CN108230666A - A kind of vehicle detection at night and method for tracing - Google Patents
A kind of vehicle detection at night and method for tracing Download PDFInfo
- Publication number
- CN108230666A CN108230666A CN201611152800.0A CN201611152800A CN108230666A CN 108230666 A CN108230666 A CN 108230666A CN 201611152800 A CN201611152800 A CN 201611152800A CN 108230666 A CN108230666 A CN 108230666A
- Authority
- CN
- China
- Prior art keywords
- car light
- car
- vehicle
- area
- frame
- 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.)
- Withdrawn
Links
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
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
-
- 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
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
-
- 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
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Length Measuring Devices By Optical Means (AREA)
- Image Analysis (AREA)
- Lighting Device Outwards From Vehicle And Optical Signal (AREA)
Abstract
A kind of vehicle detection at night and method for tracing, include the following steps：S1：Vehicle at night video image acquisition, and video is pre-processed；S2：Preceding tail-light is detected, including illumination compensation, car light segmentation and car light positioning；S3：Car light is tracked and pairing, utilizes the area degree of overlapping (E of vehiclearea) track car light, then the distance (E with vehicledistance), angle (Eangle), difference in areas (Edarea) and eccentricity (Edeformation), headlight is matched；S4：By car tracing feedback modifiers mechanism, vehicle is accurately tracked.The vehicle detection and method for tracing complexity of the present invention is low, it is easily achieved, with good real-time and robustness, it and can be under different illuminations and transportation condition, the precision of car light tracking and pairing is effectively improved, good result can be equally obtained even for the situation of change in shape, irregular motion and more car lights.
Description
Technical field
Present invention relates particularly to a kind of vehicle detection at night and method for tracing.
Background technology
In view of various countries' traffic accident takes place frequently, the research of vehicle detection and tracer technique, which has become, to become more and more important.However, research
The algorithm that personnel propose early stage was largely only focused on daytime, in recent years, just started to be detected vehicle at night.Its method master
It is divided into following four classes：1. based on night vision sensor；2. based on vehicle-mounted camera；3. based on coordinate conversion；4. it is based on
Car light and windshield characteristic.However 1. equipment manufacturing cost is expensive, and is easily influenced by car light intensity of illumination for method, method is 2. right
In detection, nearby vehicle effect is preferable, for the vehicle apart from camera farther out, due to the occlusion in traveling process, detection effect
Fruit is bad, and the coverage of in-vehicle camera is limited, therefore is not suitable for the car tracing of long range, and method is 3. computationally intensive,
It is complicated for operation, and method 4. due to car light feature influenced by scene and camera exposure it is small and at low cost, it is applied widely, become
In recent years the hot spot studied extensively.
Invention content
The technical problem to be solved in the present invention is to provide a kind of vehicle detection at night and method for tracing.
A kind of vehicle detection at night and method for tracing, include the following steps：
S1：Vehicle at night video image acquisition, and video is pre-processed；
S2：Preceding tail-light is detected, it is specific as follows：
S2-1：It is firstly introduced into homomorphic filter and carries out illumination compensation；
S2-2：Car light is divided, the result after artwork threshold value (Th=0.95) is extracted and homomorphic filtering threshold value (Th=0.7)
The result summation of extraction, while morphological erosion and expansive working are carried out to result, the vehicle body edge highlighted is removed, obtains car light
Binary map；
S2-3：Different ID numbers is distributed for each car light in first frame, the car light ID number of other frames is initialized as 50+
Frame number+car light number, while the horizontal coordinate and ordinate of all car lights in first frame can be obtained, and store these information and be used for
Subsequent processing.
S3：Car light is tracked and pairing, utilizes the area degree of overlapping (E of vehiclearea) track car light, then with vehicle away from
From (Edis tan ce), angle (Eangle), difference in areas (Edarea) and eccentricity (Edeformation), headlight is matched；
S4：By car tracing feedback modifiers mechanism, vehicle is accurately tracked.
Further, each parameter of the homomorphic filter is：Cutoff frequency D0=2000, sharpening coefficient c=1.5 are high
Frequency gain Rh=4, low-frequency gain Rl=0.1.
Further, car light tracking and pairing specific method are as follows：
S3-1：Car light is tracked, area degree of overlapping EareaThe area overlapping degree of car light i and car light j are represented, formula is as follows：
In formula, S (M, i) represents areas of the car light i in M frames, S (M+1, j) represents areas of the car light j in M+1 frames,
Soverlap(M, M+1, j) represents the area of lap between M frame car light i and M+1 frame car lights j, according to above formula, two car lights
Between area lap it is bigger, the two car lights more may be same car light, i.e. the ID number of M frames car light i passes to M+1
Frame car light j；
S3-2：Car light matches,
Distance Edis tan ceThe ratio of two car light distances between centers and car light ordinate is expressed as, formula is as follows：
In formula, y (M, i) and y (M, i) represent the center ordinate of M frame car light i and car light j respectively, x (M, i) and x (M,
J) the center abscissa of M frame car light i and car light j is represented respectively, and eps representatives are infinitely close to zero positive number；Edis tan ceCertainly
Closest car light i and j have been determined with the second priority, i.e. two car light distances between centers and ordinate ratio Edis tan ce
Smaller, the probability for belonging to same vehicle is bigger；
Angle EangleRepresent the departure degree of two car lights in vertical direction, formula is as follows：
Car light i and j in same horizontal line have maximum priority during pairing, i.e. two car lights are pressed from both sides with horizontal
Angle EangleSmaller, the probability for belonging to same vehicle is bigger；
Difference in areas EdareaThe difference in areas between car light i and car light j is represented, formula is as follows：
In formula, S (M, j) represents the area of M frame car lights j, and S (M, i) represents the area of M frame car lights i, is matched in car light
In the process, EdareaThere is two car lights i and j for determining difference in areas minimum third to match priority；
With the area of oval next approximate car light, elliptical eccentricity Ecc=P2/ (4 π × A2), wherein P and A are respectively referred to
Be elliptical diameter and area, therefore eccentricity EdeformationCalculation formula it is as follows：
EdeformationSmaller, the probability for belonging to same vehicle is bigger；
Car light pairing cost calculation formula is as follows：
EM(i, j)=α1×Eangle+α2×Edis tan ce+α3×Edarea+α4×Edeformation,
In formula, EM(i, j) represents the pairing cost between Biao M frame car lights i and j, αiIt is the constant in 0 to 1, αiSelection
It is determined by vehicle flowrate and Velicle motion velocity, and meets α1+α2+α3+α4=1.
The beneficial effects of the invention are as follows：
The vehicle detection and method for tracing of the present invention, in detection part, vehicle is extracted with homomorphic filtering and threshold technology
Lamp；In tracking and mating section, by combining for the kinetic characteristic of car light and geometric properties.This method complexity is low, is easy to
Realize there is good real-time and robustness, and car light can be effectively improved under different illuminations and transportation condition
Tracking and the precision of pairing, good effect can be equally obtained even for the situation of change in shape, irregular motion and more car lights
Fruit.
Specific embodiment
The present invention is further elaborated for specific examples below, but not as a limitation of the invention.
A kind of vehicle detection at night and method for tracing, include the following steps：
S1：Vehicle at night video image acquisition, and video is pre-processed；
S2：Preceding tail-light is detected, it is specific as follows：
S2-1：It is firstly introduced into homomorphic filter and carries out illumination compensation；
S2-2：Car light is divided, the result after artwork threshold value (Th=0.95) is extracted and homomorphic filtering threshold value (Th=0.7)
The result summation of extraction, while morphological erosion and expansive working are carried out to result, the vehicle body edge highlighted is removed, obtains car light
Binary map；
S2-3：Different ID numbers is distributed for each car light in first frame, the car light ID number of other frames is initialized as 50+
Frame number+car light number, while the horizontal coordinate and ordinate of all car lights in first frame can be obtained, and store these information and be used for
Subsequent processing.
S3：Car light is tracked and pairing, utilizes the area degree of overlapping (E of vehiclearea) track car light, then with vehicle away from
From (Edis tan ce), angle (Eangle), difference in areas (Edarea) and eccentricity (Edeformation), headlight is matched；
S4：By car tracing feedback modifiers mechanism, vehicle is accurately tracked.
Each parameter of the homomorphic filter is：Cutoff frequency D0=2000, sharpening coefficient c=1.5, high-frequency gain Rh=
4, low-frequency gain Rl=0.1.
Car light is tracked and pairing specific method is as follows：
S3-1：Car light is tracked, area degree of overlapping EareaThe area overlapping degree of car light i and car light j are represented, formula is as follows：
In formula, S (M, i) represents areas of the car light i in M frames, S (M+1, j) represents areas of the car light j in M+1 frames,
Soverlap(M, M+1, j) represents the area of lap between M frame car light i and M+1 frame car lights j, according to above formula, two car lights
Between area lap it is bigger, the two car lights more may be same car light, i.e. the ID number of M frames car light i passes to M+1
Frame car light j；
S3-2：Car light matches,
Distance Edis tan ceThe ratio of two car light distances between centers and car light ordinate is expressed as, formula is as follows：
In formula, y (M, i) and y (M, i) represent the center ordinate of M frame car light i and car light j respectively, x (M, i) and x (M,
J) the center abscissa of M frame car light i and car light j is represented respectively, and eps representatives are infinitely close to zero positive number；Edis tan ceCertainly
Closest car light i and j have been determined with the second priority, i.e. two car light distances between centers and ordinate ratio Edis tan ce
Smaller, the probability for belonging to same vehicle is bigger；
Angle EangleRepresent the departure degree of two car lights in vertical direction, formula is as follows：
Car light i and j in same horizontal line have maximum priority during pairing, i.e. two car lights are pressed from both sides with horizontal
Angle EangleSmaller, the probability for belonging to same vehicle is bigger；
Difference in areas EdareaThe difference in areas between car light i and car light j is represented, formula is as follows：
In formula, S (M, j) represents the area of M frame car lights j, and S (M, i) represents the area of M frame car lights i, is matched in car light
In the process, EdareaThere is two car lights i and j for determining difference in areas minimum third to match priority；
With the area of oval next approximate car light, elliptical eccentricity Ecc=P2/ (4 π × A2), wherein P and A are respectively referred to
Be elliptical diameter and area, therefore eccentricity EdeformationCalculation formula it is as follows：
EdeformationSmaller, the probability for belonging to same vehicle is bigger；
Car light pairing cost calculation formula is as follows：
EM(i, j)=α1×Eangle+α2×Edis tan ce+α3×Edarea+α4×Edeformation,
In formula, EM(i, j) represents the pairing cost between Biao M frame car lights i and j, αiIt is the constant in 0 to 1, αiSelection
It is determined by vehicle flowrate and Velicle motion velocity, and meets α1+α2+α3+α4=1.
Claims (3)
1. a kind of vehicle detection at night and method for tracing, which is characterized in that include the following steps：
S1：Vehicle at night video image acquisition, and video is pre-processed；
S2：Preceding tail-light is detected, it is specific as follows：
S2-1：It is firstly introduced into homomorphic filter and carries out illumination compensation；
S2-2：Car light is divided, and the result after artwork threshold value (Th=0.95) is extracted is extracted with homomorphic filtering threshold value (Th=0.7)
Result summation, while morphological erosion and expansive working are carried out to result, remove the vehicle body edge highlighted, obtain car light two-value
Figure；
S2-3：Different ID numbers is distributed for each car light in first frame, the car light ID number of other frames is initialized as 50+ frame numbers
+ car light number, while the horizontal coordinate and ordinate of all car lights in first frame can be obtained, and store these information and be used for subsequently
Processing.
S3：Car light is tracked and pairing, utilizes the area degree of overlapping (E of vehiclearea) track car light, then the distance with vehicle
(Edistance), angle (Eangle), difference in areas (Edarea) and eccentricity (Edeformation), headlight is matched；
S4：By car tracing feedback modifiers mechanism, vehicle is accurately tracked.
2. vehicle detection at night according to claim 1 and method for tracing, which is characterized in that the homomorphic filter it is each
Parameter is：Cutoff frequency D0=2000, sharpening coefficient c=1.5, high-frequency gain Rh=4, low-frequency gain Rl=0.1.
3. vehicle detection at night according to claim 1 and method for tracing, which is characterized in that car light is tracked and pairing is specific
Method is as follows：
S3-1：Car light is tracked, area degree of overlapping EareaThe area overlapping degree of car light i and car light j are represented, formula is as follows：
In formula, S (M, i) represents areas of the car light i in M frames, S (M+1, j) represents areas of the car light j in M+1 frames, Soverlap
(M, M+1, j) represents the area of lap between M frame car light i and M+1 frame car lights j, according to above formula, area between two car lights
Lap is bigger, the two car lights more may be same car light, i.e. the ID number of M frames car light i passes to M+1 frame car lights
j；
S3-2：Car light matches,
Distance EdistanceThe ratio of two car light distances between centers and car light ordinate is expressed as, formula is as follows：
In formula, y (M, i) and y (M, i) represent the center ordinate of M frame car light i and car light j respectively, x (M, i) and x (M, j) points
The center abscissa of M frame car light i and car light j is not represented, and eps representatives are infinitely close to zero positive number；EdistanceDetermine away from
There is the second priority, i.e. two car light distances between centers and ordinate ratio E from nearest car light i and jdistanceIt is smaller, belong to
It is bigger in the probability of same vehicle；
Angle EangleRepresent the departure degree of two car lights in vertical direction, formula is as follows：
Car light i and j in same horizontal line have maximum priority, i.e. two car lights and horizontal sextant angle during pairing
EangleSmaller, the probability for belonging to same vehicle is bigger；
Difference in areas EdareaThe difference in areas between car light i and car light j is represented, formula is as follows：
In formula, S (M, j) represents the area of M frame car lights j, and S (M, i) represents the area of M frame car lights i, and process is matched in car light
In, EdareaThere is two car lights i and j for determining difference in areas minimum third to match priority；
With the area of oval next approximate car light, elliptical eccentricity Ecc=P2/ (4 π × A2), wherein P and A are referred respectively to
Elliptical diameter and area, therefore eccentricity EdeformationCalculation formula it is as follows：
EdeformationSmaller, the probability for belonging to same vehicle is bigger；
Car light pairing cost calculation formula is as follows：
EM(i, j)=α1×Eangle+α2×Edistance+α3×Edarea+α4×Edeformation,
In formula, EM(i, j) represents the pairing cost between Biao M frame car lights i and j, αiIt is the constant in 0 to 1, αiSelection by vehicle
Flow and Velicle motion velocity determine, and meet α1+α2+α3+α4=1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611152800.0A CN108230666A (en) | 2016-12-14 | 2016-12-14 | A kind of vehicle detection at night and method for tracing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611152800.0A CN108230666A (en) | 2016-12-14 | 2016-12-14 | A kind of vehicle detection at night and method for tracing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108230666A true CN108230666A (en) | 2018-06-29 |
Family
ID=62638545
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611152800.0A Withdrawn CN108230666A (en) | 2016-12-14 | 2016-12-14 | A kind of vehicle detection at night and method for tracing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108230666A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101382997A (en) * | 2008-06-13 | 2009-03-11 | 青岛海信电子产业控股股份有限公司 | Vehicle detecting and tracking method and device at night |
