CN118397779B - Smoke and fire detection early warning method and system based on AI video analysis - Google Patents
Smoke and fire detection early warning method and system based on AI video analysis Download PDFInfo
- Publication number
- CN118397779B CN118397779B CN202410389130.2A CN202410389130A CN118397779B CN 118397779 B CN118397779 B CN 118397779B CN 202410389130 A CN202410389130 A CN 202410389130A CN 118397779 B CN118397779 B CN 118397779B
- Authority
- CN
- China
- Prior art keywords
- smoke
- frame
- video
- image
- fire
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 92
- 238000004458 analytical method Methods 0.000 title claims abstract description 81
- 238000000034 method Methods 0.000 title claims abstract description 29
- 239000000779 smoke Substances 0.000 title claims description 232
- 238000000605 extraction Methods 0.000 claims abstract description 66
- 238000004891 communication Methods 0.000 claims abstract description 63
- 238000007689 inspection Methods 0.000 claims abstract description 11
- 238000003032 molecular docking Methods 0.000 claims abstract description 11
- 230000005540 biological transmission Effects 0.000 claims description 35
- 238000012544 monitoring process Methods 0.000 claims description 28
- 238000005070 sampling Methods 0.000 claims description 14
- 210000001503 joint Anatomy 0.000 claims description 9
- 230000007246 mechanism Effects 0.000 claims description 8
- 230000006978 adaptation Effects 0.000 claims description 7
- 238000006243 chemical reaction Methods 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 5
- 238000013139 quantization Methods 0.000 claims description 5
- 238000007726 management method Methods 0.000 description 11
- 230000007613 environmental effect Effects 0.000 description 6
- 238000012549 training Methods 0.000 description 6
- 238000011156 evaluation Methods 0.000 description 5
- 238000013441 quality evaluation Methods 0.000 description 5
- 238000010223 real-time analysis Methods 0.000 description 5
- 239000003086 colorant Substances 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000000877 morphologic effect Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000004806 packaging method and process Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B17/00—Fire alarms; Alarms responsive to explosion
- G08B17/10—Actuation by presence of smoke or gases, e.g. automatic alarm devices for analysing flowing fluid materials by the use of optical means
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B17/00—Fire alarms; Alarms responsive to explosion
- G08B17/06—Electric actuation of the alarm, e.g. using a thermally-operated switch
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
- H04L69/08—Protocols for interworking; Protocol conversion
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Business, Economics & Management (AREA)
- Emergency Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Security & Cryptography (AREA)
- Computer Networks & Wireless Communication (AREA)
- Multimedia (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Alarm Systems (AREA)
Abstract
本发明提供了一种基于AI视频分析的烟火检测预警方法及系统，包括：将烟火预警监管平台与视频设备进行对接，并基于对接结果对视频设备中的实时视频流进行视频抽帧；基于视频抽帧结果每次将一帧图像帧基于通信接口传送至视频分析模型进行初步烟火检测，并当存在烟火时将目标数目的图像帧集合基于通信接口复传至视频分析模型进行复检；基于复检结果确定当存在烟火时，基于烟火预警监管平台显示实时告警页面，并同步向管理终端发送烟火预警通知。提高了对烟火检测的准确性以及精度，保障了烟火预警的准确率，避免了重复错误报警。
The present invention provides a method and system for detecting and warning fireworks based on AI video analysis, including: docking a fireworks warning supervision platform with a video device, and extracting video frames from the real-time video stream in the video device based on the docking result; transmitting one image frame to a video analysis model based on a communication interface for preliminary fireworks detection based on the video extraction result, and retransmitting a target number of image frames to the video analysis model based on a communication interface for re-inspection when fireworks are present; based on the re-inspection result, when fireworks are present, displaying a real-time alarm page based on the fireworks warning supervision platform, and synchronously sending a fireworks warning notification to a management terminal. The accuracy and precision of fireworks detection are improved, the accuracy of fireworks warning is guaranteed, and repeated false alarms are avoided.
Description
Technical Field
The invention relates to the technical field of image processing and monitoring, in particular to a smoke and fire detection early warning method and system based on AI video analysis.
Background
Currently, in social production and life, fire prevention is an important proposition of safety protection, so that smoke and fire monitoring needs to be emphasized to ensure safety;
the traditional image processing method is to identify and detect some characteristics such as shape, color and texture in the image, but the traditional image processing is to analyze the collected field image, so that real-time analysis cannot be realized, and meanwhile, when the similar smoke and fire condition exists in one image is detected during analysis, corresponding alarm operation is performed, so that the instantaneity and the precision of smoke and fire early warning are not good, the accuracy of smoke and fire detection results is influenced, and repeated false alarms are performed;
therefore, in order to overcome the defects, the invention provides a smoke and fire detection early warning method and system based on AI video analysis.
Disclosure of Invention
The invention provides a smoke and fire detection early warning method and a smoke and fire detection early warning system based on AI video analysis, which are used for conveniently transmitting a real-time video stream of a monitoring area acquired by video equipment to a smoke and fire early warning platform for real-time analysis, secondly, performing video sampling inspection on the real-time video stream, transmitting an obtained frame of image frame to a video analysis model for preliminary smoke and fire detection, improving the efficiency of preliminary smoke and fire detection, simultaneously, performing rechecking on an image frame set of locking target data in the real-time video stream when the preliminary smoke and fire detection judges that smoke and fire exist, improving the accuracy and precision of smoke and fire detection, and finally, performing smoke and fire early warning notification when judging that smoke and fire exist according to a rechecking result, ensuring the accuracy of smoke and fire early warning and avoiding repeated false alarm.
