Fire early warning and alarming device for battery car and working method of fire early warning and alarming device

Abstract

The invention relates to a fire early warning and alarming device for a battery car and a working method of the fire early warning and alarming device. The device comprises a bottom plate, a box body, a Raspberry Pi, a camera, an infrared temperature detection module and a buzzer. According to the method, when an electric vehicle enters a station, a real-time video stream and temperature information are captured, and then the real-time video stream is preprocessed through the Raspberry Pi to obtain video information; carrying out object detection on the video information by using a YOLO5 model, and marking a battery car area; the temperature information and the video information are processed, and the temperature change condition and whether a fire feature appears in the battery car area are observed; and if the temperature change is abnormal or a fire feature occurs, a buzzer is triggered to give an alarm, and the alarm is transmitted to a remote server. Compared with the prior art, the method has the advantages that the accuracy of fire early warning is improved, the fire false alarm rate is reduced, and the timeliness of fire early warning information is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of electric vehicle fire warning, and in particular to a battery vehicle fire warning and alarm device and a working method thereof. Background Art

[0002] With the advancement of new energy technology and the enhancement of environmental awareness, the number of electric vehicles has increased rapidly. However, electric vehicles are prone to problems such as line aging and short circuit. In addition, due to the irregular charging behavior of some users and the poor quality of individual batteries during use, there are huge safety hazards in the use of electric vehicles. Once a malfunction occurs, it is easy to cause a fire, seriously threatening people's lives and property.

[0003] The battery vehicle detection and warning in the market are mainly based on the separate use of temperature recognition system and image recognition technology. Among them, most of the battery vehicle alarm devices mainly use temperature recognition technology, while most of the battery vehicle detection systems mainly use image recognition technology based on opencv, yolo5, etc.

[0004] From the perspective of temperature recognition, the mainstream temperature recognition methods include thermal imaging temperature recognition, which uses a thermal imaging camera to perform relatively accurate temperature imaging of all surfaces of objects within the range, making it easier to know the overall temperature and relative temperature of objects within the range, and to easily identify high and low temperature areas. However, in general, it is impossible to accurately know the temperature of the object to be measured within the range.

[0005] At present, there are existing technologies that combine cameras and sensors to provide battery vehicle fire warnings, but most of them are smoke sensors or gas sensors, and there is still a large rate of false fire judgments. For example, the invention patent with publication number CN118280089B discloses a battery fire warning method and system for battery vehicles based on data analysis. The method obtains battery vehicle environment image data in the surrounding environment based on an image sensor, and obtains real-time gas data in the surrounding environment based on a gas sensor; and uses a ZigBee wireless communication module to transmit target gas data information and the training battery vehicle environment image data to the system to generate a fire warning signal. The method can comprehensively utilize image data and gas data for analysis, but the method can only prevent fires caused by batteries, and cannot accurately warn electric vehicle fires caused by other causes, and the reliability is insufficient.

[0006] The current battery vehicle fire warning and alarm system is based on only a single technology such as temperature detection or image recognition, and has problems such as single usage scenario, low recognition accuracy, and high false alarm rate. For example, in the invention patent with publication number CN117046003A, an infrared thermal imaging sensor is connected to the control unit to detect whether the temperature in the monitoring area exceeds a given threshold, and if it exceeds, the sound and light alarm unit is triggered. However, no camera is added to identify the location of the battery of the battery vehicle and the characteristics of the fire, so the recognition accuracy is not high and false alarms are prone to occur.

[0007] The current electric vehicle fire warning and alarm system still has the problem of a single control module and high module price, which makes the system difficult to promote. For example, the system in the invention patent with publication number CN114288588B includes a control module, a fire detector and an alarm. The control module uses a PLC controller, which is large in size, expensive, not suitable for promotion, and has limited applicable scenarios.

[0008] There are still some problems with the battery vehicle fire warning and alarm products on the market. The products are single-function and can only be used for fire detection, but cannot complete effective alarm. For example, the invention patent with the publication number CN118928094A includes a fire capture and detection system, a cloud control platform and a mobile platform. This patent uses a convolutional neural network to detect fires, but only transmits fire information to the mobile platform. On the administrator side, messages are easily missed, and effective alarms and warnings to administrators cannot be completed. Summary of the invention

[0009] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a battery vehicle fire warning and alarm device and a working method thereof.

[0010] The purpose of the present invention can be achieved by the following technical solutions:

[0011] According to one aspect of the present invention, a battery vehicle fire warning and alarm device is provided, comprising a bottom plate 1 and a box body 2, the bottom plate 1 is provided with four centrally symmetrical threaded holes, and the bottom plate 1 is fixedly connected to the wall by passing expansion bolts through the threaded holes; the box body 2 is a hexahedron with a hollow structure, the front end face of the box body 2 is at an angle of 60 degrees to the horizontal line of the ground, the rear end face is perpendicular to the ground and fixedly connected to the bottom plate 1, and the two side faces are right-angled trapezoids;

[0012] A camera 4 and an infrared temperature detection module 5 are fixed on one side of the front end surface inside the box body 2, and observation openings of the camera 4 and the infrared temperature detection module 5 are set. A Raspberry Pi 3 is fixed on one side of the rear end surface 1 inside the box body 2, and the Raspberry Pi 3 is electrically connected to the output end of the camera 4 and the output end of the infrared temperature detection module 5;

[0013] A buzzer 6 is installed inside the box 2 . The buzzer 6 is fixedly connected to the front side of the Raspberry Pi and is electrically connected to the Raspberry Pi 3 .

[0014] As a preferred technical solution, the diameter of the threaded hole on the bottom plate 1 is 8 mm.

[0015] As a preferred technical solution, the thickness of the six surfaces of the box body 2 is 5 mm.

