CN109191492B - Intelligent video black smoke vehicle detection method based on contour analysis - Google Patents
Intelligent video black smoke vehicle detection method based on contour analysis Download PDFInfo
- Publication number
- CN109191492B CN109191492B CN201810754421.1A CN201810754421A CN109191492B CN 109191492 B CN109191492 B CN 109191492B CN 201810754421 A CN201810754421 A CN 201810754421A CN 109191492 B CN109191492 B CN 109191492B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- contour
- 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.)
- Active
Links
Images
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)
- Image Analysis (AREA)
- Closed-Circuit Television Systems (AREA)
Abstract
The invention discloses an intelligent video black smoke vehicle detection method based on contour analysis, which comprises the following steps: (1) extracting a moving target from a road monitoring video by using a foreground detection algorithm; (2) removing non-vehicle targets and tracking the vehicle targets; (3) extracting the contour of the rear part of the vehicle target, calculating a series of static features and dynamic features based on the contour, and fusing to form a feature vector; (4) and classifying the feature vectors extracted by each frame by using an SVM classifier, identifying the black smoke vehicle through multi-frame analysis, and automatically retaining evidences such as the license plate, the passing place and the passing time of the black smoke vehicle. The method can make up the defect that the traditional manual monitoring black smoke vehicle is low in efficiency, reduce the false alarm rate and have more obvious algorithm advantages in windy weather.
Description
Technical Field
The invention relates to the technical field of smoke and fire detection, in particular to an intelligent video black smoke vehicle detection method based on contour analysis.
Background
The black smoke vehicle is a highly polluted vehicle which emits dense black smoke at the tail gas hole of the vehicle. The pollution of black smoke vehicles is always the key and difficult point of the environmental protection of motor vehicles. Particulate Matter (PM) in black smoke not only pollutes air, but also can cause damage to respiratory tracts and lungs of contacters, and harm to human health. Therefore, timely discovery of black smoke cars and further processing will help to improve city air quality.
The current methods for detecting black smoke vehicles can be roughly divided into two types: (1) the conventional method. The traditional method mainly comprises the following steps: the method comprises the following steps of mass reporting, periodic road inspection, night patrol, manual video monitoring, installation of a vehicle tail gas analysis device, sensor detection and the like. The methods reduce the pollution of black smoke vehicles to a certain extent, but because the holding quantity of motor vehicles is increased sharply and the traffic is busy, the methods usually need to invest a large amount of manpower and financial resources and have low efficiency; (2) provided is an intelligent video monitoring method. The intelligent video monitoring of the black smoke vehicle is that the black smoke vehicle is automatically detected from a mass of road monitoring videos by utilizing a computer vision technology, video related data are automatically uploaded to an environmental protection department, and evidences such as license plates, vehicle passing places and vehicle passing time of the black smoke vehicle are kept. The method belongs to remote monitoring, does not hinder traffic, can realize all-antenna online watching, is suitable for various road environments such as double lanes, multiple lanes and the like, is convenient to install, is suitable for large-range distribution and control of urban roads, is easier to form an online monitoring network for high-pollution black smoke vehicles, and improves law enforcement efficiency. However, such methods are still in the beginning of research.
The intelligent video black smoke vehicle detection method based on the contour analysis can improve the law enforcement efficiency and make up for the defect that the traditional method for manually monitoring the black smoke vehicle is low in efficiency. The invention adopts a strategy of directly analyzing the area behind the vehicle, greatly reduces the false alarm rate and avoids false detection caused by leaf shaking, cloud movement and the like. In addition, the designed characteristics are superior to those used by the traditional smoke and fire detection, and the algorithm has more obvious advantages in windy weather.
Disclosure of Invention
The invention aims to solve the technical problem of providing an intelligent video black smoke vehicle detection method based on contour analysis, which can make up the defect of low efficiency of the traditional manual black smoke vehicle monitoring, reduce the false alarm rate and have more obvious algorithm advantages in windy weather.
In order to solve the technical problem, the invention provides an intelligent video black smoke vehicle detection method based on contour analysis, which comprises the following steps:
(1) extracting a moving target from a road monitoring video by using a foreground detection algorithm;
(2) removing non-vehicle targets and tracking the vehicle targets;
(3) extracting the contour of the rear part of the vehicle target, calculating a series of static features and dynamic features based on the contour, and fusing to form a feature vector;
(4) and classifying the feature vectors extracted by each frame by using an SVM classifier, identifying the black smoke vehicle through multi-frame analysis, and automatically retaining evidences such as the license plate, the passing place and the passing time of the black smoke vehicle.
Preferably, the foreground detection algorithm in the step (1) adopts a Vibe background difference algorithm.
Preferably, the removing of the non-vehicle object in the step (2) includes the steps of:
(21) the area S of the moving object should be larger than the minimum value S of the general vehicle object area0I.e. by
S>S0;
(22) The ratio of the width to height of the circumscribed rectangle of the moving object should be within a certain range, i.e. the ratio
Wherein [ delta ]1,δ2]A range of ratios of width to height of a circumscribed rectangle that is a vehicle target;
(23) and the moving target meeting the two rules is regarded as the vehicle target.
Preferably, the tracking of the vehicle target in the step (2) includes the steps of:
(24) calculating the position of the central point of the circumscribed rectangle of each vehicle target in each frame of image;
(25) calculating Euclidean distances between the center of a certain vehicle target in the previous frame and the centers of all vehicle targets in the next frame, regarding two vehicle targets in the previous frame and the next frame with the smallest distance as the same vehicle, and performing such processing on each vehicle target in the previous frame to obtain the corresponding position of each vehicle target in the previous frame in the next frame;
(26) and (4) for each frame of the video, acquiring the front-back corresponding relation of each vehicle in two adjacent frames according to the method in the step (25), namely equivalently tracking the vehicle.
Preferably, the calculation of the series of static features in step (3) comprises the following steps:
(31) extracting a contour C of a vehicle objectobj;
(32) Extracting a vehicle contour CobjConvex hull Hobj;
(33) Calculating the vehicle rear contour C using the following equationrearStarting point coordinates P ofstart(xstart,ystart) And end point coordinates Pend(xend,yend),
Wherein, Ocenter(xcenter,ycenter) And r represents the profile C, respectivelyobjK is a coefficient for adjusting the size of the proposed contour range;
convex hull H of vehicle rear profilerearThe start point coordinate and the end point coordinate of (2) are also the point Pstart(xstart,ystart) And point Pend(xend,yend);
(34) Length L of convex hull of vehicle rear contourconvex_rearAnd length L of the vehicle rear profilecontour_rearThe ratio Rat1 can be used as an important static characteristic for judging whether black smoke exists behind the vehicle or not, namely
(35) With Scontour_rearIs represented by line segment PendOcenterLine segment OcenterPstartAnd a vehicle rear profile CrearThe area of the enclosed polygon; from Sconvex_rearIs represented by line segment PendOcenterLine segment OcenterPstartConvex hull H of contour behind vehiclerearArea of the enclosed polygon, Scontour_rearAnd Sconvex_rearCan be used as an important static characteristic for judging whether black smoke exists behind the vehicle, namely
(36) Denotes Δ P by FstartOrearPendThe center of gravity of; with SregionIs represented by line segment PstartF. Line segment FPendAnd a vehicle rear profile CrearA polygonal area, some statistical characteristics of which can be used to determine whether there is black smoke behind the vehicle, i.e. the vehicle is in a dark state
Wherein N isregionIndicating the region SregionNumber of pixels of (S)region(i, j) denotes the region SregionThe pixel value at position (i, j).
Preferably, the extracting of the series of dynamic features in step (3) includes the following steps:
(37) extracting a contour C of a vehicle objectobjAnd the center O of the circumscribed circle of the outlinecenter;
(38) Calculating the center of a circle OcenterOn the horizontal straight line and the profile CobjTwo right and left intersection points PleftAnd PrightAnd is represented by X by the line segment PrightPleftAnd a contour CobjA polygonal outline is formed by encircling;
(39) for the same vehicle with K frames apart, calculating the outline X according to the step 3.8, and respectively recording the outline X as A and B;
(311) calculating the matching degree M of the contour A and the contour Bmatch(A, B), we have designed two matching degree calculation methods, respectively,
wherein the content of the first and second substances,andseven Hu moments of invariance representing profile a and profile B, respectively, sign (x) representing a sign function, log (x) representing a logarithmic function;
(312) using the degree of matching M of contours in different frames of the same vehiclematch(A, B) to determine whether or not black smoke is present behind the vehicle.
Preferably, the Hu invariant moment h in step (310)i(i 1, 2.., 7.) is calculated by,
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=(3η21-η03)(η30+η12)[(η30+η12)2-3(η21+η03)2]+(3η21-η30)(η21+η03)[3(η30+η12)2-(η21+η03)2]
wherein, mupqIs the central moment, f (x, y) is the two-dimensional contour, pointFor the center of gravity of the contour, the calculation method is,
m in the above formulapqIs a two-dimensional p + q order conventional moment, and the calculation method is,
preferably, the SVM classifier in step (4) is obtained by training, and the classifier is used for testing new sample data.
Preferably, the recognition of a vehicle as a black smoke vehicle in step (4) requires that the following two rules be satisfied simultaneously,
rule one is as follows: omega2>α1
wherein, ω is1Is the number of frames, ω, detected in the video by the same vehicle2All detected ω1Among the frames, the number of frames identified as black smoke cars.
The invention has the beneficial effects that: (1) the law enforcement efficiency is improved, and the defect that the traditional manual monitoring black cigarette vehicle is low in efficiency is overcome; automatically detecting black smoke vehicles from a mass of road monitoring videos by using a computer vision technology, automatically uploading video related data to an environmental protection department, and simultaneously reserving evidences such as license plates, vehicle passing places, vehicle passing time and the like of the black smoke vehicles; the method belongs to remote monitoring, does not hinder traffic, can realize all-antenna online watching, is suitable for various road environments such as double lanes, multiple lanes and the like, is convenient to install, is suitable for large-range distribution and control of urban roads, is easier to form an online monitoring network for high-pollution black smoke vehicles, and improves law enforcement efficiency; (2) the false alarm rate is reduced; according to the technical scheme provided by the invention, the area behind the vehicle is directly analyzed, so that the false alarm rate is greatly reduced, and false detection caused by leaf shaking, cloud movement and the like is avoided; (3) the designed characteristics are superior to those used by the traditional smoke and fire detection, and the algorithm has more obvious advantages in windy weather; the dynamic characteristic of the matching degree provided by the technical scheme of the invention is used for describing the matching degree of the vehicle rear profile between different frames of the same vehicle, and the larger the wind is, the more obvious the change of the vehicle rear profile is, but the vehicle rear profile of the non-black smoke vehicle is not changed in the same time interval, so that the black smoke vehicle and the non-black smoke vehicle can be more easily distinguished.
Drawings
FIG. 1 is a schematic flow chart of the method of the present invention.
FIG. 2 is a schematic view of the contour and pocket of the rear of the vehicle object of the present invention.
FIG. 3 is a schematic view of the vehicle rear contour and crate extracted from a black smoke vehicle target according to the present invention.
FIG. 4 is a schematic drawing of the vehicle rear contour and crate extracted from two non-black smoke vehicle targets in accordance with the present invention.
FIG. 5 is a schematic view of the outline of the black smoke vehicle of the present invention.
FIG. 6 is a schematic view of a variation of the profile of the non-black smoke vehicle of the present invention.
Detailed Description
The invention provides an intelligent video black smoke vehicle detection method based on contour analysis, which is shown in a flow chart of figure 1 and specifically comprises the following steps:
step 1: extracting a moving target from a road monitoring video by using a foreground detection algorithm;
step 2: removing non-vehicle targets and tracking the vehicle targets;
and step 3: extracting the contour of the rear part of the vehicle target, calculating a series of static features and dynamic features based on the contour, and fusing to form a feature vector;
and 4, step 4: and classifying the feature vectors extracted by each frame by using an SVM classifier, identifying the black smoke vehicle through multi-frame analysis, and automatically retaining evidences such as the license plate, the passing place and the passing time of the black smoke vehicle.
The foreground detection algorithm in the step 1 adopts a Vibe background difference algorithm.
The removing of the non-vehicle target in the step 2 comprises the following steps:
step 2.1: the area S of the moving object should be larger than the minimum value S of the general vehicle object area0I.e. by
S>S0;
Step 2.2: the ratio of the width to height of the circumscribed rectangle of the moving object should be within a certain range, i.e. the ratio
Wherein [ delta ]1,δ2]A range of ratios of width to height of a circumscribed rectangle that is a vehicle target;
step 2.3: and the moving target meeting the two rules is regarded as the vehicle target.
The tracking of the vehicle target in the step 2 comprises the following steps:
step 2.4: calculating the position of the central point of the circumscribed rectangle of each vehicle target in each frame of image;
step 2.5: calculating Euclidean distances between the center of a certain vehicle target in the previous frame and the centers of all vehicle targets in the next frame, regarding two vehicle targets in the previous frame and the next frame with the smallest distance as the same vehicle, and performing such processing on each vehicle target in the previous frame to obtain the corresponding position of each vehicle target in the previous frame in the next frame;
step 2.6: for each frame of the video, the front-back corresponding relation of each vehicle in two adjacent frames is obtained according to the method of step 2.5, namely, the method is equivalent to tracking the vehicle.
The calculation of the series of static features in the step 3 comprises the following steps:
step 3.1: extracting a contour C of a vehicle objectobj;
Step 3.2: extracting a vehicle contour CobjConvex hull Hobj;
Step 3.3: calculating the vehicle rear contour C using the following equationrearStarting point coordinates P ofstart(xstart,ystart) And end point coordinates Pend(xend,yend),
Wherein, Ocenter(xcenter,ycenter) And r represents the profile C, respectivelyobjK is a coefficient for adjusting the size of the mentioned contour range;
Convex hull H of vehicle rear profilerearThe start point coordinate and the end point coordinate of (2) are also the point Pstart(xstart,ystart) And point Pend(xend,yend);
FIG. 2 shows a schematic representation of the contour and the convex hull of the rear of a vehicle object, FIG. 3 shows a schematic representation of the contour and the concave hull of the rear of a vehicle extracted from one black smoke vehicle object, FIG. 4 shows a schematic representation of the contour and the concave hull of the rear of a vehicle extracted from two non-black smoke vehicle objects, it can be seen that for black smoke vehicles the contour and the convex hull of the rear of a vehicle appear significantly different, whereas for non-black smoke vehicles the contour and the convex hull of the rear of a vehicle are almost coincident;
step 3.4: length L of convex hull of vehicle rear contourconvex_rearAnd length L of the vehicle rear profilecontour_rearThe ratio Rat1 can be used as an important static characteristic for judging whether black smoke exists behind the vehicle or not, namely
Step 3.5: with Scontour_rearIs represented by line segment PendOcenterLine segment OcenterPstartAnd a vehicle rear profile CrearThe area of the enclosed polygon; from Sconvex_rearIs represented by line segment PendOcenterLine segment OcenterPstartConvex hull H of contour behind vehiclerearArea of the enclosed polygon, Scontour_rearAnd Sconvex_rearCan be used as an important static characteristic for judging whether black smoke exists behind the vehicle, namely
Step 3.6: denotes Δ P by FstartOrearPendThe center of gravity of; with SregionIs represented by line segment PstartF. Line segmentFPendAnd a vehicle rear profile CrearA polygonal area, some statistical characteristics of which can be used to determine whether there is black smoke behind the vehicle, i.e. the vehicle is in a dark state
Wherein N isregionIndicating the region SregionNumber of pixels of (S)region(i, j) denotes the region SregionThe pixel value at position (i, j).
The extraction of the series of dynamic features in the step 3 comprises the following steps:
step 3.7: extracting a contour C of a vehicle objectobjAnd the center O of the circumscribed circle of the outlinecenterRespectively taking black smoke vehicles and non-black smoke vehicles as examples, fig. 5 shows a schematic change diagram of the profile of the black smoke vehicle, fig. 6 shows a schematic change diagram of the profile of the non-black smoke vehicle, and it can be seen that the black smoke vehicle is more obvious than the non-black smoke vehicle for the change of the profile behind the vehicle along with time;
step 3.8: calculating the center of a circle OcenterOn the horizontal straight line and the profile CobjTwo right and left intersection points PleftAnd PrightAnd is represented by X by the line segment PrightPleftAnd a contour CobjA polygonal outline is formed by encircling;
step 3.9: for the same vehicle with K frames apart, calculating the outline X according to the step 3.8, and respectively recording the outline X as A and B;
step 3.11: calculating the matching degree M of the contour A and the contour Bmatch(A, B), we have designed two matching degree calculation methods, respectively,
wherein the content of the first and second substances,andseven Hu moments of invariance representing profile a and profile B, respectively, sign (x) representing a sign function, log (x) representing a logarithmic function;
step 3.12: using the degree of matching M of contours in different frames of the same vehiclematch(A, B) to determine whether or not black smoke is present behind the vehicle.
Pertaining Hu invariant moment h in step 3.10i(i 1, 2.., 7.) is calculated by,
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=(3η21-η03)(η30+η12)[(η30+η12)2-3(η21+η03)2]+(3η21-η30)(η21+η03)[3(η30+η12)2-(η21+η03)2]
wherein, mupqIs the central moment, f (x, y) is the two-dimensional contour, pointFor the center of gravity of the contour, the calculation method is,
m in the above formulapqIs a two-dimensional p + q order conventional moment, and the calculation method is,
and 4, training the SVM classifier in the step 4, and using the classifier for testing new sample data.
The recognition of a vehicle as a black smoke vehicle in said step 4 requires that the following two rules are satisfied simultaneously,
rule one is as follows: omega2>α1
wherein, ω is1Is the number of frames, ω, detected in the video by the same vehicle2All detected ω1Among the frames, the number of frames identified as black smoke cars.
Claims (6)
1. An intelligent video black smoke vehicle detection method based on contour analysis is characterized by comprising the following steps:
(1) extracting a moving target from a road monitoring video by using a foreground detection algorithm;
(2) removing non-vehicle targets and tracking the vehicle targets;
(3) extracting the contour of the rear part of the vehicle target, calculating a series of static features and dynamic features based on the contour, and fusing to form a feature vector;
the calculation of a series of static features comprises the following steps:
(31) extracting a contour C of a vehicle objectobj;
(32) Extracting a vehicle contour CobjConvex hull Hobj;
(33) Calculating the vehicle rear contour C using the following equationrearStarting point coordinates P ofstart(xstart,ystart) And end point coordinates Pend(xend,yend),
Wherein, Ocenter(xcenter,ycenter) And r represents the profile C, respectivelyobjK is a coefficient for adjusting the size of the proposed contour range;
convex hull H of vehicle rear profilerearThe start point coordinate and the end point coordinate of (2) are also the point Pstart(xstart,ystart) And point Pend(xend,yend);
(34) Length L of convex hull of vehicle rear contourconvex_rearAnd length L of the vehicle rear profilecontour_rearThe ratio Rat1 can be used as an important static characteristic for judging whether black smoke exists behind the vehicle or not, namely
(35) With Scontour_rearIs represented by line segment PendOcenterLine segment OcenterPstartAnd a vehicle rear profile CrearThe area of the enclosed polygon; from Sconvex_rearIs represented by line segment PendOcenterLine segment OcenterPstartConvex hull H of contour behind vehiclerearArea of the enclosed polygon, Scontour_rearAnd Sconvex_rearCan be used as an important static characteristic for judging whether black smoke exists behind the vehicle, namely
(36) By F Δ PstartOrearPendThe center of gravity of; with SregionIs represented by line segment PstartF. Line segment FPendAnd a vehicle rear profile CrearA polygonal area, some statistical characteristics of which can be used to determine whether there is black smoke behind the vehicle, i.e. the vehicle is in a dark state
Wherein N isregionIndicating the region SregionNumber of pixels of (S)region(i, j) denotes the region SregionThe pixel value at location (i, j);
the extraction of a series of dynamic features comprises the following steps:
(37) extracting a contour C of a vehicle objectobjAnd the center O of the circumscribed circle of the outlinecenter;
(38) Calculating the center of a circle OcenterOn the horizontal straight line and the profile CobjTwo right and left intersection points PleftAnd PrightAnd is represented by X by the line segment PrightPleftAnd a contour CobjA polygonal outline is formed by encircling;
(39) calculating the contour X according to the step (38) for the same vehicle which is separated by K frames, and respectively recording the contour X as A and B;
(310) calculating the Hu invariant moment of the contour A and the contour B, and respectively recording the Hu invariant moment as hi AAnd hi B(ii) a Wherein, i is 1,2,. 7;
(311) calculating the matching degree M of the contour A and the contour Bmatch(A, B), designing two matching degree calculation methods, respectively,
Wherein the content of the first and second substances,andseven Hu invariant moments, i 1,2,.., 7, sign (x) representing a sign function, and log (x) representing a log function, representing profile a and profile B, respectively;
(312) using the degree of matching M of contours in different frames of the same vehiclematch(A, B) judging whether black smoke is present behind the vehicle;
(4) and classifying the feature vectors extracted by each frame by using an SVM classifier, identifying the black smoke vehicle through multi-frame analysis, and automatically keeping the evidence of the license plate, the passing place and the passing time of the black smoke vehicle.
2. The intelligent video black smoke vehicle detection method based on the contour analysis as claimed in claim 1, wherein the foreground detection algorithm in step (1) adopts a Vibe background difference algorithm.
3. The intelligent video black smoke vehicle detection method based on contour analysis as defined in claim 1, wherein the removing of non-vehicle targets in step (2) comprises the following steps:
(21) the area S of the moving object should be larger than the minimum value S of the general vehicle object area0I.e. by
S>S0;
(22) The ratio of the width to height of the circumscribed rectangle of the moving object should be within a certain range, i.e. the ratio
Wherein [ delta ]1,δ2]A range of ratios of width to height of a circumscribed rectangle that is a vehicle target;
(23) the moving target meeting the two rules is regarded as a vehicle target;
the tracking of the vehicle target in the step (2) comprises the following steps:
(24) calculating the position of the central point of the circumscribed rectangle of each vehicle target in each frame of image;
(25) calculating Euclidean distances between the center of a certain vehicle target in the previous frame and the centers of all vehicle targets in the next frame, regarding two vehicle targets in the previous frame and the next frame with the smallest distance as the same vehicle, and performing such processing on each vehicle target in the previous frame to obtain the corresponding position of each vehicle target in the previous frame in the next frame;
(26) and (4) for each frame of the video, acquiring the front-back corresponding relation of each vehicle in two adjacent frames according to the method in the step (25), namely equivalently tracking the vehicle.
4. The intelligent video black smoke vehicle detection method based on the contour analysis as claimed in claim 1,characterized by the Hu moment of invariance h in step (310)iThe calculation method of (a) is that,
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=(3η21-η03)(η30+η12)[(η30+η12)2-3(η21+η03)2]+(3η21-η30)(η21+η03)[3(η30+η12)2-(η21+η03)2]
wherein, i is 1,2pqIs the central moment, f (x, y) is the two-dimensional contour, pointFor the center of gravity of the contour, the calculation method is,
m in the above formulapqIs a two-dimensional p + q order conventional moment, and the calculation method is,
5. the intelligent video black smoke vehicle detection method based on contour analysis as claimed in claim 1, wherein the SVM classifier in step (4) is obtained by training, and the classifier is used for testing new sample data.
6. The intelligent video black smoke vehicle detection method based on the contour analysis as claimed in claim 1, wherein the recognition of a vehicle as a black smoke vehicle in the step (4) requires the following two rules to be satisfied simultaneously,
rule one is as follows: omega2>α1
wherein, ω is1Is the number of frames, ω, detected in the video by the same vehicle2All detected ω1In the frame, is recognized as blackNumber of frames of smoking cars.
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 CN109191492A (en) | 2019-01-11 |
CN109191492B true 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) |
Families Citing this family (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 |
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 |
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 |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101029824B (en) * | 2006-02-28 | 2011-10-26 | 东软集团股份有限公司 | Method and apparatus for positioning vehicle based on characteristics |
US8655588B2 (en) * | 2011-05-26 | 2014-02-18 | Crown Equipment Limited | Method and apparatus for providing accurate localization for an industrial vehicle |
CN102737247B (en) * | 2012-07-04 | 2014-08-13 | 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 |
-
2018
- 2018-07-11 CN CN201810754421.1A patent/CN109191492B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN109191492A (en) | 2019-01-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109191492B (en) | Intelligent video black smoke vehicle detection method based on contour analysis | |
CN111368687B (en) | Sidewalk vehicle illegal parking detection method based on target detection and semantic segmentation | |
CN109977782B (en) | Cross-store operation behavior detection method based on target position information reasoning | |
CN109190444B (en) | Method for realizing video-based toll lane vehicle feature recognition system | |
CN112101433A (en) | Automatic lane-dividing vehicle counting method based on YOLO V4 and DeepsORT | |
CN103093249B (en) | A kind of taxi identification method based on HD video and system | |
CN109086682B (en) | Intelligent video black smoke vehicle detection method based on multi-feature fusion | |
CN107665327B (en) | Lane line detection method and device | |
CN105844257A (en) | Early warning system based on machine vision driving-in-fog road denoter missing and early warning method | |
CN113850123A (en) | Video-based road monitoring method and device, storage medium and monitoring system | |
CN103714363A (en) | Motor vehicle exhaust smoke video identification system | |
CN101968924A (en) | High-resolution image-based real-time traffic flow detection method | |
CN108416316B (en) | Detection method and system for black smoke vehicle | |
CN109191495B (en) | Black smoke vehicle detection method based on self-organizing background difference model and multi-feature fusion | |
CN110929676A (en) | Deep learning-based real-time detection method for illegal turning around | |
CN109657599B (en) | Picture identification method of distance-adaptive vehicle appearance part | |
CN111523397A (en) | Intelligent lamp pole visual identification device, method and system and electronic equipment | |
CN110852177A (en) | Obstacle detection method and system based on monocular camera | |
CN107315998A (en) | Vehicle class division method and system based on lane line | |
CN111626170A (en) | Image identification method for railway slope rockfall invasion limit detection | |
CN111723704A (en) | Raspberry pie-based van body door opening monitoring method | |
CN111144301A (en) | Road pavement defect quick early warning device based on degree of depth learning | |
CN113450573A (en) | Traffic monitoring method and traffic monitoring system based on unmanned aerial vehicle image recognition | |
CN113903012A (en) | Collision early warning method and device, vehicle-mounted equipment and storage medium | |
CN113657305B (en) | Video-based intelligent detection method for black smoke vehicle and ringeman blackness level |
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 |