The invention provides a smoke and fire detection early warning method based on AI video analysis, which comprises the following steps:
S1, butting a smoke and fire early warning supervision platform with video equipment, and performing video frame extraction on a real-time video stream in the video equipment based on a butting result;
S2, transmitting one frame of image frame to a video analysis model based on a communication interface each time based on a video frame extraction result to perform preliminary smoke and fire detection, and re-transmitting the image frame set with the target number to the video analysis model based on the communication interface to perform re-detection when smoke and fire exist;
And S3, determining that when smoke exists based on the rechecking result, displaying a real-time warning page based on a smoke early warning supervision platform, and synchronously sending smoke early warning notification to the management terminal.
Preferably, in a smoke and fire detection early warning method based on AI video analysis, in S1, a smoke and fire early warning supervision platform is docked with video equipment, including:
acquiring terminal parameters of different video devices, and determining service standards and communication protocols of the different video devices based on the terminal parameters;
Performing differential classification on different video devices based on a service standard, performing parameter compatibility adaptation on preset gateway devices based on a differential classification result according to a communication protocol, and docking different types of video devices with the preset gateway devices based on the parameter compatibility adaptation;
Meanwhile, the data interfaces of the smoke and fire early warning supervision platform are subjected to interface standard unification, and the data interfaces with the unified interface standard are in butt joint with the preset gateway equipment with the compatible and matched parameters, so that the butt joint of the smoke and fire early warning supervision platform and the video equipment is completed.
Preferably, in a smoke and fire detection early warning method based on AI video analysis, in S1, video frame extraction is performed on a real-time video stream in a video device based on a docking result, including:
reading a real-time video stream in the video device based on the docking result, and determining an image frame contained in the real-time video stream based on the reading result;
extracting corresponding image content from the real-time video stream frame by frame based on the image frames, obtaining each frame of image data based on the image content, and obtaining a frame image sequence based on each frame of image data;
And determining a polling frame extraction mechanism for the frame image sequence based on the video frame extraction index, and performing video frame extraction on a real-time video stream in the video device based on the polling frame extraction mechanism.
Preferably, in a smoke and fire detection early warning method based on AI video analysis, in S2, based on a video frame extraction result, one frame of image frame is transmitted to a video analysis model based on a communication interface for preliminary smoke and fire detection, and when smoke and fire exist, a target number of image frame sets are re-transmitted to the video analysis model based on the communication interface for re-detection, including:
acquiring a frame of image frame corresponding to a video frame extraction result, and determining a data transmission rule based on a communication protocol of a communication interface;
carrying out semantic extraction on the data transmission rule, determining a grammar execution command in the data transmission rule based on the extracted semantic, and carrying out format conversion on a frame of image frame based on the grammar execution command to obtain transmissible image data;
transmitting transmissible image data to a video analysis model based on a communication interface, and dividing a frame of image frame into M sampling small images based on a transmission result;
Respectively extracting the three-primary-color gray average value of each sampling small image in a color space based on the RGB color channels, and carrying out weight summarization on the three-primary-color gray average value to obtain a color histogram corresponding to one frame of image frame;
Converting the color histogram into a low-dimensional color histogram, quantizing each color component in the low-dimensional color histogram to a normalization range, and obtaining the color duty ratio in one frame of image frame based on the quantization result;
Determining the membership degree of the color ratio relative to the reference firework color characteristic, and judging that the firework exists when the membership degree is higher than a preset membership degree threshold value;
And re-transmitting the image frame set of the target number to a video analysis model for re-inspection based on the communication interface based on the judgment result.
Preferably, a smoke and fire detection early warning method based on AI video analysis, based on the determination result, the image frame set of the target number is retransmitted to the video analysis model based on the communication interface for rechecking, including:
locking the source position of one frame of image frame corresponding to the video frame extraction result based on the judgment result, and locking a known number of forward and backward video frames in the real-time video stream by taking the one frame of image frame corresponding to the video frame extraction result as the center based on the source position, so as to obtain a target number of image frame sets;
Retransmitting the image frame set of the target number to a video analysis model based on a communication interface, and carrying out bit-by-bit classification on the image frames of the target number based on a retransmission result to obtain inter-frame differential parameters;
determining a motion area between image frame sets of a target number based on the inter-frame differential parameters, and determining a spatial span of a preliminary smoke and fire detection result in the image frame sets based on the motion area;
When the space span is larger than a preset requirement, calibrating a smoke and fire area in each image frame in the image frame set based on a preliminary smoke and fire detection result, and performing edge monitoring on smoke and fire in the smoke and fire area based on the calibration result to obtain a smoke and fire profile;
Determining smoke morphology features based on the smoke profile, and performing a recheck of the target number of image frame sets based on the smoke features.
The invention provides a smoke and fire detection early warning system based on AI video analysis, comprising:
the video frame extraction module is used for butting the smoke and fire early warning supervision platform with the video equipment and carrying out video frame extraction on a real-time video stream in the video equipment based on a butting result;
The smoke and fire detection module is used for transmitting one frame of image frame to the video analysis model based on the communication interface each time based on the video frame extraction result to perform preliminary smoke and fire detection, and retransmitting the image frame set with the target number to the video analysis model based on the communication interface to perform rechecking when smoke and fire exist;
and the early warning module is used for determining that when smoke exists based on the rechecking result, displaying a real-time warning page based on the smoke early warning supervision platform and synchronously sending smoke early warning notification to the management terminal.
Compared with the prior art, the invention has the following beneficial effects:
1. The video equipment is in butt joint with the smoke and fire early warning platform, so that real-time video streams of a monitoring area collected by the video equipment are conveniently transmitted to the smoke and fire early warning platform for real-time analysis, secondly, video sampling inspection is carried out on the real-time video streams, one frame of image frame is obtained and transmitted to the video analysis model for preliminary smoke and fire detection, the efficiency of preliminary smoke and fire detection is improved, meanwhile, when the preliminary smoke and fire detection judges that smoke and fire exist, rechecking is carried out on an image frame set of locking target data in the real-time video streams, accuracy and precision of smoke and fire detection are improved, finally, smoke and fire early warning notification is carried out when the smoke and fire exist are judged according to rechecking results, the accuracy of smoke and fire early warning is guaranteed, and repeated false alarm is avoided.
