CN108986476A - Motor vehicle does not use according to regulations high beam recognition methods, system and storage medium - Google Patents
Motor vehicle does not use according to regulations high beam recognition methods, system and storage medium Download PDFInfo
- Publication number
- CN108986476A CN108986476A CN201810889659.5A CN201810889659A CN108986476A CN 108986476 A CN108986476 A CN 108986476A CN 201810889659 A CN201810889659 A CN 201810889659A CN 108986476 A CN108986476 A CN 108986476A
- Authority
- CN
- China
- Prior art keywords
- high beam
- target area
- motor vehicle
- camera
- video stream
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 40
- 230000033228 biological regulation Effects 0.000 title claims abstract description 14
- 238000001514 detection method Methods 0.000 claims abstract description 38
- 230000009466 transformation Effects 0.000 claims abstract description 25
- 238000005286 illumination Methods 0.000 claims abstract description 24
- 238000012544 monitoring process Methods 0.000 claims abstract description 4
- 230000005764 inhibitory process Effects 0.000 claims description 34
- 238000012549 training Methods 0.000 claims description 21
- 238000003062 neural network model Methods 0.000 claims description 16
- 230000008569 process Effects 0.000 claims description 8
- 238000013135 deep learning Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 5
- 238000003384 imaging method Methods 0.000 claims description 3
- 230000001629 suppression Effects 0.000 claims description 3
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 238000009434 installation Methods 0.000 abstract 1
- KBPHJBAIARWVSC-RGZFRNHPSA-N lutein Chemical compound C([C@H](O)CC=1C)C(C)(C)C=1\C=C\C(\C)=C\C=C\C(\C)=C\C=C\C=C(/C)\C=C\C=C(/C)\C=C\[C@H]1C(C)=C[C@H](O)CC1(C)C KBPHJBAIARWVSC-RGZFRNHPSA-N 0.000 abstract 1
- 229960005375 lutein Drugs 0.000 abstract 1
- KBPHJBAIARWVSC-XQIHNALSSA-N trans-lutein Natural products CC(=C/C=C/C=C(C)/C=C/C=C(C)/C=C/C1=C(C)CC(O)CC1(C)C)C=CC=C(/C)C=CC2C(=CC(O)CC2(C)C)C KBPHJBAIARWVSC-XQIHNALSSA-N 0.000 abstract 1
- FJHBOVDFOQMZRV-XQIHNALSSA-N xanthophyll Natural products CC(=C/C=C/C=C(C)/C=C/C=C(C)/C=C/C1=C(C)CC(O)CC1(C)C)C=CC=C(/C)C=CC2C=C(C)C(O)CC2(C)C FJHBOVDFOQMZRV-XQIHNALSSA-N 0.000 abstract 1
- 235000008210 xanthophylls Nutrition 0.000 abstract 1
- 230000000694 effects Effects 0.000 description 5
- 238000011835 investigation Methods 0.000 description 5
- 206010039203 Road traffic accident Diseases 0.000 description 4
- 230000009471 action Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 241001622623 Coeliadinae Species 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
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
- G06V20/40—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
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- 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
- G06V20/63—Scene text, e.g. street names
-
- 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
- G08G1/04—Detecting movement of traffic to be counted or controlled using optical or ultrasonic detectors
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Traffic Control Systems (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The present invention announces a kind of motor vehicle and does not use according to regulations high beam recognition methods, system and storage medium, is related to the technical field of artificial intelligence machine vision, can solve currently to investigate and prosecute method larger workload, and there is technical issues that.Including being used for traffic monitoring using two cameras and installation, it obtains described two cameras and distinguishes collected video flowing, car light extra bus board objective cross detection mode is taken in the strong Xanthophyll cycle video flowing, the position for obtaining target area obtains the ground coordinate position of the target area using perspective inverse transformation；Perspective inverse transformation is recycled to obtain the target area in ordinary video flow；The intensity of illumination for detecting the target area in the ordinary video flow judges the state that high beam is opened, and according to the state that high beam is opened, judges whether the corresponding motor vehicle of the license plate uses according to regulations its high beam.Whether the present invention can be with intelligent recognition motor vehicle by proper use of high beam is provided, convenient for evidence obtaining, securely and reliably.
Description
Technical Field
The invention relates to the technical field of artificial intelligence machine vision, in particular to a method and a system for identifying that a motor vehicle uses a high beam lamp without a specified use and a storage medium.
