Unmanned fire monitoring device

Abstract

The utility model relates to an unmanned fire monitoring device in the technical field of unmanned fire monitoring. The unmanned fire monitoring device comprises a ground station and a sky station. Each mobile video acquisition terminal of the sky station comprises an unmanned aerial vehicle, and a pod, an edge computing device and a wireless image transmission device which are carried by the unmanned aerial vehicle. The edge computing device is provided with a data processing module which is originally arranged on a ground station and used for analyzing images collected by the pod to give out a fire analysis result, all the wireless image transmission devices establish a local area network, and the edge computing device covered by the local area network carries out air-ground data receiving and sending with the ground station so as to transmit the fire analysis result to the ground station. Local real-time processing of video streams is achieved, only detection results are transmitted back through a local area network of the wireless image transmission equipment assembly, the bandwidth requirement and delay are effectively reduced, and the problem that a traditional unmanned fire monitoring device cannot meet the timeliness of fire early warning is solved. And finally, the traditional unmanned fire monitoring device cannot be really applied to forest fire monitoring.

Description

Technical Field

[0001] This utility model relates to a fire monitoring device, and more particularly to an unmanned fire monitoring device. Background Technology

[0002] Forest fires are among the most destructive natural disasters globally, causing enormous ecological and economic losses every year. Traditional forest fire monitoring methods mainly include satellite remote sensing, fixed camera monitoring stations, and drone patrols. While fixed camera monitoring stations can provide continuous video surveillance, their deployment faces several limitations. First, due to the vast area of ​​forest cover, effective monitoring requires the construction of numerous base stations, resulting in high costs. Second, in complex terrain environments such as mountainous areas, the laying and maintenance of wired networks are extremely difficult, severely impacting system reliability. Furthermore, fixed cameras have limited monitoring range, with numerous blind spots, failing to achieve comprehensive coverage.

[0003] While satellite remote sensing and drone patrols offer advantages over fixed camera monitoring stations, such as mobility, flexibility, and wide coverage, satellite remote sensing images are generally not used for forest fire monitoring due to their long distances and insufficient accuracy. Drone monitoring solutions for drone patrols also have significant drawbacks. For example, a typical drone monitoring system using a DJI M300 paired with a SIYI H20, while capable of collecting high-quality video data, only functions as a video acquisition terminal. The raw video data must be transmitted back to the ground station for processing. This architecture leads to three main problems in air-to-ground data transmission: firstly, high-resolution video transmission demands high bandwidth, placing immense pressure on the wireless transmission system; secondly, the additional latency introduced by cloud processing results in end-to-end response times exceeding 5 seconds, failing to meet the timeliness requirements for early fire warnings; and thirdly, existing technologies also face significant challenges in data transmission. For example, common Wi-Fi networking methods have limited transmission distances, typically not exceeding 500 meters, and multi-hop relays introduce additional latency of over 200ms, making it difficult to meet the monitoring needs of large-scale forest areas. Meanwhile, 4G / 5G public network solutions are limited by base station coverage, making signal quality difficult to guarantee in forest areas, leading to reduced reliability of video stream transmission. Furthermore, traditional analog image transmission technology can only transmit video signals and cannot support the transmission of structured data. Existing digital image transmission technologies have not been optimized for edge computing scenarios; most pod devices on the market currently support high-definition video acquisition but lack integrated AI processing modules, resulting in a lack of local processing capabilities and an inability to meet real-time processing requirements.

[0004] In summary, existing forest fire monitoring technologies have significant shortcomings in terms of real-time performance, reliability, and deployment cost. Particularly when dealing with fire early warning in large-scale forest areas, these issues make it difficult for traditional solutions to meet the demands for rapid response and accurate identification. Consequently, traditional unmanned fire monitoring devices cannot be truly applied to forest fire monitoring. Therefore, there is an urgent need to develop a new type of unmanned fire monitoring device that effectively combines the mobility of drones with the efficiency of edge computing to solve the key bottlenecks in existing technologies. Summary of the Invention

[0005] To address the technical challenge that traditional unmanned fire monitoring devices cannot meet the timeliness requirements for early fire warnings, thus preventing their true application in forest fire monitoring, this invention provides an unmanned fire monitoring device.

