Method, device and equipment for detecting mobile fire in tunnel and medium

By obtaining real-time monitoring video data in highway tunnels and using pre-trained ignition vehicle identification models, the problem that traditional fire detectors cannot detect high-speed mobile vehicles fires in a timely manner is solved, and timely and accurate detection of fires in the tunnel is achieved, improving the Traffic safety.

CN118968700APending Publication Date: 2024-11-15SHANGHAI TENSUN TRANSMART
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411086384.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-08
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

Traditional linear temperature sensing products and point infrared flame detectors are not ideal for use in highway tunnels, and they cannot detect high-speed moving vehicle fires in time, resulting in the inability to trigger the fire alarm system.

Method used

By obtaining real-time monitoring video data in the tunnel and inputting it into the pre-trained fire vehicle identification model, identifying and determining whether there is a fire vehicle. If it exists, determine real-time fire detection information to achieve timely and timely and Accurate detection.

Benefits of technology

The ability to detect ignition vehicles in a timely and accurately detect them in a high-speed tunnel improves traffic safety in the tunnel and avoids the problem that traditional detectors cannot detect due to the rapid movement of vehicles.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118968700A_ABST
    Figure CN118968700A_ABST
Patent Text Reader

Abstract

The embodiment of the invention discloses a mobile fire detection method, device and equipment in a tunnel and a medium. The method comprises the following steps: acquiring real-time monitoring video data in a tunnel; inputting the real-time monitoring video data into a pre-trained on-fire vehicle identification model to obtain a real-time identification result; and if the real-time identification result reflects that the vehicle on fire exists in the tunnel, determining real-time fire detection information in the tunnel according to the target monitoring video data including the vehicle on fire. According to the technical scheme, the on-fire vehicle identification is carried out on the real-time monitoring video data in the tunnel through the on-fire vehicle identification model, the on-fire vehicle can be found in time, the on-fire vehicle in moving can be accurately detected, and support is provided for traffic safety in the tunnel.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

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

[0002] At present, the main products used for fire detection in tunnels include linear temperature sensing products and point infrared flame detectors. However, the application effect of these traditional fire detectors in highway tunnels is not ideal.

[0003] Linear temperature sensing products can only trigger the alarm threshold after the fire reaches a certain scale and generates heat that reaches the top of the tunnel. This means that in the early stages of a fire, linear temperature sensing products may not be able to detect the fire in time because the heat has not yet fully diffused. In addition, due to the limitations of the internal structure of the tunnel, the installation location of linear temperature sensing products may be subject to certain restrictions, thus affecting their detection effect.

[0004] Point-type infrared flame detectors are installed on the side walls of tunnels along the direction of travel. Due to the high speed of vehicles in highway tunnels, the time a vehicle stays within the detection range of each point-type infrared flame detector is short. Once a high-speed moving vehicle catches fire, it cannot be detected in time by the point-type infrared flame detector, and the fire alarm system cannot be triggered. Summary of the invention

[0005] The present invention provides a method, device, equipment and medium for detecting mobile fire in a tunnel, which can timely and accurately realize mobile fire detection in a tunnel when a fire occurs in a vehicle in a high-speed tunnel.

[0006] According to one aspect of the present invention, a method for detecting mobile fire in a tunnel is provided, the method comprising:

[0007] Obtain real-time surveillance video data in the tunnel;

[0008] Inputting the real-time monitoring video data into a pre-trained fire vehicle recognition model to obtain a real-time recognition result;

[0009] If the real-time identification result reflects that there is a vehicle on fire in the tunnel, real-time fire detection information in the tunnel is determined based on the target monitoring video data including the vehicle on fire.

[0010] According to another aspect of the present invention, there is provided a mobile fire detection device in a tunnel, comprising:

[0011] A real-time monitoring video data acquisition module is used to acquire real-time monitoring video data in the tunnel;

[0012] A real-time recognition result determination module is used to input the real-time monitoring video data into a pre-trained fire vehicle recognition model to obtain a real-time recognition result;

[0013] The real-time fire detection information determination module is used to determine the real-time fire detection information in the tunnel based on the target monitoring video data including the burning vehicle if the real-time identification result reflects that there is a burning vehicle in the tunnel.

[0014] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0015] at least one processor; and

[0016] a memory communicatively connected to the at least one processor; wherein,

[0017] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method for detecting mobile fire in a tunnel described in any embodiment of the present invention.