Background
In recent years, the road traffic industry of China is rapidly developed, and meanwhile, the problem of road traffic safety is increasingly prominent. In particular, the use of high beam lights against regulations during night driving has been a major cause of traffic accidents at night. It is known that the abuse of the high beam light can cause the driver of the opposite running vehicle to blindly instantly visually, interfere the sight of the driver of the same running vehicle, reduce the observation ability of the driver for the surrounding pedestrians, and reduce the perception and judgment of the speed, distance and width of the coming vehicle.
Abusive high beam causes many traffic accidents, but abusive high beam is difficult to check, and once a driver takes light-changing measures, law enforcement on site is difficult to fix evidence, and law enforcement disputes are easy to cause because the behavior of abusive high beam has variability, dynamics and instantaneity. And the abusive high beam causes traffic accidents of other vehicles, and the vehicles do not directly participate in the collision, so the vehicles causing the accidents can leave the scene quickly without leaving traces, which also leads to the illegal action who should undertake the accident responsibility.
At present, when a public security department carries out illegal activities of using a high beam lamp when a place is checked, the illegal activities mainly depend on road surface law enforcement, traffic policemen observe the behaviors on site and intercept and obtain evidence, the place checking mode is passive, the technical content is low, a large amount of police force is needed to participate, the workload is high, the safety of the traffic policemen is influenced to a certain extent, and the treatment effect is not obvious because the place cannot be checked continuously.
Disclosure of Invention
Technical problem to be solved
Aiming at the defects of the prior art, the invention provides a method, a system and a storage medium for identifying a high beam lamp which is not used by a motor vehicle according to the regulations, and solves the technical problems of large workload, time consumption, labor consumption and potential safety hazard of the currently used high beam lamp inspection and treatment method.
(II) technical scheme
In order to achieve the purpose, the invention is realized by the following technical scheme:
a method for intelligently identifying that a motor vehicle does not use a high beam according to regulations based on deep learning comprises the following steps:
two camera devices are adopted, wherein one camera device adopts a strong light suppression camera, and the other camera device adopts a common camera; installing the two camera devices for traffic monitoring, enabling the two camera devices to detect the same object, calibrating the two camera devices, measuring the relation between a ground coordinate system and the pixels of the camera devices, and calculating perspective transformation coefficients of the two camera devices;
acquiring video streams respectively acquired by the two camera devices, wherein the video streams comprise multi-frame images; the highlight inhibition video stream is collected by the highlight inhibition camera, and the common video stream is collected by the common camera;
detecting the high beam of each motor vehicle in each frame image in the strong light inhibition video stream by adopting a pre-trained high beam recognition neural network model to obtain the position of a target area, and obtaining the ground coordinate position of the target area by utilizing perspective inverse transformation according to the obtained position of the target area;
obtaining a pixel coordinate position in a common video stream by utilizing perspective inverse transformation according to the ground coordinate position of the target area, and determining the target area in the common video stream;
and detecting the illumination intensity of a target area in the ordinary video stream, comparing the illumination intensity with a preset illumination intensity threshold value, judging the starting state of the high beam, and judging whether the motor vehicle corresponding to the license plate uses the high beam according to the specified state of the high beam.
Further, the distance between the two cameras and the detection area is larger than the farthest distance which can be irradiated by the dipped headlight of the motor vehicle.
Further, the detecting the illumination intensity of the target area in the normal video stream specifically includes:
selecting an image in which a target area is located in a common video stream, and evaluating the light intensity level by analyzing a high-definition image color histogram and utilizing the characteristics of corresponding point gray values of R, G, B three color channels;
let Vi＝(Ri,Gi,Bi) For a color space vector of pixel i in an image, define the level of red bias for pixel i:
rismaller indicates that pixel i is more red-shifted, and if pixel i belongs to a high beam, then ri∈[0,1](ii) a Evaluating the light intensity level through the reddening level of a certain area in the image;
define reddish level of region S:
then, a light intensity threshold Ta is defined, if r > Ta indicates that the high beam is turned on, otherwise, the high beam is not turned on.
Further, the calibration is performed on the two image capturing devices, and the calibration method includes:
and drawing a rectangular frame at a preset distance between the camera equipment and the detection area on the ground, and respectively calculating perspective transformation coefficients of the two camera equipment by utilizing the corresponding relation between the imaging positions of the rectangular frame in the two camera equipment and the ground coordinates of the actual rectangular frame.
