基于ISAR图像序列的目标三维重构技术

Target 3D reconstruction based on ISAR image sequence

摘要:

从目标逆合成孔径雷达(ISAR Inverse Synthetic Aperture Radar)图像序列中获得目标的三维位置及结构信息,对目标识别与解译、空间目标监视等技术十分重要。通过对散射点特征提取与匹配及三维重构算法的研究,实现基于ISAR图像序列对目标的三维重构。首先利用LoFTR(Local Feature based Transformer)特征匹配算法提取目标强散射点关联位置信息,通过粗匹配与精匹配获得特征点二维坐标并生成观测矩阵。然后利用正交因式分解法计算目标强散射点三维位置信息,并通过重构结果融合实现较好重构效果。通过对序列ALOS卫星图像的处理,得到了目标强散射点的三维位置及形状的重构结果。结果表明,该方法能有有效的从ISAR二维图像信息中获得目标的三维空间信息。

Abstract:

Obtaining three-dimensional information of targets from inverse synthetic aperture radar image sequences is crucial. Reflected in target recognition, spatial monitoring, and other aspects. In this paper, the Feature matching algorithm based on LoFTR (Local Feature based Transformer) is used to extract the associated position information of the target strong scattering points. After obtaining the horizontal and vertical coordinate information of the target, adjust the image center uniformly to generate an observation matrix. Finally, the three-dimensional coordinate information of the target is obtained through factorization, and the shape and structure of the target are reconstructed. And achieving good reconstruction results through fusion of reconstruction results . By processing the continuous ALOS satellite image, the three-dimensional position and shape reconstruction results of the target's strong scattering points are obtained.The reconstruction results show that this method can reconstruct the three-dimensional position and shape information of the target from two-dimensional multi-angle ISAR images.