CN113205687B - Drunk driving vehicle track recognition system based on video monitoring - Google Patents

Drunk driving vehicle track recognition system based on video monitoring Download PDF

Info

Publication number
CN113205687B
CN113205687B CN202110481444.1A CN202110481444A CN113205687B CN 113205687 B CN113205687 B CN 113205687B CN 202110481444 A CN202110481444 A CN 202110481444A CN 113205687 B CN113205687 B CN 113205687B
Authority
CN
China
Prior art keywords
module
image
automobile
vehicle
line
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
Application number
CN202110481444.1A
Other languages
Chinese (zh)
Other versions
CN113205687A (en
Inventor
朱静
何伟聪
林静旖
潘梓沛
毛俊彦
尹邦政
明家辉
钟绮岚
薛穗华
赵宣博
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangzhou University
Original Assignee
Guangzhou University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Guangzhou University filed Critical Guangzhou University
Priority to CN202110481444.1A priority Critical patent/CN113205687B/en
Publication of CN113205687A publication Critical patent/CN113205687A/en
Application granted granted Critical
Publication of CN113205687B publication Critical patent/CN113205687B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/017—Detecting movement of traffic to be counted or controlled identifying vehicles
    • G08G1/0175—Detecting movement of traffic to be counted or controlled identifying vehicles by photographing vehicles, e.g. when violating traffic rules
    • 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
    • G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
    • G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
    • G06V30/10—Character recognition
    • G06V30/14—Image acquisition
    • G06V30/148—Segmentation of character regions
    • G06V30/153—Segmentation of character regions using recognition of characters or words
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/60—Type of objects
    • G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
    • G06V20/625—License plates
    • 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

Abstract

The invention relates to a drunk driving vehicle track recognition system based on video monitoring, which comprises a video acquisition module, a digital image preprocessing module, a vehicle running characteristic detection module, a drunk driving vehicle recognition module, a license plate recognition module and a traffic police reporting artificial detection module, wherein the modules are sequentially connected; the digital image preprocessing module comprises a lane line identification module, an automobile identification module and an image preprocessing module; the vehicle running characteristic detection module comprises an automobile tracking module, a turn light identification module, an angle detection module and a line pressing detection module. The invention improves the drunk driving detection efficiency of traffic polices by enlarging the drunk driving detection range, so that drunk driving drivers can not hold lucky psychology any more, and further traffic accidents caused by drunk driving are reduced.

