CN104331687B - A kind of behavioral value method of not fastening the safety belt based on vehicular video analysis - Google Patents
A kind of behavioral value method of not fastening the safety belt based on vehicular video analysis Download PDFInfo
- Publication number
- CN104331687B CN104331687B CN201410605551.0A CN201410605551A CN104331687B CN 104331687 B CN104331687 B CN 104331687B CN 201410605551 A CN201410605551 A CN 201410605551A CN 104331687 B CN104331687 B CN 104331687B
- Authority
- CN
- China
- Prior art keywords
- image
- value
- background
- safety belt
- pixel
- 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
Classifications
-
- 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
- G06V20/59—Context or environment of the image inside of a vehicle, e.g. relating to seat occupancy, driver state or inner lighting conditions
- G06V20/597—Recognising the driver's state or behaviour, e.g. attention or drowsiness
Abstract
The invention discloses a kind of behavioral value method of not fastening the safety belt based on vehicular video analysis, vehicular control unit detects vehicle movement situation in real time, it was found that after vehicle movement, then trigger Vehicular video elements capture dynamic video information, the human pilot band of position delimited after image preprocessing, lock wear safety belt target signature, start DSP video analysis programs, the behavior it was found that driver does not fasten the safety belt, starts vehicle carried module and will send information to remote supervisory central platform by remote transmission module.The situation of change of dynamic video image is analyzed and detected in real time to this method this image interframe difference algorithm, and identification and rejecting are without image sequence that is obvious or especially changing.To there is the image sequence of significantly motion or dynamic change to carry out dynamic behaviour identification and background difference identification, then by gray-level projection algorithm confirm to drive man-powered vehicle whether wear safety belt behavior, accuracy of detection is high, detection speed is fast.
Description
Technical field
The present invention relates to the analyzing and processing field of dynamic image, more particularly to it is a kind of based on vehicular video analysis
Do not fasten the safety belt behavioral value method.
Background technology
The category of video monitoring image treatment technology is very wide, in the past main research monitoring image digitlization, networking problem
Technology, improve monitoring image interconnect, information sharing, the function such as remote monitoring.
With China big and medium-sized cities quick increased security protection, technical precaution, traffic surveillance and control system, the quantity of monitoring video in recent years
In explosive increase situation, security personnel want to watch close all monitored pictures and found from the video monitoring image of magnanimity
The view data needed is hardly possible.
Traditional Safe belt detection method is many drivers by the way of vehicle electronics detection in current China market
Directly tipping safety belt in detection means " can just deceive " device, would not alarm, evade punishment.
The content of the invention
Based on vehicular video analysis it is not it is an object of the invention to overcome the deficiencies of the prior art and provide a kind of
Safety belt behavioral value method.
The present invention is achieved by the following technical solutions：A kind of behavior of not fastening the safety belt based on vehicular video analysis
Detection method, this method comprises the following steps：
(1), the real-time monitoring screen image that vehicle-mounted vidicon is arrested is obtained
(2), dynamic video image is pre-processed
(2.1) the monitoring screen image of continuous multiframe is first converted into gray level image, extracts two or three consecutive frames,
According to the relative change of present frame and the moving image and background image of adjacent front and rear frame, then based on image frame differential method,
Calculate the binary map of substantially moving target；
(2.2) binary map in step 2.1 is modified using edge detection method, then by revised binary map, drawn
It is divided into a region, if the area in some region is less than the threshold value of setting, is determined as background and behavior without significant change
Region is simultaneously abandoned, if the area in region is more than the threshold value of setting, is determined as the region of background and behavior for significant change,
It is pre- in retaining and be transferred to step (3)；
(3), do not fasten the safety belt behavioral value
An initial background is obtained, background model is built using improved gauss hybrid models, adds when there is new object
Enter the object into background or in original background to disappear, the pixel value of image is regarded as prospect Gaussian Profile and background Gauss
The mixture of distribution, if certain point pixel value of image meets prospect Gaussian Profile, the point belongs to foreground target；If image
Certain point pixel value meets background Gaussian Profile, then the point belongs to background, and carries out context update；Image sequence is averaged,
Background is updated using following equation：
Bk+1(i)=Bk(i)+(α1×Mk(i)+α2×(1-Mk(i)))(Ck(i)-Bk(i))
In formula：Bk(i) value for being pixel i in current background, Bk+1(i) it is the value of pixel i in renewal rear backdrop, Ck(i) it is
Pixel i value, α in present image1、α2To update coefficient, Mk(i) it is defined as follows：
Due to the change of noise and background, when | Ck(i)-Bk(i) | ＜ TbWhen, it is believed that current image value is background, otherwise
Current image value is prospect.