[0018] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method for detecting mobile fire in a tunnel according to any embodiment of the present invention when executed.

[0019] The technical solution of the embodiment of the present application includes: obtaining real-time monitoring video data in a tunnel; inputting the real-time monitoring video data into a pre-trained fire vehicle identification model to obtain a real-time identification result; if the real-time identification result reflects that there is a fire vehicle in the tunnel, then determining the real-time fire detection information in the tunnel based on the target monitoring video data including the fire vehicle. This technical solution uses the fire vehicle identification model to identify the fire vehicle in the real-time monitoring video data in the tunnel, so that the fire vehicle can be discovered in time, and the fire vehicle in motion can be accurately detected, providing support for traffic safety in the tunnel.

[0020] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0022] Figure 1 This is a flow chart of a method for detecting mobile fire in a tunnel provided according to Embodiment 1 of the present application;

[0023] Figure 2 This is a flow chart of a method for detecting mobile fire in a tunnel provided according to Embodiment 2 of the present application;

[0024] Figure 3 This is a specific example diagram of mobile fire detection in a tunnel provided according to the second embodiment of the present application;

[0025] Figure 4 It is a structural schematic diagram of a mobile fire detection device in a tunnel provided according to the third embodiment of the present application;

[0026] Figure 5 It is a structural schematic diagram of an electronic device for implementing a method for detecting mobile fire in a tunnel according to an embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only embodiments of a part of the present invention, not all 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.

[0028] It should be noted that the terms "first", "second", "target", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0029] Embodiment 1

[0030] Figure 1A flowchart of a mobile fire detection method in a tunnel is provided for the first embodiment of the present application. The embodiment of the present application can be applied to the case of real-time detection of fire in a tunnel. The method can be executed by a mobile fire detection device in a tunnel. The mobile fire detection device in a tunnel can be implemented in the form of hardware and / or software. The mobile fire detection device in a tunnel can be configured in an electronic device with data processing capabilities. Figure 1 As shown, the method includes:

[0031] S110, obtaining real-time monitoring video data in the tunnel.

[0032] The tunnel may be a road with certain closed properties for vehicles, pedestrians, etc. to pass through. The tunnel in the embodiment of the present application includes but is not limited to: underground excavated passages, passages in mountains, and submarine tunnels, etc. The real-time monitoring video data may be a monitoring video captured in real time by monitoring equipment in the tunnel, and the real-time monitoring video data may include image information such as vehicles and tunnel backgrounds.

[0033] Specifically, in a feasible solution, if a signal indicating that the fire detection function has been triggered is received, real-time monitoring video data captured by the monitoring equipment is continuously acquired.

[0034] In another feasible solution, the monitoring device in the tunnel itself has a storage medium, such as a hard disk, and the monitoring device will cyclically store the monitoring video in the storage medium. After the monitoring device updates the monitoring video, the real-time monitoring video data can be obtained from the storage medium.

[0035] In another feasible solution, the monitoring equipment in the tunnel uploads the captured monitoring video to the cloud, and real-time monitoring video data can be obtained from the cloud.

[0036] S120, inputting the real-time monitoring video data into a pre-trained fire vehicle recognition model to obtain a real-time recognition result.

[0037] Among them, the burning vehicle detection model is used to identify the burning vehicle in the monitoring video. The real-time recognition result reflects whether there is a burning vehicle in the real-time monitoring video data. If there is, the real-time recognition result also includes the specific information of the burning vehicle. Exemplarily, if there is no burning vehicle in the tunnel, the real-time recognition result can be: no burning vehicle is detected; if there is a burning vehicle in the tunnel, one real-time recognition result can be: there is a burning vehicle in the tunnel, and the target monitoring video data of the burning vehicle; another real-time recognition result can be: there are three burning vehicles in the tunnel, and the target monitoring video data of each burning vehicle.

[0038] Specifically, after obtaining the real-time monitoring video data, it is input into the pre-trained on-fire vehicle recognition model to obtain the real-time recognition result output by the on-fire vehicle recognition model. Since the monitoring video data is real-time, that is, the real-time monitoring video data can be continuously acquired, it is necessary to continuously input the real-time monitoring video data into the pre-trained on-fire vehicle recognition model, and the on-fire vehicle recognition model continuously outputs the real-time recognition result.

[0039] S130, if the real-time recognition result reflects that there is a vehicle on fire in the tunnel, then determine the real-time fire detection information in the tunnel based on the target monitoring video data including the vehicle on fire.