[0006] This utility model adopts the following technical solution: an unmanned fire monitoring device includes a ground station and an air station. The air station includes at least one mobile video acquisition terminal, and each mobile video acquisition terminal includes a drone and a pod carried by the drone, an edge computing device, and a wireless image transmission device. The edge computing device is equipped with a data processing module originally deployed at the ground station for analyzing images acquired by the pod to provide fire analysis results. All wireless image transmission devices form a local area network (LAN), and the edge computing devices covered by the LAN communicate with the ground station via air-to-ground data transmission to transmit the fire analysis results to the ground station.

[0007] In a preferred embodiment of this utility model, the pod uses a pod that supports the RTSP protocol to communicate directly point-to-point with the corresponding edge computing device.

[0008] In a preferred embodiment of this utility model, the pod is model SIYIZT30 and the edge computing device is model RK3588.

[0009] In a preferred embodiment of this utility model, the wireless image transmission device is model HM30, and the computing device and the wireless image transmission device communicate via a USB interface.

[0010] In a preferred embodiment of the present invention, each mobile video acquisition terminal further includes a bracket, and the pod is suspended from the bottom of the corresponding drone by the bracket.

[0011] In a preferred embodiment of this utility model, the bracket is a hollow bracket.

[0012] In a preferred embodiment of the present invention, the edge and center of the bracket form a receiving groove for mounting an edge computing device.

[0013] In a preferred embodiment of this utility model, a plurality of bolt holes are provided on the bracket, and the wireless image transmission device is fixed on the bracket by bolts.

[0014] In a preferred embodiment of this invention, the wireless image transmission device is fixed to the drone shell with adhesive.

[0015] In a preferred embodiment of this utility model, the drone model is M300 RTK.

[0016] Compared with the prior art, the present invention has the following beneficial effects:

[0017] (1) By adding edge computing devices to the UAV, the data processing module originally deployed on the ground station is deployed on the sky station to realize local real-time processing of video streams. The local area network built by the wireless image transmission device only transmits back the detection results. By building a local area network between the ground station and multiple UAVs equipped with wireless image transmission devices, the overall signal coverage distance is widened. Even if one of the UAVs flies far away from the ground station and is separated from the signal transmission distance between the single wireless image transmission device and the ground station, the UAV can still access the local area network through the network connection of the wireless image transmission devices of other nearby UAVs, and finally realize the signal transmission between the UAV and the ground station. This effectively reduces the bandwidth requirements and latency, and solves the technical problems of high bandwidth pressure and high latency caused by the reliance on cloud processing in traditional UAV monitoring schemes. Therefore, it solves the technical problem that traditional unmanned fire monitoring devices cannot meet the timeliness of early fire warning, which ultimately leads to the inability of traditional unmanned fire monitoring devices to be truly applied to forest fire monitoring.

[0018] (2) The unmanned fire monitoring device of this utility model can form a local area network (LAN) through multiple wireless image transmission devices. Each edge computing device can enter the LAN through a directly connected wireless image transmission device and transmit and receive air-to-ground data with the ground station through the LAN. Since the LAN only transmits fire analysis results, it effectively reduces bandwidth requirements and latency. Therefore, a) this utility model can replace fixed monitoring points with mobile drone patrols, with a single drone covering an area of ​​10-20 square kilometers, significantly reducing infrastructure construction costs. The unmanned fire monitoring device of this utility model effectively solves the technical problems of high deployment costs and limited coverage of fixed monitoring systems; b) this utility model uses digital image transmission technology to establish a dedicated network, realizing reliable and low-latency transmission of detection results, solving the technical problem that existing networking schemes cannot support low-latency edge computing collaboration; c) this utility model directly connects an intelligent pod supporting the RTSP protocol with edge computing devices to build an integrated "acquisition-processing-transmission" architecture, realizing true airborne real-time analysis, solving the technical problem that traditional pod equipment only supports video acquisition and lacks local processing capabilities; d) this utility model reduces dependence on public network infrastructure by combining edge computing local processing and a dedicated wireless transmission network, ensuring stable monitoring performance in remote areas such as forests, and solving the reliability problem of existing systems in complex environments. Attached Figure Description

[0019] Figure 1 The system architecture diagram of the unmanned fire monitoring device provided in Example 1 is shown.

