A fire detection analysis method and device based on video analysis technology
Through video analysis technology, the grayscale values and motion characteristics of fire video images are collected and processed, and the fire characteristics are identified, which solves the problems of low recognition accuracy and false alarms of existing image fire detection methods, achieving more accurate fire detection and timely Early warning.
Patent Information
- Application Number
- CN202210546717.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-05-19
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2042-05-19
AI Technical Summary
The recognition accuracy of existing image fire detection methods is not high, and it is prone to false recognition and false alarms, and the accuracy of the result is greatly affected by environmental characteristics.
Using a fire detection method based on video analysis technology, by collecting smoke and fire video images, obtaining the change range and motion characteristics of pixel grayscale values, identifying the diffusion and swing movement of smoke and fire characteristics relative to the static environment, combining monitoring equipment and computer processing The network connection of the device, analyze the monitoring screen in real time, and determine whether there is a fire.
It improves the accuracy of fire identification and judgment, reduces the interference of environmentally similar objects on fire detection, enhances the effect of video fire detection, and promptly detects fire hazards.
Smart Images

Figure CN115049955B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of fire detection and analysis, and in particular to a fire detection and analysis method and device based on video analysis technology. Background Art
[0002] Fire detection is achieved by using monitoring equipment to detect whether there is a fire in the corresponding area environment. Image fire detection technology uses the detection screen in the monitoring equipment to detect and analyze the flames, smoke in the monitoring image and the temperature in the monitoring environment to achieve image fire detection.
[0003] The existing image fire detection method is to perform static analysis on flame images and smoke images to identify and judge fire phenomena. However, the recognition accuracy of this image fire detection method is not high. During detection, some objects with similar features in the monitoring environment will be misidentified, resulting in false fire detection. The accuracy of image fire detection results is greatly affected by environmental characteristics. Summary of the invention
[0004] In order to solve the problem that the recognition accuracy of the above image fire detection method is not high, some objects with similar features in the monitoring environment will be misidentified during detection, and the accuracy of the image fire detection result is greatly affected by environmental characteristics, the present invention is implemented through the following technical solutions: A fire detection analysis method based on video analysis technology, comprising the following specific steps:
[0005] S1, collecting multiple groups of smoke video images and fire light video images when a fire occurs;
[0006] S2, obtaining the minimum variation range (D1min, D2min) of the pixel grayscale values of the continuous multiple frames of the smoke video image, and obtaining the minimum grayscale value Dmin and the maximum grayscale value Dmax of the image in the smoke video image;
[0007] S3, obtaining the minimum variation range (F1min, F2min) of the pixel grayscale values of the continuous multiple frames of the fire light video image, and obtaining the minimum grayscale value Fmin and the maximum grayscale value Fmax of the image in the fire light video image;
[0008] S4, identifying the motion change parameters of the smoke feature relative to the static environment feature in the smoke video image, and obtaining the diffusion motion characteristics of the smoke feature relative to the static environment feature;
[0009] S5, identifying the motion change parameters of the fire light feature relative to the static environment feature in the fire light video image, and obtaining the swing motion characteristics of the fire light feature relative to the static environment feature;
[0010] S6. Use monitoring equipment to detect the environment space, and use video analysis technology to perform real-time analysis and processing on the monitoring images of the monitoring equipment;
[0011] S7, judging and analyzing whether the image pixels are in (D1min, D2min) or (F1min, F2min) according to the processed image pixels, when the image pixels do not fall into (D1min, D2min) or (F1min, F2min), judging that there is no fire in the monitoring screen of the monitoring device, otherwise executing S8;
[0012] S8. Based on the processed dynamic image, determine and analyze whether the smoke characteristics in the dynamic image are consistent with the diffusion movement characteristics of the smoke characteristics relative to the static environment characteristics, or determine and analyze whether the fire characteristics in the dynamic image are consistent with the swing movement characteristics of the fire characteristics relative to the static environment characteristics. If either of the two is met, it is determined that there is a fire in the monitoring image of the monitoring device. Otherwise, it is determined that there is no fire in the monitoring image of the monitoring device.