[0040] The real-time fire detection information reflects the specific information of the burning vehicle, including but not limited to: the location of the burning vehicle, the identification information of the burning vehicle, and the fire situation.

[0041] Specifically, if the real-time recognition result reflects that there is a vehicle on fire in the tunnel, the target monitoring video data output by the model is obtained, the identification of the vehicle on fire (such as license plate number, vehicle brand, etc.) is determined according to the target monitoring video data, and the real-time driving trajectory of the vehicle on fire is determined.

[0042] In a tunnel, vehicles are usually in a moving state (especially the tunnel section of a highway). If a fire occurs in a moving vehicle, traditional fire sensors are limited by their own properties and are not easy to detect fires on moving vehicles. The technical solution of the embodiment of the present application detects the burning vehicle through a burning vehicle detection model. It is a detection for video images. Even if the burning vehicle is in motion, it can be accurately detected, and the real-time monitoring video data can be captured by the monitoring equipment in the tunnel. There is no need to install other equipment in the tunnel, which has the advantage of low cost.

[0043] The technical solution of the embodiment of the present application includes: obtaining real-time monitoring video data in a tunnel; inputting the real-time monitoring video data into a pre-trained fire vehicle identification model to obtain a real-time identification result; if the real-time identification result reflects that there is a fire vehicle in the tunnel, then determining the real-time fire detection information in the tunnel based on the target monitoring video data including the fire vehicle. This technical solution uses the fire vehicle identification model to identify the fire vehicle in the real-time monitoring video data in the tunnel, so that the fire vehicle can be discovered in time, and the fire vehicle in motion can be accurately detected, providing support for traffic safety in the tunnel.

[0044] Embodiment 2

[0045] Figure 2 This is a flow chart of a method for detecting mobile fire in a tunnel provided in Example 2 of the present application. The embodiment of the present application is optimized based on the above embodiment.

[0046] like Figure 2 As shown, the method of the embodiment of the present application specifically includes the following steps:

[0047] S210, obtaining real-time monitoring video data in the tunnel.

[0048] Exemplarily, real-time monitoring video data is obtained, and image information in the video is processed by noise reduction, contrast enhancement, etc. to improve image quality and provide a clearer data basis for subsequent feature extraction and recognition.

[0049] In the embodiment of the present application, optionally, the real-time monitoring video data is processed by adaptive lighting adjustment technology and anti-shadow interference algorithm, which can automatically adjust parameters to adapt to different environmental conditions, ensure the stability and reliability of fire identification, and solve problems such as complex internal environment, uneven lighting, shadow interference, etc. of the tunnel.

[0050] S220, inputting the real-time monitoring video data into a pre-trained fire vehicle recognition model to obtain a real-time recognition result.

[0051] In an embodiment of the present application, optionally, the real-time monitoring video data is input into a pre-trained fire vehicle recognition model to obtain a real-time recognition result, including: inputting the real-time monitoring video data into a pre-trained fire vehicle recognition model, and extracting vehicle features and / or flame features in the real-time monitoring video data through the fire vehicle recognition model; if it is determined that the image area where the flame features are extracted and the image area where the vehicle features are extracted are both image areas corresponding to the same vehicle, then the real-time recognition result obtained is that there is a fire vehicle; otherwise, the real-time recognition result obtained is that there is no fire vehicle.

[0052] The vehicle feature may be a feature of a moving vehicle, and the flame feature may be a feature of a moving flame.

[0053] Specifically, the real-time monitoring video data is input into a pre-trained burning vehicle recognition model. The burning vehicle recognition model can be used to extract vehicle features and / or flame features in the real-time monitoring video data. If both vehicle features and flame features are extracted, it can be further verified whether the vehicle features and flame features are extracted from the same vehicle, thereby determining whether there is a burning vehicle. In another feasible solution, the fused vehicle features and flame features can be extracted by the burning vehicle recognition model. If the extraction is successful, it is determined that there is a burning vehicle. It should be noted that the specific process of identifying a burning vehicle by the burning vehicle detection model is not limited in the embodiment of the present application.

[0054] Exemplarily, a sliding window may be used to extract features from a candidate area (the candidate area may be an area near a lane line in real-time monitoring video data), and then a nonlinear algorithm may be used to classify the features to obtain vehicle features and / or flame features.

