnecessary emergency responses caused by false alarms, saves manpower, material resources and financial resources, and reduces unnecessary alarms, avoids the troubles caused by frequent false alarms to users, and improves the user experience. On the other hand, the filtered ambient temperature data and color video data are transmitted to the environment perception module 1201 and the fire target detection module 1202 to improve the accuracy and efficiency of subsequent processing, can more accurately identify real fire signals, and improve the accuracy and reliability of fire target detection.

[0091] Optionally, in a possible implementation, the brain-inspired computing device 120 further includes: a display module 1210 .

[0092] The display module 1210 is connected to the environmental perception module 1201, the fire target detection module 1202 and the fire spread prediction module 1206. The display module 1210 is used to obtain alarm information, fire situation information, fire location, predicted direction of fire spread and predicted location of fire spread, and visualize the alarm information, fire situation information, fire location, predicted direction of fire spread and predicted location of fire spread.

[0093] In this embodiment, the display module 1210 visualizes the alarm information, fire situation information, fire location, predicted direction of fire spread, and predicted location of fire spread, greatly improving the availability and efficiency of the fire detection system, so that management personnel can quickly and intuitively understand and respond to fire incidents in emergency situations, thereby improving the real-time and reliability of the fire target detection system.

[0094] It should be noted that the above embodiments can be implemented individually or in combination. Figure 2 FIG. 4 is a structural diagram of the fire target detection system provided by the present invention. Figure 2 As shown, the fire target detection system includes: multiple fire target detection groups 10 and a unified management platform 30, each fire target detection group 10 is set in a fire detection area, and each fire target detection group 10 includes a data acquisition device 110 and a brain-like computing device 120. Among them, the brain-like computing device 120 includes: an environmental perception module 1201, a fire target detection module 1202, a parsing module 1203, an image integration module 1204, a temperature integration module 1205, a fire spread prediction module 1206, a preprocessing module 1207, a dual-light positioning module 1208, an interference suppression module 1209 and a display module 1210.

[0095] Specifically, the data acquisition device 110 is connected to the preprocessing module 1207 , and the data acquisition device 110 acquires the ambient temperature data and the color video data in the fire detection area, and transmits the acquired ambient temperature data and the color video data to the preprocessing module 1207 .

[0096] The preprocessing module 1207 is connected to the dual-light positioning module 1208 . The preprocessing module 1207 preprocesses the ambient temperature data and the color video data to obtain the preprocessed ambient temperature data and the color video data, and transmits the preprocessed ambient temperature data and the color video data to the dual-light positioning module 1208 .

[0097] The dual-light calibration module 1208 is connected to the interference suppression module 1209 , and the dual-light calibration module 1208 performs dual-light calibration on the pre-processed ambient temperature data and color video data to obtain calibrated ambient temperature data and color video data, and transmits the calibrated ambient temperature data and color video data to the interference suppression module 1209 .

[0098] The interference suppression module 1209 is connected to the environmental perception module 1201, the fire target detection module 1202 and the analysis module 1203. The interference suppression module 1209 identifies and filters the interference data in the calibrated ambient temperature data and color video data, and transmits the filtered ambient temperature data and color video data to the environmental perception module 1201, the fire target detection module 1202 and the analysis module 1203.

[0099] The environmental perception module 1201 performs temperature mutation analysis based on the ambient temperature data, and performs flame perception, flame segmentation detection and smoke perception based on the color video data to determine whether a fire occurs in the fire detection area, and outputs alarm information when a fire occurs in the fire detection area.

[0100] The fire target detection module 1202 is connected to the environment perception module 1201. The fire target detection module 1202 is used to detect the color video data in response to the alarm information to determine the fire situation information and the fire location.

[0101] The analysis module 1203 analyzes the color video data to obtain images of different positions in the fire detection area, and analyzes the ambient temperature data to obtain temperatures of different positions in the fire detection area.

[0102] The image integration module 1204 is connected to the analysis module 1203 and is used to stitch images at different positions in the fire detection area to obtain a panoramic image of the fire detection area.