[0020] Figure 2 for Figure 1 A schematic diagram of the equipment connection method and networking scheme of the unmanned fire monitoring device.

[0021] Figure 3 for Figure 2 The network topology diagram.

[0022] Figure 4 for Figure 1 A 3D view of the mobile video acquisition terminal used in the unmanned fire monitoring device.

[0023] Figure 5 for Figure 4 A magnified view of a portion of the image. Detailed Implementation

[0024] The technical solutions of the present utility model will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present utility model, and not all embodiments. Based on the embodiments of the present utility model, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of the present utility model.

[0025] Example 1

[0026] Please see Figure 1 This is a system architecture diagram of the unmanned fire monitoring device provided in this embodiment of the utility model. The unmanned fire monitoring device includes a ground station and a sky station that transmits and receives data between the ground station and the air station.

[0027] Ground station such as Figure 1 The diagram shows a ground-based equipment room. The sky station includes at least one mobile video acquisition terminal. Each mobile video acquisition terminal includes a drone 1 and a pod 2, an edge computing device 3, and a wireless image transmission device 4 carried by the drone 1. The pod 2 is used to acquire images of the monitored object, such as a forest reserve, in real time. The edge computing device 3 is equipped with a data processing module (not shown) originally deployed at the ground station for analyzing the images to provide fire analysis results. The edge computing device 3, for example, model RK3588, can acquire front-end video for AI inference. All wireless image transmission devices 4 form a local area network (LAN). The edge computing device 3 covered by the LAN communicates with the ground station via air-to-ground data transmission to transmit the fire analysis results to the ground station.

[0028] Compared to traditional unmanned fire monitoring devices, the pod of the unmanned fire monitoring device communicates wirelessly with the ground station point-to-point to transmit images to the ground station, which is then analyzed by the data processing module deployed at the ground station to provide fire analysis results. Therefore: (1) It is necessary to transmit the complete high-definition video stream back, which occupies a large amount of wireless bandwidth (4K video requires more than 20Mbps), and the end-to-end delay exceeds 5 seconds; (2) Traditional Wi-Fi networking has a short distance, high multi-hop delay, and the 4G / 5G public network signal is unstable in forest areas.

[0029] The unmanned fire monitoring device of this utility model, by adding an edge computing device 3 to the UAV 1, deploys the data processing module originally deployed at the ground station to the sky station, realizing local real-time processing of video streams, and only transmitting back the detection results, effectively reducing bandwidth requirements and latency, and solving the technical problems of high bandwidth pressure and high latency caused by the reliance on cloud processing in traditional UAV monitoring solutions. The pod 2 can be selected to support the RTSP protocol (such as the SIYI H20 smart pod), so that it can communicate directly point-to-point with the corresponding edge computing device (3) without additional processing.