[0013] Furthermore, the method for obtaining the grayscale value of the smoke video image in S2 is:
[0014] S201, obtaining multiple groups of continuous multi-frame images of the smoke video images within a continuous period of time;
[0015] S202, processing multiple frames of each group of smoke video images to obtain a gray value D of a pixel of each frame;
[0016] S203, recording the variation range (D1, D2) of the pixel grayscale values of the continuous multiple frames of images, recording the variation range of the pixel grayscale values of the continuous multiple frames of images in the multiple groups of the smoke video images;
[0017] S204, determining the minimum variation range (D1min, D2min) of the pixel grayscale values of the continuous multiple frames of images, and determining the minimum grayscale value Dmin and the maximum grayscale value Dmax of the image in the smoke video image.
[0018] Furthermore, the method for obtaining the grayscale value of the fire light video image in S3 is as follows:
[0019] S301, obtaining multiple groups of continuous multi-frame images of the fire light video images within a continuous period of time;
[0020] S302, processing each group of multiple frames of the fire light video image to obtain a gray value F of a pixel of each frame;
[0021] S303, recording the variation range (F1, F2) of the pixel grayscale values of the continuous multiple frames of images, recording the variation range of the pixel grayscale values of the continuous multiple frames of images in the multiple groups of the fire light video images;
[0022] S304, determining the minimum variation range (F1min, F2min) of the pixel grayscale values of the continuous multiple frames of images, and determining the minimum grayscale value Fmin and the maximum grayscale value Fmax of the image in the fire light video image.
[0023] Furthermore, the specific steps of obtaining the diffusion motion characteristics of the smoke characteristics relative to the static environment characteristics in S4 include:
[0024] S401, extracting a dynamic smoke image containing surrounding static environment features from the smoke video image, and obtaining a plurality of continuous frames of the dynamic smoke image;
[0025] S402, marking static environment features and smoke features in the continuous multi-frame images;
[0026] S403, using image segmentation to process the continuous multiple frames of images to obtain a static environment feature image and a smoke feature image;
[0027] S404: Identify the motion change parameters of the smoke feature in the dynamic smoke image relative to the static environment feature, and obtain the diffusion motion characteristics of the smoke feature relative to the static environment feature.
[0028] Furthermore, the specific steps of obtaining the swing motion characteristics of the fire light characteristics relative to the static environment characteristics in S5 include:
[0029] S501, extracting a dynamic fire image containing surrounding static environment features from the fire video image, and obtaining a continuous plurality of frames of the dynamic fire image;
[0030] S502, marking static environment features and fire features in the continuous multi-frame images;
[0031] S503, using image segmentation to process the continuous multiple frames of images to obtain a static environment feature image and a fire feature image;
[0032] S504, identifying the motion change parameters of the fire light feature in the dynamic fire light image relative to the static environment feature, and obtaining the swing motion characteristics of the fire light feature relative to the static environment feature.
[0033] Furthermore, the grayscale value D of the frame image in the smoke video image in S202 and the grayscale value F of the frame image in the fire video image in S302 are:
[0034]
[0035] Among them, R, G, and B are color component values.
[0036] Furthermore, the identification method of the smoke characteristic motion change parameter in S404 and the fire characteristic motion change parameter in S504 is:
[0037] Each frame of the dynamic image is divided into blocks of M×N pixels. In the matching window of (N+2w, N+2w), w is the residual value of the matching window relative to the boundary of the image. The current block is compared with the corresponding block in the previous frame. The best match is found based on the matching standard to obtain the replacement position of the current block. The position changes of the blocks of smoke features or fire features are combined to obtain the running change parameters of the smoke features or fire features in the continuous multiple frames of the dynamic image. The matching standard is the minimum mean square error MSE or the minimum absolute error MAE:
[0038]
[0039]
[0040] Where f(m,n) indicates that the position of the current block in the current frame is (m,n), and f(m+i,n+j) indicates that the position of the current block in the previous frame is (m+i,n+j).
[0041] A fire detection device based on video analysis technology, comprising a monitoring device and a computer processing device, wherein the monitoring device and the computer device are connected via a network, and an image recognition module and a dynamic video extraction module are arranged inside the computer device, wherein the image recognition module is used to recognize and obtain a smoke feature image or a fire feature image appearing in a monitoring screen of the monitoring device, and the dynamic video extraction module is used to extract a dynamic image of a smoke feature or a fire feature appearing in a monitoring screen of the monitoring device;
[0042] The computer device is provided with a grayscale value calculation module and a video image dynamic feature analysis module. The grayscale value calculation module is used to calculate the grayscale value of the smoke feature image or the fire feature image obtained by the image recognition module, and the video image dynamic feature analysis module is used to process and analyze the motion characteristics of the smoke features or the fire features in the smoke feature or the fire feature dynamic image extracted by the dynamic video extraction module.