(4) behavior of not fastening the safety belt judges
Using based on gray-level projection model algorithm, detection zone is divided into different zones, each region sets special
Levy a little, safety is laterally worn with waist wears two parts with oblique above-below direction,
(4.1) by horizontal integral projection and upright projection founding mathematical models, the level integration of binary image is calculated
Projection and vertical integral projection；
Horizontal integral projection：Vertical integral projection：
In formula：N is row valid pixel, and m is row valid pixel, and x is the abscissa of pixel in the picture, and y is pixel in figure
Ordinate as in,
(4.2) setting regions characteristic value, by the horizontal integral projection and vertical integral projection that are obtained in step 4.1 and region
Characteristic value is contrasted, and determines whether the behavior of fastening the safety belt.
As the further optimization of such scheme, factor alpha is updated1=0.1, α2When=0.01, Bk+1(i)=Bk(i)+(α1
×Mk(i)+α2×(1-Mk(i)))(Ck(i)-Bk(i) it is) optimal models, background estimating effect is best.
The present invention has advantages below compared with prior art：The present invention it is a kind of based on vehicular video analysis be not peace
Full band behavioral value method, this method fusion video image intellectual analysis and fast searching techniques, by image and event description it
Between set up a kind of mapping relations, computer is quickly offered an explanation from numerous and complicated video image, identify common-denominator target object and its
Movement tendency etc..Accuracy of identification height, clear and legible recognition result, identification intelligent degree and efficiency are increased substantially, and to hand over
Logical violation enforcement provides strong technical support.
Brief description of the drawings
Fig. 1 is a kind of structural frames of behavioral value system of not fastening the safety belt based on vehicular video analysis of the present invention
Body.
Fig. 2 is a kind of Video processing of behavioral value system of not fastening the safety belt based on vehicular video analysis of the present invention
The communication connection structural representation of unit.
Fig. 3 is a kind of flow chart of behavioral value method of not fastening the safety belt based on vehicular video analysis of the present invention.
Fig. 4 is the projection result figure of specification wear safety belt.
Fig. 5 is the projection result of wear safety belt lack of standardization
Embodiment
Embodiments of the invention are elaborated below, the present embodiment is carried out lower premised on technical solution of the present invention
Implement, give detailed embodiment and specific operating process, but protection scope of the present invention is not limited to following implementations
Example.
Referring to Fig. 1 and Fig. 2, a kind of behavioral value system of not fastening the safety belt based on vehicular video analysis, including video
Processing unit and signal transmission control unit.Video processing unit includes Vehicular video capturing unit, vehicle-mounted DSP behavioural analyses list
Member and vehicular control unit.Vehicular video capturing unit is the vehicle-mounted vidicon for being placed in in-car front end, and vehicle-mounted vidicon and car
Carry the connection of DSP behavioural analyses cell signal.MV-1394 video frequency collection cards built in vehicle-mounted vidicon.
Vehicle-mounted DSP behavioural analyses unit is provided with ADC digital-to-analogue conversion interfaces, and vehicle-mounted vidicon and vehicle-mounted DSP behavioural analyses
The ADC digital-to-analogue conversions interface signal connection of unit.Vehicle-mounted DSP behavioural analyses unit is by MS320C6713 chips and its peripheral circuit
Composition.MS320C6713 chips are provided with ADC digital-to-analogue conversion interfaces, vehicle-mounted vidicon and the ADC digital-to-analogues on MS320C6713 chips
Translation interface is communicated to connect.