Description

Drunk driving vehicle track recognition system based on video monitoring
Technical Field
The invention relates to the technical field of video monitoring and track identification, in particular to a drunk driving vehicle track identification system based on video monitoring.
Background
At present, drunk driving detection is mainly carried out by a traffic police on a highway for spot check and vehicle stopping, and a detector is used for carrying out breath alcohol detection on an automobile driver. However, as the drunk driving and drunk driving traffic accidents are continuously expanded, the traffic accidents cannot be prevented only by manual spot inspection. The existing detection mode has a small detection range, a large amount of manpower is consumed, and the working efficiency of a traffic police cannot be improved.
Therefore, a system for improving the drunk driving detection efficiency needs to be found.
Disclosure of Invention
In order to solve the technical problems in the prior art, the invention provides a drunk driving vehicle track recognition system based on video monitoring, which enlarges the drunk driving detection range and improves the drunk driving detection working efficiency of traffic polices by a video acquisition module, a digital image preprocessing module, a vehicle driving characteristic detection module, a drunk driving vehicle recognition module, a license plate recognition module and a manual traffic police reporting detection module, so that drunk driving drivers can not hold lucky psychology any more, and further traffic accidents caused by drunk driving are reduced.
The invention is realized by adopting the following technical scheme: a drunk driving vehicle track recognition system based on video monitoring comprises a video acquisition module, a digital image preprocessing module, a vehicle driving characteristic detection module, a drunk driving vehicle recognition module, a license plate recognition module and a traffic police reporting artificial detection module; the digital image preprocessing module comprises a lane line identification module, an automobile identification module and an image preprocessing module; the vehicle running characteristic detection module comprises an automobile tracking module, a steering lamp identification module, an angle detection module and a line pressing detection module; the system comprises a video acquisition module, a digital image preprocessing module, a vehicle running characteristic detection module, a drunk driving vehicle identification module, a license plate identification module and a traffic police reporting artificial detection module, wherein the modules are connected in sequence.
Compared with the prior art, the invention has the following advantages and beneficial effects:
according to the invention, through the video acquisition module, the digital image preprocessing module, the vehicle driving characteristic detection module, the drunk driving vehicle identification module, the license plate identification module and the manual traffic police reporting detection module, the drunk driving detection range is expanded, the drunk driving detection work efficiency of a traffic police is improved, so that drunk driving drivers do not hold a lucky psychology any more, and further traffic accidents caused by drunk driving are reduced.
Drawings
FIG. 1 is a schematic diagram of the system architecture of the present invention;
FIG. 2 is a schematic diagram of lane change driving and lane center line pressing driving of a vehicle;
FIG. 3 is a schematic view of traffic lights such as vehicles, which is a first vehicle, and a head line;
FIG. 4(a) is a lane driving diagram of a vehicle;
fig. 4(b) is a lane line identification acquisition diagram;
FIG. 5 is a schematic view of a vehicle position lock.
Detailed Description
The present invention will be described in further detail with reference to examples and drawings, but the embodiments of the present invention are not limited thereto.
Examples
As shown in fig. 1, the drunk driving vehicle track recognition system based on video monitoring in the embodiment includes a video acquisition module, a digital image preprocessing module, a vehicle driving characteristic detection module, a drunk driving vehicle recognition module, a license plate recognition module and an artificial traffic police reporting detection module; the digital image preprocessing module comprises a lane line identification module, an automobile identification module and an image preprocessing module; the vehicle running characteristic detection module comprises an automobile tracking module, a steering lamp identification module, an angle detection module and a line pressing detection module; the system comprises a video acquisition module, a digital image preprocessing module, a vehicle running characteristic detection module, a drunk driving vehicle identification module, a license plate identification module and a traffic police reporting artificial detection module, wherein the modules are connected in sequence.
In this embodiment, the details of each module are as follows:
the video acquisition module is used for acquiring the video image data of the vehicles coming and going, and the vehicles learn the included angle theta between the vehicle body line and the lane line when all the vehicles coming and going on the corresponding road change lanes through the monitoring camera, and count the maximum threshold value theta max.
Specifically, the method for detecting the included angle theta between the automobile body line and the lane line comprises the following steps:
setting the time from the automobile lane change starting time point t1 to the lane change ending time point t2 as t2-t 1; from t1 to t2, there are k pieces in the video set per unit time of 1 secondA picture; at any time during the lane changing period of the automobile, the included angle between the automobile body line and the lane line is theta; then there are k (t2-t1) pictures from t1 to t2, and the pictures are grouped in time series, and the number of groups is (t2-t 1); in group 1, let θ be
Figure GDA0003651685290000021
Theta in the kth picture is
Figure GDA0003651685290000022
In group 2, let θ of picture 1 be
Figure GDA0003651685290000023
Theta in the kth picture is
Figure GDA0003651685290000024
And by analogy, calculating the included angle theta of the t2-t1 group of pictures at each moment.
For the t2-t1 group of pictures, the following judgments are made:
Figure GDA0003651685290000031
Figure GDA0003651685290000032
...
and the like, counting the differences of the t2-t1 groups.
If one or more angle difference values among the t2-t1 are greater than or equal to the maximum threshold value theta max, the drunk driving is judged to be suspected.