2. The method comprises the steps of analyzing a communication protocol of a communication interface, accurately and effectively determining a data transmission rule corresponding to the communication interface, extracting semantics of the data transmission rule, locking a grammar execution command in the data transmission rule according to an extraction result, and converting a frame of image frame through the grammar execution command, so that the frame of image frame obtained by video frame extraction can be accurately and reliably transmitted to a video analysis model, finally, segmenting the transmitted frame of image frame and analyzing a color histogram, locking the color proportion of each color in the frame of image frame according to the color histogram, comparing the locked color proportion with a reference firework color characteristic, accurately and reliably detecting smoke preliminarily, guaranteeing the accuracy of preliminary detection, and timely rechecking when judging that smoke exists, and guaranteeing the timeliness and reliability of smoke detection.
Additional features and advantages of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the application. The objects and other advantages of the application may be realized and obtained by means of the instrumentalities particularly pointed out in the specification.
The technical scheme of the invention is further described in detail through the drawings and the embodiments.
Drawings
The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention. In the drawings:
FIG. 1 is a flowchart of a smoke and fire detection early warning method based on AI video analysis in an embodiment of the invention;
FIG. 2 is a flowchart of step 1 in a smoke and fire detection early warning method based on AI video analysis in an embodiment of the invention;
fig. 3 is a block diagram of a smoke and fire detection early warning system based on AI video analysis in an embodiment of the invention.
Detailed Description
The preferred embodiments of the present invention will be described below with reference to the accompanying drawings, it being understood that the preferred embodiments described herein are for illustration and explanation of the present invention only, and are not intended to limit the present invention.
Example 1:
the embodiment provides a smoke and fire detection early warning method based on AI video analysis, as shown in FIG. 1, comprising the following steps:
S1, butting a smoke and fire early warning supervision platform with video equipment, and performing video frame extraction on a real-time video stream in the video equipment based on a butting result;
S2, transmitting one frame of image frame to a video analysis model based on a communication interface each time based on a video frame extraction result to perform preliminary smoke and fire detection, and re-transmitting the image frame set with the target number to the video analysis model based on the communication interface to perform re-detection when smoke and fire exist;
And S3, determining that when smoke exists based on the rechecking result, displaying a real-time warning page based on a smoke early warning supervision platform, and synchronously sending smoke early warning notification to the management terminal.
In this embodiment, the video device is arranged in the monitoring area and is used for collecting real-time video of the monitoring area, so that real-time transmission of the monitoring video back to the monitoring area is facilitated, the video device comprises national standard equipment and non-national standard equipment, and the video device can be in butt joint with the smoke and fire early warning supervision platform, wherein the smoke and fire early warning supervision platform is set in advance.
In this embodiment, the real-time video stream refers to real-time video data in a corresponding area monitored by the video device.
In this embodiment, video frame extraction refers to extracting a frame of image from a real-time video stream for smoke and fire preliminary detection.
In the embodiment, the video analysis model is built in advance and used for explaining the received image frames, so that accurate and effective judgment on whether smoke and fire conditions exist or not is realized.
In this embodiment, the preliminary smoke and fire detection refers to performing rough analysis on an image frame obtained by video frame extraction through a video analysis model, and judging whether smoke and fire exist, that is, whether smoke and fire exist actually cannot be accurately determined, and further analysis on a real-time video stream is required.
In this embodiment, the target number is set in advance, and is typically 8 frames.
In this embodiment, the rechecking refers to performing the smoke and fire detection again on the image frame set with the target number through the video analysis model, so as to accurately determine whether smoke and fire actually exist.
The technical scheme has the advantages that the video equipment is in butt joint with the smoke and fire early warning platform, so that real-time video streams of a monitoring area acquired by the video equipment are conveniently transmitted to the smoke and fire early warning platform for real-time analysis, secondly, video sampling inspection is carried out on the real-time video streams, one frame of image frames are obtained and transmitted to the video analysis model for preliminary smoke and fire detection, the efficiency of preliminary smoke and fire detection is improved, meanwhile, when the preliminary smoke and fire detection judges that smoke and fire exist, rechecking is carried out on an image frame set of locking target data in the real-time video streams, the accuracy and precision of smoke and fire detection are improved, finally, smoke and fire early warning notification is carried out when the smoke and fire exists according to rechecking results, the accuracy of smoke and fire early warning is guaranteed, and repeated false alarm is avoided.
Example 2:
On the basis of embodiment 1, this embodiment provides a smoke and fire detection early warning method based on AI video analysis, as shown in fig. 2, in S1, the smoke and fire early warning supervision platform is docked with a video device, including:
s101, acquiring terminal parameters of different video devices, and determining service standards and communication protocols of the different video devices based on the terminal parameters;
S102, carrying out differential classification on different video devices based on a service standard, carrying out parameter compatibility adaptation on preset gateway devices according to a communication protocol based on a differential classification result, and carrying out butt joint on the video devices of different categories and the preset gateway devices based on the parameter compatibility adaptation;
And S103, simultaneously, unifying interface standards of the data interfaces of the smoke and fire early warning supervision platform, and butting the data interfaces with the unified interface standards with preset gateway equipment with compatible and adaptive parameters to finish the butting of the smoke and fire early warning supervision platform and the video equipment.
In this embodiment, the different video devices refer to national standard devices (cameras, platforms, NVR, etc.) and non-national standard (onvi, rtsp, rtmp, live devices, etc.).
In this embodiment, the terminal parameters refer to parameters such as device type parameters and working power of different video devices.
In this embodiment, the service standard refers to the working modes of different video devices, i.e. national standard and non-national standard.
In this embodiment, the communication protocol refers to a communication manner adopted by different video devices when data transmission is performed.