[0103] The temperature integration module 1205 is connected to the analysis module 1203 , and is used to combine the temperatures at different locations within the fire detection area to obtain the temperature distribution within the fire detection area.

[0104] The fire spread prediction module 1206 is connected to the image integration module 1204 and the temperature integration module 1205. The fire spread prediction module 1206 is used to perform flame segmentation detection and temperature mutation analysis based on the panoramic image within the fire detection area and the temperature distribution within the fire detection area, so as to obtain the predicted direction and predicted position of the fire spread.

[0105] The display module 1210 is connected to the environmental perception module 1201, the fire target detection module 1202 and the fire spread prediction module 1206. The display module 1210 is used to obtain alarm information, fire situation information, fire location, predicted direction of fire spread and predicted location of fire spread, and visualize the alarm information, fire situation information, fire location, predicted direction of fire spread and predicted location of fire spread.

[0106] The unified management platform 30 is connected to the brain-like computing device 120, and the brain-like computing device 120 is also used to transmit the target detection results to the unified management platform 30, wherein the target detection results include but are not limited to alarm information, fire situation information, fire location, predicted direction of fire spread, predicted location of fire spread, color video data, and ambient temperature data.

[0107] In the fire target detection system provided in the present embodiment, the fire target detection system includes multiple fire target detection groups and a unified management platform, each fire target detection group is arranged in a fire detection area, and each fire target detection group includes a data acquisition device and a brain-like computing device; the data acquisition device in each fire target detection group is connected to the brain-like computing device, and the data acquisition device is used to collect ambient temperature data and color video data in the fire detection area, and transmit the collected ambient temperature data and color video data to the brain-like computing device; the brain-like computing device is used to perceive the ambient temperature data and color video data based on the environmental perception technology, and determine whether a fire occurs in the fire detection area; and when a fire occurs in the fire detection area, the color video data is detected based on the fire target detection technology to determine the fire situation information and the fire location; each brain-like computing device is connected to the unified management platform, and the brain-like computing device is also used to transmit the fire situation information and the fire location to the unified management platform. The scheme of this embodiment, by setting a data acquisition device and a brain-like computing device in each fire detection area, the data acquisition device can timely collect the ambient temperature data and color video data in the fire detection area. The data acquisition device and the brain-like computing device are deployed in each fire detection area, which can reduce the data transmission time. The brain-like computing device can process the ambient temperature data and color video data collected by the data acquisition device in real time, and realize the rapid response of fire target detection. Furthermore, the brain-like computing device performs fire target detection based on the ambient temperature data and color video data, can determine the fire situation information and the fire location, and can realize accurate detection of the fire; further, the brain-like computing device transmits the fire situation information and the fire location to the unified management platform to realize the centralized monitoring and management of the fire. Therefore, the scheme of this embodiment improves the accuracy and real-time performance of fire target detection.

[0108] The fire target detection method provided by the present invention is described below. The fire target detection method described below and the fire target detection system described above can be referenced to each other.

[0109] The fire target detection method of this embodiment is applied to a fire target detection system. Figure 1 The fire target detection system includes: multiple fire target detection groups 10 and a unified management platform 30, each fire target detection group 10 is set in a fire detection area, and each fire target detection group 10 includes a data acquisition device 110 and a brain-like computing device 120.

[0110] Figure 3 FIG. 1 is a flow chart of the fire target detection method provided by the present invention, such as Figure 3 As shown, the fire target detection method includes step 301 and step 302.

[0111] Step 301: The data acquisition device in each fire target detection group collects ambient temperature data and color video data in the fire detection area, and transmits the collected ambient temperature data and color video data to the brain-like computing device.

[0112] Step 302: The brain-like computing device senses the ambient temperature data and color video data based on the environmental sensing technology to determine whether a fire occurs in the fire detection area; and when a fire occurs in the fire detection area, the brain-like computing device detects the color video data based on the fire target detection technology to determine the fire situation information and the fire location.

[0113] Step 303: The brain-like computing device transmits the fire situation information and fire location to the unified management platform.

[0114] In this embodiment, the fire target detection method can be applied to a wide range of fire detection scenarios, and the fire detection scenarios are divided into multiple fire detection areas. A data acquisition device 110 and a brain-like computing device 120 are respectively set in each fire detection area. It can be understood that by setting the data acquisition device 110 and the brain-like computing device 120 in the fire detection area, real-time fire detection can be achieved in each fire detection area, thereby improving the real-time and reliability of fire target detection.

[0115] The data acquisition device 110 refers to a device capable of acquiring ambient temperature data and color video data. In this embodiment, the data acquisition device 110 is not specifically limited. For example, the data acquisition device 110 may be a dual-light pan-tilt camera, a visible light camera, a multi-spectral camera, or the like.

[0116] As an example, the data acquisition device 110 is a dual-light pan-tilt camera, which is a camera device that integrates visible light and infrared light imaging technologies. It can simultaneously collect ambient temperature data and color (RGB) video data. Specifically, the visible light camera is responsible for shooting color video data, that is, RGB video data. The infrared camera is responsible for capturing thermal radiation in the scene, that is, infrared radiation. This radiation is directly related to the temperature of the object, so it can be used to measure and monitor the ambient temperature. The dual-light pan-tilt camera usually has dual-light fusion technology, which can superimpose visible light images and infrared images on the same screen, or switch the display as needed. This technology allows the operator to observe the visible light image and temperature distribution of the scene at the same time.

[0117] In practical applications, the data acquisition device 110 and the brain-like computing device 120 can be electrically connected or communicatively connected, and the data acquisition device 110 and the brain-like computing device 120 are arranged in the fire detection area. For example, the data acquisition device 110 can transmit the color video data and the ambient temperature data to the brain-like computing device 120 through the local area network. It can be understood that the data acquisition device 110 can transmit the collected ambient temperature data and color video data to the brain-like computing device 120 in real time, and the brain-like computing device 120 can process the ambient temperature data and the color video data in time to realize the detection of the fire. Therefore, the scheme of this embodiment can improve the real-time performance of the fire target detection.

[0118] Exemplarily, the data acquisition device 110 can transmit the color video data and the ambient temperature data to the brain-like computing device 120 in the form of a video stream and a data stream, respectively. The video stream refers to the continuous transmission process of the video data, and the data acquisition device 110 sends the color video data to the brain-like computing device 120 in the form of a continuous stream.

[0119] The brain-like computing device 120 is deployed in the fire detection area to identify the fire target in the fire detection area. Specifically, the brain-like computing device 120 can be a brain-like computer, a brain-like chip, a brain-like server, etc. For example, the brain-like computing device 120 can be a brain-like computer, which is a new type of computer model that simulates the operation of the brain neural network. It builds corresponding computing theories, chip architectures, and application models and algorithms by drawing on the information processing mode and structure of the biological nervous system.

[0120] In actual applications, a fire target detection model is deployed in the brain-like computing device 120. The fire target detection model is an algorithm model built and pre-trained based on brain-like computing and deep learning technology, which is used to identify and detect targets such as flames and smoke in fire images or videos. This model can analyze mutations in color video data and ambient temperature data to determine whether a fire has occurred, and when a fire occurs, determine the location of the fire and fire situation information. For different fire detection areas, deploying a brain-like computer (edge ​​computer) means setting up an edge computing node in each area where fire monitoring is required. Edge computing nodes are close to data sources and can process data quickly, reduce latency, and improve response speed.

[0121] As an example, the fire target detection model can be a model built and trained based on the YOLOv8s flame detection algorithm. Specifically, the brain-like computer is equipped with the KA200 brain-like computing chip. Based on the new storage-computing-in-one, multi-core parallel, heterogeneous fusion architecture, the AI ​​computing power reaches 32TOPS (INT8) and 16TFLOPS (FP16). It mainly receives the video stream and temperature data collected by the dual-light pan-tilt camera, calls LynDriver and LynSDK, and uses the YOLOv8s model for cross-compilation to achieve real-time detection of fire targets.

[0122] In combination with the above description, the fire target detection system can be applied to a wide range of fire detection scenarios, and the fire detection scenarios can be divided into multiple fire detection areas. A data acquisition device 110 and a brain-like computing device 120 are respectively set in each fire detection area.

[0123] The unified management platform 30 is used to uniformly manage multiple target detection areas. The unified management platform 30 relies on advanced information technology, such as the Internet of Things (IoT), big data analysis, artificial intelligence, and cloud computing, to achieve efficient and intelligent management of multiple target detection areas. For example, the unified management platform 30 can be a server, a central computer, a cloud platform (such as a smart fire protection cloud platform), etc.

[0124] In actual applications, when the brain-like computing device 120 determines that a fire has occurred in the fire detection area, it outputs alarm information and determines the fire situation information and the fire location. Furthermore, the brain-like computing device 120 combines the alarm information, the fire situation information, the fire location, the color video data, and the ambient temperature data into a file, names the file according to a pre-set naming rule, and saves the file locally.

[0125] Furthermore, the brain-like computing device 120 transmits the fire situation information and the fire location to the unified management platform 30. Optionally, the brain-like computing device 120 may transmit the alarm information, the fire situation information, the fire location, the color video data and the ambient temperature data, or transmit the locally stored files to the unified management platform 30.

[0126] In actual applications, when the brain-like computing device 120 in the fire target detection system detects a fire in the corresponding fire detection area, it will promptly notify the user through a real-time notification mechanism and synchronize the data to the unified management platform 30 to achieve centralized monitoring and management of the fire.

[0127] As an example, in one possible implementation, the unified management platform 30 includes an integrated management server; the step 303 specifically includes: the brain-like computing device transmits the fire situation information and the fire location to the integrated management server.

[0128] Specifically, the integrated management server is a centralized computing platform that plays a core role in an organization or system, responsible for processing, analyzing and storing data from different sources and providing decision support. For example, the integrated management server can be a physical server, a cloud server, a virtualized server, etc.

[0129] The scheme of this embodiment, by setting a data acquisition device 110 and a brain-like computing device 120 in each fire detection area, the data acquisition device 110 can timely collect the ambient temperature data and color video data in the fire detection area. The data acquisition device 110 and the brain-like computing device 120 are deployed in each fire detection area, which can reduce the data transmission time. The brain-like computing device 120 can process the ambient temperature data and color video data collected by the data acquisition device 110 in real time, and realize the rapid response of fire target detection. Further, the brain-like computing device 120 performs fire target detection based on the ambient temperature data and color video data, can determine the fire situation information and the fire location, and can realize accurate detection of the fire; further, the brain-like computing device 120 transmits the fire situation information and the fire location to the unified management platform 30 to realize the centralized monitoring and management of the fire. Therefore, the scheme of this embodiment improves the accuracy and real-time performance of fire target detection.

[0130] In addition, in some implementations, the above step 302 includes: The environment sensing module 1201 performs temperature mutation analysis based on the environment temperature data, and performs flame sensing, flame segmentation detection and smoke sensing based on the color video data to determine whether a fire occurs in the fire detection area, and outputs alarm information when a fire occurs in the fire detection area; The fire target detection module 1202 detects the color video data in response to the alarm information to determine the fire situation information and the fire location.

[0131] In addition, in some embodiments, the above method further comprises: The parsing module 1203 parses the color video data to obtain images of different positions in the fire detection area, and parses the ambient temperature data to obtain temperatures of different positions in the fire detection area; The image integration module 1204 stitches images at different locations within the fire detection area to obtain a panoramic image within the fire detection area; The temperature integration module 1205 combines the temperatures at different locations within the fire detection area to obtain the temperature distribution within the fire detection area; The fire spread prediction module 1206 performs flame segmentation detection and temperature mutation analysis based on the panoramic image within the fire detection area and the temperature distribution within the fire detection area to obtain the predicted direction and position of the fire spread.