[0055] In an embodiment of the present application, optionally, the training process of the fire vehicle recognition model includes: using video data with a tunnel as the background and including vehicles as training samples; determining labels of the training samples, and training the fire vehicle detection model to be trained based on the training samples to obtain a trained fire vehicle detection model.

[0056] The training sample is a tunnel video, which may include a moving vehicle, a moving vehicle on fire, etc. The training sample can be obtained by shooting in a shooting state, or can be a generated video.

[0057] In an embodiment of the present application, optionally, the process of determining the training samples includes: acquiring a small target image; the small target includes a moving flame, an unburned vehicle, and a burned vehicle; fusing the small target image into a tunnel background video to obtain a candidate sample; and processing the candidate sample through an image data enhancement algorithm to obtain a training sample.

[0058] Specifically, after obtaining the small target image, it is fused into the tunnel background video to obtain candidate samples, and then for the candidate samples including the burning vehicle, the video including the marked frame of the burning vehicle is used as a label to obtain a training sample; in this process, the video can be enhanced at any stage to obtain more training samples under different light intensities and different environmental conditions.

[0059] For example, we obtain test videos, mark samples, and include more relevant small target data in the data set; we use image data enhancement algorithms to enrich the sample data set to obtain more small target images, and fuse the images of small targets with more background information to generate more data sets, and calibrate the corresponding targets and position information through manually annotated pictures, and continuously optimize the algorithm model. Through technical means such as ResNet and CSPNet, the feature extraction and expression capabilities of the constructed network are enhanced, and the convergence ability is stronger; the bidirectional FPN technology is used to make the feature fusion effect more obvious. On the feature map finally output by the network, each anchor point is responsible for the prediction of the specified anchor point box.

[0060] This scheme is set up in this way, so that the types and number of training samples are greatly increased, the convergence degree of the fire vehicle detection model is higher, and the identification of the fire vehicle is more accurate. Since the training samples can be obtained by fusing the small target image and the tunnel background video, there is no need to collect a large number of samples through monitoring equipment, which reduces the difficulty and workload of collecting samples and reduces the cost of sample collection.

[0061] S230, if the real-time recognition result reflects that there is a vehicle on fire in the tunnel, the actual position corresponding to the marking box is determined based on the correspondence between the pre-calibrated spatial position in the tunnel and the pixel points of the video frame of the monitoring video.

[0062] In the embodiment of the present application, if the real-time recognition result reflects that there is a vehicle on fire in the tunnel, the real-time recognition result includes target monitoring video data, and each video frame of the target monitoring video data includes a marking box; the marking box includes the vehicle on fire.

[0063] Specifically, if the real-time recognition result reflects that there is no vehicle on fire in the tunnel, the real-time monitoring video data is continuously detected through the fire vehicle detection model; if the real-time recognition result reflects that there is a vehicle on fire in the tunnel, the target monitoring video data output by the fire vehicle detection model is obtained, and then based on the correspondence between the pre-calibrated spatial position in the tunnel and the pixel points of the video frame of the monitoring video, the actual position corresponding to each marking box in the target monitoring video data is determined, and the actual positions where the fire vehicle has traveled are obtained.

[0064] In the embodiment of the present application, since the position of the monitoring device is usually fixed, that is, the actual spatial position corresponding to each pixel in the image taken by the monitoring device is almost unchanged, the correspondence between the spatial position in the tunnel and the pixel points of the video frame of the monitoring video can be pre-calibrated so that it can be used directly after obtaining the target monitoring video data.

[0065] Exemplarily, after the monitoring product is installed and fixed, the background space is calibrated through external calibration points, and three-dimensional relational coordinates are established to form a coordinate mapping model; the model reflects the correspondence between the spatial position in the tunnel and the pixel points of the video frame of the monitoring video.

[0066] S240: Determine the real-time position information of the vehicle on fire according to the actual position corresponding to each marking box.

[0067] Specifically, after obtaining the actual positions of the vehicles on fire, real-time position information of the vehicles on fire may be generated according to the actual positions, and the real-time position information may include a real-time trajectory.

[0068] In an embodiment of the present application, optionally, the real-time position information of the burning vehicle is determined according to the actual position corresponding to each marking box, including: determining the acquisition time of the video frame corresponding to the burning vehicle as the timestamp of the burning vehicle at each actual position; and generating the real-time driving trajectory of the burning vehicle according to the actual position and timestamp of the burning vehicle.

[0069] Specifically, the actual position is determined based on the marking box. The shooting time (i.e., acquisition time) of the video frame where the marking box is located is the timestamp of the actual position. After obtaining the timestamp of each actual position, the real-time driving trajectory of the burning vehicle can be generated.

[0070] This scheme is set up in such a way that the locations where the burning vehicle has traveled and the latest location can be reflected in the real-time driving trajectory, which makes it convenient for the rescue party to carry out rescue according to the real-time driving trajectory during subsequent fire rescue.

[0071] In an embodiment of the present application, optionally, the method also includes: determining the driving direction of the burning vehicle based on the lane line information where the burning vehicle is located; if the real-time driving trajectory does not match the actual driving direction of the burning vehicle, determining the abnormal video frame where the abnormal position point in the real-time driving trajectory is located; if it is determined that the burning vehicle in the abnormal video frame is inconsistent with the burning vehicles corresponding to the remaining video frames in the target video frame, deleting the abnormal position point and regenerating the real-time driving trajectory of the burning vehicle.

[0072] In the embodiment of the present application, in order to avoid the situation where the detection model of the burning vehicle has recognition errors, the real-time driving trajectory can be compared with the driving direction of the burning vehicle. If they do not match, manual intervention can be performed, or an abnormal position point can be determined in the real-time driving trajectory, and then the abnormal video frame where the abnormal position point is located can be determined; if it is determined that the burning vehicle in the abnormal video frame is inconsistent with the burning vehicle corresponding to the remaining video frames in the target video frame, the abnormal position point is deleted and the real-time driving trajectory of the burning vehicle is regenerated. When determining whether the real-time driving trajectory matches the driving direction of the burning vehicle, the lane line direction of the lane where the burning vehicle is located can be determined as the driving direction, and then it is determined whether the driving direction is consistent with the direction corresponding to the real-time driving trajectory. If they are inconsistent, it is determined that the real-time driving trajectory does not match the driving direction of the burning vehicle.

[0073] This scheme is configured in such a way that the robustness of the burning vehicle tracking process is enhanced through the spatial information of the burning vehicle, and when a non-burning vehicle is identified as a burning vehicle, the interference of the non-burning vehicle can be eliminated.

[0074] In an embodiment of the present application, optionally, in a specific achievable embodiment, the step of determining the real-time driving trajectory can be implemented through a model. Specifically, the BBOX regression algorithm can be used to perform coordinate regression of the target frame to obtain the specific coordinate information of the target, and through the mapping model of the tunnel space, the target location can be accurately obtained, and then fitted to the real-time driving trajectory.

[0075] For fire sources in high-speed movement, such as fires inside or on top of vehicles, the outer flames of the flames are suppressed by airflow due to the rapid movement of the vehicle, and infrared radiation is difficult to be quickly captured by infrared detection sensors during rapid movement. This application uses a fire vehicle detection model with video as input data to identify fires on fast-moving vehicles in highway tunnel scenes, retaining many advantages of image recognition technology in fire detection. By combining vehicle features and flame features, the recognition effect for special situations of vehicle fires in highway tunnels is better.

[0076] It should be noted that when identifying a vehicle on fire, multiple vehicles can be identified at the same time to obtain identification results for multiple vehicles on fire, so that the fire prevention and control capabilities in the tunnel can be greatly improved. By continuously collecting and analyzing data in the fire identification process, the embodiment of the present application can continuously optimize its own performance and improve the accuracy and speed of identification. At the same time, these data can also provide valuable reference information for relevant parties, so that relevant parties can better understand the fire risk situation in the tunnel and formulate more effective preventive measures.

[0077] In a specific example, if Figure 3 The specific example diagram of mobile fire detection in a tunnel shown in the figure can establish a spatial model in advance, which reflects the correspondence between the spatial position in the tunnel and the pixel points of the video frame of the monitoring video. The fire vehicle detection model is trained in advance. During the training, the simulated test video can be labeled. The label includes whether there is a fire vehicle and the marking box of the fire vehicle. The training is completed until the loss function meets certain conditions. When performing fire detection in real time, fire detection is performed on the real-time video. If a fire vehicle is detected by the fire vehicle detection model, the location point of the fire vehicle is determined based on the spatial model, and then a real-time trajectory is generated.