In this embodiment, the gateway device is preset in advance, and is used for connecting different video devices and intermediaries of the smoke and fire early warning supervision platform.
In this embodiment, the unified interface standard refers to that data transmitted from different video devices to the pyrotechnic early warning supervision platform is subjected to data unified operation, so that transmission and management are facilitated, that is, data in different formats is received by adopting the unified standard.
The technical scheme has the advantages that different video devices are subjected to differential classification according to terminal parameters of the different video devices, so that parameter compatibility adaptation is conveniently carried out on preset gateway devices according to different types of video devices, secondly, the different types of video devices are conveniently docked with the preset gateway devices according to differential classification results, so that the different video devices are conveniently docked with the smoke and fire early warning supervision platform through the preset gateway devices, finally, after the interface standard of the data interface of the smoke and fire early warning supervision platform is unified, the different video devices are effectively docked with the smoke and fire early warning supervision platform through the preset gateway devices, and accordingly, real-time video stream data monitored by the different video devices are conveniently received and analyzed through the smoke and fire early warning supervision platform, and the instantaneity of smoke and fire analysis is guaranteed.
Example 3:
on the basis of embodiment 1, this embodiment provides a smoke and fire detection early warning method based on AI video analysis, in S1, the smoke and fire early warning supervision platform is docked with the video device, including:
extracting terminal monitoring areas of different video devices based on the docking result, and determining identity tags of video streams of the different video devices in a smoke and fire early warning supervision platform based on environmental characteristics of the terminal monitoring areas;
And performing personalized division on the cache space in the smoke and fire early warning supervision platform based on the terminal quantity of the video equipment and the coverage area of the corresponding terminal monitoring area to obtain a personalized cache space, and packaging the identity tag and the corresponding personalized cache space.
In this embodiment, the environmental characteristics refer to buildings, distribution of articles, and the like included in the terminal monitoring area.
In this embodiment, the identity tag refers to a tag capable of characterizing the terminal monitoring area corresponding to different video streams.
In this embodiment, personalized division refers to splitting a buffer space in a smoke and fire early warning supervision platform according to the number of video devices and the data amount of each terminal monitoring area and the monitored video stream, so that video streams monitored by different video devices can be buffered in corresponding buffer spaces, wherein the personalized buffer space is a result obtained after personalized division of the buffer spaces.
The technical scheme has the beneficial effects that the identity tags corresponding to different video streams are determined through determining the terminal monitoring areas corresponding to different video devices and according to the environmental characteristics of the terminal monitoring areas, the buffer space in the smoke and fire early warning supervision platform is individually divided according to the terminal quantity of the video devices and the coverage area of the corresponding terminal monitoring areas, and finally the obtained identity tags and the corresponding individual buffer space are packaged, so that the video streams monitored by the different video devices can be effectively stored, and the smoke and fire conditions of the different terminal monitoring areas can be effectively analyzed in real time.
Example 4:
on the basis of embodiment 1, the present embodiment provides a smoke and fire detection early warning method based on AI video analysis, in S1, video frame extraction is performed on a real-time video stream in a video device based on a docking result, including:
reading a real-time video stream in the video device based on the docking result, and determining an image frame contained in the real-time video stream based on the reading result;
extracting corresponding image content from the real-time video stream frame by frame based on the image frames, obtaining each frame of image data based on the image content, and obtaining a frame image sequence based on each frame of image data;
And determining a polling frame extraction mechanism for the frame image sequence based on the video frame extraction index, and performing video frame extraction on a real-time video stream in the video device based on the polling frame extraction mechanism.
In this embodiment, an image frame refers to a plurality of still images contained in a real-time video stream.
In this embodiment, each frame of image data refers to specific recording object or the like information contained in each image frame.
In this embodiment, the frame image sequence refers to a result of splitting the real-time video stream into a plurality of consecutive image frames.
In this embodiment, the video frame extraction index is known in advance, and is used to characterize the standard that needs to be used in video frame extraction, for example, it may be that the frame is extracted randomly every time the frame is extracted, or extracted 30 seconds from the first frame.
In this embodiment, the polling frame-extracting mechanism refers to a rule for performing video frame extraction on the real-time video stream, which is formulated according to the video frame-extracting index.
The technical scheme has the advantages that the real-time video stream in the video equipment is read, the image frames contained in the real-time video stream are determined, the image content in each image frame is extracted, the accurate and effective determination of the frame image sequence is realized, and finally, the video frame extraction is realized on the real-time video stream according to the formulated polling frame extraction mechanism, so that the smoke and fire preliminary detection is facilitated, and the efficiency of the smoke and fire preliminary detection is improved.
Example 5:
On the basis of embodiment 4, the present embodiment provides a smoke and fire detection early warning method based on AI video analysis, and obtains a frame image sequence based on each frame image data, including:
Acquiring each frame of image data extracted frame by frame, and carrying out uniform block division on each frame of image data based on a reference division size;
Training a preset training image based on a multi-scale evaluation index, constructing a support vector regression model, and analyzing the image data of each block based on the support vector regression model to obtain a feature vector of each image block;
determining a quality evaluation value of each image block based on the feature vector, determining an image quality loss degree of each frame image based on the quality evaluation value of each image block, and judging the frame image with the image quality loss degree larger than a preset threshold value as a distorted frame image;
And determining similar structural features based on the adjacent frame image data, and repairing the distorted frame image based on the similar structural features.
In this embodiment, the reference division size is set in advance for dividing the frame image data.
In this embodiment, the multi-scale evaluation index is set in advance, is a basis for evaluating the quality of the frame image, and may be, for example, sharpness or the like, and is not unique.
In this embodiment, the preset training image is known in advance.
In this embodiment, the support vector regression model is constructed after training a preset training image according to a multi-scale evaluation index, and is used for extracting feature information of image data, where the feature vector is image feature information contained in an image block obtained by analyzing the image block through the support vector regression model.