Specifically, for a video set of lane change of vehicles drunk or drunk on a road, when the vehicles change lanes, the maximum threshold value θ max of the body line and the lane line is counted, and the specific statistical method is as follows:
setting the time from the automobile lane change starting time point t1 to the lane change ending time point t2 as t2-t 1; within its video set from t1 to t2, there are k pictures per unit time of 1 second; ren during lane change of automobileAt the moment, the included angle between the vehicle body line and the lane line is theta; then there are k (t2-t1) pictures from t1 to t2, and the pictures are grouped in time series, and the number of groups is (t2-t 1); in group 1, let θ be
Figure GDA0003651685290000033
Theta in the kth picture is
Figure GDA0003651685290000034
In group 2, let θ of picture 1 be
Figure GDA0003651685290000035
Theta in the kth picture is
Figure GDA0003651685290000036
And by analogy, calculating the included angle theta of the t2-t1 group of pictures at each moment.
For the t2-t1 group of pictures, the following calculations were performed:
Figure GDA0003651685290000037
Figure GDA0003651685290000038
...
comparing the included angle difference of the t2-t1 groups, and recording the statistical maximum value as theta0。
By analogy, the difference values of the t2-t1 groups are counted, and data statistics of the maximum value of the included angle between the body line and the lane line is carried out on vehicles which are drunk or drunk to drive on the road by using a statistical learning method, so that the maximum threshold value theta max of the included angle between the body line and the lane line is calculated when the vehicles change the lane on the road.
The digital image preprocessing module is used for acquiring an original image of a vehicle by using a contour recognition method for the vehicle in the monitoring video, segmenting the vehicle and lane lines in the image, and carrying out image preprocessing on the acquired image, wherein the image preprocessing comprises the steps of filtering, binaryzation, filling, modification, thinning and the like to acquire a contour image only containing the vehicle and the lane lines; the median filtering and the mean filtering in the filtering are performed by sliding a 3 x 3 filtering template on an image, sequencing 9 pixel gray values in a window of the image which slides to the image, and replacing the gray values of pixels corresponding to the center point of the window with the gray values of three pixels in the middle of the sequencing.
Specifically, the digital image preprocessing module comprises a lane line identification module, an automobile identification module and an image preprocessing module, and the details of the modules are as follows:
as shown in fig. 4(a) and 4(b), in this embodiment, the lane line identification module performs edge detection using a Sobel operator, a Laplacian operator, and a Canny operator, performs region-of-interest detection using fillconvex, outputs a matrix with 0 elements, and draws a mask image to obtain a region-of-interest; independent pixel points forming the lane line are connected into a straight line by Hough transformation, and then the addWeighted function is used for realizing image weighted superposition.
As shown in fig. 5, in this embodiment, the car identification module is used for determining a car and locking the car position; the identification is carried out by using an algorithm combining the histogram of oriented gradient HOG and the support vector machine SVM, and the specific implementation process is as follows:
step 1: generating feature data of the image by extracting a Histogram of Oriented Gradients (HOG) descriptor of the image, and normalizing HOG feature vectors of the histogram of oriented gradients; specifically, the principle of the directional gradient calculation is as follows:
G(x,y)＝gx(x,y)+gy(x,y) (1)
gx(x,y)＝I(x+1,y)-I(x,y) (2)
gy(x,y)＝I(x,y+1)-I(x,y) (3)
Figure GDA0003651685290000041
β＝arctan(gy/gx) (5)
wherein, (x, y) is the coordinates of the pixel; g (x, y) is the sum of the levels at the pixel points (x, y) in the input imageThe sum of the vertical gradients; gx(x, y) is the horizontal gradient at the pixel point (x, y) in the input image; g is a radical of formulay(x, y) is the vertical gradient at the pixel point (x, y) in the input image; i (x, y) is a pixel value of the image at (x, y), I (x +1, y) is a pixel value of the image at (x +1, y); i (x, y +1) is the pixel value of the image at (x, y + 1); g is the gradient amplitude of the pixel point; gxThe gradient of the pixel points in the horizontal direction is obtained; gyThe gradient of the pixel points in the vertical direction is obtained; beta is the gradient direction of the pixel point;
step 2: connecting HOG characteristics in series, dividing the image into a plurality of connection areas, and performing convolution kernel operations on each connection area in the x and y directions to obtain a gradient integral image;
and step 3: regularizing the gradient vector by calculating characteristic dimensions to obtain final HOG characteristics;
and 4, step 4: and judging the automobile by combining with an SVM classifier, locking the position of the automobile, establishing a rectangular coordinate system for the image, detecting and acquiring the center point of the automobile, and acquiring the current coordinate and the width of the automobile.
The image preprocessing module is used for framing the video and enabling the video to perform the next image processing in the form of each frame; preprocessing an image, removing noise points of each frame of image by Gaussian filtering, carrying out gray level processing by gray level transformation, converting an RGB image into a gray level image by a cvtColor function, and carrying out binarization processing on the obtained image.
The vehicle running characteristic detection module is used for detecting whether an included angle theta between a vehicle body and a lane line exceeds a maximum threshold value theta max; whether the automobile runs through the pressing line or not; when the automobile waits for the traffic light and is the first automobile, whether the head of the automobile presses a line or not is judged; this module mainly used deals with the wine and drives the driver, because intake alcohol, to the outside condition that leads to after reaction ability and the controllability decline, if: slow response to signal lights; swinging is indefinite; snake driving, etc., the vehicle running characteristics are shown in fig. 2 and 3.
Specifically, the vehicle driving characteristic detection module comprises an automobile tracking module, a turn light identification module, an angle detection module and a line pressing detection module, and the details of the modules are as follows;
