• DocumentCode
    1979799
  • Title

    Feature extraction for rescue target detection based on multi-spectral image analysis

  • Author

    Xin Ran ; Jinbiao Chen

  • Author_Institution
    Dept. of Navig., Shanghai Maritime Univ., Shanghai, China
  • fYear
    2015
  • fDate
    25-28 June 2015
  • Firstpage
    579
  • Lastpage
    582
  • Abstract
    In order to solve the problems with rescue target detection under complex sea environment, the research on object detection on the sea surface based on multi-spectral analysis and machine vision is presented in this paper. Firstly the difficulties and features of weak target detection and its complex sea background are discussed and the multi-spectral image acquisition and pre-processing method suitable for this task are proposed. Then the spectral features of targets and sea surface are extracted by analyzing visual light spectra and near-infrared spectra, which can distinguish the ships or survivors from islands or sea wave. The main purpose of this task is to detect lifeboats or survivors on the sea by analyzing the multi-spectral information, which can propose a target detection method based on image analysis and machine vision to find the survivor in the case that a marine casualty occurs, and then help the rescuer find the survivor more efficiently.
  • Keywords
    emergency management; feature extraction; object detection; feature extraction; machine vision; marine casualty; multispectral image acquisition; multispectral image analysis; multispectral image preprocessing method; rescue target detection; sea background; sea surface; Cameras; Feature extraction; Object detection; Reconstruction algorithms; Reflectivity; Sea surface; Surface treatment; feature extraction; maritime rescue; multi-spectral image; target detection;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Transportation Information and Safety (ICTIS), 2015 International Conference on
  • Conference_Location
    Wuhan
  • Print_ISBN
    978-1-4799-8693-4
  • Type

    conf

  • DOI
    10.1109/ICTIS.2015.7232204
  • Filename
    7232204