Further, the training process of the high beam recognition neural network model includes:
constructing a training sample set, wherein the training sample set comprises a plurality of pictures for marking combination of car lights and license plates, and the pictures are obtained from a highlight inhibition video stream;
and training a plurality of pictures in the training sample set by adopting a yolo algorithm and taking the combination of the vehicle lamp and the license plate as training data to obtain each model parameter of the high beam recognition neural network model.
Further, the distance between the two camera devices and the detection area is 50 meters.
An intelligent recognition system for motor vehicle non-use-as-specified high beam based on deep learning, comprising: the device comprises an acquisition module, a first detection module, a second detection module and a judgment module; wherein,
the acquisition module comprises two camera devices, wherein one camera device adopts a strong light inhibition camera and is used for detecting the automobile lamps and license plates of the motor vehicle; the other camera device adopts a common camera for detecting the light intensity degree of the lamp light of the motor vehicle;
the acquisition module is used for acquiring video streams respectively acquired by the two cameras, the highlight inhibition cameras acquire highlight inhibition video streams, and the common cameras acquire common video streams; the first detection module is used for detecting the high beam of each motor vehicle in each frame image in the strong light inhibition video stream by adopting a pre-trained high beam recognition neural network model in the strong light inhibition video stream to obtain the position of a target area, and obtaining the ground coordinate position of the target area by utilizing perspective inverse transformation according to the obtained position of the target area;
the second detection module is used for obtaining a pixel coordinate position in the common video stream by utilizing perspective inverse transformation according to the ground coordinate position of the target area obtained by the first detection module, and determining the target area in the common video stream;
the judgment module is used for firstly detecting the illumination intensity of the target area in the ordinary video stream determined by the second detection module, comparing the illumination intensity with a preset illumination intensity threshold value, judging the starting state of the high beam, and judging whether the motor vehicle corresponding to the license plate uses the high beam according to the specified state of the high beam.
In a third aspect, the present invention provides a storage medium having stored thereon a computer program which, when executed by a processor, is operable to carry out the method described above.
(III) advantageous effects
The invention provides a method, a system and a storage medium for recognizing a high beam lamp which is not used by a motor vehicle according to regulations, wherein two camera devices are adopted, one camera device adopts a strong light inhibition camera for detecting a lamp and a license plate of the motor vehicle, and the other camera device adopts a common camera for detecting the light intensity degree of the lamp light of the lamp of the motor vehicle; the method comprises the steps of acquiring video streams respectively acquired by two camera devices, detecting a position of a target area by adopting a pre-trained high beam recognition neural network model based on a yolo algorithm in a highlight inhibition video stream, obtaining the target area in a common video stream by utilizing reverse perspective transformation, detecting the illumination intensity of the target area in the common video stream, judging the starting state of a high beam, and further judging whether a motor vehicle corresponding to a license plate uses the high beam according to the regulation. The invention can intelligently identify whether the motor vehicle correctly uses the high beam according to the regulations, is convenient for obtaining evidence, reduces the difficulty of the investigation of the high beam at traffic control places, increases the management intensity of the use of the high beam, prevents various traffic accidents caused by incorrect use of the high beam during driving on a highway, reduces traffic casualties and traffic injuries, and practically ensures the life and property safety of the nation, the society and the masses.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a schematic flow chart of a method according to an embodiment of the present invention;
FIG. 2 is a schematic distance diagram of a camera from a detection area according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a corresponding relationship between ground coordinates of four endpoints of a rectangular frame and images formed by two cameras in the calibration method according to the embodiment of the present invention;
FIG. 4 is a schematic diagram of a combined detection method for vehicle lights and license plates according to an embodiment of the present invention;
fig. 5 is a schematic view of a video stream detection process according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example (b):
because the current public security office traffic control department mainly relies on the road surface to enforce law when the act of offence that uses the high beam lamp according to the regulation is not in the investigation department, through the on-the-spot observation of traffic police, intercept and collect evidence again, the mode of investigation department is comparatively passive, and technical content is lower, needs a large amount of police force to participate in, and work load is very big, also has certain influence to traffic police's safety, owing to can not continuous investigation department, the treatment effect is also not showing in addition.
