基于光电融合的目标三维结构反演

针对同传感器视角下的空天目标光电特征融合问题, 提出一种综合利用散射中心参数及轮廓角点特征的目标三维结构反演方法。首先, 基于正交匹配追踪(orthogonal matching pursuit, OMP)算法估计出序列雷达图像中的二维散射中心特征, 并利用Freeman链码提取连续可见光图像的角点特征。然后, 基于图论思想分别对雷达及可见光图像中的二维特征点进行关联。最后, 通过特征点三维重构来实现散射中心特征与光学角点特征之间的融合, 从而反演出较单源特征更为丰富准确的目标结构信息。仿真试验结果验证了所提方法的有效性。

Target 3-D structure inversion based on electro-optical fusion

Aiming at the problem of electro-optical feature fusion simultaneous interpreting the sky and sky targets from the same sensor angle, a three dimensional structure inversion method is proposed, which integrates the parameters of scattering centers and the features of contour points. Firstly, based on orthogonal matching pursuit (OMP) algorithm, the feature of two-dimensional scattering center in radar image sequence is estimated, and the corner feature of continuous visible image is extracted by Freeman chain code. Then, based on graph theory, two-dimensional feature points in radar and visible images are associated. Finally, the fusion of scattering center feature and optical corner feature is realized by three dimensional reconstruction of feature points, so as to reconstruct more abundant and accurate target structure information than single source feature. Simulation results verify the effectiveness of the proposed method.