In this embodiment, the quality evaluation value refers to a parameter capable of characterizing whether the image block is good or bad, and the higher the evaluation value is, the better the quality of the image block is.
In this embodiment, the preset threshold is set in advance, is a criterion for measuring whether the image block meets the minimum requirement, and can be modified.
In this embodiment, the similar structural features refer to the similarities existing between the image data of the adjacent frames, including the recording object, the image composition, and the like.
The technical scheme has the advantages that the frame image data are uniformly partitioned, the preset training image is trained through the multi-scale evaluation index, the support vector regression model is accurately and effectively constructed, the constructed support vector regression model is used for analyzing the image data of each block, the quality evaluation value of each image block is accurately and effectively determined, finally, the distorted frame image existing in the frame image is locked through the quality evaluation value, the distorted frame image is repaired through the similar structural characteristics in the adjacent frame image data, the accuracy and the reliability of the obtained frame image are ensured, and convenience and guarantee are provided for improving the smoke and fire detection accuracy.
Example 6:
On the basis of embodiment 1, the present embodiment provides a smoke and fire detection early warning method based on AI video analysis, in S2, based on a video frame extraction result, each time, one frame of image frame is transmitted to a video analysis model based on a communication interface to perform preliminary smoke and fire detection, and when smoke and fire exist, a target number of image frame sets are re-transmitted to the video analysis model based on the communication interface to perform re-detection, including:
acquiring a frame of image frame corresponding to a video frame extraction result, and determining a data transmission rule based on a communication protocol of a communication interface;
carrying out semantic extraction on the data transmission rule, determining a grammar execution command in the data transmission rule based on the extracted semantic, and carrying out format conversion on a frame of image frame based on the grammar execution command to obtain transmissible image data;
transmitting transmissible image data to a video analysis model based on a communication interface, and dividing a frame of image frame into M sampling small images based on a transmission result;
Respectively extracting the three-primary-color gray average value of each sampling small image in a color space based on the RGB color channels, and carrying out weight summarization on the three-primary-color gray average value to obtain a color histogram corresponding to one frame of image frame;
Converting the color histogram into a low-dimensional color histogram, quantizing each color component in the low-dimensional color histogram to a normalization range, and obtaining the color duty ratio in one frame of image frame based on the quantization result;
Determining the membership degree of the color ratio relative to the reference firework color characteristic, and judging that the firework exists when the membership degree is higher than a preset membership degree threshold value;
And re-transmitting the image frame set of the target number to a video analysis model for re-inspection based on the communication interface based on the judgment result.
In this embodiment, the communication protocol refers to a data transmission mode corresponding to the communication interface, where the data transmission rule is a data transmission mode and is determined according to the communication protocol of the communication interface.
In this embodiment, the grammar execution command refers to specific requirement parameters corresponding to data transmission in the determined data transmission rule after semantic analysis is performed on the data transmission rule, including a data format and the like.
In this embodiment, the transmissible image data refers to a result obtained by format-converting one frame of image frame by a syntax-execution command.
In this embodiment, the sampling thumbnail refers to a plurality of image blocks obtained by dividing an image frame extracted from a video.
In this embodiment, the RGB color channels refer to three primary colors of red, green and blue, and the average value of the gray levels of the three primary colors of each sampling plot is determined through the RGB color channels, so as to determine a color histogram corresponding to one frame of image frame.
In this embodiment, the color histogram is a state and distribution that characterizes the presence of image colors in the sample plot, thereby facilitating a determination of whether a smoke is present based on the color histogram, wherein the smoke includes smoke and fire.
In this embodiment, the low-dimensional color histogram refers to the color histogram being subjected to a dimensional transformation in order to more accurately and intuitively determine the color condition contained in the image.
In this embodiment, each color component refers to the content of each color in one image frame.
In this embodiment, the reference pyrotechnic color characteristics are known in advance for characterizing the pyrotechnic corresponding color conditions.
In this embodiment, the membership is used to characterize the amount of likelihood that a smoke is contained in an image frame determined according to the color duty cycle, and a higher membership indicates a higher likelihood of smoke being present.
In this embodiment, the preset membership threshold is set in advance.
In this embodiment, the normalization range is set in advance, and is used to normalize and unify the quantization of each color component, so as to facilitate determination of the duty ratio of different colors.
The technical scheme has the advantages that the communication protocol of the communication interface is analyzed, the data transmission rule corresponding to the communication interface is accurately and effectively determined, the data transmission rule is subjected to semantic extraction, the grammar execution command in the data transmission rule is locked according to the extraction result, the format conversion is carried out on a frame of image frame through the grammar execution command, the frame of image frame obtained by video frame extraction is accurately and reliably transmitted to the video analysis model, finally, the color ratio of each color in the frame of image frame is locked according to the color histogram through segmentation and color histogram analysis, the locked color ratio is compared with the color characteristics of a reference smoke and fire, the accurate and reliable preliminary detection of smoke and fire is realized, the accuracy of the preliminary detection of smoke and fire is ensured, and meanwhile, when the smoke and fire are judged to exist, the timeliness and the reliability of smoke and fire detection are timely rechecked.
Example 7:
On the basis of embodiment 6, the present embodiment provides a smoke and fire detection early warning method based on AI video analysis, and based on a determination result, the method includes that a set of image frames with a target number is re-transmitted to a video analysis model for re-detection based on a communication interface, including:
locking the source position of one frame of image frame corresponding to the video frame extraction result based on the judgment result, and locking a known number of forward and backward video frames in the real-time video stream by taking the one frame of image frame corresponding to the video frame extraction result as the center based on the source position, so as to obtain a target number of image frame sets;
Retransmitting the image frame set of the target number to a video analysis model based on a communication interface, and carrying out bit-by-bit classification on the image frames of the target number based on a retransmission result to obtain inter-frame differential parameters;
determining a motion area between image frame sets of a target number based on the inter-frame differential parameters, and determining a spatial span of a preliminary smoke and fire detection result in the image frame sets based on the motion area;
When the space span is larger than a preset requirement, calibrating a smoke and fire area in each image frame in the image frame set based on a preliminary smoke and fire detection result, and performing edge monitoring on smoke and fire in the smoke and fire area based on the calibration result to obtain a smoke and fire profile;
Determining smoke morphology features based on the smoke profile, and performing a recheck of the target number of image frame sets based on the smoke features.
