CN116453064B - Method for identifying abnormal road conditions of tunnel road section based on monitoring data - Google Patents
Method for identifying abnormal road conditions of tunnel road section based on monitoring data Download PDFInfo
- Publication number
- CN116453064B CN116453064B CN202310714673.2A CN202310714673A CN116453064B CN 116453064 B CN116453064 B CN 116453064B CN 202310714673 A CN202310714673 A CN 202310714673A CN 116453064 B CN116453064 B CN 116453064B
- Authority
- CN
- China
- Prior art keywords
- smoke
- video frame
- image
- area
- value
- 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
- 238000000034 method Methods 0.000 title claims abstract description 35
- 238000012544 monitoring process Methods 0.000 title claims abstract description 26
- 230000002159 abnormal effect Effects 0.000 title claims abstract description 17
- 239000000779 smoke Substances 0.000 claims abstract description 239
- 238000001514 detection method Methods 0.000 claims abstract description 56
- 238000012545 processing Methods 0.000 claims abstract description 7
- 230000033001 locomotion Effects 0.000 claims description 42
- 239000013598 vector Substances 0.000 claims description 38
- 238000012216 screening Methods 0.000 claims description 14
- 238000002834 transmittance Methods 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 8
- 230000005856 abnormality Effects 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 4
- 239000000428 dust Substances 0.000 description 36
- 230000008859 change Effects 0.000 description 15
- 239000004973 liquid crystal related substance Substances 0.000 description 4
- 238000009792 diffusion process Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 206010033799 Paralysis Diseases 0.000 description 1
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/588—Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of data processing, in particular to a method for identifying road condition anomalies of a tunnel road section based on monitoring data, which comprises the following steps: acquiring a region in which smoke possibly exists in a video frame image through an identification image, analyzing and processing characteristic data of the suspected smoke region, obtaining a first possibility index of the suspected smoke region according to the distance between a pixel point on the edge of the suspected smoke region and a detection point and a first coefficient, and obtaining a second possibility index according to the difference of the area of the suspected smoke region between adjacent frame images and a second coefficient; and combining the first possibility index and the second possibility index to obtain a global possibility index, obtaining an actual smoke area, and further obtaining the abnormal condition of the road condition of the tunnel road. The invention adopts a recognition mode and carries out related data processing, so that the accuracy of actual smoke recognition can be improved, and further the accurate fire disaster abnormal condition of the road condition of the tunnel road can be obtained.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a method for identifying road condition anomalies of a tunnel road section based on monitoring data.
Background
The fire hazard of the highway tunnel is extremely high, and once the fire hazard happens, traffic paralysis, commercial and trade logistics interruption and even group death and group injury are easily caused, so that bad social response is caused. Therefore, it is important to identify the fire abnormality of the highway tunnel. At present, the purpose of identifying fire abnormality of a tunnel is generally achieved by detecting smoke conditions in a highway tunnel. When detecting the smoke condition in the tunnel, the existing method obtains the smoke condition in the tunnel through the smoke sensor, but the smoke sensor is limited by the installation position of the smoke sensor greatly, and the space in the tunnel is large, so that the more accurate smoke condition cannot be obtained.
At present, along with the continuous development of monitoring technology, the smoke condition in the tunnel is obtained by using tunnel monitoring data. However, when detecting the smoke condition in the tunnel monitoring video image, the vehicle running in the tunnel is easy to generate dust, and the actual smoke condition is influenced, so that the result of identifying the fire abnormality of the tunnel is less accurate.
Disclosure of Invention
In order to solve the technical problem that the result of fire disaster anomaly identification on a tunnel is inaccurate, the invention aims to provide a tunnel road section road condition anomaly identification method based on monitoring data, and the adopted technical scheme is as follows:
Acquiring at least two frames of video frame images in the video monitoring data of the tunnel section, and screening each frame of video frame image to obtain a smoke suspicious image; extracting a suspected smoke area in a smoke suspected image, and determining a detection point according to the running condition of a vehicle in the smoke suspected image;
calculating a first coefficient according to the corresponding moving direction of the suspected smoke area, and obtaining a first possibility index of the suspected smoke area according to the distance between the pixel points on the edge of the suspected smoke area and the detection points and the first coefficient;
obtaining a second coefficient according to the difference between the moving direction corresponding to the suspected smoke area and the set direction, and obtaining a second possibility index according to the difference between the areas of the suspected smoke areas and the adjacent frame images and the second coefficient;
and determining a global possibility index of the suspected smoke area according to the first possibility index and the second possibility index, screening the suspected smoke area according to the global possibility index to obtain an actual smoke area, and further obtaining the abnormal condition of the road condition of the tunnel road section.
Preferably, the method for obtaining the first likelihood index specifically includes:
for any one smoke suspicious image, acquiring a continuous preset number Zhang Shipin of images by taking the smoke suspicious image as a starting frame image to form a first characteristic image set of the smoke suspicious image; acquiring a vehicle running path according to the video frame image, and marking two detection points symmetrical about the vehicle running path as detection point pairs;
Recording any one smoke suspicious region in the smoke suspicious images as a target region, and acquiring target regions at positions corresponding to all video frame images in the first characteristic image set; extracting a vehicle region on a video frame image in a first characteristic image set;
for any one target area, calculating the distance between the vehicle areas on the set number of frames of video frame images after the smoke suspicious image, and recording the vehicle corresponding to the vehicle area with the minimum distance as a target vehicle, wherein the vehicle running path corresponding to the target vehicle is a target path;
on any video frame image of the first characteristic image set, for any detection point pair of the target path, calculating the average value of the distance between the edge pixel point of the target area and the detection point, taking the inverse of the sum of the absolute value of the difference value of the two average values and a preset first numerical value as the characteristic value of the detection point pair, and calculating the average value of the characteristic values of all the detection point pairs on the target path to obtain the characteristic average value;
and calculating the product of the average number of the feature mean values corresponding to all the video frame images in the first feature image set and the first coefficient to obtain a first possibility index of the target area.
