基于匹配点相似度引导采样的图像多平面检测.pdfVIP

优先出版 计 算 机 应 用 研 究 第32 卷 基于匹配点相似度引导采样的图像多平面检测 1 2 曹 芳 ，秦 川 (1．上海海事大学 信息工程学院，上海 200135 ；2 ．上海理工大学 光电信息与计算机工程学院，上海 200093) 摘 要：针对实际广泛存在的具有深度变化的多平面场景，提出一种基于匹配点引导采样的多平面检测算法。根据平面 结构中匹配点的相似度，改进匹配点的采样规则，通过对匹配点的残差信息进行排序和加权分析，获取场景中各个平面 结构对应的单应矩阵，从而实现了场景中多平面区域的准确检测。实验结果表明，与现有典型的算法相比，本文算法可 获得较高的平面检测准确率。 关键词：多平面检测；引导采样；匹配点；单应矩阵 中图分类号：TP391 文献标志码：A Multi-planar detection for digital images based on guided sampling of matching point similarity 1 2 FANG Cao , CHUAN Qin (1. College of Information Engineering, Shanghai Maritime University, Shanghai 200135, China; 2. School of Optical-Electrical Computer Engineering, University of Shanghaifor Science Technology, Shanghai 200093, China) Abstract: This paper proposes a multi-planar detection algorithm based on guided sampling of matching point similarity for the widely existing multi-planar scenes with changing depths. According to the similarity of matching points, the sampling rule is improved. By sorting and weighted analyzing for the residual information, the homographic matrix corresponding to each planar structure can be obtained, and all the planar structures can be detected successfully. Experimental results demonstrate that, compared with the existing typical algorithms, the proposed algorithm can achieve higher accuracy of the multi-planar detection. Key Words: multi-planar detection; guided sampling; matching point; homographic matrix