In this embodiment, the source position refers to a specific position of a frame of image corresponding to the video frame extraction result in the original video stream.
In this embodiment, performing the known number of forward and backward video frame locks based on the source position in the real-time video stream with respect to the one frame image frame corresponding to the video frame extraction result as a center refers to selecting a corresponding number of multiple image frames in the real-time video stream according to the one frame image frame corresponding to the video frame extraction result, including the image frame before the video frame extraction result and the image frame after the video frame extraction result.
In this embodiment, the target number is the same as the known number, and is set in advance.
In this embodiment, the inter-frame difference parameter refers to an image difference existing between adjacent image frames in the set of image frames of the target number.
In this embodiment, the motion area refers to a monitoring area corresponding to a set of image frames of the target number.
In this embodiment, the spatial span refers to the proportion of the preliminary smoke detection result in the image frame set, that is, specifically, several image frames contain the preliminary smoke detection result.
In this embodiment, the preset requirements are set in advance.
In this embodiment, the pyrotechnic contours refer to the shape in which the pyrotechnic is present in the image frame.
In this embodiment, the pyrotechnic morphology features refer to the specific morphology of the pyrotechnic presence determined from the pyrotechnic profile.
In this embodiment, the completion of the re-inspection of the target number of image frame sets based on the pyrotechnic features means that the obtained pyrotechnic morphology features are compared with the standard pyrotechnic morphology, and if the two meet a preset similarity criterion, it is determined that there is indeed a pyrotechnic.
The technical scheme has the advantages that the image frames with the target number are locked according to the source position of one frame of image frame corresponding to the video frame extraction result, the locked image frame set with the target number is repeatedly transmitted to the video analysis model for analysis, the space span of the preliminary smoke and fire detection result in the image frame set and the smoke and fire outline are locked, finally, the smoke and fire morphological characteristics contained in the image frame set are determined according to the smoke and fire outline, the accurate and reliable re-detection of smoke and fire according to the smoke and fire morphological characteristics is realized, the accuracy and reliability of the re-detection are ensured, and repeated false alarm is avoided.
Example 8:
On the basis of embodiment 1, the present embodiment provides a smoke and fire detection early warning method based on AI video analysis, in S3, when determining that smoke and fire exists based on the rechecking result, displaying a real-time warning page based on a smoke and fire early warning supervision platform, including:
acquiring a rechecking result, and determining a video monitoring area corresponding to the real-time video stream based on the image frame set when the rechecking result judges that smoke exists;
extracting environmental characteristics of the video monitoring area, determining a smoke and fire position based on the environmental characteristics, determining a smoke and fire severity level based on the re-detection result, and generating smoke and fire early warning information based on the smoke and fire position and the smoke and fire severity level;
And feeding back the smoke and fire early warning information to a smoke and fire early warning supervision platform, and generating a smoke and fire early warning page based on the smoke and fire early warning information at the smoke and fire early warning supervision platform for smoke and fire early warning display.
In this embodiment, pyrotechnic location refers to the specific location at which a pyrotechnic event occurs.
The technical scheme has the advantages that the video monitoring area where the smoke and fire occur is determined according to the rechecking result, the smoke and fire position is accurately and effectively locked according to the environmental characteristics of the video monitoring area, the smoke and fire serious grade is determined according to the rechecking result, corresponding smoke and fire early warning information is generated according to the smoke and fire serious grade and the smoke and fire position, and finally the smoke and fire early warning information is fed back to the smoke and fire early warning supervision platform, so that the smoke and fire early warning is accurately and effectively displayed.
Example 9:
On the basis of embodiment 1, the present embodiment provides a smoke and fire detection early warning method based on AI video analysis, in S3, the method synchronously sends smoke and fire early warning notification to a management terminal, including:
Acquiring a communication address of a management terminal and a terminal address of a smoke and fire early warning supervision platform, and generating a distributed transmission link based on the communication address and the terminal address;
and synchronously transmitting the smoke and fire early warning information to the management terminal based on the distributed transmission link to carry out smoke and fire early warning notification.
The technical scheme has the beneficial effects that the distributed transmission link is constructed according to the communication address of the management terminal and the terminal address of the smoke fire early warning supervision platform, and smoke fire early warning information is synchronously sent to the management terminal through the distributed transmission link to carry out smoke fire early warning notification, so that the management terminal can conveniently know whether smoke fire occurs in time, and the timeliness of smoke fire discovery is improved.
Example 10:
the embodiment provides a smoke and fire detection early warning system based on AI video analysis, as shown in FIG. 3, including:
the video frame extraction module is used for butting the smoke and fire early warning supervision platform with the video equipment and carrying out video frame extraction on a real-time video stream in the video equipment based on a butting result;
The smoke and fire detection module is used for transmitting one frame of image frame to the video analysis model based on the communication interface each time based on the video frame extraction result to perform preliminary smoke and fire detection, and retransmitting the image frame set with the target number to the video analysis model based on the communication interface to perform rechecking when smoke and fire exist;
and the early warning module is used for determining that when smoke exists based on the rechecking result, displaying a real-time warning page based on the smoke early warning supervision platform and synchronously sending smoke early warning notification to the management terminal.
