CN105893962A - Method for counting passenger flow at airport security check counter - Google Patents

Method for counting passenger flow at airport security check counter Download PDF

Info

Publication number
CN105893962A
CN105893962A CN201610196768.XA CN201610196768A CN105893962A CN 105893962 A CN105893962 A CN 105893962A CN 201610196768 A CN201610196768 A CN 201610196768A CN 105893962 A CN105893962 A CN 105893962A
Authority
CN
China
Prior art keywords
connected domain
people
image
video
pedestrian
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
Application number
CN201610196768.XA
Other languages
Chinese (zh)
Inventor
郭德全
朱家新
兰时勇
潘卫军
梁海军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Sichuan University
Chengdu University of Information Technology
Civil Aviation Flight University of China
Original Assignee
Sichuan University
Chengdu University of Information Technology
Civil Aviation Flight University of China
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Sichuan University, Chengdu University of Information Technology, Civil Aviation Flight University of China filed Critical Sichuan University
Priority to CN201610196768.XA priority Critical patent/CN105893962A/en
Publication of CN105893962A publication Critical patent/CN105893962A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • 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
    • 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
    • G06T2207/10024—Color image
    • 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
    • G06T2207/20—Special algorithmic details
    • G06T2207/20212—Image combination
    • G06T2207/20224—Image subtraction
    • 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
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30196—Human being; Person
    • 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
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30242—Counting objects in image
    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C11/00—Arrangements, systems or apparatus for checking, e.g. the occurrence of a condition, not provided for elsewhere
    • G07C2011/04—Arrangements, systems or apparatus for checking, e.g. the occurrence of a condition, not provided for elsewhere related to queuing systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Image Analysis (AREA)
  • Closed-Circuit Television Systems (AREA)

Abstract

The invention discloses a method for counting passenger flow at an airport security check counter. The method comprises steps of: converting an image in a video sequence into a grayscale image, extracting a foreground image by a frame difference method and binaryzing the foreground image; extracting a head part from the foreground image by using a head grayscale characteristic and connected domain detection algorithm; in a head tracking and counting algorithm, tracking and counting head characteristics by using the specificity of the head connected domain in videos and a container; and counting the number of people by using the tracking of the head characteristics. The method may automatically count the passenger flow at a place, and can be used at other fixed places similar to the airports and with dense people flow, such as railway stations, bus stations, and tourist resorts.