[0030] Secondly, each mobile video acquisition terminal of the unmanned fire monitoring device of this invention is equipped with a wireless image transmission device 4. The wireless image transmission device 4 is inherently used for wireless transmission, and multiple wireless image transmission devices 4 can form a local area network (LAN), such as the HM30 wireless digital image transmission device. Therefore, each edge computing device 3 can access the LAN through the directly connected wireless image transmission device 4, and conduct air-to-ground data transmission and reception with the ground station via the LAN. If traditional Wi-Fi networking is used, even if a drone 1 is used for inspection, it is impossible to achieve wide-area signal coverage because even if it flies over, the distance between the drone 1 and the ground station is too great, causing the signal to not be transmitted back in time, resulting in high latency due to multiple hops. This invention's unmanned fire monitoring device establishes a local area network (LAN) between a ground station and multiple drones (1) equipped with wireless image transmission devices (4), thus extending the overall signal coverage distance. Even if one drone (1) flies far away from the ground station, exceeding the signal transmission distance between a single wireless image transmission device (4) and the ground station, it can still access the LAN through the network connection of other nearby drones (1) with their wireless image transmission devices (4), ultimately achieving signal transmission between the drone (1) and the ground station. Since the LAN only transmits fire analysis results, it effectively reduces bandwidth requirements and latency. Therefore, (1) This utility model replaces fixed monitoring points with mobile patrols by drones, and a single drone can cover an area of ​​10-20 square kilometers, which significantly reduces the cost of infrastructure construction. The unmanned fire monitoring device of this utility model effectively solves the technical problems of high deployment cost and limited coverage of fixed monitoring systems; (2) This utility model uses digital image transmission technology to establish a dedicated network to achieve reliable and low-latency transmission of detection results, which solves the technical problem that the existing networking scheme cannot support low-latency edge computing collaboration; (3) This utility model directly connects the smart pod that supports the RTSP protocol with the edge computing device to build an integrated architecture of "acquisition-processing-transmission", which realizes true airborne real-time analysis and solves the technical problem that traditional pod equipment only supports video acquisition and lacks local processing capabilities; (4) This utility model reduces the dependence on public network infrastructure by combining edge computing local processing and dedicated wireless transmission network, ensuring that stable monitoring performance can still be maintained in remote areas such as forests, which solves the reliability problem of existing systems in complex environments.

[0031] Since the drone 1 needs to carry the pod 2, edge computing device 3, and wireless image transmission device 4, the selected model needs to have a certain carrying capacity. For example, the M300 RTK model can carry the pod 2, edge computing device 3, and wireless image transmission device 4. In summary, the hardware configuration of the unmanned fire monitoring device can be referenced in the table below.

[0032]

[0033] The IP address allocation is shown in the table below.

[0034] equipment IP address ZT30 pod 192.168.144.25 RK3588 192.168.144.125 ground station 192.168.144.100

[0035] The pod direct connection solution based on the RTSP protocol abandons the traditional vendor SDK docking method (such as DJI MSDK) and adopts private LAN communication technology to realize point-to-point direct connection between the pod (192.168.144.25) and the edge device (192.168.144.125). Platform: DJI M300 RTK drone. Installation location: bottom payload compartment. Equipment list: (1) Edge computing device 3: RK3588 development board; (2) pod 2: SIYI ZT30 pod; (3) Wireless image transmission device 4: HM30 digital image transmission.

[0036] Next, using the aforementioned hardware configuration table as an example, we will provide detailed evidence to demonstrate the feasibility of this utility model.

[0037] Please see Figure 2 and Figure 3 , Figure 2 The diagram shows the device connection methods and networking schemes corresponding to the hardware configuration table above; Figure 3 The corresponding network topology diagram is shown below. (1) Video acquisition and transmission: The SIYI ZT30 pod carried by the UAV acquires high-definition video streams (such as 4K / 30fps) in real time through the RTSP protocol and transmits the video streams to the airborne edge computing device (RK3588) through the local area network. The pod and the edge computing device are directly connected via Ethernet to form a low-latency local data transmission channel, avoiding the bandwidth pressure of video backhaul in the traditional solution. (2) Edge computing real-time processing: After receiving the video stream, the edge computing device completes fire detection through the following steps: a. Video decoding: Hardware acceleration decoding is performed using the built-in VPU of RK3588 to convert the video stream into a processable frame sequence; b. AI inference: The lightweight fire detection model (such as the RKNN model converted by YOLOv11n) deployed on the NPU performs real-time analysis on each frame image, detects features such as flames and smoke, and outputs structured data (including fire location, confidence level and bounding box coordinates). (3) Wireless data transmission: a. The processed data is transmitted to the ground station through the HM30 digital image transmission module; b. Local area network communication: the pod 2 and the edge computing device 3 form a private local area network (such as 192.168.144.0 / 24) with the ground station through the wireless image transmission device 4 to ensure low-latency communication; c. Data backhaul: only the structured detection results (such as fire location and confidence level) and detection frames are transmitted, rather than the original video stream, which greatly reduces the wireless transmission bandwidth requirements; d. Reliability guarantee: the ACK confirmation mechanism is adopted to ensure that key data is not lost, and the data is automatically cached when the signal is poor and retransmitted after the signal is restored.