| CN103150898A (en) * | 2013-01-25 | 2013-06-12 | 大唐移动通信设备有限公司 | Method and device for detection of vehicle at night and method and device for tracking of vehicle at night |
| CN103208185A (en) * | 2013-03-19 | 2013-07-17 | 东南大学 | Method and system for nighttime vehicle detection on basis of vehicle light identification |
| DE102013226599A1 (en) * | 2013-12-19 | 2015-06-25 | Robert Bosch Gmbh | Method and apparatus for adaptive distance and speed control of a motor vehicle |
| CN105303160A (en) * | 2015-09-21 | 2016-02-03 | 中电海康集团有限公司 | Method for detecting and tracking vehicles at night |
-
2016
- 2016-12-14 CN CN201611152800.0A patent/CN108230666A/en not_active Withdrawn
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101382997A (en) * | 2008-06-13 | 2009-03-11 | 青岛海信电子产业控股股份有限公司 | Vehicle detecting and tracking method and device at night |
| CN103150898A (en) * | 2013-01-25 | 2013-06-12 | 大唐移动通信设备有限公司 | Method and device for detection of vehicle at night and method and device for tracking of vehicle at night |
| CN103208185A (en) * | 2013-03-19 | 2013-07-17 | 东南大学 | Method and system for nighttime vehicle detection on basis of vehicle light identification |
| DE102013226599A1 (en) * | 2013-12-19 | 2015-06-25 | Robert Bosch Gmbh | Method and apparatus for adaptive distance and speed control of a motor vehicle |
| CN105303160A (en) * | 2015-09-21 | 2016-02-03 | 中电海康集团有限公司 | Method for detecting and tracking vehicles at night |
Non-Patent Citations (1)
| Title |
|---|
| 汤春明: "基于最小特征匹配代价的夜间车辆检测和追踪", 《计算机应用与软件》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Narote et al. | A review of recent advances in lane detection and departure warning system | |
| US9558412B2 (en) | Vehicle exterior environment recognition device | |
| US8712097B2 (en) | Approaching object detection system | |
| CN106489174B (en) | Method, camera system and the motor vehicles of vehicle are tracked by camera system | |
| CN109190523B (en) | Vehicle detection tracking early warning method based on vision | |
| CN105981042B (en) | Vehicle detection system and method | |
| CN104916163B (en) | Parking space detection method | |
| US10891738B2 (en) | Boundary line recognition apparatus and branch road determination apparatus | |
| US9665781B2 (en) | Moving body detection device and moving body detection method | |
| CN109784344A (en) | A kind of non-targeted filtering method of image for ground level mark identification | |
| CN102865824B (en) | A kind of method and apparatus calculating relative distance between vehicle | |
| CN107787496B (en) | Vanishing point correcting device and method | |
| WO2013136827A1 (en) | Vehicle periphery monitor device | |
| JP5482672B2 (en) | Moving object detection device | |
| KR101727054B1 (en) | Method for detecting and recognizing traffic lights signal based on features | |
| KR20170104287A (en) | Driving area recognition apparatus and method for recognizing driving area thereof | |
| EP3115933A1 (en) | Image processing device, image capturing device, mobile body control system, image processing method, and computer-readable recording medium | |
| JPWO2017047282A1 (en) | Image processing apparatus, object recognition apparatus, device control system, image processing method and program | |
| CN110659552A (en) | Tramcar obstacle detection and alarm method | |
| JPWO2011055418A1 (en) | Distance measuring device and distance measuring method | |
| KR20160108344A (en) | Vehicle detection system and method thereof | |
| CN108230666A (en) | A kind of vehicle detection at night and method for tracing | |
| Chiu et al. | Real-Time Front Vehicle Detection Algorithm for an Asynchronous Binocular System. | |
| Chen et al. | Robust rear light status recognition using symmetrical surfs | |
| CN108572650B (en) | Adaptive headlamp steering control algorithm based on lane line detection |
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 | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20180629 |