WO2005062256A1 - Procede permettant l'extraction automatisee retroactive d'objets en mouvement, d'une image prise au moyen d'un appareil de representation d'image, systeme de traitement d'image et produit de programme informatique - Google Patents

Procede permettant l'extraction automatisee retroactive d'objets en mouvement, d'une image prise au moyen d'un appareil de representation d'image, systeme de traitement d'image et produit de programme informatique Download PDF

Info

Publication number
WO2005062256A1
WO2005062256A1 PCT/EP2004/053510 EP2004053510W WO2005062256A1 WO 2005062256 A1 WO2005062256 A1 WO 2005062256A1 EP 2004053510 W EP2004053510 W EP 2004053510W WO 2005062256 A1 WO2005062256 A1 WO 2005062256A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
images
sequence
moving objects
areas
Prior art date
Application number
PCT/EP2004/053510
Other languages
German (de)
English (en)
Inventor
Gero BÄSE
Robert Kutka
Norbert Oertel
Original Assignee
Siemens Aktiengesellschaft
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 Siemens Aktiengesellschaft filed Critical Siemens Aktiengesellschaft
Publication of WO2005062256A1 publication Critical patent/WO2005062256A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/77—Retouching; Inpainting; Scratch removal
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G06T7/215—Motion-based segmentation
    • 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
    • G06T2207/20221—Image fusion; Image merging

Definitions

  • the invention relates to a method for the automated, subsequent removal of moving objects from an image recorded with an image recording device.
  • the invention further relates to an image processing system for performing the method according to the invention and a computer program product.
  • a still or static object such as a building with an image recording device, e.g. a camera or a video cell phone
  • an image recording device e.g. a camera or a video cell phone
  • moving objects e.g. People or passing vehicles
  • the coverage complicates e.g. the further processing of the image if it is used to construct a virtual environment.
  • the user of the image-recording device has therefore until now been forced to wait until the still object to be captured is free of a covering of one or more moving objects.
  • this is difficult to do, especially in busy places or under time pressure.
  • the object of the present invention is therefore to provide a method with which moving objects can subsequently be removed from an image recorded with an image recording device in a simple and inexpensive manner. Another object is to provide an image processing system with which the method can be carried out. These objects are achieved with a method with the features of claim 1 and with an image processing system with the features of claim 15.
  • the method according to the invention for the automated, subsequent removal of moving objects from an image recorded with an image recording device initially comprises the recording of a sequence of images with at least one still or static object which is temporarily and / or partially covered by at least one moving object , Corresponding image areas of the sequence of images, each with an image value, are evaluated. The determined image values are then acted upon by a statistical function in order to find the image areas with only immovable objects from the sequence of images and to assemble them into a result image which is cleaned of the moving objects.
  • the invention proposes to use a preferably digital image recording device such as e.g. a video cell phone, a PDA (Personal Digital Assistant), a digital camera, etc., in order to automatically distinguish and filter out the moving from the unmoving objects from the plurality of available recordings.
  • a preferably digital image recording device such as e.g. a video cell phone, a PDA (Personal Digital Assistant), a digital camera, etc.
  • a digital image recording device such as e.g. a video cell phone, a PDA (Personal Digital Assistant), a digital camera, etc.
  • the image processing system has an image recording device, a memory coupled to it for storing a plurality of images and an evaluation unit which can access the data stored in the memory and which is set up to carry out the method according to the invention.
  • the evaluation of corresponding image areas with a respective image value is carried out iteratively.
  • each image of the sequence of images is divided into a number of blocks, which are preferably arranged regularly in a coordinate system. Each of these blocks corresponds to an image area.
  • an image area is selected in the first image and evaluated with the corresponding image areas of the other images of the sequence of images.
  • the image area adjacent to the image area just evaluated is selected, for example.
  • an image area corresponds to a pixel or pixel.
  • the statistics function uses a median operator known from the statistics.
  • the median operator is executed area by area. For example, are aj, j, ⁇ , ..., a ⁇ , j, n the values of the image areas at position i, j in the coordinate system for a total of n images and are ei, -, the values of the pixels at position i, j for the result image to be calculated, the result image is calculated as follows:
  • the statistical function uses a histogram operator instead of the median operator.
  • the histogram operator specifies the most common image value "of respective corresponding image areas. In this case, the result image is calculated accordingly
  • a frequency distribution of the image values of corresponding image areas is expediently created, the image value with the most frequent occurrence being selected as the result value.
  • Either the brightness of an image area or one or more color values of an image area can expediently be evaluated as the image value and thus as a criterion for the decision and, in the case of several color values, superimposed.
  • the image noise is expediently carried out as a further method step for each image in the sequence of images in order to improve the accuracy of the method according to the invention. Due to the image noise that is common in digital image recording devices, it is unlikely that two image values are exactly the same, so that evaluation using the statistical function is made more difficult. It is therefore advisable to reduce the image noise before loading the image values with the statistical function. According to an embodiment of the invention, this can be done by means of low-pass filtering. Alternatively, the quantization of the image areas is conceivable.
  • the sequence of images is preferably recorded from the same perspective. However, if the perspective does not exactly match, then after the sequence of images has been recorded, matching image sections of the sequence of images are first determined, so that corresponding image regions from different images come to lie in the same position in a coordinate system. So-called block matching methods are known for this method step, which detects and compensates for the displacement of the image background, which comprises the static or stationary objects. After this method step, a sequence of images is present, with corresponding image areas from the different images of the sequence of images coming to rest in the same position in a coordinate system.
  • the method according to the invention can already be carried out when corresponding image areas of the static or stationary object are visible in only two images.
  • the sequence of images thus comprises at least three images.
  • the number of images in the sequence of images is preferably selected as a function of the moving objects in front of the still object or the still objects. In other words, the more moving objects temporarily and / or partially cover the still object, the greater the number of images in the sequence of images must be in order to achieve a reliable result.
  • FIGS. 2a to 2c image values assigned to the respective images, which are subjected to a further statistical evaluation
  • 3 shows the result image determined with the method according to the invention from the sequence of images
  • Figure 4 shows an image processing system
  • FIG. 1 shows a sequence of images that were taken in succession from the same perspective at different times.
  • a static object 2 in the form of three houses arranged side by side can be seen in each of the figures la to lc.
  • a motor vehicle which represents a moving object 3, moves from the left edge of the image in the direction of the right edge of the image.
  • the number n of the sequence of images is arbitrary, but must not be less than 3.
  • the number of the sequence of images is essentially determined by the number of moving objects that temporarily and / or partially obscure a static object.
  • the figures la to lc show the ideal case in which the sequence of n images is recorded from the same perspective. If this is not possible for certain reasons, matching image sections must first be found, for example by means of a block matching method for displacement or disparity estimation.
  • the block matching methods detect the displacement of the static objects, or more generally the image background, in order to compensate for the displacement. After carrying out these methods, which are well known from the prior art, image sections of corresponding size are available. If, as can be seen in FIGS. 1 a to 1 c, a grid with image areas of the same size is placed over a respective image, then corresponding image areas 5 of the sequence of images are located at the same position in the coordinate system.
  • the size of an image area is preferably one image point (pixel), since the method thereby produces the greatest accuracy.
  • the image value can represent the brightness of an image area.
  • one or more color values of an image area can be evaluated as the image value and, in the case of several color values, superimposed.
  • FIGS. 2a to 2c Such an evaluation with an image value at position j along the x-axis of respective images is shown in FIGS. 2a to 2c.
  • the image value curves shown are only chosen as an example for the explanation. It is assumed here that the image value of an image with only static objects has an approximately uniform course. Due to the vehicle moving from left to right, the image recording device records a lower brightness value. The movement of the vehicle or generally of the moving object can therefore be traced in the image value curve by the pot-shaped drop running from left to right.
  • Such image value profiles are determined not only at point j, but over the entire y-axis, so that the result would be an area in an x-y coordinate system instead of a curve.
  • the image values a ⁇ , j of each image area 6a, 6b, ... 6n of the sequence of images are compared with one another. This comparison takes place for each of the image areas 5 of the sequence of images.
  • the image values a - ,, 1, ... a- ,, n are sorted in size and the image value with the middle index is used as the result value.
  • the method according to the invention thus generates a complete image of the static object from various individual images.
  • the method according to the invention can be used, for example, in mobile imaging terminals, such as, for example, video cell phones, PDAs, digital cameras and the like.
  • the method enables hidden objects to be hidden. objects, such as people or vehicles, in front of the desired static object.
  • augmented reality is a combination of real and artificial objects.
  • To create the virtual worlds it is necessary to take panoramas of natural surroundings, which then serve as background images for them. People or other moving objects in the natural environment can be easily removed or filtered out.
  • FIG. 4 shows in schematic form an image processing system 10.
  • This has an image recording device, such as e.g. a camera, which is connected to a memory 12.
  • the memory 12 which can be of any type, serves to store the sequence of images which have been recorded with the image recording device 11.
  • An evaluation unit 13, which can read out the images stored in the memory 12, is also connected to the memory 12.
  • the evaluation unit is set up to carry out the method described above and gives the result e.g. the result image adjusted for the moving objects is returned to a monitor or the memory 12.
  • the image-recording device and evaluation unit 13 do not necessarily have to be different devices, as shown in FIG. 4. Rather, the evaluation unit 13 can also be contained in the image recording device 11 itself, as can the memory 12.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)

Abstract

La présente invention concerne un procédé permettant l'extraction automatisée rétroactive d'objets en mouvement (3), d'une image prise au moyen d'un appareil de représentation d'image. Le procédé de l'invention comprend les étapes suivantes: prise d'une suite d'images (1a, 1b, , 1n) comprenant au moins un objet statique (2) qui peut être recouvert par intermittence et/ou partiellement par au moins un objet en mouvement (3); évaluation des zones d'images correspondantes (6a, 6b, , 6n) de la suite d'images (1a, 1b, , 1n) avec au moins une valeur d'image respective (a); et application aux valeurs d'image déterminées (a) d'une fonction statistique afin de mettre en évidence dans la suite d'images, les zones d'image comprenant des objets exclusivement statiques (2), et de les rassembler pour composer une image dans laquelle les objets en mouvement ont été supprimés.
PCT/EP2004/053510 2003-12-19 2004-12-15 Procede permettant l'extraction automatisee retroactive d'objets en mouvement, d'une image prise au moyen d'un appareil de representation d'image, systeme de traitement d'image et produit de programme informatique WO2005062256A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
DE10359928.2 2003-12-19
DE10359928A DE10359928A1 (de) 2003-12-19 2003-12-19 Verfahren zum automatisierten, nachträglichen Entfernen von bewegten Objekten aus einem mit einem Bildaufzeichenden Gerät aufgenommenen Bild, Bildverarbeitungssystem und Computerprogrammprodukt

Publications (1)

Publication Number Publication Date
WO2005062256A1 true WO2005062256A1 (fr) 2005-07-07

Family

ID=34706367

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/EP2004/053510 WO2005062256A1 (fr) 2003-12-19 2004-12-15 Procede permettant l'extraction automatisee retroactive d'objets en mouvement, d'une image prise au moyen d'un appareil de representation d'image, systeme de traitement d'image et produit de programme informatique

Country Status (2)

Country Link
DE (1) DE10359928A1 (fr)
WO (1) WO2005062256A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2011014901A3 (fr) * 2009-08-06 2011-04-07 Florian Matusek Procédé d'analyse vidéo

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105933604A (zh) * 2016-05-20 2016-09-07 珠海市魅族科技有限公司 一种图像获取的方法及终端

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
HASE H ET AL: "Real-time snowfall noise elimination", IMAGE PROCESSING, 1999. ICIP 99. PROCEEDINGS. 1999 INTERNATIONAL CONFERENCE ON KOBE, JAPAN 24-28 OCT. 1999, PISCATAWAY, NJ, USA,IEEE, US, vol. 2, 24 October 1999 (1999-10-24), pages 406 - 409, XP010368992, ISBN: 0-7803-5467-2 *
HSU C-T ET AL: "Mosaics of video sequences with moving objects", SIGNAL PROCESSING. IMAGE COMMUNICATION, ELSEVIER SCIENCE PUBLISHERS, AMSTERDAM, NL, vol. 19, no. 1, January 2004 (2004-01-01), pages 81 - 98, XP004476840, ISSN: 0923-5965, Retrieved from the Internet <URL:doi:10.1016/j.image.2003.10.001> [retrieved on 20050217] *
YAMASHITA A ET AL: "A virtual wiper - Restoration of deteriorated images by using multiple cameras", PROCEEDINGS OF THE 2003 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS. (IROS 2003). LAS VEGAS, NV, OCT. 27 - 31, 2003, IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS, NEW YORK, NY : IEEE, US, vol. VOL. 4 OF 4, 27 October 2003 (2003-10-27), pages 3126 - 3131, XP010671183, ISBN: 0-7803-7860-1 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2011014901A3 (fr) * 2009-08-06 2011-04-07 Florian Matusek Procédé d'analyse vidéo

Also Published As

Publication number Publication date
DE10359928A1 (de) 2005-07-28

Similar Documents

Publication Publication Date Title
DE69628282T2 (de) Verfahren zur kompression mehrerer videobilder
DE2937282C2 (fr)
DE19815861A1 (de) Verfahren und Vorrichtung für eine auf Blöcken basierende Bildverarbeitung
DE19531004C2 (de) Verfahren und Vorrichtung zur wahrnehmungsoptimierten Übertragung von Video- und Audio-Daten
DE102007034010A1 (de) Verfahren und Vorrichtung zur Bearbeitung von Videodaten
DE3603552C2 (fr)
EP0897247A2 (fr) Procédé de calcul de vecteurs de mouvement
WO2005062256A1 (fr) Procede permettant l&#39;extraction automatisee retroactive d&#39;objets en mouvement, d&#39;une image prise au moyen d&#39;un appareil de representation d&#39;image, systeme de traitement d&#39;image et produit de programme informatique
EP1279291B1 (fr) Procede et dispositif de mise en memoire et de traitement d&#39;informations d&#39;images successives dans le temps
DE69833836T2 (de) Verfahren und Gerät zur massstäblichen Umformung von Bildern
EP1729258A2 (fr) Procédé et dispositif destinés à la détermination de vecteurs de déplacement
DE102020109789A1 (de) Verfahren zum Durchführen einer Selbstlokalisierung eines Fahrzeugs auf der Grundlage einer reduzierten digitalen Umgebungskarte, Computerprogrammprodukt sowie ein Selbstlokalisierungssystem
WO2019057473A1 (fr) Procédé et dispositif d&#39;évaluation d&#39;images, procédé d&#39;aide à la commande et dispositif de commande
EP0614555B1 (fr) Procede de segmentation d&#39;une zone continue a bord diffus dans une image-ecran monochrome
DE19938845A1 (de) Vidiobild-Überwachungsverfahren
DE10301572B4 (de) Verfahren zum Komprimieren einer Folge von Bildern
EP0496051B1 (fr) Méthode pour estimer des vecteurs de mouvement en vue de la transmission d&#39;information d&#39;images digitales
EP0786186A1 (fr) Codage par transformation orthogonale bidimensionnelle pouvant s&#39;adapter aux segments, pour un systeme de television numerique
DE3917085C2 (de) Schaltungsanordnung zur Filterung eines Videosignals
WO2020187695A1 (fr) Réduction de matériel vidéo aux portions de mouvement
DE60221075T2 (de) Verfahren zur Sequenzierung von Videoinformationen
DE10238576B4 (de) Bildbearbeitungsverfahren
DE102011016094A1 (de) Verfahren zur Anpassung von Bildern mit unterschiedlicher Aufnahme- und Belichtungsdynamik
WO1999017551A1 (fr) Procede et dispositif de memorisation d&#39;au moins une image numerique
DE69927774T2 (de) Verfahren und Vorrichtung zur Verarbeitung eines digitalen Bildes

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BW BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE EG ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NA NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): BW GH GM KE LS MW MZ NA SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LT LU MC NL PL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG

121 Ep: the epo has been informed by wipo that ep was designated in this application
NENP Non-entry into the national phase

Ref country code: DE

WWW Wipo information: withdrawn in national office

Country of ref document: DE

122 Ep: pct application non-entry in european phase