In view of the above, the present embodiment provides a method for recognizing a high beam not used as intended by a motor vehicle, as shown in fig. 1, comprising the steps of:
s101, two camera devices are adopted, wherein one camera device adopts a highlight suppression camera, and the other camera device adopts a common camera; installing the two camera devices for traffic monitoring, enabling the two camera devices to detect the same object, calibrating the two camera devices, measuring the relation between a ground coordinate system and the pixels of the camera devices, and calculating perspective transformation coefficients of the two camera devices;
s102, video streams respectively collected by the two camera devices are obtained, and the video streams comprise multi-frame images; the highlight inhibition video stream is collected by the highlight inhibition camera, and the common video stream is collected by the common camera;
s103, detecting the high beam of each motor vehicle in each frame image in the strong light inhibition video stream by adopting a pre-trained high beam recognition neural network model to obtain the position of a target area, and obtaining the ground coordinate position of the target area by utilizing perspective inverse transformation according to the obtained position of the target area;
s104, obtaining a pixel coordinate position in a common video stream by utilizing perspective inverse transformation according to the ground coordinate position of the target area, and determining the target area in the common video stream;
s105, detecting the illumination intensity of a target area in the ordinary video stream, comparing the illumination intensity with a preset illumination intensity threshold value, judging the starting state of the high beam, and judging whether the motor vehicle corresponding to the license plate uses the high beam according to the specified state of the high beam.
The invention is described in detail below with reference to the accompanying drawings:
since the detection of the high beam is often interfered by the dipped headlight, in order to distinguish the influence of the dipped headlight from the influence of the high beam, as shown in fig. 2, the distance W between the two cameras and the detection area is set to be greater than the farthest distance that the dipped headlight of the motor vehicle can illuminate, and the effective detection range of the embodiment is set to be about 50 meters, because the coverage range of the ordinary dipped headlight is 0-40 meters, and the coverage range of the high beam is more than 80 meters; the detection at the position of 50 meters can effectively remove the interference of the dipped headlights.
The embodiment of the invention adopts double cameras, wherein one camera is a strong light inhibition camera and is used for detecting the car light, the license plate and the car distance; the other camera is a common camera used for detecting the light intensity degree of the high beam. In order to enable the two cameras to detect the same object, the two cameras need to be calibrated. And the calibration is used for measuring the relation between a ground coordinate system and the pixels of the camera, and the calibration method comprises the steps of drawing a rectangular frame at a position 50 meters away from the camera on the ground after the camera is installed, and calculating the perspective transformation coefficient of the camera by utilizing the corresponding relation between the imaging of the rectangular frame and the ground coordinate of the actual rectangular frame.
The calibration method specifically includes that as shown in fig. 3, four end points a, b, c, d of a rectangular frame, corresponding pixels in two cameras are a1, b1, c1, d1 and a2, b2, c2, d 2; then, the coefficients of perspective transformation of the two cameras can be obtained according to the matrix of perspective transformation.
Because the light interference in the city at night is large, in order to avoid the interference of other light rays, the embodiment of the invention not only detects the light, but also adopts a mode of combining the car light and the license plate for detection, as shown in fig. 4. Only the left and right lamps and the middle license plate can be seen at the same time, and the combination of the lamps and the license plate is trained and detected by adopting the yolo network in the video stream acquired by the highlight inhibition camera. The yolo network is a very effective target detection tool, the vehicle lamp and license plate combination is used as training data, then training is carried out according to the training steps, and the obtained network model can be used for detecting the combination of the lamp and the license plate.
As shown in fig. 5, the video stream detection process according to the embodiment of the present invention is as follows:
firstly, detecting the target combination of the vehicle lamp and the license plate by using a yolo algorithm in a strong light inhibition video stream, wherein the method can effectively remove urban stray light interference;
firstly, a high beam recognition neural network model is trained in advance, and the training process of the high beam recognition neural network model comprises the following steps: constructing a training sample set, wherein the training sample set comprises a plurality of pictures for marking combination of car lights and license plates, and the pictures are obtained from a highlight inhibition video stream; training a plurality of pictures in the training sample set by adopting a yolo algorithm and taking a vehicle lamp and license plate combination as training data to obtain each model parameter of the high beam recognition neural network model so as to obtain the high beam recognition neural network model;
then adopting a pre-trained high beam recognition neural network model based on the yolo algorithm in the strong light inhibition video stream to detect the position of a target area obtained by the vehicle lamp and license plate target combination,
obtaining the position of a target area, then obtaining the ground coordinate position of the target area by utilizing perspective inverse transformation, and then obtaining the pixel coordinate in the common video stream by utilizing the perspective inverse transformation;
detecting the illumination intensity of a target area from the common video stream, if the intensity is greater than a threshold value T, considering that the high beam is turned on, otherwise, not considering that the high beam is turned on; and judging whether the motor vehicle corresponding to the license plate uses the high beam according to the specified rule or not according to the starting state of the high beam.
Further, the light intensity detection steps of the target area of the normal video stream according to the embodiment of the present invention are as follows:
through analysis of a high-definition image color histogram, the light intensity level can be evaluated by utilizing the characteristics of corresponding point gray values of R, G, B three color channels;
let Vi＝(Ri,Gi,Bi) For a color space vector of pixel i in an image, define the level of red bias for pixel i:
rismaller indicates that pixel i is more red-shifted, and if pixel i belongs to a high beam, then ri∈[0,1](ii) a The embodiment of the invention evaluates the light intensity level through the reddish level of a certain area in the image, defines the reddish level of the area S:
and then defining a light intensity threshold Ta, if r is greater than Ta, indicating that the high beam is turned on, otherwise, considering that the high beam is not turned on, and finally, judging whether the vehicle uses the high beam according to the rule according to the turning-on state of the high beam of the target vehicle. The threshold Ta is selected mainly according to experimental parameters of collected samples, for example, 1 ten thousand images are collected, then r1 for turning on the high beam and r2 for turning off the high beam are respectively calculated, and then Ta is (r1+ r 2)/2.
Since the high beam is not allowed to be turned on at the relevant road section, the camera device of the embodiment is arranged at the road section where the high beam is not allowed to be turned on, so that the traffic violation of the vehicle driver can be determined only by detecting and identifying the turning-on state of the high beam, namely whether the motor vehicle driver uses the high beam illegally according to the regulations.
In a second aspect, an embodiment of the present invention further provides an identification system for a motor vehicle that does not use a high beam as specified, including: the device comprises an acquisition module, a first detection module, a second detection module and a judgment module; wherein,
the acquisition module comprises two camera devices, wherein one camera device adopts a strong light inhibition camera and is used for detecting the automobile lamps and license plates of the motor vehicle; the other camera device adopts a common camera for detecting the light intensity degree of the lamp light of the motor vehicle;
the acquisition module is used for acquiring video streams respectively acquired by the two cameras, the highlight inhibition cameras acquire highlight inhibition video streams, and the common cameras acquire common video streams;
the first detection module is used for detecting the high beam of each motor vehicle in each frame image in the strong light inhibition video stream by adopting a pre-trained high beam recognition neural network model in the strong light inhibition video stream to obtain the position of a target area, and obtaining the ground coordinate position of the target area by utilizing perspective inverse transformation according to the obtained position of the target area;
the second detection module is used for obtaining a pixel coordinate position in the common video stream by utilizing perspective inverse transformation according to the ground coordinate position of the target area obtained by the first detection module, and determining the target area in the common video stream;
the judgment module is used for firstly detecting the illumination intensity of the target area in the ordinary video stream determined by the second detection module, comparing the illumination intensity with a preset illumination intensity threshold value, judging the starting state of the high beam, and judging whether the motor vehicle corresponding to the license plate uses the high beam according to the specified state of the high beam.
It is understood that the system provided by the embodiment of the present invention corresponds to the method provided by the embodiment of the present invention, and the explanation, the example and the beneficial effects of the related contents can refer to the corresponding parts in the method.
In a third aspect, an embodiment of the present invention further provides a storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method described above can be implemented.