the automobile tracking module acquires the automobile motion direction and the automobile motion line by combining Kalman filtering and Hungary algorithm, and the specific realization process is as follows:
step 1: setting the current central point coordinate (x, y) of the automobile and the central point coordinate (x) at the last momentt-1,yt-1) Then its euclidean distance:
Figure GDA0003651685290000051
Figure GDA0003651685290000052
wherein D ist|t-1A coordinate set of the central point detected at the last moment;
step 2: acquiring corner points by using goodffeatureToTrack of an OpenCV library, and calculating average points of the corner points to further determine the motion direction;
and step 3: and drawing the driving road strength of the automobile according to the automobile motion direction and the automobile central point, and simulating the automobile body into a straight line.
The turn signal lamp identification module identifies the position of the turn signal lamp by using the YOLO, and then identifies the color of the turn signal lamp by using the OpenCV, and the specific implementation process is as follows:
step 1: acquiring image data acquired by a camera, inputting the image data into a YOLO (YOLO), carrying out target detection, and acquiring a target list;
step 2: checking whether the target type of the turn light exists in the target list;
and step 3: extracting a picture area with a target of a turn light by utilizing OpenCV, converting the color of each pixel point from RGB space to HSV space, and identifying whether the turn light with yellow color exists or not.
The angle detection module is used for calculating the angles of two straight lines of an automobile and a lane line, performing edge detection on the image by using a Canny operator, counting Hough LinesP transformed by Hough, solving all lines, and calculating the maximum included angle between vectors; the Canny edge detection is implemented by carrying out convolution on the image through a Canny function by utilizing a sober operator, and the gradient of the image in the x direction and the y direction is calculated.
The pressing line detection module is used for detecting the pressing line behavior of the running automobile so as to judge the suspicion of drunk driving of the automobile; namely, when the automobile runs on a solid lane on a road, if the automobile has a linear running behavior, the automobile is judged to be suspected to be drunk to drive; specifically, whether a lane line in a video set under a monitoring camera is a solid line or not is judged, and whether a line pressing behavior exists on a road when an automobile runs is judged, wherein the specific judgment process is as follows:
step 1: judging by a solid dotted line, and defining an interval through an ROI (region of interest), namely forming a rectangular region for the four acquired points, wherein a lane line in the image needs to be framed in the region, and counting the length of the ROI; then, a straight line composed of pixel points in the image, namely a lane line, is found out by utilizing Hough transformation; further, storing pixel points of each row with white lane lines to corresponding positions of the list, judging whether the first row of the list is a non-blank area, if so, storing the lane lines at the initial position of the list, and recording the initial position s; otherwise, directly judging that the lane line is a dotted line; then judging where the position of the list has no lane line area, and recording an end position s'; then judging whether the list only has a section of continuous area with a lane line; if not, directly judging that the lane line is a dotted line; and finally, judging whether the distance difference between the end position and the initial position of the lane line is greater than 60% of the ROI length, if so, judging that the lane line is a solid line, and otherwise, judging that the lane line is a dotted line.
Step 2: judging the line pressing behavior, counting the width of the automobile which is judged to be driven on the solid line road in the video set, marking as x0, and then counting the distance between the center line of the automobile body and the right side and the left side of the road, and marking as xr and xl; judging whether the absolute value of the difference value of xr and xl is approximately equal to 0, if so, judging that the vehicle runs normally; otherwise, judging whether xr or xl is less than or equal to x0/2, if so, judging that the vehicle runs under the condition of pressing the line, and otherwise, judging that the vehicle runs normally.
The drunk driving vehicle identification module is used for confirming suspected drunk driving vehicles, and if the vehicle driving characteristic detection module confirms lane change and lighting; the change rate of the angle formed by the central line of the automobile body and the lane line exceeds the average value when the automobile changes lanes; the automobile runs on the middle of the two lanes by pressing a line; when the vehicle is waiting for the traffic light and is the first vehicle, the vehicle head presses the line, and the suspected drunk driving vehicle is confirmed.
The license plate recognition module is used for carrying out character segmentation and character recognition by detecting the license plate so as to acquire the number of the license plate of the vehicle, and the specific implementation process is as follows:
step 1: detecting a license plate, deleting irrelevant details in an image by using bilateral filtering, carrying out edge detection, calculating a gradient by using a Canny detection principle, finding out a contour, traversing a result, masking a picture, and leaving a license plate part;
step 2: character segmentation, cutting out an image of the region of interest;
and step 3: and character recognition, namely graying the image, performing edge detection by using a Sobel operator, performing image corrosion, filling and filtering processing, extracting a license plate, and performing character segmentation and recognition.
And the vehicle license plate number is sent to the traffic police manual detection module, and the vehicle license plate number identified by the vehicle license plate identification module is sent to the traffic police for manual detection and confirmation, so that the drunk driving vehicle identification process is completed.
The above embodiments are preferred embodiments of the present invention, but the present invention is not limited to the above embodiments, and any other changes, modifications, substitutions, combinations, and simplifications which do not depart from the spirit and principle of the present invention should be construed as equivalents thereof, and all such changes, modifications, substitutions, combinations, and simplifications are intended to be included in the scope of the present invention.

