中科微点-全球专利检索

  1. Login
  2. Sign Up

Search Results

20 records (took 0.025 seconds)

    PROCEDE DE DETERMINATION D'UNE DIRECTION D'UN OBJET A PARTIR D'UNE IMAGE DE L'OBJET
    11.
    发明公开
    PROCEDE DE DETERMINATION D'UNE DIRECTION D'UN OBJET A PARTIR D'UNE IMAGE DE L'OBJET 审中-公开
    从物体图像确定物体方向的方法

    公开(公告)号：EP3278301A1

    公开(公告)日：2018-02-07

    申请号：EP16716504.2

    申请日：2016-03-30

    Applicant: Thales

    Inventor： SIMON, Alain

    IPC: G06T7/00

    CPC classification number: G06T7/74 ,  G06T7/80 ,  G06T2207/10016 ,  G06T2207/10032 ,  G06T2207/10048 ,  G06T2207/20081 ,  G06T2207/30244 ,  G06T2207/30252 ,  H04N5/2258 ,  H04N5/23238

    Abstract: The invention relates to a method of determining the absolute direction of an object of a scene (1), with a predetermined desired performance. It comprises a learning phase and an on-line operation phase, the learning phase comprising the following steps: - acquisition by circular scanning by means of a first optronic imaging device of determined fixed position, of a series of partially overlapping optronic images (2), including an image or several images of the scene (step A1), - automatic extraction from the images, of descriptors defined by their image coordinates and their radiometric characteristics, with at least one descriptor of unknown direction in each overlap (21) of images (step B1), - on the basis of the descriptors extracted from the overlaps between images, automatic estimation of the relative rotation of the images and mapping of the descriptors extracted from the overlaps (step C1), - identification in the images, of at least one known geographical reference direction (22) of precision compatible with the desired performance, and determination of the image coordinates of each reference (step D1), - on the basis of the descriptors extracted from the overlaps and mapped, of the direction and image coordinates of each reference, automatic estimation of the attitude of each image, termed the fine registration step (step E1), - on the basis of the attitude of each image, of the position and of internal parameters of the first imaging device, and of the image coordinates of each descriptor, calculation of the absolute directions of the descriptors according to a predetermined model of picture capture of the imaging device (step F1), - the on-line operation phase comprising the following steps: - acquisition of at least one image of the object termed current image (20), on the basis of a second imaging device of determined fixed position (step A2), - extraction of the descriptors from each current image (step B2), - mapping of the descriptors of each current image with the descriptors whose absolute direction was calculated during the learning phase, so as to determine the absolute direction of the descriptors of each current image (step C2), - on the basis of the absolute directions of the descriptors of each current image, estimation of the attitude of each current image (step D2), - on the basis of the image coordinates of the object in each current image, of the attitude of each current image, of the position and of predetermined internal parameters of the second imaging device, calculation of the absolute direction of the object according to a predetermined model of picture capture of each current image (step E2).

    PROCEDE DE CALIBRAGE D'UN INSTRUMENT DE MESURE D'UN SYSTEME OPTRONIQUE
    13.
    发明公开
    PROCEDE DE CALIBRAGE D'UN INSTRUMENT DE MESURE D'UN SYSTEME OPTRONIQUE 有权
    方法CALIBRATING光电系统的一个测量仪器

    公开(公告)号：EP2513664A1

    公开(公告)日：2012-10-24

    申请号：EP10792894.7

    申请日：2010-12-13

    Applicant: Thales

    Inventor： SIMON, Alain

    IPC: G01S5/16 ,  G01S11/12

    CPC classification number: G01S5/163 ,  G01S11/12

    Abstract: The invention relates to a method for calibrating measurement instruments of a moving optronic system occupying positions P
    1 , P
    2 ,..., P
    i ,..., said optronic system comprising: a device (10) for acquiring images of a scene including a fixed object G
    0 ; means (15) for tracking the fixed object G
    0 during the acquisition of the images; means (20) for obtaining positions P
    1 , P
    2 ,...; and at least one distance measuring instrument (25) and/or an instrument (30) for measuring angles of orientation and/or attitude between the measurement instrument and the fixed object G
    0 , along a line of sight (LdV). The invention includes the following steps: acquisition of at least two images at times t
    1 , t
    2 ,..., whereby each image is acquired from different system positions P
    1 , P
    2 ,... and the fixed object G
    0 is targeted in each image, but the position thereof is unknown; acquisition of distance and/or angle measurements at times t'
    1 , t'
    2 ,...; synchronisation of distance and/or angle measurements with positions P
    1 , P
    2 ,... established at times t
    1 , t
    2 ,...; and estimation of measurement errors that minimise the dispersion of at least two points of intersection G
    ij between the line of sight (LdV) at position P
    i and the line of sight (LdV) at position P
    j, as a function of said measurements and known positions, P
    i , P
    j , of the system.

    PROCEDE DE CLASSIFICATION D'OBJETS DANS UN SYSTEME DE VEILLE PAR IMAGERIE
    15.
    发明公开
    PROCEDE DE CLASSIFICATION D'OBJETS DANS UN SYSTEME DE VEILLE PAR IMAGERIE 有权
    方法OBJECKTKLASSIFIZIERUNG在一个图像观察系统

    公开(公告)号：EP2517152A1

    公开(公告)日：2012-10-31

    申请号：EP10798992.3

    申请日：2010-12-08

    Applicant: Thales

    Inventor： FOURNIER, Jérôme ,  SIMON, Alain ,  PAYOT, Etienne ,  HENAFF, Gilles

    IPC: G06K9/62

    CPC classification number: G06K9/6217

    Abstract: The invention relates to a method for developing a knowledge base of object images obtained by an imaging device provided with a sensor, which includes a step of defining N classes each comprising a set of objects represented by a label and a set of features, and a step of learning decision-making rules associated with said classes. Said method further includes the following steps: defining K items of contextual image background information f
    k , with k varying from 1 to K and K>1; associating one of said pieces of image background information f
    k with each object; dividing the objects into M new classes, with N k and the set of features of said objects, defining, for each background f
    k , a subset Q
    k of said M classes associated with said image background, and learning decision-making rules for each of said subsets Q
    k .

Patent Agency Ranking