Preferably, the method for obtaining the first coefficient specifically includes:
for any two adjacent video frame images in the first characteristic image set, calculating the absolute value of the difference value of the areas of the target areas on the two video frame images; according to the two video frame images, obtaining a motion vector of a target area, according to a running direction corresponding to a target path of a vehicle, calculating a value of an included angle between a unit vector of the running direction and the motion vector, calculating the reciprocal of the sum of the value and a preset second value, and taking the product of the absolute value and the reciprocal of the difference of the areas as a first direction characteristic value of the target area on two adjacent video frame images; taking the average value of the first direction characteristic values of the target area on all any two adjacent video frame images in the first characteristic image set as a first coefficient.
Preferably, the obtaining a second coefficient according to the difference between the moving direction and the set direction corresponding to the suspected smoke area, and obtaining a second likelihood indicator according to the difference between the areas of the suspected smoke areas and the adjacent frame images and the second coefficient specifically includes:
for any two adjacent video frame images in the first characteristic image set, acquiring a value of an included angle between a motion vector of a target area obtained according to the two video frame images and a unit vector of a set direction, and taking the reciprocal of the sum of the value and a preset third value as a second direction characteristic value of the target area on the two adjacent video frame images; taking the average value of the second direction characteristic values of the target area on all any two adjacent video frame images in the first characteristic image set as a second coefficient;
For any video frame image in the first characteristic image set, acquiring a video frame image of a target vehicle corresponding to a target area, marking the video frame image as a selected frame image, and acquiring a certain number of Zhang Shipin frame images by taking the selected frame image as a starting frame, marking the video frame image as a second characteristic image set;
calculating the absolute value of the difference between the area of the target area on the video frame image and the area of the target area on the video frame image in the second characteristic image set, and calculating the reciprocal of the sum of the absolute value of the difference and a preset fourth numerical value; and obtaining the reciprocal of the time difference between the video frame image and the video frame image in the second characteristic image set, and obtaining a second possibility index of the target area according to the two reciprocal and the second coefficient.
Preferably, the calculation formula of the second likelihood index is specifically:
where PB denotes a second likelihood indicator for the target region, N0 denotes the total number of images in the first set of feature images,representing the motion vectors corresponding to the j-1 th frame video frame image of the target area in the first characteristic image set, Y representing the unit vector in which the set direction is located, and +.>Representing a second direction characteristic value of the target area on two adjacent video frame images; n1 represents the total number of images in the second set of feature images,/- >Area of target area in j-th frame video frame image in first characteristic image set, +.>Representing the area of the target area in the video frame image of the e-th frame in the second set of feature images,/->Representing the difference in time between the image of the j-th frame video frame and the image of the e-th frame video frame,is a second coefficient>Wherein 1 is a third value, +.>And 1 in (2) is a fourth value.
Preferably, the determining the global likelihood indicator of the suspected smoke area according to the first likelihood indicator and the second likelihood indicator specifically includes:
and calculating a difference value between the preset fifth numerical value and the normalized value of the first likelihood index, wherein the product of the difference value and the normalized value of the second likelihood index is the global likelihood index.
Preferably, the screening the suspected smoke area according to the global likelihood index to obtain the actual smoke area specifically includes:
and (3) recording a suspected smoke area corresponding to the global possibility index being larger than a preset possibility threshold as an actual smoke area.
Preferably, the filtering the video frame image of each frame to obtain the suspicious image of the smoke specifically includes:
for any frame of video frame image, acquiring the light transmittance of the video frame image in the process of processing the video frame image by using a dark channel defogging algorithm, and if the light transmittance is smaller than a preset light transmittance threshold value, marking the video frame image as a smoke suspicious image; and screening each frame of video frame image to obtain all the smoke suspicious images.
Preferably, the determining the detection point according to the running condition of the vehicle in the smoke suspicious image includes:
and acquiring a vehicle driving path according to the video frame image, and setting detection points at equal intervals at positions which are set at a distance from two sides of the vehicle driving path, so as to acquire detection points in the smoke suspicious image.
The embodiment of the invention has at least the following beneficial effects:
according to the method, an area with the possibility of existence of smoke in a video frame image is obtained through an identification image, further, characteristic data of the suspected smoke area are analyzed and processed, a first coefficient is calculated according to a moving direction corresponding to the suspected smoke area, the change characteristic of the suspected smoke area along the moving direction along the time is considered, a first possibility index of the suspected smoke area is obtained according to the distance between a pixel point on the edge of the suspected smoke area and a detection point and the first coefficient, and the position characteristic between the suspected smoke area and a running vehicle is reflected by utilizing distance data; obtaining a second coefficient according to the difference between the corresponding moving direction and the set direction of the suspected smoke area, taking the difference between the moving direction and the set direction of the suspected smoke area into consideration, obtaining a second possibility index according to the difference between the areas of the suspected smoke areas and the adjacent frame images and the second coefficient, and reflecting the area change characteristics of the suspected smoke areas along with the time by using the area data; by combining the first possibility index and the second possibility index, the change condition of the characteristic data of the suspected smoke area is considered from multiple aspects, and the actual smoke area is obtained.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flow chart of a method for identifying abnormal road conditions of a tunnel section based on monitoring data.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following detailed description is given below of a method for identifying abnormal road conditions of a tunnel section based on monitoring data according to the invention, which is provided by combining the accompanying drawings and the preferred embodiment. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following specifically describes a specific scheme of the method for identifying the road condition abnormality of the tunnel section based on the monitoring data provided by the invention with reference to the accompanying drawings.
Examples:
the main purpose of the invention is as follows: in the process of identifying the smoke situation in the tunnel, firstly, rough smoke identification is carried out on the video frame image to obtain a smoke suspicious image, and at the moment, dust generated by a running vehicle has a great influence on the identification of actual smoke, namely, the dust generated by the vehicle is easily identified as the actual smoke by mistake, so that a smoke identification result is inaccurate. Therefore, the invention eliminates the influence of dust on smoke identification by analyzing the characteristic conditions of dust and actual smoke, thereby obtaining an accurate identification result.
Referring to fig. 1, a method flowchart of a method for identifying abnormal road conditions of a tunnel road section based on monitoring data according to an embodiment of the present invention is shown, the method includes the following steps:
step one, acquiring at least two frames of video frame images in the video monitoring data of a tunnel road section, and screening each frame of video frame image to obtain a smoke suspicious image; and extracting a suspected smoke area in the smoke suspected image, and determining a detection point according to the running condition of the vehicle in the smoke suspected image.
Firstly, when the tunnel monitoring video is utilized to conduct tunnel fire disaster anomaly identification, an original monitoring video of a tunnel section needs to be obtained, a plurality of cameras need to be installed in the tunnel at the moment, pictures of the tunnel section are shot in real time, and the monitoring video pictures can cover the whole tunnel. After the multi-frame video frame image of the tunnel section monitoring video is obtained, video frame image data are transmitted to a video monitoring center, then the data in the video frame image are analyzed, and whether fire disaster occurs in the tunnel is judged through smoke recognition so as to obtain the abnormal condition of the tunnel section.
When the tunnel monitoring video data is used for identifying the smoke abnormal condition, the video frame images are mainly relied on, before smoke identification is carried out, the smoke condition in all the video frame images is required to be analyzed, and the video frame images possibly with smoke are screened out, so that the calculated amount is reduced. In general, smoke in a video frame image shows a fog feature, that is, there is a shielding to light in the video frame image, based on which analysis can be performed according to the light transmittance in the image to determine whether there is suspicious smoke in the video frame image.
Screening each frame of video frame image to obtain a smoke suspicious image, specifically, for any frame of video frame image, acquiring the light transmittance of the video frame image in the process of processing the video frame image by using a dark channel defogging algorithm, and if the light transmittance is smaller than a preset light transmittance threshold value, marking the video frame image as the smoke suspicious image; and screening each frame of video frame image to obtain all the smoke suspicious images.
The dark channel defogging algorithm is known calculation, and when the image is processed by the dark channel defogging algorithm, the transmissivity of the image can be obtained, and in the embodiment, the transmissivity of the video frame image is the transmissivity, so that whether suspicious smoke exists in the video frame image is represented. When the light transmittance of the video frame image is smaller than the light transmittance threshold, it is indicated that fog may exist in the video frame image, so the video frame image is marked as a smoke suspicious image, and in this embodiment, the light transmittance threshold has a value of 0.6, and an operator can set the video frame image according to actual situations.
Further, the suspicious image of the smoke is analyzed to obtain a region with a smoke part in the image, namely, a suspicious smoke region in the suspicious image of the smoke is extracted, and in the embodiment, the suspicious smoke region in the suspicious image of the smoke is extracted by using the BP neural network.
In an actual tunnel, dust generated by running of a vehicle may exist, and thus the recognition result of smoke is affected, so that it is necessary to analyze the characteristics of an area where smoke may exist. The difference between the smoke and the dust is mainly represented by the movement characteristics of the area, because when no large wind force is influenced in the tunnel, the smoke is generally diffused from the fire occurrence point to the periphery, but the dust is generated by the running of the vehicle, and the dust moves along with the movement of the vehicle, so that the smoke or the dust can be judged according to the movement characteristics of the suspected smoke area.
Based on the method, the running condition of the vehicle in the smoke suspicious image is analyzed, and the possibility that the suspicious smoke area is the area where dust is located is judged according to the relation between the suspicious smoke area and the running vehicle. And determining detection points according to the running condition of the vehicle in the smoke suspicious image, namely acquiring a running path of the vehicle according to the video frame image, and setting the detection points at equal intervals at positions which are set at a distance from two sides of the running path so as to obtain the detection points in the smoke suspicious image.
In this embodiment, firstly, a BP neural network is used to extract a vehicle region in a smoke suspicious image, and since dust is generally generated along with the vehicle after the vehicle runs, and the movement direction of the dust in the monitoring video is consistent with the running direction of the vehicle, at this time, the relationship between the suspicious smoke region and the vehicle running can be judged according to the continuous multi-frame video frame images before and after the vehicle runs, so as to obtain the possibility that the suspicious smoke region is the region where the dust is located.
And determining the running direction and the running path of the vehicle corresponding to the vehicle region in the continuous multi-frame video frame images, setting a detection position according to the running path of the vehicle, and judging the relation between the suspected smoke region and the running of the vehicle by utilizing the distance relation between the suspected smoke region and the detection position. The method for acquiring the traveling direction of the vehicle and the traveling path of the vehicle is a known technique, and the traveling path of the vehicle is a traveling track of the vehicle.
For example, for any vehicle region, if the vehicle running path corresponding to the vehicle region is a line segment, the straight line where the line segment is located is taken as a symmetry axis, detection points are set at equal intervals at positions where the distance k between the left side and the right side of the line segment is equal to the distance k from the line segment, the value of the distance k is set according to the road width in the tunnel, and an implementer can set according to a specific implementation scenario, in this embodiment, the value of k is 20, the value of the interval distance between adjacent detection points located on the same side is 40, and meanwhile, the detection points located at the corresponding positions on both sides of the vehicle running path are symmetrical with respect to the vehicle running path. The detection points at the left side of the vehicle driving path are respectively recorded as，/>The n-th left detection point is denoted by +.>，/>The nth right detection point is indicated, and the nth left detection point and the nth right detection point are symmetrical with respect to the vehicle travel path.
Calculating a first coefficient according to the movement direction corresponding to the suspected smoke area, and obtaining a first possibility index of the suspected smoke area according to the distance between the pixel points on the edge of the suspected smoke area and the detection points and the first coefficient.
First, it is to be noted that, in recognizing the smoke abnormality in the tunnel, there may be dust affecting the recognition result, and dust is generated due to the running of the vehicle, so that the dust is mainly distributed in the vicinity of the running path of the vehicle, and the dust is diffused in the running direction of the vehicle. According to the embodiment of the invention, the suspected smoke area is analyzed according to the movement characteristics of the dust, and the possibility that the suspected smoke area is the area where the dust is located is judged.
And analyzing a suspected smoke area in any one smoke suspicious image, wherein if the suspected smoke area is an area where dust is actually located, the suspected smoke area moves along with the movement of the vehicle in the video frame images of a plurality of continuous frames, and the movement direction of the suspected smoke area is relatively close to the running direction of the vehicle. Meanwhile, for a suspected smoke area, a suspected smoke area exists in the continuous multi-frame video frame images at the corresponding position, the suspected smoke areas are not identical, if the suspected smoke area in the suspected smoke image is the area where dust is located, the suspected smoke area in the corresponding position in the continuous multi-frame images moves along with the movement of a vehicle after the suspected smoke image, and the suspected smoke area can change to different degrees along with the running of the vehicle.