[0078] Embodiment 3

[0079] Figure 4 This is a schematic diagram of the structure of a mobile fire detection device in a tunnel provided in the third embodiment of the present application. The device can execute the mobile fire detection method in a tunnel provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method. Figure 4 As shown, the device comprises:

[0080] A real-time monitoring video data acquisition module 310 is used to acquire real-time monitoring video data in the tunnel;

[0081] A real-time recognition result determination module 320 is used to input the real-time monitoring video data into a pre-trained fire vehicle recognition model to obtain a real-time recognition result;

[0082] The real-time fire detection information determination module 330 is used to determine the real-time fire detection information in the tunnel based on the target monitoring video data including the vehicle on fire if the real-time identification result reflects that there is a vehicle on fire in the tunnel.

[0083] The technical solution of the embodiment of the present application includes: a real-time monitoring video data acquisition module 310, which is used to acquire real-time monitoring video data in the tunnel; a real-time recognition result determination module 320, which is used to input the real-time monitoring video data into a pre-trained fire vehicle recognition model to obtain a real-time recognition result; a real-time fire detection information determination module 330, which is used to determine the real-time fire detection information in the tunnel according to the target monitoring video data including the fire vehicle if the real-time recognition result reflects that there is a fire vehicle in the tunnel. This technical solution uses the fire vehicle recognition model to identify the fire vehicle in the real-time monitoring video data in the tunnel, which can detect the fire vehicle in time, and can accurately detect the moving fire vehicle, providing support for traffic safety in the tunnel.

[0084] In the embodiment of the present application, optionally, each video frame of the target monitoring video data includes a marking frame; the marking frame includes a vehicle on fire;

[0085] The real-time fire detection information determination module 330 includes:

[0086] An actual position determination unit, used to determine the actual position corresponding to the marking frame based on the correspondence between the pre-calibrated spatial position in the tunnel and the pixel points of the video frame of the monitoring video;

[0087] The real-time position information determining unit is used to determine the real-time position information of the vehicle on fire according to the actual position corresponding to each marking frame.

[0088] In the embodiment of the present application, optionally, the real-time location information determining unit includes:

[0089] A timestamp determination subunit, used to determine the acquisition time of the video frame corresponding to the burning vehicle as the timestamp of each actual position of the burning vehicle;

[0090] The real-time driving trajectory determination subunit is used to generate the real-time driving trajectory of the burning vehicle according to the actual position and timestamp of the burning vehicle.

[0091] In the embodiment of the present application, optionally, the device further includes:

[0092] A driving direction determination module is used to determine the driving direction of the burning vehicle according to the lane line information of the burning vehicle;

[0093] an abnormal video frame determination module, for determining an abnormal video frame where an abnormal position point in the real-time driving trajectory is located if the real-time driving trajectory does not match the actual driving direction of the vehicle on fire;

[0094] The real-time driving trajectory re-determination module is used to delete the abnormal position point and regenerate the real-time driving trajectory of the burning vehicle if it is determined that the burning vehicle in the abnormal video frame is inconsistent with the burning vehicles corresponding to the remaining video frames in the target video frame.

[0095] In the embodiment of the present application, optionally, the real-time recognition result determination module 320 includes:

[0096] A feature extraction unit, used for inputting the real-time monitoring video data into a pre-trained on-fire vehicle recognition model, and extracting vehicle features and / or flame features in the real-time monitoring video data through the on-fire vehicle recognition model;

[0097] A fire vehicle judgment unit, for obtaining a real-time recognition result that there is a fire vehicle if it is determined that the image area where the flame feature is extracted and the image area where the vehicle feature is extracted are both image areas corresponding to the same vehicle;

[0098] Otherwise, the real-time recognition result is that there is no vehicle on fire.

[0099] In the embodiment of the present application, optionally, the device further includes: a fire vehicle identification model training module, including:

[0100] A training sample determination unit, used to use video data with a tunnel as the background and including a vehicle as a training sample;

[0101] The fire vehicle recognition model determination unit is used to determine the label of the training sample, and train the fire vehicle detection model to be trained based on the training sample to obtain the trained fire vehicle detection model.

[0102] In an embodiment of the present application, optionally, the training sample determination unit includes:

[0103] A small target image acquisition subunit, used to acquire small target images; the small targets include moving flames, unburned vehicles, and burned vehicles;

[0104] A candidate sample determination subunit is used to fuse the small target image into the tunnel background video to obtain a candidate sample;

[0105] The training sample determination subunit is used to process the candidate samples through the image data enhancement algorithm to obtain the training samples.