The technical scheme has the advantages that the video equipment is in butt joint with the smoke and fire early warning platform, so that real-time video streams of a monitoring area acquired by the video equipment are conveniently transmitted to the smoke and fire early warning platform for real-time analysis, secondly, video sampling inspection is carried out on the real-time video streams, one frame of image frames are obtained and transmitted to the video analysis model for preliminary smoke and fire detection, the efficiency of preliminary smoke and fire detection is improved, meanwhile, when the preliminary smoke and fire detection judges that smoke and fire exist, rechecking is carried out on an image frame set of locking target data in the real-time video streams, the accuracy and precision of smoke and fire detection are improved, finally, smoke and fire early warning notification is carried out when the smoke and fire exists according to rechecking results, the accuracy of smoke and fire early warning is guaranteed, and repeated false alarm is avoided.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, it is intended that the present invention also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.
Claims (5)
1. A smoke and fire detection early warning method based on AI video analysis is characterized by comprising the following steps:
S1, butting a smoke and fire early warning supervision platform with video equipment, and performing video frame extraction on a real-time video stream in the video equipment based on a butting result;
S2, transmitting one frame of image frame to a video analysis model based on a communication interface each time based on a video frame extraction result to perform preliminary smoke and fire detection, and re-transmitting the image frame set with the target number to the video analysis model based on the communication interface to perform re-detection when smoke and fire exist;
S3, determining that when smoke exists based on the rechecking result, displaying a real-time warning page based on a smoke early warning supervision platform, and synchronously sending smoke early warning notification to a management terminal;
in S2, based on the video frame extraction result, transmitting one frame of image frame to the video analysis model based on the communication interface for preliminary smoke and fire detection, and when smoke and fire exist, retransmitting the image frame set with the target number to the video analysis model based on the communication interface for rechecking, including:
acquiring a frame of image frame corresponding to a video frame extraction result, and determining a data transmission rule based on a communication protocol of a communication interface;
carrying out semantic extraction on the data transmission rule, determining a grammar execution command in the data transmission rule based on the extracted semantic, and carrying out format conversion on a frame of image frame based on the grammar execution command to obtain transmissible image data;
transmitting transmissible image data to a video analysis model based on a communication interface, and dividing a frame of image frame into M sampling small images based on a transmission result;
Respectively extracting the three-primary-color gray average value of each sampling small image in a color space based on the RGB color channels, and carrying out weight summarization on the three-primary-color gray average value to obtain a color histogram corresponding to one frame of image frame;
Converting the color histogram into a low-dimensional color histogram, quantizing each color component in the low-dimensional color histogram to a normalization range, and obtaining the color duty ratio in one frame of image frame based on the quantization result;
Determining the membership degree of the color ratio relative to the reference firework color characteristic, and judging that the firework exists when the membership degree is higher than a preset membership degree threshold value;
And re-transmitting the image frame set of the target number to a video analysis model for re-inspection based on the communication interface based on the judgment result.
2. The smoke and fire detection and early warning method based on AI video analysis of claim 1, wherein in S1, interfacing the smoke and fire early warning supervision platform with the video device comprises:
acquiring terminal parameters of different video devices, and determining service standards and communication protocols of the different video devices based on the terminal parameters;
Performing differential classification on different video devices based on a service standard, performing parameter compatibility adaptation on preset gateway devices based on a differential classification result according to a communication protocol, and docking different types of video devices with the preset gateway devices based on the parameter compatibility adaptation;
Meanwhile, the data interfaces of the smoke and fire early warning supervision platform are subjected to interface standard unification, and the data interfaces with the unified interface standard are in butt joint with the preset gateway equipment with the compatible and matched parameters, so that the butt joint of the smoke and fire early warning supervision platform and the video equipment is completed.
3. The smoke and fire detection and early warning method based on AI video analysis of claim 1, wherein in S1, video frame extraction is performed on a real-time video stream in a video device based on a docking result, comprising:
reading a real-time video stream in the video device based on the docking result, and determining an image frame contained in the real-time video stream based on the reading result;
extracting corresponding image content from the real-time video stream frame by frame based on the image frames, obtaining each frame of image data based on the image content, and obtaining a frame image sequence based on each frame of image data;
And determining a polling frame extraction mechanism for the frame image sequence based on the video frame extraction index, and performing video frame extraction on a real-time video stream in the video device based on the polling frame extraction mechanism.
4. The smoke and fire detection early warning method based on AI video analysis according to claim 1, wherein the step of re-transmitting the image frame set of the target number to the video analysis model for re-detection based on the communication interface based on the determination result comprises the steps of:
locking the source position of one frame of image frame corresponding to the video frame extraction result based on the judgment result, and locking a known number of forward and backward video frames in the real-time video stream by taking the one frame of image frame corresponding to the video frame extraction result as the center based on the source position, so as to obtain a target number of image frame sets;
Retransmitting the image frame set of the target number to a video analysis model based on a communication interface, and carrying out bit-by-bit classification on the image frames of the target number based on a retransmission result to obtain inter-frame differential parameters;
determining a motion area between image frame sets of a target number based on the inter-frame differential parameters, and determining a spatial span of a preliminary smoke and fire detection result in the image frame sets based on the motion area;
When the space span is larger than a preset requirement, calibrating a smoke and fire area in each image frame in the image frame set based on a preliminary smoke and fire detection result, and performing edge monitoring on smoke and fire in the smoke and fire area based on the calibration result to obtain a smoke and fire profile;
Determining smoke morphology features based on the smoke profile, and performing a recheck of the target number of image frame sets based on the smoke features.