In summary, the embodiment of the invention adopts two cameras, wherein one camera is a strong light inhibition camera and is used for detecting the car light, the license plate and the car distance; another camera is the light intensity degree that ordinary camera is used for detecting the high beam, and the analysis the video stream that two camera equipment gathered respectively locks the target vehicle to detect the illumination intensity of target area correspondence image, judge the state that the target vehicle high beam was opened, whether can intelligent recognition motor vehicle correctly uses the high beam according to the regulation, the collection evidence of being convenient for alleviates the traffic control department and locates the degree of difficulty to the investigation of high beam. The embodiment of the invention sets an effective detection range to remove the interference of the dipped headlights, and simultaneously adopts a combined detection mode of the car lights and the vehicle license plate to ensure that the measurement result is more accurate and efficient.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention 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 solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (8)
1. A motor vehicle intelligent recognition method for not using high beam according to regulations based on deep learning is characterized by comprising the following steps:
two camera devices are adopted, wherein one camera device adopts a strong light suppression camera, and the other camera device adopts a common camera; installing the two camera devices for traffic monitoring, enabling the two camera devices to detect the same object, calibrating the two camera devices, measuring the relation between a ground coordinate system and the pixels of the camera devices, and calculating perspective transformation coefficients of the two camera devices;
acquiring video streams respectively acquired by the two camera devices, wherein the video streams comprise multi-frame images; the highlight inhibition video stream is collected by the highlight inhibition camera, and the common video stream is collected by the common camera;
detecting the high beam of each motor vehicle in each frame image in the strong light inhibition video stream by adopting a pre-trained high beam recognition neural network model to obtain the position of a target area, and obtaining the ground coordinate position of the target area by utilizing perspective inverse transformation according to the obtained position of the target area;
obtaining a pixel coordinate position in a common video stream by utilizing perspective inverse transformation according to the ground coordinate position of the target area, and determining the target area in the common video stream;
and detecting the illumination intensity of a target area in the ordinary video stream, comparing the illumination intensity with a preset illumination intensity threshold value, judging the starting state of the high beam, and judging whether the motor vehicle corresponding to the license plate uses the high beam according to the specified state of the high beam.
2. The intelligent recognition method for the non-specified use of high beam lights of the motor vehicle based on the deep learning as claimed in claim 1, wherein the distance between the two cameras and the detection area is greater than the maximum distance that the low beam lights of the motor vehicle can illuminate.
3. The intelligent recognition method for the motor vehicle not using the high beam as specified in claim 2, wherein the detecting the illumination intensity of the target area in the normal video stream specifically comprises:
selecting an image in which a target area is located in a common video stream, and evaluating the light intensity level by analyzing a high-definition image color histogram and utilizing the characteristics of corresponding point gray values of R, G, B three color channels;
let Vi＝(Ri,Gi,Bi) Defining the red bias of a pixel i for a color space vector of the pixel i in an imageLevel:
rismaller indicates that pixel i is more red-shifted, and if pixel i belongs to a high beam, then ri∈[0,1]；
Evaluating the light intensity level through the reddening level of a certain area in the image;
define reddish level of region S:
then, a light intensity threshold Ta is defined, if r > Ta indicates that the high beam is turned on, otherwise, the high beam is not turned on.
4. The intelligent recognition method for the motor vehicle not using the high beam as specified in claim 1, wherein the calibration is performed on the two cameras, and the calibration method comprises the following steps:
and drawing a rectangular frame at a preset distance between the camera equipment and the detection area on the ground, and respectively calculating perspective transformation coefficients of the two camera equipment by utilizing the corresponding relation between the imaging positions of the rectangular frame in the two camera equipment and the ground coordinates of the actual rectangular frame.
5. The intelligent recognition method for the non-specified use of the high beam for the motor vehicle based on the deep learning of any one of claims 1 to 4, wherein the training process of the high beam recognition neural network model comprises the following steps:
constructing a training sample set, wherein the training sample set comprises a plurality of pictures for marking combination of car lights and license plates, and the pictures are obtained from a highlight inhibition video stream;
and training a plurality of pictures in the training sample set by adopting a yolo algorithm and taking the combination of the vehicle lamp and the license plate as training data to obtain each model parameter of the high beam recognition neural network model.
6. The intelligent recognition method for the non-specified use of the high beam of the motor vehicle based on the deep learning as claimed in claim 5, wherein the distance between the two camera devices and the detection area is 50 meters.
7. An intelligent recognition system for motor vehicles not using high beam as specified based on deep learning, comprising: the device comprises an acquisition module, a first detection module, a second detection module and a judgment module; wherein,
the acquisition module comprises two camera devices, wherein one camera device adopts a strong light inhibition camera and is used for detecting the automobile lamps and license plates of the motor vehicle; the other camera device adopts a common camera for detecting the light intensity degree of the lamp light of the motor vehicle;
the acquisition module is used for acquiring video streams respectively acquired by the two cameras, the highlight inhibition cameras acquire highlight inhibition video streams, and the common cameras acquire common video streams;
the first detection module is used for detecting the high beam of each motor vehicle in each frame image in the strong light inhibition video stream by adopting a pre-trained high beam recognition neural network model in the strong light inhibition video stream to obtain the position of a target area, and obtaining the ground coordinate position of the target area by utilizing perspective inverse transformation according to the obtained position of the target area;
the second detection module is used for obtaining a pixel coordinate position in the common video stream by utilizing perspective inverse transformation according to the ground coordinate position of the target area obtained by the first detection module, and determining the target area in the common video stream;
the judgment module is used for firstly detecting the illumination intensity of the target area in the ordinary video stream determined by the second detection module, comparing the illumination intensity with a preset illumination intensity threshold value, judging the starting state of the high beam, and judging whether the motor vehicle corresponding to the license plate uses the high beam according to the specified state of the high beam.
8. A storage medium having a computer program stored thereon, wherein the computer program when executed by a processor is adapted to implement the method of any of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810889659.5A CN108986476B (en) | 2018-08-07 | 2018-08-07 | method, system and storage medium for recognizing non-use of high beam by motor vehicle according to regulations |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810889659.5A CN108986476B (en) | 2018-08-07 | 2018-08-07 | method, system and storage medium for recognizing non-use of high beam by motor vehicle according to regulations |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108986476A true CN108986476A (en) | 2018-12-11 |
| CN108986476B CN108986476B (en) | 2019-12-06 |
Family
ID=64555397
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810889659.5A Active CN108986476B (en) | 2018-08-07 | 2018-08-07 | method, system and storage medium for recognizing non-use of high beam by motor vehicle according to regulations |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108986476B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110135343A (en) * | 2019-05-15 | 2019-08-16 | 南通大学 | A kind of street lamp intelligent measurement and night status judgment method |
| CN110877517A (en) * | 2019-12-06 | 2020-03-13 | 集美大学 | Method for preventing strong light stimulation of high beam, terminal equipment and storage medium |
| CN111310738A (en) * | 2020-03-31 | 2020-06-19 | 青岛讯极科技有限公司 | High beam vehicle snapshot method based on deep learning |
| CN111524360A (en) * | 2020-04-21 | 2020-08-11 | 刘哲安 | Embedded license plate identification and screening system and control method thereof |
| CN113628447A (en) * | 2020-05-06 | 2021-11-09 | 杭州海康威视数字技术股份有限公司 | High beam light starting detection method, device, equipment and system |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008143421A (en) * | 2006-12-12 | 2008-06-26 | Denso Corp | Headlight control device |
| CN102298845A (en) * | 2011-08-29 | 2011-12-28 | 安徽超远信息技术有限公司 | Far-beam light video detection method and system |
| CN104008650A (en) * | 2014-05-20 | 2014-08-27 | 宁辉 | Motor vehicle high beam monitoring device and identification method thereof |
| CN105184227A (en) * | 2015-08-13 | 2015-12-23 | 安徽超远信息技术有限公司 | Traffic signal control system with automobile high beam light detection function |
| CN205003858U (en) * | 2015-08-13 | 2016-01-27 | 安徽超远信息技术有限公司 | Possesses traffic signal controller that car far -reaching headlamp detected function |
| CN105575128A (en) * | 2016-01-20 | 2016-05-11 | 王威立 | High beam role breaking detection and snapshot system for motor vehicles |
| CN106934378A (en) * | 2017-03-16 | 2017-07-07 | 山东建筑大学 | A kind of dazzle light identifying system and method based on video depth study |
| CN108230690A (en) * | 2018-02-09 | 2018-06-29 | 浙江安谐智能科技有限公司 | A kind of high beam based on convolutional neural networks continues the method for discrimination of opening |
-
2018
- 2018-08-07 CN CN201810889659.5A patent/CN108986476B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008143421A (en) * | 2006-12-12 | 2008-06-26 | Denso Corp | Headlight control device |
| CN102298845A (en) * | 2011-08-29 | 2011-12-28 | 安徽超远信息技术有限公司 | Far-beam light video detection method and system |
| CN104008650A (en) * | 2014-05-20 | 2014-08-27 | 宁辉 | Motor vehicle high beam monitoring device and identification method thereof |
| CN105184227A (en) * | 2015-08-13 | 2015-12-23 | 安徽超远信息技术有限公司 | Traffic signal control system with automobile high beam light detection function |
| CN205003858U (en) * | 2015-08-13 | 2016-01-27 | 安徽超远信息技术有限公司 | Possesses traffic signal controller that car far -reaching headlamp detected function |
| CN105575128A (en) * | 2016-01-20 | 2016-05-11 | 王威立 | High beam role breaking detection and snapshot system for motor vehicles |
| CN106934378A (en) * | 2017-03-16 | 2017-07-07 | 山东建筑大学 | A kind of dazzle light identifying system and method based on video depth study |
| CN108230690A (en) * | 2018-02-09 | 2018-06-29 | 浙江安谐智能科技有限公司 | A kind of high beam based on convolutional neural networks continues the method for discrimination of opening |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110135343A (en) * | 2019-05-15 | 2019-08-16 | 南通大学 | A kind of street lamp intelligent measurement and night status judgment method |
| CN110877517A (en) * | 2019-12-06 | 2020-03-13 | 集美大学 | Method for preventing strong light stimulation of high beam, terminal equipment and storage medium |
| CN110877517B (en) * | 2019-12-06 | 2021-08-17 | 集美大学 | Method for preventing strong light stimulation of high beam, terminal equipment and storage medium |
| CN111310738A (en) * | 2020-03-31 | 2020-06-19 | 青岛讯极科技有限公司 | High beam vehicle snapshot method based on deep learning |
| CN111524360A (en) * | 2020-04-21 | 2020-08-11 | 刘哲安 | Embedded license plate identification and screening system and control method thereof |
| CN113628447A (en) * | 2020-05-06 | 2021-11-09 | 杭州海康威视数字技术股份有限公司 | High beam light starting detection method, device, equipment and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108986476B (en) | 2019-12-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108986476B (en) | method, system and storage medium for recognizing non-use of high beam by motor vehicle according to regulations | |
| CN107967806B (en) | Vehicle fake-license detection method, device, readable storage medium storing program for executing and electronic equipment | |
| CN106652468B (en) | The detection and from vehicle violation early warning alarm set and method in violation of rules and regulations of road vehicle front truck | |
| CN106373426B (en) | Parking stall based on computer vision and violation road occupation for parking monitoring method | |
| CA2885019C (en) | Robust windshield detection via landmark localization | |
| US20150371094A1 (en) | A road marking analyser and a method of analysis of road markings and an apparatus and method for detecting vehicle weave | |
| EP2710573B1 (en) | Method and apparatus for identifying a possible collision object | |
| CN109711264A (en) | A kind of bus zone road occupying detection method and device | |
| CN107886034B (en) | Driving reminding method and device and vehicle | |
| CN110866479A (en) | Method, device and system for detecting that motorcycle driver does not wear helmet | |
| CN110929676A (en) | Deep learning-based real-time detection method for illegal turning around | |
| DE102006055906A1 (en) | Retro reflector and vehicle light identifying method, involves classifying objects as retro-reflector or light, based on movement of light spots in image and time-dependant intensity gradient of spots and position of spots in image | |
| DE102017122432A1 (en) | Method for detecting a road condition of a roadway for a motor vehicle, driver assistance system and motor vehicle | |
| CN106203273A (en) | The lane detection system of multiple features fusion, method and senior drive assist system | |
| CN105117696B (en) | A kind of automatic testing method of not fastening the safety belt based on small line analysis | |
| JP2011174794A (en) | Device and method for detecting state of road surface | |
| CN103294990B (en) | Select the visibility region in nighttime image to perform clear path detection | |
| CN110766949B (en) | Violation snapshot method and device | |
| CN103680148B (en) | A kind of taxi recognition methods | |
| DE102015216927B4 (en) | Method and device for automatic documentation of traffic violations | |
| CN113870185A (en) | Image processing method based on image snapshot, terminal and storage medium | |
| CN109191855B (en) | Method and system for identifying motor vehicle reflecting mark not stuck according to regulations and storage medium | |
| CN114494938A (en) | Non-motor vehicle behavior identification method and related device | |
| CN112200148A (en) | Decentralized traffic integrity evaluation system | |
| CN113177443A (en) | Method for intelligently identifying road traffic violation based on image vision |
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 |