Claims (5)

1. A drunk driving vehicle track recognition system based on video monitoring is characterized by comprising a video acquisition module, a digital image preprocessing module, a vehicle running characteristic detection module, a drunk driving vehicle recognition module, a license plate recognition module and a traffic police reporting artificial detection module; the digital image preprocessing module comprises a lane line identification module, an automobile identification module and an image preprocessing module; the vehicle running characteristic detection module comprises an automobile tracking module, a steering lamp identification module, an angle detection module and a line pressing detection module; the system comprises a video acquisition module, a digital image preprocessing module, a vehicle running characteristic detection module, a drunk driving vehicle identification module, a license plate identification module and a traffic police reporting artificial detection module, wherein all the modules are connected in sequence;
the digital image preprocessing module is used for acquiring an original image of a vehicle by using a contour recognition method for the vehicle in the monitoring video, segmenting the vehicle and lane lines in the image, and carrying out image preprocessing on the acquired image, wherein the image preprocessing comprises filtering, binaryzation, filling, modification and thinning processing to acquire a contour image containing the vehicle and the lane lines;
the specific implementation processes of the lane line identification module, the automobile identification module and the image preprocessing module in the digital image preprocessing module are as follows:
the lane line identification module is used for performing edge detection by using a Sobel operator, a Laplacian operator and a Canny operator, performing region-of-interest detection by using fillConvexColy, outputting an element matrix, drawing a mask image and acquiring a region-of-interest; independent pixel points forming the lane line are connected into a straight line by Hough transformation, and then the addWeighted function is used for realizing image weighted superposition;
the automobile identification module is used for judging an automobile and locking the position of the automobile; the method comprises the following steps of identifying by using an algorithm combining a Histogram of Oriented Gradients (HOG) and a Support Vector Machine (SVM), and specifically realizing the following steps:
s11, generating feature data of the image by extracting a Histogram of Oriented Gradients (HOG) descriptor of the image, and normalizing HOG feature vectors of the histogram of oriented gradients; the calculation principle of the directional gradient is as follows:
G(x,y)＝gx(x,y)+gy(x,y) (1)
gx(x,y)＝I(x+1,y)-I(x,y) (2)
gy(x,y)＝I(x,y+1)-I(x,y) (3)
Figure FDA0003651685280000011
β＝arctan(gy/gx) (5)
wherein, (x, y) is the coordinates of the pixel; g (x, y) is the sum of horizontal and vertical gradients at a pixel point (x, y) in the input image; gx(x, y) is the horizontal gradient at the pixel point (x, y) in the input image; g is a radical of formulay(x, y) is the vertical gradient at the pixel point (x, y) in the input image; i (x, y) is a pixel value of the image at (x, y), I (x +1, y) is a pixel value of the image at (x +1, y); i (x, y +1) is the pixel value of the image at (x, y + 1); g is the gradient amplitude of the pixel point; gxThe gradient of the pixel points in the horizontal direction is obtained; gyThe gradient of the pixel points in the vertical direction is obtained; beta is the gradient direction of the pixel point;
s12, connecting HOG features in series, dividing the image into a plurality of connection areas, and performing convolution kernel operations on each connection area in the x and y directions to obtain a gradient integral image;
s13, regularizing the gradient vector by calculating feature dimensions to obtain final HOG features;
s14, judging the automobile by combining with an SVM classifier, locking the position of the automobile, establishing a rectangular coordinate system for the image, detecting and acquiring the center point of the automobile, and acquiring the current coordinate and the width of the automobile;
the image preprocessing module is used for framing the video and enabling the video to perform the next image processing in the form of each frame; preprocessing an image, removing noise points of each frame of image by Gaussian filtering, performing graying processing by utilizing gray level transformation, converting an RGB image into a gray level image by a cvtColor function, and performing binarization processing on the obtained image;
the vehicle driving characteristic detection module is used for detecting an included angle between a vehicle body and a lane line; the automobile runs through the pressing line; when the automobile waits for the traffic light and is the first automobile, the line pressing condition of the automobile head is met;