5. A smoke and fire detection early warning system based on AI video analysis, characterized by comprising:
the video frame extraction module is used for butting the smoke and fire early warning supervision platform with the video equipment and carrying out video frame extraction on a real-time video stream in the video equipment based on a butting result;
The smoke and fire detection module is used for transmitting one frame of image frame to the video analysis model based on the communication interface each time based on the video frame extraction result to perform preliminary smoke and fire detection, and retransmitting the image frame set with the target number to the video analysis model based on the communication interface to perform rechecking when smoke and fire exist;
The early warning module is used for determining that when smoke exists based on the rechecking result, displaying a real-time warning page based on the smoke early warning supervision platform and synchronously sending smoke early warning notification to the management terminal;
Wherein, smoke and fire detection module includes:
acquiring a frame of image frame corresponding to a video frame extraction result, and determining a data transmission rule based on a communication protocol of a communication interface;
carrying out semantic extraction on the data transmission rule, determining a grammar execution command in the data transmission rule based on the extracted semantic, and carrying out format conversion on a frame of image frame based on the grammar execution command to obtain transmissible image data;
transmitting transmissible image data to a video analysis model based on a communication interface, and dividing a frame of image frame into M sampling small images based on a transmission result;
Respectively extracting the three-primary-color gray average value of each sampling small image in a color space based on the RGB color channels, and carrying out weight summarization on the three-primary-color gray average value to obtain a color histogram corresponding to one frame of image frame;
Converting the color histogram into a low-dimensional color histogram, quantizing each color component in the low-dimensional color histogram to a normalization range, and obtaining the color duty ratio in one frame of image frame based on the quantization result;
Determining the membership degree of the color ratio relative to the reference firework color characteristic, and judging that the firework exists when the membership degree is higher than a preset membership degree threshold value;
And re-transmitting the image frame set of the target number to a video analysis model for re-inspection based on the communication interface based on the judgment result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410389130.2A CN118397779B (en) | 2024-04-02 | 2024-04-02 | Smoke and fire detection early warning method and system based on AI video analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410389130.2A CN118397779B (en) | 2024-04-02 | 2024-04-02 | Smoke and fire detection early warning method and system based on AI video analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118397779A CN118397779A (en) | 2024-07-26 |
| CN118397779B true CN118397779B (en) | 2024-12-10 |
Family
ID=91985921
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410389130.2A Active CN118397779B (en) | 2024-04-02 | 2024-04-02 | Smoke and fire detection early warning method and system based on AI video analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118397779B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110084166A (en) * | 2019-04-19 | 2019-08-02 | 山东大学 | Substation's smoke and fire intelligent based on deep learning identifies monitoring method |
| CN211702069U (en) * | 2020-04-30 | 2020-10-16 | 浙江省机电设计研究院有限公司 | Internet of things terminal with multi-protocol adaptation |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7046842B2 (en) * | 1999-08-17 | 2006-05-16 | National Instruments Corporation | System and method for color characterization using fuzzy pixel classification with application in color matching and color match location |
| JP4345648B2 (en) * | 2004-11-24 | 2009-10-14 | ノーリツ鋼機株式会社 | Photographic image processing method and apparatus |
| KR101460440B1 (en) * | 2013-04-26 | 2014-11-12 | 최두현 | Camera with fire detection function, and fire detection method thereof |
| KR101649593B1 (en) * | 2015-12-18 | 2016-08-30 | 주식회사 싸인텔레콤 | The appartus and method for tunnel smoke detection based on vehicle fire |
| CN113783749B (en) * | 2021-08-11 | 2022-11-01 | 太原市高远时代科技有限公司 | Method for detecting equipment fault based on network port |
-
2024
- 2024-04-02 CN CN202410389130.2A patent/CN118397779B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110084166A (en) * | 2019-04-19 | 2019-08-02 | 山东大学 | Substation's smoke and fire intelligent based on deep learning identifies monitoring method |
| CN211702069U (en) * | 2020-04-30 | 2020-10-16 | 浙江省机电设计研究院有限公司 | Internet of things terminal with multi-protocol adaptation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118397779A (en) | 2024-07-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR102121958B1 (en) | Method, system and computer program for providing defect analysis service of concrete structure | |
| CN110096945B (en) | Indoor monitoring video key frame real-time extraction method based on machine learning | |
| CN117152161A (en) | Shaving board quality detection method and system based on image recognition | |
| CN113642474A (en) | Hazardous area personnel monitoring method based on YOLOV5 | |
| CN111222478A (en) | Construction site safety protection detection method and system | |
| CN116862712A (en) | Electric power construction potential safety risk detection method and system based on thunder fusion | |
| CN118747988A (en) | Liquid crystal display function detection method, device, equipment and storage medium | |
| CN118135480A (en) | Visual image processing method and system for electromechanical construction of tunnel | |
| CN113743235B (en) | Electric power inspection image processing method, device and equipment based on edge calculation | |
| CN118506273A (en) | Risk feature quantity twin method and system based on power monitoring video stream | |
| CN112712514A (en) | Product detection system and auxiliary detection method | |
| CN115049955A (en) | Fire detection analysis method and device based on video analysis technology | |
| CN118912528A (en) | Intelligent flame monitoring system and method based on industrial vision technology | |
| CN117670867A (en) | Industrial defect detection method, device, electronic equipment and storage medium | |
| CN118397779B (en) | Smoke and fire detection early warning method and system based on AI video analysis | |
| CN114820461B (en) | Surface defect quality inspection platform and method based on image recognition algorithm | |
| CN118711188B (en) | A method and system for recognizing steel stamp characters | |
| CN113673406B (en) | Curtain wall glass burst detection method, system, electronic equipment and storage medium | |
| CN115565171A (en) | Method and device for associated labeling of security inspection images | |
| CN115713758B (en) | Carriage identification method, system, device and storage medium | |
| CN112001237A (en) | Video monitoring and broadcasting automatic testing method and device | |
| CN117575382A (en) | Automatic intelligent image inspection quality control management system | |
| CN116704440A (en) | Intelligent comprehensive acquisition and analysis system based on big data | |
| CN114429421A (en) | A Knife Gate Status Detection Method Oriented to Scheduling Business Scenarios | |
| CN114630112B (en) | Video playing test method, device and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |