EP2356627A1 - Procédé de détection d'un processus d'arrêt de véhicules - Google Patents

Procédé de détection d'un processus d'arrêt de véhicules

Info

Publication number
EP2356627A1
EP2356627A1 EP09774608A EP09774608A EP2356627A1 EP 2356627 A1 EP2356627 A1 EP 2356627A1 EP 09774608 A EP09774608 A EP 09774608A EP 09774608 A EP09774608 A EP 09774608A EP 2356627 A1 EP2356627 A1 EP 2356627A1
Authority
EP
European Patent Office
Prior art keywords
vehicle
image
passage
vehicles
images
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.)
Withdrawn
Application number
EP09774608A
Other languages
German (de)
English (en)
Inventor
Dieter Schmidradler
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.)
Smartspector Artificial Perception Engineering GmbH
Original Assignee
Smartspector Artificial Perception Engineering GmbH
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 Smartspector Artificial Perception Engineering GmbH filed Critical Smartspector Artificial Perception Engineering GmbH
Publication of EP2356627A1 publication Critical patent/EP2356627A1/fr
Withdrawn legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • G06V20/54—Surveillance or monitoring of activities, e.g. for recognising suspicious objects of traffic, e.g. cars on the road, trains or boats
    • 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
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30236—Traffic on road, railway or crossing

Definitions

  • the invention relates to a method for detecting the stopping of vehicles.
  • the invention relates to a device for detecting the stopping of vehicles, comprising a camera which cyclically detects images of passing vehicles obliquely to the direction of travel from an angle of view, as well as an evaluation unit which takes over these images.
  • Holding Bid Compliance with the Holding Bid is an essential basis for avoiding accidents and dangerous situations in traffic.
  • the specification of the holding bid is usually either static, e.g. by means of a stop table and a stop line or time-dependent, e.g. by means of a traffic light state "red".
  • Automated methods are based on known methods of speed measurement (typically radar, laser), wherein the undershooting of a measured minimum speed is rated as "stopping".
  • imaging measurement methods which recognize vehicles by means of known methods of image sequence analysis as foreground objects and derived movement trajectories from these foreground objects.
  • image sequence analysis For the automatic evaluation of the traffic flow today also imaging measurement methods are used, which recognize vehicles by means of known methods of image sequence analysis as foreground objects and derived movement trajectories from these foreground objects.
  • shadowing by neighboring Vehicles Through changing light and shadow conditions, shadowing by neighboring Vehicles, adverse weather conditions and ambiguities in the image often result in such imaging measuring methods to significant losses in measurement accuracy and tracking errors.
  • the process must be documented on the basis of at least one suitable camera and the vehicle registration number (hereinafter referred to as "number plate") of the vehicle must be recorded.
  • Another known system avoids the technical disadvantages of tracking general vehicle contours described above by using evaluation results of a license plate reading to carry out a passage recognition on the basis of a cycle-related localization and the tracking of the license plate over the entire observation area (WO 2007107875A2).
  • the said system is based on stereoscopic measuring methods and requires appropriate calibration procedures and sufficiently rigid mechanical couplings between at least two cameras.
  • a predominant part of today's known camera systems for the automatic reading of license plates is only used to determine the identity of vehicles based on their license plate.
  • both foreign-triggered systems which receive a control signal for image acquisition and / or evaluation of an external detector (eg camera in speed monitoring), and free-running, cyclic measuring camera systems that automatically recognize a passage using known methods of image processing and thus ensure that only one passenger message with the evaluation result of the license plate reading is generated for a passage of a vehicle.
  • This object is achieved by a method mentioned above in that according to the invention, a camera from an angle obliquely to the direction of cyclic captured images of passing vehicles, an evaluation takes over these images, this evaluation unit in the images structural elements that are the number plate associated, within observes the position of at least one uniquely located structural element of the license plate in the image of a defined observation area and detects the stopping of the vehicle on the basis of the presence of accumulation values of the at least one structural element at at least one position in the observation area in successive measurement cycles.
  • the evaluation unit is set up to segment structural elements belonging to the license plate into the images, furthermore within a defined observation area across cycles the position of at least one uniquely located structural element of the license plate in the image and to detect the stopping of the vehicle on the basis of the presence of accumulation values of the at least one structural element at at least one position in the observation area in successive measuring cycles.
  • the viewing area can either cover the entire field of view of the camera or as a partial area, e.g. be defined by a trigger line or as part of a trigger line, by a circumscribing rectangle or by a polygon within the field of view of the camera.
  • images are recorded with a suitably selected cycle duration and not the general vehicle contours, but precisely those unique structural elements of the registration mark are tracked, which makes possible a precise and clearly traceable determination of the vehicle position.
  • the developed method is characterized in that in a obliquely to the optical axis of the camera passing vehicle with registration marks by a vanishing difference segmented elements of the plate over two or more cycles, the stopping of a vehicle is determined beyond doubt.
  • a non-stop can be determined beyond doubt if the position of the segmented elements has shifted noticeably with each measurement cycle.
  • results are made anonymous by making found marks in the result image hidden or made anonymous (eg unrecognizable);
  • a spatial position of the license plate position is determined from known real dimensions of structural elements of the license plate and known optical imaging ratios of the camera, wherein in particular an average speed of two different measurement cycles of a passage is determined by the change of the determined spatial position with respect to the time difference of the two Measuring cycles is set; It may be advantageous if a cycle-resolved speed curve is calculated by calculating the average speeds successive measuring cycles is determined;
  • the evaluation unit for documenting the temporal course of motion for individual vehicle passages generates a time-coded overlay image, wherein in particular the time-coded overlay image, an intensity or color-coded time axis is generated with the recording times, and wherein after complete passage based on the actual number of measuring cycles, a dynamic spread the intensity assignment of the time-coded sub-picture image is made in order to achieve a high-contrast representation;
  • the evaluation unit for documenting the temporal course of motion for individual vehicle passages for displaying a clustering each generates a summation image in which prior to the first detection of a new vehicle, the entire image area is initialized with the value 0 and with each measurement cycle those pixels of the summation image increments, in the in the current measuring cycle is not disappearing segmentation result, wherein in particular after complete passage on the basis of the actually occurring maximum accumulation values, a dynamic spread of the intensity assignment of the overlay image is made in order to achieve a high-contrast representation;
  • the passage is treated in relation to at least one external signal, it being possible for the time characteristic of the external signal to be assigned via a time diagram to the pictorial passport documentation;
  • FIG. 1a shows an example of a measuring arrangement in a perspective view
  • FIG. 1b shows an example of the measuring arrangement in a view from above
  • FIG. 2 shows an example of the cross-cycle tracking of unique ones
  • FIGS. 3a-3d show different velocity profiles in relation to image acquisition and analysis
  • FIG. 4 compares the real license plate geometry with the distorted license plate image.
  • FIGS. 5a-5f make different based on a simple segmentation result
  • FIGS. 6a-6d show the visualization possibilities on a real example
  • FIGS. 7a-7d show the visualization of movement processes including the stopping process with reference to two examples.
  • Figures 8a-8b show an implementation example for the consideration of an external state for the temporal evaluation of a passage situation.
  • FIG. 1 shows an example of an advantageous sensor arrangement in two views: a camera (1) detects the vehicle front, including license plate (3) within a defined measuring range (4), from a passing vehicle (2). These images are processed by an evaluation unit (5) integrated in the camera or connected to the camera.
  • FIG. 2 shows the detection of a characteristic (3) at discrete measurement times (9). Precisely because of the uniqueness of the mark as a whole, it is possible to determine trajectories (10) for suitably selected contour features of the registration mark very precisely over the entire observation period.
  • the velocity component is calculated normal to the optical axis
  • V q (t n) h I dpjxel ⁇ pjxel I (t n -t n .i).
  • the symbol height can either be known a priori or can be derived from geometrical conditions at the location (for example nominal measurement distance of a mark to a defined trigger line).
  • FIGS. 3a-3d show possible real speed profiles v (t) (12) and detections of a passage, as well as discrete-time determined velocities normal to the optical axis v q (t n ) (13). While FIGS. 3 a, 3 b and 3 c show a measurement with a sufficient sampling rate, in FIG. 3 d a sampling rate that is too low is set, which in the assumed case leads to a misdetection of the measuring system.
  • the measuring system detects a steadily reducing velocity v q (t n ), which corresponds to a monotonous decrease in speed qualitatively with the real speed decrease.
  • v q (t n ) is thus a correct lower bound for the real mean passage velocity in the time interval [t n , t n . ⁇ ⁇ .
  • a stop of the vehicle apparently does not take place throughout the passage.
  • the measuring system correctly interprets a vehicle passage without stopping.
  • the vehicle stops, the vehicle / license plate is located at an identical position at the sampling times t 2 and t 3 , so that the stopping of the vehicle can be determined correctly on the basis of the speed component v q (t 3 ).