Based on the above, for any one smoke suspicious image, acquiring a continuous preset number Zhang Shipin of images by taking the smoke suspicious image as a starting frame image, and forming a first characteristic image set of the smoke suspicious image; recording any one smoke suspicious region in the smoke suspicious images as a target region, and acquiring target regions at positions corresponding to all video frame images in the first characteristic image set; and extracting a vehicle region on the video frame image in the first characteristic image set, calculating the distance between a target region on the smoke suspicious image and the vehicle region, and marking a vehicle driving path corresponding to the vehicle region with the minimum distance as the target path.
The preset number of values is 7, and the method for acquiring the vehicle region on the video frame image in the first feature image set is the same as the method for acquiring the vehicle region in the first step. If the target area is the area where dust is located, it is necessary to find dust generated by which vehicle is running in the image in the first feature image set, so that the target area exists, and when the vehicle is running to generate dust, the dust is located near the vehicle, so the distance between the vehicle area and the area where dust is located should be small.
Based on this, the target vehicle is acquired from the distance between the target area and the vehicle area, the distance between the vehicle areas on the set number of frames of video frame images after the smoke suspicious image is calculated for any one target area, the vehicle corresponding to the vehicle area with the minimum distance is recorded as the target vehicle, and the vehicle travel path corresponding to the target vehicle is the target path. For example, if the smoke suspicious image is the ith frame of video frame image, selecting a set number of frames of video frame images as a basis after the ith frame of video frame image, acquiring a target vehicle corresponding to a target area in the ith frame of video frame image, and setting the value of the set number to be 5. That is, one target area corresponds to one target vehicle, the image of the target vehicle is closer to the image of the target area, and the position of the target vehicle is closer to the position of the target area.
It should be noted that, the distance between the target area and the vehicle area may be calculated by calculating the distance between the pixel point where the centroid of the target area is located and the centroid pixel point of the vehicle area, and the distance between the two pixel points may be calculated by acquiring the pixel coordinates of the pixel points, thereby calculating the distance between the two points based on the pixel coordinates. The practitioner may also select other methods to calculate the distance depending on the particular implementation scenario.
And calculating the absolute value of the difference value of the areas of the target areas on any two adjacent video frame images in the first characteristic image set.
According to the motion vector of the target area acquired by two video frame images, the motion vector of the target macro block acquired by the video frame images is known, in this embodiment, n2 edge pixel points are selected at equal intervals on the edge of the target area on the j-1 th video frame image in the first feature image set, n1 edge pixel points are selected at equal intervals on the edge of the target area on the j-1 th video frame image, and the vector between the n2 edge pixel points and the n1 edge pixel points is acquired, and the calculation is performedThe v-th edge pixel point representing the target area on the j-1-th frame video frame image and the +. >The vector sum of the edge pixels, i.e. the motion vector of the target area relative to the v-th edge pixel. Further use->And representing the vector sum between the edge pixel points of the target area of the j-1 frame video frame image and the edge pixel points of the target area on the j frame video frame image, and representing the motion vector of the target area.
In this embodiment, the pixel coordinates of the edge pixel points are obtained, and then the vector determined by the two edge pixel points is obtained by using a coordinate method. The practitioner may select other suitable methods to obtain the two-point determined vector depending on the particular implementation scenario. Meanwhile, the values of n1 and n2 are different according to the total number of the edge pixel points of the target area, and an implementer needs to set according to a specific implementation scene.
According to the video frame image, acquiring a running direction corresponding to a target path of the vehicle, calculating a value of an included angle between a unit vector and a motion vector of the running direction, calculating the reciprocal of the sum of the value and a preset second value, and obtaining a first direction characteristic value of a target area on two adjacent video frame images according to the product of the absolute value and the reciprocal; taking the average value of the first direction characteristic values of the target area on all any two adjacent video frame images in the first characteristic image set as a first coefficient.
The calculation formula of the first coefficient is expressed as:
wherein, the liquid crystal display device comprises a liquid crystal display device,a first coefficient corresponding to the target region, N0 representing the total number of images in the first set of feature images, +.>Representing the area of the target area in the image of the j-th frame of video frames in the first set of feature images,/->Representing the area of a target area in a j-1 th frame of video frame image in the first characteristic image set, X represents a unit vector formed by the running direction of a target vehicle, +.>The second value is 1 to prevent the denominator from being 0, representing the motion vector of the target region corresponding to the j-1 th video frame image.
Representing the surface of the target area in the j-1 frame video frame image and the j frame video frame image in the first characteristic image setThe larger the difference between the products, the larger the difference is, which indicates that the target area is changed greatly with the passage of time or the running of the vehicle, and further indicates that the larger the possibility that the target area is the area where dust is located, the larger the corresponding value of the first coefficient is.
A value representing an angle between a motion vector of the target area and a traveling direction vector of the vehicle, for example, an angle of 30 DEG between the two vectors, then +. >The value of (2) is 30. The larger the value of the included angle is, the larger the difference between the moving direction of the target area and the running direction of the vehicle is, and further the target area does not move along with the movement of the vehicle, the smaller the possibility that the target area is the area where dust is located is, and the smaller the value of the corresponding first coefficient is.
The first direction characteristic value of the target area on the j-1 frame video frame image and the j frame video frame image represents the change of the area and the movement direction of the target area in two continuous frames of video frame images, and further the first coefficient represents the change of the area and the movement direction of the target area in the continuous frames of video frame images.
Further, if one suspected target area is the area where dust is located, the suspected target area moves along with the running of the vehicle, that is, the movement track of the suspected smoke area is relatively close to the running path of the vehicle, and the detection points are symmetrical about the running path of the vehicle, so that the positions of the detection points about the suspected smoke area are also symmetrical, and the distances from the suspected smoke area to two mutually symmetrical monitoring points are relatively close or equal.
Recording two detection points symmetrical about a vehicle driving path as detection point pairs, calculating the average value of the distance between the edge pixel point of a target area and the detection point for any one detection point pair of a target path on any one video frame image of a first characteristic image set, taking the inverse of the sum of the absolute value of the difference value of the two average values and a preset first numerical value as the characteristic value of the detection point pair, and calculating the average value of the characteristic values of all the detection point pairs on the target path to obtain a characteristic average value; and further calculating the product of the average number of the feature mean values corresponding to all the video frame images in the first feature image set and the first coefficient to obtain a first possibility index of the target area.
The calculation formula of the first likelihood index is expressed as:
wherein PC represents a first likelihood index of the target area, namely a first likelihood index of the suspected smoke area, N0 represents the total number of images in the first characteristic image set, m3 represents the total number of detection point pairs on a target path of a target vehicle corresponding to the target area in a j-th frame of video frame image in the first characteristic image set,representing the average value of the distance between the edge pixel point of the target area in the j-th video frame image and the first detection point in the t-th detection point pair,/for>Representing the average value of the distance between the edge pixel point of the target area in the j-th video frame image and the second detection point in the t-th detection point pair,/for>For the characteristic value of the t-th detection point pair, < >>For the characteristic mean value corresponding to the target area on the j-th frame video frame image, the value of the first numerical value is 1, and the first numerical value is used for preventing the denominator from being 0, < >>Representing the first coefficient corresponding to the target region.
The difference representing the distance between the target area and the detection points at the two sides of the same position of the vehicle driving path is smaller, which means that the more symmetrical the t detection point pair is about the target area, the more the target area is likely to be the dust. / >The position relation between the target area and the detection point pair is represented, the degree of consistency between the movement track of the target area and the vehicle running path is reflected, and the larger the value is, the more consistent the movement track of the target area and the vehicle running path are, and the greater the possibility that the target area is the area where dust is located is further.
The consistency degree of the movement track of the target area in the first characteristic image set and the vehicle running path is reflected in the aspect of position information, and the greater the movement track of the target area is consistent with the vehicle running path, the greater the corresponding value of the first probability index is, which indicates that the greater the probability that the target area is the dust area is. The first coefficient characterizes the change of the area and the movement direction of the target area in continuous multi-frame video frame images, and the larger the value of the first coefficient is, the larger the value of the corresponding first probability index is, which indicates that the larger the probability that the target area is the area where dust is located is. I.e. the first likelihood indicator characterizes the likelihood that the target area is the area where dust is located.
And thirdly, obtaining a second coefficient according to the difference between the moving direction corresponding to the suspected smoke area and the set direction, and obtaining a second possibility index according to the difference between the areas of the suspected smoke areas and the adjacent frame images and the second coefficient.
First, the actual smoke generated by the occurrence of a fire is spread from the point of occurrence of the fire, and the direction of spread of the smoke is mainly upward, and the relationship between the area where the smoke is located and the vehicle and the traveling path of the vehicle is weak, that is, the area where the smoke is located is mainly characterized by the movement characteristics of the area itself. The likelihood that the suspected smoke area is the area where the smoke is located can be determined by utilizing the position relation between the suspected smoke area and the vehicle driving path.
If the suspected smoke area is the area where the smoke is located, the motion direction of the suspected smoke area in the video frame images of the continuous multiframe is relatively close to the vertical upward direction. For any two adjacent video frame images in the first characteristic image set, acquiring a value of an included angle between a motion vector of a target area obtained according to the two video frame images and a unit vector of a set direction, and taking the reciprocal of the sum of the value and a preset third value as a second direction characteristic value of the target area on the two adjacent video frame images; and taking the average value of the second direction characteristic values of the target area on all any two adjacent video frame images in the first characteristic image set as a second coefficient. In this embodiment, the set direction is a vertically upward direction.
The calculation formula of the second coefficient is expressed as:
wherein, the liquid crystal display device comprises a liquid crystal display device,representing a second coefficient corresponding to the target region, N0 representing the total number of images in the first set of feature images, +.>The motion vectors corresponding to the j-1 th frame video frame image in the first characteristic image set of the target region are represented, Y represents a unit vector in which a set direction is located, namely a unit vector formed in a vertical upward direction, the value of a third numerical value is 1, and the third numerical value is used for preventing a denominator from being 0.
A value representing the angle between the motion vector of the target area and the vertically upward direction vector, e.g. the angle between the two vectors is 30 °, then->The value of (2) is 30. The smaller the value of the included angle, the smaller the difference between the movement direction of the target area and the vertical upward direction is, namely the more the suspected smoke area corresponding to the target area moves upwards along with the time or the running of the vehicle, the greater the possibility that the target area is the area where the smoke is located, and the greater the value of the corresponding second coefficient is.
The second direction characteristic value of the target area from the j-1 frame video frame image to the j frame video frame image reflects the change of the movement direction of the target area in two continuous frames of video frame images, and the second coefficient further represents the change of the movement direction of the target area in multiple continuous frames of video frame images.
Further, for a suspected smoke area, there is a suspected smoke area in the continuous multi-frame video frame image at a corresponding position, the suspected smoke areas are not identical, if the suspected smoke area in the suspected smoke image is the area where the smoke is located, the suspected smoke area in the corresponding position in the continuous multi-frame image after the suspected smoke image does not change greatly along with the running of the vehicle, but the suspected smoke area may diffuse along with the time. When the time is long enough, the area of the region where the smoke is located may become large due to the diffusion of the smoke.
For any video frame image in the first characteristic image set, acquiring a video frame image of a target vehicle corresponding to a target area, marking the video frame image as a selected frame image, and acquiring a certain number of Zhang Shipin frame images by taking the selected frame image as a starting frame, marking the video frame image as a second characteristic image set;
calculating the absolute value of the difference between the area of the target area on the video frame image and the area of the target area on the video frame image in the second characteristic image set, and calculating the reciprocal of the sum of the absolute value of the difference and a preset fourth numerical value; and obtaining the reciprocal of the time difference between the video frame image and the selected frame image, and obtaining a second possibility index of the target area according to the two reciprocal and a second coefficient.