[0016] As a preferred technical solution, Raspberry Pi 3 includes a Raspberry Pi main control module and a Raspberry Pi shell. The Raspberry Pi main control module is connected to a power cord. The Raspberry Pi shell is a hollow structure made of plastic. A WiFi module is set in the Raspberry Pi 3, and the Raspberry Pi 3 is wirelessly connected to the remote server via the WiFi module.

[0017] As a preferred technical solution, a power interface 301 is provided on the side of the box body 2, and the power line of the Raspberry Pi main control module passes through the power interface 301 to be electrically connected to the external power supply.

[0018] According to another aspect of the present invention, a working method of a battery vehicle fire warning and alarm device is provided. The battery vehicle fire warning and alarm device as described above is operated by using the method, and the method comprises the following steps:

[0019] S1, when the electric vehicle enters the station, capture the real-time video stream and temperature information and transmit it to the Raspberry Pi;

[0020] S2, preprocessing each frame of the real-time video stream through the Raspberry Pi to obtain video information;

[0021] S3, apply the YOLO5 model to detect objects in the current frame of the video information, identify and mark the battery vehicle area;

[0022] S4, using image recognition technology to process video information, observe whether there are fire characteristics in the battery vehicle area, if so, proceed to step S5; process temperature information, observe temperature changes, if the temperature changes are abnormal, proceed to step S5;

[0023] S5. Trigger the buzzer alarm and transmit the alarm to the remote server.

[0024] As a preferred technical solution, OpenCV technology is selected as the image processing technology in S4.

[0025] As a preferred technical solution, the fire characteristics in S4 include smoke and flames.

[0026] As a preferred technical solution, the processing of temperature information in S4 specifically includes analyzing the change of temperature information over time on the basis of ambient temperature compensation, and recording the normal increase rate of temperature.

[0027] As a preferred technical solution, the abnormal temperature change in S4 includes: the temperature increase rate exceeds the normal increase rate by 10% or more, or the temperature suddenly rises for more than 1 minute after the electric vehicle is stably charged.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] 1. In a battery vehicle fire warning and alarm device of the present invention, a camera 4 and an infrared temperature detection module 5 are fixed on the front of the box, and a Raspberry Pi 3 is fixed on one side of the bottom plate 1 inside the box 2, and the Raspberry Pi 3 is electrically connected to the output end of the camera 4 and the output end of the infrared temperature detection module 5. This allows the camera and the infrared temperature detection module to directly cooperate with the Raspberry Pi to monitor the charging environment and status of the battery vehicle in real time, combining the image recognition algorithm of yolov5 with temperature detection, and realizing double protection and early warning through image processing or temperature abnormality recognition mechanism, breaking through the limitations of traditional single detection, effectively improving the accuracy of fire warning, and reducing the false alarm rate of fire.

[0030] 2. In a battery vehicle fire warning and alarm device of the present invention, the buzzer 6 is electrically connected to the Raspberry Pi 3 and the Raspberry Pi is wirelessly connected to the remote server via a WiFi module. Alarms are issued in two ways to ensure that the administrator receives notification at the first time. When a fire occurs, the alarm is immediately sounded and the administrator is warned. The system is more diverse, which helps to reduce the losses caused by fire and ensures the timeliness of fire warning information.

[0031] 3. In the working method of a battery vehicle fire warning and alarm device in the present invention, real-time video stream and temperature information are jointly processed, that is, video information is processed using image recognition technology to observe whether there are fire characteristics in the battery vehicle area, and temperature information is processed at the same time to observe temperature changes. If the temperature changes abnormally or fire characteristics appear, the buzzer alarm is triggered and the alarm is transmitted to the remote server. This makes it possible to detect fire characteristics or temperature anomalies through images to achieve early warning, especially in the early stage of a fire, and effectively improve the response speed of fire warning.

[0032] 4. The battery vehicle fire warning and alarm device of the present invention uses Raspberry Pi 3 as the core processor, integrates a camera 4 and an infrared temperature detection module 5, so that it does not need to rely on the cloud, has the characteristics of low latency and fast response speed, and is suitable for fire scenes with high response time requirements. In addition, Raspberry Pi 3 has low power consumption and strong hardware performance. Compared with similar products, it realizes functions at a lower cost, which is more conducive to promotion and application.

[0033] 5. In the present invention, the Raspberry Pi 3 is wirelessly connected to the remote server via the WiFi module, so that the invention can store video data and upload it to the remote server, realize video stream monitoring, and the alarm signal will also remotely notify the administrator via the WiFi module, which is more reliable and helps to reduce the losses caused by fire.

[0034] 6. When processing temperature information in the present invention, ambient temperature compensation is performed to eliminate the interference of ambient temperature, and only the temperature change of the core part of the battery vehicle is retained, ensuring the accuracy of temperature detection. In the present invention, image processing is used to identify electric vehicle objects, making fire warning more targeted and effectively eliminating the influence of environmental factors on the accuracy of fire warning.

[0035] 7. The image processing technology in the present invention uses OpenCV technology, which significantly improves the response speed of image processing, and the result of image processing has high accuracy, making the method efficient and highly practical. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is an axonometric view of the appearance of a battery vehicle fire warning and alarm device in the present invention;

[0037] Figure 2 A cross-sectional view of a battery vehicle fire warning and alarm device in the present invention;

[0038] Figure 3 This is a schematic diagram of the camera and infrared temperature detection module inside the box of the present invention;

[0039] Figure 4 A side view of one side of the power interface of the present invention;

[0040] Figure 5 The present invention is a flowchart of the working method of a battery vehicle fire early warning and alarm device.

[0041] In the figure, 1 is the base plate, 2 is the box, 3 is the Raspberry Pi, 301 is, 4 is the camera, 5 is the infrared temperature detection module, and 6 is the buzzer. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments 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 should fall within the scope of protection of the present invention.

[0043] Nowadays, temperature recognition systems based on infrared temperature sensors and image recognition systems based on image recognition technology principles such as opencv and yolo5 are the mainstream technologies used in the basic devices and systems of battery vehicle detection and early warning systems on the market. Among them, most battery vehicle alarm devices mainly use only temperature recognition technology, while most battery vehicle detection systems mainly use only image recognition technology. In recent years, with the development of software and hardware technology, the market has more mature, more powerful and cheaper embedded development boards, and it has become more convenient to combine the functions of image recognition and temperature recognition into the battery vehicle fire warning system.

[0044] From the perspective of temperature recognition, there are two mainstream temperature recognition methods. First, the use of infrared temperature sensors, such as body temperature monitoring guns, is to use infrared temperature sensors to remotely identify the target temperature. The advantage of this method is that the application cost is relatively cheap, and the cost of infrared temperature sensors is low. Second, thermal imaging temperature recognition uses thermal imaging cameras to perform relatively accurate temperature imaging of all object surfaces within the range. The advantage is that it is easy to know the overall object temperature and relative temperature within the range, and it is easier to see the high temperature and low temperature areas. The disadvantage is that under normal circumstances, it is impossible to accurately know the temperature of the object to be measured within the range.

[0045] At present, the technologies that can effectively integrate the advantages of image recognition technology and temperature recognition technology mainly include: registration and fusion of infrared thermal imaging and optical images, and the main application is infrared face recognition: using the thermal image captured by the infrared camera, using infrared thermal imaging and image recognition technology, combined with the face database, to achieve a high accuracy of infrared face recognition. Infrared thermal imaging and image recognition technology can also be applied to battery vehicle identification, used to monitor the temperature of battery vehicles, and to warn of battery vehicle fires; infrared temperature sensors combined with image recognition technology can be used in situations where the object being measured is relatively static, such as when the battery vehicle is charging, the battery vehicle is stationary, and the parking direction is relatively regular. At this time, the camera can be used to identify the position of the battery vehicle on the screen, and the temperature sensor can obtain the temperature data of the battery vehicle.

[0046] Regardless of which of the above methods is used, the key lies in the comprehensive application of image recognition technology and temperature recognition technology.

[0047] The current battery vehicle fire warning and alarm systems on the market are based on only a single technology such as temperature detection or image recognition, and have problems such as single usage scenarios, low recognition accuracy, and high false alarm rate. This solution uses a dual mechanism of image processing and temperature detection, and innovatively combines the image recognition algorithm such as yolov5 with temperature detection, breaking through the limitations of traditional single detection and ensuring the accuracy of fire identification.

[0048] In this solution, the combined judgment of image recognition and temperature detection can eliminate false alarms and greatly reduce the false alarm rate. For example, when a high-temperature object passes by, if it does not cause an abnormal temperature change in the battery vehicle, the system will not trigger an alarm. The combined judgment of image recognition and temperature detection greatly improves the reliability and accuracy of the system.

[0049] This solution also uses Raspberry Pi 3 as the core processor, integrates camera 4 and infrared temperature detection module 5, and uses yolov5 algorithm to realize local data processing. Compared with similar products on the market, it does not need to rely on the cloud, has the characteristics of low latency and fast response speed, and is suitable for fire scenes with high response time requirements. In addition, Raspberry Pi has low power consumption and strong hardware performance. Compared with similar products, it realizes functions at a lower cost, which is more conducive to promotion and application.

[0050] Such products on the market have a single function and can only be used for fire detection. However, this solution can store video data and upload it to a remote server to achieve video stream monitoring. The alarm signal will also notify the administrator remotely, which is more reliable.

[0051] The specific implementation of this scheme is described in Examples 1-3.

[0052] Example 1

[0053] This embodiment applies a battery vehicle fire warning and alarm device, which can not only accurately detect temperature anomalies when the battery vehicle is charging, but also determine whether the battery vehicle has fire hazards, such as battery smoke, flames, etc. through image recognition technology, thereby greatly improving the accuracy of fire warning.

[0054] The device comprises a bottom plate 1 and a box body 2. The structure of the device is as follows Figure 1-Figure 4 As shown, Figure 1 This is the axonometric view of the device; Figure 2 is a cross-sectional view of the device; Figure 3 This is a schematic diagram of the camera and infrared temperature detection module inside the box; Figure 4 This is a side view of the power interface.

[0055] The bottom plate 1 of the device is provided with four centrally symmetrical threaded holes with a diameter of 8 mm, through which expansion bolts are passed to fix the bottom plate 1 to the wall. The box 2 is a hexahedron with a hollow structure, and the thickness of the six faces is 5 mm. Its front face is at an angle of 60 degrees to the horizontal line of the ground, which is convenient for the camera 4 to capture the panoramic image of the battery vehicle while avoiding dust accumulation or other external interference. The rear face is perpendicular to the ground and fixedly connected to the bottom plate 1, and the two side faces are right-angled trapezoids. Sufficient space is reserved in the box to place the Raspberry Pi fixed therein.

[0056] In this embodiment, a camera 4 and an infrared temperature detection module 5 are fixed on one side of the front end surface inside the box body 2, and observation openings of the camera 4 and the infrared temperature detection module 5 are set. A Raspberry Pi 3 is fixed on one side of the rear end surface 1 inside the box body 2, and the Raspberry Pi 3 is electrically connected to the output end of the camera 4 and the output end of the infrared temperature detection module 5.

[0057] In this embodiment, the Raspberry Pi 3 is the core module of the entire device, which is used for calculation and control. It includes a Raspberry Pi main control module and a Raspberry Pi shell. The Raspberry Pi shell is a hollow structure and is made of plastic. The Raspberry Pi 3 processes the video stream captured from the camera through a program written in Python, and analyzes the temperature data collected by the infrared temperature detection module. The Raspberry Pi main control module is connected to a power cord, and a power interface 301 is provided on the side of the box 2. In this embodiment, the power cord of the Raspberry Pi main control module passes through the power interface 301 and is connected to a 5V3A external power supply through a USB power interface.

[0058] In this embodiment, a WiFi module is set in the Raspberry Pi 3, and the Raspberry Pi 3 is wirelessly connected to the remote server via the WiFi module. When the temperature detects an abnormality or the image recognizes a fire risk, the system sends the alarm information to the remote server through the Raspberry Pi's Wi-Fi module. At this time, the manager can receive the alarm information through the monitoring interface of the mobile phone or PC, and can view the on-site video and make decisions in time.

[0059] In this embodiment, the camera 4 captures the image of the battery vehicle parking area in real time, and uses the OpenCV library to perform object detection and position tracking. Through image analysis, the state of the battery vehicle is monitored in real time to determine whether there are abnormalities such as flames and smoke.

[0060] In this embodiment, the infrared temperature detection module 5 is responsible for measuring the temperature near the battery vehicle. The Raspberry Pi 3 receives the data from the module. Once it detects that the temperature increase rate exceeds the original detected increase rate by 10% or more, or the temperature suddenly rises for more than 1 minute when it has stabilized during charging, the Raspberry Pi 3 will send an alarm signal to the buzzer 6 and the remote server. In order to ensure the accuracy of temperature detection, the system will detect the ambient temperature to eliminate the interference of the ambient temperature, and only retain the temperature changes in the core parts of the battery vehicle.

[0061] In this embodiment, the buzzer 6 is arranged inside the box 2, which is fixedly connected to the front side of the Raspberry Pi and is electrically connected to the Raspberry Pi 3. After the camera 4 and the Raspberry Pi 3 jointly perform target recognition and complete the temperature detection of the target, if the alarm requirement is met, that is, the temperature detected by the infrared temperature detection module 4 shows an abnormal change, the Raspberry Pi 3 controls the buzzer 6 through the GPIO to sound an alarm to remind the user to deal with it in time.

[0062] In this embodiment, the specific working method of the device is as follows: Figure 5 As shown, the method realizes the joint judgment of image recognition and temperature detection, and minimizes the false alarm rate by comprehensively analyzing the dual data of the camera and the infrared temperature detection module. The method includes the following steps:

[0063] S1, when the electric vehicle enters the station, capture the real-time video stream and temperature information and transmit it to the Raspberry Pi;

[0064] S2, preprocessing each frame of the real-time video stream through the Raspberry Pi to obtain video information;

[0065] S3, apply the YOLO5 model to detect objects in the current frame of the video information, identify and mark the battery vehicle area;

[0066] S4, using image recognition technology to process video information, observe whether there are fire characteristics in the battery vehicle area, if so, proceed to step S5; process temperature information, observe temperature changes, if the temperature changes are abnormal, proceed to step S5;

[0067] S5. Trigger the buzzer alarm and transmit the alarm to the remote server.

[0068] Among them, the image processing technology in S4 uses OpenCV technology; fire characteristics include smoke and flames.

[0069] The processing of the temperature information in S4 specifically includes analyzing the change of the temperature information over time and performing ambient temperature compensation to reduce the influence of the ambient temperature change on the measurement result.

[0070] The abnormal temperature change in S4 includes: the temperature increase rate exceeds the originally detected temperature increase rate by 10% or more, or the temperature suddenly rises for more than 1 minute after the electric vehicle is stably charged.

[0071] In this embodiment, the device is fixed to the wall of the charging garage through a base plate, and the camera and the infrared temperature detection module are installed and fixed on the wall at a height of about 1.6m from the ground.

[0072] In this embodiment, first, the device is connected to a 5V3A power supply through a USB power interface, and it is confirmed that the Raspberry Pi 3, the camera 4 and the infrared temperature detection module 5 are started and connected normally.

[0073] In this embodiment, the battery in an uncharged battery vehicle is taken out to simulate the condition of stable temperature of the battery vehicle, and a sponge soaked in boiling water is fixed on the side of the seat of the battery vehicle to simulate the condition of abnormal temperature rise.

[0074] In this embodiment, when the battery car enters the monitoring area, the camera 4 captures the real-time video stream to pre-process the image (grayscale, denoising). At this time, the camera 4 captures the real-time video stream to pre-process the image and detect the battery car target. The infrared temperature detection module 5 collects the battery car and ambient temperature data, and records it once per second. The system performs ambient temperature compensation on the battery car temperature data to remove the interference of external temperature fluctuations, ensuring that the monitored temperature changes are as close to the actual temperature state of the battery car as possible. To improve the efficiency and accuracy of subsequent processing; then the video data is transmitted to the Raspberry Pi 3 via the OpenCV interface, and the Raspberry Pi 3 uses the YOLO5 model to detect the object in the current frame, identify the position and boundary box of the battery car, and complete the target detection of the battery car.

[0075] Raspberry Pi 3 monitors the status of the battery vehicle in real time through image recognition to determine whether there are abnormalities such as flames and smoke;

[0076] At the same time, the infrared temperature detection module 5 collects the battery vehicle and the ambient temperature data, records it once per second, and converts the temperature data into an electrical signal and transmits it to the Raspberry Pi 3. The Raspberry Pi 3 receives the data from the module. Once it detects that the temperature increase rate exceeds the original detected increase rate by 10% or more, or the temperature suddenly rises for more than 1 minute when it has stabilized during charging, the Raspberry Pi 3 will send an alarm signal to the buzzer and the remote server. In order to ensure the accuracy of temperature detection, the system will detect the ambient temperature to eliminate the interference of the ambient temperature, and only retain the temperature changes in the core parts of the battery vehicle.

[0077] On the one hand, the Raspberry Pi 3 uses its built-in YOLO5 model to process video data to detect electric vehicle targets, and detects whether there is flame or smoke in the entire image area through image recognition; on the other hand, it detects temperature changes. Once it is detected that the temperature increase rate exceeds the original detected increase rate by 10% or more, or the temperature suddenly rises for more than 1 minute when it has stabilized during charging, an alarm signal is generated.

[0078] In this embodiment, what is simulated is a situation where the temperature suddenly rises for more than 1 minute after it has stabilized during charging, and an alarm signal is generated at this time.

[0079] When an alarm signal is generated, the Raspberry Pi 3 controls the buzzer 6 through GPIO, and the buzzer 6 sounds an alarm to remind the user to deal with it in time. At the same time, the alarm signal is sent to the remote server through the Wi-Fi module of the Raspberry Pi. At this time, the manager can receive the alarm information through the monitoring interface of the mobile phone or PC, and can view the on-site video and make decisions in time.

[0080] In this embodiment, after the device is started, the buzzer immediately sounds an alarm after about 62 seconds, starting from the time the sponge is fixed, and the remote server also immediately receives the alarm after about 64 seconds, and stores the alarm time in the database. This solution can not only quickly detect fire risks, but also quickly transmit alarm information to the monitoring center through a wireless network, so that management personnel can respond and handle it in time, ultimately achieving early warning and efficient handling of fires.

[0081] In summary, in this embodiment, the scheme enables an effective alarm to be sounded when the battery car temperature rises rapidly. When the battery car temperature rises abnormally for one minute, the system can trigger the buzzer alarm in time, and the alarm information can be uploaded to the server within 5 seconds. The device has the reliability of warning battery car fires in abnormal situations and alarming when a fire has already occurred.

[0082] Example 2

[0083] This embodiment applies a battery vehicle fire warning and alarm device, which can not only accurately detect temperature anomalies when the battery vehicle is charging, but also determine whether the battery vehicle has fire hazards, such as battery smoke, flames, etc. through image recognition technology, thereby greatly improving the accuracy of fire warning.

[0084] The device comprises a bottom plate 1 and a box body 2. The structure of the device is as follows Figure 1-Figure 4 As shown, Figure 1 This is the axonometric view of the device; Figure 2 is a cross-sectional view of the device; Figure 3 This is a schematic diagram of the camera and infrared temperature detection module inside the box; Figure 4 This is a side view of the power interface.

[0085] The bottom plate 1 of the device is provided with four centrally symmetrical threaded holes with a diameter of 8 mm, through which expansion bolts are passed to fix the bottom plate 1 to the wall. The box 2 is a hexahedron with a hollow structure, and the thickness of the six faces is 5 mm. Its front face is at an angle of 60 degrees to the horizontal line of the ground, which is convenient for the camera 4 to capture the panoramic image of the battery vehicle while avoiding dust accumulation or other external interference. The rear face is perpendicular to the ground and fixedly connected to the bottom plate 1, and the two side faces are right-angled trapezoids. Sufficient space is reserved in the box to place the Raspberry Pi fixed therein.

[0086] In this embodiment, a camera 4 and an infrared temperature detection module 5 are fixed on one side of the front end surface inside the box body 2, and observation openings of the camera 4 and the infrared temperature detection module 5 are set. A Raspberry Pi 3 is fixed on one side of the rear end surface 1 inside the box body 2, and the Raspberry Pi 3 is electrically connected to the output end of the camera 4 and the output end of the infrared temperature detection module 5.

[0087] In this embodiment, the Raspberry Pi 3 is the core module of the entire device, which is used for calculation and control. It includes a Raspberry Pi main control module and a Raspberry Pi shell. The Raspberry Pi shell is a hollow structure and is made of plastic. The Raspberry Pi 3 processes the video stream captured from the camera through a program written in Python, and analyzes the temperature data collected by the infrared temperature detection module. The Raspberry Pi main control module is connected to a power cord, and a power interface 301 is provided on the side of the box 2. In this embodiment, the power cord of the Raspberry Pi main control module passes through the power interface 301 and is connected to a 5V3A external power supply through a USB power interface.

[0088] In this embodiment, a WiFi module is set in the Raspberry Pi 3, and the Raspberry Pi 3 is wirelessly connected to the remote server via the WiFi module. When the temperature detects an abnormality or the image recognizes a fire risk, the system sends the alarm information to the remote server through the Raspberry Pi's Wi-Fi module. At this time, the manager can receive the alarm information through the monitoring interface of the mobile phone or PC, and can view the on-site video and make decisions in time.

[0089] In this embodiment, the camera 4 captures the image of the battery vehicle parking area in real time, and uses the OpenCV library to perform object detection and position tracking. Through image analysis, the state of the battery vehicle is monitored in real time to determine whether there are abnormalities such as flames and smoke.

[0090] In this embodiment, the infrared temperature detection module 5 is responsible for measuring the temperature near the battery vehicle. The Raspberry Pi 3 receives the data from the module. Once it detects that the temperature increase rate exceeds the original detected increase rate by 10% or more, or the temperature suddenly rises for more than 1 minute when it has stabilized during charging, the Raspberry Pi 3 will send an alarm signal to the buzzer 6 and the remote server. In order to ensure the accuracy of temperature detection, the system will detect the ambient temperature to eliminate the interference of the ambient temperature, and only retain the temperature changes in the core parts of the battery vehicle.

[0091] In this embodiment, the buzzer 6 is arranged inside the box 2, which is fixedly connected to the front side of the Raspberry Pi and is electrically connected to the Raspberry Pi 3. After the camera 4 and the Raspberry Pi 3 jointly perform target recognition and complete the temperature detection of the target, if the alarm requirement is met, that is, the temperature detected by the infrared temperature detection module 4 shows an abnormal change, the Raspberry Pi 3 controls the buzzer 6 through the GPIO to sound an alarm to remind the user to deal with it in time.