[0132] In addition, in some embodiments, the color video data includes a plurality of sub-video data and a plurality of sub-temperature data; the data acquisition device 110 includes: a plurality of dual-light pan-tilt cameras; the step 301 specifically includes: Each dual-light PTZ camera shoots different positions within the fire detection area to obtain multiple sub-video data and multiple sub-temperature data.

[0133] In addition, in some embodiments, the data acquisition device 110 includes: a dual-light pan-tilt camera; the step 301 specifically includes: The dual-light PTZ camera adjusts the shooting angle and focal length according to the preset shooting requirements, shoots different positions in the fire detection area, and obtains color video data and ambient temperature data.

[0134] In addition, in one embodiment, before the above step 301, the above method further includes: The preprocessing module 1207 preprocesses the ambient temperature data and the color video data to obtain the preprocessed ambient temperature data and the color video data, and transmits the preprocessed ambient temperature data and the color video data to the environment perception module 1201 and the fire target detection module 1202 .

[0135] In addition, in one embodiment, before the above step 301, the above method further includes: The dual-light calibration module 1208 performs dual-light calibration on the ambient temperature data and the color video data to obtain calibrated ambient temperature data and color video data, and transmits the calibrated ambient temperature data and color video data to the environment perception module 1201 and the fire target detection module 1202 .

[0136] In addition, in one embodiment, before the above step 301, the above method further includes: The interference suppression module 1209 identifies and filters interference data in the ambient temperature data and the color video data, and transmits the filtered ambient temperature data and the color video data to the environment perception module 1201 and the fire target detection module 1202 .

[0137] Optionally, in a possible implementation manner, after the above step 302, the above method further includes: The display module 1210 obtains alarm information, fire situation information, fire location, predicted direction of fire spread and predicted location of fire spread, and visualizes the alarm information, fire situation information, fire location, predicted direction of fire spread and predicted location of fire spread.

[0138] The scheme of this embodiment, by setting a data acquisition device and a brain-like computing device in each fire detection area, the data acquisition device can timely collect the ambient temperature data and color video data in the fire detection area. The data acquisition device and the brain-like computing device are deployed in each fire detection area, which can reduce the data transmission time. The brain-like computing device can process the ambient temperature data and color video data collected by the data acquisition device in real time, and realize the rapid response of fire target detection. Furthermore, the brain-like computing device performs fire target detection based on the ambient temperature data and color video data, can determine the fire situation information and the fire location, and can realize accurate detection of the fire; further, the brain-like computing device transmits the fire situation information and the fire location to the unified management platform to realize the centralized monitoring and management of the fire. Therefore, the scheme of this application improves the accuracy and real-time performance of fire target detection.

[0139] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A fire target detection system, characterized in that: include: Multiple fire target detection groups and a unified management platform, each of the fire target detection groups is arranged in a fire detection area, and each of the fire target detection groups includes a data acquisition device and a brain-like computing device; The data acquisition device in each of the fire target detection groups is connected to the brain-like computing device, and the data acquisition device is used to collect ambient temperature data and color video data in the fire detection area, and transmit the collected ambient temperature data and color video data to the brain-like computing device; The brain-like computing device is used to sense the ambient temperature data and the color video data based on the environmental sensing technology to determine whether a fire occurs in the fire detection area; And, when a fire occurs in the fire detection area, the color video data is detected based on the fire target detection technology to determine the fire situation information and the fire location; Each of the brain-like computing devices is connected to the unified management platform, and the brain-like computing device is also used to transmit the fire situation information and the fire location to the unified management platform.

2. The fire target detection system according to claim 1, characterized in that: The brain-like computing device includes: an environment perception module and a fire target detection module; The environment sensing module is connected to the data acquisition device, and is used to perform temperature mutation analysis based on the ambient temperature data, and flame perception, flame segmentation detection and smoke perception based on the color video data to determine whether a fire occurs in the fire detection area, and output alarm information when a fire occurs in the fire detection area; The fire target detection module is connected to the environment perception module and the data acquisition device. The fire target detection module is used to detect the color video data in response to the alarm information to determine the fire situation information and the fire location.