The calculation formula of the second likelihood index is expressed as:
or expressed as:
where PB denotes the second likelihood indicator of the target area,representing a second coefficient corresponding to the target region, N0 representing the total number of images in the first feature image set, N1 representing the total number of images in the second feature image set, < >>Area of target area in j-th frame video frame image in first characteristic image set, +.>Representing the area of the target area in the video frame image of the e-th frame in the second set of feature images,/->Representing the difference in time between the image of the j-th frame video frame and the image of the e-th frame video frame. Meanwhile, the fourth value is 1, and the fourth value is to prevent the denominator from being 0.
Representing a j-th frame video frame imageThe smaller the area difference is, the smaller the relation between the target area and the vehicle running is, namely, the area of the target area cannot be changed along with the vehicle running, and the larger the possibility that the target area is the area where the smoke is located is, the larger the corresponding value of the second possibility index is.
The time difference between the target area and the target vehicle is reflected, and the larger the time difference is, the more likely the target area is changed, so that when the time difference is large, the influence of the area change of the target area in different video frame images on the possibility that the target area is the area where the smoke is located is small. / >When the j-th frame image and the target vehicle pass through the multi-frame image, the change condition of the area of the target area reflects the relation between the target area and the vehicle running.
The area change relation between the target area in the continuous N0 frame video frame images and the target area in the N1 frame video frame images after the target vehicle passes through is reflected, the larger the value is, the larger the value of the corresponding second possibility index is, and the greater the possibility that the suspected smoke area corresponding to the target area is the area where the smoke is located is indicated. The second coefficient characterizes the change of the motion direction of the target area in the continuous multi-frame video frame images, and the larger the value of the second coefficient is, the larger the value of the corresponding second possibility index is, which indicates that the possibility that the suspected smoke area corresponding to the target area is the area where the smoke is located is higher.
The second possibility index is used for determining the possibility that the suspected smoke area is the area where the smoke is located according to the change direction of the suspected smoke area and the influence of the vehicle passing before and after the suspected smoke area. The method comprises the steps of firstly considering that actual smoke has natural diffusion characteristics, reflecting the possibility that a suspected smoke area is the area where the smoke is located from the diffusion direction of the smoke, then considering the influence of vehicle driving on the suspected smoke area, and considering the influence of time intervals between images on the relation between a target area and the vehicle driving, so as to avoid misjudgment of the actual smoke by the vehicle driving to the greatest extent, and obtaining the possibility that the suspected smoke area is the area where the smoke is located more accurately.
And fourthly, determining a global possibility index of the suspected smoke area according to the first possibility index and the second possibility index, screening the suspected smoke area according to the global possibility index to obtain an actual smoke area, and further obtaining the abnormal condition of the road condition of the tunnel road section.
The first likelihood index characterizes the likelihood that the suspected smoke area is the area where dust is located, so the greater the value of the first likelihood index is, the less the likelihood that the suspected smoke area is the area where smoke is located. The second likelihood indicator characterizes the likelihood that the suspected smoke area is the area where the smoke is located, so the smaller the value of the second likelihood indicator is, the smaller the likelihood that the suspected smoke area is the area where the smoke is located is. Based on this, a difference between the preset fifth value and the normalized value of the first likelihood indicator is calculated, and the product between the difference and the normalized value of the second likelihood indicator is a global likelihood indicator, expressed by a formula:
wherein PF is a global likelihood index corresponding to a suspected smoke region,for the normalized value of the first likelihood indicator,/->Is the normalized value of the second likelihood indicator. />Indicating the possibility of the suspected smoke area being the area where the smoke is located,/- >The probability that the suspected smoke area is the area where the smoke is located is indicated, the probability that the suspected smoke area finally belongs to the area where the smoke is located is indicated by the PF, and the greater the value of the PF is, the greater the probability that the suspected smoke area finally is the area where the actual smoke is located.
And screening the suspected smoke area according to the global possibility index of the suspected smoke area to obtain an actual smoke area. Specifically, when the global likelihood index of the suspected smoke area is greater than the preset likelihood threshold, the likelihood that the suspected smoke area is the area where the smoke is located is indicated to be greater, so that the suspected smoke area corresponding to the global likelihood index being greater than the preset likelihood threshold is recorded as the actual smoke area. Wherein, the value of the probability threshold is 0.7, and the implementer can set according to the specific implementation scene.
When the fact that the smoke suspicious image has the actual smoke area is detected, the fact that the fire disaster abnormal situation occurs in the tunnel section at the moment is indicated, and in order to ensure the safety of the tunnel and the running vehicles in the tunnel, relevant staff should be immediately informed to check the abnormal situation of the tunnel.
The above embodiments are only for illustrating the technical solution of the present application, and not for limiting the same; although the application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the application and are intended to be included within the scope of the application.
Claims (4)
1. The method for identifying the road condition abnormality of the tunnel road section based on the monitoring data is characterized by comprising the following steps:
acquiring at least two frames of video frame images in the video monitoring data of the tunnel section, and screening each frame of video frame image to obtain a smoke suspicious image; extracting a suspected smoke area in a smoke suspected image, and determining a detection point according to the running condition of a vehicle in the smoke suspected image;
calculating a first coefficient according to the corresponding moving direction of the suspected smoke area, and obtaining a first possibility index of the suspected smoke area according to the distance between the pixel points on the edge of the suspected smoke area and the detection points and the first coefficient;
obtaining a second coefficient according to the difference between the moving direction corresponding to the suspected smoke area and the set direction, and obtaining a second possibility index according to the difference between the areas of the suspected smoke areas and the adjacent frame images and the second coefficient;
determining a global possibility index of the suspected smoke area according to the first possibility index and the second possibility index, screening the suspected smoke area according to the global possibility index to obtain an actual smoke area, and further obtaining the abnormal condition of the road condition of the tunnel road section;
The method for acquiring the first possibility index specifically comprises the following steps:
for any one smoke suspicious image, acquiring a continuous preset number Zhang Shipin of images by taking the smoke suspicious image as a starting frame image to form a first characteristic image set of the smoke suspicious image; acquiring a vehicle running path according to the video frame image, and marking two detection points symmetrical about the vehicle running path as detection point pairs;
recording any one smoke suspicious region in the smoke suspicious images as a target region, and acquiring target regions at positions corresponding to all video frame images in the first characteristic image set; extracting a vehicle region on a video frame image in a first characteristic image set;
for any one target area, calculating the distance between the vehicle areas on the set number of frames of video frame images after the smoke suspicious image, and recording the vehicle corresponding to the vehicle area with the minimum distance as a target vehicle, wherein the vehicle running path corresponding to the target vehicle is a target path;
on any video frame image of the first characteristic image set, for any detection point pair of the target path, calculating the average value of the distance between the edge pixel point of the target area and the detection point, taking the inverse of the sum of the absolute value of the difference value of the two average values and a preset first numerical value as the characteristic value of the detection point pair, and calculating the average value of the characteristic values of all the detection point pairs on the target path to obtain the characteristic average value;
Calculating the product of the average number of the feature mean values corresponding to all the video frame images in the first feature image set and the first coefficient to obtain a first possibility index of the target area;
the method for acquiring the first coefficient specifically comprises the following steps:
for any two adjacent video frame images in the first characteristic image set, calculating the absolute value of the difference value of the areas of the target areas on the two video frame images; according to the two video frame images, obtaining a motion vector of a target area, according to a running direction corresponding to a target path of a vehicle, calculating a value of an included angle between a unit vector of the running direction and the motion vector, calculating the reciprocal of the sum of the value and a preset second value, and taking the product of the absolute value and the reciprocal of the difference of the areas as a first direction characteristic value of the target area on two adjacent video frame images;
taking the average value of the first direction characteristic values of the target area on all any two adjacent video frame images in the first characteristic image set as a first coefficient;
obtaining a second coefficient according to the difference between the moving direction corresponding to the suspected smoke area and the set direction, and obtaining a second possibility index according to the difference between the areas of the suspected smoke areas and the adjacent frame images and the second coefficient, wherein the method specifically comprises the following steps:
For any two adjacent video frame images in the first characteristic image set, acquiring a value of an included angle between a motion vector of a target area obtained according to the two video frame images and a unit vector of a set direction, and taking the reciprocal of the sum of the value and a preset third value as a second direction characteristic value of the target area on the two adjacent video frame images; taking the average value of the second direction characteristic values of the target area on all any two adjacent video frame images in the first characteristic image set as a second coefficient;
for any video frame image in the first characteristic image set, acquiring video frame images of a target vehicle corresponding to a target area, and marking the video frame images as selected frame images, and acquiring a first preset number of video frame images by taking the selected frame images as initial frames, and marking the first preset number of video frame images as a second characteristic image set;
calculating the absolute value of the difference between the area of the target area on the video frame image and the area of the target area on the video frame image in the second characteristic image set, and calculating the reciprocal of the sum of the absolute value of the difference and a preset fourth numerical value; obtaining the reciprocal of the time difference between the video frame image and the video frame image in the second characteristic image set, and obtaining a second possibility index of the target area according to the two reciprocal and a second coefficient;
The calculation formula of the second possibility index specifically includes:
where PB denotes a second likelihood indicator for the target region, N0 denotes the total number of images in the first set of feature images,representing the motion vectors corresponding to the j-1 th frame video frame image of the target area in the first characteristic image set, Y representing the unit vector in which the set direction is located, and +.>Representing a second direction characteristic value of the target area on two adjacent video frame images; n1 represents the total number of images in the second set of feature images,/->Area of target area in j-th frame video frame image in first characteristic image set, +.>Representing a second characteristic diagramArea of target area in video frame image of e-th frame in image set, < >>Representing the difference in time between the image of the j-th frame video frame and the image of the e-th frame video frame,is a second coefficient>Wherein the value 1 is a third value, +.>The value 1 in (2) is a fourth value;
the determining the global likelihood indicator of the suspected smoke area according to the first likelihood indicator and the second likelihood indicator specifically includes:
and calculating a difference value between the preset fifth numerical value and the normalized value of the first likelihood index, wherein the product of the difference value and the normalized value of the second likelihood index is the global likelihood index.
2. The method for identifying abnormal road conditions of a tunnel road section based on monitoring data according to claim 1, wherein the step of screening the suspected smoke area according to the global likelihood index to obtain an actual smoke area is specifically as follows:
and (3) recording a suspected smoke area corresponding to the global possibility index being larger than a preset possibility threshold as an actual smoke area.
3. The method for identifying abnormal road conditions of a tunnel road section based on monitoring data according to claim 1, wherein the step of screening each frame of video frame image to obtain a smoke suspicious image is specifically as follows:
for any frame of video frame image, acquiring the light transmittance of the video frame image in the process of processing the video frame image by using a dark channel defogging algorithm, and if the light transmittance is smaller than a preset light transmittance threshold value, marking the video frame image as a smoke suspicious image; and screening each frame of video frame image to obtain all the smoke suspicious images.