[0038] Therefore, the workflow of the unmanned fire monitoring device of this utility model can be divided into four stages:

[0039] (1) Device self-organizing network stage: After the UAV 1 takes off, it automatically completes the connection between the pod 2 and the edge computing device 3 and the ground station via the wireless image transmission device 4 to achieve local area network connection.

[0040] (2) Video acquisition stage: Pod 2 acquires high-definition video streams in real time.

[0041] (3) Edge computing stage: Edge computing device 3 completes the entire process of video decoding, fire detection and result encoding.

[0042] (4) Data backhaul stage: Wireless image transmission device 4 only transmits structured detection results and detection frames to achieve efficient transmission.

[0043] The data processing module deployed on edge computing device 3 can adopt the data processing module deployed on ground stations by traditional unmanned fire monitoring devices. The model deployment flowchart of the data processing module is as follows: smoke and fire detection model training (such as a trained .pth model), ONNX model conversion (i.e., ONNX conversion), RKNN model conversion (i.e., RKNN quantization), RKNN model inference code, and RKNN model deployment (i.e., RK3588 NPU deployment).

[0044] The key parameters are shown in the table below:

[0045] step Tools / Parameters Output Model conversion torch.onnx.export() model.onnx Quantization calibration rknn-toolkit2 (INT8 quantization) model.rknn Inference engine initialization rknn.init_runtime() NPU acceleration instance

[0046] The video stream processing of the data processing module is as follows: (1) Video stream access stage: the video stream is obtained from the SIYI ZT30 pod through the RTSP protocol; (2) Video decoding stage: hardware decoding is performed using the built-in VPU of RK3588; (3) Target detection stage: NPU is used to accelerate inference.

[0047] Please see Figure 4 This is a 3D view of a mobile video acquisition terminal, each of which may also include a support bracket 5. To improve the drone's battery life, the support bracket 5 can be configured as a hollow bracket, such as... Figure 5 As shown, the edge and center of the bracket 5 can form receiving slots for mounting the edge computing device 3. The hollowed-out bracket can reduce the energy consumption of the drone 1. The receiving slots facilitate the improvement of the installation stability of the edge computing device 3, ensuring the drone 1's endurance while protecting the device's safety.

[0048] A multi-angle rotating pod 2 is mounted at the bottom of the support frame 5. This pod 2 is a drone pod, and its model is SIYIZT30. Video acquisition and transmission: The SIYI ZT30 pod acquires high-definition video streams, such as 4K / 30fps, in real time via the RTSP protocol and transmits the video stream to the onboard edge computing device RK3588 via a local area network. Pod 2 has a first real-time streaming media protocol interface, and the edge computing device 3 has a second real-time streaming media protocol interface. The two ends of the first data transmission line are connected to the first and second real-time streaming media protocol interfaces, respectively. Traditional Wi-Fi networks suffer from short range, high latency due to multiple hops, and unstable 4G / 5G public network signals in forest areas. Existing pods, such as the SIYIH20, are only used as video acquisition terminals. This solution, however, uses pods supporting the RTSP protocol and edge computing devices directly connected via Ethernet to build an integrated "acquisition-processing-transmission" architecture. This enables true airborne real-time analysis, forming a low-latency local data transmission channel, avoiding the bandwidth pressure of video backhaul in traditional solutions, improving response speed, and solving the technical problem that traditional unmanned fire monitoring devices cannot meet the timeliness requirements for early fire warnings, ultimately preventing their application in forest fire monitoring. An optimized digital image transmission technology is used to establish a dedicated network, achieving reliable, low-latency backhaul of detection results; significantly reducing system deployment and maintenance costs, and solving the technical problem that existing networking solutions cannot support low-latency edge computing collaboration.