[0092] In this embodiment, the specific working method of the device is as follows: Figure 5 As shown, the method realizes the joint judgment of image recognition and temperature detection, and minimizes the false alarm rate by comprehensively analyzing the dual data of the camera and the infrared temperature detection module. The method includes the following steps:

[0093] S1, when the electric vehicle enters the station, capture the real-time video stream and temperature information and transmit it to the Raspberry Pi;

[0094] S2, preprocessing each frame of the real-time video stream through the Raspberry Pi to obtain video information;

[0095] S3, apply the YOLO5 model to detect objects in the current frame of the video information, identify and mark the battery vehicle area;

[0096] S4, using image recognition technology to process video information, observe whether there are fire characteristics in the battery vehicle area, if so, proceed to step S5; process temperature information, observe temperature changes, if the temperature changes are abnormal, proceed to step S5;

[0097] S5. Trigger the buzzer alarm and transmit the alarm to the remote server.

[0098] Among them, the image processing technology in S4 uses OpenCV technology; fire characteristics include smoke and flames.

[0099] The processing of the temperature information in S4 specifically includes analyzing the change of the temperature information over time and performing ambient temperature compensation to reduce the influence of the ambient temperature change on the measurement result.

[0100] The abnormal temperature change in S4 includes: the temperature increase rate exceeds the originally detected temperature increase rate by 10% or more, or the temperature suddenly rises for more than 1 minute after the electric vehicle is stably charged.

[0101] In this embodiment, the device is fixed to the wall of the charging garage through a base plate, and the camera and the infrared temperature detection module are installed and fixed on the wall at a height of about 1.6m from the ground.

[0102] In this embodiment, first, the device is connected to a 5V3A power supply through a USB power interface, and it is confirmed that the Raspberry Pi 3, the camera 4 and the infrared temperature detection module 5 are started and connected normally.

[0103] In this embodiment, a lighter is ignited within the camera range to simulate the battery vehicle being on fire.

[0104] In this embodiment, when the battery car enters the monitoring area, the camera 4 captures the real-time video stream to pre-process the image (grayscale, denoising). At this time, the camera 4 captures the real-time video stream to pre-process the image and detect the battery car target. The infrared temperature detection module 5 collects the battery car and ambient temperature data, and records it once per second. The system performs ambient temperature compensation on the battery car temperature data to remove the interference of external temperature fluctuations, ensuring that the monitored temperature changes are as close to the actual temperature state of the battery car as possible. To improve the efficiency and accuracy of subsequent processing; then the video data is transmitted to the Raspberry Pi 3 via the OpenCV interface, and the Raspberry Pi 3 uses the YOLO5 model to detect the object in the current frame, identify the position and boundary box of the battery car, and complete the target detection of the battery car.

[0105] Raspberry Pi 3 monitors the status of the battery vehicle in real time through image recognition to determine whether there are abnormalities such as flames and smoke;

[0106] At the same time, the infrared temperature detection module 5 collects the battery vehicle and the ambient temperature data, records it once per second, and converts the temperature data into an electrical signal and transmits it to the Raspberry Pi 3. The Raspberry Pi 3 receives the data from the module. Once it detects that the temperature increase rate exceeds the original detected increase rate by 10% or more, or the temperature suddenly rises for more than 1 minute when it has stabilized during charging, the Raspberry Pi 3 will send an alarm signal to the buzzer and the remote server. In order to ensure the accuracy of temperature detection, the system will detect the ambient temperature to eliminate the interference of the ambient temperature, and only retain the temperature changes in the core parts of the battery vehicle.

[0107] On the one hand, the Raspberry Pi 3 uses its built-in YOLO5 model to process video data to detect electric vehicle targets, and detects whether there is flame or smoke in the entire image area through image recognition; on the other hand, it detects temperature changes. Once it is detected that the temperature increase rate exceeds the original detected increase rate by 10% or more, or the temperature suddenly rises for more than 1 minute when it has stabilized during charging, an alarm signal is generated.

[0108] In this embodiment, what is simulated is the situation where flames appear in the image area, and an alarm signal is generated at this time.

[0109] When an alarm signal is generated, the Raspberry Pi 3 controls the buzzer 6 through GPIO, and the buzzer 6 sounds an alarm to remind the user to deal with it in time. At the same time, the alarm signal is sent to the remote server through the Wi-Fi module of the Raspberry Pi. At this time, the manager can receive the alarm information through the monitoring interface of the mobile phone or PC, and can view the on-site video and make decisions in time.

[0110] In this embodiment, after the device is started, the buzzer immediately sounds an alarm after about 2 seconds, and the remote server also immediately receives the alarm and stores the alarm time in the database.

[0111] In summary, in this embodiment, the device can effectively sound an alarm when a flame image is detected. In the case of a simulated battery car on fire, the system can quickly trigger a buzzer alarm, and the alarm information can be uploaded to the server within 5 seconds, ensuring a timely response and having the reliability of warning battery car fires in abnormal situations and alarming when a fire has already occurred.

[0112] Example 3

[0113] This embodiment applies a battery vehicle fire warning and alarm device, which can not only accurately detect temperature anomalies when the battery vehicle is charging, but also determine whether the battery vehicle has fire hazards, such as battery smoke, flames, etc. through image recognition technology, thereby greatly improving the accuracy of fire warning.

[0114] The device comprises a bottom plate 1 and a box body 2. The structure of the device is as follows Figure 1-Figure 4 As shown, Figure 1 This is the axonometric view of the device; Figure 2 is a cross-sectional view of the device; Figure 3 This is a schematic diagram of the camera and infrared temperature detection module inside the box; Figure 4 This is a side view of the power interface.

[0115] The bottom plate 1 of the device is provided with four centrally symmetrical threaded holes with a diameter of 8 mm, through which expansion bolts are passed to fix the bottom plate 1 to the wall. The box 2 is a hexahedron with a hollow structure, and the thickness of the six faces is 5 mm. Its front face is at an angle of 60 degrees to the horizontal line of the ground, which is convenient for the camera 4 to capture the panoramic image of the battery vehicle while avoiding dust accumulation or other external interference. The rear face is perpendicular to the ground and fixedly connected to the bottom plate 1, and the two side faces are right-angled trapezoids. Sufficient space is reserved in the box to place the Raspberry Pi fixed therein.

[0116] In this embodiment, a camera 4 and an infrared temperature detection module 5 are fixed on one side of the front end surface inside the box body 2, and observation openings of the camera 4 and the infrared temperature detection module 5 are set. A Raspberry Pi 3 is fixed on one side of the rear end surface 1 inside the box body 2, and the Raspberry Pi 3 is electrically connected to the output end of the camera 4 and the output end of the infrared temperature detection module 5.

[0117] In this embodiment, the Raspberry Pi 3 is the core module of the entire device, which is used for calculation and control. It includes a Raspberry Pi main control module and a Raspberry Pi shell. The Raspberry Pi shell is a hollow structure and is made of plastic. The Raspberry Pi 3 processes the video stream captured from the camera through a program written in Python, and analyzes the temperature data collected by the infrared temperature detection module. The Raspberry Pi main control module is connected to a power cord, and a power interface 301 is provided on the side of the box 2. In this embodiment, the power cord of the Raspberry Pi main control module passes through the power interface 301 and is connected to a 5V3A external power supply through a USB power interface.

[0118] In this embodiment, a WiFi module is set in the Raspberry Pi 3, and the Raspberry Pi 3 is wirelessly connected to the remote server via the WiFi module. When the temperature detects an abnormality or the image recognizes a fire risk, the system sends the alarm information to the remote server through the Raspberry Pi's Wi-Fi module. At this time, the manager can receive the alarm information through the monitoring interface of the mobile phone or PC, and can view the on-site video and make decisions in time.

[0119] In this embodiment, the camera 4 captures the image of the battery vehicle parking area in real time, and uses the OpenCV library to perform object detection and position tracking. Through image analysis, the state of the battery vehicle is monitored in real time to determine whether there are abnormalities such as flames and smoke.

[0120] In this embodiment, the infrared temperature detection module 5 is responsible for measuring the temperature near the battery vehicle. The Raspberry Pi 3 receives the data from the module. Once it detects that the temperature increase rate exceeds the original detected increase rate by 10% or more, or the temperature suddenly rises for more than 1 minute when it has stabilized during charging, the Raspberry Pi 3 will send an alarm signal to the buzzer 6 and the remote server. In order to ensure the accuracy of temperature detection, the system will detect the ambient temperature to eliminate the interference of the ambient temperature, and only retain the temperature changes in the core parts of the battery vehicle.

[0121] In this embodiment, the buzzer 6 is arranged inside the box 2, which is fixedly connected to the front side of the Raspberry Pi and is electrically connected to the Raspberry Pi 3. After the camera 4 and the Raspberry Pi 3 jointly perform target recognition and complete the temperature detection of the target, if the alarm requirement is met, that is, the temperature detected by the infrared temperature detection module 4 shows an abnormal change, the Raspberry Pi 3 controls the buzzer 6 through the GPIO to sound an alarm to remind the user to deal with it in time.

[0122] In this embodiment, the specific working method of the device is as follows: Figure 5 As shown, the method realizes the joint judgment of image recognition and temperature detection, and minimizes the false alarm rate by comprehensively analyzing the dual data of the camera and the infrared temperature detection module. The method includes the following steps:

[0123] S1, when the electric vehicle enters the station, capture the real-time video stream and temperature information and transmit it to the Raspberry Pi;

[0124] S2, preprocessing each frame of the real-time video stream through the Raspberry Pi to obtain video information;

[0125] S3, apply the YOLO5 model to detect objects in the current frame of the video information, identify and mark the battery vehicle area;

[0126] S4, using image recognition technology to process video information, observe whether there are fire characteristics in the battery vehicle area, if so, proceed to step S5; process temperature information, observe temperature changes, if the temperature changes are abnormal, proceed to step S5;

[0127] S5. Trigger the buzzer alarm and transmit the alarm to the remote server.

[0128] Among them, the image processing technology in S4 uses OpenCV technology; fire characteristics include smoke and flames.

[0129] The processing of the temperature information in S4 specifically includes analyzing the change of the temperature information over time and performing ambient temperature compensation to reduce the influence of the ambient temperature change on the measurement result.

[0130] The abnormal temperature change in S4 includes: the temperature increase rate exceeds the originally detected temperature increase rate by 10% or more, or the temperature suddenly rises for more than 1 minute after the electric vehicle is stably charged.

[0131] In this embodiment, the device is fixed to the wall of the charging garage through a base plate, and the camera and the infrared temperature detection module are installed and fixed on the wall at a height of about 1.6m from the ground.

[0132] In this embodiment, first, the device is connected to a 5V3A power supply through a USB power interface, and it is confirmed that the Raspberry Pi 3, the camera 4 and the infrared temperature detection module 5 are started and connected normally.

[0133] In this embodiment, the battery in an uncharged battery car is taken out to simulate the condition that the temperature of the battery car is stable, and a cup of boiling water is placed on the ground about 80 cm next to the battery car.

[0134] In this embodiment, when the battery car enters the monitoring area, the camera 4 captures the real-time video stream to pre-process the image (grayscale, denoising). At this time, the camera 4 captures the real-time video stream to pre-process the image and detect the battery car target. The infrared temperature detection module 5 collects the battery car and ambient temperature data, and records it once per second. The system performs ambient temperature compensation on the battery car temperature data to remove the interference of external temperature fluctuations, ensuring that the monitored temperature changes are as close to the actual temperature state of the battery car as possible. To improve the efficiency and accuracy of subsequent processing; then the video data is transmitted to the Raspberry Pi 3 via the OpenCV interface, and the Raspberry Pi 3 uses the YOLO5 model to detect the object in the current frame, identify the position and boundary box of the battery car, and complete the target detection of the battery car.

[0135] Raspberry Pi 3 monitors the status of the battery vehicle in real time through image recognition to determine whether there are abnormalities such as flames and smoke;

[0136] At the same time, the infrared temperature detection module 5 collects the battery vehicle and the ambient temperature data, records it once per second, and converts the temperature data into an electrical signal and transmits it to the Raspberry Pi 3. The Raspberry Pi 3 receives the data from the module. Once it detects that the temperature increase rate exceeds the original detected increase rate by 10% or more, or the temperature suddenly rises for more than 1 minute when it has stabilized during charging, the Raspberry Pi 3 will send an alarm signal to the buzzer and the remote server. In order to ensure the accuracy of temperature detection, the system will detect the ambient temperature to eliminate the interference of the ambient temperature, and only retain the temperature changes in the core parts of the battery vehicle.

[0137] On the one hand, the Raspberry Pi 3 uses its built-in YOLO5 model to process video data to detect electric vehicle targets, and detects whether there is flame or smoke in the entire image area through image recognition; on the other hand, it detects temperature changes. Once it is detected that the temperature increase rate exceeds the original detected increase rate by 10% or more, or the temperature suddenly rises for more than 1 minute when it has stabilized during charging, an alarm signal is generated.

[0138] In this embodiment, what is simulated is the situation where the temperature of other objects near the electric vehicle is abnormal, that is, the situation where other methods are prone to misjudgment, and no alarm signal is generated at this time.

[0139] When an alarm signal is generated, the Raspberry Pi 3 controls the buzzer 6 through GPIO, and the buzzer 6 sounds an alarm to remind the user to deal with it in time. At the same time, the alarm signal is sent to the remote server through the Wi-Fi module of the Raspberry Pi. At this time, the manager can receive the alarm information through the monitoring interface of the mobile phone or PC, and can view the on-site video and make decisions in time.

[0140] In this embodiment, since the temperature of the battery vehicle is stable, after the device is started, the buzzer does not sound an alarm within 10 minutes, and the remote server does not receive any alarm.

[0141] In summary, in this embodiment, the device can avoid alarming when the source of abnormal temperature change is not from the battery vehicle, thereby successfully reducing the false alarm rate and not triggering false alarms due to external high temperature interference.

[0142] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A battery vehicle fire warning and alarm device, comprising a bottom plate (1) and a box body (2), wherein the bottom plate (1) is provided with four centrally symmetrical threaded holes, and expansion bolts are passed through the threaded holes to fix the bottom plate (1) to a wall; the box body (2) is a hexahedron with a hollow structure, the front end face of the box body (2) is at an angle of 60 degrees to the horizontal line of the ground, the rear end face is perpendicular to the ground and fixedly connected to the bottom plate (1), and the two side faces are right-angled trapezoids; It is characterized in that A camera (4) and an infrared temperature detection module (5) are fixed on one side of the front end surface inside the box (2), and observation openings of the camera (4) and the infrared temperature detection module (5) are provided. A Raspberry Pi (3) is fixed on one side of the rear end surface (1) inside the box (2), and the Raspberry Pi (3) is electrically connected to the output end of the camera (4) and the output end of the infrared temperature detection module (5); A buzzer (6) is installed inside the box (2), and the buzzer (6) is fixedly connected to the front side of the Raspberry Pi and the buzzer (6) is electrically connected to the Raspberry Pi (3).

2. A battery vehicle fire early warning and alarm device according to claim 1, characterized in that: The diameter of the threaded hole on the bottom plate (1) is 8 mm.

3. A battery vehicle fire early warning and alarm device according to claim 1, characterized in that: The thickness of the six surfaces of the box body (2) is 5 mm.

4. A battery vehicle fire early warning and alarm device according to claim 1, characterized in that: The Raspberry Pi (3) comprises a Raspberry Pi main control module and a Raspberry Pi shell. The Raspberry Pi main control module is connected to a power line. The Raspberry Pi shell is hollow and made of plastic. A WiFi module is arranged in the Raspberry Pi (3). The Raspberry Pi (3) is wirelessly connected to a remote server via the WiFi module.

5. A battery vehicle fire early warning and alarm device according to claim 4, characterized in that: A power interface (301) is provided on the side of the box (2), and a power line of the Raspberry Pi main control module passes through the power interface (301) to be electrically connected to an external power source.

6. A working method of a battery vehicle fire warning and alarm device, characterized in that: A battery vehicle fire warning and alarm device as claimed in any one of claims 1 to 5 is operated by applying this method, and the method comprises the following steps: S1, when the electric vehicle enters the station, capture the real-time video stream and temperature information and transmit it to the Raspberry Pi; S2, preprocessing each frame of the real-time video stream through the Raspberry Pi to obtain video information; S3, apply the YOLO5 model to detect objects in the current frame of the video information, identify and mark the battery vehicle area; S4, using image recognition technology to process video information, observe whether there are fire characteristics in the battery vehicle area, if so, proceed to step S5; process temperature information, observe temperature changes, if the temperature changes are abnormal, proceed to step S5; S5. Trigger the buzzer alarm and transmit the alarm to the remote server.

7. A working method of a battery vehicle fire warning and alarm device according to claim 6, characterized in that: The image processing technology in S4 uses OpenCV technology.

8. The working method of the battery vehicle fire early warning and alarm device according to claim 6 is characterized in that: The fire characteristics in S4 include smoke and flames.

9. The working method of the battery vehicle fire warning and alarm device according to claim 6, characterized in that: The processing of the temperature information in S4 specifically includes analyzing the change of the temperature information over time on the basis of performing ambient temperature compensation, and recording the normal increase rate of the temperature.

10. A working method of a battery vehicle fire early warning and alarm device according to claim 9, characterized in that: The abnormal temperature change in S4 includes: the temperature increase rate exceeds the normal increase rate by 10% or more, or the temperature suddenly rises for more than 1 minute after the electric vehicle is stably charged.

Images