杨长久, 李双, 裘桢炜, 洪津, 乔延利. 同时偏振成像探测系统的偏振图像配准研究[J]. 红外与激光工程, 2013, 42(1): 262-267.
引用本文: 杨长久, 李双, 裘桢炜, 洪津, 乔延利. 同时偏振成像探测系统的偏振图像配准研究[J]. 红外与激光工程, 2013, 42(1): 262-267.
Yang Changjiu, Li Shuang, Qiu Zhenwei, Hong Jin, Qiao Yanli. Study on image registration of simultaneous imaging polarization system[J]. Infrared and Laser Engineering, 2013, 42(1): 262-267.
Citation: Yang Changjiu, Li Shuang, Qiu Zhenwei, Hong Jin, Qiao Yanli. Study on image registration of simultaneous imaging polarization system[J]. Infrared and Laser Engineering, 2013, 42(1): 262-267.

同时偏振成像探测系统的偏振图像配准研究

Study on image registration of simultaneous imaging polarization system

  • 摘要: 同时偏振成像探测技术是一种新型的偏振成像探测技术,它能在同一个探测器上同时获得被探测目标0、45、90、135 4个偏振方向的偏振强度图像。为了准确获取被探测目标的偏振信息,这4幅图像的配准显得尤为重要。为了使配准精度达到0.1个像元,结合同时偏振成像探测系统的成像方式,提出了一种基于空域和频域互相关的偏振图像配准方法。该方法使用硬件和软件相结合的方式完成图像的配准。首先使用空域互相关的算法实现图像的像元级粗配准;然后使用频域互相关的像元级配准算法进行像元级精配准;最后使用频域互相关的亚像元配准算法实现偏振图像的亚像元级配准。

     

    Abstract: Detection technology of imaging polarization has potential applications in target detection and remote sensing, with this technology the polarization state of light emitted from and/or reflected from scene can be measured. Simultaneous Imaging Polarization (SIP) is a new imaging polarization technology, which can avoid false polarization effects in time-sequential polarization detection system caused by the motion of the platform or target. The detection system can simultaneously obtain polarization intensity images of the same target among 0, 45, 90 and 135 on the same detector. Consequently, there is mis-registration among these four images, though the registration parameters are constant after the system is assembled. So, in order to obtain accurate polarization information of the detected target, registration accuracy of four images must reach 0.1 pixel. The polarization image registration method which was based on the spatial and frequency domain cross-correlation was presented. Firstly, pretreatment of these four images was prepared for image registration. Secondly, cross- correlation in spatial domain was used to achieve the pixel level of image registration. At last, cross-correlation in frequency-domain was used to achieve the sub-pixel level image registration. In order to validate the accuracy of this registration method, many experiments have been done and its registration accuracy can meet this SIP system's requirement.

     

/

返回文章
返回