the automobile tracking module, the turn light identification module, the angle detection module and the line pressing detection module in the vehicle running characteristic detection module are specifically realized in the following processes:
the automobile tracking module acquires the automobile motion direction and the automobile motion line by combining Kalman filtering and Hungary algorithm, and the specific realization process is as follows:
s21, setting the coordinates (x, y) of the current center point of the automobile, and setting the coordinates (x) of the center point at the last momentt-1,yt-1) Then its euclidean distance:
Figure FDA0003651685280000021
Figure FDA0003651685280000022
wherein D ist|t-1A coordinate set of the central point detected at the last moment;
s22, acquiring corner points by using goodFeatureToTrack of an OpenCV library, and calculating average points of the corner points to further determine the motion direction;
s23, drawing the driving road strength of the automobile according to the automobile movement direction and the automobile central point, and simulating the automobile body into a straight line;
the turn signal lamp identification module identifies the position of the turn signal lamp by using the YOLO, and then identifies the color of the turn signal lamp by using the OpenCV, and the specific implementation process is as follows:
s31, acquiring image data acquired by the camera, inputting the image data into YOLO, carrying out target detection, and acquiring a target list;
s32, checking the target type of the turn light in the target list;
s33, extracting a picture area with a target as a turn light by utilizing OpenCV, converting the color of each pixel point from RGB space to HSV space, and identifying a yellow turn light;
the angle detection module is used for calculating the angles of two straight lines of an automobile and a lane line, performing edge detection on the image by using a Canny operator, counting Hough LinesP transformed by Hough, solving all lines, and calculating the maximum included angle between vectors; the Canny edge detection is carried out on the image by utilizing a sober operator through a Canny function, and the gradient of the image in the x direction and the y direction is calculated;
the pressing line detection module is used for detecting the pressing line behavior of the automobile running and further judging the suspicion of the automobile drunk driving.
2. The drunk-driving vehicle track recognition system based on video monitoring as claimed in claim 1, wherein the video acquisition module is used for acquiring video image data of the automobile, learning an included angle between an automobile body line and a lane line when the automobile on the road changes lanes by the monitoring camera, and counting a maximum threshold value.
3. The drunk-driving vehicle track recognition system based on video monitoring as claimed in claim 1, wherein the drunk-driving vehicle recognition module is used for confirming drunk-driving vehicles, and if lane-changing lighting is confirmed through the vehicle driving characteristic detection module; the change rate of the angle formed by the central line of the automobile body and the lane line exceeds the average value when the automobile changes lanes; the automobile runs in the middle of the two lanes by pressing a line; when the vehicle is waiting for the traffic light and is the first vehicle, the vehicle head presses the line, and the drunk driving vehicle is determined.
4. The drunk-driving vehicle track recognition system based on video monitoring as claimed in claim 1, wherein the license plate recognition module is used for obtaining the number of the vehicle license plate by detecting the license plate, performing character segmentation and character recognition, and the specific implementation process is as follows:
s41, detecting the license plate, deleting other details in the image by using bilateral filtering, carrying out edge detection, calculating gradient by using a Canny detection principle, finding out a contour, traversing a result, masking a picture, and leaving a license plate part;
s42, segmenting characters, and cutting out an image of the region of interest;
and S43, character recognition, namely graying the image, performing edge detection by using a Sobel operator, performing image corrosion, filling and filtering processing, extracting the license plate, and performing character segmentation and recognition.
5. The drunk-driving vehicle track recognition system based on video monitoring as recited in claim 1, wherein a traffic police manual detection module is reported, and a license plate number recognized by the license plate recognition module is sent to a traffic police for manual detection and confirmation, so that the drunk-driving vehicle recognition process is completed.
CN202110481444.1A 2021-04-30 2021-04-30 Drunk driving vehicle track recognition system based on video monitoring Active CN113205687B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110481444.1A CN113205687B (en) 2021-04-30 2021-04-30 Drunk driving vehicle track recognition system based on video monitoring