Vehicular control unit is connected with vehicle-mounted DSP behavioural analyses unit by serial communication port.Vehicular control unit by
MC9S12G48 chips and its peripheral circuit composition.MS320C6713 built-in chip type image analysis softwares, obtain MV-1394 videos
Whether the vehicle-mounted control information in the dynamic video information and MC9S12G48 chips of capture card, analysis calculates to derive and fastens the safety belt.
Signal transmission control unit includes vehicle-mounted GPD navigators, vehicle carried transmission unit and car alarm.It is vehicle-mounted
GPD navigators are connected with vehicle carried transmission unit serial communication, and vehicle-mounted DSP behavioural analyses unit is single by vehicle carried transmission
Member is connected with car alarm signal.A kind of behavioral value system of not fastening the safety belt based on vehicular video analysis passes through nothing
Line communicates with remote monitoring platform.
A kind of behavioral value system of not fastening the safety belt based on vehicular video analysis of the present invention, control unit is examined in real time
Measuring car motion conditions, find to trigger Vehicular video elements capture information after vehicle movement, delimit driving after image preprocessing
Personnel positions region, locks wear safety belt target signature, starts DSP video analysis programs, it is found that driver does not fasten the safety belt row
To start vehicle carried module and will send information to remote supervisory central platform by remote transmission module.
A kind of behavioral value method of not fastening the safety belt based on vehicular video analysis, referring to Fig. 3, this method includes as follows
Step：
(1), the real-time monitoring screen image that vehicle-mounted vidicon is arrested is obtained
Vehicular control unit detects vehicle movement situation in real time, when finding vehicle movement, then triggers Vehicular video unit and catch
Moving-picture information is obtained, and obtains the real-time monitoring screen image that vehicle-mounted vidicon is arrested, into step (2).
(2), dynamic video image is pre-processed
(2.1) the monitoring screen image of continuous multiframe is first converted into gray level image, extracts two or three consecutive frames,
According to the relative change of present frame and the moving image and background image of adjacent front and rear frame, then based on image frame differential method,
Calculate the binary map of substantially moving target.
Frame differential method has stronger adaptivity for dynamic environment, and processing speed is fast, with real-time.And should
Method is insensitive to homochromy object, therefore can solve shadow problem.
(2.2) binary map in step 2.1 is modified using edge detection method, then by revised binary map, drawn
It is divided into a region, if the area in some region is less than the threshold value of setting, is determined as background and behavior without significant change
Region is simultaneously abandoned, if the area in region is more than the threshold value of setting, is determined as the region of background and behavior for significant change,
It is pre- in retaining and be transferred to step (3).
(3), do not fasten the safety belt behavioral value
An initial background is obtained, background model is built using improved gauss hybrid models, adds when there is new object
Enter the object into background or in original background to disappear, the pixel value of image is regarded as prospect Gaussian Profile and background Gauss
The mixture of distribution, if certain point pixel value of image meets prospect Gaussian Profile, the point belongs to foreground target；If image
Certain point pixel value meets background Gaussian Profile, then the point belongs to background, and carries out context update.Image sequence is averaged,
Image averaging model is improved using more new formulas of the Karmann based on Kalman filtering：
Bk+1(i)=Bk(i)+(α1×Mk(i)+α2×(1-Mk(i)))(Ck(i)-Bk(i))
In formula：Bk(i) value for being pixel i in current background, Bk+1(i) it is the value of pixel i in renewal rear backdrop, Ck(i) it is
Pixel i value, α in present image1、α2To update coefficient, Mk(i) it is defined as follows：
Due to the change of noise and background, when | Ck(i)-Bk(i) | ＜ TbWhen, it is believed that current image value is background, otherwise
Current image value is prospect.