4. The method for identifying abnormal road conditions of a tunnel road section based on monitoring data according to claim 1, wherein the determining the detection point according to the running condition of the vehicle in the smoke suspicious image comprises:
And acquiring a vehicle driving path according to the video frame image, and setting detection points at equal intervals at positions which are set at a distance from two sides of the vehicle driving path, so as to acquire detection points in the smoke suspicious image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310714673.2A CN116453064B (en) | 2023-06-16 | 2023-06-16 | Method for identifying abnormal road conditions of tunnel road section based on monitoring data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310714673.2A CN116453064B (en) | 2023-06-16 | 2023-06-16 | Method for identifying abnormal road conditions of tunnel road section based on monitoring data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116453064A CN116453064A (en) | 2023-07-18 |
| CN116453064B true CN116453064B (en) | 2023-08-18 |
Family
ID=87124170
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310714673.2A Active CN116453064B (en) | 2023-06-16 | 2023-06-16 | Method for identifying abnormal road conditions of tunnel road section based on monitoring data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116453064B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117079219B (en) * | 2023-10-08 | 2024-01-09 | 山东车拖车网络科技有限公司 | Vehicle running condition monitoring method and device applied to trailer service |
| CN117423068B (en) * | 2023-12-18 | 2024-04-02 | 东莞市杰瑞智能科技有限公司 | Vehicle fire detection method and system for smart city |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110458047A (en) * | 2019-07-23 | 2019-11-15 | 北京理工大学 | A kind of country scene recognition method and system based on deep learning |
| CN110516609A (en) * | 2019-08-28 | 2019-11-29 | 南京邮电大学 | A kind of fire video detection and method for early warning based on image multiple features fusion |
| KR102062500B1 (en) * | 2019-09-09 | 2020-01-03 | 윤영복 | Forest fire disaster detection server |
| CN113102305A (en) * | 2021-04-02 | 2021-07-13 | 重庆万重山智能科技有限公司 | Cleaning mechanism and forest fire monitoring system based on unmanned aerial vehicle |
| WO2021212443A1 (en) * | 2020-04-20 | 2021-10-28 | 南京邮电大学 | Smoke video detection method and system based on lightweight 3d-rdnet model |
| WO2022121129A1 (en) * | 2020-12-12 | 2022-06-16 | 南方电网调峰调频发电有限公司 | Fire recognition method and apparatus, and computer device and storage medium |
| KR102441874B1 (en) * | 2022-04-06 | 2022-09-14 | 주식회사 맥서브 | Fire fighting system in tunnel using CCTV |
| CN115223105A (en) * | 2022-09-20 | 2022-10-21 | 万链指数(青岛)信息科技有限公司 | Big data based risk information monitoring and analyzing method and system |
| CN115761658A (en) * | 2023-01-07 | 2023-03-07 | 山东高速股份有限公司 | Highway pavement condition detection method based on artificial intelligence |
| CN115830790A (en) * | 2022-11-16 | 2023-03-21 | 安徽兆之翔智能科技有限公司 | Corridor safety monitoring system based on intelligent monitoring equipment |
| CN116259005A (en) * | 2023-02-04 | 2023-06-13 | 国网浙江省电力有限公司嘉兴供电公司 | Intelligent monitoring system based on roof photovoltaic fire control |
-
2023
- 2023-06-16 CN CN202310714673.2A patent/CN116453064B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110458047A (en) * | 2019-07-23 | 2019-11-15 | 北京理工大学 | A kind of country scene recognition method and system based on deep learning |
| CN110516609A (en) * | 2019-08-28 | 2019-11-29 | 南京邮电大学 | A kind of fire video detection and method for early warning based on image multiple features fusion |
| KR102062500B1 (en) * | 2019-09-09 | 2020-01-03 | 윤영복 | Forest fire disaster detection server |
| WO2021212443A1 (en) * | 2020-04-20 | 2021-10-28 | 南京邮电大学 | Smoke video detection method and system based on lightweight 3d-rdnet model |
| WO2022121129A1 (en) * | 2020-12-12 | 2022-06-16 | 南方电网调峰调频发电有限公司 | Fire recognition method and apparatus, and computer device and storage medium |
| CN113102305A (en) * | 2021-04-02 | 2021-07-13 | 重庆万重山智能科技有限公司 | Cleaning mechanism and forest fire monitoring system based on unmanned aerial vehicle |
| KR102441874B1 (en) * | 2022-04-06 | 2022-09-14 | 주식회사 맥서브 | Fire fighting system in tunnel using CCTV |
| CN115223105A (en) * | 2022-09-20 | 2022-10-21 | 万链指数(青岛)信息科技有限公司 | Big data based risk information monitoring and analyzing method and system |
| CN115830790A (en) * | 2022-11-16 | 2023-03-21 | 安徽兆之翔智能科技有限公司 | Corridor safety monitoring system based on intelligent monitoring equipment |
| CN115761658A (en) * | 2023-01-07 | 2023-03-07 | 山东高速股份有限公司 | Highway pavement condition detection method based on artificial intelligence |
| CN116259005A (en) * | 2023-02-04 | 2023-06-13 | 国网浙江省电力有限公司嘉兴供电公司 | Intelligent monitoring system based on roof photovoltaic fire control |
Non-Patent Citations (1)
| Title |
|---|
| Development of sensor and optimal placement for smoke detection in an electric vehicle battery pack;Mukund Kulkarni 等;《2015 IEEE International Transportation Electrification Conference (ITEC)》;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116453064A (en) | 2023-07-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116453064B (en) | Method for identifying abnormal road conditions of tunnel road section based on monitoring data | |
| CN108647572B (en) | Lane departure early warning method based on Hough transform | |
| CN105389567B (en) | Group abnormality detection method based on dense optical flow histogram | |
| US9047518B2 (en) | Method for the detection and tracking of lane markings | |
| JP5297078B2 (en) | Method for detecting moving object in blind spot of vehicle, and blind spot detection device | |
| US20220004818A1 (en) | Systems and Methods for Evaluating Perception System Quality | |
| Hadjidemetriou et al. | Automated detection of pavement patches utilizing support vector machine classification | |
| CN111199647B (en) | Monitoring video detection method for continuous lane changing and illegal turning of road vehicles | |
| US20080100473A1 (en) | Spatial-temporal Image Analysis in Vehicle Detection Systems | |
| CN104966304A (en) | Kalman filtering and nonparametric background model-based multi-target detection tracking method | |
| CN115862259B (en) | Fire alarm early warning system based on temperature monitoring | |
| CN111783700B (en) | Automatic recognition and early warning method and system for pavement foreign matters | |
| CN116385948B (en) | System and method for early warning railway side slope abnormality | |
| Li et al. | PETS 2015: Datasets and challenge | |
| CN114663859A (en) | Sensitive and accurate complex road condition lane deviation real-time early warning system | |
| CN116110006B (en) | Scenic spot tourist abnormal behavior identification method for intelligent tourism system | |
| KR20150002040A (en) | The way of Real-time Pedestrian Recognition and Tracking using Kalman Filter and Clustering Algorithm based on Cascade Method by HOG | |
| CN114373162B (en) | Dangerous area personnel intrusion detection method and system for transformer substation video monitoring | |
| CN115331258A (en) | Industrial-grade applied danger forbidden zone personnel detection system and method | |
| KR20220120211A (en) | Drone image analysis system based on deep learning for traffic measurement | |
| CN112507902A (en) | Traffic sign abnormality detection method, computer device, and storage medium | |
| CN112634299A (en) | Remnant detection method for eliminating interference of winged insects | |
| CN115546737B (en) | Machine room monitoring method | |
| CN114821795B (en) | Personnel running detection and early warning method and system based on ReiD technology | |
| Abukhait et al. | Occlusion-invariant tilt angle computation for automated road sign condition assessment |
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 |