CN109191492A - A kind of intelligent video black smoke vehicle detection method based on edge analysis - Google Patents
A kind of intelligent video black smoke vehicle detection method based on edge analysis Download PDFInfo
- Publication number
- CN109191492A CN109191492A CN201810754421.1A CN201810754421A CN109191492A CN 109191492 A CN109191492 A CN 109191492A CN 201810754421 A CN201810754421 A CN 201810754421A CN 109191492 A CN109191492 A CN 109191492A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- profile
- black smoke
- center
- target
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Closed-Circuit Television Systems (AREA)
- Image Analysis (AREA)
Abstract
The intelligent video black smoke vehicle detection method based on edge analysis that the invention discloses a kind of, includes the following steps: that (1) Utilization prospects detection algorithm extracts moving target from traffic surveillance videos；(2) non-vehicle target is removed, and vehicle target is tracked；(3) profile for extracting vehicle target rear, calculates a series of static natures and behavioral characteristics based on profile, and fusion forms a feature vector；(4) classified using SVM classifier to the mentioned feature vector of every frame, black smoke vehicle is identified by the analysis of multiframe, and retained the license plate of black smoke vehicle automatically, cross vehicle place, cross the evidences such as vehicle time.The deficiency that can make up for it traditional artificial monitoring black smoke vehicle inefficiency, reduces rate of false alarm, has algorithm advantage under wind weather to become apparent from.
Description
Technical field
The present invention relates to pyrotechnics detection technique field, especially a kind of intelligent video black smoke car test based on edge analysis is surveyed
Method.
Background technique
Black smoke vehicle is the high pollution vehicle that a kind of Vehicular exhaust hole emits dense black smoke.The pollution of black smoke vehicle is always motor vehicle
The key points and difficulties of environmental protection.Particulate matter (PM) in black smoke not only pollutes air, can also cause to the respiratory tract and lung of contactee
Damage damages human health.Therefore, black smoke vehicle is found in time and will be helpful to improve urban air-quality for further processing.
The method surveyed at present about black smoke car test, is broadly divided into two classes: (1) conventional method.Conventional method specifically includes that
Reports, the inspection of regular road, night inspection, manual video monitoring, installation Vehicular exhaust analytical equipment, sensor detection etc..This
A little methods reduce the pollution of black smoke vehicle to a certain extent, but due to the sharp increase of vehicle guaranteeding organic quantity, traffic it is busy
Deng such methods generally require to put into a large amount of manpower and financial resources, and efficiency is very low；(2) intelligent video monitoring method.Black smoke
Vehicle intelligent video monitoring refers to detects black smoke vehicle, video phase using computer vision technique automatically from magnanimity traffic surveillance videos
It closes data and uploads environmental protection administration automatically, while retaining the licence plate of black smoke vehicle, cross vehicle place, cross the evidences such as vehicle time.Such methods
Belong to remote monitor, does not block traffic, it can be achieved that whole day is on duty online, be adapted to the various roads ring such as two-way traffic and multilane
Border, and it is easy for installation, it is suitble to deploying to ensure effective monitoring and control of illegal activities on a large scale for urban road, it is easier to form the on-line monitoring for being directed to high pollution black smoke vehicle
Network improves law enforcement efficiency.But such methods are at present still in the starting stage of research.
Intelligent video black smoke vehicle detection method proposed by the present invention based on edge analysis, can be improved law enforcement efficiency, more
Mend the deficiency of traditional artificial monitoring black smoke vehicle inefficiency.The present invention uses the strategy for directly analyzing the region at vehicle rear, significantly
Rate of false alarm is reduced, the erroneous detection as caused by leaf shaking, white clouds movement etc. is avoided.In addition, designed feature is better than biography
The pyrotechnics of system detects feature used, has algorithm advantage under wind weather to become apparent from.
Summary of the invention
Technical problem to be solved by the present invention lies in provide a kind of intelligent video black smoke car test survey based on edge analysis
Method can make up for it the deficiency of traditional artificial monitoring black smoke vehicle inefficiency, reduce rate of false alarm, have under wind weather algorithm advantage more
Obviously.
In order to solve the above technical problems, the present invention provides a kind of intelligent video black smoke car test survey side based on edge analysis
Method includes the following steps:
(1) Utilization prospects detection algorithm extracts moving target from traffic surveillance videos；
(2) non-vehicle target is removed, and vehicle target is tracked；
(3) profile for extracting vehicle target rear calculates a series of static natures and behavioral characteristics based on profile, fusion
Form a feature vector；
(4) classified using SVM classifier to the mentioned feature vector of every frame, black smoke is identified by the analysis of multiframe
Vehicle, and retain the license plate of black smoke vehicle automatically, cross vehicle place, cross the evidences such as vehicle time.
Preferably, the foreground detection algorithm in step (1) uses Vibe background difference algorithm.
Preferably, the removal non-vehicle target in step (2) the following steps are included:
(21) the area S of moving target should be greater than the minimum value S of usual vehicle target area0, i.e.,
S>S0；
(22) ratio of the wide width and high height of the boundary rectangle of moving target, should be within the scope of some, i.e.,
Wherein, [δ1,δ2] be vehicle target boundary rectangle width and high ratio range；
(23) moving target for meeting above-mentioned two rule is considered as vehicle target.
Preferably, vehicle target is tracked in step (2) the following steps are included:
(24) center position of the boundary rectangle of each vehicle target in each frame image is calculated；
(25) Euclidean at each vehicle target center in the center and a later frame of some vehicle target of former frame is calculated
Distance, two vehicle targets apart from the smallest two frame of front and back are considered as same vehicle, do to each vehicle target of former frame
Such processing can be obtained corresponding position of each vehicle target in a later frame of former frame；
(26) to each frame of video, the correspondence of each vehicle in adjacent two frame is obtained according to the method for step (25)
Relationship is equal to and tracks to vehicle.
Preferably, a series of static natures in step (3) calculating the following steps are included:
(31) the profile C of vehicle target is extractedobj；
(32) vehicle's contour C is extractedobjConvex closure Hobj；
(33) rear of vehicle profile C is calculated using following formularearStarting point coordinate Pstart(xstart,ystart) and terminal point coordinate
Pend(xend,yend),
Wherein, Ocenter(xcenter,ycenter) and r respectively represent profile CobjMinimum circumscribed circle the center of circle and radius, k be
Adjust the coefficient of mentioned profile size；
The convex closure H of rear of vehicle profilerearStarting point coordinate and terminal point coordinate be similarly point Pstart(xstart,ystart) and
Point Pend(xend,yend)；
(34) the length L of the convex closure of rear of vehicle profileconvex_rearWith the length L of rear of vehicle profilecontour_rearRatio
Value Rat1 can be used as an important static nature for judging whether the rear of vehicle has black smoke, i.e.,
(35) S is usedcontour_rearIt indicates by line segment PendOcenter, line segment OcenterPstartWith vehicle rear profile CrearIt surrounds
The area of polygon；By Sconvex_rearIt indicates by line segment PendOcenter, line segment OcenterPstartWith the convex closure H of vehicle rear profilerear
The area of the polygon surrounded, then Scontour_rearAnd Sconvex_rearRatio can be used as and judge whether the rear of vehicle has black smoke
An important static nature, i.e.,
(36) Δ P is indicated with FstartOrearPendCenter of gravity；Use SregionIt indicates by line segment PstartF, line segment FPendAfter vehicle
Square wheel exterior feature CrearThe polygonal region surrounded, some statistical natures in the region can be used to judge whether the rear of vehicle has black
Cigarette, i.e.,
Wherein, NregionIndicate region SregionNumber of pixels, Sregion(i, j) indicates region SregionAt position (i, j)
The pixel value at place.
Preferably, a series of behavioral characteristics in step (3) extraction the following steps are included:
(37) the profile C of vehicle target is extractedobjAnd the circumscribed circle center of circle O of profilecenter；
(38) center of circle O is calculatedcenterPlace horizontal linear and profile CobjThe intersection point P of left and right twoleftAnd Pright, with X table
Show by line segment PrightPleftWith profile CobjThe polygonal profile surrounded；
(39) it is denoted as A and B respectively according to step 3.8 calculating profile X for being separated by the same vehicle of K frame；
(310) the Hu not bending moment for calculating profile A and profile B, is denoted as respectivelyWith
(311) the matching degree M of profile A and profile B is calculatedmatch(A, B), we devise two kinds of matching degree calculation methods,
Respectively,
Wherein,WithSeven Hu of profile A and profile B not bending moment is respectively indicated, sign (x) indicates symbol
Number function, log (x) indicate logarithmic function；
(312) the matching degree M of the profile in same vehicle different frame is utilizedmatchWhether (A, B) judges the rear of vehicle
There is black smoke.
Preferably, the not bending moment h of the Hu in step (310)i(i=1,2 ..., calculation method 7.) be,
h1=η20+η02
h3=(η30-3η12)2+(3η21-η03)2
h4=(η30+η12)2+(η21+η03)2
h5=(η30-3η12)(η30+η12)[(η30+η12)2-3(η21+η03)2]+(3η21-η03)(η21+η03)[3(η30+η12)2-
(η21+η03)2]
h6=(η20-η02)[(η30+η12)2-(η21+η03)2]+4η11(η30+η12)(η21+η03)
h7=(3 η21-η03)(η30+η12)[(η30+η12)2-3(η21+η03)2]+(3η21-η30)(η21+η03)[3(η30+η12)2-
(η21+η03)2]
Wherein, μpqIt is central moment, f (x, y) is two-dimensional silhouette, pointFor the center of gravity of profile, calculation method is,
M in above formulapqFor two-dimentional p+q rank routine square, calculation method is,
Preferably, the SVM classifier in step (4) is obtained by training, and is used to test new samples number for the classifier
According to.
Preferably, a vehicle is identified as black smoke vehicle needs while meeting following two rules in step (4),
Rule one: ω2>α1
Rule two:
Wherein, ω1It is the frame number that same vehicle is detected in video, ω2Refer in all ω being detected1In frame,
It is identified as the frame number of black smoke vehicle.
The invention has the benefit that (1) improves law enforcement efficiency, traditional artificial monitoring black smoke vehicle inefficiency is made up not
Foot；Black smoke vehicle is detected automatically from magnanimity traffic surveillance videos using computer vision technique, and video related data uploads automatically
Environmental protection administration, while retaining the licence plate of black smoke vehicle, cross vehicle place, cross the evidences such as vehicle time；This method belongs to remote monitor, no
Block traffic, it can be achieved that whole day is on duty online, be adapted to the various roads environment such as two-way traffic and multilane, and easy for installation,
It is suitble to deploying to ensure effective monitoring and control of illegal activities on a large scale for urban road, it is easier to form the on-line monitoring network for being directed to high pollution black smoke vehicle, improve law enforcement effect
Rate；(2) rate of false alarm is reduced；Technical solution proposed by the present invention greatly reduces wrong report using the region for directly analyzing vehicle rear
Rate avoids the erroneous detection as caused by leaf shaking, white clouds movement etc.；(3) feature designed by is detected better than traditional pyrotechnics
Feature used has algorithm advantage under wind weather to become apparent from；This behavioral characteristics of the matching degree proposed in technical solution of the present invention,
For portraying the matching degree of vehicle rear profile between same vehicle different frame, in same time interval, wind is bigger, square wheel after vehicle
Wide variation is more obvious, rather than what the vehicle rear profile of black smoke vehicle was no variation in, this allow for more easily discriminating black smoke vehicle and
Non- black smoke vehicle.
Detailed description of the invention
Fig. 1 is method flow schematic diagram of the invention.
Fig. 2 is the profile and recessed packet schematic diagram at vehicle target rear of the present invention.
Fig. 3 is the profile and recessed packet schematic diagram at the vehicle rear that the present invention is extracted from a black smoke vehicle target.
Fig. 4 is the profile and recessed packet schematic diagram at the vehicle rear that the present invention is extracted from two non-black smoke vehicle targets.
Fig. 5 is the variation schematic diagram of black smoke wheel exterior feature of the present invention.
Fig. 6 is the variation schematic diagram of the non-black smoke wheel exterior feature of the present invention.
Specific embodiment
The present invention provides a kind of intelligent video black smoke vehicle detection method based on edge analysis, flow chart as shown in Figure 1,
Specifically follow the steps below:
Step 1: Utilization prospects detection algorithm extracts moving target from traffic surveillance videos；
Step 2: removal non-vehicle target, and vehicle target is tracked；
Step 3: the profile at vehicle target rear is extracted, a series of static natures and behavioral characteristics based on profile are calculated,
Fusion forms a feature vector；
Step 4: being classified using SVM classifier to the mentioned feature vector of every frame, identified by the analysis of multiframe black
Cigarette vehicle, and retain the license plate of black smoke vehicle automatically, cross vehicle place, cross the evidences such as vehicle time.
Foreground detection algorithm in the step 1 uses Vibe background difference algorithm.
Removal non-vehicle target in the step 2 the following steps are included:
Step 2.1: the area S of moving target should be greater than the minimum value S of usual vehicle target area0, i.e.,
S>S0；
Step 2.2: the ratio of the wide width and high height of the boundary rectangle of moving target, should within the scope of some,
I.e.
Wherein, [δ1,δ2] be vehicle target boundary rectangle width and high ratio range；
Step 2.3: the moving target for meeting above-mentioned two rule is considered as vehicle target.
Vehicle target is tracked in the step 2 the following steps are included:
Step 2.4: calculating the center position of the boundary rectangle of each vehicle target in each frame image；
Step 2.5: calculating each vehicle target center in the center and a later frame of some vehicle target of former frame
Euclidean distance, two vehicle targets apart from the smallest two frame of front and back are considered as same vehicle, to each vehicle target of former frame
Such processing is all done, can be obtained corresponding position of each vehicle target in a later frame of former frame；
Step 2.6: to each frame of video, the front and back of each vehicle in adjacent two frame is obtained according to the method for step 2.5
Corresponding relationship is equal to and tracks to vehicle.
The calculating of static natures in the step 3 a series of the following steps are included:
Step 3.1: extracting the profile C of vehicle targetobj；
Step 3.2: extracting vehicle's contour CobjConvex closure Hobj；
Step 3.3: calculating rear of vehicle profile C using following formularearStarting point coordinate Pstart(xstart,ystart) and terminal
Coordinate Pend(xend,yend),
Wherein, Ocenter(xcenter,ycenter) and r respectively represent profile CobjMinimum circumscribed circle the center of circle and radius, k be
Adjust the coefficient of mentioned profile size；
The convex closure H of rear of vehicle profilerearStarting point coordinate and terminal point coordinate be similarly point Pstart(xstart,ystart) and
Point Pend(xend,yend)；
Fig. 2 shows that the profile and recessed packet schematic diagram at vehicle target rear, Fig. 3 show and mention from a black smoke vehicle target
The profile at the vehicle rear taken and recessed packet schematic diagram, Fig. 4 show the profile at the vehicle rear extracted from two non-black smoke vehicle targets
With recessed packet schematic diagram, it can be seen that for black smoke vehicle, apparent difference is presented in the profile and convex closure at vehicle rear, and for non-black smoke
Vehicle, the profile and convex closure at vehicle rear are almost to be overlapped；
Step 3.4: the length L of the convex closure of rear of vehicle profileconvex_rearWith the length L of rear of vehicle profilecontour_rear
Ratio R at1, can be used as an important static nature for judging whether the rear of vehicle has black smoke, i.e.,
Step 3.5: using Scontour_rearIt indicates by line segment PendOcenter, line segment OcenterPstartWith vehicle rear profile CrearIt encloses
At polygon area；By Sconvex_rearIt indicates by line segment PendOcenter, line segment OcenterPstartIt is convex with vehicle rear profile
Wrap HrearThe area of the polygon surrounded, then Scontour_rearAnd Sconvex_rearRatio can be used as whether judge the rear of vehicle
There is an important static nature of black smoke, i.e.,
Step 3.6: indicating Δ P with FstartOrearPendCenter of gravity；Use SregionIt indicates by line segment PstartF, line segment FPendWith
Vehicle rear profile CrearThe polygonal region surrounded, some statistical natures in the region can be used to judge whether the rear of vehicle has
Black smoke, i.e.,
Wherein, NregionIndicate region SregionNumber of pixels, Sregion(i, j) indicates region SregionAt position (i, j)
The pixel value at place.
The extractions of behavioral characteristics in the step 3 a series of the following steps are included:
Step 3.7: extracting the profile C of vehicle targetobjAnd the circumscribed circle center of circle O of profilecenter, respectively with black smoke vehicle and
For non-black smoke vehicle, Fig. 5 shows that the variation schematic diagram of black smoke wheel exterior feature, Fig. 6 show the variation signal of non-black smoke wheel exterior feature
Figure, it can be seen that change with time for vehicle rear profile, black smoke vehicle becomes apparent from than non-black smoke vehicle；
Step 3.8: calculating center of circle OcenterPlace horizontal linear and profile CobjThe intersection point P of left and right twoleftAnd Pright, use
X is indicated by line segment PrightPleftWith profile CobjThe polygonal profile surrounded；
Step 3.9: the same vehicle for being separated by K frame calculates profile X according to step 3.8, is denoted as A and B respectively；
Step 3.10: calculating the Hu not bending moment of profile A and profile B, be denoted as respectivelyWith
Step 3.11: calculating the matching degree M of profile A and profile Bmatch(A, B), we devise two kinds of matching degree calculating sides
Method, respectively,
Wherein,WithSeven Hu of profile A and profile B not bending moment is respectively indicated, sign (x) indicates symbol
Number function, log (x) indicate logarithmic function；
Step 3.12: utilizing the matching degree M of the profile in same vehicle different framematch(A, B) judges the rear of vehicle
Whether black smoke is had.
Hu in affiliated step 3.10 not bending moment hi(i=1,2 ..., calculation method 7.) be,
h1=η20+η02
h3=(η30-3η12)2+(3η21-η03)2
h4=(η30+η12)2+(η21+η03)2
h5=(η30-3η12)(η30+η12)[(η30+η12)2-3(η21+η03)2]+(3η21-η03)(η21+η03)[3(η30+η12)2-
(η21+η03)2]
h6=(η20-η02)[(η30+η12)2-(η21+η03)2]+4η11(η30+η12)(η21+η03)
h7=(3 η21-η03)(η30+η12)[(η30+η12)2-3(η21+η03)2]+(3η21-η30)(η21+η03)[3(η30+η12)2-
(η21+η03)2]
Wherein, μpqIt is central moment, f (x, y) is two-dimensional silhouette, pointFor the center of gravity of profile, calculation method is,
M in above formulapqFor two-dimentional p+q rank routine square, calculation method is,
SVM classifier in the step 4 is obtained by training, and is used to test new samples data for the classifier.
A vehicle is identified as black smoke vehicle needs while meeting following two rules in the step 4,
Rule one: ω2>α1
Rule two:
Wherein, ω1It is the frame number that same vehicle is detected in video, ω2Refer in all ω being detected1In frame,
It is identified as the frame number of black smoke vehicle.
Claims (8)
1. a kind of intelligent video black smoke vehicle detection method based on edge analysis, which comprises the steps of:
(1) Utilization prospects detection algorithm extracts moving target from traffic surveillance videos；
(2) non-vehicle target is removed, and vehicle target is tracked；
(3) profile for extracting vehicle target rear, calculates a series of static natures and behavioral characteristics based on profile, and fusion is formed
One feature vector；
(4) classified using SVM classifier to the mentioned feature vector of every frame, black smoke vehicle is identified by the analysis of multiframe, and
The automatic license plate for retaining black smoke vehicle crosses vehicle place, crosses the evidences such as vehicle time.
2. the intelligent video black smoke vehicle detection method based on edge analysis as described in claim 1, which is characterized in that step
(1) the foreground detection algorithm in uses Vibe background difference algorithm.
3. the intelligent video black smoke vehicle detection method based on edge analysis as described in claim 1, which is characterized in that step
(2) removal non-vehicle target in the following steps are included:
(21) the area S of moving target should be greater than the minimum value S of usual vehicle target area0, i.e.,
S>S0；
(22) ratio of the wide width and high height of the boundary rectangle of moving target, should be within the scope of some, i.e.,
Wherein, [δ1,δ2] be vehicle target boundary rectangle width and high ratio range；
(23) moving target for meeting above-mentioned two rule is considered as vehicle target.
Preferably, vehicle target is tracked in step (2) the following steps are included:
(24) center position of the boundary rectangle of each vehicle target in each frame image is calculated；
(25) Euclidean distance at each vehicle target center in the center and a later frame of some vehicle target of former frame is calculated,
Two vehicle targets apart from the smallest two frame of front and back are considered as same vehicle, do to each vehicle target of former frame such
Processing, can be obtained corresponding position of each vehicle target in a later frame of former frame；
(26) it to each frame of video, is closed according to the correspondence that the method for step (25) obtains each vehicle in adjacent two frame
System, that is, be equal to and track to vehicle.
4. the intelligent video black smoke vehicle detection method based on edge analysis as described in claim 1, which is characterized in that step
(3) calculating of static natures in a series of the following steps are included:
(31) the profile C of vehicle target is extractedobj；
(32) vehicle's contour C is extractedobjConvex closure Hobj；
(33) rear of vehicle profile C is calculated using following formularearStarting point coordinate Pstart(xstart,ystart) and terminal point coordinate Pend
(xend,yend),
Wherein, Ocenter(xcenter,ycenter) and r respectively represent profile CobjMinimum circumscribed circle the center of circle and radius, k be adjust
The coefficient of mentioned profile size；
The convex closure H of rear of vehicle profilerearStarting point coordinate and terminal point coordinate be similarly point Pstart(xstart,ystart) and point Pend
(xend,yend)；
(34) the length L of the convex closure of rear of vehicle profileconvex_rearWith the length L of rear of vehicle profilecontour_rearRatio
Rat1 can be used as an important static nature for judging whether the rear of vehicle has black smoke, i.e.,
(35) S is usedcontour_rearIt indicates by line segment PendOcenter, line segment OcenterPstartWith vehicle rear profile CrearWhat is surrounded is polygon
The area of shape；By Sconvex_rearIt indicates by line segment PendOcenter, line segment OcenterPstartWith the convex closure H of vehicle rear profilerearIt surrounds
Polygon area, then Scontour_rearAnd Sconvex_rearRatio can be used as and judge whether the rear of vehicle has black smoke one
A important static nature, i.e.,
(36) Δ P is indicated with FstartOrearPendCenter of gravity；Use SregionIt indicates by line segment PstartF, line segment FPendWith square wheel after vehicle
Wide CrearThe polygonal region surrounded, some statistical natures in the region can be used to judge whether the rear of vehicle has black smoke, i.e.,
Wherein, NregionIndicate region SregionNumber of pixels, Sregion(i, j) indicates region SregionAt position (i, j)
Pixel value.
5. the intelligent video black smoke vehicle detection method based on edge analysis as described in claim 1, which is characterized in that step
(3) extractions of behavioral characteristics in a series of the following steps are included:
(37) the profile C of vehicle target is extractedobjAnd the circumscribed circle center of circle O of profilecenter；
(38) center of circle O is calculatedcenterPlace horizontal linear and profile CobjThe intersection point P of left and right twoleftAnd Pright, with X indicate by
Line segment PrightPleftWith profile CobjThe polygonal profile surrounded；
(39) it is denoted as A and B respectively according to step 3.8 calculating profile X for being separated by the same vehicle of K frame；
(310) the Hu not bending moment for calculating profile A and profile B, is denoted as respectivelyWith
(311) the matching degree M of profile A and profile B is calculatedmatch(A, B), we devise two kinds of matching degree calculation methods, respectively
For,
Or
Wherein,WithSeven Hu of profile A and profile B not bending moment is respectively indicated, sign (x) indicates symbol letter
Number, log (x) indicate logarithmic function；
(312) the matching degree M of the profile in same vehicle different frame is utilizedmatch(A, B) judges that it is black whether the rear of vehicle has
Cigarette.
6. the intelligent video black smoke vehicle detection method based on edge analysis as claimed in claim 5, which is characterized in that step
(310) Hu in not bending moment hi(i=1,2 ..., calculation method 7.) be,
h1=η20+η02
h3=(η30-3η12)2+(3η21-η03)2
h4=(η30+η12)2+(η21+η03)2
h5=(η30-3η12)(η30+η12)[(η30+η12)2-3(η21+η03)2]+(3η21-η03)(η21+η03)[3(η30+η12)2-(η21+
η03)2]
h6=(η20-η02)[(η30+η12)2-(η21+η03)2]+4η11(η30+η12)(η21+η03)
h7=(3 η21-η03)(η30+η12)[(η30+η12)2-3(η21+η03)2]+(3η21-η30)(η21+η03)[3(η30+η12)2-(η21+
η03)2]
Wherein, μpqIt is central moment, f (x, y) is two-dimensional silhouette, pointFor the center of gravity of profile, calculation method is,
M in above formulapqFor two-dimentional p+q rank routine square, calculation method is,
7. the intelligent video black smoke vehicle detection method based on edge analysis as described in claim 1, which is characterized in that step
(4) SVM classifier in is obtained by training, and is used to test new samples data for the classifier.
8. the intelligent video black smoke vehicle detection method based on edge analysis as described in claim 1, which is characterized in that step
(4) vehicle is identified as black smoke vehicle needs while meeting following two rules in,
Rule one: ω2>α1
Rule two:
Wherein, ω1It is the frame number that same vehicle is detected in video, ω2Refer in all ω being detected1In frame, known
Not Wei black smoke vehicle frame number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810754421.1A CN109191492B (en) | 2018-07-11 | 2018-07-11 | Intelligent video black smoke vehicle detection method based on contour analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810754421.1A CN109191492B (en) | 2018-07-11 | 2018-07-11 | Intelligent video black smoke vehicle detection method based on contour analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109191492A true CN109191492A (en) | 2019-01-11 |
| CN109191492B CN109191492B (en) | 2021-07-27 |
Family
ID=64936480
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810754421.1A Active CN109191492B (en) | 2018-07-11 | 2018-07-11 | Intelligent video black smoke vehicle detection method based on contour analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109191492B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110428443A (en) * | 2019-08-11 | 2019-11-08 | 上海天诚比集科技有限公司 | A kind of intelligence community Vehicle tracing method |
| CN111539343A (en) * | 2020-04-26 | 2020-08-14 | 安徽皖仪科技股份有限公司 | Black smoke vehicle detection method based on convolution attention network |
| CN113378629A (en) * | 2021-04-27 | 2021-09-10 | 阿里云计算有限公司 | Method and device for detecting abnormal vehicle in smoke discharge |
| US20220076032A1 (en) * | 2020-09-08 | 2022-03-10 | Nvidia Corporation | Adaptive object tracking algorithm for autonomous machine applications |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101029824A (en) * | 2006-02-28 | 2007-09-05 | 沈阳东软软件股份有限公司 | Method and apparatus for positioning vehicle based on characteristics |
| CN102737247A (en) * | 2012-07-04 | 2012-10-17 | 中国科学技术大学 | Identification system of smoke intensity image of tail gas of diesel vehicle |
| CN103733084A (en) * | 2011-05-26 | 2014-04-16 | 克朗设备有限公司 | Method and apparatus for providing accurate localization for an industrial vehicle |
| CN106951821A (en) * | 2016-12-27 | 2017-07-14 | 湘潭大学 | A kind of black smoke car intelligent monitoring recognition methods based on image processing techniques |
-
2018
- 2018-07-11 CN CN201810754421.1A patent/CN109191492B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101029824A (en) * | 2006-02-28 | 2007-09-05 | 沈阳东软软件股份有限公司 | Method and apparatus for positioning vehicle based on characteristics |
| CN103733084A (en) * | 2011-05-26 | 2014-04-16 | 克朗设备有限公司 | Method and apparatus for providing accurate localization for an industrial vehicle |
| CN102737247A (en) * | 2012-07-04 | 2012-10-17 | 中国科学技术大学 | Identification system of smoke intensity image of tail gas of diesel vehicle |
| CN106951821A (en) * | 2016-12-27 | 2017-07-14 | 湘潭大学 | A kind of black smoke car intelligent monitoring recognition methods based on image processing techniques |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110428443A (en) * | 2019-08-11 | 2019-11-08 | 上海天诚比集科技有限公司 | A kind of intelligence community Vehicle tracing method |
| CN111539343A (en) * | 2020-04-26 | 2020-08-14 | 安徽皖仪科技股份有限公司 | Black smoke vehicle detection method based on convolution attention network |
| CN111539343B (en) * | 2020-04-26 | 2023-05-23 | 安徽皖仪科技股份有限公司 | Black smoke vehicle detection method based on convolution attention network |
| US20220076032A1 (en) * | 2020-09-08 | 2022-03-10 | Nvidia Corporation | Adaptive object tracking algorithm for autonomous machine applications |
| US11636689B2 (en) * | 2020-09-08 | 2023-04-25 | Nvidia Corporation | Adaptive object tracking algorithm for autonomous machine applications |
| CN113378629A (en) * | 2021-04-27 | 2021-09-10 | 阿里云计算有限公司 | Method and device for detecting abnormal vehicle in smoke discharge |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109191492B (en) | 2021-07-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106935035B (en) | Parking offense vehicle real-time detection method based on SSD neural network | |
| CN105512623B (en) | Based on multisensor travelling in fog day vision enhancement and visibility early warning system and method | |
| CN102819764B (en) | Method for counting pedestrian flow from multiple views under complex scene of traffic junction | |
| CN109191492A (en) | A kind of intelligent video black smoke vehicle detection method based on edge analysis | |
| CN102903124B (en) | A kind of moving target detecting method | |
| CN100545867C (en) | Aerial shooting traffic video frequency vehicle rapid checking method | |
| CN103324930B (en) | A kind of registration number character dividing method based on grey level histogram binaryzation | |
| CN111800507A (en) | Traffic monitoring method and traffic monitoring system | |
| CN104599502A (en) | Method for traffic flow statistics based on video monitoring | |
| CN113850123A (en) | Video-based road monitoring method and device, storage medium and monitoring system | |
| CN108830246B (en) | Multi-dimensional motion feature visual extraction method for pedestrians in traffic environment | |
| CN104504908A (en) | Monitoring method and system for illegal parking | |
| CN111626170B (en) | Image recognition method for railway side slope falling stone intrusion detection | |
| CN101968924A (en) | High-resolution image-based real-time traffic flow detection method | |
| CN102867417A (en) | Taxi anti-forgery system and taxi anti-forgery method | |
| CN107644206A (en) | A kind of road abnormal behaviour action detection device | |
| CN109086682A (en) | A kind of intelligent video black smoke vehicle detection method based on multi-feature fusion | |
| Park et al. | Real-time signal light detection | |
| CN109147393A (en) | Vehicle lane change detection method based on video analysis | |
| CN111008574A (en) | Key person track analysis method based on body shape recognition technology | |
| CN113505638B (en) | Method and device for monitoring traffic flow and computer readable storage medium | |
| CN106339657A (en) | Straw incineration monitoring method and device based on monitoring video | |
| CN111523397A (en) | Intelligent lamp pole visual identification device, method and system and electronic equipment | |
| CN105957341A (en) | Wide area traffic jam detection method based on unmanned plane airborne platform | |
| CN111144301A (en) | Road pavement defect quick early warning device based on degree of depth learning |
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 |