It is that background or prospect take different renewal coefficients according to current image value：α1、α2.Current image value belongs to
During background, α is used1Update, when current image value belongs to prospect, use α2Update.If considering α simultaneously1、α2Change over time, obtain
Update factor alpha1(t)、α2(t).If consideration takes different parameters in different image-regions, obtain updating factor alpha1(t,x,y)、
α2(t, x, y), wherein t are time coordinate, and x, y are the coordinate of pixel in the picture.Generally, fixed α1、α2Value,
Take α1=0.1, α2=0.01 can just obtain preferable effect.Image averaging model is simple and easy to apply, and amount of calculation is small, background estimating
Effect is good.(4) behavior of not fastening the safety belt judges
Using based on gray-level projection model algorithm, detection zone is divided into different zones, each region sets special
Levy a little, safety is laterally worn with waist wears two parts with oblique above-below direction.
(4.1) by horizontal integral projection and upright projection founding mathematical models, the level integration of binary image is calculated
Projection and vertical integral projection；
Horizontal integral projection：Vertical integral projection：
In formula：N is row valid pixel, and m is row valid pixel, and x is the abscissa of pixel in the picture, and y is pixel in figure
Ordinate as in.
(4.2) setting regions characteristic value, by the horizontal integral projection and vertical integral projection that are obtained in step 4.1 and region
Characteristic value is contrasted, and is met regional characteristic value and is then illustrated wear safety belt, does not meet regional characteristic value and just illustrate non-wear safety belt.
Floor projection integral algorithm is that the valid pixel of image toward horizontal direction is done project, meets pendant when detecting
When wearing safety belt feature, start detection zone positioning feature point, characteristic point is calculated by gray-level projection model algorithm
Coordinate simultaneously resolves to the result whether worn.Similarly, upright projection integral algorithm is to the integral projection distribution in vertical direction
Feature is detected to it, when image vertical direction valid pixel gray average changes, and this change can be vertical
Reflected in integral projection value.
From horizontal integral projection and vertical integral projection formula, floor projection is by all pixels of a line
Drawn again after gray value progress is cumulative, vertical integral projection is exactly to be shown again after the pixel gray value of vertical direction is added up.
According to horizontal and vertical gray-level projection curve, can whether specification wear safety belt.The projection knot of specification wear safety belt
Really, referring to accompanying drawing 4.The projection result of wear safety belt lack of standardization, referring to accompanying drawing 5.
A kind of behavioral value method of not fastening the safety belt based on vehicular video analysis of the present invention, passes through image frame-to-frame differences
Algorithm is divided to analyze and detect in real time the situation of change of dynamic video image, identification and rejecting are without image sequence that is obvious or especially changing
Row.To there is the image sequence of significantly motion or dynamic change to carry out dynamic behaviour identification and background difference identification, then pass through gray scale
Integral projection algorithm confirm to drive man-powered vehicle whether wear safety belt behavior, accuracy of detection is high, detection speed is fast.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
Any modifications, equivalent substitutions and improvements made within refreshing and principle etc., should be included in the scope of the protection.
Claims (2)
1. a kind of behavioral value method of not fastening the safety belt based on vehicular video analysis, it is characterised in that：This method is included such as
Lower step：
(1), the real-time monitoring screen image that vehicle-mounted vidicon is arrested is obtained
(2), dynamic video image is pre-processed
(2.1) the monitoring screen image of continuous multiframe is first converted into gray level image, extracts two or three consecutive frames, according to
The relative change of the moving image and background image of present frame and adjacent front and rear frame, then based on image frame differential method, calculate
The substantially binary map of moving target；
(2.2) binary map in step 2.1 is modified using edge detection method, then by revised binary map, be divided into