Description

A kind of airport security bayonet socket stream of people's statistical method
Technical field
The invention belongs to control and field of instrumentation technology, be specifically related to the design of a kind of airport security bayonet socket stream of people's statistical method.
Background technology
Due to the effect that the airport national economy in various countries is the most important with performance in productive life, the safety on airport and operation Efficiency is by bigger attention.Video monitoring, as the means of a kind of " visible ", can get information about airport real-time situation, Become the requisite safety guarantee in each size airport.In the extensively application of airport video monitoring system, utilize computer graphic As technology automatically processes information in video monitoring, reduce artificial participation, be video monitoring and the direction of data analysis future development. Demographics has in actual life and is widely applied very much, such as, large stadium passenger flow quantitative statistics, as railway station, airport and Bus station etc.；The guest flow statistics of public bus network；Supermarket, museum, the statistics etc. of some public place numbers such as exhibition center Deng.These statistics aid decision making person can preferably carry out MRP and management, the most at the train station or airport can root According to the open ticketing of the data of different periods guest flow statistics or the quantity of ticket-checking pass way；Traffic department can be according to the visitor of public bus network Traffic statistics are planned public bus network and arrange the quantity of bus；Exhibition can be planned according to the data of the volume of the flow of passengers in public place Looking at arranging an order according to class and grade of route, mode, open hour or work service personnel etc., the monitoring that safely provides for personnel compact district simultaneously is protected Barrier.
The identification of human body target and tracking are the difficult points of current airport security bayonet socket Video processing, and between human body, morphological differences is big, and Attitude is changeable, and this brings big challenge to human body target identification and tracking.At present, human body target identification and tracking in video are had Following several method: the people counting algorithm of feature based；People counting algorithm based on region；Number based on template matches Statistic algorithm.
The people counting algorithm of feature based chooses certain or some local feature of target, and depending on as human bioequivalence and tracking According to.2009, Albiol used the angle point number calculating scene as scene characteristic statistical number of person.First with Harris angle point The angle point of detector detection scene, then distinguishes Corner and background angle point by corners Matching, and Albiol thinks every two field picture Child's hair twisted in a knot-childhood count and total number of persons direct proportionality, carry out statistical number of person by this algorithm.Some scholar changes on this basis Entering, use SURF characteristic point to replace angle point to add up number, in conjunction with SURF characteristic point, shooting distance is classified, In order to improve the precision of statistical number of person.Wherein, according to pedestrian's occupied area division in image and merging prospect, in the ideal case There is reasonable counting effect.People counting algorithm basic thought based on region is: utilize the template comprising target to enter image Row segmentation, template is usually slightly larger than the rectangle of target, it is possible to the most irregularly shaped for other；Then in sequence image, Use related algorithm follow the tracks of target, gray level image can be used based on texture relevant with feature, to coloured image utilization based on Being correlated with of color.Current research person uses the marginal information of target that target is tracked counting, utilizes shape and the color of the number of people People's head region is detected by information, and it comprises two steps: the extraction of black region and shape analysis, by image HSV V passage pixel placement threshold value in space obtains black region, and this method can resist illumination variation and shadow effect.Part is learned Person proposes one and utilizes local feature track algorithm, by multiscale analysis, Target Segmentation becomes multiple region, each region by One agglomerate represents, each agglomerate contains the mean value of all pixels, shape and the position of respective regions, passes through blob match Follow the tracks of target.People counting algorithm based on template matches is that the image sequence of input is converted into a kind of static in shape pattern, During identification with prestore edition comparison before.
Summary of the invention
The invention aims to solve in the existing stream of people's statistical technique human body target identification and the accuracy rate of tracking and real-time relatively The problem of difference, it is proposed that a kind of airport security bayonet socket stream of people's statistical method.
The technical scheme is that a kind of airport security bayonet socket stream of people's statistical method, comprise the following steps:
S1, by being arranged on vertical camera collection stream of people's video at bayonet socket top；
S2, the stream of people's video collected in step S1 is pre-processed；
S3, carry out foreground extraction to step S2 pre-processes the video obtained, obtain sport foreground image；
S4, the sport foreground image obtained in step S3 is carried out binary conversion treatment；
S5, the image after binary conversion treatment in step S4 is carried out morphology processing；