[0043] Furthermore, both the monitoring device and the computer device are provided with an early warning device, and the early warning device is connected to the video image dynamic feature analysis module via an electrical signal.
[0044] Furthermore, a driving and adjusting module is provided inside the monitoring device, and the driving and adjusting module is used to control the monitoring angle and monitoring focal length of the monitoring device.
[0045] Compared with the prior art, the present invention has the following beneficial effects:
[0046] 1. The fire detection analysis method and device based on video analysis technology can perform real-time fire detection in the monitoring environment by collecting and processing the image features of smoke and fire light when a fire occurs, as well as the dynamic change features of smoke and fire light as comparison standards for food fire detection, identification and judgment, reduce the interference of similar objects in the environment on fire detection, effectively improve the accuracy of fire identification and judgment, and help improve the effect of video fire detection.
[0047] 2. The fire detection and analysis method and device based on video analysis technology can perform real-time fire detection in the monitoring area of the monitoring equipment through the network connection between the monitoring equipment and the computer equipment, the coordinated use of the image recognition module and the gray value calculation module inside the computer equipment, and the coordinated use of the dynamic video extraction module and the video image dynamic feature analysis module, which is conducive to timely discovery of fire hazards. BRIEF DESCRIPTION OF THE DRAWINGS
[0048] Figure 1 This is a flow chart of the fire detection and analysis method of the present invention;
[0049] Figure 2 It is a structural schematic diagram of the fire detection and analysis device of the present invention. DETAILED DESCRIPTION
[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0051] The embodiments of the fire detection analysis method and device based on video analysis technology are as follows:
[0052] See also Figure 1 , a fire detection analysis method based on video analysis technology, comprising the following specific steps:
[0053] S1, collecting multiple groups of smoke video images and fire light video images when a fire occurs;
[0054] S2. Obtain the minimum variation range (D1min, D2min) of the pixel grayscale values of the continuous multiple frames in the smoke video image, and obtain the minimum grayscale value Dmin and the maximum grayscale value Dmax of the image in the smoke video image:
[0055] S201, obtaining multiple sets of continuous multi-frame images of smoke video images within a continuous period of time;
[0056] S202, processing multiple frames of each group of smoke video images to obtain the gray value D of each frame pixel:
[0057]
[0058] Among them, R, G, and B are color component values.
[0059] S203, recording the variation range (D1, D2) of the pixel grayscale values of the continuous multiple frames of images, recording the variation range of the pixel grayscale values of the continuous multiple frames of images in the multiple groups of smoke video images;
[0060] S204, determining the minimum variation range (D1min, D2min) of the pixel grayscale values of the continuous multiple frame images, and determining the minimum grayscale value Dmin and the maximum grayscale value Dmax of the image in the smoke video image.
[0061] S3, obtain the minimum variation range (F1min, F2min) of the pixel grayscale values of the continuous multiple frames in the fire light video image, and obtain the minimum grayscale value Fmin and the maximum grayscale value Fmax of the image in the fire light video image:
[0062] S301, obtaining multiple sets of continuous multiple frame images of fire light video images within a continuous period of time;
[0063] S302, process multiple frames of each group of fire light video images to obtain the gray value F of each frame pixel:
[0064]
[0065] Among them, R, G, and B are color component values.
[0066] S303, recording the variation range (F1, F2) of the pixel grayscale values of the continuous multiple frames of images, recording the variation range of the pixel grayscale values of the continuous multiple frames of images in the multiple groups of fire light video images;
[0067] S304, determine the minimum variation range (F1min, F2min) of the pixel grayscale values of the continuous multiple frame images, and determine the minimum grayscale value Fmin and the maximum grayscale value Fmax of the image in the fire light video image.