  • the structural elements of the plate Due to the matching position of the vehicle at successive sampling times, the structural elements of the plate also occur repeatedly on successive measuring cycles at a constant image position.
  • the standstill of the vehicle is thus synonymous with the occurrence of a characteristic maximum of the frequency distribution of the structural elements of the plate at the holding position in the picture.
  • An accumulation value assigned to the structure element at a specific image position thus characterizes the stopping of the vehicle, the accumulation value corresponding to the number of measurement cycles in which the vehicle stops.
  • Figure 3c shows an example of a passage in which the path difference between successive measuring cycles is so small that the measuring system detects a stop of the vehicle Thresholds for the positional tolerance of a stationary object due to the spatial quantization and possible vibrations at the measuring location, as well as due to the selected sampling rate, a lower limit speed v % m ⁇ n can be specified, from which a stopping process can be detected despite not disappearing speed.
  • the speed v (f) of a vehicle is at least a time-continuous physical quantity with a correspondingly limited positive and negative rate of change.
  • a time interval [f SfO p, t stop + i] can be defined within which an acceleration from standstill at time t stop results in only a minimal position shift In any case, the measuring system would still be interpreted as a "standstill".
  • FIG. 3d shows a case example of a sampling rate that is too low. After the sampling cycle at time t 2 , the vehicle reduces its speed to a stop and accelerates before the next sampling at time t 3 . It can be seen from the timing diagram that the measuring system erroneously detects a violation in this situation.
  • the velocity component v (t) along the optical axis can be determined by determining the change in distance between camera and license plate in addition to a velocity component normal to the optical axis . If we first consider distortions and location-dependent aberrations of the optics, we can use the height of the symbol in the image area, hp Xe ⁇ t), (11), and a subsuming global optical correction factor k opt to determine a time-dependent distance d (t) between the camera and characteristics are calculated:
  • V ⁇ (t n ) k O pf (H / hp iX ⁇ 1 (t n ) -H / hpi X ⁇ 1 (t n . 1 ))
  • a position C (t n ) (17) of the code can be obtained from the object coordinate system (u'.v'.w 1 ) are converted by means of coordinate transformation into a time-independent camera coordinate system:
  • Vehicle characteristic the average passage speed can be derived: v m ⁇ t m , t k )
  • the velocity vector results directly from the positional shifts of the characteristic position C (t) on which the evaluation is based.
  • a quasicontinuous velocity profile for the entire process can be determined on the basis of the considerations previously made with regard to the required sampling rate and mass inertia Derive observation area.
  • the location of the license plate position in the camera coordinate system now also opens up the possibility of establishing a spatial context for further vehicle front object areas, in particular in conjunction with the speed vector detectable across the cycle. This also means that other vehicle features on the front of the vehicle are interpreted spatially, and thus dimensions of the vehicle can be determined.
  • FIG. 5a defines an image coordinate system to illustrate advantageous visualization techniques.
  • FIG. 4b shows a segmented symbol in the image coordinate system at different acquisition times.
  • FIG. 5d shows the principle of a time-coded representation.
  • all pixels of the result image are initialized with O.
  • the intensity values n + ⁇ are assigned to all segmented pixels in the result image. This makes it possible to present the sequence of movement based on the segmentation result in a temporal context. For zones of lower passage speed, segmentation results of the same structural element are close to each other, so that the intensity changes rapidly over the location in the result image. At standstill, the segmentation results of the same structural element are substantially congruent, the intensity profile changes abruptly.
  • the time profile can be determined on the basis of the actually occurring intensity values l.sub.Org (x, y), ie on the basis of the number of measurement cycles for the passage, via a suitable assignment rule within the representable intensity range (17 ) in the result image / res (x, y).
  • the well-known principle of false-color representation can subsequently be applied to the result image l res .
  • FIG. 5e shows a representation in frequency-coded form. If the symbols overlap at a correspondingly low passage speed, such an overlay image gives a particularly clear impression of speed changes and stopping.
  • the entire result image is initialized with O values before the start of detection. Subsequently, in each measurement cycle, all pixels are incremented to currently segmented regions in the result image. In particular, in those pictures, therefore, those panel positions in which the vehicle has been stopped are clearly highlighted.
  • the frequency-coded image according to FIG. 5f can be assigned on the basis of the actually occurring accumulation values l or g (x, y) via a suitable assignment rule within the representable intensity range in the result image / res (x > y).
  • the well-known principle of false-color representation can subsequently be applied to the result image l res . Since the time-coded visualization as coding of the trajectories, and the frequency coding with the clear emphasis of stationary segmentation areas provides essential impressions for a subjective interpretation, a combined application of different visualization techniques is particularly advantageous.
  • FIG. 6a-6d show a combined visualization of photorealistic representation and the proposed overlay images on a real example.
  • FIG. 6a shows a frequency coding with non-overlapping segmentation regions.
  • a legend (18) represents the relation between frequency p (x, y) and intensity value or false color.
  • FIG. 6b shows the same segmentation regions in time-coded representation.
  • the time information is assigned to the image in a legend (19).
  • Figure 6c shows an overall picture illustrating some of the particular advantages of the present invention:
  • a text box (20) relevant data for the passage is documented.
  • a photorealistic image (21) with the insertion of the defined trigger line (22) establishes a clear relationship between the position of the mark and the position of a real stop line.
  • a color mapping table (24) shows the association between intensity values and representation in image areas (21) and (23).
  • the overlay images (25) and (26) are highlighted in the real composition image.
  • FIG. 6d shows an inverse representation of the overall image of FIG. 6c that is more suitable for printing.
  • Figures 7a-7d show two passages where the vehicle stops within the observation area.
  • FIGS. 7a and 7b show the same passage in normal and inverse representation, wherein the vehicle stops exactly once in the measuring range (27).
  • a further advantage of the invention according to FIGS. 7c and 7d is also that both the method and the image material derivable therefrom allow a direct interpretation of multiple speed changes and stopping processes (28) as well as a determination of time-dependent directions of travel.
  • the measuring system is to have a time-dependent holding requirement, e.g. At controlled intersections, the signal indicative of the hold in the measurement range can be scanned either optically by means of another camera or by means of a photodetector, or electrically within the signal system, and correlated with observed passages. If there is currently no hold bid, detected passages can be ignored while those passages detected during the upright hold bid are documented.
  • the timing of the stop signal provided can be compared directly with the time coding of the license plate. This way results for the viewer again, in a single picture, a clear local-temporal reference for the assessment of a transgression.
  • Figures 8a and 8b show the compilation of the vehicle passage in the context of an external signal by way of example.
  • a rectangle (29) the photo-realistic image (21) is set in a clear temporal relation to the passage.
  • the crossing of the stop line can be set and documented in relation to the time at which the holding request occurs.
  • the detailed passport documentation together with the provision of a speed profile can in particular also be used to determine whether a vehicle driver has committed the transgression with constant, ascending or descending speed. This is a very important indicator of whether road users have not or too late perceived the holding requirement due to distraction, or whether a breach of the Holding Bid was deliberately committed.
  • the proposed measuring system is to function, for example, as the data source of a dynamic traffic control, it may be advantageous not to forward individual detections, but to interpret a plurality of passages together. Accordingly, it may be useful, for example, to generate a traffic jam message only if a repeated stop of passing vehicles was detected within a predetermined observation period.
  • the measurement result can be displayed, for example, in the text area (20) of the result image or output as a separate result message.

Abstract

L'invention concerne un procédé de détection d'un processus d'arrêt de véhicules. Selon l'invention, une caméra prend cycliquement des images de véhicules qui passent, en biais par rapport à la direction de déplacement. Une unité d'analyse reçoit ces images, y isole des éléments structurels qui sont associés au numéro minéralogique et observe dans l'image, dans un domaine d'observation prédéterminé, au-delà des cycles, la position d'au moins un élément structurel de la plaque minéralogique qui est localisé de manière univoque. Sur la base de l'accumulation dudit au moins un élément structurel dans au moins une position dans le domaine d'observation au cours de cycles successifs, l'arrêt du véhicule est détecté.
EP09774608A 2008-11-14 2009-11-12 Procédé de détection d'un processus d'arrêt de véhicules Withdrawn EP2356627A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
AT0177308A AT507457B1 (de) 2008-11-14 2008-11-14 Verfahren zur automatischen feststellung des anhaltens von kraftfahrzeugen
PCT/AT2009/000432 WO2010054417A1 (fr) 2008-11-14 2009-11-12 Procédé de détection d'un processus d'arrêt de véhicules

Publications (1)

Publication Number Publication Date
EP2356627A1 true EP2356627A1 (fr) 2011-08-17

Family

ID=42008547

Family Applications (1)

Application Number Title Priority Date Filing Date
EP09774608A Withdrawn EP2356627A1 (fr) 2008-11-14 2009-11-12 Procédé de détection d'un processus d'arrêt de véhicules

Country Status (3)

Country Link
EP (1) EP2356627A1 (fr)
AT (1) AT507457B1 (fr)
WO (1) WO2010054417A1 (fr)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2889842A1 (fr) * 2013-12-31 2015-07-01 Patents Factory Ltd. Sp. z o.o. Procédé permettant d'estimer la dynamique de mouvement dans une image vidéo
CN111178224B (zh) * 2019-12-25 2024-04-05 浙江大华技术股份有限公司 物体规则判断方法、装置、计算机设备和存储介质

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2854359B2 (ja) * 1990-01-24 1999-02-03 富士通株式会社 画像処理システム
JP3435623B2 (ja) * 1996-05-15 2003-08-11 株式会社日立製作所 交通流監視装置
JP4800455B2 (ja) * 1999-02-19 2011-10-26 富士通株式会社 車速計測方法および装置
KR100459476B1 (ko) * 2002-04-04 2004-12-03 엘지산전 주식회사 차량의 대기 길이 측정 장치 및 방법
ITTO20060214A1 (it) 2006-03-22 2007-09-23 Kria S R L Sistema di rilevamento di veicoli

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See references of WO2010054417A1 *

Also Published As

Publication number Publication date
WO2010054417A1 (fr) 2010-05-20
AT507457B1 (de) 2011-01-15
AT507457A1 (de) 2010-05-15

Similar Documents

Publication Publication Date Title
EP1192597B1 (fr) Procede pour detecter des objets dans le champ peripherique d'un vehicule routier jusqu'a une grande distance
EP1825276B1 (fr) Procede et dispositif pour determiner la vitesse d'un vehicule
WO2013053438A2 (fr) Procédé d'intégration d'objets virtuels dans des affichages de véhicule
WO2011110204A1 (fr) Dispositif d'assistance au conducteur à représentation optique d'objets détectés
DE102007001649A1 (de) Verfahren, Vorrichtung und Computerprogramm zur Selbstkalibrierung einer Überwachungskamera
DE102009005553A1 (de) Verfahren und Vorrichtung zur Erkennung der Umgebung eines Kraftfahrzeugs
DE102013210836A1 (de) Erkennen eines bewegungslosen Gegenstands auf Kreuzungen von Verkehrswegen
EP2249580A1 (fr) Procédé de calibrage de l'image d'une caméra
DE102015122172A1 (de) Scheinwerferbasierte Projetion von Mustern zur Vermessung räumlicher Eigenschaften einer Fahrzeugumgebung
WO2015032399A1 (fr) Procédé et dispositif de détection d'objets d'après des données d'image ayant une résolution de profondeur
DE102009022278A1 (de) Verfahren zur Ermittlung eines hindernisfreien Raums
DE102012023060A1 (de) Verfahren zum Detektieren eines beweglichen Objekts mithilfe eines Histogramms anhand von Bildern einer Kamera und Kamerasystem für ein Kraftfahrzeug
DE102013104425B3 (de) Verfahren zur Erfassung von Geschwindigkeitsverstößen mit restriktiver Datenspeicherung
DE102010042026A1 (de) Verfahren und Vorrichtung zum Erzeugen eines Abbildes mindestens eines Objekts in einer Umgebung eines Fahrzeugs
DE102011075674A1 (de) Abstandsbestimmung mittels eines Kamerasensors
DE102007025147A1 (de) System zur Spurverlassenswarnung und/oder Spurhaltefunktion
AT507457B1 (de) Verfahren zur automatischen feststellung des anhaltens von kraftfahrzeugen
DE102007041333B4 (de) Berührungslos arbeitendes Zählsystem
DE102009052868A1 (de) Verfahren zum Betrieb eines Fahrassistenzsystems eines Fahrzeugs
WO2016096446A1 (fr) Différenciation entre premier plan et arrière-plan dans un véhicule
WO2011137973A2 (fr) Procédé de détection de l'environnement d'un véhicule
DE102016118538A1 (de) Verfahren zum Klassifizieren eines Verkehrszeichens in einem Umgebungsbereich eines Kraftfahrzeugs, Rechenvorrichtung, Fahrerassistenzsystem sowie Kraftfahrzeug
DE102015226749A1 (de) Überwachungsanlage
DE102018107212A1 (de) Verfahren zum Erfassen eines Verkehrsbereichs-Belegungserkennungs-Hilfselements in einem Umgebungsbereich eines Kraftfahrzeugs, Computerprogrammprodukt sowie Fahrzeugführungssystem
EP2355007A2 (fr) Système et procédé destinés à la mesure d'objet tridimensionnelle

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

17P Request for examination filed

Effective date: 20110604

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO SE SI SK SM TR

DAX Request for extension of the european patent (deleted)
STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION IS DEEMED TO BE WITHDRAWN

18D Application deemed to be withdrawn

Effective date: 20130601