Feature point bidirectional matching and 3D reconstruction of sequence ISAR image based on SFIT and RANSAC method

In this letter, a matching and three-dimensional (3D) reconstruction method is presented for space target by using inverse synthetic aperture radar image sequences. Firstly, The ISAR images are preprocessed by morphology. In order to realize and improve the association of feature points of ISAR image sequences, the paper proposes a bidirectional matching approach to increase the number of matching-point, based on Scale Invariant Feature Transform (SIFT) and Random Sample Consensus (RANSAC) method. Finally, use the matrix factorization method to realize the 3D reconstruction. Experiment confirms the effectiveness of the proposal.