[0068] S4. Identify the motion change parameters of the smoke feature relative to the static environment feature in the smoke video image, and obtain the diffusion motion characteristics of the smoke feature relative to the static environment feature:
[0069] S401, extracting a dynamic smoke image containing surrounding static environment features from a smoke video image, and obtaining a continuous plurality of frames of the dynamic smoke image;
[0070] S402, marking static environment features and smoke features in the continuous multi-frame images;
[0071] S403, using image segmentation to process the continuous multiple frames of images to obtain a static environment feature image and a smoke feature image;
[0072] S404, identifying the motion change parameters of the smoke features in the dynamic smoke image relative to the static environment features, and obtaining the diffusion motion features of the smoke features relative to the static environment features:
[0073] Each frame of the dynamic smoke image is divided into blocks of M×N pixels. In the matching window of (N+2w, N+2w), w is the residual value of the matching window relative to the boundary of the image. The current block is compared with the corresponding block in the previous frame. The best match is found based on the matching standard to obtain the replacement position of the current block. The position changes of the blocks of smoke features or fire features are combined to obtain the running change parameters of the smoke features or fire features in the continuous multiple frames of the dynamic image. The matching standard is the minimum mean square error MSE or the minimum absolute error MAE:
[0074]
[0075]
[0076] Where f(m,n) indicates that the position of the current block in the current frame is (m,n), and f(m+i,n+j) indicates that the position of the current block in the previous frame is (m+i,n+j).
[0077] S5. Identify the motion change parameters of the fire light feature relative to the static environment feature in the fire light video image, and obtain the swing motion feature of the fire light feature relative to the static environment feature:
[0078] S501, extracting a dynamic fire image containing surrounding static environment features from the fire video image, and obtaining a continuous multi-frame image of the dynamic fire image;
[0079] S502, marking static environment features and fire features in the continuous multi-frame images;
[0080] S503, using image segmentation to process the continuous multiple frames of images to obtain a static environment feature image and a fire feature image;
[0081] S504, identifying the motion change parameters of the fire light feature relative to the static environment feature in the dynamic fire light image, and obtaining the swing motion feature of the fire light feature relative to the static environment feature:
[0082] Each frame of the dynamic fire image is divided into blocks of M×N pixels. In the matching window of (N+2w, N+2w), w is the residual value of the matching window relative to the boundary of the image. The current block is compared with the corresponding block in the previous frame. The best match is found based on the matching standard to obtain the replacement position of the current block. The position changes of the blocks of smoke features or fire features are combined to obtain the running change parameters of the smoke features or fire features in the continuous multiple frames of the dynamic image. The matching standard is the minimum mean square error MSE or the minimum absolute error MAE:
[0083]
[0084]
[0085] Where f(m,n) indicates that the position of the current block in the current frame is (m,n), and f(m+i,n+j) indicates that the position of the current block in the previous frame is (m+i,n+j).
[0086] S6. Use monitoring equipment to detect the environment space, and use video analysis technology to perform real-time analysis and processing on the monitoring images of the monitoring equipment;
[0087] S7, judging and analyzing whether the image pixels are in (D1min, D2min) or (F1min, F2min) according to the processed image pixels, when the image pixels do not fall into (D1min, D2min) or (F1min, F2min), judging that there is no fire in the monitoring screen of the monitoring device, otherwise executing S8;
[0088] S8. Based on the processed dynamic image, determine and analyze whether the smoke characteristics in the dynamic image are consistent with the diffusion movement characteristics of the smoke characteristics relative to the static environment characteristics, or determine and analyze whether the fire characteristics in the dynamic image are consistent with the swing movement characteristics of the fire characteristics relative to the static environment characteristics. If either of the two is met, it is determined that there is a fire in the monitoring image of the monitoring device. Otherwise, it is determined that there is no fire in the monitoring image of the monitoring device.
[0089] See also Figure 2 , a fire detection device based on video analysis technology, including a monitoring device and a computer processing device, the monitoring device and the computer device are connected by a network, an image recognition module and a dynamic video extraction module are arranged inside the computer device, the image recognition module is used to recognize and obtain a smoke feature image or a fire feature image appearing in a monitoring screen of the monitoring device, and the dynamic video extraction module is used to extract a dynamic image of a smoke feature or a fire feature appearing in a monitoring screen of the monitoring device;
[0090] The computer device is provided with a grayscale value calculation module and a video image dynamic feature analysis module. The grayscale value calculation module is used to calculate the grayscale value of the smoke feature image or the fire feature image obtained by the image recognition module. The video image dynamic feature analysis module is used to process and analyze the motion characteristics of the smoke features or the fire features in the dynamic image of the smoke features or the fire features extracted by the dynamic video extraction module.