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110481444.1A CN113205687B (en) 2021-04-30 2021-04-30 Drunk driving vehicle track recognition system based on video monitoring

Publications (2)

Publication Number Publication Date
CN113205687A CN113205687A (en) 2021-08-03
CN113205687B true CN113205687B (en) 2022-07-05

Family

ID=77028117

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110481444.1A Active CN113205687B (en) 2021-04-30 2021-04-30 Drunk driving vehicle track recognition system based on video monitoring

Country Status (1)

Country Link
CN (1) CN113205687B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111047908B (en) * 2018-10-12 2021-11-02 富士通株式会社 Detection device and method for cross-line vehicle and video monitoring equipment

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104408932B (en) * 2014-11-03 2016-08-24 河海大学常州校区 A kind of drunk driving vehicle detecting system based on video monitoring
CN106004884A (en) * 2016-07-11 2016-10-12 南昌工学院 Method and system for realizing real-time identification and danger judgment of road conditions based on complex sensing
CN106774328A (en) * 2016-12-26 2017-05-31 广州大学 A kind of automated driving system and method based on road Identification
CN108091136A (en) * 2017-12-14 2018-05-29 阜阳裕晟电子科技有限公司 A kind of drunk driving vehicle detecting system based on video monitoring
CN112257539A (en) * 2020-10-16 2021-01-22 广州大学 Method, system and storage medium for detecting position relation between vehicle and lane line
CN112580736A (en) * 2020-12-26 2021-03-30 浙江天行健智能科技有限公司 Drunk driving vehicle identification method based on SVM algorithm

Also Published As

Publication number Publication date
CN113205687A (en) 2021-08-03

Similar Documents

Publication Publication Date Title
CN106709436B (en) Track traffic panoramic monitoring-oriented cross-camera suspicious pedestrian target tracking system
Malik Fast vehicle detection with probabilistic feature grouping and its application to vehicle tracking
CN108564814B (en) Image-based parking lot parking space detection method and device
CN109299674B (en) Tunnel illegal lane change detection method based on car lamp
CN109740595B (en) Oblique vehicle detection and tracking system and method based on machine vision
US8902053B2 (en) Method and system for lane departure warning
CN112819094B (en) Target detection and identification method based on structural similarity measurement
CN109800752B (en) Automobile license plate character segmentation and recognition algorithm based on machine vision
CN108256521B (en) Effective area positioning method for vehicle body color identification
WO2010006361A1 (en) Detection of vehicles in images of a night time scene
CN102880863B (en) Method for positioning license number and face of driver on basis of deformable part model
Bedruz et al. Real-time vehicle detection and tracking using a mean-shift based blob analysis and tracking approach
Kortli et al. A novel illumination-invariant lane detection system
CN109190483B (en) Lane line detection method based on vision
CN106887004A (en) A kind of method for detecting lane lines based on Block- matching
Siogkas et al. Random-walker monocular road detection in adverse conditions using automated spatiotemporal seed selection
Cai et al. Real-time arrow traffic light recognition system for intelligent vehicle
CN103324958B (en) Based on the license plate locating method of sciagraphy and SVM under a kind of complex background
CN111553214B (en) Method and system for detecting smoking behavior of driver
Chang et al. An efficient method for lane-mark extraction in complex conditions
Kortli et al. Efficient implementation of a real-time lane departure warning system
CN101369312B (en) Method and equipment for detecting intersection in image
CN113205687B (en) Drunk driving vehicle track recognition system based on video monitoring
Devane et al. Lane detection techniques using image processing
CN106803064B (en) Traffic light rapid identification method

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