3. The fire target detection system according to claim 2, characterized in that: The brain-like computing device also includes: an analysis module, an image integration module, a temperature integration module and a fire spread prediction module; The analysis module is connected to the data acquisition device, and is used to analyze the color video data to obtain images of different positions in the fire detection area, and to analyze the ambient temperature data to obtain temperatures of different positions in the fire detection area; The image integration module is connected to the analysis module, and the image integration module is used to stitch images at different positions within the fire detection area to obtain a panoramic image within the fire detection area; The temperature integration module is connected to the analysis module, and is used to combine the temperatures at different locations within the fire detection area to obtain the temperature distribution within the fire detection area; The fire spread prediction module is connected to the image integration module and the temperature integration module. The fire spread prediction module is used to perform flame segmentation detection and temperature mutation analysis based on the panoramic image within the fire detection area and the temperature distribution within the fire detection area to obtain the predicted direction and predicted position of fire spread.

4. The fire target detection system according to claim 1, characterized in that: The color video data includes a plurality of sub-video data and a plurality of sub-temperature data; the data acquisition device includes: a plurality of dual-light PTZ cameras; Each of the dual-light PTZ cameras is used to shoot different positions in the fire detection area to obtain the multiple sub-video data and the multiple sub-temperature data; or, The data acquisition device comprises: a dual-light PTZ camera; The dual-light pan-tilt camera is used to adjust the shooting angle and focal length according to preset shooting requirements, shoot different positions in the fire detection area, and obtain the color video data and the ambient temperature data.

5. The fire target detection system according to claim 2, characterized in that: The brain-like computing device further includes: a preprocessing module; The preprocessing module is connected to the data acquisition device, the environmental perception module and the fire target detection module. The preprocessing module is used to preprocess the ambient temperature data and the color video data to obtain the preprocessed ambient temperature data and color video data, and transmit the preprocessed ambient temperature data and color video data to the environmental perception module and the fire target detection module.

6. The fire target detection system according to claim 2, characterized in that: The brain-like computing device further includes: a dual-light positioning module; The dual-light calibration module is connected to the data acquisition device, the environmental perception module and the fire target detection module. The dual-light calibration module is used to perform dual-light calibration on the ambient temperature data and the color video data to obtain calibrated ambient temperature data and color video data, and transmit the calibrated ambient temperature data and color video data to the environmental perception module and the fire target detection module.

7. The fire target detection system according to claim 2, characterized in that: The brain-like computing device further includes: an interference suppression module; The interference suppression module is connected to the data acquisition device, the environmental perception module and the fire target detection module. The interference suppression module is used to identify and filter interference data in the ambient temperature data and the color video data, and transmit the filtered ambient temperature data and color video data to the environmental perception module and the fire target detection module.

8. The fire target detection system according to claim 3, characterized in that: The brain-like computing device further includes: a display module; The display module is connected to the environmental perception module, the fire target detection module and the fire spread prediction module, and the display module is used to obtain the alarm information, the fire situation information, the fire location, the predicted direction of the fire spread and the predicted position of the fire spread, and to visualize the alarm information, the fire situation information, the fire location, the predicted direction of the fire spread and the predicted position of the fire spread.

9. The fire target detection system according to any one of claims 1 to 8, characterized in that: The unified management platform includes: an integrated management server.

10. A fire target detection method, characterized in that: Applied to a fire target detection system, the fire target detection system comprises: a plurality of fire target detection groups and a unified management platform, each of the fire target detection groups is arranged in a fire detection area, and each of the fire target detection groups comprises a data acquisition device and a brain-like computing device; the method comprises: The data acquisition device in each of the fire target detection groups collects ambient temperature data and color video data within the fire detection area, and transmits the collected ambient temperature data and color video data to the brain-like computing device; The brain-like computing device senses the ambient temperature data and the color video data based on the environmental sensing technology to determine whether a fire occurs in the fire detection area; and when a fire occurs in the fire detection area, the brain-like computing device detects the color video data based on the fire target detection technology to determine the fire situation information and the fire location; The brain-like computing device transmits the fire situation information and the fire location to the unified management platform.

Images