S6, employing method based on connected domain detection carry out number of people feature identification and screening to the image after processing in step S5；
S7, utilize number of people connected domain particularity in video and container that number of people feature is tracked counting, by special to the number of people The tracking levied realizes demographics.
Further, step S2 particularly as follows:
Judge whether the stream of people's video collected in step S1 is coloured image, if being then translated into gray level image and adjustment regards Frequently size, the most directly adjusts video size.
Further, step S3 particularly as follows:
For step S2 pre-processes the video obtained, process a frame every two frames, use frame difference method that its prospect is extracted, Directly the intercropping difference to present frame Yu former frame, obtains sport foreground image.
Further, step S5 specifically include following step by step:
S51, expansion: add pixel to the border of the objects in images after binary conversion treatment, when image is carried out expansive working, Output pixel value is the maximum of all pixels in the neighborhood of pixels corresponding to input picture；
S52, corrosion: fritter pixel discrete in image after deletion expansion process, when image is carried out etching operation, export picture The minimum of a value of all pixels in element value neighborhood of pixels corresponding to input picture.
Further, step S6 particularly as follows:
Image after processing in step S5 is carried out connected domain detection based on lower edges Point matching, calculates the area C of connected domains With Aspect Ratio Ct, it is desirable to Cs> C, Ca≤Ct≤Cb, wherein C is the minimum of a value of the number of people part connected domain area specified, Ca And CbIt is respectively minimum of a value and maximum, C, C of nominator's head region Aspect Ratioa、CbValue all in accordance with actual conditions adjust； According to this condition, reject and do not meet the connected domain that size and shape requires, and for the connected domain of detection, return to initial Image carries out matches color information, when the colouring information of connected domain corresponding region meets the colouring information of head zone, is considered as What this connected domain represented is exactly the number of people feature of pedestrian.
Further, step S7 specifically include following step by step:
S71, read the video of a certain frame；
S72, judging whether the foreground image of this frame video has connected domain, if then entering step S73, otherwise entering step S74；
S73, judging whether the foreground image of this frame video has untreated connected domain, if then entering step S75, otherwise entering Enter step S74；
S74, the pedestrian information updated in container S, return step S71；
S75, judge that whether pedestrian corresponding to this connected domain be the pedestrian of newly entering monitored area, if then entering step S76, no Then enter step S77；
S76, corresponding for this connected domain pedestrian information being added and be saved in container S, total number of persons counting adds 1 simultaneously, returns step S73；
S77, pedestrian information corresponding for this connected domain is updated to current pedestrian information；
Whether S78, the response intensity judging this connected domain are 1, if then entering step S79, otherwise return step S73；
S79, judge that pedestrian's velocity corresponding to this connected domain, whether more than 0, if the counting that then pedestrian goes out adds 1, returns Step S73；Otherwise pedestrian enters counting and adds 1, returns step S73.
The invention has the beneficial effects as follows:
(1) employing is arranged on vertical camera collection stream of people's video at bayonet socket top, and when pedestrian is close to each other, the number of people is substantially without sending out Life is blocked, and accuracy of data acquisition is high.
(2) amount of calculation is little, can meet the requirement of real-time statistics number.
(3) affected little by illumination variation and shade.
(4) under condition of doing what is required of social etiquette, it is not necessary to carry out image segmentation more.
(5) applicability is wide, can be used in several scenes.
Accompanying drawing explanation
A kind of airport security bayonet socket stream of people's statistical method flow chart that Fig. 1 provides for the present invention.
Fig. 2 is the source video image of the embodiment of the present invention.
Fig. 3 is the foreground image of the embodiment of the present invention.
Fig. 4 is the prospect binary image of the embodiment of the present invention.
Fig. 5 is the foreground image after the morphology processing of the embodiment of the present invention.
Fig. 6 is the number of people connected component labeling image of the embodiment of the present invention.
Fig. 7 is the flow chart step by step of step S7 of the present invention.
Detailed description of the invention
Below in conjunction with the accompanying drawings embodiments of the invention are further described.
The invention provides a kind of airport security bayonet socket stream of people's statistical method, as it is shown in figure 1, comprise the following steps:
S1, by being arranged on vertical camera collection stream of people's video at bayonet socket top.
Find by observing substantial amounts of video data, when using vertical camera collection stream of people's video, number of people when pedestrian is close to each other Substantially without blocking.Therefore by number of people feature in foreground image is distinguished and extracts, can be substantially accurately to pedestrian Count.
S2, the stream of people's video collected in step S1 is pre-processed.
Judge whether the stream of people's video collected in step S1 is coloured image, if as in figure 2 it is shown, being then translated into gray scale Image also adjusts video size, the most directly adjusts video size.So can reduce amount of calculation, quickening processes the speed of video.
S3, carry out foreground extraction to step S2 pre-processes the video obtained, obtain sport foreground image.
For step S2 pre-processes the video obtained, process a frame every two frames, use frame difference method that its prospect is extracted, Directly the intercropping difference to present frame Yu former frame, obtains sport foreground image, as shown in Figure 3.Direct differential is represented by:
D (x, y)=| a (x, y)-b (x, y) | (1)
Wherein, a (x, y) and b (x, y) be current frame image respectively and previous frame image be positioned at coordinate (x, y) grey scale pixel value at place, d (x, y) It it is sport foreground image gray value of pixel at corresponding coordinate.
S4, the sport foreground image obtained in step S3 is carried out binary conversion treatment.
There is a lot of noise jamming in the sport foreground image obtained in step S3.At this moment, sport foreground image is carried out binaryzation Process, filter noise, as shown in Figure 4, obtain motion candidates region.Binary conversion treatment is expressed as:
g ( x , y ) = { 0 , d ( x , y ) < B 255 , d ( x , y ) &GreaterEqual; B - - - ( 2 )
Wherein, (x, being y) sport foreground image, (x, y) grey scale pixel value at place, (x y) is the binary picture of relevant position to g to d at coordinate As value, value is " 0 " or " 255 ", and B is threshold value, uses varimax to be calculated.
S5, the image after binary conversion treatment in step S4 is carried out morphology processing.
Based on OpenCV, the image after binary conversion treatment in step S4 is carried out morphology processing, reduce further noise and Increase the feature of number of people candidate region.This step specifically include following step by step:
S51, expansion: add pixel to the border of the objects in images after binary conversion treatment, when image is carried out expansive working, Output pixel value is the maximum of all pixels in the neighborhood of pixels corresponding to input picture.With structural element b to input picture f Carry out gray scale and expand mark, be denoted asIt is defined as:
f &CirclePlus; b = max { f ( s - x , t - y ) + b ( x , y ) | ( s - x , t - y ) &Element; D f , ( x , y ) &Element; D b } - - - ( 3 )
Wherein DfAnd DbBeing respectively the definition territory of f and b, s represents gray level, and t is threshold value.Dilatometer is at last from structural element b The field determined is tried to achieveMaximum.By this method, the number of people features of foreground image after binaryzation can be strengthened Point.
S52, corrosion: fritter pixel discrete in image after deletion expansion process, when image is carried out etching operation, export picture The minimum of a value of all pixels in element value neighborhood of pixels corresponding to input picture.With structural element b, input picture f is carried out ash Degree etching mark, is denoted asIt is defined as:
f &CircleTimes; b = min { f ( s - x , t - y ) + b ( x , y ) | ( s - x , t - y ) &Element; D f , ( x , y ) &Element; D b } - - - ( 4 )
Wherein DfAnd DbBeing respectively the definition territory of f and b, s represents gray level, and t is threshold value.Corrosometer is at last from structural element b The neighborhood determined is tried to achieveMinimum of a value.By this method, the noise spot of some sudden changes can be filtered.
Carry out the image effect after morphology processing as shown in Figure 5.
S6, employing method based on connected domain detection carry out number of people feature identification and screening to the image after processing in step S5.
Under the experimental situation of vertical camera, the head of people has the shape of approximate circle, but there is also between the head of different people Difference.Image after processing in step S5 is carried out connected domain detection based on lower edges Point matching, calculates the area of connected domain CsWith Aspect Ratio Ct, it is desirable to Cs> C, Ca≤Ct≤Cb, wherein C is the minimum of a value of the number of people part connected domain area specified, CaAnd CbIt is respectively minimum of a value and maximum, C, C of nominator's head region Aspect Ratioa、CbValue all in accordance with actual conditions adjust Whole；According to this condition, reject and do not meet the connected domain that size and shape requires, and for the connected domain of detection, return to Initial pictures carries out matches color information, when the colouring information of connected domain corresponding region meets the colouring information of head zone, just Think that what this connected domain represented is exactly the number of people feature of pedestrian.Detect and screen the image after connected domain as shown in Figure 6.
S7, utilize number of people connected domain particularity in video and container that number of people feature is tracked counting, by special to the number of people The tracking levied realizes demographics.
The sport foreground image of each frame is carried out connected domain detection, detects that several connected domain i.e. indicates that several pedestrian is currently The picture of frame occurs.In the embodiment of the present invention, identify the pedestrian position at present frame with the coordinate of connected domain central point, adopt Pedestrian T: the frame number T of video when this connected domain being detected uniquely is identified by the several information below connected domainf, in connected domain The coordinate T of heart pointp, it is consecutively detected the frame number T of connected domain corresponding to this pedestriann(in order to judge response intensity), the speed of pedestrian Degree vector Tv.The information of each pedestrian is saved in container S, before and after processing every two field picture, to row all in container The information of people updates dynamically.
As it is shown in fig. 7, this step specifically include following step by step:
S71, read the video of a certain frame, when the foreground image at a certain frame detects connected domain, first its information is carried out Initialize, wherein Tf, TpAll can be by the information acquisition of connected domain, Tn=0, Tv=0.
S72, judging whether the foreground image of this frame video has connected domain, if then entering step S73, otherwise entering step S74。
S73, judging whether the foreground image of this frame video has untreated connected domain, if then entering step S75, otherwise entering Enter step S74.
S74, the pedestrian information updated in container S, return step S71.
Due to circumstance complication in reality, pedestrian, when by video surveillance region, is not that every frame can be detected.Processing After complete each two field picture, can process container S does not has the element of more fresh information, represent that the pedestrian that this element is corresponding is " real Border " occur in video and be detected.According to the frame number S [n] of video when nth elements detects connected domain in container Sf(hold The frame number of video when nth elements detects connected domain in device S) judge S [n] element the most consistent with the frame number of current video be No it is updated.
Assuming that the information of S [n] is not updated at present frame, it can be done following renewal and process:
Wherein, S [n]f、S[n]p、S[n]n、S[n]vRepresent respectively and update when connected domain being detected of nth elements in front container S The frame number of video, the coordinate of connected domain central point, it is consecutively detected the frame number of connected domain corresponding to this element, velocity；S[n]f ′、 S[n]p′、S[n]n′、S[n]v' represent respectively after updating nth elements in container S the frame number of video when connected domain being detected, The coordinate of connected domain central point, it is consecutively detected the frame number of connected domain corresponding to this element, velocity.
When renewal position detects, after being to ensure that, the connected domain having detected pedestrian more timely, new pedestrian will not be mistaken for, Reduce error.After having updated information, if S [n]n=0, represent that the pedestrian that this connected domain represents has walked out video surveillance region, then Delete the information of this pedestrian from container.
The present invention uses the method processed once every two frames to process video, on the one hand accelerates the processing speed of video, The error produced when on the other hand reducing counting, the particularly counting to turnover pedestrian.
S75, judge that whether pedestrian corresponding to this connected domain be the pedestrian of newly entering monitored area, if then entering step S76, no Then enter step S77.
The principle judged is as follows, if the element number that container is now is n, then:
A = min { ( T p x - S &lsqb; m &rsqb; p x ) 2 + ( T p y - S &lsqb; m &rsqb; p y ) 2 | m &Element; &lsqb; 0 , n ) } , n > 0 D + 100 , n = 0 - - - ( 6 )
Wherein Tpx、TpyRepresent abscissa and the ordinate of currently detected connected domain central point, S [m]px、S[m]pyRepresent container The abscissa of m-th element center point and ordinate in S, D represents that two connected domains of judgement are the minimum of a values representing two pedestrians.
S76, as A > D time, represent that pedestrian corresponding to this connected domain is newly entering monitored area personnel, by corresponding for this connected domain Pedestrian information interpolation is saved in container S, and total number of persons counting adds 1 simultaneously, returns step S73.
S77, pedestrian information corresponding for this connected domain is updated to current pedestrian information.
When A < during D, represents that pedestrian corresponding to this connected domain is the most being detected and is counting, generally, in container only There is an element S [m] to make A < D, now need the information that information updating is current connected domain of S [m].The most updated Journey is as follows:
Wherein, S [m]p、S[m]nRepresent respectively and update the coordinate of connected domain central point of m-th element, company in front container S Continue the frame number of the connected domain detecting that this element is corresponding；S[m]f′、S[m]p′、S[m]n′、S[m]v' represent container after renewal respectively The frame number of video, the coordinate of connected domain central point when connected domain being detected of m-th element in S, it is consecutively detected this element pair The frame number of the connected domain answered, velocity.
S78, judge the response intensity S [m] of this connected domainnWhether being 1, if then entering step S79, otherwise returning step S73；
S79, judge pedestrian's velocity corresponding to this connected domain whether more than 0, if then representing that pedestrian goes out, corresponding pedestrian The counting gone out adds 1, returns step S73；Otherwise represent that pedestrian enters, then pedestrian enters counting and adds 1, returns step S73.
Those of ordinary skill in the art is it will be appreciated that embodiment described here is to aid in the former of the reader understanding present invention Reason, it should be understood that protection scope of the present invention is not limited to such special statement and embodiment.The ordinary skill of this area Personnel can according to these technology disclosed by the invention enlightenment make various other various concrete deformation without departing from essence of the present invention and Combination, these deformation and combination are the most within the scope of the present invention.

Claims (6)

1. airport security bayonet socket stream of people's statistical method, it is characterised in that comprise the following steps:
S1, by being arranged on vertical camera collection stream of people's video at bayonet socket top；
S2, the stream of people's video collected in step S1 is pre-processed；
S3, carry out foreground extraction to step S2 pre-processes the video obtained, obtain sport foreground image；
S4, the sport foreground image obtained in step S3 is carried out binary conversion treatment；
S5, the image after binary conversion treatment in step S4 is carried out morphology processing；
S6, employing method based on connected domain detection carry out number of people feature identification and screening to the image after processing in step S5；
S7, number of people connected domain particularity in video and container is utilized number of people feature to be tracked counting, by the number of people The tracking of feature realizes demographics.
Airport security bayonet socket stream of people's statistical method the most according to claim 1, it is characterised in that described step S2 has Body is:
Judge whether the stream of people's video collected in step S1 is coloured image, if being then translated into gray level image and adjusting Video size, the most directly adjusts video size.
Airport security bayonet socket stream of people's statistical method the most according to claim 1, it is characterised in that described step S3 has Body is:
For step S2 pre-processes the video obtained, process a frame every two frames, use frame difference method that its prospect is extracted, Directly the intercropping difference to present frame Yu former frame, obtains sport foreground image.
Airport security bayonet socket stream of people's statistical method the most according to claim 1, it is characterised in that described step S5 has Body include following step by step:
S51, expansion: add pixel to the border of the objects in images after binary conversion treatment, when image is carried out expansive working, Output pixel value is the maximum of all pixels in the neighborhood of pixels corresponding to input picture；
S52, corrosion: fritter pixel discrete in image after deletion expansion process, when image is carried out etching operation, output Pixel value is the minimum of a value of all pixels in the neighborhood of pixels corresponding to input picture.
Airport security bayonet socket stream of people's statistical method the most according to claim 1, it is characterised in that described step S6 has Body is:
Image after processing in step S5 is carried out connected domain detection based on lower edges Point matching, calculates the area of connected domain CsWith Aspect Ratio Ct, it is desirable to Cs> C, Ca≤Ct≤Cb, wherein C is the minimum of a value of the number of people part connected domain area specified, CaAnd CbIt is respectively minimum of a value and maximum, C, C of nominator's head region Aspect Ratioa、CbValue all in accordance with actual conditions Adjust；According to this condition, reject and do not meet the connected domain that size and shape requires, and for the connected domain of detection, return Return to initial pictures and carry out matches color information, when the colouring information of connected domain corresponding region meets the colouring information of head zone Time, be considered as the representative of this connected domain is exactly the number of people feature of pedestrian.
Airport security bayonet socket stream of people's statistical method the most according to claim 1, it is characterised in that described step S7 has Body include following step by step:
S71, read the video of a certain frame；
S72, judging whether the foreground image of this frame video has connected domain, if then entering step S73, otherwise entering step S74；
S73, judge whether the foreground image of this frame video has untreated connected domain, if then entering step S75, otherwise Enter step S74；
S74, the pedestrian information updated in container S, return step S71；
S75, judge that whether pedestrian corresponding to this connected domain be the pedestrian of newly entering monitored area, if then entering step S76, Otherwise enter step S77；
S76, corresponding for this connected domain pedestrian information being added and be saved in container S, total number of persons counting adds 1 simultaneously, returns step Rapid S73；
S77, pedestrian information corresponding for this connected domain is updated to current pedestrian information；
Whether S78, the response intensity judging this connected domain are 1, if then entering step S79, otherwise return step S73；
S79, judge that pedestrian's velocity corresponding to this connected domain, whether more than 0, if the counting that then pedestrian goes out adds 1, is returned Return step S73；The counting that otherwise pedestrian enters adds 1, returns step S73.
CN201610196768.XA 2016-03-31 2016-03-31 Method for counting passenger flow at airport security check counter Pending CN105893962A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610196768.XA CN105893962A (en) 2016-03-31 2016-03-31 Method for counting passenger flow at airport security check counter