[0091] Both the monitoring equipment and the computer equipment are equipped with early warning devices, and the early warning devices are connected to the video image dynamic feature analysis module via electrical signals. A driving and adjusting module is provided inside the monitoring equipment, and the driving and adjusting module is used to control the monitoring angle and monitoring focal length of the monitoring equipment.
[0092] Working principle:
[0093] The monitoring equipment is used to monitor the environmental space that needs to be monitored in real time. The images monitored by the monitoring equipment will be transmitted to the computer equipment in real time through the network system. The image recognition module in the computer equipment will recognize and obtain the smoke characteristic image or fire characteristic image appearing in the monitoring image of the monitoring equipment. The grayscale value calculation module will calculate the grayscale value of the smoke characteristic image or fire characteristic image obtained by the image recognition module, thereby identifying whether there is a smoke characteristic image or fire characteristic image in the monitoring image.
[0094] When a smoke feature image or a fire feature image is identified in the monitoring image, the dynamic video extraction module will extract the dynamic image of the smoke feature or fire feature appearing in the monitoring screen of the monitoring device, and the video image dynamic feature analysis module will process and analyze the motion characteristics of the smoke feature or fire feature in the dynamic image of the smoke feature or fire feature extracted by the dynamic video extraction module, and determine whether the smoke feature image or fire feature image recognized by the image recognition module is real, thereby determining whether there is a fire in the monitoring area of the monitoring device.
[0095] When there is a fire, the video image dynamic feature analysis module will transmit information to the early warning device in the monitoring equipment and computer equipment through electrical signals. The early warning device will sound an alarm. At the same time, the driving adjustment module inside the monitoring device will control the monitoring angle of the monitoring device and adjust the monitoring focal length, so as to further grasp the details of the fire in the monitoring area of the monitoring device.
[0096] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A fire detection and analysis method based on video analysis technology, characterized in that: The specific steps include: S1, collecting multiple groups of smoke video images and fire light video images when a fire occurs; S2, obtaining the minimum variation range (D1min, D2min) of the pixel grayscale values of the continuous multiple frames of the smoke video image, and obtaining the minimum grayscale value Dmin and the maximum grayscale value Dmax of the image in the smoke video image; S3, obtaining the minimum variation range (F1min, F2min) of the pixel grayscale values of the continuous multiple frames of the fire light video image, and obtaining the minimum grayscale value Fmin and the maximum grayscale value Fmax of the image in the fire light video image; S4, identifying the motion change parameters of the smoke feature relative to the static environment feature in the smoke video image, and obtaining the diffusion motion characteristics of the smoke feature relative to the static environment feature; S5, identifying the motion change parameters of the fire light feature relative to the static environment feature in the fire light video image, and obtaining the swing motion characteristics of the fire light feature relative to the static environment feature; S6. Use monitoring equipment to detect the environment space, and use video analysis technology to perform real-time analysis and processing on the monitoring images of the monitoring equipment; S7, judging and analyzing whether the image pixels are in (D1min, D2min) or (F1min, F2min) according to the processed image pixels, when the image pixels do not fall into (D1min, D2min) or (F1min, F2min), judging that there is no fire in the monitoring screen of the monitoring device, otherwise executing S8; S8. According to the processed dynamic image, determine and analyze whether the smoke characteristics in the dynamic image meet the diffusion motion characteristics of the smoke characteristics relative to the static environment characteristics, or determine and analyze whether the fire characteristics in the dynamic image meet the swing motion characteristics of the fire characteristics relative to the static environment characteristics. If one of the two is satisfied, it is determined that there is a fire in the monitoring image of the monitoring device. Otherwise, it is determined that there is no fire in the monitoring image of the monitoring device. The specific steps of obtaining the diffusion motion characteristics of the smoke characteristics relative to the static environment characteristics in S4 include: S401, extracting a dynamic smoke image containing surrounding static environment features from the smoke video image, and obtaining a plurality of continuous frames of the dynamic smoke image; S402, marking static environment features and smoke features in the continuous multi-frame images; S403, using image segmentation to process the continuous multiple frames of images to obtain a static environment feature image and a smoke feature image; S404, identifying the motion change parameters of the smoke feature in the dynamic smoke image relative to the static environment feature, and obtaining the diffusion motion characteristics of the smoke feature relative to the static environment feature; The specific steps of obtaining the swing motion characteristics of the fire light characteristics relative to the static environment characteristics in S5 include: S501, extracting a dynamic fire image containing surrounding static environment features from the fire video image, and obtaining a continuous plurality of frames of the dynamic fire image; S502, marking static environment features and fire features in the continuous multi-frame images; S503, using image segmentation to process the continuous multiple frames of images to obtain a static environment feature image and a fire feature image; S504, identifying the motion change parameters of the fire light feature in the dynamic fire light image relative to the static environment feature, and obtaining the swing motion characteristics of the fire light feature relative to the static environment feature; The identification method of the smoke characteristic motion change parameter in S404 and the fire characteristic motion change parameter in S504 is: Each frame of the dynamic image is divided into blocks of M×N pixels. In the matching window of (N+2w, N+2w), w is the residual value of the matching window relative to the boundary of the image. The current block is compared with the corresponding block in the previous frame. The best match is found based on the matching standard to obtain the replacement position of the current block. The position changes of the blocks of smoke features or fire features are combined to obtain the running change parameters of the smoke features or fire features in the continuous multiple frames of the dynamic image. The matching standard is the minimum mean square error MSE or the minimum absolute error MAE: ,-w≤i,j≤w ,-w≤i,j≤w in Indicates that the position of the current block in the current frame is (m, n), Indicates that the position of the current block in the previous frame is (m+i, n+j).
2. The fire detection and analysis method based on video analysis technology according to claim 1, characterized in that: The method for obtaining the grayscale value of the smoke video image described in S2: S201, obtaining multiple groups of continuous multi-frame images of the smoke video images within a continuous period of time; S202, processing multiple frames of each group of smoke video images to obtain a gray value D of a pixel of each frame; S203, recording the variation range (D1, D2) of the pixel grayscale values of the continuous multiple frames of images, recording the variation range of the pixel grayscale values of the continuous multiple frames of images in the multiple groups of the smoke video images; S204, determining a minimum variation range (D1min, D2min) of the pixel grayscale values of the continuous multiple frames of images, and determining a minimum grayscale value Dmin and a maximum grayscale value Dmax of the image in the smoke video image.
3. The fire detection and analysis method based on video analysis technology according to claim 2 is characterized in that: The method for obtaining the grayscale value of the fire light video image in S3: S301, obtaining multiple groups of continuous multi-frame images of the fire light video images within a continuous period of time; S302, processing each group of multiple frames of the fire light video image to obtain a gray value F of a pixel of each frame; S303, recording the variation range (F1, F2) of the pixel grayscale values of the continuous multiple frames of images, recording the variation range of the pixel grayscale values of the continuous multiple frames of images in the multiple groups of the fire light video images; S304, determining the minimum variation range (F1min, F2min) of the pixel grayscale values of the continuous multiple frames of images, and determining the minimum grayscale value Fmin and the maximum grayscale value Fmax of the image in the fire light video image.
4. A fire detection device based on video analysis technology, applied to a fire detection analysis method based on video analysis technology as claimed in any one of claims 1 to 3, comprising a monitoring device and a computer processing device, characterized in that: The monitoring device and the computer device are connected via a network. The computer device is internally provided with an image recognition module and a dynamic video extraction module. The image recognition module is used to recognize and obtain a smoke feature image or a fire feature image appearing in a monitoring screen of the monitoring device. The dynamic video extraction module is used to extract a dynamic image of a smoke feature or a fire feature appearing in a monitoring screen of the monitoring device. The computer device is provided with a grayscale value calculation module and a video image dynamic feature analysis module. The grayscale value calculation module is used to calculate the grayscale value of the smoke feature image or the fire feature image obtained by the image recognition module, and the video image dynamic feature analysis module is used to process and analyze the motion characteristics of the smoke features or the fire features in the smoke feature or the fire feature dynamic image extracted by the dynamic video extraction module.
5. A fire detection device based on video analysis technology according to claim 4, characterized in that: The monitoring device and the computer device are both provided with an early warning device, and the early warning device is connected to the video image dynamic feature analysis module via an electrical signal.
6. A fire detection device based on video analysis technology according to claim 5, characterized in that: The monitoring device is provided with a driving and adjusting module inside, and the driving and adjusting module is used to control the monitoring angle and monitoring focal length of the monitoring device.