[0106] A mobile fire detection device in a tunnel provided in an embodiment of the present application can execute a mobile fire detection method in a tunnel provided in any embodiment of the present invention, and has functional modules and beneficial effects corresponding to the execution method.

[0107] Embodiment 4

[0108] Figure 5 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0109] like Figure 5 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0110] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0111] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a mobile fire detection method in a tunnel.

[0112] In some embodiments, the mobile fire detection method in a tunnel may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the mobile fire detection method in a tunnel described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the mobile fire detection method in a tunnel in any other appropriate manner (for example, by means of firmware).

[0113] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0114] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0115] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0116] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0117] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0118] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0119] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0120] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for detecting mobile fire in a tunnel, It is characterized in that include: Obtain real-time surveillance video data in the tunnel; Inputting the real-time monitoring video data into a pre-trained fire vehicle recognition model to obtain a real-time recognition result; If the real-time identification result reflects that there is a vehicle on fire in the tunnel, real-time fire detection information in the tunnel is determined based on the target monitoring video data including the vehicle on fire.

2. The method according to claim 1, It is characterized in that Each video frame of the target monitoring video data includes a marking frame; The marking frame includes a vehicle on fire; Determine real-time fire detection information in the tunnel based on target surveillance video data including vehicles on fire, including: Based on the correspondence between the pre-calibrated spatial position in the tunnel and the pixel points of the video frame of the surveillance video, the actual position corresponding to the marking frame is determined; According to the actual position corresponding to each marking box, the real-time position information of the burning vehicle is determined.

3. The method according to claim 2, It is characterized in that According to the actual position corresponding to each marked box, the real-time position information of the burning vehicle is determined, including: The acquisition time of the video frame corresponding to the burning vehicle is determined as the timestamp of each actual position of the burning vehicle; According to the actual position and timestamp of the burning vehicle, the real-time driving trajectory of the burning vehicle is generated.

4. The method according to claim 3, It is characterized in that The method further comprises: Determine the traveling direction of the burning vehicle according to the lane line information of the burning vehicle; If the real-time driving trajectory does not match the actual driving direction of the vehicle on fire, determining an abnormal video frame where an abnormal position point in the real-time driving trajectory is located; If it is determined that the burning vehicle in the abnormal video frame is inconsistent with the burning vehicles corresponding to the remaining video frames in the target video frame, the abnormal position point is deleted and the real-time driving trajectory of the burning vehicle is regenerated.

5. The method according to claim 1, It is characterized in that The real-time monitoring video data is input into a pre-trained fire vehicle recognition model to obtain a real-time recognition result, including: Inputting the real-time monitoring video data into a pre-trained on-fire vehicle recognition model, and extracting vehicle features and / or flame features in the real-time monitoring video data through the on-fire vehicle recognition model; If it is determined that the image area where the flame feature is extracted and the image area where the vehicle feature is extracted are both image areas corresponding to the same vehicle, then the real-time recognition result obtained is that there is a vehicle on fire; Otherwise, the real-time recognition result is that there is no vehicle on fire.

6. The method according to claim 1, It is characterized in that The training process of the fire vehicle recognition model includes: The video data with the tunnel as the background and vehicles included are used as training samples; The labels of the training samples are determined, and the fire vehicle detection model to be trained is trained based on the training samples to obtain a trained fire vehicle detection model.

7. The method according to claim 6, It is characterized in that The process of determining the training samples includes: Acquire small target images; the small targets include moving flames, unburned vehicles, and burned vehicles; The small target image is fused into the tunnel background video to obtain a candidate sample; The candidate samples are processed by the image data enhancement algorithm to obtain training samples.

8. A mobile fire detection device in a tunnel, It is characterized in that include: A real-time monitoring video data acquisition module is used to acquire real-time monitoring video data in the tunnel; A real-time recognition result determination module is used to input the real-time monitoring video data into a pre-trained fire vehicle recognition model to obtain a real-time recognition result; The real-time fire detection information determination module is used to determine the real-time fire detection information in the tunnel based on the target monitoring video data including the burning vehicle if the real-time identification result reflects that there is a burning vehicle in the tunnel.

9. An electronic device, It is characterized in that The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method for detecting mobile fire in a tunnel according to any one of claims 1 to 7.

10. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method for detecting mobile fire in a tunnel according to any one of claims 1 to 7 when executed.

Citation Information

Cited By

  • Fire accident video analysis method, device, equipment, storage medium and product

    CN119888309A