• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:

    公开号:SE1230112A1
    申请号:SE1230112
    申请日:2012-11-01
    公开日:2014-05-02
    发明作者:Emanuel Johansson；Odd Larson
    申请人:Flir Systems Ab；
    IPC主号:
    专利说明:

    15 20 25 30 the folders were made. This shift from the integration time when the folders were made gives rise to a low frequency image unevenness. An extra spatial, fixed-pattern noise, sptial, fix-pattern noise, is added which is visible in low-contrast scenes. A previously known solution to this has been to only allow a change in the integration time in connection with the offset folder being updated at a so-called NUC, non uniformity correction.
    In order to function adequately, the solution requires that a NUC be made as soon as the scene has changed significantly, for example during panning.
    For examples of prior art where the integration time changes, reference can be made to WO 2008/1071 17 A1 which describes a method for changing the integration time depending on the temperature of an IR sensor.
    SUMMARY OF THE INVENTION The object of the present invention is to provide a method that dynamically adjusts the integration time based on scene content to achieve maximum performance in each scene without the need for constant folder updates through N UC processing and without storing folders for a large number of fixed integration times.
    The object of the invention is achieved by a method characterized by: a) that at least two reference images are taken into production against a black body radiator at the same temperature at two or fl different integration times. b) that when updating an offset folder is stored for which integration time the offset folder was updated, c) that during operation the current integration time is compared with the integration time for the last update of the offset folder, d) that the recorded reference image closest to the current integration time is selected, e) that a compensation per pixel for the integration time change is calculated by linear interpolation between the selected reference image and the most recently updated offset folder resulting in a compensation folder, f) adding the compensation folder to an incoming video image. 10 15 20 25 30 By utilizing reference images, the latest update of the offset folder and its integration time, as well as linear interpolation, a compensation folder can be created that closely corresponds to the current integration time. The procedure enables the integration time to be changed continuously and completely automatically without having to undergo a renewed NUC-pfOCCSS.
    Advantageously, the compensation folder is normalized before adding by subtracting the average value of the compensation folder from all pixels for centering the compensation folder around zero.
    Alternatively, the reference images can be normalized before the interpolation by subtracting the average value of the reference image from all pixels to center the reference image around zero.
    According to an advantageous method, two reference images are specifically recorded in production against a black body radiator at the same temperature at two different integration times. By holding down the number of reference images to two, the mapping process can be kept relatively simple in terms of storage and calculation.
    Suitably, according to the proposed method, the reference images can be recorded for integration times within the millisecond range with a difference in integration time of the order of 10 milliseconds.
    The invention also relates to a computer program comprising program code, which when said program code is executed in a computer causes said computer to perform methods as above for achieving the object of the invention, and a computer program product comprising a computer readable medium and a computer program as above, said computer program being included in said computer readable medium Brief Description of the Drawings The invention will be further described in the form exemplified below with reference to the accompanying drawings in which: Figure 1 schematically shows examples of a curve showing scene temperature as a function of integration time for optimized image performance.
    Figure 2 schematically shows an example in block diagram form of the principles for mapping according to the method of the invention.
    Figure 3 schematically shows an example of a curve showing signal per pixel as a function of integration time for a pixel which has undergone the mapping according to the invention method.
    Detailed design description The curve l shown in figure 1 shows an example of stage temperature T as a function of integration time t for optimized image performance. In the figure, two fixed integration times have t1 and t; inserted. According to a known method, a separate folder is linked to each integration time. In the event that another suitable integration time is used, one will benefit from the integration times for which the folders were made and thus low-frequency image irregularities occur. According to the proposed method for mapping and which is described in more detail with reference to Figure 2, compensation folders are created which offer relevant mapping along the stage temperature curve between fixed references.
    The principles of mapping according to the method of the invention are now described with reference to Figure 2.
    Through a block 2, an incoming video fate is available. The flow can be delivered from the sensor part of an IR camera, which sensor part can comprise, for example, one or more focal plane arrays. For video video, the current integration time has been denoted by tp.
    A block 3 marks the availability of two reference images taken in production. The images were taken on a blackbody radiator at the same temperature at two different integration times. The first reference image is called Breg here and is taken with an integration time of three, which in an example here can amount to 3 ms. The second reference image is called Bref; and is taken with an integration time three; which in the example here can amount to 14 ms. The reference image Brefr is marked stored in block 4, while the reference image Breg is marked stored in block 5.
    During the operation of an IR camera, calibrations are performed by the camera during a known process called NUC, non uniformity correction. A block 6 indicates this N UC process.
    In this NUC process, in addition to the latest folder, the integration time with which the NUC process was performed is also saved. Block 6 thus provides both the latest folder and the latest integration time.
    A block 7 handles the selection of the nearest reference image. Om itrefj - tpi <i trefg - fp | the reference image Brefr is selected because it is closest to the new current integration time tp. Otherwise, the reference image Breg is selected because it is then considered to be closest to the new current integration time tp.
    When the integration time is to be changed, the current integration time tr is compared, with the integration time trrrre from the latest NUC and as a reference image the reference closest to the new current integration time tr is selected, and is provided by block 7.
    In block 8, a compensation per pixel is calculated by linear interpolation between the selected reference image, either the reference image Brefr or Brefg, and the folder from the latest NUC. The result of the calculation is a folder in this description called the compensation folder. A block 9 performs standardization of the correction smap before it is added via an adder 10 to the incoming video output from block 2 to deliver a video output out through a block II. Video output can be connected to, for example, a display for display or a storage medium for storage. Display and memory medium are not shown in the figure. Alternatively, the normalization of the reference image can be performed before the interpolation and in Figure 2 indicated by dashed blocks 13 and 15 two possible locations of a block for normalizing the reference image Brefr. Correspondingly, two dashed blocks 14 and 15 indicate two possible locations of a block for normalizing the reference image Bref2- Figure 3 schematically illustrates the result of linear interpolation per pixel. Curve 12 shows the signal S per pixel as a function of integration time t. Along the time axis t, the integration times trefr and trefg for the reference images Brefr and Brefg have been entered as well as the integration time trrrre for the latest NUC. In addition, the current integration time tr, has been entered. In the example shown, the reference image Brefr is closer to the current integration time tr, than the reference image Breg. In analogy with what is described with reference to block 8 in Figure 2, a linear interpolation is performed between the reference image Brefr with the integration time trefr and the map from the latest NUC with the integration time tllllC 'The invention has been described above with reference to a schematic block structure shown in Figure 2. This description should only be seen as an explanation of the basic structure. The performance of the described functions can be done in many ways within the scope of the invention and in particular the use of a computer program for the implementation of the functions is pointed out.
    The invention is not limited to the methods described above by way of example, but may be subject to modifications within the scope of the appended claims.
    权利要求:
    Claims (7)
    [1]
    A method for mapping video recordings by means of a camera, such as an IR camera, characterized by a) that at least two reference images are taken in production against a black body radiator at the same temperature at two or fl different integration times. b) that when updating an offset folder is stored for which integration time the offset folder was updated, c) that during operation the current integration time is compared with the integration time for the last update of the offset folder, d) that the recorded reference image closest to the current integration time is selected, e) that a compensation per pixel for the integration time change is calculated by linear interpolation between the selected reference image and the most recently updated offset folder resulting in a compensation folder, f) adding the compensation folder to an incoming video image.
    [2]
    Method according to claim 1, characterized in that the compensation folder before addition is normalized by subtracting the average value of the compensation folder from all pixels for centering the compensation folder around zero.
    [3]
    Method according to claim 1, characterized in that the reference images are normalized before the interpolation by subtracting the average value of the reference image from all pixels for centering the reference image around zero.
    [4]
    Method according to one of the preceding claims, characterized in that two reference images are taken up in production against a blackbody radiator at the same temperature at two different integration times.
    [5]
    Method according to one of the preceding claims, characterized in that the reference images are recorded for integration times within the millisecond range with a difference in integration time of the order of 10 milliseconds.
    [6]
    Computer program comprising program code, which when said program code is executed in a computer causes said computer to perform the method according to any one of claims 1-4.
    [7]
    A computer program product comprising a computer readable medium and a computer program according to claim 6, wherein said computer program is included in said computer readable medium.
    类似技术:
    公开号 | 公开日 | 专利标题
    US9332186B2|2016-05-03|Method for compensating motion blur in connection with vibrations
    US10116918B2|2018-10-30|Disparity image generating device, disparity image generating method, and image
    CN103327245B|2016-08-10|A kind of Atomatic focusing method of infrared imaging system
    JP2014160213A5|2016-03-17|Microscope system, program, and method for correcting spherical aberration
    US20120044246A1|2012-02-23|Image Processing Device, Method, and Program
    JP2017129788A5|2019-03-07|
    JP2013038651A|2013-02-21|Image processing device, image processing method and imaging device
    SE1230112A1|2014-05-02|Procedure for mapping when recording video streams by camera
    US9466001B1|2016-10-11|Image processing apparatus and computer-readable storage medium
    Laixian et al.2013|Auto focusing algorithm based on largest gray gradient summation
    US11082631B2|2021-08-03|Image processing device
    SE1130099A1|2013-01-15|Procedure for Gain Folder Generation in an IR Camera, and IR Camera for Performing Gain Folder Generation
    JP2013197965A|2013-09-30|Imaging apparatus
    US20170091946A1|2017-03-30|Stereo image depth map generation device and method
    CN101124612B|2012-06-20|Image processing method, image processing system, image processing device
    JP6975003B2|2021-12-01|Peripheral monitoring device and its calibration method
    SE536252C2|2013-07-16|Method for setting integration time for an IR detector, and IR camera for performing the procedure
    WO2011146436A3|2012-03-01|Method and apparatus for auto-convergence based on auto focus point for stereoscopic frame
    US20170034437A1|2017-02-02|Image processing apparatus and method for operating image processing apparatus
    US20200034953A1|2020-01-30|Method and System for Representation of Vehicle Surroundings with Reduced Artefacts
    JP2013197964A|2013-09-30|Imaging apparatus
    JP2006236022A|2006-09-07|Image processor
    US10247551B2|2019-04-02|Vehicle image processing device for environment recognition
    JP2015118321A5|2017-01-19|Shake correction device, shake correction method and program, and imaging device
    JP5364531B2|2013-12-11|In-vehicle image recognition system
    同族专利:
    公开号 | 公开日
    WO2014070062A1|2014-05-08|
    US9648253B2|2017-05-09|
    US20150288891A1|2015-10-08|
    SE536679C2|2014-05-20|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    CA2110368C|1992-12-07|1999-11-23|Gary M. Lindgren|Wide dynamic range non-uniformity compensation for infrared focal plane arrays|
    US6897446B2|2002-03-26|2005-05-24|Lockheed Martin Corporation|Method and system for target detection using an infra-red sensor|
    FR2859279B1|2003-09-03|2005-11-25|Jobin Yvon Sas|DEVICE AND METHOD FOR SPECTROSCOPIC MEASUREMENT WITH AN IMAGING DEVICE COMPRISING A PHOTODETECTORS MATRIX|
    DE102007010649B8|2007-03-02|2009-01-22|Thermosensorik Gmbh|Method and apparatus for adaptively changing the integration time of an infrared sensor|
    SE532353C2|2008-06-19|2009-12-22|Flir Systems Ab|Device for imaging in the IR field|
    US8373757B1|2009-02-23|2013-02-12|Flir Systems, Inc.|Flat field correction for infrared cameras|
    FR2951895B1|2009-10-22|2012-07-13|Ulis|METHOD OF CORRECTING IMAGES DELIVERED BY A DETECTOR NOT REGULATED IN TEMPERATURE, AND DETECTOR USING SUCH A METHOD|
    US20120133776A1|2010-02-24|2012-05-31|Nova Research, Inc.|Low contrast midwave flir implementation|
    JP5672147B2|2011-05-24|2015-02-18|コニカミノルタ株式会社|Chest diagnosis support information generation system|US11198494B2|2018-11-01|2021-12-14|Brunswick Corporation|Methods and systems for controlling propulsion of a marine vessel to enhance proximity sensing in a marine environment|
    US11257378B2|2019-01-31|2022-02-22|Brunswick Corporation|Marine propulsion control system and method|
    法律状态:
    优先权:
    申请号 | 申请日 | 专利标题
    SE1230112A|SE536679C2|2012-11-01|2012-11-01|Procedure for mapping when recording video streams by camera|SE1230112A| SE536679C2|2012-11-01|2012-11-01|Procedure for mapping when recording video streams by camera|
    PCT/SE2013/000166| WO2014070062A1|2012-11-01|2013-10-30|Procedure for mapping when capturing video streams by means of a camera|
    US14/439,941| US9648253B2|2012-11-01|2013-10-30|Procedure for mapping when capturing video streams by means of a camera|
    [返回顶部]