[0049] In addition, a wireless image transmission device 4, equipped with wireless LAN setup and data transmission / reception capabilities, is installed on one side of the drone 1; the model of the wireless image transmission device 4 is HM30. The wireless image transmission device 4 is installed on one side of the drone 1 and electrically connected to the edge computing device 3 via a second data transmission line. The wireless image transmission device 4 is used to acquire fire analysis results via the second data transmission line. The wireless image transmission device 4 can establish a LAN for convenient air-to-ground data transmission and reception. Other image transmission devices adapted to the pod can also be used instead. The pod 2, edge computing device 3, and ground base station can form a private LAN such as 192.168.144.0 / 24 to ensure low-latency communication. Only structured detection results such as fire location, confidence level, and detection frames are transmitted, rather than the raw video stream, significantly reducing the wireless transmission bandwidth requirements. Reliability assurance: An ACK confirmation mechanism is adopted to ensure that critical data is not lost, and data is automatically cached when the signal is poor and retransmitted after the signal is restored. The edge computing device 3 achieves local processing of the video stream, with end-to-end latency controlled within 200ms, improving response speed by 25 times compared to traditional cloud processing solutions. By only transmitting structured detection results and detection frame streams, the bandwidth requirement is significantly reduced compared to transmitting the original video stream, effectively improving the real-time performance of fire detection. Existing solutions require the complete transmission of high-definition video streams, consuming a large amount of wireless bandwidth (4K video requires over 20Mbps), and the end-to-end latency exceeds 5 seconds. This solution, however, deploys edge computing devices on the drone to achieve local real-time processing of the video stream, transmitting only the detection results, effectively reducing bandwidth requirements and latency. This solves the technical problems of high bandwidth pressure and high latency caused by traditional drone monitoring solutions relying on cloud processing. A private LAN is used to directly connect the pod 2 and the edge computing device 3, avoiding transmission interruptions caused by unstable public network signals. A dedicated wireless image transmission device 4 provides stable transmission bandwidth in forest environments, significantly improving coverage compared to the 500-meter limitation of traditional Wi-Fi solutions, and solving the reliability issues of existing systems in complex environments. The combination of local edge computing processing and a dedicated wireless transmission network reduces dependence on public network infrastructure, ensuring stable monitoring performance in remote areas such as forests. This enhances reliability in complex environments. The bracket has several bolt holes. The wireless image transmission device 4 is fixed to the bracket 5 by bolts, or the wireless image transmission device 4 is fixed to the shell of the drone 1 by glue.

[0050] Edge computing device 3 is mounted on top of bracket 5 and is electrically connected to pod 2 via a first data transmission line. Edge computing device 3 is used to acquire, process, and analyze image data collected by pod 2 via the first data transmission line, providing fire analysis results. Edge computing device 3 uses an RK3588 development board. Edge computing device 3 also has a first USB 3.0 interface, and wireless image transmission device 4 has a second USB 3.0 interface. Data transmission between the two interfaces can be achieved by connecting the data transmission line.

[0051] Compared with the prior art, this utility model has the following significant advantages:

[0052] (1) Significantly reduce system deployment and maintenance costs

[0053] Through the architectural design of this utility model, a single UAV 1 can cover an area of ​​10-20 square kilometers, reducing infrastructure construction costs compared to traditional fixed camera solutions. By employing an edge computing device 3 to deploy a data processing module originally deployed at the ground station to analyze the images and provide fire analysis results, local processing is achieved, eliminating the need to rely on cloud servers and saving cloud computing resources and long-term operation and maintenance costs.

[0054] (2) Effectively improve the real-time performance of fire detection

[0055] Airborne edge computing enables local video stream processing with end-to-end latency controlled within 200ms, achieving a 25-fold improvement in response speed compared to traditional cloud processing solutions (latency > 5 seconds). Only structured detection results and detection frame streams are transmitted back, significantly reducing bandwidth requirements compared to transmitting the original video stream.

[0056] (3) Enhance reliability in complex environments

[0057] A private LAN is used to directly connect the pod 2 and the edge computing device 3, avoiding transmission interruptions caused by unstable public network signals. A dedicated wireless image transmission device 4 (HM30) provides stable transmission bandwidth in forest environments, significantly improving coverage compared to traditional Wi-Fi solutions (limited to 500 meters).

[0058] (4) Achieve all-weather autonomous monitoring

[0059] Drone 1 can be preset with inspection routes, has a single flight time of more than 2 hours, and can achieve 24-hour uninterrupted monitoring with automatic return to recharge. Edge computing device 3 supports wide temperature range of 0℃~60℃, adapting to the diurnal temperature variation in forest areas.

[0060] In summary, this invention employs a pod supporting the RTSP protocol. Therefore, it can achieve point-to-point direct communication between the pod and the edge computing device (e.g., the SIYIZT30 pod uses its built-in RTSP protocol to achieve point-to-point direct connection between the pod (192.168.144.25) and the edge device (192.168.144.125), supporting rapid adaptation and combination with industrial drone platforms such as DJI M300). This invention also uses an edge computing device with edge computing capabilities installed on the drone (e.g., the RK3588 edge computing device completes the entire process of video decoding (VPU) → AI inference (NPU) → result encoding (JSON), and uses lightweight data encapsulation technology to compress the detection results). This invention also employs a wireless image transmission device capable of forming a local area network (LAN) installed on the drone (such as the HM30 model wireless image transmission device) to achieve real-time collaborative work among the pod, edge computing device, and wireless image transmission device. The pod of the sky station collects images of the forest reserve and transmits them to the edge computing device for image processing via point-to-point direct communication using the RTSP protocol. The image processing results are then transmitted to the ground station via the wireless image transmission device through the LAN. This effectively combines the mobility of the drone with the efficiency of edge computing, achieving rapid response and accurate identification of forest fires. This solves the technical problem that traditional unmanned fire monitoring devices cannot meet the timeliness requirements for early fire warnings, ultimately preventing their application in forest fire monitoring.

[0061] In the description of this utility model, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "front," and "rear," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this utility model and simplifying the description, and do not indicate or imply that the device or component referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this utility model. In this utility model, it should also be noted that the terms "installation" and "connection" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, integral connections, mechanical connections, or indirect connections through an intermediate medium. The specific meaning of the terms in this utility model can be understood according to the specific circumstances.

[0062] The embodiments described above are merely illustrative of several implementations of this utility model, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the utility model patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this utility model, and these all fall within the protection scope of this utility model. Therefore, the protection scope of this utility model patent should be determined by the appended claims.

Claims

1. An unmanned fire monitoring device, comprising: Ground station; Sky station, which includes at least one mobile video acquisition terminal, each mobile video acquisition terminal including a drone (1) and a pod (2) carried by the drone (1) for real-time image acquisition. The feature is that each mobile video acquisition terminal also includes an edge computing device (3) and a wireless image transmission device (4) carried by a drone (1). The edge computing device (3) is equipped with a data processing module originally deployed at the ground station for analyzing the images to provide fire analysis results. All wireless image transmission devices (4) form a local area network. The edge computing device (3) covered by the local area network communicates with the ground station via air-to-ground data transmission and reception to transmit the fire analysis results to the ground station.

2. The unmanned fire monitoring device of claim 1, wherein, The pod (2) uses a pod that supports the RTSP protocol to communicate directly point-to-point with the corresponding edge computing device (3).

3. The unmanned fire monitoring device of claim 2, wherein, The pod (2) is model SIYIZT30, and the edge computing device (3) is model RK3588.

4. The unmanned fire monitoring device of claim 1, wherein, The wireless image transmission device (4) is model HM30. The computing device (3) and the wireless image transmission device (4) communicate via a USB interface.

5. The unmanned fire monitoring device of claim 1, wherein, Each mobile video acquisition terminal also includes a bracket (5), and the pod (2) is suspended from the bottom of the corresponding drone (1) via the bracket (5).

6. The unmanned fire monitoring device of claim 5, wherein, The bracket (5) is a hollow bracket.

7. The unmanned fire monitoring device of claim 5, wherein, The edges and center of the bracket (5) form a receiving groove for mounting the edge computing device (3).

8. The unmanned fire monitoring device of claim 5, wherein, The bracket (5) has several bolt holes, and the wireless image transmission device (4) is fixed to the bracket (5) by bolts.

9. The unmanned fire monitoring device of claim 1, wherein, The wireless image transmission device (4) is fixed to the shell of the drone (1) with glue.

10. The unmanned fire monitoring device of claim 1, wherein, The model of the drone (1) is M300RTK.

Images