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610196768.XA CN105893962A (en) 2016-03-31 2016-03-31 Method for counting passenger flow at airport security check counter

Publications (1)

Publication Number Publication Date
CN105893962A true CN105893962A (en) 2016-08-24

Family

ID=57011946

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610196768.XA Pending CN105893962A (en) 2016-03-31 2016-03-31 Method for counting passenger flow at airport security check counter

Country Status (1)

Country Link
CN (1) CN105893962A (en)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106485217A (en) * 2016-09-30 2017-03-08 四川世纪云道科技有限公司 A kind of method and system of identification visit scenic spot stream of people's saturation degree
CN106846837A (en) * 2017-03-27 2017-06-13 广州大学 A kind of traffic light intelligent control system, traffic lights intelligent control method and device
EP3306573A1 (en) * 2016-10-05 2018-04-11 Wirelesswerx International, Inc. Flow control in a defined location
CN108428326A (en) * 2018-03-14 2018-08-21 海南师范大学 A kind of eco-tour assessment management precaution information system
CN108596377A (en) * 2018-04-16 2018-09-28 世纪美映影院技术服务（北京）有限公司 A kind of intelligent processing method for cinema projection operation management
CN109241851A (en) * 2018-08-07 2019-01-18 电子科技大学 A kind of the Oil Leakage Detecting method and its system of view-based access control model image
US10288737B2 (en) 2017-09-19 2019-05-14 Wirelesswerx International, Inc. LiDAR sensing system
CN111339873A (en) * 2020-02-18 2020-06-26 南京甄视智能科技有限公司 Passenger flow statistical method and device, storage medium and computing equipment
CN113269745A (en) * 2021-05-21 2021-08-17 燕山大学 Aerial photography automobile counting method based on OpenCv
CN113642362A (en) * 2020-05-11 2021-11-12 广东毓秀科技有限公司 Crowd density estimation method for intelligent escape in dense place
CN114140542A (en) * 2021-12-09 2022-03-04 成都智元汇信息技术股份有限公司 Picture cutting method based on color compensation, electronic equipment and medium

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101567097A (en) * 2009-06-05 2009-10-28 上海大学 Bus passenger flow automatic counting method based on two-way parallactic space-time diagram and system thereof
CN103049788A (en) * 2012-12-24 2013-04-17 南京航空航天大学 Computer-vision-based system and method for detecting number of pedestrians waiting to cross crosswalk

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101567097A (en) * 2009-06-05 2009-10-28 上海大学 Bus passenger flow automatic counting method based on two-way parallactic space-time diagram and system thereof
CN103049788A (en) * 2012-12-24 2013-04-17 南京航空航天大学 Computer-vision-based system and method for detecting number of pedestrians waiting to cross crosswalk

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
朱家新 等: "基于帧间特征和连通域检测的人数统计", 《计算机技术与发展》 *

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106485217A (en) * 2016-09-30 2017-03-08 四川世纪云道科技有限公司 A kind of method and system of identification visit scenic spot stream of people's saturation degree
EP3306573A1 (en) * 2016-10-05 2018-04-11 Wirelesswerx International, Inc. Flow control in a defined location
CN106846837A (en) * 2017-03-27 2017-06-13 广州大学 A kind of traffic light intelligent control system, traffic lights intelligent control method and device
US10288737B2 (en) 2017-09-19 2019-05-14 Wirelesswerx International, Inc. LiDAR sensing system
CN108428326A (en) * 2018-03-14 2018-08-21 海南师范大学 A kind of eco-tour assessment management precaution information system
CN108596377A (en) * 2018-04-16 2018-09-28 世纪美映影院技术服务（北京）有限公司 A kind of intelligent processing method for cinema projection operation management
CN109241851A (en) * 2018-08-07 2019-01-18 电子科技大学 A kind of the Oil Leakage Detecting method and its system of view-based access control model image
CN111339873A (en) * 2020-02-18 2020-06-26 南京甄视智能科技有限公司 Passenger flow statistical method and device, storage medium and computing equipment
CN111339873B (en) * 2020-02-18 2021-04-20 南京甄视智能科技有限公司 Passenger flow statistical method and device, storage medium and computing equipment
CN113642362A (en) * 2020-05-11 2021-11-12 广东毓秀科技有限公司 Crowd density estimation method for intelligent escape in dense place
CN113269745A (en) * 2021-05-21 2021-08-17 燕山大学 Aerial photography automobile counting method based on OpenCv
CN114140542A (en) * 2021-12-09 2022-03-04 成都智元汇信息技术股份有限公司 Picture cutting method based on color compensation, electronic equipment and medium

Similar Documents

Publication Publication Date Title
CN105893962A (en) Method for counting passenger flow at airport security check counter
CN103971380B (en) Pedestrian based on RGB-D trails detection method
CN104933710B (en) Based on the shop stream of people track intelligent analysis method under monitor video
CN104978567B (en) Vehicle checking method based on scene classification
CN102214291B (en) Method for quickly and accurately detecting and tracking human face based on video sequence
CN112183471A (en) Automatic detection method and system for standard wearing of epidemic prevention mask of field personnel
CN106204640A (en) A kind of moving object detection system and method
CN102915433B (en) Character combination-based license plate positioning and identifying method
CN107622258A (en) A kind of rapid pedestrian detection method of combination static state low-level image feature and movable information
CN106203513A (en) A kind of based on pedestrian&#39;s head and shoulder multi-target detection and the statistical method of tracking
CN104866843B (en) A kind of masked method for detecting human face towards monitor video
CN107301378A (en) The pedestrian detection method and system of Multi-classifers integrated in image
US10650249B2 (en) Method and device for counting pedestrians based on identification of head top of human body
CN109919053A (en) A kind of deep learning vehicle parking detection method based on monitor video
CN103136537B (en) Vehicle type identification method based on support vector machine
CN106296677A (en) A kind of remnant object detection method of double mask context updates based on double-background model
CN113095263B (en) Training method and device for pedestrian re-recognition model under shielding and pedestrian re-recognition method and device under shielding
CN102867188A (en) Method for detecting seat state in meeting place based on cascade structure
CN105160297A (en) Masked man event automatic detection method based on skin color characteristics
Karpagavalli et al. Estimating the density of the people and counting the number of people in a crowd environment for human safety
CN110347870A (en) The video frequency abstract generation method of view-based access control model conspicuousness detection and hierarchical clustering method
CN104851288B (en) Traffic light positioning method
CN108038432A (en) Bus pedestrian flow statistical method and system based on optical flow counting
CN103902981A (en) Method and system for identifying license plate characters based on character fusion features
CN112183472A (en) Method for detecting whether test field personnel wear work clothes or not based on improved RetinaNet

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: 20160824

RJ01 Rejection of invention patent application after publication