Individual region, if the area in some region is less than the threshold value of setting, is determined as the region of background and behavior without significant change
And abandon, if the area in region is more than the threshold value of setting, be determined as the region that background and behavior are significant change, in advance in
Retain and be transferred to step (3)；
(3), do not fasten the safety belt behavioral value
An initial background is obtained, background model is built using improved gauss hybrid models, is added to when there is new object
Object in background or in original background is disappeared, and the pixel value of image is regarded as prospect Gaussian Profile and background Gaussian Profile
Mixture, if certain point pixel value of image is when meeting prospect Gaussian Profile, the point belongs to foreground target；If certain point of image
Pixel value meets background Gaussian Profile, then the point belongs to background, and carries out context update using the renewal based on Kalman filtering
Formula improves image averaging model：
Bk+1(i)=Bk(i)+(α1×Mk(i)+α2×(1-Mk(i)))(Ck(i)-Bk(i))
In formula：Bk(i) value for being pixel i in current background, Bk+1(i) it is the value of pixel i in renewal rear backdrop, Ck(i) it is current
Pixel i value, α in image1、α2To update coefficient, Mk(i) it is defined as follows：
<mrow>
<msub>
<mi>M</mi>
<mi>k</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mn>1</mn>
</mtd>
<mtd>
<mrow>
<mi>i</mi>
<mi>f</mi>
<mrow>
<mo>(</mo>
<mo>|</mo>
<msub>
<mi>C</mi>
<mi>k</mi>
</msub>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
<mo>-</mo>
<msub>
<mi>B</mi>
<mi>k</mi>
</msub>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
<mo>|</mo>
<mo><</mo>
<msub>
<mi>T</mi>
<mi>b</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mn>0</mn>
</mtd>
<mtd>
<mrow>
<mi>e</mi>
<mi>l</mi>
<mi>s</mi>
<mi>e</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
</mrow>
Due to the change of noise and background, when | Ck(i)-Bk(i) | ＜ TbWhen, it is believed that current image value is background, otherwise currently
Image value be prospect；
(4) behavior of not fastening the safety belt judges
Using based on gray-level projection model algorithm, detection zone is divided into different zones, each region sets characteristic point,
Safety is laterally worn with waist wears two parts with oblique above-below direction,
(4.1) by horizontal integral projection and upright projection founding mathematical models, the horizontal integral projection of binary image is calculated
And vertical integral projection；
Horizontal integral projection：Vertical integral projection：
In formula：N be row valid pixel, m be row valid pixel, x be the abscissa of pixel in the picture, y be pixel in the picture
Ordinate,
(4.2) setting regions characteristic value, by the horizontal integral projection and vertical integral projection and provincial characteristics that are obtained in step 4.1
Value contrast, meets regional characteristic value and then illustrates wear safety belt, do not meet regional characteristic value and just illustrate non-wear safety belt.
2. a kind of behavioral value method of not fastening the safety belt based on vehicular video analysis according to claim 1, it is special
Levy and be：
Update factor alpha1=0.1, α2When=0.01,
Bk+1(i)=Bk(i)+(α1×Mk(i)+α2×(1-Mk(i)))(Ck(i)-Bk(i)) it is optimal models, background estimating effect
It is optimal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410605551.0A CN104331687B (en) | 2014-10-30 | 2014-10-30 | A kind of behavioral value method of not fastening the safety belt based on vehicular video analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410605551.0A CN104331687B (en) | 2014-10-30 | 2014-10-30 | A kind of behavioral value method of not fastening the safety belt based on vehicular video analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104331687A CN104331687A (en) | 2015-02-04 |
| CN104331687B true CN104331687B (en) | 2017-10-13 |
Family
ID=52406409
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410605551.0A Active CN104331687B (en) | 2014-10-30 | 2014-10-30 | A kind of behavioral value method of not fastening the safety belt based on vehicular video analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104331687B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104680156A (en) * | 2015-03-23 | 2015-06-03 | 山东农业大学 | System and method for identifying unfastened state of safety belt in front row of motor vehicle based on machine version |
| CN105427343A (en) * | 2015-11-18 | 2016-03-23 | 中国科学院信息工程研究所 | Moving target detecting and tracking method based on DM8148 development board |
