点击次数：

论文类型：期刊论文

发表时间：2018-09-01

发表刊物：PATTERN RECOGNITION

收录刊物：SCIE

卷号：81

页面范围：471-483

ISSN号：0031-3203

关键字：Line matching; Characteristic number; Line-points projective invariant

摘要：Line matching across views is a fundamental task in many applications. Existing methods are hardly applicable to across view scenarios due to the limitation of line descriptors and matching strategy. In this paper, we present a novel line-points invariant based on a new projective invariant named characteristic number. The invariant is able to reflect the intrinsic geometry of line and points, which keeps invariant across views. The construction of this invariant uses intersections of coplanar lines instead of endpoints, rendering more robust matching across views. Accurate homography recovered from the invariant allows all lines, even those without interest points around them, a chance to be matched. Extensive comparisons with the state-of-the-art validate the matching accuracy and robustness of the proposed method to projective transformations. The method also performs well for image pairs with similar textures and those of low textures. (C) 2018 Elsevier Ltd. All rights reserved.