Abstract: Compared with point features, high-level features, especially free-form linear features, contain abundant information, which are more beneficial to the consequent image interpretation and surface reconstruction, etc. But, the linear features that can be matched successfully with typical method are few. Concerning the weakness of typical matching methods, perceptual organization theory was introduced into the matching process, and an image matching method based on perceptual organization of linear features in stereoscopic images was proposed. According to the results of the experiment, this algorithm advances the success ratio of free-form linear features matching and increases the length of free-form linear features, which facilitates the subsequent manual objects extraction.

Key words: Linear Feature, Image Matching, Perceptual Organization

摘要： 与点特征相比，高级别的特征，特别是自由形状线特征包含了更加丰富的信息，这对于其后的影像解译和表面重建等处理是更为有利的。针对传统自由形状线特征匹配方法的弊端——匹配成功的线特征数量较少，引入了感知编组理论，提出了一种基于双视线特征感知编组的影像匹配方法。实验证明该方法不仅可以提高成功匹配的自由形状线特征数量，还可增加自由形状线特征的长度，方便后续人工地物（如建筑物）的提取。

关键词: 线特征, 影像匹配, 感知编组