CN105574502A - Automatic detection method for violation behaviors of self-service card sender - Google Patents
Automatic detection method for violation behaviors of self-service card sender Download PDFInfo
- Publication number
- CN105574502A CN105574502A CN201510941325.4A CN201510941325A CN105574502A CN 105574502 A CN105574502 A CN 105574502A CN 201510941325 A CN201510941325 A CN 201510941325A CN 105574502 A CN105574502 A CN 105574502A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- target
- image
- value
- rsqb
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/35—Categorising the entire scene, e.g. birthday party or wedding scene
- G06V20/38—Outdoor scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/584—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of vehicle lights or traffic lights
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses an automatic detection method for violation behaviors of a self-service card sender. The behaviors of vehicles and pedestrians in an area of the self-service card sender are automatically identified by video monitoring means; the automatic detection method comprises three main steps, which are object segmentation, violation behavior identification and violation behavior detection respectively; and the treatment steps in the automatic detection method all are treated in sequence aiming at monitored video images of the area of the self-service card sender.
Description
Technical field
The invention belongs to technical field of intelligent traffic, particularly a kind of Self-service card sender act of violating regulations automatic testing method.
Background technology
Along with socioeconomic fast development, China's highway traffic volume is growing, and charge station faces staff wretched insufficiency, the problem that efficiency is too low.Therefore, be provided with Self-help card-distributing track in the charge station of some staffing deficiencies, facilitate the self-service fast card-taking of driver, quickly through.
But high-tech, bringing us easily simultaneously, also progressively shows some leaks.Some lawless persons utilize the leak of automatic card dispenser, repeatedly card taking, profiteering high speed visa card, thus carry out a series of illegal activity such as fee evasion, payment omitted.This behavior allows national economy suffer a loss on the one hand, also upsets normal current order at a high speed on the other hand, causes traffic safety hidden danger.
Along with computer video analytical technology is in the widespread use of all trades and professions, intelligent analysis system can allow front-end camera in real time automatically " discovery situation ", and the supervision target in active " analysis " visual field, judge whether the behavior of these monitored targets exists abnormal operation simultaneously, to the abnormal behaviour occurring maybe will occurring, give the alarm to operator on duty in time, conscientiously improve the safe precaution ability of highway, realize Intellectualized monitoring.
Summary of the invention
The object of this invention is to provide a kind of Self-service card sender act of violating regulations automatic testing method, to solve for the automatic identification for the illegal activities of Self-service card sender in highway.
Technical scheme of the present invention is, a kind of Self-service card sender act of violating regulations automatic testing method, adopt video monitoring means, the vehicle in Self-service card sender region and the behavior of pedestrian are identified automatically, described automatic testing method comprises 3 main steps, be Target Segmentation respectively, act of violating regulations identification and act of violating regulations detect, the treatment step in described automatic testing method be all for Self-service card sender area monitoring video image according to sequencing process
Described Target Segmentation, comprises the background extracting that begins step by step, frame difference cuts target and morphologic filtering,
The performing step that described initial background is extracted is:
A1) add up N continuous frame video situation of change, in recording pixel there is situation in a some gray scale
Wherein, P (x, y, k) represents that pixel (x, y) place brightness value is the number of times that k occurs, image
i(x, y, m) represents that a certain two field picture pixel (x, y) place brightness value is m,
A2) by N continuous frame point gray scale frequency of occurrences maximal value, as the gray-scale value of this point, i.e. initial background gray-scale value.
Background(x,y)=max(P(x,y,k))k=0,1,2…255(1.2)
The specific implementation step that described frame difference cuts target is:
B1) for convenience of subsequent calculations, first the result after sampling is carried out blocking, if original image width and be highly W, H, the size of block is w, h, then the image size after blocking is
B2) present image and background image is used to carry out difference, to obtain moving target, wherein DifGray
ifor the gray-scale value of certain block after background difference, Gray
nfor grey scale pixel value in present frame block, Background
nfor grey scale pixel value in corresponding blocks in background
B3) the iterative Threshold selection method of the choice for use of binary-state threshold,
A) select the intermediate value of the gray scale in video image as estimation threshold value T first
0;
B) the threshold value T starting to estimate is utilized
0the gray-scale value of image is divided into two different regions: R
1, R
2, according to formula (1.5) zoning R
1and R
2the average u of gray scale
1and u
2:
C) u is calculated
1and u
2after, calculate the threshold value T made new advances
i+1:
D) repeat step B), C), until T
i+1and T
iduring infinite approach, its value is binary-state threshold T,
B4) when difference result is greater than threshold value T, then this agllutination fruit is set to 255, otherwise is set to 0, the binaryzation of realize target,
The step that described morphologic filtering realizes is:
C1) transversal scanning is carried out to binarization segmentation result, when adjacent two white interblock gaps are less than 2 blocks, then the black patch of zone line is set to 255, otherwise retain initial value;
C2) longitudinal scanning is carried out to binarization segmentation result, when adjacent two white interblock gaps are less than 2 blocks, then the black patch of zone line is set to 255, otherwise retain initial value;
Described act of violating regulations identification, comprises vehicle location step by step and vehicle cab recognition,
The performing step of described vehicle location is:
D1) within the scope of statistical picture each capable be not 0 pixel, record its number, obtain the width of this row image,
D2) statistical picture scope Nei Gelie is not the pixel of 0, records its number, obtains the height of this row image,
According to d1) and statistics d2), the positional information of target and the Width x Height information of target can being obtained, judging whether it captures region, as then identified vehicle in surveyed area in act of violating regulations according to the positional information of target, otherwise, it is not detected;
Described vehicle cab recognition performing step is:
For detection target machine motor-car, motorcycle and pedestrian, motorcycle and pedestrian are not distinguished, determines motor vehicle and motorcycle object module, namely in image display area, the masterplate size of motor vehicle and motorcycle is set, to facilitate subsequent detection,
Motor vehicle:
S
machine-S
machine mould| <20 (1.8)
Motorcycle:
S
rub-S
rub mould| <20 (1.9)
Wherein, η is the ratio of target area Width x Height, and obtain according to many experiments for motor vehicle, motorcycle η span, S=width*height is target area area, according to itself and the comparison of stencil area size, judges type belonging to target;
Described act of violating regulations detects, and comprises and in surveyed area, steals card behavioral value to drive in the wrong direction behavioral value and motorcycle and pedestrian of motor vehicles,
Motor vehicles drive in the wrong direction behavioral value, and by following the tracks of vehicle target track, judged whether that retrograde behavior occurs according to its trajectory direction, specific implementation step is:
E1) adopt angle point track algorithm, in object run process, with change template forward minimal difference Matching pursuitalgorithm, the synchronous position found angle point and occur in every two field picture, whenever finding new target location, then upgrades masterplate data with current location information;
E2) carry out the demarcation of the correct travel direction of vehicle to every row pixel, due to the restriction of sample frequency and car speed, being set in its movement locus in adjacent three pixel coverages is even variation, and now the error of calculation is less;
E3) demarcate each row pixel travel direction respectively from the row pixel of image lower edge the 0th, for being the section travelling positive dirction away from camera direction, its row k direction vector is respectively:
For driving towards the section that camera direction is traveling positive dirction, its row k direction vector is respectively:
E4) basis is for retrograde vehicle, the angle of its trajectory direction and the correct travel direction of vehicle has certain feature, when being occurred to drive in the wrong direction by vehicle, the angle of pursuit path and the correct travel direction of calibration vehicle diatom is defined as 120 °, regulation is for every objective movement locus, when it and the track quantity of normal travel direction angle more than 120 ° is greater than the threshold value of setting time, then judge that this movement locus is as retrograde track, namely there is behavior of driving in the wrong direction in tracked vehicle;
E5)
for the vehicle of calibration vehicle diatom correctly travels vector,
for track actual motion vector, wherein,
then the angle theta of the correct travel direction of the vehicle of vehicle actual travel direction and regulation can be expressed as:
E6) in actual testing process, as θ >120 °, then corresponding counter is added 1, when meeting condition quantity of driving in the wrong direction and being greater than setting threshold value, then can judge that current tracking vehicle drives in the wrong direction;
Described card behavioral value of stealing refers to, for the target being defined as motorcycle and pedestrian in monitor video image detection region, then directly to report to the police to it, records its license board information;
For the automotive vehicle determined, if it has retrograde behavior, think that it has card behavior steathily, then it is reported to the police.
Further, before described Target Segmentation, act of violating regulations identification and act of violating regulations detect, also comprise pre-treatment step,
Described pre-service refers to the high resolving power due to monitor video, for reducing operand, carries out horizontal and longitudinal sample process to the image collected.
Compared with prior art, the present invention is adopted to have following technique effect:
1, large scene is monitored in real time
High-definition camera selected by front end camera, ensures that in whole monitoring range, vehicle image is clear, ensures can see the car plate entering guarded region vehicle clearly, so that the follow-up punishment for vehicles peccancy.
2, high precision test
This method carries out retrograde vehicle detection in conjunction with angle point tracking, is judged retrograde vehicle by orbiting motion direction.Use angle point tracking effectively can improve Detection accuracy, improve algorithm accuracy of detection, effectively eliminate the interference of other factors such as shade shake.
3, strong adaptability detects
This method is on traditional background estimating basis, Bring out Background update algorithm, the change of better adaptation scene, add change trickle in the discovery background that background update method can be sharp, effective elimination light, shake impact on small target deteection, the accurate segmentation of realize target, the judgement for successor provides foundation accurately.
Accompanying drawing explanation
Fig. 1 is CCTV camera scheme of installation in the embodiment of the present invention;
Fig. 2 is camera supervised realistic picture in the embodiment of the present invention;
Fig. 3 is Self-service card sender act of violating regulations automatic testing method principle schematic of the present invention;
Fig. 4 is background estimating result figure in the embodiment of the present invention;
Fig. 5 is initial target segmentation result figure in the embodiment of the present invention;
Fig. 6 is image vertical and horizontal perspective view in the embodiment of the present invention;
Fig. 7 is motorcycle and motor vehicle size comparison figure in the embodiment of the present invention;
Fig. 8 is matched jamming schematic diagram of the present invention;
Fig. 9 is that the correct travel direction in complex scene track of the present invention arranges schematic diagram.
Embodiment
In Specific construction process, video camera is arranged near the awning lamp in automatic card track, adopts the mounting means irradiating the tailstock, and as shown in Figure 1, video camera installation and monitoring range outdoor scene are as shown in Figure 2.
See Fig. 3, the present invention primarily of four part compositions, is specially Image semantic classification part, Target Segmentation part, act of violating regulations identification division, act of violating regulations detecting portion in implementation procedure.Because HD video resolution is higher, acquisition data message enriches, and for reducing operand, carries out transverse direction and longitudinally sampling at image pre-processing phase to the image collected, namely original image size is W*H, then the image size of actual treatment is (W/2) * (H/2).When carrying out Target Segmentation in conjunction with background grey scale change situation, real-time update background image, and carry out the difference of present image and background image, so that segmentation object more accurately, corresponding morphologic filtering is carried out to segmentation result simultaneously, adjacent block is carried out connected domain merging simultaneously, to obtain full object block, obtain the length of object block, width simultaneously, be convenient to the follow-up calculating such as length breadth ratio, area carrying out suspicious object block.To determine target vehicle, be motor vehicle or motorcycle, pedestrian, the basis obtaining object block will be followed the tracks of motor vehicle, and then acquisition target trajectory, this method adopts the trace tracking method based on angle point, whether be retrograde vehicle, and then follow-up judgement is carried out to illegal activities by the direction of motion determination target of track.
Implementation method of the present invention comprises:
One, Target Segmentation
1. initial background is extracted
This method adopts the histogrammic method of Corpus--based Method to carry out initial background estimation, this algorithm to move the hypothesis that can not rest on for a long time on a position based on moving target in scene, in section sometime, the pixel value that the video sequence specific pixel location place frequency of occurrences is the highest is exactly this background pixel value, adopts histogram method to carry out statistics to video sequence gray scale frequency of occurrences information.Its specific implementation step is:
1) N continuous frame is added up (by the test to multiple different scene and different timing statistics, finally determine, when N gets 200, to have and test effect preferably, therefore in actual use, get N=200) video situation of change, in recording pixel there is situation in individual some gray scale
Wherein, P (x, y, k) represents that pixel (x, y) place brightness value is the number of times that k occurs, image
i(x, y, m) represents that a certain two field picture pixel (x, y) place brightness value is m.
2) by N continuous frame point gray scale frequency of occurrences maximal value, as the gray-scale value of this point, i.e. initial background gray-scale value.
Background(x,y)=max(P(x,y,k))k=0,1,2…255(1.2)
Adopt histogram method to carry out the result of context update as shown in Figure 4, wherein N is 200.
2. frame difference cuts target
Acquisition real-time background image after, be find pedestrian target, need to split target, this algorithm adopt based on background subtraction method segmentation moving target, and to be partitioned into result carry out Mathematical morphology filter involve connected domain calculate, to obtain full moving target.Specific implementation step is as follows:
1) for convenience of subsequent calculations, first the result after sampling is carried out blocking, if original image width and be highly W, H, the size of block is w, h, then the image size after blocking is
2) present image and background image is used to carry out difference, to obtain moving target, wherein DifGray
ifor the gray-scale value of certain block after background difference, Gray
nfor grey scale pixel value in present frame block, Background
nfor grey scale pixel value in corresponding blocks in background
3) the iterative Threshold selection method of the choice for use of binary-state threshold
A) select the intermediate value of the gray scale in video image as estimation threshold value T first
0;
B) the threshold value T starting to estimate is utilized
0the gray-scale value of image is divided into two different regions: R
1, R
2, according to formula (1.5) zoning R
1and R
2the average u of gray scale
1and u
2:
C) u is calculated
1and u
2after, calculate the threshold value T made new advances
i+1:
D) repeat step B), C), until T
i+1and T
iduring infinite approach, its value is binary-state threshold T.
B4) when difference result is greater than threshold value T, then this agllutination fruit is set to 255, otherwise is set to 0, the binaryzation of realize target.
Use the method carries out the result after Target Segmentation as shown in Figure 5.
3. morphologic filtering
This part uses and fills adjacent binaryzation region blank position nearly based on the method for transverse direction and longitudinal scanning.
1) transversal scanning is carried out to binarization segmentation result, when adjacent two white interblock gaps are less than 2 blocks, then the black patch of zone line is set to 255, otherwise retain initial value;
2) longitudinal scanning is carried out to binarization segmentation result, when adjacent two white interblock gaps are less than 2 blocks, then the black patch of zone line is set to 255, otherwise retain initial value;
Two, act of violating regulations identification
1. vehicle location
This method, at the beginning of realization, need arrange illegal activities surveyed area, judge the behavior of surveyed area internal object.
1) within the scope of statistical picture each capable be not 0 pixel, record its number, obtain the width of this row image,
2) statistical picture scope Nei Gelie is not the pixel of 0, records its number, obtains the height of this row image,
Its result is as shown in Figure 6 after horizontal longitudinal scanning for target.
According to above-mentioned statistics, the positional information of target and the Width x Height information of target can be obtained, judge whether it captures region in act of violating regulations according to the positional information of target, as then identified vehicle in surveyed area, otherwise, it is not detected.
2. vehicle cab recognition
In actual use procedure, motor vehicle, motorcycle size are as shown in Figure 7.This method, at the beginning of realization, needs the size of Offered target object, namely determines the masterplate of motor vehicle, motorcycle and pedestrian target, so that when program realizes and the comparison of masterplate size, and then determine target type, vehicle is identified, in the method concrete differentiation is not done to motorcycle and pedestrian.
The target that the present invention mainly detects is motor vehicle, motorcycle and pedestrian, determines each object module, namely in image display area, arrange the masterplate size of motor vehicle and motorcycle according to detection target, to facilitate subsequent detection,
Motor vehicle:
S
machine-S
machine mould| <20 (1.8)
Motorcycle:
S
rub-S
rub mould| <20 (1.9)
In above formula, η is the ratio of target area Width x Height, and obtain according to many experiments for motor vehicle, motorcycle η span, S=width*height is target area area, according to itself and the comparison of stencil area size, judges type belonging to target.
Three, act of violating regulations detects
1. drive in the wrong direction behavioral value
1) angle point track algorithm is adopted.In object run process, with change template forward minimal difference Matching pursuitalgorithm, the synchronous position found angle point and occur in every two field picture, whenever finding new target location, then upgrade masterplate data with current location information, its schematic diagram as shown in Figure 8.
2) carry out the demarcation of the correct travel direction of vehicle to every row pixel, due to the restriction of sample frequency and car speed, can assert that its movement locus is even variation in adjacent three pixel coverages, now the error of calculation is less.
3) demarcate each row pixel travel direction respectively from the row pixel of image lower edge the 0th, for being the section travelling positive dirction away from camera direction, its row k direction vector is respectively:
For driving towards the section that camera direction is traveling positive dirction, its row k direction vector is respectively:
4) for retrograde vehicle, the angle of its trajectory direction and the correct travel direction of vehicle has certain feature, vehicle is in normal driving process, violent transformation can't be there is suddenly in its travel direction, in complex scene, the correct travel direction in track establishes intention as shown in Figure 9, when vehicle occurs to drive in the wrong direction by the present invention, the angle of pursuit path and the correct travel direction of calibration vehicle diatom is defined as 120 °, for the accurate detection of realization event, regulation is for every objective movement locus, when it and the track quantity of normal travel direction angle more than 120 ° is greater than the threshold value of setting time, then judge that this movement locus is as retrograde track, namely there is behavior of driving in the wrong direction in tracked vehicle.
5)
for the vehicle demarcated correctly travels vector,
for track actual motion vector, wherein,
then the angle theta of the correct travel direction of the vehicle of vehicle actual travel direction and regulation can be expressed as:
6) in actual testing process, as θ >120 °, then corresponding counter is added 1, when meeting condition quantity of driving in the wrong direction and being greater than setting threshold value, then can judge that current tracking vehicle drives in the wrong direction.
2. steal card behavioral value
For the target being defined as motorcycle and pedestrian in surveyed area, then direct it to be reported to the police, record its license board information; For the automotive vehicle determined, if it has retrograde behavior, think that it has card behavior steathily, then it is reported to the police.
Claims (2)
1. a Self-service card sender act of violating regulations automatic testing method, adopt video monitoring means, the vehicle in Self-service card sender region and the behavior of pedestrian are identified automatically, it is characterized in that, described automatic testing method comprises 3 main steps, be Target Segmentation respectively, act of violating regulations identification and act of violating regulations detect, the treatment step in described automatic testing method be all for Self-service card sender area monitoring video image according to sequencing process
Described Target Segmentation, comprises the background extracting that begins step by step, frame difference cuts target and morphologic filtering,
The performing step that described initial background is extracted is:
A1) add up N continuous frame video situation of change, in recording pixel there is situation in a some gray scale
Wherein, P (x, y, k) represents that pixel (x, y) place brightness value is the number of times that k occurs, image
i(x, y, m) represents that a certain two field picture pixel (x, y) place brightness value is m,
A2) by N continuous frame point gray scale frequency of occurrences maximal value, as the gray-scale value of this point, i.e. initial background gray-scale value.
Background (x, y)=max (P (x, y, k)) k=0,1,2 ... the specific implementation step that frame difference described in 255 (1.2) cuts target is:
B1) for convenience of subsequent calculations, first the result after sampling is carried out blocking, if original image width and be highly W, H, the size of block is w, h, then the image size after blocking is
B2) present image and background image is used to carry out difference, to obtain moving target, wherein DifGray
ifor the gray-scale value of certain block after background difference, Gray
nfor grey scale pixel value in present frame block, Background
nfor grey scale pixel value in corresponding blocks in background
B3) the iterative Threshold selection method of the choice for use of binary-state threshold,
A) select the intermediate value of the gray scale in video image as estimation threshold value T first
0;
B) the threshold value T starting to estimate is utilized
0the gray-scale value of image is divided into two different regions: R
1, R
2, according to formula (1.5) zoning R
1and R
2the average u of gray scale
1and u
2:
C) u is calculated
1and u
2after, calculate the threshold value T made new advances
i+1:
D) repeat step B), C), until T
i+1and T
iduring infinite approach, its value is binary-state threshold T,
B4) when difference result is greater than threshold value T, then this agllutination fruit is set to 255, otherwise is set to 0, the binaryzation of realize target,
The step that described morphologic filtering realizes is:
C1) transversal scanning is carried out to binarization segmentation result, when adjacent two white interblock gaps are less than 2 blocks, then the black patch of zone line is set to 255, otherwise retain initial value;
C2) longitudinal scanning is carried out to binarization segmentation result, when adjacent two white interblock gaps are less than 2 blocks, then the black patch of zone line is set to 255, otherwise retain initial value;
Described act of violating regulations identification, comprises vehicle location step by step and vehicle cab recognition,
The performing step of described vehicle location is:
D1) within the scope of statistical picture each capable be not 0 pixel, record its number, obtain the width of this row image,
D2) statistical picture scope Nei Gelie is not the pixel of 0, records its number, obtains the height of this row image,
According to d1) and statistics d2), the positional information of target and the Width x Height information of target can being obtained, judging whether it captures region, as then identified vehicle in surveyed area in act of violating regulations according to the positional information of target, otherwise, it is not detected;
Described vehicle cab recognition performing step is:
For detection target machine motor-car, motorcycle and pedestrian, motorcycle and pedestrian are not distinguished, determines motor vehicle and motorcycle object module, namely in image display area, the masterplate size of motor vehicle and motorcycle is set, to facilitate subsequent detection,
Motor vehicle:
| S
machine-S
machine mould| < 20 (1.8)
Motorcycle:
| S
rub-S
rub mould| < 20 (1.9)
Wherein, η is the ratio of target area Width x Height, and obtain according to many experiments for motor vehicle, motorcycle η span, S=width*height is target area area, according to itself and the comparison of stencil area size, judges type belonging to target;
Described act of violating regulations detects, and comprises and in surveyed area, steals card behavioral value to drive in the wrong direction behavioral value and motorcycle and pedestrian of motor vehicles,
Motor vehicles drive in the wrong direction behavioral value, and by following the tracks of vehicle target track, judged whether that retrograde behavior occurs according to its trajectory direction, specific implementation step is:
E1) adopt angle point track algorithm, in object run process, with change template forward minimal difference Matching pursuitalgorithm, the synchronous position found angle point and occur in every two field picture, whenever finding new target location, then upgrades masterplate data with current location information;
E2) carry out the demarcation of the correct travel direction of vehicle to every row pixel, due to the restriction of sample frequency and car speed, being set in its movement locus in adjacent three pixel coverages is even variation, and now the error of calculation is less;
E3) demarcate each row pixel travel direction respectively from the row pixel of image lower edge the 0th, for being the section travelling positive dirction away from camera direction, its row k direction vector is respectively:
For driving towards the section that camera direction is traveling positive dirction, its row k direction vector is respectively:
E4) basis is for retrograde vehicle, the angle of its trajectory direction and the correct travel direction of vehicle has certain feature, when being occurred to drive in the wrong direction by vehicle, the angle of pursuit path and the correct travel direction of calibration vehicle diatom is defined as 120 °, regulation is for every objective movement locus, when it and the track quantity of normal travel direction angle more than 120 ° is greater than the threshold value of setting time, then judge that this movement locus is as retrograde track, namely there is behavior of driving in the wrong direction in tracked vehicle;
E5)
for the vehicle of calibration vehicle diatom correctly travels vector,
for track actual motion vector, wherein,
then the angle theta of the correct travel direction of the vehicle of vehicle actual travel direction and regulation can be expressed as:
E6) in actual testing process, as θ > 120 °, then corresponding counter is added 1, when meeting condition quantity of driving in the wrong direction and being greater than setting threshold value, then can judge that current tracking vehicle drives in the wrong direction;
Described card behavioral value of stealing refers to, for the target being defined as motorcycle and pedestrian in monitor video image detection region, then directly to report to the police to it, records its license board information;
For the automotive vehicle determined, if it has retrograde behavior, think that it has card behavior steathily, then it is reported to the police.
2. Self-service card sender act of violating regulations automatic testing method as claimed in claim 1, is characterized in that, before described Target Segmentation, act of violating regulations identification and act of violating regulations detect, also comprises pre-treatment step,
Described pre-service refers to the high resolving power due to monitor video, for reducing operand, carries out horizontal and longitudinal sample process to the image collected.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510941325.4A CN105574502A (en) | 2015-12-15 | 2015-12-15 | Automatic detection method for violation behaviors of self-service card sender |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510941325.4A CN105574502A (en) | 2015-12-15 | 2015-12-15 | Automatic detection method for violation behaviors of self-service card sender |
Publications (1)
Publication Number | Publication Date |
---|---|
CN105574502A true CN105574502A (en) | 2016-05-11 |
Family
ID=55884610
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510941325.4A Pending CN105574502A (en) | 2015-12-15 | 2015-12-15 | Automatic detection method for violation behaviors of self-service card sender |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105574502A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106204596A (en) * | 2016-07-14 | 2016-12-07 | 南京航空航天大学 | A kind of panchromatic wave-band remote sensing image cloud detection method of optic estimated with fuzzy hybrid based on Gauss curve fitting function |
CN107133610A (en) * | 2017-06-01 | 2017-09-05 | 电子科技大学 | A kind of descending vehicle flowrate vision-based detection of complex road condition and method of counting |
CN110348332A (en) * | 2019-06-24 | 2019-10-18 | 长沙理工大学 | The inhuman multiple target real-time track extracting method of machine under a kind of traffic video scene |
CN110517506A (en) * | 2019-08-26 | 2019-11-29 | 重庆同济同枥信息技术有限公司 | Method, apparatus and storage medium based on traffic video image detection Parking |
CN110929676A (en) * | 2019-12-04 | 2020-03-27 | 浙江工业大学 | Deep learning-based real-time detection method for illegal turning around |
CN111611901A (en) * | 2020-05-15 | 2020-09-01 | 北京百度网讯科技有限公司 | Vehicle reverse running detection method, device, equipment and storage medium |
CN113177509A (en) * | 2021-05-19 | 2021-07-27 | 浙江大华技术股份有限公司 | Method and device for recognizing backing behavior |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20040257224A1 (en) * | 2003-06-17 | 2004-12-23 | United Security Applications Id, Inc. | Electronic security system for monitoring and recording activity and data relating to institutions and clients thereof |
CN104463903A (en) * | 2014-06-24 | 2015-03-25 | 中海网络科技股份有限公司 | Pedestrian image real-time detection method based on target behavior analysis |
-
2015
- 2015-12-15 CN CN201510941325.4A patent/CN105574502A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20040257224A1 (en) * | 2003-06-17 | 2004-12-23 | United Security Applications Id, Inc. | Electronic security system for monitoring and recording activity and data relating to institutions and clients thereof |
CN104463903A (en) * | 2014-06-24 | 2015-03-25 | 中海网络科技股份有限公司 | Pedestrian image real-time detection method based on target behavior analysis |
Non-Patent Citations (1)
Title |
---|
汪文斌 等: "基于视频的自助发卡车道异常行为检测", 《上海船舶运输科学研究所学报》 * |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106204596A (en) * | 2016-07-14 | 2016-12-07 | 南京航空航天大学 | A kind of panchromatic wave-band remote sensing image cloud detection method of optic estimated with fuzzy hybrid based on Gauss curve fitting function |
CN106204596B (en) * | 2016-07-14 | 2020-12-29 | 南京航空航天大学 | Panchromatic waveband remote sensing image cloud detection method based on Gaussian fitting function and fuzzy mixed estimation |
CN107133610A (en) * | 2017-06-01 | 2017-09-05 | 电子科技大学 | A kind of descending vehicle flowrate vision-based detection of complex road condition and method of counting |
CN107133610B (en) * | 2017-06-01 | 2020-09-01 | 电子科技大学 | Visual detection and counting method for traffic flow under complex road conditions |
CN110348332A (en) * | 2019-06-24 | 2019-10-18 | 长沙理工大学 | The inhuman multiple target real-time track extracting method of machine under a kind of traffic video scene |
CN110517506A (en) * | 2019-08-26 | 2019-11-29 | 重庆同济同枥信息技术有限公司 | Method, apparatus and storage medium based on traffic video image detection Parking |
CN110929676A (en) * | 2019-12-04 | 2020-03-27 | 浙江工业大学 | Deep learning-based real-time detection method for illegal turning around |
CN111611901A (en) * | 2020-05-15 | 2020-09-01 | 北京百度网讯科技有限公司 | Vehicle reverse running detection method, device, equipment and storage medium |
CN111611901B (en) * | 2020-05-15 | 2023-10-03 | 北京百度网讯科技有限公司 | Vehicle reverse running detection method, device, equipment and storage medium |
CN113177509A (en) * | 2021-05-19 | 2021-07-27 | 浙江大华技术股份有限公司 | Method and device for recognizing backing behavior |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105574502A (en) | Automatic detection method for violation behaviors of self-service card sender | |
CN109829403B (en) | Vehicle anti-collision early warning method and system based on deep learning | |
CN104008645B (en) | One is applicable to the prediction of urban road lane line and method for early warning | |
CN112700470B (en) | Target detection and track extraction method based on traffic video stream | |
CN106652445B (en) | A kind of road traffic accident method of discrimination and device | |
CN104537841B (en) | Unlicensed vehicle violation detection method and detection system thereof | |
CN110298300B (en) | Method for detecting vehicle illegal line pressing | |
CN108335489A (en) | Vehicle on highway behavior semantic analysis and abnormal behaviour monitoring system and method | |
CN103605967A (en) | Subway fare evasion prevention system and working method thereof based on image recognition | |
CN105632186A (en) | Method and device for detecting vehicle queue jumping behavior | |
CN104408932A (en) | Drunk driving vehicle detection system based on video monitoring | |
CN102332209A (en) | Automobile violation video monitoring method | |
CN105744232A (en) | Method for preventing power transmission line from being externally broken through video based on behaviour analysis technology | |
CN113850123A (en) | Video-based road monitoring method and device, storage medium and monitoring system | |
CN105825185A (en) | Early warning method and device against collision of vehicles | |
CN101231756A (en) | Method and apparatus for detecting moving goal shade | |
CN103021186A (en) | Vehicle monitoring method and vehicle monitoring system | |
CN1862620A (en) | Intelligent detecting prewarning method for expressway automobile running and prewaring system thereof | |
CN114781479A (en) | Traffic incident detection method and device | |
KR20090130977A (en) | System of traffic conflict decision for signalized intersections using image processing technique | |
CN101727748A (en) | Method, system and equipment for monitoring vehicles based on vehicle taillight detection | |
CN105809184A (en) | Vehicle real-time identification tracking and parking space occupancy determining method suitable for gas station | |
CN105654734A (en) | Data processing method and device for vehicle violations | |
CN108154146A (en) | A kind of car tracing method based on image identification | |
CN113033275B (en) | Vehicle lane-changing non-turn signal lamp analysis system based on deep learning |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160511 |
|
RJ01 | Rejection of invention patent application after publication |