| CN106845393A (en) * | 2017-01-19 | 2017-06-13 | 博康智能信息技术有限公司北京海淀分公司 | Safety belt identification model construction method and device |
| CN108647708A (en) * | 2018-04-28 | 2018-10-12 | 清华-伯克利深圳学院筹备办公室 | Driver evaluation's method, apparatus, equipment and storage medium |
| CN110084123A (en) * | 2019-03-28 | 2019-08-02 | 上海拍拍贷金融信息服务有限公司 | Human body behavioral value method and system, computer readable storage medium |
| CN114529729B (en) * | 2022-04-22 | 2022-08-23 | 珠海视熙科技有限公司 | Strobe detection and elimination method, device, camera and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103150556A (en) * | 2013-02-20 | 2013-06-12 | 西安理工大学 | Safety belt automatic detection method for monitoring road traffic |
| CN103522982A (en) * | 2013-10-25 | 2014-01-22 | 公安部第三研究所 | Vehicle safety belt detection method and device based on image analysis |
| CN103871052A (en) * | 2014-02-21 | 2014-06-18 | 杭州奥视图像技术有限公司 | Straight line detection-based safety belt detecting algorithm |
| CN104112141A (en) * | 2014-06-29 | 2014-10-22 | 中南大学 | Method for detecting lorry safety belt hanging state based on road monitoring equipment |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060170769A1 (en) * | 2005-01-31 | 2006-08-03 | Jianpeng Zhou | Human and object recognition in digital video |
-
2014
- 2014-10-30 CN CN201410605551.0A patent/CN104331687B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103150556A (en) * | 2013-02-20 | 2013-06-12 | 西安理工大学 | Safety belt automatic detection method for monitoring road traffic |
| CN103522982A (en) * | 2013-10-25 | 2014-01-22 | 公安部第三研究所 | Vehicle safety belt detection method and device based on image analysis |
| CN103871052A (en) * | 2014-02-21 | 2014-06-18 | 杭州奥视图像技术有限公司 | Straight line detection-based safety belt detecting algorithm |
| CN104112141A (en) * | 2014-06-29 | 2014-10-22 | 中南大学 | Method for detecting lorry safety belt hanging state based on road monitoring equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104331687A (en) | 2015-02-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104331687B (en) | A kind of behavioral value method of not fastening the safety belt based on vehicular video analysis | |
| CN104013414B (en) | A kind of Study in Driver Fatigue State Surveillance System based on intelligent movable mobile phone | |
| CN102765365B (en) | Pedestrian detection method based on machine vision and pedestrian anti-collision warning system based on machine vision | |
| CN105512720B (en) | A kind of public transit vehicle passenger flow statistics method and system | |
| CN108596129A (en) | A kind of vehicle based on intelligent video analysis technology gets over line detecting method | |
| CN103400111B (en) | Method for detecting fire accident on expressway or in tunnel based on video detection technology | |
| CN106203274A (en) | Pedestrian's real-time detecting system and method in a kind of video monitoring | |
| Dai et al. | Multi-task faster R-CNN for nighttime pedestrian detection and distance estimation | |
| CN103310444B (en) | A kind of method of the monitoring people counting based on overhead camera head | |
| CN105931467B (en) | A kind of method and device tracking target | |
| CN106204640A (en) | A kind of moving object detection system and method | |
| CN101783076A (en) | Method for quick vehicle type recognition under video monitoring mode | |
| CN104809470B (en) | A kind of vehicle based on SVM drives in the wrong direction detection device and detection method | |
| CN103617410A (en) | Highway tunnel parking detection method based on video detection technology | |
| CN110969131B (en) | Subway people flow counting method based on scene flow | |
| CN106529496A (en) | Locomotive driver real-time video fatigue detection method | |
| CN103400113B (en) | Freeway tunnel pedestrian detection method based on image procossing | |
| CN103164958B (en) | Method and system for vehicle monitoring | |
| CN103310206B (en) | A kind of vehicle using motor detection method based on many features and multiframe information fusion | |
| CN104361317B (en) | A kind of bayonet type video analysis is not fastened the safety belt behavioral value system and method | |
| CN104658272A (en) | Street traffic volume statistics and sped measurement method based on binocular stereo vision | |
| CN103927875B (en) | Based on the traffic overflow state identification method of video | |
| CN104537649A (en) | Vehicle steering judgment method and system based on image ambiguity comparison | |
| CN103794050A (en) | Real-time transport vehicle detecting and tracking method | |
| CN109508659A (en) | A kind of face identification system and method for crossing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |